Hospital Medicine

From The Ohio State University College of Medicine Department of Family and Community Medicine, Columbus, OH (Candy Magaña, Jná Báez, Christine Junk, Drs. Ahmad, Conroy, and Olayiwola); The Ohio State University College of Medicine Center for Primary Care Innovation and Transformation (Candy Magaña, Jná Báez, and Dr. Olayiwola); and The Ohio State University Wexner Medical Center (Christine Harsh, Erica Esposito).

Much has been discussed about the growing crisis of professional dissatisfaction among physicians, with increasing efforts being made to incorporate physician wellness into health system strategies that move from the Triple to the Quadruple Aim.¹ For many years, our health care system has been focused on improving the health of populations, optimizing the patient experience, and reducing the cost of care (Triple Aim). The inclusion of the fourth aim, improving the experience of the teams that deliver care, has become paramount in achieving the other aims.

An area often overlooked in this focus on wellness, however, is the importance of the earliest days of employment to shape and predict long-term career contentment. This is a missed opportunity, as data suggest that organizations with standardized onboarding programs boast a 62% increased productivity rate and a 50% greater retention rate among new hires.^2,3 Moreover, a study by the International Institute for Management Development found that businesses lose an estimated $37 billion annually because employees do not fully understand their jobs.⁴ The report ties losses to “actions taken by employees who have misunderstood or misinterpreted company policies, business processes, job function, or a combination of the three.” Additionally, onboarding programs that focus strictly on technical or functional orientation tasks miss important opportunities for culture integration during the onboarding process.⁵ It is therefore imperative to look to effective models of employee onboarding to develop systems that position physicians and practices for success.

Challenges With Traditional Physician Onboarding

In recent years, the Department of Family and Community Medicine at The Ohio State University College of Medicine has experienced rapid organizational change. Like many primary care systems nationwide responding to disruption in health care and changing demands on the clinical workforce, the department has hired new leadership, revised strategic priorities, and witnessed an influx of faculty and staff. It has also planned an expansion of ambulatory services that will more than double the clinical workforce over the next 3 years. While an exciting time, there has been a growing need to align strategy, culture, and human capital during these changes.

As we entered this phase of transformation, we recognized that our highly individualized, ad hoc orientation system presented shortcomings. During the act of revamping our physician recruitment process, stakeholder workgroup members specifically noted that improvement efforts were needed regarding new physician orientation, as no consistent structures were previously in place. New physician orientation had been a major gap for years, resulting in dissatisfaction in the first few months of physician practice, early physician turnover, and staff frustration. For physicians, we continued to learn about their frustration and unanswered questions regarding expectations, norms, structures, and processes.

Many new hires were left with a kind of “trial by fire” entry into their roles. On the first day of clinic, a new physician would most likely need to simultaneously see patients, learn the nuances of the electronic health record (EHR), figure out where the break room was located, and quickly learn population health issues for the patients they were serving. Opportunities to meet key clinic site leadership would be at random, and new physicians might not have the opportunity to meet leadership or staff until months into their tenure; this did not allow for a sense of belonging or understanding of the many resources available to them. We learned that the quality of these ad hoc orientations also varied based on the experience and priorities of each practice’s clinic and administrative leaders, who themselves felt ill-equipped to provide a consistent, robust, and confidence-building experience. In addition, practice site management was rarely given advance time to prepare for the arrival of new physicians, which resulted in physicians perceiving practices to be unwelcoming and disorganized. Their first days were often spent with patients in clinic with no structured orientation and without understanding workflows or having systems practice knowledge.

Institutionally, the interview process satisfied some transfer of knowledge, but we were unclear of what was being consistently shared and understood in the multiple ambulatory locations where our physicians enter practice. More importantly, we knew we were missing a critical opportunity to use orientation to imbue other values of diversity and inclusion, health equity, and operational excellence into the workforce. Based on anecdotal insights from employees and our own review of successful onboarding approaches from other industries, we also knew a more structured welcoming process would predict greater long-term career satisfaction for physicians and create a foundation for providing optimal care for patients when clinical encounters began.

Reengineering Physician Onboarding

In 2019, our department developed a multipronged approach to physician onboarding, which is already paying dividends in easing acculturation and fostering team cohesion. The department tapped its Center for Primary Care Innovation and Transformation (PCIT) to direct this effort, based on its expertise in practice transformation, clinical transformation and adaptations, and workflow efficiency through process and quality improvement. The PCIT team provides support to the department and the entire health system focused on technology and innovation, health equity, and health care efficiency.⁶ They applied many of the tools used in the Clinical Transformation in Technology approach to lead this initiative.⁷

The PCIT team began identifying key stakeholders (department, clinical and ambulatory leadership, clinicians and clinical staff, community partners, human resources, and resident physicians), and then engaging those individuals in dialogue surrounding orientation needs. During scheduled in-person and virtual work sessions, stakeholders were asked to provide input on pain points for new physicians and clinic leadership and were then empowered to create an onboarding program. Applying health care quality improvement techniques, we leveraged workflow mapping, current and future state planning, and goal setting, led by the skilled process improvement and clinical transformation specialists. We coordinated a multidisciplinary process improvement team that included clinic administrators, medical directors, human resources, administrative staff, ambulatory and resident leadership, clinical leadership, and recruitment liaisons. This diverse group of leadership and staff was brought together to address these critical identified gaps and weaknesses in new physician onboarding.

Through a series of learning sessions, the workgroup provided input that was used to form an itemized physician onboarding schedule, which was then leveraged to develop Plan-Do-Study-Act (PDSA) cycles, collecting feedback in real time. Some issues that seem small can cause major distress for new physicians. For example, in our inaugural orientation implementation, a physician provided feedback that they wanted to obtain information on setting up their work email on their personal devices and was having considerable trouble figuring out how to do so. This particular topic was not initially included in the first iteration of the Department’s orientation program. We rapidly sought out different ways to embed that into the onboarding experience. The first PDSA involved integrating the university information technology team (IT) into the process but was not successful because it required extra work for the new physician and reliance on the IT schedule. The next attempt was to have IT train a department staff member, but again, this still required that the physician find time to connect with that staff member. Finally, we decided to obtain a useful tip sheet that clearly outlined the process and could be included in orientation materials. This gave the new physicians control over how and when they would work on this issue. Based on these learnings, this was incorporated as a standing agenda item and resource for incoming physicians.

Essential Elements of Effective Onboarding

The new physician onboarding program consists of 5 key elements: (1) 2-week acclimation period; (2) peer learning and connection; (3) training before beginning patient care; (4) standardization, transparency, and accountability in all processes; (5) ongoing feedback for continued program improvement with individual support (Figure).

The program begins with a 2-week period of intentional investment in individual success, during which time no patients are scheduled. In week 1, we work with new hires to set expectations for performance, understand departmental norms, and introduce culture. Physicians meet formally and informally with department and institutional leadership, as well as attend team meetings and trainings that include a range of administrative and compliance requirements, such as quality standards and expectations, compliance, billing and coding specific to family medicine, EHR management, and institutionally mandated orientations. We are also adding implicit bias and antiracism training during this period, which are essential to creating a culture of unity and belonging.

During week 2, we focus on clinic-level orientation, assigning new hires an orientation buddy and a department sponsor, such as a physician lead or medical director. Physicians spend time with leadership at their clinic as they nurture relationships important for mentorship, sponsorship, and peer support. They also meet care team members, including front desk associates, medical assistants, behavioral health clinicians, nutritionists, social workers, pharmacists, and other key colleagues and care team members. This introduces the physician to the clinical environment and physical space as well as acclimates the physician to workflows and feedback loops for regular interaction.

When physicians ultimately begin patient care, they begin with an expected productivity rate of 50%, followed by an expected productivity rate of 75%, and then an expected productivity rate of 100%. This steady increase occurs over 3 to 4 weeks depending on the physician’s comfort level. They are also provided monthly reports on work relative value unit performance so that they can track and adapt practice patterns as necessary.More details on the program can be found in Appendix 1.

Takeaways From the Implementation of the New Program

Give time for new physicians to focus on acclimating to the role and environment.

The initial 2-week period of transition—without direct patient care—ensures that physicians feel comfortable in their new ecosystem. This also supports personal transitions, as many new hires are managing relocation and acclimating themselves and their families to new settings. Even residents from our training program who returned as attending physicians found this flexibility and slow reentry essential. This also gives the clinic time to orient to an additional provider, nurture them into the team culture, and develop relationships with the care team.

Cultivate spaces for shared learning, problem-solving, and peer connection.

Orientation is delivered primarily through group learning sessions with cohorts of new physicians, thus developing spaces for networking, fostering psychological safety, encouraging personal and professional rapport, emphasizing interactive learning, and reinforcing scheduling blocks at the departmental level. New hires also participate in peer shadowing to develop clinical competencies and are assigned a workplace buddy to foster a sense of belonging and create opportunities for additional knowledge sharing and cross-training.

Strengthen physician knowledge base, confidence, and comfort in the workplace before beginning direct patient care.

Without fluency in the workflows, culture, and operations of a practice, the urgency to have physicians begin clinical care can result in frustration for the physician, patients, and clinical and administrative staff. Therefore, we complete essential training prior to seeing any patients. This includes clinical workflows, referral processes, use of alternate modalities of care (eg, telehealth, eConsults), billing protocols, population health training, patient resources, office resources, and other essential daily processes and tools. This creates efficiency in administrative management, increased productivity, and better understanding of resources available for patients’ medical, social, and behavioral needs when patient care begins.

Embrace standardization, transparency, and accountability in as many processes as possible.

Standardized knowledge-sharing and checklists are mandated at every step of the orientation process, requiring sign off from the physician lead, practice manager, and new physicians upon completion. This offers all parties the opportunity to play a role in the delivery of and accountability for skills transfer and empowers new hires to press pause if they feel unsure about any domain in the training. It is also essential in guaranteeing that all physicians—regardless of which ambulatory location they practice in—receive consistent information and expectations. A sample checklist can be found in Appendix 2.

Commit to collecting and acting on feedback for continued program improvement and individual support.

As physicians complete the program, it is necessary to create structures to measure and enhance its impact, as well as evaluate how physicians are faring following the program. Each physician completes surveys at the end of the orientation program, attends a 90-day post-program check-in with the department chair, and receives follow-up trainings on advanced topics as they become more deeply embedded in the organization.

Lessons Learned

Feedback from surveys and 90-day check-ins with leadership and physicians reflect a high degree of clarity on job roles and duties, a sense of team camaraderie, easier system navigation, and a strong sense of support. We do recognize that sustaining change takes time and our study is limited by data demonstrating the impact of these efforts. We look forward to sharing more robust data from surveys and qualitative interviews with physicians, clinical leadership, and staff in the future. Our team will conduct interviews at 90-day and 180-day checkpoints with new physicians who have gone through this program, followed by a check-in after 1 year. Additionally, new physicians as well as key stakeholders, such as physician leads, practice managers, and members of the recruitment team, have started to participate in short surveys. These are designed to better understand their experiences, what worked well, what can be improved, and the overall satisfaction of the physician and other members of the extended care team.

What follows are some comments made by the initial group of physicians that went through this program and participated in follow-up interviews:

“I really feel like part of a bigger team.”

“I knew exactly what do to when I walked into the exam room on clinic Day 1.”

“It was great to make deep connections during the early process of joining.”

“Having a buddy to direct questions and ideas to is amazing and empowering.”

“Even though the orientation was long, I felt that I learned so much that I would not have otherwise.”

“Thank you for not letting me crash and burn!”

“Great culture! I love understanding our values of health equity, diversity, and inclusion.”

In the months since our endeavor began, we have learned just how essential it is to fully and effectively integrate new hires into the organization for their own satisfaction and success—and ours. Indeed, we cannot expect to achieve the Quadruple Aim without investing in the kind of transparent and intentional orientation process that defines expectations, aligns cultural values, mitigates costly and stressful operational misunderstandings, and communicates to physicians that, not only do they belong, but their sense of belonging is our priority. While we have yet to understand the impact of this program on the fourth aim of the Quadruple Aim, we are hopeful that the benefits will be far-reaching.

It is our ultimate hope that programs like this: (1) give physicians the confidence needed to create impactful patient-centered experiences; (2) enable physicians to become more cost-effective and efficient in care delivery; (3) allow physicians to understand the populations they are serving and access tools available to mitigate health disparities and other barriers; and (4) improve the collective experience of every member of the care team, practice leadership, and clinician-patient partnership.

Corresponding author: J. Nwando Olayiwola, MD, MPH, FAAFP, The Ohio State University College of Medicine, Department of Family and Community Medicine, 2231 N High St, Ste 250, Columbus, OH 43210; [email protected].

Financial disclosures: None.

Keywords: physician onboarding; Quadruple Aim; leadership; clinician satisfaction; care team satisfaction.

From The Ohio State University College of Medicine Department of Family and Community Medicine, Columbus, OH (Candy Magaña, Jná Báez, Christine Junk, Drs. Ahmad, Conroy, and Olayiwola); The Ohio State University College of Medicine Center for Primary Care Innovation and Transformation (Candy Magaña, Jná Báez, and Dr. Olayiwola); and The Ohio State University Wexner Medical Center (Christine Harsh, Erica Esposito).

Much has been discussed about the growing crisis of professional dissatisfaction among physicians, with increasing efforts being made to incorporate physician wellness into health system strategies that move from the Triple to the Quadruple Aim.¹ For many years, our health care system has been focused on improving the health of populations, optimizing the patient experience, and reducing the cost of care (Triple Aim). The inclusion of the fourth aim, improving the experience of the teams that deliver care, has become paramount in achieving the other aims.

An area often overlooked in this focus on wellness, however, is the importance of the earliest days of employment to shape and predict long-term career contentment. This is a missed opportunity, as data suggest that organizations with standardized onboarding programs boast a 62% increased productivity rate and a 50% greater retention rate among new hires.^2,3 Moreover, a study by the International Institute for Management Development found that businesses lose an estimated $37 billion annually because employees do not fully understand their jobs.⁴ The report ties losses to “actions taken by employees who have misunderstood or misinterpreted company policies, business processes, job function, or a combination of the three.” Additionally, onboarding programs that focus strictly on technical or functional orientation tasks miss important opportunities for culture integration during the onboarding process.⁵ It is therefore imperative to look to effective models of employee onboarding to develop systems that position physicians and practices for success.

Challenges With Traditional Physician Onboarding

In recent years, the Department of Family and Community Medicine at The Ohio State University College of Medicine has experienced rapid organizational change. Like many primary care systems nationwide responding to disruption in health care and changing demands on the clinical workforce, the department has hired new leadership, revised strategic priorities, and witnessed an influx of faculty and staff. It has also planned an expansion of ambulatory services that will more than double the clinical workforce over the next 3 years. While an exciting time, there has been a growing need to align strategy, culture, and human capital during these changes.

As we entered this phase of transformation, we recognized that our highly individualized, ad hoc orientation system presented shortcomings. During the act of revamping our physician recruitment process, stakeholder workgroup members specifically noted that improvement efforts were needed regarding new physician orientation, as no consistent structures were previously in place. New physician orientation had been a major gap for years, resulting in dissatisfaction in the first few months of physician practice, early physician turnover, and staff frustration. For physicians, we continued to learn about their frustration and unanswered questions regarding expectations, norms, structures, and processes.

Many new hires were left with a kind of “trial by fire” entry into their roles. On the first day of clinic, a new physician would most likely need to simultaneously see patients, learn the nuances of the electronic health record (EHR), figure out where the break room was located, and quickly learn population health issues for the patients they were serving. Opportunities to meet key clinic site leadership would be at random, and new physicians might not have the opportunity to meet leadership or staff until months into their tenure; this did not allow for a sense of belonging or understanding of the many resources available to them. We learned that the quality of these ad hoc orientations also varied based on the experience and priorities of each practice’s clinic and administrative leaders, who themselves felt ill-equipped to provide a consistent, robust, and confidence-building experience. In addition, practice site management was rarely given advance time to prepare for the arrival of new physicians, which resulted in physicians perceiving practices to be unwelcoming and disorganized. Their first days were often spent with patients in clinic with no structured orientation and without understanding workflows or having systems practice knowledge.

Institutionally, the interview process satisfied some transfer of knowledge, but we were unclear of what was being consistently shared and understood in the multiple ambulatory locations where our physicians enter practice. More importantly, we knew we were missing a critical opportunity to use orientation to imbue other values of diversity and inclusion, health equity, and operational excellence into the workforce. Based on anecdotal insights from employees and our own review of successful onboarding approaches from other industries, we also knew a more structured welcoming process would predict greater long-term career satisfaction for physicians and create a foundation for providing optimal care for patients when clinical encounters began.

Reengineering Physician Onboarding

In 2019, our department developed a multipronged approach to physician onboarding, which is already paying dividends in easing acculturation and fostering team cohesion. The department tapped its Center for Primary Care Innovation and Transformation (PCIT) to direct this effort, based on its expertise in practice transformation, clinical transformation and adaptations, and workflow efficiency through process and quality improvement. The PCIT team provides support to the department and the entire health system focused on technology and innovation, health equity, and health care efficiency.⁶ They applied many of the tools used in the Clinical Transformation in Technology approach to lead this initiative.⁷

The PCIT team began identifying key stakeholders (department, clinical and ambulatory leadership, clinicians and clinical staff, community partners, human resources, and resident physicians), and then engaging those individuals in dialogue surrounding orientation needs. During scheduled in-person and virtual work sessions, stakeholders were asked to provide input on pain points for new physicians and clinic leadership and were then empowered to create an onboarding program. Applying health care quality improvement techniques, we leveraged workflow mapping, current and future state planning, and goal setting, led by the skilled process improvement and clinical transformation specialists. We coordinated a multidisciplinary process improvement team that included clinic administrators, medical directors, human resources, administrative staff, ambulatory and resident leadership, clinical leadership, and recruitment liaisons. This diverse group of leadership and staff was brought together to address these critical identified gaps and weaknesses in new physician onboarding.

Through a series of learning sessions, the workgroup provided input that was used to form an itemized physician onboarding schedule, which was then leveraged to develop Plan-Do-Study-Act (PDSA) cycles, collecting feedback in real time. Some issues that seem small can cause major distress for new physicians. For example, in our inaugural orientation implementation, a physician provided feedback that they wanted to obtain information on setting up their work email on their personal devices and was having considerable trouble figuring out how to do so. This particular topic was not initially included in the first iteration of the Department’s orientation program. We rapidly sought out different ways to embed that into the onboarding experience. The first PDSA involved integrating the university information technology team (IT) into the process but was not successful because it required extra work for the new physician and reliance on the IT schedule. The next attempt was to have IT train a department staff member, but again, this still required that the physician find time to connect with that staff member. Finally, we decided to obtain a useful tip sheet that clearly outlined the process and could be included in orientation materials. This gave the new physicians control over how and when they would work on this issue. Based on these learnings, this was incorporated as a standing agenda item and resource for incoming physicians.

Essential Elements of Effective Onboarding

The new physician onboarding program consists of 5 key elements: (1) 2-week acclimation period; (2) peer learning and connection; (3) training before beginning patient care; (4) standardization, transparency, and accountability in all processes; (5) ongoing feedback for continued program improvement with individual support (Figure).

The program begins with a 2-week period of intentional investment in individual success, during which time no patients are scheduled. In week 1, we work with new hires to set expectations for performance, understand departmental norms, and introduce culture. Physicians meet formally and informally with department and institutional leadership, as well as attend team meetings and trainings that include a range of administrative and compliance requirements, such as quality standards and expectations, compliance, billing and coding specific to family medicine, EHR management, and institutionally mandated orientations. We are also adding implicit bias and antiracism training during this period, which are essential to creating a culture of unity and belonging.

During week 2, we focus on clinic-level orientation, assigning new hires an orientation buddy and a department sponsor, such as a physician lead or medical director. Physicians spend time with leadership at their clinic as they nurture relationships important for mentorship, sponsorship, and peer support. They also meet care team members, including front desk associates, medical assistants, behavioral health clinicians, nutritionists, social workers, pharmacists, and other key colleagues and care team members. This introduces the physician to the clinical environment and physical space as well as acclimates the physician to workflows and feedback loops for regular interaction.

When physicians ultimately begin patient care, they begin with an expected productivity rate of 50%, followed by an expected productivity rate of 75%, and then an expected productivity rate of 100%. This steady increase occurs over 3 to 4 weeks depending on the physician’s comfort level. They are also provided monthly reports on work relative value unit performance so that they can track and adapt practice patterns as necessary.More details on the program can be found in Appendix 1.

Takeaways From the Implementation of the New Program

Give time for new physicians to focus on acclimating to the role and environment.

The initial 2-week period of transition—without direct patient care—ensures that physicians feel comfortable in their new ecosystem. This also supports personal transitions, as many new hires are managing relocation and acclimating themselves and their families to new settings. Even residents from our training program who returned as attending physicians found this flexibility and slow reentry essential. This also gives the clinic time to orient to an additional provider, nurture them into the team culture, and develop relationships with the care team.

Cultivate spaces for shared learning, problem-solving, and peer connection.

Orientation is delivered primarily through group learning sessions with cohorts of new physicians, thus developing spaces for networking, fostering psychological safety, encouraging personal and professional rapport, emphasizing interactive learning, and reinforcing scheduling blocks at the departmental level. New hires also participate in peer shadowing to develop clinical competencies and are assigned a workplace buddy to foster a sense of belonging and create opportunities for additional knowledge sharing and cross-training.

Strengthen physician knowledge base, confidence, and comfort in the workplace before beginning direct patient care.

Without fluency in the workflows, culture, and operations of a practice, the urgency to have physicians begin clinical care can result in frustration for the physician, patients, and clinical and administrative staff. Therefore, we complete essential training prior to seeing any patients. This includes clinical workflows, referral processes, use of alternate modalities of care (eg, telehealth, eConsults), billing protocols, population health training, patient resources, office resources, and other essential daily processes and tools. This creates efficiency in administrative management, increased productivity, and better understanding of resources available for patients’ medical, social, and behavioral needs when patient care begins.

Embrace standardization, transparency, and accountability in as many processes as possible.

Standardized knowledge-sharing and checklists are mandated at every step of the orientation process, requiring sign off from the physician lead, practice manager, and new physicians upon completion. This offers all parties the opportunity to play a role in the delivery of and accountability for skills transfer and empowers new hires to press pause if they feel unsure about any domain in the training. It is also essential in guaranteeing that all physicians—regardless of which ambulatory location they practice in—receive consistent information and expectations. A sample checklist can be found in Appendix 2.

Commit to collecting and acting on feedback for continued program improvement and individual support.

As physicians complete the program, it is necessary to create structures to measure and enhance its impact, as well as evaluate how physicians are faring following the program. Each physician completes surveys at the end of the orientation program, attends a 90-day post-program check-in with the department chair, and receives follow-up trainings on advanced topics as they become more deeply embedded in the organization.

Lessons Learned

Feedback from surveys and 90-day check-ins with leadership and physicians reflect a high degree of clarity on job roles and duties, a sense of team camaraderie, easier system navigation, and a strong sense of support. We do recognize that sustaining change takes time and our study is limited by data demonstrating the impact of these efforts. We look forward to sharing more robust data from surveys and qualitative interviews with physicians, clinical leadership, and staff in the future. Our team will conduct interviews at 90-day and 180-day checkpoints with new physicians who have gone through this program, followed by a check-in after 1 year. Additionally, new physicians as well as key stakeholders, such as physician leads, practice managers, and members of the recruitment team, have started to participate in short surveys. These are designed to better understand their experiences, what worked well, what can be improved, and the overall satisfaction of the physician and other members of the extended care team.

What follows are some comments made by the initial group of physicians that went through this program and participated in follow-up interviews:

“I really feel like part of a bigger team.”

“I knew exactly what do to when I walked into the exam room on clinic Day 1.”

“It was great to make deep connections during the early process of joining.”

“Having a buddy to direct questions and ideas to is amazing and empowering.”

“Even though the orientation was long, I felt that I learned so much that I would not have otherwise.”

“Thank you for not letting me crash and burn!”

“Great culture! I love understanding our values of health equity, diversity, and inclusion.”

In the months since our endeavor began, we have learned just how essential it is to fully and effectively integrate new hires into the organization for their own satisfaction and success—and ours. Indeed, we cannot expect to achieve the Quadruple Aim without investing in the kind of transparent and intentional orientation process that defines expectations, aligns cultural values, mitigates costly and stressful operational misunderstandings, and communicates to physicians that, not only do they belong, but their sense of belonging is our priority. While we have yet to understand the impact of this program on the fourth aim of the Quadruple Aim, we are hopeful that the benefits will be far-reaching.

It is our ultimate hope that programs like this: (1) give physicians the confidence needed to create impactful patient-centered experiences; (2) enable physicians to become more cost-effective and efficient in care delivery; (3) allow physicians to understand the populations they are serving and access tools available to mitigate health disparities and other barriers; and (4) improve the collective experience of every member of the care team, practice leadership, and clinician-patient partnership.

Corresponding author: J. Nwando Olayiwola, MD, MPH, FAAFP, The Ohio State University College of Medicine, Department of Family and Community Medicine, 2231 N High St, Ste 250, Columbus, OH 43210; [email protected].

Financial disclosures: None.

Keywords: physician onboarding; Quadruple Aim; leadership; clinician satisfaction; care team satisfaction.

From The Ohio State University College of Medicine Department of Family and Community Medicine, Columbus, OH (Candy Magaña, Jná Báez, Christine Junk, Drs. Ahmad, Conroy, and Olayiwola); The Ohio State University College of Medicine Center for Primary Care Innovation and Transformation (Candy Magaña, Jná Báez, and Dr. Olayiwola); and The Ohio State University Wexner Medical Center (Christine Harsh, Erica Esposito).

Much has been discussed about the growing crisis of professional dissatisfaction among physicians, with increasing efforts being made to incorporate physician wellness into health system strategies that move from the Triple to the Quadruple Aim.¹ For many years, our health care system has been focused on improving the health of populations, optimizing the patient experience, and reducing the cost of care (Triple Aim). The inclusion of the fourth aim, improving the experience of the teams that deliver care, has become paramount in achieving the other aims.

An area often overlooked in this focus on wellness, however, is the importance of the earliest days of employment to shape and predict long-term career contentment. This is a missed opportunity, as data suggest that organizations with standardized onboarding programs boast a 62% increased productivity rate and a 50% greater retention rate among new hires.^2,3 Moreover, a study by the International Institute for Management Development found that businesses lose an estimated $37 billion annually because employees do not fully understand their jobs.⁴ The report ties losses to “actions taken by employees who have misunderstood or misinterpreted company policies, business processes, job function, or a combination of the three.” Additionally, onboarding programs that focus strictly on technical or functional orientation tasks miss important opportunities for culture integration during the onboarding process.⁵ It is therefore imperative to look to effective models of employee onboarding to develop systems that position physicians and practices for success.

Challenges With Traditional Physician Onboarding

In recent years, the Department of Family and Community Medicine at The Ohio State University College of Medicine has experienced rapid organizational change. Like many primary care systems nationwide responding to disruption in health care and changing demands on the clinical workforce, the department has hired new leadership, revised strategic priorities, and witnessed an influx of faculty and staff. It has also planned an expansion of ambulatory services that will more than double the clinical workforce over the next 3 years. While an exciting time, there has been a growing need to align strategy, culture, and human capital during these changes.

As we entered this phase of transformation, we recognized that our highly individualized, ad hoc orientation system presented shortcomings. During the act of revamping our physician recruitment process, stakeholder workgroup members specifically noted that improvement efforts were needed regarding new physician orientation, as no consistent structures were previously in place. New physician orientation had been a major gap for years, resulting in dissatisfaction in the first few months of physician practice, early physician turnover, and staff frustration. For physicians, we continued to learn about their frustration and unanswered questions regarding expectations, norms, structures, and processes.

Many new hires were left with a kind of “trial by fire” entry into their roles. On the first day of clinic, a new physician would most likely need to simultaneously see patients, learn the nuances of the electronic health record (EHR), figure out where the break room was located, and quickly learn population health issues for the patients they were serving. Opportunities to meet key clinic site leadership would be at random, and new physicians might not have the opportunity to meet leadership or staff until months into their tenure; this did not allow for a sense of belonging or understanding of the many resources available to them. We learned that the quality of these ad hoc orientations also varied based on the experience and priorities of each practice’s clinic and administrative leaders, who themselves felt ill-equipped to provide a consistent, robust, and confidence-building experience. In addition, practice site management was rarely given advance time to prepare for the arrival of new physicians, which resulted in physicians perceiving practices to be unwelcoming and disorganized. Their first days were often spent with patients in clinic with no structured orientation and without understanding workflows or having systems practice knowledge.

Institutionally, the interview process satisfied some transfer of knowledge, but we were unclear of what was being consistently shared and understood in the multiple ambulatory locations where our physicians enter practice. More importantly, we knew we were missing a critical opportunity to use orientation to imbue other values of diversity and inclusion, health equity, and operational excellence into the workforce. Based on anecdotal insights from employees and our own review of successful onboarding approaches from other industries, we also knew a more structured welcoming process would predict greater long-term career satisfaction for physicians and create a foundation for providing optimal care for patients when clinical encounters began.

Reengineering Physician Onboarding

In 2019, our department developed a multipronged approach to physician onboarding, which is already paying dividends in easing acculturation and fostering team cohesion. The department tapped its Center for Primary Care Innovation and Transformation (PCIT) to direct this effort, based on its expertise in practice transformation, clinical transformation and adaptations, and workflow efficiency through process and quality improvement. The PCIT team provides support to the department and the entire health system focused on technology and innovation, health equity, and health care efficiency.⁶ They applied many of the tools used in the Clinical Transformation in Technology approach to lead this initiative.⁷

The PCIT team began identifying key stakeholders (department, clinical and ambulatory leadership, clinicians and clinical staff, community partners, human resources, and resident physicians), and then engaging those individuals in dialogue surrounding orientation needs. During scheduled in-person and virtual work sessions, stakeholders were asked to provide input on pain points for new physicians and clinic leadership and were then empowered to create an onboarding program. Applying health care quality improvement techniques, we leveraged workflow mapping, current and future state planning, and goal setting, led by the skilled process improvement and clinical transformation specialists. We coordinated a multidisciplinary process improvement team that included clinic administrators, medical directors, human resources, administrative staff, ambulatory and resident leadership, clinical leadership, and recruitment liaisons. This diverse group of leadership and staff was brought together to address these critical identified gaps and weaknesses in new physician onboarding.

Through a series of learning sessions, the workgroup provided input that was used to form an itemized physician onboarding schedule, which was then leveraged to develop Plan-Do-Study-Act (PDSA) cycles, collecting feedback in real time. Some issues that seem small can cause major distress for new physicians. For example, in our inaugural orientation implementation, a physician provided feedback that they wanted to obtain information on setting up their work email on their personal devices and was having considerable trouble figuring out how to do so. This particular topic was not initially included in the first iteration of the Department’s orientation program. We rapidly sought out different ways to embed that into the onboarding experience. The first PDSA involved integrating the university information technology team (IT) into the process but was not successful because it required extra work for the new physician and reliance on the IT schedule. The next attempt was to have IT train a department staff member, but again, this still required that the physician find time to connect with that staff member. Finally, we decided to obtain a useful tip sheet that clearly outlined the process and could be included in orientation materials. This gave the new physicians control over how and when they would work on this issue. Based on these learnings, this was incorporated as a standing agenda item and resource for incoming physicians.

Essential Elements of Effective Onboarding

The new physician onboarding program consists of 5 key elements: (1) 2-week acclimation period; (2) peer learning and connection; (3) training before beginning patient care; (4) standardization, transparency, and accountability in all processes; (5) ongoing feedback for continued program improvement with individual support (Figure).

The program begins with a 2-week period of intentional investment in individual success, during which time no patients are scheduled. In week 1, we work with new hires to set expectations for performance, understand departmental norms, and introduce culture. Physicians meet formally and informally with department and institutional leadership, as well as attend team meetings and trainings that include a range of administrative and compliance requirements, such as quality standards and expectations, compliance, billing and coding specific to family medicine, EHR management, and institutionally mandated orientations. We are also adding implicit bias and antiracism training during this period, which are essential to creating a culture of unity and belonging.

During week 2, we focus on clinic-level orientation, assigning new hires an orientation buddy and a department sponsor, such as a physician lead or medical director. Physicians spend time with leadership at their clinic as they nurture relationships important for mentorship, sponsorship, and peer support. They also meet care team members, including front desk associates, medical assistants, behavioral health clinicians, nutritionists, social workers, pharmacists, and other key colleagues and care team members. This introduces the physician to the clinical environment and physical space as well as acclimates the physician to workflows and feedback loops for regular interaction.

When physicians ultimately begin patient care, they begin with an expected productivity rate of 50%, followed by an expected productivity rate of 75%, and then an expected productivity rate of 100%. This steady increase occurs over 3 to 4 weeks depending on the physician’s comfort level. They are also provided monthly reports on work relative value unit performance so that they can track and adapt practice patterns as necessary.More details on the program can be found in Appendix 1.

Takeaways From the Implementation of the New Program

Give time for new physicians to focus on acclimating to the role and environment.

The initial 2-week period of transition—without direct patient care—ensures that physicians feel comfortable in their new ecosystem. This also supports personal transitions, as many new hires are managing relocation and acclimating themselves and their families to new settings. Even residents from our training program who returned as attending physicians found this flexibility and slow reentry essential. This also gives the clinic time to orient to an additional provider, nurture them into the team culture, and develop relationships with the care team.

Cultivate spaces for shared learning, problem-solving, and peer connection.

Orientation is delivered primarily through group learning sessions with cohorts of new physicians, thus developing spaces for networking, fostering psychological safety, encouraging personal and professional rapport, emphasizing interactive learning, and reinforcing scheduling blocks at the departmental level. New hires also participate in peer shadowing to develop clinical competencies and are assigned a workplace buddy to foster a sense of belonging and create opportunities for additional knowledge sharing and cross-training.

Strengthen physician knowledge base, confidence, and comfort in the workplace before beginning direct patient care.

Without fluency in the workflows, culture, and operations of a practice, the urgency to have physicians begin clinical care can result in frustration for the physician, patients, and clinical and administrative staff. Therefore, we complete essential training prior to seeing any patients. This includes clinical workflows, referral processes, use of alternate modalities of care (eg, telehealth, eConsults), billing protocols, population health training, patient resources, office resources, and other essential daily processes and tools. This creates efficiency in administrative management, increased productivity, and better understanding of resources available for patients’ medical, social, and behavioral needs when patient care begins.

Embrace standardization, transparency, and accountability in as many processes as possible.

Standardized knowledge-sharing and checklists are mandated at every step of the orientation process, requiring sign off from the physician lead, practice manager, and new physicians upon completion. This offers all parties the opportunity to play a role in the delivery of and accountability for skills transfer and empowers new hires to press pause if they feel unsure about any domain in the training. It is also essential in guaranteeing that all physicians—regardless of which ambulatory location they practice in—receive consistent information and expectations. A sample checklist can be found in Appendix 2.

Commit to collecting and acting on feedback for continued program improvement and individual support.

As physicians complete the program, it is necessary to create structures to measure and enhance its impact, as well as evaluate how physicians are faring following the program. Each physician completes surveys at the end of the orientation program, attends a 90-day post-program check-in with the department chair, and receives follow-up trainings on advanced topics as they become more deeply embedded in the organization.

Lessons Learned

Feedback from surveys and 90-day check-ins with leadership and physicians reflect a high degree of clarity on job roles and duties, a sense of team camaraderie, easier system navigation, and a strong sense of support. We do recognize that sustaining change takes time and our study is limited by data demonstrating the impact of these efforts. We look forward to sharing more robust data from surveys and qualitative interviews with physicians, clinical leadership, and staff in the future. Our team will conduct interviews at 90-day and 180-day checkpoints with new physicians who have gone through this program, followed by a check-in after 1 year. Additionally, new physicians as well as key stakeholders, such as physician leads, practice managers, and members of the recruitment team, have started to participate in short surveys. These are designed to better understand their experiences, what worked well, what can be improved, and the overall satisfaction of the physician and other members of the extended care team.

What follows are some comments made by the initial group of physicians that went through this program and participated in follow-up interviews:

“I really feel like part of a bigger team.”

“I knew exactly what do to when I walked into the exam room on clinic Day 1.”

“It was great to make deep connections during the early process of joining.”

“Having a buddy to direct questions and ideas to is amazing and empowering.”

“Even though the orientation was long, I felt that I learned so much that I would not have otherwise.”

“Thank you for not letting me crash and burn!”

“Great culture! I love understanding our values of health equity, diversity, and inclusion.”

In the months since our endeavor began, we have learned just how essential it is to fully and effectively integrate new hires into the organization for their own satisfaction and success—and ours. Indeed, we cannot expect to achieve the Quadruple Aim without investing in the kind of transparent and intentional orientation process that defines expectations, aligns cultural values, mitigates costly and stressful operational misunderstandings, and communicates to physicians that, not only do they belong, but their sense of belonging is our priority. While we have yet to understand the impact of this program on the fourth aim of the Quadruple Aim, we are hopeful that the benefits will be far-reaching.

It is our ultimate hope that programs like this: (1) give physicians the confidence needed to create impactful patient-centered experiences; (2) enable physicians to become more cost-effective and efficient in care delivery; (3) allow physicians to understand the populations they are serving and access tools available to mitigate health disparities and other barriers; and (4) improve the collective experience of every member of the care team, practice leadership, and clinician-patient partnership.

Corresponding author: J. Nwando Olayiwola, MD, MPH, FAAFP, The Ohio State University College of Medicine, Department of Family and Community Medicine, 2231 N High St, Ste 250, Columbus, OH 43210; [email protected].

Financial disclosures: None.

Keywords: physician onboarding; Quadruple Aim; leadership; clinician satisfaction; care team satisfaction.

From Adelante Healthcare, Mesa, AZ (Dr. Chin), University Hospitals of Cleveland, Cleveland, OH (Drs. Delozier, Bascug, Levine, Bejanishvili, and Wynbrandt and Janet C. Peachey, Rachel M. Cerminara, and Sharon M. Darkovich), and Houston Methodist Hospitals, Houston, TX (Dr. Bhakta).

Abstract

Background: Resident physicians play an active role in the reporting of errors that occur in patient care. Previous studies indicate that residents significantly underreport errors in patient care.

Methods: Fifty-four of 80 eligible residents enrolled at University Hospitals–Regional Hospitals (UH-RH) during the 2018-2019 academic year completed a survey assessing their knowledge and experience in completing Patient Advocacy and Shared Stories (PASS) reports, which serve as incident reports in the UH health system in reporting errors in patient care. A series of interventions aimed at educating residents about the PASS report system were then conducted. The 54 residents who completed the first survey received it again 4 months later.

Results: Residents demonstrated greater understanding of when filing PASS reports was appropriate after the intervention, as significantly more residents reported having been involved in a situation where they should have filed a PASS report but did not (P = 0.036).

Conclusion: In this study, residents often did not report errors in patient care because they simply did not know the process for doing so. In addition, many residents often felt that the reporting of patient errors could be used as a form of retaliation.

Keywords: resident physicians; quality improvement; high-value care; medical errors; patient safety.

Resident physicians play a critical role in patient care. Residents undergo extensive supervised training in order to one day be able to practice medicine in an unsupervised setting, with the goal of providing the highest quality of care possible. One study reported that primary care provided by residents in a training program is of similar or higher quality than that provided by attending physicians.¹

Besides providing high-quality care, it is important that residents play an active role in the reporting of errors that occur regarding patient care as well as in identifying events that may compromise patient safety and quality.² In fact, increased reporting of patient errors has been shown to decrease liability-related costs for hospitals.³ Unfortunately, physicians, and residents in particular, have historically been poor reporters of errors in patient care.⁴ This is especially true when comparing physicians to other health professionals, such as nurses, in error reporting.⁵

Several studies have examined the involvement of residents in reporting errors in patient care. One recent study showed that a graduate medical education financial incentive program significantly increased the number of patient safety events reported by residents and fellows.⁶ This study, along with several others, supports the concept of using incentives to help improve the reporting of errors in patient care for physicians in training.^7-10 Another study used Quality Improvement Knowledge Assessment Tool (QIKAT) scores to assess quality improvement (QI) knowledge. The study demonstrated that self-assessment scores of QI skills using QIKAT scores improved following a targeted intervention.¹¹ Because further information on the involvement and attitudes of residents in reporting errors in patient care is needed, University Hospitals of Cleveland (UH) designed and implemented a QI study during the 2018-2019 academic year. This prospective study used anonymous surveys to objectively examine the involvement and attitudes of residents in reporting errors in patient care.

Methods

The UH health system uses Patient Advocacy and Shared Stories (PASS) reports as incident reports to not only disclose errors in patient care but also to identify any events that may compromise patient safety and quality. Based on preliminary review, nurses, ancillary staff, and administrators file the majority of PASS reports.

The study group consisted of residents at University Hospitals–Regional Hospitals (UH-RH), which is comprised of 2 hospitals: University Hospitals–Richmond Medical Center (UH-RMC) and University Hospitals –Bedford Medical Center (UH-BMC). UH-RMC and UH-BMC are 2 medium-sized university-affiliated community hospitals located in the Cleveland metropolitan area in Northeast Ohio. Both serve as clinical training sites for Case Western Reserve University School of Medicine and Lake Erie College of Osteopathic Medicine, the latter of which helped fund this study. The study was submitted to the Institutional Review Board (IRB) of University Hospitals of Cleveland and granted “not human subjects research” status as a QI study.

Surveys

UH-RH offers residency programs in dermatology, emergency medicine, family medicine, internal medicine, orthopedic surgery, and physical medicine and rehabilitation, along with a 1-year transitional/preliminary year. A total of 80 residents enrolled at UH-RH during the 2018-2019 academic year. All 80 residents at UH-RH received an email in December 2018 asking them to complete an anonymous survey regarding the PASS report system. The survey was administered using the REDCap software system and consisted of 15 multiple-choice questions. As an incentive for completing the survey, residents were offered a $10 Amazon gift card. The gift cards were funded through a research grant from Lake Erie College of Osteopathic Medicine. Residents were given 1 week to complete the survey. At the end of the week, 54 of 80 residents completed the first survey.

Following the first survey, efforts were undertaken by the study authors, in conjunction with the quality improvement department at UH-RH, to educate residents about the PASS report system. These interventions included giving a lecture on the PASS report system during resident didactic sessions, sending an email to all residents about the PASS report system, and providing residents an opportunity to complete an optional online training course regarding the PASS report system. As an incentive for completing the online training course, residents were offered a $10 Amazon gift card. As before, the gift cards were funded through a research grant from Lake Erie College of Osteopathic Medicine.

A second survey was administered in April 2019, 4 months after the first survey. To determine whether the intervention made an impact on the involvement and attitudes of residents in the reporting errors in patient care, only residents who completed the first survey were sent the second survey. The second survey consisted of the same questions as the first survey and was also administered using the REDCap software system. As an incentive for completing the survey, residents were offered another $10 Amazon gift card, again were funded through a research grant from Lake Erie College of Osteopathic Medicine. Residents were given 1 week to complete the survey.

Analysis

Chi-square analyses were utilized to examine differences between preintervention and postintervention responses across categories. All analyses were conducted using R statistical software, version 3.6.1 (R Foundation for Statistical Computing).

Results

A total of 54 of 80 eligible residents responded to the first survey (Table). Twenty-nine of 54 eligible residents responded to the second survey. Postintervention, significantly more residents indicated being involved in a situation where they should have filed a PASS report but did not (58.6% vs 53.7%; P = 0.036). Improvement was seen in PASS knowledge postintervention, where fewer residents reported not knowing how to file a PASS report (31.5% vs 55.2%; P = 0.059). No other improvements were significant, nor were there significant differences in responses between any other categories pre- and postintervention.

Discussion

Errors in patient care are a common occurrence in the hospital setting. Reporting errors when they happen is important for hospitals to gain data and better care for patients, but studies show that patient errors are usually underreported. This is concerning, as data on errors and other aspects of patient care are needed to inform quality improvement programs.

This study measured residents’ attitudes and knowledge regarding the filing of a PASS report. It also aimed to increase both the frequency of and knowledge about filing a PASS report with interventions. The results from each survey indicated a statistically significant increase in knowledge of when to file a PASS report. In the first survey, 53.7% of residents responded they they were involved in an instance where they should have filed a PASS report but did not. In the second survey, 58.5% of residents reported being involved in an instance where they should have filed a PASS report but did not. This difference was statistically significant (P = 0.036), sugesting that the intervention was successful at increasing residents’ knowledge regarding PASS reports and the appropriate times to file a PASS report.

The survey results also showed a trend toward increasing aggregate knowledge level of how to file PASS reports on the first survey and second surveys (from 31.5% vs 55.2%. This demonstrates an increase in knowledge of how to file a PASS report among residents at our hospital after the intervention. It should be noted that the intervention that was performed in this study was simple, easy to perform, and can be completed at any hospital system that uses a similar system for reporting patient errors.

Another important trend indicating the effectiveness of the intervention was a 15% increase in knowledge of what the PASS report acronym stands for, along with a 13.1% aggregate increase in the number of residents who filed a PASS report. This indicated that residents may have wanted to file a PASS report previously but simply did not know how to until the intervention. In addition, there was also a decrease in the aggregate percentages of residents who had never filed a PASS report and an increase in how many PASS reports were filed.

While PASS reports are a great way for hospitals to gain data and insight into problems at their sites, there was also a negative view of PASS reports. For example, a large percentage of residents indicated that filing a PASS report would not make any difference and that PASS reports are often used as a form of retaliation, either against themselves as the submitter or the person(s) mentioned in the PASS report. More specifically, more than 50% of residents felt that PASS reports were sometimes or often used as a form of retaliation against others. While many residents correctly identified in the survey that PASS reports are not equivalent to a “write-up,” it is concerning that they still feel there is a strong potential for retaliation when filing a PASS report. This finding is unfortunate but matches the results of a multicenter study that found that 44.6% of residents felt uncomfortable reporting patient errors, possibly secondary to fear of retaliation, along with issues with the reporting system.¹²

It is interesting to note that a minority of residents indicated that they feel that PASS reports are filed as often as they should be (25.9% on first survey and 24.1% on second survey). This is concerning, as the data gathered through PASS reports is used to improve patient care. However, the percentage reported in our study, although low, is higher than that reported in a similar study involving patients with Medicare insurance, which showed that only 14% of patient safety events were reported.¹³ These results demonstrate that further interventions are necessary in order to ensure that a PASS report is filed each time a patient safety event occurs.

Another finding of note is that the majority of residents also feel that the process of filing a PASS report is too time consuming. The majority of residents who have completed a PASS report stated that it took them between 10 and 20 minutes to complete a PASS report, but those same individuals also feel that it should take < 10 minutes to complete a PASS report. This is an important issue for hospital systems to address. Reducing the time it takes to file a PASS report may facilitate an increase in the amount of PASS reports filed.

We administered our surveys using email outreach to residents asking them to complete an anonymous online survey regarding the PASS report system using the REDCap software system. Researchers have various ways of administering surveys, ranging from paper surveys, emails, and even mobile apps. One study showed that online surveys tend to have higher response rates compared to non-online surveys, such as paper surveys and telephone surveys, which is likely due to the ease of use of online surveys.¹⁴ At the same time, unsolicited email surveys have been shown to have a negative influence on response rates. Mobile apps are a new way of administering surveys. However, research has not found any significant difference in the time required to complete the survey using mobile apps compared to other forms of administering surveys. In addition, surveys using mobile apps did not have increased response rates compared to other forms of administering surveys.¹⁵

To increase the response rate of our surveys, we offered gift cards to the study population for completing the survey. Studies have shown that surveys that offer incentives tend to have higher response rates than surveys that do not.¹⁶ Also, in addition to serving as a method for gathering data from our study population, we used our surveys as an intervention to increase awareness of PASS reporting, as reported in other studies. For example, another study used the HABITS questionnaire to not only gather information about children’s diet, but also to promote behavioral change towards healthy eating habits.¹⁷

This study had several limitations. First, the study was conducted using an anonymous online survey, which means we could not clarify questions that residents found confusing or needed further explanation. For example, 17 residents indicated in the first survey that they knew how to PASS report, but 19 residents indicated in the same survey that they have filed a PASS report in the past.

A second limitation of the study was that fewer residents completed the second survey (29 of 54 eligible residents) compared to the first survey (54 of 80 eligible residents). This may have impacted the results of the analysis, as certain findings were not statistically significant, despite trends in the data.

A third limitation of the study is that not all of the residents that completed the first and second surveys completed the entire intervention. For example, some residents did not attend the didactic lecture discussing PASS reports, and as such may not have received the appropriate training prior to completing the second survey.

The findings from this study can be used by the residency programs at UH-RH and by residency programs across the country to improve the involvement and attitudes of residents in reporting errors in patient care. Hospital staff need to be encouraged and educated on how to better report patient errors and the importance of reporting these errors. It would benefit hospital systems to provide continued and targeted training to familiarize physicians with the process of reporting patient errors, and take steps to reduce the time it takes to report patient errors. By increasing the reporting of errors, hospitals will be able to improve patient care through initiatives aimed at preventing errors.

Conclusion

Residents play an important role in providing high-quality care for patients. Part of providing high-quality care is the reporting of errors in patient care when they occur. Physicians, and in particular, residents, have historically underreported errors in patient care. Part of this underreporting results from residents not knowing or understanding the process of filing a report and feeling that the reports could be used as a form of retaliation. For hospital systems to continue to improve patient care, it is important for residents to not only know how to report errors in patient care but to feel comfortable doing so.

Corresponding author: Andrew J. Chin, DO, MS, MPH, Department of Internal Medicine, Adelante Healthcare, 1705 W Main St, Mesa, AZ 85201; [email protected].

Financial disclosures: None.

Funding: This study was funded by a research grant provided by Lake Eric College of Osteopathic Medicine to Andrew J. Chin and Anish Bhakta.

From Adelante Healthcare, Mesa, AZ (Dr. Chin), University Hospitals of Cleveland, Cleveland, OH (Drs. Delozier, Bascug, Levine, Bejanishvili, and Wynbrandt and Janet C. Peachey, Rachel M. Cerminara, and Sharon M. Darkovich), and Houston Methodist Hospitals, Houston, TX (Dr. Bhakta).

Abstract

Background: Resident physicians play an active role in the reporting of errors that occur in patient care. Previous studies indicate that residents significantly underreport errors in patient care.

Methods: Fifty-four of 80 eligible residents enrolled at University Hospitals–Regional Hospitals (UH-RH) during the 2018-2019 academic year completed a survey assessing their knowledge and experience in completing Patient Advocacy and Shared Stories (PASS) reports, which serve as incident reports in the UH health system in reporting errors in patient care. A series of interventions aimed at educating residents about the PASS report system were then conducted. The 54 residents who completed the first survey received it again 4 months later.

Results: Residents demonstrated greater understanding of when filing PASS reports was appropriate after the intervention, as significantly more residents reported having been involved in a situation where they should have filed a PASS report but did not (P = 0.036).

Conclusion: In this study, residents often did not report errors in patient care because they simply did not know the process for doing so. In addition, many residents often felt that the reporting of patient errors could be used as a form of retaliation.

Keywords: resident physicians; quality improvement; high-value care; medical errors; patient safety.

Resident physicians play a critical role in patient care. Residents undergo extensive supervised training in order to one day be able to practice medicine in an unsupervised setting, with the goal of providing the highest quality of care possible. One study reported that primary care provided by residents in a training program is of similar or higher quality than that provided by attending physicians.¹

Besides providing high-quality care, it is important that residents play an active role in the reporting of errors that occur regarding patient care as well as in identifying events that may compromise patient safety and quality.² In fact, increased reporting of patient errors has been shown to decrease liability-related costs for hospitals.³ Unfortunately, physicians, and residents in particular, have historically been poor reporters of errors in patient care.⁴ This is especially true when comparing physicians to other health professionals, such as nurses, in error reporting.⁵

Several studies have examined the involvement of residents in reporting errors in patient care. One recent study showed that a graduate medical education financial incentive program significantly increased the number of patient safety events reported by residents and fellows.⁶ This study, along with several others, supports the concept of using incentives to help improve the reporting of errors in patient care for physicians in training.^7-10 Another study used Quality Improvement Knowledge Assessment Tool (QIKAT) scores to assess quality improvement (QI) knowledge. The study demonstrated that self-assessment scores of QI skills using QIKAT scores improved following a targeted intervention.¹¹ Because further information on the involvement and attitudes of residents in reporting errors in patient care is needed, University Hospitals of Cleveland (UH) designed and implemented a QI study during the 2018-2019 academic year. This prospective study used anonymous surveys to objectively examine the involvement and attitudes of residents in reporting errors in patient care.

Methods

The UH health system uses Patient Advocacy and Shared Stories (PASS) reports as incident reports to not only disclose errors in patient care but also to identify any events that may compromise patient safety and quality. Based on preliminary review, nurses, ancillary staff, and administrators file the majority of PASS reports.

The study group consisted of residents at University Hospitals–Regional Hospitals (UH-RH), which is comprised of 2 hospitals: University Hospitals–Richmond Medical Center (UH-RMC) and University Hospitals –Bedford Medical Center (UH-BMC). UH-RMC and UH-BMC are 2 medium-sized university-affiliated community hospitals located in the Cleveland metropolitan area in Northeast Ohio. Both serve as clinical training sites for Case Western Reserve University School of Medicine and Lake Erie College of Osteopathic Medicine, the latter of which helped fund this study. The study was submitted to the Institutional Review Board (IRB) of University Hospitals of Cleveland and granted “not human subjects research” status as a QI study.

Surveys

UH-RH offers residency programs in dermatology, emergency medicine, family medicine, internal medicine, orthopedic surgery, and physical medicine and rehabilitation, along with a 1-year transitional/preliminary year. A total of 80 residents enrolled at UH-RH during the 2018-2019 academic year. All 80 residents at UH-RH received an email in December 2018 asking them to complete an anonymous survey regarding the PASS report system. The survey was administered using the REDCap software system and consisted of 15 multiple-choice questions. As an incentive for completing the survey, residents were offered a $10 Amazon gift card. The gift cards were funded through a research grant from Lake Erie College of Osteopathic Medicine. Residents were given 1 week to complete the survey. At the end of the week, 54 of 80 residents completed the first survey.

Following the first survey, efforts were undertaken by the study authors, in conjunction with the quality improvement department at UH-RH, to educate residents about the PASS report system. These interventions included giving a lecture on the PASS report system during resident didactic sessions, sending an email to all residents about the PASS report system, and providing residents an opportunity to complete an optional online training course regarding the PASS report system. As an incentive for completing the online training course, residents were offered a $10 Amazon gift card. As before, the gift cards were funded through a research grant from Lake Erie College of Osteopathic Medicine.

A second survey was administered in April 2019, 4 months after the first survey. To determine whether the intervention made an impact on the involvement and attitudes of residents in the reporting errors in patient care, only residents who completed the first survey were sent the second survey. The second survey consisted of the same questions as the first survey and was also administered using the REDCap software system. As an incentive for completing the survey, residents were offered another $10 Amazon gift card, again were funded through a research grant from Lake Erie College of Osteopathic Medicine. Residents were given 1 week to complete the survey.

Analysis

Chi-square analyses were utilized to examine differences between preintervention and postintervention responses across categories. All analyses were conducted using R statistical software, version 3.6.1 (R Foundation for Statistical Computing).

Results

A total of 54 of 80 eligible residents responded to the first survey (Table). Twenty-nine of 54 eligible residents responded to the second survey. Postintervention, significantly more residents indicated being involved in a situation where they should have filed a PASS report but did not (58.6% vs 53.7%; P = 0.036). Improvement was seen in PASS knowledge postintervention, where fewer residents reported not knowing how to file a PASS report (31.5% vs 55.2%; P = 0.059). No other improvements were significant, nor were there significant differences in responses between any other categories pre- and postintervention.

Discussion

Errors in patient care are a common occurrence in the hospital setting. Reporting errors when they happen is important for hospitals to gain data and better care for patients, but studies show that patient errors are usually underreported. This is concerning, as data on errors and other aspects of patient care are needed to inform quality improvement programs.

This study measured residents’ attitudes and knowledge regarding the filing of a PASS report. It also aimed to increase both the frequency of and knowledge about filing a PASS report with interventions. The results from each survey indicated a statistically significant increase in knowledge of when to file a PASS report. In the first survey, 53.7% of residents responded they they were involved in an instance where they should have filed a PASS report but did not. In the second survey, 58.5% of residents reported being involved in an instance where they should have filed a PASS report but did not. This difference was statistically significant (P = 0.036), sugesting that the intervention was successful at increasing residents’ knowledge regarding PASS reports and the appropriate times to file a PASS report.

The survey results also showed a trend toward increasing aggregate knowledge level of how to file PASS reports on the first survey and second surveys (from 31.5% vs 55.2%. This demonstrates an increase in knowledge of how to file a PASS report among residents at our hospital after the intervention. It should be noted that the intervention that was performed in this study was simple, easy to perform, and can be completed at any hospital system that uses a similar system for reporting patient errors.

Another important trend indicating the effectiveness of the intervention was a 15% increase in knowledge of what the PASS report acronym stands for, along with a 13.1% aggregate increase in the number of residents who filed a PASS report. This indicated that residents may have wanted to file a PASS report previously but simply did not know how to until the intervention. In addition, there was also a decrease in the aggregate percentages of residents who had never filed a PASS report and an increase in how many PASS reports were filed.

While PASS reports are a great way for hospitals to gain data and insight into problems at their sites, there was also a negative view of PASS reports. For example, a large percentage of residents indicated that filing a PASS report would not make any difference and that PASS reports are often used as a form of retaliation, either against themselves as the submitter or the person(s) mentioned in the PASS report. More specifically, more than 50% of residents felt that PASS reports were sometimes or often used as a form of retaliation against others. While many residents correctly identified in the survey that PASS reports are not equivalent to a “write-up,” it is concerning that they still feel there is a strong potential for retaliation when filing a PASS report. This finding is unfortunate but matches the results of a multicenter study that found that 44.6% of residents felt uncomfortable reporting patient errors, possibly secondary to fear of retaliation, along with issues with the reporting system.¹²

It is interesting to note that a minority of residents indicated that they feel that PASS reports are filed as often as they should be (25.9% on first survey and 24.1% on second survey). This is concerning, as the data gathered through PASS reports is used to improve patient care. However, the percentage reported in our study, although low, is higher than that reported in a similar study involving patients with Medicare insurance, which showed that only 14% of patient safety events were reported.¹³ These results demonstrate that further interventions are necessary in order to ensure that a PASS report is filed each time a patient safety event occurs.

Another finding of note is that the majority of residents also feel that the process of filing a PASS report is too time consuming. The majority of residents who have completed a PASS report stated that it took them between 10 and 20 minutes to complete a PASS report, but those same individuals also feel that it should take < 10 minutes to complete a PASS report. This is an important issue for hospital systems to address. Reducing the time it takes to file a PASS report may facilitate an increase in the amount of PASS reports filed.

We administered our surveys using email outreach to residents asking them to complete an anonymous online survey regarding the PASS report system using the REDCap software system. Researchers have various ways of administering surveys, ranging from paper surveys, emails, and even mobile apps. One study showed that online surveys tend to have higher response rates compared to non-online surveys, such as paper surveys and telephone surveys, which is likely due to the ease of use of online surveys.¹⁴ At the same time, unsolicited email surveys have been shown to have a negative influence on response rates. Mobile apps are a new way of administering surveys. However, research has not found any significant difference in the time required to complete the survey using mobile apps compared to other forms of administering surveys. In addition, surveys using mobile apps did not have increased response rates compared to other forms of administering surveys.¹⁵

To increase the response rate of our surveys, we offered gift cards to the study population for completing the survey. Studies have shown that surveys that offer incentives tend to have higher response rates than surveys that do not.¹⁶ Also, in addition to serving as a method for gathering data from our study population, we used our surveys as an intervention to increase awareness of PASS reporting, as reported in other studies. For example, another study used the HABITS questionnaire to not only gather information about children’s diet, but also to promote behavioral change towards healthy eating habits.¹⁷

This study had several limitations. First, the study was conducted using an anonymous online survey, which means we could not clarify questions that residents found confusing or needed further explanation. For example, 17 residents indicated in the first survey that they knew how to PASS report, but 19 residents indicated in the same survey that they have filed a PASS report in the past.

A second limitation of the study was that fewer residents completed the second survey (29 of 54 eligible residents) compared to the first survey (54 of 80 eligible residents). This may have impacted the results of the analysis, as certain findings were not statistically significant, despite trends in the data.

A third limitation of the study is that not all of the residents that completed the first and second surveys completed the entire intervention. For example, some residents did not attend the didactic lecture discussing PASS reports, and as such may not have received the appropriate training prior to completing the second survey.

The findings from this study can be used by the residency programs at UH-RH and by residency programs across the country to improve the involvement and attitudes of residents in reporting errors in patient care. Hospital staff need to be encouraged and educated on how to better report patient errors and the importance of reporting these errors. It would benefit hospital systems to provide continued and targeted training to familiarize physicians with the process of reporting patient errors, and take steps to reduce the time it takes to report patient errors. By increasing the reporting of errors, hospitals will be able to improve patient care through initiatives aimed at preventing errors.

Conclusion

Residents play an important role in providing high-quality care for patients. Part of providing high-quality care is the reporting of errors in patient care when they occur. Physicians, and in particular, residents, have historically underreported errors in patient care. Part of this underreporting results from residents not knowing or understanding the process of filing a report and feeling that the reports could be used as a form of retaliation. For hospital systems to continue to improve patient care, it is important for residents to not only know how to report errors in patient care but to feel comfortable doing so.

Corresponding author: Andrew J. Chin, DO, MS, MPH, Department of Internal Medicine, Adelante Healthcare, 1705 W Main St, Mesa, AZ 85201; [email protected].

Financial disclosures: None.

Funding: This study was funded by a research grant provided by Lake Eric College of Osteopathic Medicine to Andrew J. Chin and Anish Bhakta.

From Adelante Healthcare, Mesa, AZ (Dr. Chin), University Hospitals of Cleveland, Cleveland, OH (Drs. Delozier, Bascug, Levine, Bejanishvili, and Wynbrandt and Janet C. Peachey, Rachel M. Cerminara, and Sharon M. Darkovich), and Houston Methodist Hospitals, Houston, TX (Dr. Bhakta).

Abstract

Background: Resident physicians play an active role in the reporting of errors that occur in patient care. Previous studies indicate that residents significantly underreport errors in patient care.

Methods: Fifty-four of 80 eligible residents enrolled at University Hospitals–Regional Hospitals (UH-RH) during the 2018-2019 academic year completed a survey assessing their knowledge and experience in completing Patient Advocacy and Shared Stories (PASS) reports, which serve as incident reports in the UH health system in reporting errors in patient care. A series of interventions aimed at educating residents about the PASS report system were then conducted. The 54 residents who completed the first survey received it again 4 months later.

Results: Residents demonstrated greater understanding of when filing PASS reports was appropriate after the intervention, as significantly more residents reported having been involved in a situation where they should have filed a PASS report but did not (P = 0.036).

Conclusion: In this study, residents often did not report errors in patient care because they simply did not know the process for doing so. In addition, many residents often felt that the reporting of patient errors could be used as a form of retaliation.

Keywords: resident physicians; quality improvement; high-value care; medical errors; patient safety.

Resident physicians play a critical role in patient care. Residents undergo extensive supervised training in order to one day be able to practice medicine in an unsupervised setting, with the goal of providing the highest quality of care possible. One study reported that primary care provided by residents in a training program is of similar or higher quality than that provided by attending physicians.¹

Besides providing high-quality care, it is important that residents play an active role in the reporting of errors that occur regarding patient care as well as in identifying events that may compromise patient safety and quality.² In fact, increased reporting of patient errors has been shown to decrease liability-related costs for hospitals.³ Unfortunately, physicians, and residents in particular, have historically been poor reporters of errors in patient care.⁴ This is especially true when comparing physicians to other health professionals, such as nurses, in error reporting.⁵

Several studies have examined the involvement of residents in reporting errors in patient care. One recent study showed that a graduate medical education financial incentive program significantly increased the number of patient safety events reported by residents and fellows.⁶ This study, along with several others, supports the concept of using incentives to help improve the reporting of errors in patient care for physicians in training.^7-10 Another study used Quality Improvement Knowledge Assessment Tool (QIKAT) scores to assess quality improvement (QI) knowledge. The study demonstrated that self-assessment scores of QI skills using QIKAT scores improved following a targeted intervention.¹¹ Because further information on the involvement and attitudes of residents in reporting errors in patient care is needed, University Hospitals of Cleveland (UH) designed and implemented a QI study during the 2018-2019 academic year. This prospective study used anonymous surveys to objectively examine the involvement and attitudes of residents in reporting errors in patient care.

Methods

The UH health system uses Patient Advocacy and Shared Stories (PASS) reports as incident reports to not only disclose errors in patient care but also to identify any events that may compromise patient safety and quality. Based on preliminary review, nurses, ancillary staff, and administrators file the majority of PASS reports.

The study group consisted of residents at University Hospitals–Regional Hospitals (UH-RH), which is comprised of 2 hospitals: University Hospitals–Richmond Medical Center (UH-RMC) and University Hospitals –Bedford Medical Center (UH-BMC). UH-RMC and UH-BMC are 2 medium-sized university-affiliated community hospitals located in the Cleveland metropolitan area in Northeast Ohio. Both serve as clinical training sites for Case Western Reserve University School of Medicine and Lake Erie College of Osteopathic Medicine, the latter of which helped fund this study. The study was submitted to the Institutional Review Board (IRB) of University Hospitals of Cleveland and granted “not human subjects research” status as a QI study.

Surveys

UH-RH offers residency programs in dermatology, emergency medicine, family medicine, internal medicine, orthopedic surgery, and physical medicine and rehabilitation, along with a 1-year transitional/preliminary year. A total of 80 residents enrolled at UH-RH during the 2018-2019 academic year. All 80 residents at UH-RH received an email in December 2018 asking them to complete an anonymous survey regarding the PASS report system. The survey was administered using the REDCap software system and consisted of 15 multiple-choice questions. As an incentive for completing the survey, residents were offered a $10 Amazon gift card. The gift cards were funded through a research grant from Lake Erie College of Osteopathic Medicine. Residents were given 1 week to complete the survey. At the end of the week, 54 of 80 residents completed the first survey.

Following the first survey, efforts were undertaken by the study authors, in conjunction with the quality improvement department at UH-RH, to educate residents about the PASS report system. These interventions included giving a lecture on the PASS report system during resident didactic sessions, sending an email to all residents about the PASS report system, and providing residents an opportunity to complete an optional online training course regarding the PASS report system. As an incentive for completing the online training course, residents were offered a $10 Amazon gift card. As before, the gift cards were funded through a research grant from Lake Erie College of Osteopathic Medicine.

A second survey was administered in April 2019, 4 months after the first survey. To determine whether the intervention made an impact on the involvement and attitudes of residents in the reporting errors in patient care, only residents who completed the first survey were sent the second survey. The second survey consisted of the same questions as the first survey and was also administered using the REDCap software system. As an incentive for completing the survey, residents were offered another $10 Amazon gift card, again were funded through a research grant from Lake Erie College of Osteopathic Medicine. Residents were given 1 week to complete the survey.

Analysis

Chi-square analyses were utilized to examine differences between preintervention and postintervention responses across categories. All analyses were conducted using R statistical software, version 3.6.1 (R Foundation for Statistical Computing).

Results

A total of 54 of 80 eligible residents responded to the first survey (Table). Twenty-nine of 54 eligible residents responded to the second survey. Postintervention, significantly more residents indicated being involved in a situation where they should have filed a PASS report but did not (58.6% vs 53.7%; P = 0.036). Improvement was seen in PASS knowledge postintervention, where fewer residents reported not knowing how to file a PASS report (31.5% vs 55.2%; P = 0.059). No other improvements were significant, nor were there significant differences in responses between any other categories pre- and postintervention.

Discussion

Errors in patient care are a common occurrence in the hospital setting. Reporting errors when they happen is important for hospitals to gain data and better care for patients, but studies show that patient errors are usually underreported. This is concerning, as data on errors and other aspects of patient care are needed to inform quality improvement programs.

This study measured residents’ attitudes and knowledge regarding the filing of a PASS report. It also aimed to increase both the frequency of and knowledge about filing a PASS report with interventions. The results from each survey indicated a statistically significant increase in knowledge of when to file a PASS report. In the first survey, 53.7% of residents responded they they were involved in an instance where they should have filed a PASS report but did not. In the second survey, 58.5% of residents reported being involved in an instance where they should have filed a PASS report but did not. This difference was statistically significant (P = 0.036), sugesting that the intervention was successful at increasing residents’ knowledge regarding PASS reports and the appropriate times to file a PASS report.

The survey results also showed a trend toward increasing aggregate knowledge level of how to file PASS reports on the first survey and second surveys (from 31.5% vs 55.2%. This demonstrates an increase in knowledge of how to file a PASS report among residents at our hospital after the intervention. It should be noted that the intervention that was performed in this study was simple, easy to perform, and can be completed at any hospital system that uses a similar system for reporting patient errors.

Another important trend indicating the effectiveness of the intervention was a 15% increase in knowledge of what the PASS report acronym stands for, along with a 13.1% aggregate increase in the number of residents who filed a PASS report. This indicated that residents may have wanted to file a PASS report previously but simply did not know how to until the intervention. In addition, there was also a decrease in the aggregate percentages of residents who had never filed a PASS report and an increase in how many PASS reports were filed.

While PASS reports are a great way for hospitals to gain data and insight into problems at their sites, there was also a negative view of PASS reports. For example, a large percentage of residents indicated that filing a PASS report would not make any difference and that PASS reports are often used as a form of retaliation, either against themselves as the submitter or the person(s) mentioned in the PASS report. More specifically, more than 50% of residents felt that PASS reports were sometimes or often used as a form of retaliation against others. While many residents correctly identified in the survey that PASS reports are not equivalent to a “write-up,” it is concerning that they still feel there is a strong potential for retaliation when filing a PASS report. This finding is unfortunate but matches the results of a multicenter study that found that 44.6% of residents felt uncomfortable reporting patient errors, possibly secondary to fear of retaliation, along with issues with the reporting system.¹²

It is interesting to note that a minority of residents indicated that they feel that PASS reports are filed as often as they should be (25.9% on first survey and 24.1% on second survey). This is concerning, as the data gathered through PASS reports is used to improve patient care. However, the percentage reported in our study, although low, is higher than that reported in a similar study involving patients with Medicare insurance, which showed that only 14% of patient safety events were reported.¹³ These results demonstrate that further interventions are necessary in order to ensure that a PASS report is filed each time a patient safety event occurs.

Another finding of note is that the majority of residents also feel that the process of filing a PASS report is too time consuming. The majority of residents who have completed a PASS report stated that it took them between 10 and 20 minutes to complete a PASS report, but those same individuals also feel that it should take < 10 minutes to complete a PASS report. This is an important issue for hospital systems to address. Reducing the time it takes to file a PASS report may facilitate an increase in the amount of PASS reports filed.

We administered our surveys using email outreach to residents asking them to complete an anonymous online survey regarding the PASS report system using the REDCap software system. Researchers have various ways of administering surveys, ranging from paper surveys, emails, and even mobile apps. One study showed that online surveys tend to have higher response rates compared to non-online surveys, such as paper surveys and telephone surveys, which is likely due to the ease of use of online surveys.¹⁴ At the same time, unsolicited email surveys have been shown to have a negative influence on response rates. Mobile apps are a new way of administering surveys. However, research has not found any significant difference in the time required to complete the survey using mobile apps compared to other forms of administering surveys. In addition, surveys using mobile apps did not have increased response rates compared to other forms of administering surveys.¹⁵

To increase the response rate of our surveys, we offered gift cards to the study population for completing the survey. Studies have shown that surveys that offer incentives tend to have higher response rates than surveys that do not.¹⁶ Also, in addition to serving as a method for gathering data from our study population, we used our surveys as an intervention to increase awareness of PASS reporting, as reported in other studies. For example, another study used the HABITS questionnaire to not only gather information about children’s diet, but also to promote behavioral change towards healthy eating habits.¹⁷

This study had several limitations. First, the study was conducted using an anonymous online survey, which means we could not clarify questions that residents found confusing or needed further explanation. For example, 17 residents indicated in the first survey that they knew how to PASS report, but 19 residents indicated in the same survey that they have filed a PASS report in the past.

A second limitation of the study was that fewer residents completed the second survey (29 of 54 eligible residents) compared to the first survey (54 of 80 eligible residents). This may have impacted the results of the analysis, as certain findings were not statistically significant, despite trends in the data.

A third limitation of the study is that not all of the residents that completed the first and second surveys completed the entire intervention. For example, some residents did not attend the didactic lecture discussing PASS reports, and as such may not have received the appropriate training prior to completing the second survey.

The findings from this study can be used by the residency programs at UH-RH and by residency programs across the country to improve the involvement and attitudes of residents in reporting errors in patient care. Hospital staff need to be encouraged and educated on how to better report patient errors and the importance of reporting these errors. It would benefit hospital systems to provide continued and targeted training to familiarize physicians with the process of reporting patient errors, and take steps to reduce the time it takes to report patient errors. By increasing the reporting of errors, hospitals will be able to improve patient care through initiatives aimed at preventing errors.

Conclusion

Residents play an important role in providing high-quality care for patients. Part of providing high-quality care is the reporting of errors in patient care when they occur. Physicians, and in particular, residents, have historically underreported errors in patient care. Part of this underreporting results from residents not knowing or understanding the process of filing a report and feeling that the reports could be used as a form of retaliation. For hospital systems to continue to improve patient care, it is important for residents to not only know how to report errors in patient care but to feel comfortable doing so.

Corresponding author: Andrew J. Chin, DO, MS, MPH, Department of Internal Medicine, Adelante Healthcare, 1705 W Main St, Mesa, AZ 85201; [email protected].

Financial disclosures: None.

Funding: This study was funded by a research grant provided by Lake Eric College of Osteopathic Medicine to Andrew J. Chin and Anish Bhakta.

From Mount Sinai Medical Center, Miami Beach, FL.

Abstract

Objective: To develop and implement a process for administering COVID-19 monoclonal antibody infusions for outpatients with mild or moderate COVID-19 at high risk for hospitalization, using multidisciplinary collaboration, US Food and Drug Administration (FDA) guidance, and infection prevention standards.

Methods: When monoclonal antibody therapy became available for mild or moderate COVID-19 outpatients via Emergency Use Authorization (EUA), our institution sought to provide this therapy option to our patients. We describe the process for planning, implementing, and maintaining a successful program for administering novel therapies based on FDA guidance and infection prevention standards. Key components of our implementation process were multidisciplinary planning involving decision makers and stakeholders; setting realistic goals in the process; team communication; and measuring and reporting quality improvement on a regular basis.

Results: A total of 790 COVID-19 monoclonal antibody infusions were administered from November 20, 2020 to March 5, 2021. Steps to minimize the likelihood of adverse drug reactions were implemented and a low incidence (< 1%) has occurred. There has been no concern from staff regarding infection during the process. Rarely, patients have raised cost-related concerns, typically due to incomplete communication regarding billing prior to the infusion. Patients, families, nursing staff, physicians, pharmacy, and hospital administration have expressed satisfaction with the program.

Conclusion: This process can provide a template for other hospitals or health care delivery facilities to provide novel therapies to patients with mild or moderate COVID-19 in a safe and effective manner.

Keywords: COVID-19; monoclonal antibody; infusion; emergency use authorization.

SARS-CoV-2 and the disease it causes, COVID-19, have transformed from scientific vernacular to common household terms. It began with a cluster of pneumonia cases of unknown etiology in December 2019 in Wuhan, China, with physicians there reporting a novel coronavirus strain (2019-nCoV), now referred to as SARS-CoV-2. Rapid spread of this virus resulted in the World Health Organization (WHO) declaring an international public health emergency. Since this time, the virus has evolved into a worldwide pandemic. COVID-19 has dramatically impacted our society, resulting in more than 2.63 million global deaths as of this writing, of which more than 527,000 deaths have occurred in the United States.¹ This novel virus has resulted in a flurry of literature, research, therapies, and collaboration across multiple disciplines in an effort to prevent, treat, and mitigate cases and complications of this disease.

On November 9, 2020, and November 21, 2020, the US Food and Drug Administration (FDA) issued Emergency Use Authorizations (EUA) for 2 novel COVID-19 monoclonal therapies, bamlanivimab^2-3 and casirivimab/imdevimab,^3-4 respectively. The EUAs granted permission for these therapies to be administered for the treatment of mild to moderate COVID-19 in adult and pediatric patients (≥ 12 years and weighing at least 40 kg) with positive results of direct SARS-CoV-2 viral testing and who are at high risk for progressing to severe COVID-19 and/or hospitalization. The therapies work by targeting the SARS-CoV-2 spike protein and subsequent attachment to human angiotensin-converting enzyme 2 receptors. Clinical trial data leading to the EUA demonstrated a reduction in viral load, safe outcome, and most importantly, fewer hospitalization and emergency room visits, as compared to the placebo group.^5-7 The use of monoclonal antibodies is not new and gained recognition during the Ebola crisis, when the monoclonal antibody to the Ebola virus showed a significant survival benefit.⁸ Providing monoclonal antibody therapy soon after symptom onset aligns with a shift from the onset of the pandemic to the current focus on the administration of pharmaceutical therapy early in the disease course. This shift prevents progression to severe COVID-19, with the goal of reducing patient mortality, hospitalizations, and strain on health care systems.

The availability of novel neutralizing monoclonal antibodies for COVID-19 led to discussions of how to incorporate these therapies as new options for patients. Our institution networked with colleagues from multiple disciplines to discuss processes and policies for the safe administration of the monoclonal antibody infusion therapies. Federal health leaders urge more use of monoclonal antibodies, but many hospitals have been unable to successfully implement infusions due to staff and logistical challenges.⁹ This article presents a viable process that hospitals can use to provide these novel therapies to outpatients with mild to moderate COVID-19.

The Mount Sinai Medical Center, Florida Experience

Mount Sinai Medical Center in Miami Beach, Florida, is the largest private, independent, not-for-profit teaching hospital in South Florida, comprising 672 licensed beds and supporting 150,000 emergency department (ED) visits annually. Per the EUA criteria for use, COVID-19 monoclonal antibody therapies are not authorized for patients who are hospitalized or who require oxygen therapy due to COVID-19. Therefore, options for outpatient administration needed to be evaluated. Directly following the first EUA press release, a task force of key stakeholders was assembled to brainstorm and develop a process to offer this therapy to the community. A multidisciplinary task force with representation from the ED, nursing, primary care, hospital medicine, pharmacy, risk management, billing, information technology, infection prevention, and senior level leadership participated (Table).

The task force reviewed institutional outpatient locations to determine whether offering this service would be feasible (eg, ED, ambulatory care facilities, cancer center). The ED was selected because it would offer the largest array of appointment times to meet the community needs with around-the-clock availability. While Mount Sinai Medical Center offers care in 3 emergency center locations in Aventura, Hialeah, and Miami Beach, it was determined to initiate the infusions at the main campus center in Miami Beach only. The main campus affords an onsite pharmacy with suitable staffing to prepare the anticipated volume of infusions in a timely manner, as both therapies have short stabilities following preparation. Thus, it was decided that patients from freestanding emergency centers in Aventura and Hialeah would be moved to the Miami Beach ED location to receive therapy. Operating at a single site also allowed for more rapid implementation, monitoring, and ability to make modifications more easily. Discussions for the possible expansion of COVID-19 monoclonal antibody infusions at satellite locations are underway.

On November 20, 2020, 11 days after the formation of the multidisciplinary task force, the first COVID-19 monoclonal infusion was successfully administered. Figure 1 depicts the timeline from assessment to program implementation. Critical to implementation was the involvement of decision makers from all necessary departments early in the planning process to ensure that standard operating procedures were followed and that the patients, community, and organization had a positive experience. This allowed for simultaneous planning of electronic health record (Epic; EHR) builds, departmental workflows, and staff education, as described in the following section. Figure 2 shows the patient safety activities included in the implementation process.

On the day of bamlanivimab EUA release, email communication was shared among hospital leadership with details of the press release. Departments were quickly involved to initiate a task force to assess if and how this therapy could be offered at Mount Sinai Medical Center. The following sections explain the role of each stakeholder and their essential role to operationalize these novel EUA treatment options. The task force was organized and led by our chief medical officer and chief nursing officer.

Information Technology

Medication Ordering and Documentation EHR and Smart Pumps. Early in the pandemic, the antimicrobial stewardship (ASP) clinical coordinator became the designated point person for pharmacy assessment of novel COVID-19 therapies. As such, this pharmacist began reviewing the bamlanivimab and, later, the casirivimab/imdevimab EUA Fact Sheet for Health Care Providers. All necessary elements for the complete and safe ordering and dispensing of the medication were developed and reviewed by pharmacy administration and ED nursing leadership for input, prior to submitting to the information technology team for implementation. Building the COVID-19 monoclonal medication records into the EHR allowed for detailed direction (ie, administration and preparation instructions) to be consistently applied. The medication records were also built into hospital smart pumps so that nurses could access prepopulated, accurate volumes and infusion rates to minimize errors.

Order Set Development. The pharmacy medication build was added to a comprehensive order set (Figure 3), which was then developed to guide prescribers and standardize the process around ordering of COVID-19 monoclonal therapies. While these therapies are new, oncology monoclonal therapies are regularly administered to outpatients at Mount Sinai Cancer Center. The cancer center was therefore consulted on their process surrounding best practices in administration of monoclonal antibody therapies. This included protocols for medications used in pretreatment and management of hypersensitivity reactions and potential adverse drug reactions of both COVID-19 monoclonal therapies. These medication orders were selected by default in the order set to ensure that all patients received premedications aimed at minimizing the risk of hypersensitivity reaction, and had as-needed medication orders, in the event a hypersensitivity reaction occurred. Reducing hypersensitivity reaction risk is important as well to increase the likelihood that the patient would receive full therapy, as management of this adverse drug reactions involves possible cessation of therapy depending on the level of severity. The pharmacy department also ensured these medications were stocked in ED automated dispensing cabinets to promote quick access. In addition to the aforementioned nursing orders, we added EUA criteria for use and hyperlinks to the Fact Sheets for Patients and Caregivers and Health Care Providers for each monoclonal therapy, and restricted ordering to ED physicians, nurse practitioners, and physician assistants.

The order set underwent multidisciplinary review by pharmacy administration, the chair of emergency medicine, physicians, and ED nursing leadership prior to presentation and approval by the Pharmacy and Therapeutics Committee. Lastly, at time of implementation, the order set was added to the ED preference list, preventing inpatient access. Additionally, as a patient safety action, free- standing orders of COVID-19 monoclonal therapies were disabled, so providers could only order therapies via the approved, comprehensive order set.

Preliminary Assessment Tool. A provider assessment tool was developed to document patient-specific EUA criteria for use during initial assessment (Figure 4). This tool serves as a checklist and is visible to the full multidisciplinary team in the patient’s EHR. It is used as a resource at the time of pharmacist verification and ED physician assessment to ensure criteria for use are met.

Patient Referral. Patients with symptoms or concerns of COVID-19 exposure can make physician appointments via telemedicine or in person at Mount Sinai Medical Center’s primary care and specialty offices. At the time of patient encounter, physicians suspecting a COVID-19 diagnosis will refer patients for outpatient COVID-19 polymerase chain reaction (PCR) laboratory testing, which has an approximate 24-hour turnaround to results. Physicians also assess whether the patient meets EUA criteria for use, pending results of testing. In the event a patient meets EUA criteria for use, the physician provides patient counseling and requests verbal consent. Following this, the physician enters a note in the EHR describing the patient’s condition, criteria for use evaluation, and the patient’s verbal agreement to therapy. This preliminary screening is beneficial to begin planning with both the patient and ED to minimize delays. Patients are notified of the results of their test once available. If the COVID-19 PCR test returns positive, the physician will call the ED at the main campus and schedule the patient for COVID-19 monoclonal therapy. As the desired timeframe for administering COVID-19 monoclonal therapies is within less than 10 days of symptom onset, timely scheduling of appointments is crucial. Infusion appointments are typically provided the same or next day. The patients are informed that they must bring documentation of their positive COVID-19 PCR test to their ED visit. Lastly, because patients are pretreated with medication that may potentially impair driving, they are instructed that they cannot drive themselves home; ride shares also are not allowed in order to limit the spread of infection.

Emergency Department

Patient Arrival and Screening. A COVID-19 patient can be evaluated in the ED 1 of 2 ways. The first option is via outpatient office referral, as described previously. Upon arrival to the ED, a second screening is performed to ensure the patient still meets EUA criteria for use and the positive COVID-19 PCR test result is confirmed. If the patient no longer meets criteria, the patient is triaged accordingly, including evaluation for higher-level care (eg, supplemental oxygen, hospital admission). The second optoion is via new patient walk-ins without outpatient physician referral (Figure 4). In these cases, an initial screening is performed, documenting EUA criteria for use in the preliminary assessment (Figure 5). Physicians will consider an outside COVID-19 test as valid, so long as documentation is readily available confirming a positive PCR result. Otherwise, an in-house COVID-19 PCR test will be performed, which has a 2-hour turnaround time.

Infusion Schedule. The ED offers a total of 16 COVID-19 monoclonal infusions slots daily. These are broken up into 4 infusion time blocks (eg, 8 am, 12 pm, 4 pm, 8 pm), with each infusion time block consisting of 4 available patient appointments. A list of scheduled infusions for the day is emailed to the pharmacy department every morning, and patients are instructed to arrive 1 hour prior to their appointment time. This allows time for patient registration, assessment, and pharmacy notification in advance of order entry. For logistical purposes, and as a patient safety initiative to reduce the likelihood of medication errors, each of the available COVID-19 monoclonal antibodies is offered on a designated day. Bamlanivimab is offered on Tuesday, Thursday, Saturday, and Sunday, while casirivimab/imdevimab is offered Monday, Wednesday, and Friday. This provides flexibility to adjust should supply deviate based on Department of Health allocation or should new therapy options within this class of medication become available.

Patient Education. Prior to administration of the monoclonal therapy, physician and nursing staff obtain a formal, written patient consent for therapy and provide patients with the option of participating in the institutional review board (IRB) approved study. Details of this are discussed in the risk management and IRB sections of the article. Nursing staff also provides the medication-specific Fact Sheet for Patients and Caregivers in either Spanish or English, which is also included as a hyperlink on the COVID-19 Monoclonal Antibody Order Set for ease of access. Interpreter services are available for patients who speak other languages. An ED decentralized pharmacist is also available onsite Monday through Friday from 12 pm to 8:30 pm to supplement education and serve as a resource for any questions.

Infusion Ordering. Once the patient is ready to begin therapy, the he/she is brought to a dedicated overflow area of the ED. There are few, if any, patients in this location, and it is adjacent to the main emergency center for easy access by the patients, nurses, pharmacists, and physicians. The physician then enters orders in the EHR using the COVID-19 Monoclonal Antibody Order Set (Figure 3). Three discrete questions were built into the medication order: (1) Was patient consent obtained? (2) Was the Fact Sheet for Patient/Caregiver provided to the patient? (3) Is the patient COVID-19 PCR-positive? These questions were built as hard stops so that the medication orders cannot be placed without a response. This serves as another double-check to ensure processes are followed and helps facilitate timely verification by the pharmacist.

Medication Administration. One nurse is dedicated to administering the monoclonal therapies scheduled at 8 am and 12 pm and another at 4 pm and 8 pm. Each appointment block is 4 hours in duration to allow adequate time for patient registration, infusion, and postinfusion observation. The nurse administers the premedications and COVID-19 monoclonal therapy, and observes the patient for the required 1-hour postadministration observation period. Nursing orders detailing monitoring parameters for mild, moderate, and severe reactions, along with associated medication orders to administer in the event they occur, are detailed in the nursing orders of the COVID-19 Monoclonal Antibody Order Set (Figure 3). Prior to administration, the nurse scans each medication and the patient’s wrist identification band, and documents the time of administration within the EHR medication administration report.

Pharmacy Department

Medication Receipt Process. Inventory is currently allocated biweekly from the state department of health and will soon be transitioning to a direct order system. The pharmacy technician in charge of deliveries notifies the pharmacy Antimicrobial Stewardship Program (ASP) clinical coordinator upon receipt of the monoclonal therapies. Bamlanivimab is supplied as 1 vial per dose, whereas casirivimab/imdevimab is supplied as 4 vials or 8 vials per dose, depending how it is shipped. To reduce the likelihood of medication errors, the ASP clinical coordinator assembles each of the casirivimab/imdevimab vials into kits, where 1 kit equals 1 dose. Labels are then affixed to each kit indicating the medication name, number of vials which equal a full dose, and pharmacist signature. The kits are stored in a dedicated refrigerator, and inventory logs are affixed to the outside of the refrigerator and updated daily. This inventory is also communicated daily to ED physician, nursing, and pharmacy leadership, as well as the director of patient safety, who reports weekly usage to the state Department of Health and Human Services. These weekly reports are used to determine allocation amounts.

Medication Verification and Delivery. The Mount Sinai Medical Center pharmacist staffing model consists of centralized order entry and specialized, decentralized positions. All orders are verified by the ED pharmacist when scheduled (not a 24/7 service) and by the designated pharmacist for all other times. At the time of medication verification, the pharmacist documents patient-specific EUA criteria for use and confirms that consent was obtained and the Fact Sheet for Patients/Caregivers was provided. A pharmacist intervention was developed to assist with this documentation. Pharmacists input smart text “.COVIDmonoclonal” and a drop-down menu of EUA criteria for use appears. The pharmacist reviews the patient care notes and medication order question responses to ascertain this information, contacting the ED prescriber if further clarification is required. This verification serves as another check to ensure processes put in place are followed. Lastly, intravenous preparation and delivery are electronically recorded in the EHR, and the medications require nursing signature at the time of delivery to ensure a formal chain of custody.

Risk Management

At Mount Sinai Medical Center, all EUA and investigational therapies require patient consent. Consistent with this requirement, a COVID-19 monoclonal specific consent was developed by risk management. This is provided to every patient receiving a COVID-19 monoclonal infusion, in addition to the FDA EUA Fact Sheet for Patients and Caregivers, and documented as part of their EHR. The questions providers must answer are built into the order set to ensure this process is followed and these patient safety checks are incorporated into the workflow.

Billing and Finance Department

In alignment with Mount Sinai Medical Center’s mission to provide high-quality health care to its diverse community through teaching, research, charity care, and financial responsibility, it was determined that this therapy would be provided to all patients regardless of insurance type, including those who are uninsured. The billing and finance department was consulted prior to this service being offered, to provide patients with accurate and pertinent information. The billing and finance department provided guidance on how to document patient encounters at time of registration to facilitate appropriate billing. At this time, the medication is free of charge, but nonmedication-related ED fees apply. This is explained to patients so there is a clear understanding prior to booking their appointment.

Infection Prevention

As patients receiving COVID-19 monoclonal therapies can transmit the virus to others, measures to ensure protection for other patients and staff are vital. To minimize exposure, specific nursing and physician staff from the ED are assigned to the treatment of these patients, and patients receive infusions and postobservation monitoring in a designated wing of the ED. Additionally, all staff who interact with these patients are required to don full personal protective equipment. This includes not only physicians and nurses but all specialties such as physician assistants, nurse practitioners, pharmacists, and laboratory technicians. Moreover, patients are not permitted to go home in a ride share and are counseled on Centers for Disease Control and Prevention quarantining following infusion.

Measurement of Process and Outcomes and Reporting

IRB approval was sought and obtained early during initiation of this service, allowing study consent to be offered to patients at the time general consent was obtained, which maximized patient recruitment and streamlined workflow. The study is a prospective observational research study to determine the impact of administration of COVID-19 monoclonal antibody therapy on length of symptoms, chronic illness, and rate of hospitalization. Most patients were eager to participate and offer their assistance to the scientific community during this pandemic.

Staff Education

In order to successfully implement this multidisciplinary EUA treatment option, comprehensive staff education was paramount after the workflow was developed. Prior to the first day of infusions, nurses and pharmacists were provided education during multiple huddle announcements. The pharmacy team also provided screen captures via email to the pharmacists so they could become familiar with the order set, intervention documentation, and location of the preliminary assessment of EUA criteria for use at the time of order verification. The emergency medicine department chair and chief medical officer also provided education via several virtual meetings and email to referring physicians (specialists and primary care) and residents in the emergency centers involved in COVID-19 monoclonal therapy-related patient care.

Factors Contributing to Success

We believe the reasons for continued success of this process are multifactorial and include the following key elements. Multidisciplinary planning, which included decision makers and all stakeholders, began at the time the idea was conceived. This allowed quick implementation of this service by efficiently navigating barriers to engaging impacted staff early on. Throughout this process, the authors set realistic step-wise goals. While navigating through the many details to implementation described, we also kept in mind the big picture, which was to provide this potentially lifesaving therapy to as many qualifying members of our community as possible. This included being flexible with the process and adapting when needed to achieve this ultimate goal. A focus on safety remained a priority to minimize possible errors and enhance patient and staff satisfaction. The optimization of the EHR streamlined workflow, provided point-of-care resources, and enhanced patient safety. Additionally, the target date set for implementation allowed staff and department leads adequate time to plan for and anticipate the changes. Serving only 1 patient on the first day allowed time for staff to experience this new process hands-on and provided opportunity for focused education. This team communication was essential to implementing this project, including staff training of processes and procedures prior to go-live. Early incorporation of IRB approval allowed the experience to be assessed and considered for contribution to the scientific literature to tackle this novel virus that has impacted our communities locally, nationally, and abroad. Moreover, continued measurement and reporting on a regular basis leads to performance improvement. The process outlined here can be adapted to incorporate other new therapies in the future, such as the recent February 9, 2021, EUA of the COVID-19 monoclonal antibody combination bamlanivimab and etesevimab.¹⁰

Conclusion

We administered 790 COVID-19 monoclonal antibody infusions between November 20, 2020 and March 5, 2021. Steps to minimize the likelihood of hypersensitivity reactions were implemented, and a low incidence (< 1%) has been observed. There has been no incidence of infection, concern from staff about infection prevention, or risk of infection during the processes. There have been very infrequent cost-related concerns raised by patients, typically due to incomplete communication regarding billing prior to the infusion. To address these issues, staff education has been provided to enhance patient instruction on this topic. The program has provided patient and family satisfaction, as well nursing, physician, pharmacist, clinical staff, and hospital administration pride and gratification. Setting up a new program to provide a 4-hour patient encounter to infuse therapy to high-risk patients with COVID-19 requires commitment and effort. This article describes the experience, ideas, and formula others may consider using to set up such a program. Through networking and formal phone calls and meetings about monoclonal antibody therapy, we have heard about other institutions who have not been able to institute this program due to various barriers to implementation. We hope our experience serves as a resource for others to provide this therapy to their patients and expand access in an effort to mitigate COVID-19 consequences and cases affecting our communities.

Corresponding author: Kathleen Jodoin, PharmD, BCPS, Mount Sinai Medical Center, 4300 Alton Rd, Miami Beach, FL 33140; [email protected].

Financial disclosures: None.

From Mount Sinai Medical Center, Miami Beach, FL.

Abstract

Objective: To develop and implement a process for administering COVID-19 monoclonal antibody infusions for outpatients with mild or moderate COVID-19 at high risk for hospitalization, using multidisciplinary collaboration, US Food and Drug Administration (FDA) guidance, and infection prevention standards.

Methods: When monoclonal antibody therapy became available for mild or moderate COVID-19 outpatients via Emergency Use Authorization (EUA), our institution sought to provide this therapy option to our patients. We describe the process for planning, implementing, and maintaining a successful program for administering novel therapies based on FDA guidance and infection prevention standards. Key components of our implementation process were multidisciplinary planning involving decision makers and stakeholders; setting realistic goals in the process; team communication; and measuring and reporting quality improvement on a regular basis.

Results: A total of 790 COVID-19 monoclonal antibody infusions were administered from November 20, 2020 to March 5, 2021. Steps to minimize the likelihood of adverse drug reactions were implemented and a low incidence (< 1%) has occurred. There has been no concern from staff regarding infection during the process. Rarely, patients have raised cost-related concerns, typically due to incomplete communication regarding billing prior to the infusion. Patients, families, nursing staff, physicians, pharmacy, and hospital administration have expressed satisfaction with the program.

Conclusion: This process can provide a template for other hospitals or health care delivery facilities to provide novel therapies to patients with mild or moderate COVID-19 in a safe and effective manner.

Keywords: COVID-19; monoclonal antibody; infusion; emergency use authorization.

SARS-CoV-2 and the disease it causes, COVID-19, have transformed from scientific vernacular to common household terms. It began with a cluster of pneumonia cases of unknown etiology in December 2019 in Wuhan, China, with physicians there reporting a novel coronavirus strain (2019-nCoV), now referred to as SARS-CoV-2. Rapid spread of this virus resulted in the World Health Organization (WHO) declaring an international public health emergency. Since this time, the virus has evolved into a worldwide pandemic. COVID-19 has dramatically impacted our society, resulting in more than 2.63 million global deaths as of this writing, of which more than 527,000 deaths have occurred in the United States.¹ This novel virus has resulted in a flurry of literature, research, therapies, and collaboration across multiple disciplines in an effort to prevent, treat, and mitigate cases and complications of this disease.

On November 9, 2020, and November 21, 2020, the US Food and Drug Administration (FDA) issued Emergency Use Authorizations (EUA) for 2 novel COVID-19 monoclonal therapies, bamlanivimab^2-3 and casirivimab/imdevimab,^3-4 respectively. The EUAs granted permission for these therapies to be administered for the treatment of mild to moderate COVID-19 in adult and pediatric patients (≥ 12 years and weighing at least 40 kg) with positive results of direct SARS-CoV-2 viral testing and who are at high risk for progressing to severe COVID-19 and/or hospitalization. The therapies work by targeting the SARS-CoV-2 spike protein and subsequent attachment to human angiotensin-converting enzyme 2 receptors. Clinical trial data leading to the EUA demonstrated a reduction in viral load, safe outcome, and most importantly, fewer hospitalization and emergency room visits, as compared to the placebo group.^5-7 The use of monoclonal antibodies is not new and gained recognition during the Ebola crisis, when the monoclonal antibody to the Ebola virus showed a significant survival benefit.⁸ Providing monoclonal antibody therapy soon after symptom onset aligns with a shift from the onset of the pandemic to the current focus on the administration of pharmaceutical therapy early in the disease course. This shift prevents progression to severe COVID-19, with the goal of reducing patient mortality, hospitalizations, and strain on health care systems.

The availability of novel neutralizing monoclonal antibodies for COVID-19 led to discussions of how to incorporate these therapies as new options for patients. Our institution networked with colleagues from multiple disciplines to discuss processes and policies for the safe administration of the monoclonal antibody infusion therapies. Federal health leaders urge more use of monoclonal antibodies, but many hospitals have been unable to successfully implement infusions due to staff and logistical challenges.⁹ This article presents a viable process that hospitals can use to provide these novel therapies to outpatients with mild to moderate COVID-19.

The Mount Sinai Medical Center, Florida Experience

Mount Sinai Medical Center in Miami Beach, Florida, is the largest private, independent, not-for-profit teaching hospital in South Florida, comprising 672 licensed beds and supporting 150,000 emergency department (ED) visits annually. Per the EUA criteria for use, COVID-19 monoclonal antibody therapies are not authorized for patients who are hospitalized or who require oxygen therapy due to COVID-19. Therefore, options for outpatient administration needed to be evaluated. Directly following the first EUA press release, a task force of key stakeholders was assembled to brainstorm and develop a process to offer this therapy to the community. A multidisciplinary task force with representation from the ED, nursing, primary care, hospital medicine, pharmacy, risk management, billing, information technology, infection prevention, and senior level leadership participated (Table).

The task force reviewed institutional outpatient locations to determine whether offering this service would be feasible (eg, ED, ambulatory care facilities, cancer center). The ED was selected because it would offer the largest array of appointment times to meet the community needs with around-the-clock availability. While Mount Sinai Medical Center offers care in 3 emergency center locations in Aventura, Hialeah, and Miami Beach, it was determined to initiate the infusions at the main campus center in Miami Beach only. The main campus affords an onsite pharmacy with suitable staffing to prepare the anticipated volume of infusions in a timely manner, as both therapies have short stabilities following preparation. Thus, it was decided that patients from freestanding emergency centers in Aventura and Hialeah would be moved to the Miami Beach ED location to receive therapy. Operating at a single site also allowed for more rapid implementation, monitoring, and ability to make modifications more easily. Discussions for the possible expansion of COVID-19 monoclonal antibody infusions at satellite locations are underway.

On November 20, 2020, 11 days after the formation of the multidisciplinary task force, the first COVID-19 monoclonal infusion was successfully administered. Figure 1 depicts the timeline from assessment to program implementation. Critical to implementation was the involvement of decision makers from all necessary departments early in the planning process to ensure that standard operating procedures were followed and that the patients, community, and organization had a positive experience. This allowed for simultaneous planning of electronic health record (Epic; EHR) builds, departmental workflows, and staff education, as described in the following section. Figure 2 shows the patient safety activities included in the implementation process.

On the day of bamlanivimab EUA release, email communication was shared among hospital leadership with details of the press release. Departments were quickly involved to initiate a task force to assess if and how this therapy could be offered at Mount Sinai Medical Center. The following sections explain the role of each stakeholder and their essential role to operationalize these novel EUA treatment options. The task force was organized and led by our chief medical officer and chief nursing officer.

Information Technology

Medication Ordering and Documentation EHR and Smart Pumps. Early in the pandemic, the antimicrobial stewardship (ASP) clinical coordinator became the designated point person for pharmacy assessment of novel COVID-19 therapies. As such, this pharmacist began reviewing the bamlanivimab and, later, the casirivimab/imdevimab EUA Fact Sheet for Health Care Providers. All necessary elements for the complete and safe ordering and dispensing of the medication were developed and reviewed by pharmacy administration and ED nursing leadership for input, prior to submitting to the information technology team for implementation. Building the COVID-19 monoclonal medication records into the EHR allowed for detailed direction (ie, administration and preparation instructions) to be consistently applied. The medication records were also built into hospital smart pumps so that nurses could access prepopulated, accurate volumes and infusion rates to minimize errors.

Order Set Development. The pharmacy medication build was added to a comprehensive order set (Figure 3), which was then developed to guide prescribers and standardize the process around ordering of COVID-19 monoclonal therapies. While these therapies are new, oncology monoclonal therapies are regularly administered to outpatients at Mount Sinai Cancer Center. The cancer center was therefore consulted on their process surrounding best practices in administration of monoclonal antibody therapies. This included protocols for medications used in pretreatment and management of hypersensitivity reactions and potential adverse drug reactions of both COVID-19 monoclonal therapies. These medication orders were selected by default in the order set to ensure that all patients received premedications aimed at minimizing the risk of hypersensitivity reaction, and had as-needed medication orders, in the event a hypersensitivity reaction occurred. Reducing hypersensitivity reaction risk is important as well to increase the likelihood that the patient would receive full therapy, as management of this adverse drug reactions involves possible cessation of therapy depending on the level of severity. The pharmacy department also ensured these medications were stocked in ED automated dispensing cabinets to promote quick access. In addition to the aforementioned nursing orders, we added EUA criteria for use and hyperlinks to the Fact Sheets for Patients and Caregivers and Health Care Providers for each monoclonal therapy, and restricted ordering to ED physicians, nurse practitioners, and physician assistants.

The order set underwent multidisciplinary review by pharmacy administration, the chair of emergency medicine, physicians, and ED nursing leadership prior to presentation and approval by the Pharmacy and Therapeutics Committee. Lastly, at time of implementation, the order set was added to the ED preference list, preventing inpatient access. Additionally, as a patient safety action, free- standing orders of COVID-19 monoclonal therapies were disabled, so providers could only order therapies via the approved, comprehensive order set.

Preliminary Assessment Tool. A provider assessment tool was developed to document patient-specific EUA criteria for use during initial assessment (Figure 4). This tool serves as a checklist and is visible to the full multidisciplinary team in the patient’s EHR. It is used as a resource at the time of pharmacist verification and ED physician assessment to ensure criteria for use are met.

Patient Referral. Patients with symptoms or concerns of COVID-19 exposure can make physician appointments via telemedicine or in person at Mount Sinai Medical Center’s primary care and specialty offices. At the time of patient encounter, physicians suspecting a COVID-19 diagnosis will refer patients for outpatient COVID-19 polymerase chain reaction (PCR) laboratory testing, which has an approximate 24-hour turnaround to results. Physicians also assess whether the patient meets EUA criteria for use, pending results of testing. In the event a patient meets EUA criteria for use, the physician provides patient counseling and requests verbal consent. Following this, the physician enters a note in the EHR describing the patient’s condition, criteria for use evaluation, and the patient’s verbal agreement to therapy. This preliminary screening is beneficial to begin planning with both the patient and ED to minimize delays. Patients are notified of the results of their test once available. If the COVID-19 PCR test returns positive, the physician will call the ED at the main campus and schedule the patient for COVID-19 monoclonal therapy. As the desired timeframe for administering COVID-19 monoclonal therapies is within less than 10 days of symptom onset, timely scheduling of appointments is crucial. Infusion appointments are typically provided the same or next day. The patients are informed that they must bring documentation of their positive COVID-19 PCR test to their ED visit. Lastly, because patients are pretreated with medication that may potentially impair driving, they are instructed that they cannot drive themselves home; ride shares also are not allowed in order to limit the spread of infection.

Emergency Department

Patient Arrival and Screening. A COVID-19 patient can be evaluated in the ED 1 of 2 ways. The first option is via outpatient office referral, as described previously. Upon arrival to the ED, a second screening is performed to ensure the patient still meets EUA criteria for use and the positive COVID-19 PCR test result is confirmed. If the patient no longer meets criteria, the patient is triaged accordingly, including evaluation for higher-level care (eg, supplemental oxygen, hospital admission). The second optoion is via new patient walk-ins without outpatient physician referral (Figure 4). In these cases, an initial screening is performed, documenting EUA criteria for use in the preliminary assessment (Figure 5). Physicians will consider an outside COVID-19 test as valid, so long as documentation is readily available confirming a positive PCR result. Otherwise, an in-house COVID-19 PCR test will be performed, which has a 2-hour turnaround time.

Infusion Schedule. The ED offers a total of 16 COVID-19 monoclonal infusions slots daily. These are broken up into 4 infusion time blocks (eg, 8 am, 12 pm, 4 pm, 8 pm), with each infusion time block consisting of 4 available patient appointments. A list of scheduled infusions for the day is emailed to the pharmacy department every morning, and patients are instructed to arrive 1 hour prior to their appointment time. This allows time for patient registration, assessment, and pharmacy notification in advance of order entry. For logistical purposes, and as a patient safety initiative to reduce the likelihood of medication errors, each of the available COVID-19 monoclonal antibodies is offered on a designated day. Bamlanivimab is offered on Tuesday, Thursday, Saturday, and Sunday, while casirivimab/imdevimab is offered Monday, Wednesday, and Friday. This provides flexibility to adjust should supply deviate based on Department of Health allocation or should new therapy options within this class of medication become available.

Patient Education. Prior to administration of the monoclonal therapy, physician and nursing staff obtain a formal, written patient consent for therapy and provide patients with the option of participating in the institutional review board (IRB) approved study. Details of this are discussed in the risk management and IRB sections of the article. Nursing staff also provides the medication-specific Fact Sheet for Patients and Caregivers in either Spanish or English, which is also included as a hyperlink on the COVID-19 Monoclonal Antibody Order Set for ease of access. Interpreter services are available for patients who speak other languages. An ED decentralized pharmacist is also available onsite Monday through Friday from 12 pm to 8:30 pm to supplement education and serve as a resource for any questions.

Infusion Ordering. Once the patient is ready to begin therapy, the he/she is brought to a dedicated overflow area of the ED. There are few, if any, patients in this location, and it is adjacent to the main emergency center for easy access by the patients, nurses, pharmacists, and physicians. The physician then enters orders in the EHR using the COVID-19 Monoclonal Antibody Order Set (Figure 3). Three discrete questions were built into the medication order: (1) Was patient consent obtained? (2) Was the Fact Sheet for Patient/Caregiver provided to the patient? (3) Is the patient COVID-19 PCR-positive? These questions were built as hard stops so that the medication orders cannot be placed without a response. This serves as another double-check to ensure processes are followed and helps facilitate timely verification by the pharmacist.

Medication Administration. One nurse is dedicated to administering the monoclonal therapies scheduled at 8 am and 12 pm and another at 4 pm and 8 pm. Each appointment block is 4 hours in duration to allow adequate time for patient registration, infusion, and postinfusion observation. The nurse administers the premedications and COVID-19 monoclonal therapy, and observes the patient for the required 1-hour postadministration observation period. Nursing orders detailing monitoring parameters for mild, moderate, and severe reactions, along with associated medication orders to administer in the event they occur, are detailed in the nursing orders of the COVID-19 Monoclonal Antibody Order Set (Figure 3). Prior to administration, the nurse scans each medication and the patient’s wrist identification band, and documents the time of administration within the EHR medication administration report.

Pharmacy Department

Medication Receipt Process. Inventory is currently allocated biweekly from the state department of health and will soon be transitioning to a direct order system. The pharmacy technician in charge of deliveries notifies the pharmacy Antimicrobial Stewardship Program (ASP) clinical coordinator upon receipt of the monoclonal therapies. Bamlanivimab is supplied as 1 vial per dose, whereas casirivimab/imdevimab is supplied as 4 vials or 8 vials per dose, depending how it is shipped. To reduce the likelihood of medication errors, the ASP clinical coordinator assembles each of the casirivimab/imdevimab vials into kits, where 1 kit equals 1 dose. Labels are then affixed to each kit indicating the medication name, number of vials which equal a full dose, and pharmacist signature. The kits are stored in a dedicated refrigerator, and inventory logs are affixed to the outside of the refrigerator and updated daily. This inventory is also communicated daily to ED physician, nursing, and pharmacy leadership, as well as the director of patient safety, who reports weekly usage to the state Department of Health and Human Services. These weekly reports are used to determine allocation amounts.

Medication Verification and Delivery. The Mount Sinai Medical Center pharmacist staffing model consists of centralized order entry and specialized, decentralized positions. All orders are verified by the ED pharmacist when scheduled (not a 24/7 service) and by the designated pharmacist for all other times. At the time of medication verification, the pharmacist documents patient-specific EUA criteria for use and confirms that consent was obtained and the Fact Sheet for Patients/Caregivers was provided. A pharmacist intervention was developed to assist with this documentation. Pharmacists input smart text “.COVIDmonoclonal” and a drop-down menu of EUA criteria for use appears. The pharmacist reviews the patient care notes and medication order question responses to ascertain this information, contacting the ED prescriber if further clarification is required. This verification serves as another check to ensure processes put in place are followed. Lastly, intravenous preparation and delivery are electronically recorded in the EHR, and the medications require nursing signature at the time of delivery to ensure a formal chain of custody.

Risk Management

At Mount Sinai Medical Center, all EUA and investigational therapies require patient consent. Consistent with this requirement, a COVID-19 monoclonal specific consent was developed by risk management. This is provided to every patient receiving a COVID-19 monoclonal infusion, in addition to the FDA EUA Fact Sheet for Patients and Caregivers, and documented as part of their EHR. The questions providers must answer are built into the order set to ensure this process is followed and these patient safety checks are incorporated into the workflow.

Billing and Finance Department

In alignment with Mount Sinai Medical Center’s mission to provide high-quality health care to its diverse community through teaching, research, charity care, and financial responsibility, it was determined that this therapy would be provided to all patients regardless of insurance type, including those who are uninsured. The billing and finance department was consulted prior to this service being offered, to provide patients with accurate and pertinent information. The billing and finance department provided guidance on how to document patient encounters at time of registration to facilitate appropriate billing. At this time, the medication is free of charge, but nonmedication-related ED fees apply. This is explained to patients so there is a clear understanding prior to booking their appointment.

Infection Prevention

As patients receiving COVID-19 monoclonal therapies can transmit the virus to others, measures to ensure protection for other patients and staff are vital. To minimize exposure, specific nursing and physician staff from the ED are assigned to the treatment of these patients, and patients receive infusions and postobservation monitoring in a designated wing of the ED. Additionally, all staff who interact with these patients are required to don full personal protective equipment. This includes not only physicians and nurses but all specialties such as physician assistants, nurse practitioners, pharmacists, and laboratory technicians. Moreover, patients are not permitted to go home in a ride share and are counseled on Centers for Disease Control and Prevention quarantining following infusion.

Measurement of Process and Outcomes and Reporting

IRB approval was sought and obtained early during initiation of this service, allowing study consent to be offered to patients at the time general consent was obtained, which maximized patient recruitment and streamlined workflow. The study is a prospective observational research study to determine the impact of administration of COVID-19 monoclonal antibody therapy on length of symptoms, chronic illness, and rate of hospitalization. Most patients were eager to participate and offer their assistance to the scientific community during this pandemic.

Staff Education

In order to successfully implement this multidisciplinary EUA treatment option, comprehensive staff education was paramount after the workflow was developed. Prior to the first day of infusions, nurses and pharmacists were provided education during multiple huddle announcements. The pharmacy team also provided screen captures via email to the pharmacists so they could become familiar with the order set, intervention documentation, and location of the preliminary assessment of EUA criteria for use at the time of order verification. The emergency medicine department chair and chief medical officer also provided education via several virtual meetings and email to referring physicians (specialists and primary care) and residents in the emergency centers involved in COVID-19 monoclonal therapy-related patient care.

Factors Contributing to Success

We believe the reasons for continued success of this process are multifactorial and include the following key elements. Multidisciplinary planning, which included decision makers and all stakeholders, began at the time the idea was conceived. This allowed quick implementation of this service by efficiently navigating barriers to engaging impacted staff early on. Throughout this process, the authors set realistic step-wise goals. While navigating through the many details to implementation described, we also kept in mind the big picture, which was to provide this potentially lifesaving therapy to as many qualifying members of our community as possible. This included being flexible with the process and adapting when needed to achieve this ultimate goal. A focus on safety remained a priority to minimize possible errors and enhance patient and staff satisfaction. The optimization of the EHR streamlined workflow, provided point-of-care resources, and enhanced patient safety. Additionally, the target date set for implementation allowed staff and department leads adequate time to plan for and anticipate the changes. Serving only 1 patient on the first day allowed time for staff to experience this new process hands-on and provided opportunity for focused education. This team communication was essential to implementing this project, including staff training of processes and procedures prior to go-live. Early incorporation of IRB approval allowed the experience to be assessed and considered for contribution to the scientific literature to tackle this novel virus that has impacted our communities locally, nationally, and abroad. Moreover, continued measurement and reporting on a regular basis leads to performance improvement. The process outlined here can be adapted to incorporate other new therapies in the future, such as the recent February 9, 2021, EUA of the COVID-19 monoclonal antibody combination bamlanivimab and etesevimab.¹⁰

Conclusion

We administered 790 COVID-19 monoclonal antibody infusions between November 20, 2020 and March 5, 2021. Steps to minimize the likelihood of hypersensitivity reactions were implemented, and a low incidence (< 1%) has been observed. There has been no incidence of infection, concern from staff about infection prevention, or risk of infection during the processes. There have been very infrequent cost-related concerns raised by patients, typically due to incomplete communication regarding billing prior to the infusion. To address these issues, staff education has been provided to enhance patient instruction on this topic. The program has provided patient and family satisfaction, as well nursing, physician, pharmacist, clinical staff, and hospital administration pride and gratification. Setting up a new program to provide a 4-hour patient encounter to infuse therapy to high-risk patients with COVID-19 requires commitment and effort. This article describes the experience, ideas, and formula others may consider using to set up such a program. Through networking and formal phone calls and meetings about monoclonal antibody therapy, we have heard about other institutions who have not been able to institute this program due to various barriers to implementation. We hope our experience serves as a resource for others to provide this therapy to their patients and expand access in an effort to mitigate COVID-19 consequences and cases affecting our communities.

Corresponding author: Kathleen Jodoin, PharmD, BCPS, Mount Sinai Medical Center, 4300 Alton Rd, Miami Beach, FL 33140; [email protected].

Financial disclosures: None.

From Mount Sinai Medical Center, Miami Beach, FL.

Abstract

Objective: To develop and implement a process for administering COVID-19 monoclonal antibody infusions for outpatients with mild or moderate COVID-19 at high risk for hospitalization, using multidisciplinary collaboration, US Food and Drug Administration (FDA) guidance, and infection prevention standards.

Methods: When monoclonal antibody therapy became available for mild or moderate COVID-19 outpatients via Emergency Use Authorization (EUA), our institution sought to provide this therapy option to our patients. We describe the process for planning, implementing, and maintaining a successful program for administering novel therapies based on FDA guidance and infection prevention standards. Key components of our implementation process were multidisciplinary planning involving decision makers and stakeholders; setting realistic goals in the process; team communication; and measuring and reporting quality improvement on a regular basis.

Results: A total of 790 COVID-19 monoclonal antibody infusions were administered from November 20, 2020 to March 5, 2021. Steps to minimize the likelihood of adverse drug reactions were implemented and a low incidence (< 1%) has occurred. There has been no concern from staff regarding infection during the process. Rarely, patients have raised cost-related concerns, typically due to incomplete communication regarding billing prior to the infusion. Patients, families, nursing staff, physicians, pharmacy, and hospital administration have expressed satisfaction with the program.

Conclusion: This process can provide a template for other hospitals or health care delivery facilities to provide novel therapies to patients with mild or moderate COVID-19 in a safe and effective manner.

Keywords: COVID-19; monoclonal antibody; infusion; emergency use authorization.

SARS-CoV-2 and the disease it causes, COVID-19, have transformed from scientific vernacular to common household terms. It began with a cluster of pneumonia cases of unknown etiology in December 2019 in Wuhan, China, with physicians there reporting a novel coronavirus strain (2019-nCoV), now referred to as SARS-CoV-2. Rapid spread of this virus resulted in the World Health Organization (WHO) declaring an international public health emergency. Since this time, the virus has evolved into a worldwide pandemic. COVID-19 has dramatically impacted our society, resulting in more than 2.63 million global deaths as of this writing, of which more than 527,000 deaths have occurred in the United States.¹ This novel virus has resulted in a flurry of literature, research, therapies, and collaboration across multiple disciplines in an effort to prevent, treat, and mitigate cases and complications of this disease.

On November 9, 2020, and November 21, 2020, the US Food and Drug Administration (FDA) issued Emergency Use Authorizations (EUA) for 2 novel COVID-19 monoclonal therapies, bamlanivimab^2-3 and casirivimab/imdevimab,^3-4 respectively. The EUAs granted permission for these therapies to be administered for the treatment of mild to moderate COVID-19 in adult and pediatric patients (≥ 12 years and weighing at least 40 kg) with positive results of direct SARS-CoV-2 viral testing and who are at high risk for progressing to severe COVID-19 and/or hospitalization. The therapies work by targeting the SARS-CoV-2 spike protein and subsequent attachment to human angiotensin-converting enzyme 2 receptors. Clinical trial data leading to the EUA demonstrated a reduction in viral load, safe outcome, and most importantly, fewer hospitalization and emergency room visits, as compared to the placebo group.^5-7 The use of monoclonal antibodies is not new and gained recognition during the Ebola crisis, when the monoclonal antibody to the Ebola virus showed a significant survival benefit.⁸ Providing monoclonal antibody therapy soon after symptom onset aligns with a shift from the onset of the pandemic to the current focus on the administration of pharmaceutical therapy early in the disease course. This shift prevents progression to severe COVID-19, with the goal of reducing patient mortality, hospitalizations, and strain on health care systems.

The availability of novel neutralizing monoclonal antibodies for COVID-19 led to discussions of how to incorporate these therapies as new options for patients. Our institution networked with colleagues from multiple disciplines to discuss processes and policies for the safe administration of the monoclonal antibody infusion therapies. Federal health leaders urge more use of monoclonal antibodies, but many hospitals have been unable to successfully implement infusions due to staff and logistical challenges.⁹ This article presents a viable process that hospitals can use to provide these novel therapies to outpatients with mild to moderate COVID-19.

The Mount Sinai Medical Center, Florida Experience

Mount Sinai Medical Center in Miami Beach, Florida, is the largest private, independent, not-for-profit teaching hospital in South Florida, comprising 672 licensed beds and supporting 150,000 emergency department (ED) visits annually. Per the EUA criteria for use, COVID-19 monoclonal antibody therapies are not authorized for patients who are hospitalized or who require oxygen therapy due to COVID-19. Therefore, options for outpatient administration needed to be evaluated. Directly following the first EUA press release, a task force of key stakeholders was assembled to brainstorm and develop a process to offer this therapy to the community. A multidisciplinary task force with representation from the ED, nursing, primary care, hospital medicine, pharmacy, risk management, billing, information technology, infection prevention, and senior level leadership participated (Table).

The task force reviewed institutional outpatient locations to determine whether offering this service would be feasible (eg, ED, ambulatory care facilities, cancer center). The ED was selected because it would offer the largest array of appointment times to meet the community needs with around-the-clock availability. While Mount Sinai Medical Center offers care in 3 emergency center locations in Aventura, Hialeah, and Miami Beach, it was determined to initiate the infusions at the main campus center in Miami Beach only. The main campus affords an onsite pharmacy with suitable staffing to prepare the anticipated volume of infusions in a timely manner, as both therapies have short stabilities following preparation. Thus, it was decided that patients from freestanding emergency centers in Aventura and Hialeah would be moved to the Miami Beach ED location to receive therapy. Operating at a single site also allowed for more rapid implementation, monitoring, and ability to make modifications more easily. Discussions for the possible expansion of COVID-19 monoclonal antibody infusions at satellite locations are underway.

On November 20, 2020, 11 days after the formation of the multidisciplinary task force, the first COVID-19 monoclonal infusion was successfully administered. Figure 1 depicts the timeline from assessment to program implementation. Critical to implementation was the involvement of decision makers from all necessary departments early in the planning process to ensure that standard operating procedures were followed and that the patients, community, and organization had a positive experience. This allowed for simultaneous planning of electronic health record (Epic; EHR) builds, departmental workflows, and staff education, as described in the following section. Figure 2 shows the patient safety activities included in the implementation process.

On the day of bamlanivimab EUA release, email communication was shared among hospital leadership with details of the press release. Departments were quickly involved to initiate a task force to assess if and how this therapy could be offered at Mount Sinai Medical Center. The following sections explain the role of each stakeholder and their essential role to operationalize these novel EUA treatment options. The task force was organized and led by our chief medical officer and chief nursing officer.

Information Technology

Medication Ordering and Documentation EHR and Smart Pumps. Early in the pandemic, the antimicrobial stewardship (ASP) clinical coordinator became the designated point person for pharmacy assessment of novel COVID-19 therapies. As such, this pharmacist began reviewing the bamlanivimab and, later, the casirivimab/imdevimab EUA Fact Sheet for Health Care Providers. All necessary elements for the complete and safe ordering and dispensing of the medication were developed and reviewed by pharmacy administration and ED nursing leadership for input, prior to submitting to the information technology team for implementation. Building the COVID-19 monoclonal medication records into the EHR allowed for detailed direction (ie, administration and preparation instructions) to be consistently applied. The medication records were also built into hospital smart pumps so that nurses could access prepopulated, accurate volumes and infusion rates to minimize errors.

Order Set Development. The pharmacy medication build was added to a comprehensive order set (Figure 3), which was then developed to guide prescribers and standardize the process around ordering of COVID-19 monoclonal therapies. While these therapies are new, oncology monoclonal therapies are regularly administered to outpatients at Mount Sinai Cancer Center. The cancer center was therefore consulted on their process surrounding best practices in administration of monoclonal antibody therapies. This included protocols for medications used in pretreatment and management of hypersensitivity reactions and potential adverse drug reactions of both COVID-19 monoclonal therapies. These medication orders were selected by default in the order set to ensure that all patients received premedications aimed at minimizing the risk of hypersensitivity reaction, and had as-needed medication orders, in the event a hypersensitivity reaction occurred. Reducing hypersensitivity reaction risk is important as well to increase the likelihood that the patient would receive full therapy, as management of this adverse drug reactions involves possible cessation of therapy depending on the level of severity. The pharmacy department also ensured these medications were stocked in ED automated dispensing cabinets to promote quick access. In addition to the aforementioned nursing orders, we added EUA criteria for use and hyperlinks to the Fact Sheets for Patients and Caregivers and Health Care Providers for each monoclonal therapy, and restricted ordering to ED physicians, nurse practitioners, and physician assistants.

The order set underwent multidisciplinary review by pharmacy administration, the chair of emergency medicine, physicians, and ED nursing leadership prior to presentation and approval by the Pharmacy and Therapeutics Committee. Lastly, at time of implementation, the order set was added to the ED preference list, preventing inpatient access. Additionally, as a patient safety action, free- standing orders of COVID-19 monoclonal therapies were disabled, so providers could only order therapies via the approved, comprehensive order set.

Preliminary Assessment Tool. A provider assessment tool was developed to document patient-specific EUA criteria for use during initial assessment (Figure 4). This tool serves as a checklist and is visible to the full multidisciplinary team in the patient’s EHR. It is used as a resource at the time of pharmacist verification and ED physician assessment to ensure criteria for use are met.

Patient Referral. Patients with symptoms or concerns of COVID-19 exposure can make physician appointments via telemedicine or in person at Mount Sinai Medical Center’s primary care and specialty offices. At the time of patient encounter, physicians suspecting a COVID-19 diagnosis will refer patients for outpatient COVID-19 polymerase chain reaction (PCR) laboratory testing, which has an approximate 24-hour turnaround to results. Physicians also assess whether the patient meets EUA criteria for use, pending results of testing. In the event a patient meets EUA criteria for use, the physician provides patient counseling and requests verbal consent. Following this, the physician enters a note in the EHR describing the patient’s condition, criteria for use evaluation, and the patient’s verbal agreement to therapy. This preliminary screening is beneficial to begin planning with both the patient and ED to minimize delays. Patients are notified of the results of their test once available. If the COVID-19 PCR test returns positive, the physician will call the ED at the main campus and schedule the patient for COVID-19 monoclonal therapy. As the desired timeframe for administering COVID-19 monoclonal therapies is within less than 10 days of symptom onset, timely scheduling of appointments is crucial. Infusion appointments are typically provided the same or next day. The patients are informed that they must bring documentation of their positive COVID-19 PCR test to their ED visit. Lastly, because patients are pretreated with medication that may potentially impair driving, they are instructed that they cannot drive themselves home; ride shares also are not allowed in order to limit the spread of infection.

Emergency Department

Patient Arrival and Screening. A COVID-19 patient can be evaluated in the ED 1 of 2 ways. The first option is via outpatient office referral, as described previously. Upon arrival to the ED, a second screening is performed to ensure the patient still meets EUA criteria for use and the positive COVID-19 PCR test result is confirmed. If the patient no longer meets criteria, the patient is triaged accordingly, including evaluation for higher-level care (eg, supplemental oxygen, hospital admission). The second optoion is via new patient walk-ins without outpatient physician referral (Figure 4). In these cases, an initial screening is performed, documenting EUA criteria for use in the preliminary assessment (Figure 5). Physicians will consider an outside COVID-19 test as valid, so long as documentation is readily available confirming a positive PCR result. Otherwise, an in-house COVID-19 PCR test will be performed, which has a 2-hour turnaround time.

Infusion Schedule. The ED offers a total of 16 COVID-19 monoclonal infusions slots daily. These are broken up into 4 infusion time blocks (eg, 8 am, 12 pm, 4 pm, 8 pm), with each infusion time block consisting of 4 available patient appointments. A list of scheduled infusions for the day is emailed to the pharmacy department every morning, and patients are instructed to arrive 1 hour prior to their appointment time. This allows time for patient registration, assessment, and pharmacy notification in advance of order entry. For logistical purposes, and as a patient safety initiative to reduce the likelihood of medication errors, each of the available COVID-19 monoclonal antibodies is offered on a designated day. Bamlanivimab is offered on Tuesday, Thursday, Saturday, and Sunday, while casirivimab/imdevimab is offered Monday, Wednesday, and Friday. This provides flexibility to adjust should supply deviate based on Department of Health allocation or should new therapy options within this class of medication become available.

Patient Education. Prior to administration of the monoclonal therapy, physician and nursing staff obtain a formal, written patient consent for therapy and provide patients with the option of participating in the institutional review board (IRB) approved study. Details of this are discussed in the risk management and IRB sections of the article. Nursing staff also provides the medication-specific Fact Sheet for Patients and Caregivers in either Spanish or English, which is also included as a hyperlink on the COVID-19 Monoclonal Antibody Order Set for ease of access. Interpreter services are available for patients who speak other languages. An ED decentralized pharmacist is also available onsite Monday through Friday from 12 pm to 8:30 pm to supplement education and serve as a resource for any questions.

Infusion Ordering. Once the patient is ready to begin therapy, the he/she is brought to a dedicated overflow area of the ED. There are few, if any, patients in this location, and it is adjacent to the main emergency center for easy access by the patients, nurses, pharmacists, and physicians. The physician then enters orders in the EHR using the COVID-19 Monoclonal Antibody Order Set (Figure 3). Three discrete questions were built into the medication order: (1) Was patient consent obtained? (2) Was the Fact Sheet for Patient/Caregiver provided to the patient? (3) Is the patient COVID-19 PCR-positive? These questions were built as hard stops so that the medication orders cannot be placed without a response. This serves as another double-check to ensure processes are followed and helps facilitate timely verification by the pharmacist.

Medication Administration. One nurse is dedicated to administering the monoclonal therapies scheduled at 8 am and 12 pm and another at 4 pm and 8 pm. Each appointment block is 4 hours in duration to allow adequate time for patient registration, infusion, and postinfusion observation. The nurse administers the premedications and COVID-19 monoclonal therapy, and observes the patient for the required 1-hour postadministration observation period. Nursing orders detailing monitoring parameters for mild, moderate, and severe reactions, along with associated medication orders to administer in the event they occur, are detailed in the nursing orders of the COVID-19 Monoclonal Antibody Order Set (Figure 3). Prior to administration, the nurse scans each medication and the patient’s wrist identification band, and documents the time of administration within the EHR medication administration report.

Pharmacy Department

Medication Receipt Process. Inventory is currently allocated biweekly from the state department of health and will soon be transitioning to a direct order system. The pharmacy technician in charge of deliveries notifies the pharmacy Antimicrobial Stewardship Program (ASP) clinical coordinator upon receipt of the monoclonal therapies. Bamlanivimab is supplied as 1 vial per dose, whereas casirivimab/imdevimab is supplied as 4 vials or 8 vials per dose, depending how it is shipped. To reduce the likelihood of medication errors, the ASP clinical coordinator assembles each of the casirivimab/imdevimab vials into kits, where 1 kit equals 1 dose. Labels are then affixed to each kit indicating the medication name, number of vials which equal a full dose, and pharmacist signature. The kits are stored in a dedicated refrigerator, and inventory logs are affixed to the outside of the refrigerator and updated daily. This inventory is also communicated daily to ED physician, nursing, and pharmacy leadership, as well as the director of patient safety, who reports weekly usage to the state Department of Health and Human Services. These weekly reports are used to determine allocation amounts.

Medication Verification and Delivery. The Mount Sinai Medical Center pharmacist staffing model consists of centralized order entry and specialized, decentralized positions. All orders are verified by the ED pharmacist when scheduled (not a 24/7 service) and by the designated pharmacist for all other times. At the time of medication verification, the pharmacist documents patient-specific EUA criteria for use and confirms that consent was obtained and the Fact Sheet for Patients/Caregivers was provided. A pharmacist intervention was developed to assist with this documentation. Pharmacists input smart text “.COVIDmonoclonal” and a drop-down menu of EUA criteria for use appears. The pharmacist reviews the patient care notes and medication order question responses to ascertain this information, contacting the ED prescriber if further clarification is required. This verification serves as another check to ensure processes put in place are followed. Lastly, intravenous preparation and delivery are electronically recorded in the EHR, and the medications require nursing signature at the time of delivery to ensure a formal chain of custody.

Risk Management

At Mount Sinai Medical Center, all EUA and investigational therapies require patient consent. Consistent with this requirement, a COVID-19 monoclonal specific consent was developed by risk management. This is provided to every patient receiving a COVID-19 monoclonal infusion, in addition to the FDA EUA Fact Sheet for Patients and Caregivers, and documented as part of their EHR. The questions providers must answer are built into the order set to ensure this process is followed and these patient safety checks are incorporated into the workflow.

Billing and Finance Department

In alignment with Mount Sinai Medical Center’s mission to provide high-quality health care to its diverse community through teaching, research, charity care, and financial responsibility, it was determined that this therapy would be provided to all patients regardless of insurance type, including those who are uninsured. The billing and finance department was consulted prior to this service being offered, to provide patients with accurate and pertinent information. The billing and finance department provided guidance on how to document patient encounters at time of registration to facilitate appropriate billing. At this time, the medication is free of charge, but nonmedication-related ED fees apply. This is explained to patients so there is a clear understanding prior to booking their appointment.

Infection Prevention

As patients receiving COVID-19 monoclonal therapies can transmit the virus to others, measures to ensure protection for other patients and staff are vital. To minimize exposure, specific nursing and physician staff from the ED are assigned to the treatment of these patients, and patients receive infusions and postobservation monitoring in a designated wing of the ED. Additionally, all staff who interact with these patients are required to don full personal protective equipment. This includes not only physicians and nurses but all specialties such as physician assistants, nurse practitioners, pharmacists, and laboratory technicians. Moreover, patients are not permitted to go home in a ride share and are counseled on Centers for Disease Control and Prevention quarantining following infusion.

Measurement of Process and Outcomes and Reporting

IRB approval was sought and obtained early during initiation of this service, allowing study consent to be offered to patients at the time general consent was obtained, which maximized patient recruitment and streamlined workflow. The study is a prospective observational research study to determine the impact of administration of COVID-19 monoclonal antibody therapy on length of symptoms, chronic illness, and rate of hospitalization. Most patients were eager to participate and offer their assistance to the scientific community during this pandemic.

Staff Education

In order to successfully implement this multidisciplinary EUA treatment option, comprehensive staff education was paramount after the workflow was developed. Prior to the first day of infusions, nurses and pharmacists were provided education during multiple huddle announcements. The pharmacy team also provided screen captures via email to the pharmacists so they could become familiar with the order set, intervention documentation, and location of the preliminary assessment of EUA criteria for use at the time of order verification. The emergency medicine department chair and chief medical officer also provided education via several virtual meetings and email to referring physicians (specialists and primary care) and residents in the emergency centers involved in COVID-19 monoclonal therapy-related patient care.

Factors Contributing to Success

We believe the reasons for continued success of this process are multifactorial and include the following key elements. Multidisciplinary planning, which included decision makers and all stakeholders, began at the time the idea was conceived. This allowed quick implementation of this service by efficiently navigating barriers to engaging impacted staff early on. Throughout this process, the authors set realistic step-wise goals. While navigating through the many details to implementation described, we also kept in mind the big picture, which was to provide this potentially lifesaving therapy to as many qualifying members of our community as possible. This included being flexible with the process and adapting when needed to achieve this ultimate goal. A focus on safety remained a priority to minimize possible errors and enhance patient and staff satisfaction. The optimization of the EHR streamlined workflow, provided point-of-care resources, and enhanced patient safety. Additionally, the target date set for implementation allowed staff and department leads adequate time to plan for and anticipate the changes. Serving only 1 patient on the first day allowed time for staff to experience this new process hands-on and provided opportunity for focused education. This team communication was essential to implementing this project, including staff training of processes and procedures prior to go-live. Early incorporation of IRB approval allowed the experience to be assessed and considered for contribution to the scientific literature to tackle this novel virus that has impacted our communities locally, nationally, and abroad. Moreover, continued measurement and reporting on a regular basis leads to performance improvement. The process outlined here can be adapted to incorporate other new therapies in the future, such as the recent February 9, 2021, EUA of the COVID-19 monoclonal antibody combination bamlanivimab and etesevimab.¹⁰

Conclusion

We administered 790 COVID-19 monoclonal antibody infusions between November 20, 2020 and March 5, 2021. Steps to minimize the likelihood of hypersensitivity reactions were implemented, and a low incidence (< 1%) has been observed. There has been no incidence of infection, concern from staff about infection prevention, or risk of infection during the processes. There have been very infrequent cost-related concerns raised by patients, typically due to incomplete communication regarding billing prior to the infusion. To address these issues, staff education has been provided to enhance patient instruction on this topic. The program has provided patient and family satisfaction, as well nursing, physician, pharmacist, clinical staff, and hospital administration pride and gratification. Setting up a new program to provide a 4-hour patient encounter to infuse therapy to high-risk patients with COVID-19 requires commitment and effort. This article describes the experience, ideas, and formula others may consider using to set up such a program. Through networking and formal phone calls and meetings about monoclonal antibody therapy, we have heard about other institutions who have not been able to institute this program due to various barriers to implementation. We hope our experience serves as a resource for others to provide this therapy to their patients and expand access in an effort to mitigate COVID-19 consequences and cases affecting our communities.

Corresponding author: Kathleen Jodoin, PharmD, BCPS, Mount Sinai Medical Center, 4300 Alton Rd, Miami Beach, FL 33140; [email protected].

Financial disclosures: None.

From Baylor College of Medicine, Houston, TX (Drs. Spezia-Lindner, Montealegre, Muldrew, and Suarez) and Harris Health System, Houston, TX (Shanna L. Harris, Maria Daheri, and Drs. Muldrew and Suarez).

Abstract

Objective: To characterize and analyze the prevalence, indications for, and outcomes of fecal immunochemical testing (FIT) in acute patient care within a safety net health care system’s emergency departments (EDs) and inpatient settings.

Design: Retrospective cohort study derived from administrative data.

Setting: A large, urban, safety net health care delivery system in Texas. The data gathered were from the health care system’s 2 primary hospitals and their associated EDs. This health care system utilizes FIT exclusively for fecal occult blood testing.

Participants: Adults ≥18 years who underwent FIT in the ED or inpatient setting between August 2016 and March 2017. Chart review abstractions were performed on a sample (n = 382) from the larger subset.

Measurements: Primary data points included total FITs performed in acute patient care during the study period, basic demographic data, FIT indications, FIT result, receipt of invasive diagnostic follow-up, and result of invasive diagnostic follow-up. Multivariable log-binomial regression was used to calculate risk ratios (RRs) to assess the association between FIT result and receipt of diagnostic follow-up. Chi-square analysis was used to compare the proportion of abnormal findings on diagnostic follow-up by FIT result.

Results: During the 8-month study period, 2718 FITs were performed in the ED and inpatient setting, comprising 5.7% of system-wide FITs. Of the 382 patients included in the chart review who underwent acute care FIT, a majority had their test performed in the ED (304, 79.6%), 133 of which were positive (34.8%). The most common indication for FIT was evidence of overt gastrointestinal (GI) bleed (207, 54.2%), followed by anemia (84, 22.0%). While a positive FIT result was significantly associated with obtaining a diagnostic exam in multivariate analysis (RR, 1.72; P < 0.001), having signs of overt GI bleeding was a stronger predictor of diagnostic follow-up (RR, 2.00; P = 0.003). Of patients who underwent FIT and received diagnostic follow-up (n = 110), 48.2% were FIT negative. These patients were just as likely to have an abnormal finding as FIT-positive patients (90.6% vs 91.2%; P = 0.86). Of the 382 patients in the study, 4 (1.0%) were subsequently diagnosed with colorectal cancer (CRC). Of those 4 patients, 1 (25%) was FIT positive.

Conclusion: FIT is being utilized in acute patient care outside of its established indication for CRC screening in asymptomatic, average-risk adults. Our study demonstrates that FIT is not useful in acute patient care.

Keywords: FOBT; FIT; fecal immunochemical testing; inpatient.

Despite the exclusive validation of FOBTs for use in CRC screening, studies have demonstrated that they are commonly used for a multitude of additional indications in emergency department (ED) and inpatient settings, most aimed at detecting or confirming GI blood loss. This may lead to inappropriate patient management, including the receipt of unnecessary follow-up procedures, which can incur significant costs to the patient and the health system.^13-19 These costs may be particularly burdensome in safety net health systems (ie, those that offer access to care regardless of the patient’s ability to pay), which serve a large proportion of socioeconomically disadvantaged individuals in the United States.^20,21 To our knowledge, no published study to date has specifically investigated the role of FIT in acute patient management.

This study characterizes the use of FIT in acute patient care within a large, urban, safety net health care system. Through a retrospective review of administrative data and patient charts, we evaluated FIT use prevalence, indications, and patient outcomes in the ED and inpatient settings.

Methods

Setting

This study was conducted in a large, urban, county-based integrated delivery system in Houston, Texas, that provides health care services to one of the largest uninsured and underinsured populations in the country.²² The health system includes 2 main hospitals and more than 20 ambulatory care clinics. Within its ambulatory care clinics, the health system implements a population-based screening strategy using stool-based testing. All adults aged 50 years or older who are due for FIT are identified through the health-maintenance module of the electronic medical record (EMR) and offered a take-home FIT. The health system utilizes FIT exclusively (OC-Light S FIT, Polymedco, Cortlandt Manor, NY); no guaiac-based assays are available.

Design and Data Collection

We began by using administrative records to determine the proportion of FITs conducted health system-wide that were ordered and completed in the acute care setting over the study period (August 2016-March 2017). Specifically, we used aggregate quality metric reports, which quantify the number of FITs conducted at each health system clinic and hospital each month, to calculate the proportion of FITs done in the ED and inpatient hospital setting.

We then conducted a retrospective cohort study of 382 adult patients who received FIT in the EDs and inpatient wards in both of the health system’s hospitals over the study period. All data were collected by retrospective chart review in Epic (Madison, WI) EMRs. Sampling was performed by selecting the medical record numbers corresponding to the first 50 completed FITs chronologically each month over the 8-month period, with a total of 400 charts reviewed.

Data collected included basic patient demographics, location of FIT ordering (ED vs inpatient), primary service ordering FIT, FIT indication, FIT result, and receipt and results of invasive diagnostic follow-up. Demographics collected included age, biological sex, race (self-selected), and insurance coverage.

FIT indication was determined based on resident or attending physician notes. The history of present illness, physical exam, and assessment and plan section of notes were reviewed by the lead author for a specific statement of indication for FIT or for evidence of clinical presentation for which FIT could reasonably be ordered. Indications were iteratively reviewed and collapsed into 6 different categories: anemia, iron deficiency with or without anemia, overt GIB, suspected GIB/miscellaneous, non-bloody diarrhea, and no indication identified. Overt GIB was defined as reported or witnessed hematemesis, coffee-ground emesis, hematochezia, bright red blood per rectum, or melena irrespective of time frame (current or remote) or chronicity (acute, subacute, or chronic). In cases where signs of overt bleed were not witnessed by medical professionals, determination of conditions such as melena or coffee-ground emesis were made based on health care providers’ assessment of patient history as documented in his or her notes. Suspected GIB/miscellaneous was defined with the following parameters: any new drop in hemoglobin, abdominal pain, anorectal pain, non-bloody vomiting, hemoptysis, isolated rising blood urea nitrogen, or patient noticing blood on self, clothing, or in the commode without an identified source. Patients who were anemic and found to have iron deficiency on recent lab studies (within 6 months) were reflexively categorized into iron deficiency with or without anemia as opposed to the “anemia” category, which was comprised of any anemia without recent iron studies or non-iron deficient anemia. FIT result was determined by test result entry in Epic, with results either reading positive or negative.

Diagnostic follow-up, for our purposes, was defined as receipt of an invasive procedure or surgery, including esophagogastroduodenoscopy (EGD), colonoscopy, flexible sigmoidoscopy, diagnostic and/or therapeutic abdominal surgical intervention, or any combination of these. Results of diagnostic follow-up were coded as normal or abnormal. A normal result was determined if all procedures performed were listed as normal or as “no pathological findings” on the operative or endoscopic report. Any reported pathologic findings on the operative/endoscopic report were coded as abnormal.

Statistical Analysis

Proportions were used to describe demographic characteristics of patients who received a FIT in acute hospital settings. Bivariable tables and Chi-square tests were used to compare indications and outcomes for FIT-positive and FIT-negative patients. The association between receipt of an invasive diagnostic follow-up (outcome) and the results of an inpatient FIT (predictor) was assessed using multivariable log-binomial regression to calculate risk ratios (RRs) and corresponding 95% confidence intervals. Log-binomial regression was used over logistic regression given that adjusted odds ratios generated by logistic regression often overestimate the association between the risk factor and the outcome when the outcome is common,²³ as in the case of diagnostic follow-up. The model was adjusted for variables selected a priori, specifically, age, gender, and FIT indication. Chi-square analysis was used to compare the proportion of abnormal findings on diagnostic follow-up by FIT result (negative vs positive).

Results

During the 8-month study period, there were 2718 FITs ordered and completed in the acute care setting, compared to 44,662 FITs ordered and completed in the outpatient setting (5.7% performed during acute care).

Among the 400 charts reviewed, 7 were excluded from the analysis because they were duplicates from the same patient, and 11 were excluded due to insufficient information in the patient’s medical record, resulting in 382 patients included in the analysis. Patient demographic characteristics are described in Table 1. Patients were predominantly Hispanic/Latino or Black/African American (51.0% and 32.5%, respectively), a majority had insurance through the county health system (50.5%), and most were male (58.1%). The average age of those receiving FIT was 52 years (standard deviation, 14.8 years), with 40.8% being under the age of 50. For a majority of patients, FIT was ordered in the ED by emergency medicine providers (79.8%). The remaining FITs were ordered by providers in 12 different inpatient departments. Of the FITs ordered, 35.1% were positive.

Indications for ordering FIT are listed in Table 2. The largest proportion of FITs were ordered for overt signs of GIB (54.2%), followed by anemia (22.0%), suspected GIB/miscellaneous reasons (12.3%), iron deficiency with or without anemia (7.6%), and non-bloody diarrhea (2.1%). In 1.8% of cases, no indication for FIT was found in the EMR. No FITs were ordered for the indication of CRC detection. Of these indication categories, overt GIB yielded the highest percentage of FIT positive results (44.0%), and non-bloody diarrhea yielded the lowest (0%).

A total of 110 patients (28.7%) underwent FIT and received invasive diagnostic follow-up. Of these 110 patients, 57 (51.8%) underwent EGD (2 of whom had further surgical intervention), 21 (19.1%) underwent colonoscopy (1 of whom had further surgical intervention), 25 (22.7%) underwent dual EGD and colonoscopy, 1 (0.9%) underwent flexible sigmoidoscopy, and 6 (5.5%) directly underwent abdominal surgical intervention. There was a significantly higher rate of diagnostic follow-up for FIT-positive vs FIT-negative patients (42.9% vs 21.3%; P < 0.001). However, of the 110 patients who underwent subsequent diagnostic follow-up, 48.2% were FIT negative. FIT-negative patients who received diagnostic follow-up were just as likely to have an abnormal finding as FIT-positive patients (90.6% vs 91.2%; P = 0.86).

Of the 382 patients in the study, 4 were diagnosed with CRC through diagnostic follow-up (1.0%). Of those 4 patients, 1 was FIT positive.

The results of the multivariable analyses to evaluate predictors of diagnostic colonoscopy are described in Table 3. Variables in the final model were FITresult, age, and FIT indication. After adjusting for other variables in the model, receipt of diagnostic follow-up was significantly associated with having a positive FIT (adjusted RR, 1.72; P < 0.001) and an overt GIB as an indication (adjusted RR, 2.00; P < 0.01).

Discussion

During the time frame of our study, 5.7% of all FITs ordered within our health system were ordered in the acute patient care setting at our hospitals. The most common indication was overt GIB, which was the indication for 54.2% of patients. Of note, none of the FITs ordered in the acute patient care setting were ordered for CRC screening. These findings support the evidence in the literature that stool-based screening tests, including FIT, are commonly used in US health care systems for diagnostic purposes and risk stratification in acute patient care to detect GIBs.^13-18

Our data suggest that FIT was not a clinically useful test in determining a patient’s need for diagnostic follow-up. While having a positive FIT was significantly associated with obtaining a diagnostic exam in multivariate analysis (RR, 1.72), having signs of overt GI bleeding was a stronger predictor of diagnostic follow-up (RR, 2.00). This salient finding is evidence that a thorough clinical history and physical exam may more strongly predict whether a patient will undergo endoscopy or other follow-up than a FIT result. These findings support other studies in the literature that have called into question the utility of FOBTs in these acute settings.^13-19 Under such circumstances, FOBTs have been shown to rarely influence patient management and thus represent an unnecessary expense.^13–17 Additionally, in some cases, FOBT use in these settings may negatively affect patient outcomes. Such adverse effects include delaying treatment until results are returned or obfuscating indicated management with the results (eg, a patient with indications for colonoscopy not being referred due to a negative FOBT).^13,14,17

We found that, for patients who subsequently went on to have diagnostic follow-up (most commonly endoscopy), there was no difference in the likelihood of FIT-positive and FIT-negative patients to have an abnormality discovered (91.2% vs 90.6%; P = 0.86). This analysis demonstrates no post-hoc support for FIT positivity as a predictor of presence of pathology in patients who were discriminately selected for diagnostic follow-up on clinical grounds by gastroenterologists and surgeons. It does, however, further support that clinical judgment about the need for diagnostic follow-up—irrespective of FIT result—has a very high yield for discovery of pathology in the acute setting.

There are multiple reasons why FOBTs, and specifically FIT, contribute little in management decisions for patients with suspected GI blood loss. Use of FIT raises concern for both false-negatives and false-positives when used outside of its indication. Regarding false- negatives, FIT is an unreliable test for detection of blood loss from the upper GI tract. As FITs utilize antibodies to detect the presence of globin, a byproduct of red blood cell breakdown, it is expected that FIT would fail to detect many cases of upper GI bleeding, as globin is broken down in the upper GI tract.²⁴ This fact is part of what has made FIT a more effective CRC screening test than its guaiac-based counterparts—it has greater specificity for lower GI tract blood loss compared to tests relying on detection of heme.⁸ While guaiac-based assays like Hemoccult have also been shown to be poor tests in acute patient care, they may more frequently, though still unreliably, detect blood of upper GI origin. We believe that part of the ongoing use of FIT in patients with a suspected upper GIB may be from lack of understanding among providers on the mechanistic difference between gFOBTs and FITs, even though gFOBTs also yield highly unreliable results.

FIT does not have the same risk of false-positive results that guaiac-based tests have, which can yield positive results with extra-intestinal blood ingestion, aspirin, or alcohol use; insignificant GI bleeding; and consumption of peroxidase-containing foods.^13,17,25 However, from a clinical standpoint, there are several scenarios of insignificant bleeding that would yield a positive FIT result, such as hemorrhoids, which are common in the US population.^26,27 Additionally, in the ED, where most FITs were performed in our study, it is possible that samples for FITs are being obtained via digital rectal exam (DRE) given patients’ acuity of medical conditions and time constraints. However, FIT has been validated when using a formed stool sample. Obtaining FIT via DRE may lead to microtrauma to the rectum, which could hypothetically yield a positive FIT.

Strengths of this study include its use of in-depth chart data on a large number of FIT-positive patients, which allowed us to discern indications, outcomes, and other clinical data that may have influenced clinical decision-making. Additionally, whereas other studies that address FOBT use in acute patient care have focused on guaiac-based assays, our findings regarding the lack of utility of FIT are novel and have particular relevance as FITs continue to grow in popularity. Nonetheless, there are certain limitations future research should seek to address. In this study, the diagnostic follow-up result was coded by presence or absence of pathologic findings but did not qualify findings by severity or attempt to determine whether the pathology noted on diagnostic follow-up was the definitive source of the suspected GI bleed. These variables could help determine whether there was a difference in severity of bleeding between FIT-positive and FIT-negative patients and could potentially be studied with a prospective research design. Our own study was not designed to address the question of whether FIT result informs patient management decisions. To answer this directly, interviews would have to be conducted with those making the follow-up decision (ie, endoscopists and surgeons). Additionally, this study was not adequately powered to make determinations on the efficacy of FIT in the acute care setting for detection of CRC. As mentioned, only 1 of the 4 patients (25%) who went on to be diagnosed with CRC on follow-up was initially FIT-positive. This would require further investigation.

Conclusion

FIT is being utilized for diagnostic purposes in the acute care of symptomatic patients, which is a misuse of an established screening test for CRC. While our study was not designed to answer whether and how often a FIT result informs subsequent patient management, our results indicate that FIT is an ineffective diagnostic and risk-stratification tool when used in the acute care setting. Our findings add to existing evidence that indicates FOBTs should not be used in acute patient care.

Taken as a whole, the results of our study add to a growing body of evidence demonstrating no role for FOBTs, and specifically FIT, in acute patient care. In light of this evidence, some health care systems have already demonstrated success with system-wide disinvestment from the test in acute patient care settings, with one group publishing about their disinvestment process.²⁸ After completion of our study, our preliminary data were presented to leadership from the internal medicine, emergency medicine, and laboratory divisions within our health care delivery system to galvanize complete disinvestment of FIT from acute care at our hospitals, a policy that was put into effect in July 2019.

Corresponding author: Nathaniel J. Spezia-Lindner, MD, Baylor College of Medicine, 7200 Cambridge St, BCM 903, Ste A10.197, Houston, TX 77030; [email protected].

Financial disclosures: None.

Funding: Cancer Prevention and Research Institute of Texas, CPRIT (PP170094, PDs: ML Jibaja-Weiss and JR Montealegre).

From Baylor College of Medicine, Houston, TX (Drs. Spezia-Lindner, Montealegre, Muldrew, and Suarez) and Harris Health System, Houston, TX (Shanna L. Harris, Maria Daheri, and Drs. Muldrew and Suarez).

Abstract

Objective: To characterize and analyze the prevalence, indications for, and outcomes of fecal immunochemical testing (FIT) in acute patient care within a safety net health care system’s emergency departments (EDs) and inpatient settings.

Design: Retrospective cohort study derived from administrative data.

Setting: A large, urban, safety net health care delivery system in Texas. The data gathered were from the health care system’s 2 primary hospitals and their associated EDs. This health care system utilizes FIT exclusively for fecal occult blood testing.

Participants: Adults ≥18 years who underwent FIT in the ED or inpatient setting between August 2016 and March 2017. Chart review abstractions were performed on a sample (n = 382) from the larger subset.

Measurements: Primary data points included total FITs performed in acute patient care during the study period, basic demographic data, FIT indications, FIT result, receipt of invasive diagnostic follow-up, and result of invasive diagnostic follow-up. Multivariable log-binomial regression was used to calculate risk ratios (RRs) to assess the association between FIT result and receipt of diagnostic follow-up. Chi-square analysis was used to compare the proportion of abnormal findings on diagnostic follow-up by FIT result.

Results: During the 8-month study period, 2718 FITs were performed in the ED and inpatient setting, comprising 5.7% of system-wide FITs. Of the 382 patients included in the chart review who underwent acute care FIT, a majority had their test performed in the ED (304, 79.6%), 133 of which were positive (34.8%). The most common indication for FIT was evidence of overt gastrointestinal (GI) bleed (207, 54.2%), followed by anemia (84, 22.0%). While a positive FIT result was significantly associated with obtaining a diagnostic exam in multivariate analysis (RR, 1.72; P < 0.001), having signs of overt GI bleeding was a stronger predictor of diagnostic follow-up (RR, 2.00; P = 0.003). Of patients who underwent FIT and received diagnostic follow-up (n = 110), 48.2% were FIT negative. These patients were just as likely to have an abnormal finding as FIT-positive patients (90.6% vs 91.2%; P = 0.86). Of the 382 patients in the study, 4 (1.0%) were subsequently diagnosed with colorectal cancer (CRC). Of those 4 patients, 1 (25%) was FIT positive.

Conclusion: FIT is being utilized in acute patient care outside of its established indication for CRC screening in asymptomatic, average-risk adults. Our study demonstrates that FIT is not useful in acute patient care.

Keywords: FOBT; FIT; fecal immunochemical testing; inpatient.

Despite the exclusive validation of FOBTs for use in CRC screening, studies have demonstrated that they are commonly used for a multitude of additional indications in emergency department (ED) and inpatient settings, most aimed at detecting or confirming GI blood loss. This may lead to inappropriate patient management, including the receipt of unnecessary follow-up procedures, which can incur significant costs to the patient and the health system.^13-19 These costs may be particularly burdensome in safety net health systems (ie, those that offer access to care regardless of the patient’s ability to pay), which serve a large proportion of socioeconomically disadvantaged individuals in the United States.^20,21 To our knowledge, no published study to date has specifically investigated the role of FIT in acute patient management.

This study characterizes the use of FIT in acute patient care within a large, urban, safety net health care system. Through a retrospective review of administrative data and patient charts, we evaluated FIT use prevalence, indications, and patient outcomes in the ED and inpatient settings.

Methods

Setting

This study was conducted in a large, urban, county-based integrated delivery system in Houston, Texas, that provides health care services to one of the largest uninsured and underinsured populations in the country.²² The health system includes 2 main hospitals and more than 20 ambulatory care clinics. Within its ambulatory care clinics, the health system implements a population-based screening strategy using stool-based testing. All adults aged 50 years or older who are due for FIT are identified through the health-maintenance module of the electronic medical record (EMR) and offered a take-home FIT. The health system utilizes FIT exclusively (OC-Light S FIT, Polymedco, Cortlandt Manor, NY); no guaiac-based assays are available.

Design and Data Collection

We began by using administrative records to determine the proportion of FITs conducted health system-wide that were ordered and completed in the acute care setting over the study period (August 2016-March 2017). Specifically, we used aggregate quality metric reports, which quantify the number of FITs conducted at each health system clinic and hospital each month, to calculate the proportion of FITs done in the ED and inpatient hospital setting.

We then conducted a retrospective cohort study of 382 adult patients who received FIT in the EDs and inpatient wards in both of the health system’s hospitals over the study period. All data were collected by retrospective chart review in Epic (Madison, WI) EMRs. Sampling was performed by selecting the medical record numbers corresponding to the first 50 completed FITs chronologically each month over the 8-month period, with a total of 400 charts reviewed.

Data collected included basic patient demographics, location of FIT ordering (ED vs inpatient), primary service ordering FIT, FIT indication, FIT result, and receipt and results of invasive diagnostic follow-up. Demographics collected included age, biological sex, race (self-selected), and insurance coverage.

FIT indication was determined based on resident or attending physician notes. The history of present illness, physical exam, and assessment and plan section of notes were reviewed by the lead author for a specific statement of indication for FIT or for evidence of clinical presentation for which FIT could reasonably be ordered. Indications were iteratively reviewed and collapsed into 6 different categories: anemia, iron deficiency with or without anemia, overt GIB, suspected GIB/miscellaneous, non-bloody diarrhea, and no indication identified. Overt GIB was defined as reported or witnessed hematemesis, coffee-ground emesis, hematochezia, bright red blood per rectum, or melena irrespective of time frame (current or remote) or chronicity (acute, subacute, or chronic). In cases where signs of overt bleed were not witnessed by medical professionals, determination of conditions such as melena or coffee-ground emesis were made based on health care providers’ assessment of patient history as documented in his or her notes. Suspected GIB/miscellaneous was defined with the following parameters: any new drop in hemoglobin, abdominal pain, anorectal pain, non-bloody vomiting, hemoptysis, isolated rising blood urea nitrogen, or patient noticing blood on self, clothing, or in the commode without an identified source. Patients who were anemic and found to have iron deficiency on recent lab studies (within 6 months) were reflexively categorized into iron deficiency with or without anemia as opposed to the “anemia” category, which was comprised of any anemia without recent iron studies or non-iron deficient anemia. FIT result was determined by test result entry in Epic, with results either reading positive or negative.

Diagnostic follow-up, for our purposes, was defined as receipt of an invasive procedure or surgery, including esophagogastroduodenoscopy (EGD), colonoscopy, flexible sigmoidoscopy, diagnostic and/or therapeutic abdominal surgical intervention, or any combination of these. Results of diagnostic follow-up were coded as normal or abnormal. A normal result was determined if all procedures performed were listed as normal or as “no pathological findings” on the operative or endoscopic report. Any reported pathologic findings on the operative/endoscopic report were coded as abnormal.

Statistical Analysis

Proportions were used to describe demographic characteristics of patients who received a FIT in acute hospital settings. Bivariable tables and Chi-square tests were used to compare indications and outcomes for FIT-positive and FIT-negative patients. The association between receipt of an invasive diagnostic follow-up (outcome) and the results of an inpatient FIT (predictor) was assessed using multivariable log-binomial regression to calculate risk ratios (RRs) and corresponding 95% confidence intervals. Log-binomial regression was used over logistic regression given that adjusted odds ratios generated by logistic regression often overestimate the association between the risk factor and the outcome when the outcome is common,²³ as in the case of diagnostic follow-up. The model was adjusted for variables selected a priori, specifically, age, gender, and FIT indication. Chi-square analysis was used to compare the proportion of abnormal findings on diagnostic follow-up by FIT result (negative vs positive).

Results

During the 8-month study period, there were 2718 FITs ordered and completed in the acute care setting, compared to 44,662 FITs ordered and completed in the outpatient setting (5.7% performed during acute care).

Among the 400 charts reviewed, 7 were excluded from the analysis because they were duplicates from the same patient, and 11 were excluded due to insufficient information in the patient’s medical record, resulting in 382 patients included in the analysis. Patient demographic characteristics are described in Table 1. Patients were predominantly Hispanic/Latino or Black/African American (51.0% and 32.5%, respectively), a majority had insurance through the county health system (50.5%), and most were male (58.1%). The average age of those receiving FIT was 52 years (standard deviation, 14.8 years), with 40.8% being under the age of 50. For a majority of patients, FIT was ordered in the ED by emergency medicine providers (79.8%). The remaining FITs were ordered by providers in 12 different inpatient departments. Of the FITs ordered, 35.1% were positive.

Indications for ordering FIT are listed in Table 2. The largest proportion of FITs were ordered for overt signs of GIB (54.2%), followed by anemia (22.0%), suspected GIB/miscellaneous reasons (12.3%), iron deficiency with or without anemia (7.6%), and non-bloody diarrhea (2.1%). In 1.8% of cases, no indication for FIT was found in the EMR. No FITs were ordered for the indication of CRC detection. Of these indication categories, overt GIB yielded the highest percentage of FIT positive results (44.0%), and non-bloody diarrhea yielded the lowest (0%).

A total of 110 patients (28.7%) underwent FIT and received invasive diagnostic follow-up. Of these 110 patients, 57 (51.8%) underwent EGD (2 of whom had further surgical intervention), 21 (19.1%) underwent colonoscopy (1 of whom had further surgical intervention), 25 (22.7%) underwent dual EGD and colonoscopy, 1 (0.9%) underwent flexible sigmoidoscopy, and 6 (5.5%) directly underwent abdominal surgical intervention. There was a significantly higher rate of diagnostic follow-up for FIT-positive vs FIT-negative patients (42.9% vs 21.3%; P < 0.001). However, of the 110 patients who underwent subsequent diagnostic follow-up, 48.2% were FIT negative. FIT-negative patients who received diagnostic follow-up were just as likely to have an abnormal finding as FIT-positive patients (90.6% vs 91.2%; P = 0.86).

Of the 382 patients in the study, 4 were diagnosed with CRC through diagnostic follow-up (1.0%). Of those 4 patients, 1 was FIT positive.

The results of the multivariable analyses to evaluate predictors of diagnostic colonoscopy are described in Table 3. Variables in the final model were FITresult, age, and FIT indication. After adjusting for other variables in the model, receipt of diagnostic follow-up was significantly associated with having a positive FIT (adjusted RR, 1.72; P < 0.001) and an overt GIB as an indication (adjusted RR, 2.00; P < 0.01).

Discussion

During the time frame of our study, 5.7% of all FITs ordered within our health system were ordered in the acute patient care setting at our hospitals. The most common indication was overt GIB, which was the indication for 54.2% of patients. Of note, none of the FITs ordered in the acute patient care setting were ordered for CRC screening. These findings support the evidence in the literature that stool-based screening tests, including FIT, are commonly used in US health care systems for diagnostic purposes and risk stratification in acute patient care to detect GIBs.^13-18

Our data suggest that FIT was not a clinically useful test in determining a patient’s need for diagnostic follow-up. While having a positive FIT was significantly associated with obtaining a diagnostic exam in multivariate analysis (RR, 1.72), having signs of overt GI bleeding was a stronger predictor of diagnostic follow-up (RR, 2.00). This salient finding is evidence that a thorough clinical history and physical exam may more strongly predict whether a patient will undergo endoscopy or other follow-up than a FIT result. These findings support other studies in the literature that have called into question the utility of FOBTs in these acute settings.^13-19 Under such circumstances, FOBTs have been shown to rarely influence patient management and thus represent an unnecessary expense.^13–17 Additionally, in some cases, FOBT use in these settings may negatively affect patient outcomes. Such adverse effects include delaying treatment until results are returned or obfuscating indicated management with the results (eg, a patient with indications for colonoscopy not being referred due to a negative FOBT).^13,14,17

We found that, for patients who subsequently went on to have diagnostic follow-up (most commonly endoscopy), there was no difference in the likelihood of FIT-positive and FIT-negative patients to have an abnormality discovered (91.2% vs 90.6%; P = 0.86). This analysis demonstrates no post-hoc support for FIT positivity as a predictor of presence of pathology in patients who were discriminately selected for diagnostic follow-up on clinical grounds by gastroenterologists and surgeons. It does, however, further support that clinical judgment about the need for diagnostic follow-up—irrespective of FIT result—has a very high yield for discovery of pathology in the acute setting.

There are multiple reasons why FOBTs, and specifically FIT, contribute little in management decisions for patients with suspected GI blood loss. Use of FIT raises concern for both false-negatives and false-positives when used outside of its indication. Regarding false- negatives, FIT is an unreliable test for detection of blood loss from the upper GI tract. As FITs utilize antibodies to detect the presence of globin, a byproduct of red blood cell breakdown, it is expected that FIT would fail to detect many cases of upper GI bleeding, as globin is broken down in the upper GI tract.²⁴ This fact is part of what has made FIT a more effective CRC screening test than its guaiac-based counterparts—it has greater specificity for lower GI tract blood loss compared to tests relying on detection of heme.⁸ While guaiac-based assays like Hemoccult have also been shown to be poor tests in acute patient care, they may more frequently, though still unreliably, detect blood of upper GI origin. We believe that part of the ongoing use of FIT in patients with a suspected upper GIB may be from lack of understanding among providers on the mechanistic difference between gFOBTs and FITs, even though gFOBTs also yield highly unreliable results.

FIT does not have the same risk of false-positive results that guaiac-based tests have, which can yield positive results with extra-intestinal blood ingestion, aspirin, or alcohol use; insignificant GI bleeding; and consumption of peroxidase-containing foods.^13,17,25 However, from a clinical standpoint, there are several scenarios of insignificant bleeding that would yield a positive FIT result, such as hemorrhoids, which are common in the US population.^26,27 Additionally, in the ED, where most FITs were performed in our study, it is possible that samples for FITs are being obtained via digital rectal exam (DRE) given patients’ acuity of medical conditions and time constraints. However, FIT has been validated when using a formed stool sample. Obtaining FIT via DRE may lead to microtrauma to the rectum, which could hypothetically yield a positive FIT.

Strengths of this study include its use of in-depth chart data on a large number of FIT-positive patients, which allowed us to discern indications, outcomes, and other clinical data that may have influenced clinical decision-making. Additionally, whereas other studies that address FOBT use in acute patient care have focused on guaiac-based assays, our findings regarding the lack of utility of FIT are novel and have particular relevance as FITs continue to grow in popularity. Nonetheless, there are certain limitations future research should seek to address. In this study, the diagnostic follow-up result was coded by presence or absence of pathologic findings but did not qualify findings by severity or attempt to determine whether the pathology noted on diagnostic follow-up was the definitive source of the suspected GI bleed. These variables could help determine whether there was a difference in severity of bleeding between FIT-positive and FIT-negative patients and could potentially be studied with a prospective research design. Our own study was not designed to address the question of whether FIT result informs patient management decisions. To answer this directly, interviews would have to be conducted with those making the follow-up decision (ie, endoscopists and surgeons). Additionally, this study was not adequately powered to make determinations on the efficacy of FIT in the acute care setting for detection of CRC. As mentioned, only 1 of the 4 patients (25%) who went on to be diagnosed with CRC on follow-up was initially FIT-positive. This would require further investigation.

Conclusion

FIT is being utilized for diagnostic purposes in the acute care of symptomatic patients, which is a misuse of an established screening test for CRC. While our study was not designed to answer whether and how often a FIT result informs subsequent patient management, our results indicate that FIT is an ineffective diagnostic and risk-stratification tool when used in the acute care setting. Our findings add to existing evidence that indicates FOBTs should not be used in acute patient care.

Taken as a whole, the results of our study add to a growing body of evidence demonstrating no role for FOBTs, and specifically FIT, in acute patient care. In light of this evidence, some health care systems have already demonstrated success with system-wide disinvestment from the test in acute patient care settings, with one group publishing about their disinvestment process.²⁸ After completion of our study, our preliminary data were presented to leadership from the internal medicine, emergency medicine, and laboratory divisions within our health care delivery system to galvanize complete disinvestment of FIT from acute care at our hospitals, a policy that was put into effect in July 2019.

Corresponding author: Nathaniel J. Spezia-Lindner, MD, Baylor College of Medicine, 7200 Cambridge St, BCM 903, Ste A10.197, Houston, TX 77030; [email protected].

Financial disclosures: None.

Funding: Cancer Prevention and Research Institute of Texas, CPRIT (PP170094, PDs: ML Jibaja-Weiss and JR Montealegre).

From Baylor College of Medicine, Houston, TX (Drs. Spezia-Lindner, Montealegre, Muldrew, and Suarez) and Harris Health System, Houston, TX (Shanna L. Harris, Maria Daheri, and Drs. Muldrew and Suarez).

Abstract

Objective: To characterize and analyze the prevalence, indications for, and outcomes of fecal immunochemical testing (FIT) in acute patient care within a safety net health care system’s emergency departments (EDs) and inpatient settings.

Design: Retrospective cohort study derived from administrative data.

Setting: A large, urban, safety net health care delivery system in Texas. The data gathered were from the health care system’s 2 primary hospitals and their associated EDs. This health care system utilizes FIT exclusively for fecal occult blood testing.

Participants: Adults ≥18 years who underwent FIT in the ED or inpatient setting between August 2016 and March 2017. Chart review abstractions were performed on a sample (n = 382) from the larger subset.

Measurements: Primary data points included total FITs performed in acute patient care during the study period, basic demographic data, FIT indications, FIT result, receipt of invasive diagnostic follow-up, and result of invasive diagnostic follow-up. Multivariable log-binomial regression was used to calculate risk ratios (RRs) to assess the association between FIT result and receipt of diagnostic follow-up. Chi-square analysis was used to compare the proportion of abnormal findings on diagnostic follow-up by FIT result.

Results: During the 8-month study period, 2718 FITs were performed in the ED and inpatient setting, comprising 5.7% of system-wide FITs. Of the 382 patients included in the chart review who underwent acute care FIT, a majority had their test performed in the ED (304, 79.6%), 133 of which were positive (34.8%). The most common indication for FIT was evidence of overt gastrointestinal (GI) bleed (207, 54.2%), followed by anemia (84, 22.0%). While a positive FIT result was significantly associated with obtaining a diagnostic exam in multivariate analysis (RR, 1.72; P < 0.001), having signs of overt GI bleeding was a stronger predictor of diagnostic follow-up (RR, 2.00; P = 0.003). Of patients who underwent FIT and received diagnostic follow-up (n = 110), 48.2% were FIT negative. These patients were just as likely to have an abnormal finding as FIT-positive patients (90.6% vs 91.2%; P = 0.86). Of the 382 patients in the study, 4 (1.0%) were subsequently diagnosed with colorectal cancer (CRC). Of those 4 patients, 1 (25%) was FIT positive.

Conclusion: FIT is being utilized in acute patient care outside of its established indication for CRC screening in asymptomatic, average-risk adults. Our study demonstrates that FIT is not useful in acute patient care.

Keywords: FOBT; FIT; fecal immunochemical testing; inpatient.

Despite the exclusive validation of FOBTs for use in CRC screening, studies have demonstrated that they are commonly used for a multitude of additional indications in emergency department (ED) and inpatient settings, most aimed at detecting or confirming GI blood loss. This may lead to inappropriate patient management, including the receipt of unnecessary follow-up procedures, which can incur significant costs to the patient and the health system.^13-19 These costs may be particularly burdensome in safety net health systems (ie, those that offer access to care regardless of the patient’s ability to pay), which serve a large proportion of socioeconomically disadvantaged individuals in the United States.^20,21 To our knowledge, no published study to date has specifically investigated the role of FIT in acute patient management.

This study characterizes the use of FIT in acute patient care within a large, urban, safety net health care system. Through a retrospective review of administrative data and patient charts, we evaluated FIT use prevalence, indications, and patient outcomes in the ED and inpatient settings.

Methods

Setting

This study was conducted in a large, urban, county-based integrated delivery system in Houston, Texas, that provides health care services to one of the largest uninsured and underinsured populations in the country.²² The health system includes 2 main hospitals and more than 20 ambulatory care clinics. Within its ambulatory care clinics, the health system implements a population-based screening strategy using stool-based testing. All adults aged 50 years or older who are due for FIT are identified through the health-maintenance module of the electronic medical record (EMR) and offered a take-home FIT. The health system utilizes FIT exclusively (OC-Light S FIT, Polymedco, Cortlandt Manor, NY); no guaiac-based assays are available.

Design and Data Collection

We began by using administrative records to determine the proportion of FITs conducted health system-wide that were ordered and completed in the acute care setting over the study period (August 2016-March 2017). Specifically, we used aggregate quality metric reports, which quantify the number of FITs conducted at each health system clinic and hospital each month, to calculate the proportion of FITs done in the ED and inpatient hospital setting.

We then conducted a retrospective cohort study of 382 adult patients who received FIT in the EDs and inpatient wards in both of the health system’s hospitals over the study period. All data were collected by retrospective chart review in Epic (Madison, WI) EMRs. Sampling was performed by selecting the medical record numbers corresponding to the first 50 completed FITs chronologically each month over the 8-month period, with a total of 400 charts reviewed.

Data collected included basic patient demographics, location of FIT ordering (ED vs inpatient), primary service ordering FIT, FIT indication, FIT result, and receipt and results of invasive diagnostic follow-up. Demographics collected included age, biological sex, race (self-selected), and insurance coverage.

FIT indication was determined based on resident or attending physician notes. The history of present illness, physical exam, and assessment and plan section of notes were reviewed by the lead author for a specific statement of indication for FIT or for evidence of clinical presentation for which FIT could reasonably be ordered. Indications were iteratively reviewed and collapsed into 6 different categories: anemia, iron deficiency with or without anemia, overt GIB, suspected GIB/miscellaneous, non-bloody diarrhea, and no indication identified. Overt GIB was defined as reported or witnessed hematemesis, coffee-ground emesis, hematochezia, bright red blood per rectum, or melena irrespective of time frame (current or remote) or chronicity (acute, subacute, or chronic). In cases where signs of overt bleed were not witnessed by medical professionals, determination of conditions such as melena or coffee-ground emesis were made based on health care providers’ assessment of patient history as documented in his or her notes. Suspected GIB/miscellaneous was defined with the following parameters: any new drop in hemoglobin, abdominal pain, anorectal pain, non-bloody vomiting, hemoptysis, isolated rising blood urea nitrogen, or patient noticing blood on self, clothing, or in the commode without an identified source. Patients who were anemic and found to have iron deficiency on recent lab studies (within 6 months) were reflexively categorized into iron deficiency with or without anemia as opposed to the “anemia” category, which was comprised of any anemia without recent iron studies or non-iron deficient anemia. FIT result was determined by test result entry in Epic, with results either reading positive or negative.

Diagnostic follow-up, for our purposes, was defined as receipt of an invasive procedure or surgery, including esophagogastroduodenoscopy (EGD), colonoscopy, flexible sigmoidoscopy, diagnostic and/or therapeutic abdominal surgical intervention, or any combination of these. Results of diagnostic follow-up were coded as normal or abnormal. A normal result was determined if all procedures performed were listed as normal or as “no pathological findings” on the operative or endoscopic report. Any reported pathologic findings on the operative/endoscopic report were coded as abnormal.

Statistical Analysis

Proportions were used to describe demographic characteristics of patients who received a FIT in acute hospital settings. Bivariable tables and Chi-square tests were used to compare indications and outcomes for FIT-positive and FIT-negative patients. The association between receipt of an invasive diagnostic follow-up (outcome) and the results of an inpatient FIT (predictor) was assessed using multivariable log-binomial regression to calculate risk ratios (RRs) and corresponding 95% confidence intervals. Log-binomial regression was used over logistic regression given that adjusted odds ratios generated by logistic regression often overestimate the association between the risk factor and the outcome when the outcome is common,²³ as in the case of diagnostic follow-up. The model was adjusted for variables selected a priori, specifically, age, gender, and FIT indication. Chi-square analysis was used to compare the proportion of abnormal findings on diagnostic follow-up by FIT result (negative vs positive).

Results

During the 8-month study period, there were 2718 FITs ordered and completed in the acute care setting, compared to 44,662 FITs ordered and completed in the outpatient setting (5.7% performed during acute care).

Among the 400 charts reviewed, 7 were excluded from the analysis because they were duplicates from the same patient, and 11 were excluded due to insufficient information in the patient’s medical record, resulting in 382 patients included in the analysis. Patient demographic characteristics are described in Table 1. Patients were predominantly Hispanic/Latino or Black/African American (51.0% and 32.5%, respectively), a majority had insurance through the county health system (50.5%), and most were male (58.1%). The average age of those receiving FIT was 52 years (standard deviation, 14.8 years), with 40.8% being under the age of 50. For a majority of patients, FIT was ordered in the ED by emergency medicine providers (79.8%). The remaining FITs were ordered by providers in 12 different inpatient departments. Of the FITs ordered, 35.1% were positive.

Indications for ordering FIT are listed in Table 2. The largest proportion of FITs were ordered for overt signs of GIB (54.2%), followed by anemia (22.0%), suspected GIB/miscellaneous reasons (12.3%), iron deficiency with or without anemia (7.6%), and non-bloody diarrhea (2.1%). In 1.8% of cases, no indication for FIT was found in the EMR. No FITs were ordered for the indication of CRC detection. Of these indication categories, overt GIB yielded the highest percentage of FIT positive results (44.0%), and non-bloody diarrhea yielded the lowest (0%).

A total of 110 patients (28.7%) underwent FIT and received invasive diagnostic follow-up. Of these 110 patients, 57 (51.8%) underwent EGD (2 of whom had further surgical intervention), 21 (19.1%) underwent colonoscopy (1 of whom had further surgical intervention), 25 (22.7%) underwent dual EGD and colonoscopy, 1 (0.9%) underwent flexible sigmoidoscopy, and 6 (5.5%) directly underwent abdominal surgical intervention. There was a significantly higher rate of diagnostic follow-up for FIT-positive vs FIT-negative patients (42.9% vs 21.3%; P < 0.001). However, of the 110 patients who underwent subsequent diagnostic follow-up, 48.2% were FIT negative. FIT-negative patients who received diagnostic follow-up were just as likely to have an abnormal finding as FIT-positive patients (90.6% vs 91.2%; P = 0.86).

Of the 382 patients in the study, 4 were diagnosed with CRC through diagnostic follow-up (1.0%). Of those 4 patients, 1 was FIT positive.

The results of the multivariable analyses to evaluate predictors of diagnostic colonoscopy are described in Table 3. Variables in the final model were FITresult, age, and FIT indication. After adjusting for other variables in the model, receipt of diagnostic follow-up was significantly associated with having a positive FIT (adjusted RR, 1.72; P < 0.001) and an overt GIB as an indication (adjusted RR, 2.00; P < 0.01).

Discussion

During the time frame of our study, 5.7% of all FITs ordered within our health system were ordered in the acute patient care setting at our hospitals. The most common indication was overt GIB, which was the indication for 54.2% of patients. Of note, none of the FITs ordered in the acute patient care setting were ordered for CRC screening. These findings support the evidence in the literature that stool-based screening tests, including FIT, are commonly used in US health care systems for diagnostic purposes and risk stratification in acute patient care to detect GIBs.^13-18

Our data suggest that FIT was not a clinically useful test in determining a patient’s need for diagnostic follow-up. While having a positive FIT was significantly associated with obtaining a diagnostic exam in multivariate analysis (RR, 1.72), having signs of overt GI bleeding was a stronger predictor of diagnostic follow-up (RR, 2.00). This salient finding is evidence that a thorough clinical history and physical exam may more strongly predict whether a patient will undergo endoscopy or other follow-up than a FIT result. These findings support other studies in the literature that have called into question the utility of FOBTs in these acute settings.^13-19 Under such circumstances, FOBTs have been shown to rarely influence patient management and thus represent an unnecessary expense.^13–17 Additionally, in some cases, FOBT use in these settings may negatively affect patient outcomes. Such adverse effects include delaying treatment until results are returned or obfuscating indicated management with the results (eg, a patient with indications for colonoscopy not being referred due to a negative FOBT).^13,14,17

We found that, for patients who subsequently went on to have diagnostic follow-up (most commonly endoscopy), there was no difference in the likelihood of FIT-positive and FIT-negative patients to have an abnormality discovered (91.2% vs 90.6%; P = 0.86). This analysis demonstrates no post-hoc support for FIT positivity as a predictor of presence of pathology in patients who were discriminately selected for diagnostic follow-up on clinical grounds by gastroenterologists and surgeons. It does, however, further support that clinical judgment about the need for diagnostic follow-up—irrespective of FIT result—has a very high yield for discovery of pathology in the acute setting.

There are multiple reasons why FOBTs, and specifically FIT, contribute little in management decisions for patients with suspected GI blood loss. Use of FIT raises concern for both false-negatives and false-positives when used outside of its indication. Regarding false- negatives, FIT is an unreliable test for detection of blood loss from the upper GI tract. As FITs utilize antibodies to detect the presence of globin, a byproduct of red blood cell breakdown, it is expected that FIT would fail to detect many cases of upper GI bleeding, as globin is broken down in the upper GI tract.²⁴ This fact is part of what has made FIT a more effective CRC screening test than its guaiac-based counterparts—it has greater specificity for lower GI tract blood loss compared to tests relying on detection of heme.⁸ While guaiac-based assays like Hemoccult have also been shown to be poor tests in acute patient care, they may more frequently, though still unreliably, detect blood of upper GI origin. We believe that part of the ongoing use of FIT in patients with a suspected upper GIB may be from lack of understanding among providers on the mechanistic difference between gFOBTs and FITs, even though gFOBTs also yield highly unreliable results.

FIT does not have the same risk of false-positive results that guaiac-based tests have, which can yield positive results with extra-intestinal blood ingestion, aspirin, or alcohol use; insignificant GI bleeding; and consumption of peroxidase-containing foods.^13,17,25 However, from a clinical standpoint, there are several scenarios of insignificant bleeding that would yield a positive FIT result, such as hemorrhoids, which are common in the US population.^26,27 Additionally, in the ED, where most FITs were performed in our study, it is possible that samples for FITs are being obtained via digital rectal exam (DRE) given patients’ acuity of medical conditions and time constraints. However, FIT has been validated when using a formed stool sample. Obtaining FIT via DRE may lead to microtrauma to the rectum, which could hypothetically yield a positive FIT.

Strengths of this study include its use of in-depth chart data on a large number of FIT-positive patients, which allowed us to discern indications, outcomes, and other clinical data that may have influenced clinical decision-making. Additionally, whereas other studies that address FOBT use in acute patient care have focused on guaiac-based assays, our findings regarding the lack of utility of FIT are novel and have particular relevance as FITs continue to grow in popularity. Nonetheless, there are certain limitations future research should seek to address. In this study, the diagnostic follow-up result was coded by presence or absence of pathologic findings but did not qualify findings by severity or attempt to determine whether the pathology noted on diagnostic follow-up was the definitive source of the suspected GI bleed. These variables could help determine whether there was a difference in severity of bleeding between FIT-positive and FIT-negative patients and could potentially be studied with a prospective research design. Our own study was not designed to address the question of whether FIT result informs patient management decisions. To answer this directly, interviews would have to be conducted with those making the follow-up decision (ie, endoscopists and surgeons). Additionally, this study was not adequately powered to make determinations on the efficacy of FIT in the acute care setting for detection of CRC. As mentioned, only 1 of the 4 patients (25%) who went on to be diagnosed with CRC on follow-up was initially FIT-positive. This would require further investigation.

Conclusion

FIT is being utilized for diagnostic purposes in the acute care of symptomatic patients, which is a misuse of an established screening test for CRC. While our study was not designed to answer whether and how often a FIT result informs subsequent patient management, our results indicate that FIT is an ineffective diagnostic and risk-stratification tool when used in the acute care setting. Our findings add to existing evidence that indicates FOBTs should not be used in acute patient care.

Taken as a whole, the results of our study add to a growing body of evidence demonstrating no role for FOBTs, and specifically FIT, in acute patient care. In light of this evidence, some health care systems have already demonstrated success with system-wide disinvestment from the test in acute patient care settings, with one group publishing about their disinvestment process.²⁸ After completion of our study, our preliminary data were presented to leadership from the internal medicine, emergency medicine, and laboratory divisions within our health care delivery system to galvanize complete disinvestment of FIT from acute care at our hospitals, a policy that was put into effect in July 2019.

Corresponding author: Nathaniel J. Spezia-Lindner, MD, Baylor College of Medicine, 7200 Cambridge St, BCM 903, Ste A10.197, Houston, TX 77030; [email protected].

Financial disclosures: None.

Funding: Cancer Prevention and Research Institute of Texas, CPRIT (PP170094, PDs: ML Jibaja-Weiss and JR Montealegre).

From The Johns Hopkins Hospital, Baltimore, MD (Dr. Muganlinskaya and Dr. Skojec, retired); The George Washington University, Washington, DC (Dr. Posey); and Johns Hopkins University, Baltimore, MD (Dr. Resar).

Abstract

Objective: Assessing the risk characteristics of patients with acute myocardial infarction (MI) can help providers make appropriate referral decisions. This quality improvement project sought to improve timely, appropriate referrals among patients with type I MI by adding a risk assessment, the AMI READMITS score, to the existing referral protocol.

Methods: Patients’ chart data were analyzed to assess changes in referrals and timely follow-up appointments from pre-intervention to intervention. A survey assessed providers’ satisfaction with the new referral protocol.

Results: Among 57 patients (n = 29 preintervention; n = 28 intervention), documented referrals increased significantly from 66% to 89% (χ2 = 4.571, df = 1, P = 0.033); and timely appointments increased by 10%, which was not significant (χ2 = 3.550, df = 2, P = 0.169). Most providers agreed that the new protocol was easy to use, useful in making referral decisions, and improved the referral process. All agreed the risk score should be incorporated into electronic clinical notes. Provider opinions related to implementing the risk score in clinical practice were mixed. Qualitative feedback suggests this was due to limited validation of the AMI READMITS score in reducing readmissions.

Conclusions: Our risk-based referral protocol helped to increase appropriate referrals among patients with type I MI. Provider adoption may be enhanced by incorporating the protocol into electronic clinical notes. Research to further validate the accuracy of the AMI READMITS score in predicting readmissions may support adoption of the protocol in clinical practice.

Keywords: quality improvement; type I myocardial infarction; referral process; readmission risk; risk assessment; chart review.

Early follow-up after discharge is an important strategy to reduce the risk of unplanned hospital readmissions among patients with various conditions.^1-3 While patient confounding factors, such as chronic health problems, environment, socioeconomic status, and literacy, make it difficult to avoid all unplanned readmissions, early follow-up may help providers identify and appropriately manage some health-related issues, and as such is a pivotal element of a readmission prevention strategy.⁴ There is evidence that patients with non-ST elevation myocardial infarction (NSTEMI) who have an outpatient appointment with a physician within 7 days after discharge have a lower risk of 30-day readmission.⁵

Our hospital’s postmyocardial infarction clinic was created to prevent unplanned readmissions within 30 days after discharge among patients with type I myocardial infarction (MI). Since inception, the number of referrals has been much lower than expected. In 2018, the total number of patients discharged from the hospital with type I MI and any troponin I level above 0.40 ng/mL was 313. Most of these patients were discharged from the hospital’s cardiac units; however, only 91 referrals were made. To increase referrals, the cardiology nurse practitioners (NPs) developed a post-MI referral protocol (Figure 1). However, this protocol was not consistently used and referrals to the clinic remained low.

Evidence-based risk assessment tools have the potential to increase effective patient management. For example, cardiology providers at the hospital utilize various scores, such as CHA₂DS₂-VASc⁶ and the Society of Thoracic Surgery risk score,⁷ to plan patient management. Among the scores used to predict unplanned readmissions for MI patients, the most promising is the AMI READMITS score.⁸ Unlike other nonspecific prediction models, the AMI READMITS score was developed based on variables extracted from the electronic health records (EHRs) of patients who were hospitalized for MI and readmitted within 30 days after discharge. Recognizing the potential to increase referrals by integrating an MI-specific risk assessment, this quality improvement study modified the existing referral protocol to include the patients’ AMI READMITS score and recommendations for follow-up.

Currently, there are no clear recommendations on how soon after discharge patients with MI should undergo follow-up. As research data vary, we selected 7 days follow-up for patients from high risk groups based on the “See you in 7” initiative for patients with heart failure (HF) and MI,^9,10 as well as evidence that patients with NSTEMI have a lower risk of 30-day readmission if they have follow-up within 7 days after discharge⁵; and we selected 14 days follow-up for patients from low-risk groups based on evidence that postdischarge follow-up within 14 days reduces risk of 30-day readmission in patients with acute myocardial infarction (AMI) and/or acutely decompensated HF.¹¹

Methods

This project was designed to answer the following question: For adult patients with type I MI, does implementation of a readmission risk assessment referral protocol increase the percentage of referrals and appointments scheduled within a recommended time? Anticipated outcomes included: (1) increased referrals to a cardiologist or the post-MI clinic; (2) increased scheduled follow-up appointments within 7 to 14 days; (3) provider satisfaction with the usability and usefulness of the new protocol; and (4) consistent provider adoption of the new risk assessment referral protocol.

To evaluate the degree to which these outcomes were achieved, we reviewed patient charts for 2 months prior and 2 months during implementation of the new referral protocol. As shown in Figure 2, the new protocol added the following process steps to the existing protocol: calculation of the AMI READMITS score, recommendations for follow-up based on patients’ risk score, and guidance to refer patients to the post-MI clinic if patients did not have an appointment with a cardiologist within 7 to 14 days after discharge. Patients’ risk assessment scores were obtained from forms completed by clinicians during the intervention. Clinician’s perceptions related to the usability and usefulness of the new protocol and feedback related to its long-term adoption were assessed using a descriptive survey.

Population

The project population included all adult patients (≥ 18 years old) with type I MI who were admitted or transferred to the hospital, had a percutaneous coronary intervention (PCI), or were managed without PCI and discharged from the hospital’s cardiac care unit (CCU) and progressive cardiac care unit (PCCU). The criteria for type I MI included the “detection of a rise and/or fall of cardiac troponin with at least 1 value above the 99th percentile and with at least 1 of the following: symptoms of acute myocardial ischemia; new ischemic electrocardiographic (ECG) changes; development of new pathological Q waves; imaging evidence of new loss of viable myocardium or new regional wall motion abnormality in a pattern consistent with an ischemic etiology; identification of a coronary thrombus by angiography including intracoronary imaging or by autopsy.”¹² The study excluded patients with type I MI who were referred for coronary bypass surgery.

Intervention

The revised risk assessment protocol was implemented within the CCU and PCCU. The lead investigator met with each provider to discuss the role of the post-MI clinic, current referral rates, the purpose of the project, and the new referral process to be completed during the project for each patient discharged with type I MI. Cardiology NPs, fellows, and residents were asked to use the risk-assessment form to calculate patients’ risk for readmission, and refer patients to the post-MI clinic if an appointment with a cardiologist was not available within 7 to 14 days after discharge. Every week during the intervention phase, the investigator sent reminder emails to ensure form completion. Providers were asked to calculate and write the score, the discharge and referral dates, where referrals were made (a cardiologist or the post-MI clinic), date of appointment, and reason for not scheduling an appointment or not referring on the risk assessment form, and to drop the completed forms in specific labeled boxes located at the CCU and PCCU work stations. The investigator collected the completed forms weekly. When the number of discharged patients did not match the number of completed forms, the investigator followed up with discharging providers to understand why.

Data and Data Collection

Data to determine whether the use of the new protocol increased discharge referrals among patients with type I MI within the recommended timeframes were collected by electronic chart review. Data included discharging unit, patients’ age, gender, admission and discharge date, diagnosis, referral to a cardiologist and the post-MI clinic, and appointment date. Clinical data needed to calculate the AMI READMITS score was also collected: PCI within 24 hours, serum creatinine, systolic blood pressure (SBP), brain natriuretic peptide (BNP), and diabetes status.

Data to assess provider satisfaction with the usability and usefulness of the new protocol were gathered through an online survey. The survey included 1 question related to the providers’ role, 1 question asking whether they used the risk assessment for each patient, and 5 Likert-items assessing the ease of usage. An additional open-ended question asked providers to share feedback related to integrating the AMI READMITS risk assessment score to the post-MI referral protocol long term.

To evaluate how consistently providers utilized the new referral protocol when discharging patients with type I MI, the number of completed forms was compared with the number of those patients who were discharged.

Statistical Analysis

Descriptive statistics were used to summarize patient demographics and to calculate the frequency of referrals before and during the intervention. Chi-square statistics were calculated to determine whether the change in percentage of referrals and timely referrals was significant. Descriptive statistics were used to determine the level of provider satisfaction related to each survey item. A content analysis method was used to synthesize themes from the open-ended question asking clinicians to share their feedback related to the new protocol.

Results

Fifty-seven patients met the study inclusion criteria: 29 patients during the preintervention phase and 28 patients during the intervention phase. There were 35 male (61.4%) and 22 female (38.6%) patients. Twenty-five patients (43.9%) were from age groups 41 through 60 years and 61 through 80 years, respectively, representing the majority of included patients. Seven patients (12.3%) were from the 81 years and older age group. There were no patients in the age group 18 through 40 years. Based on the AMI READMITS score calculation, 57.9% (n = 33) patients were from a low-risk group (includes extremely low and low risk for readmission) and 42.1% (n = 24) were from a high-risk group (includes moderate, high, and extremely high risk for readmission).

Provider adoption of the new protocol during the intervention was high. Referral forms were completed for 82% (n = 23) of the 28 patients during the intervention. Analysis findings showed a statistically significant increase in documented referrals after implementing the new referral protocol. During the preintervention phase, 66% (n = 19) of patients with type I MI were referred to see a cardiologist or an NP at a post-MI clinic and there was no documented referral for 34% (n = 10) of patients. During the intervention phase, 89% (n = 25) of patients were referred and there was no documented referral for 11% (n = 3) of patients. Chi-square results indicated that the increase in referrals was significant (χ2 = 4.571, df = 1, P = 0.033).

Data analysis examined whether patient referrals fell within the recommended timeframe of 7 days for the high-risk group (included moderate-to-extremely high risk) and 14 days for the low-risk group (included low-to-extremely low risk). During the preintervention phase, 31% (n = 9) of patient referrals were scheduled as recommended; 28% (n = 8) of patient referrals were scheduled but delayed; and there was no referral date documented for 41% (n = 12) of patients. During the intervention phase, referrals scheduled as recommended increased to 53% (n = 15); 25% (n = 7) of referrals were scheduled but delayed; and there was no referral date documented for 21.4% (n = 6) of patients. The change in appointments scheduled as recommended was not significant (χ2 = 3.550, df = 2, P = 0.169).

Surveys were emailed to 25 cardiology fellows and 3 cardiology NPs who participated in this study. Eighteen of the 28 clinicians (15 cardiology fellows and 3 cardiology NPs) responded for a response rate of 64%. One of several residents who rotated through the CCU and PCCU during the intervention also completed the survey, for a total of 19 participants. When asked if the protocol was easy to use, 79% agreed or strongly agreed. Eighteen of the 19 participants (95%) agreed or strongly agreed that the protocol was useful in making referral decisions. Sixty-eight percent agreed or strongly agreed that the AMI READMITS risk assessment score improves referral process. All participants agreed or strongly agreed that there should be an option to incorporate the AMI READMITS risk assessment score into electronic clinical notes. When asked whether the AMI READMITS risk score should be implemented in clinical practice, responses were mixed (Figure 3). A common theme among the 4 participants who responded with comments was the need for additional data to validate the usefulness of the AMI READMITS to reduce readmissions. In addition, 1 participant commented that “manual calculation [of the risk score] is not ideal.”

Discussion

This project demonstrated that implementing an evidence-based referral protocol integrating the AMI-READMITS score can increase timely postdischarge referrals among patients with type I MI. The percentage of appropriately scheduled appointments increased during the intervention phase; however, a relatively high number of appointments were scheduled outside of the recommended timeframe, similar to preintervention. Thus, while the new protocol increased referrals and provider documentation of these referrals, it appears that challenges in scheduling timely referral appointments remained. This project did not examine the reasons for delayed appointments.

The survey findings indicated that providers were generally satisfied with the usability and usefulness of the new risk assessment protocol. A large majority agreed or strongly agreed that it was easy to use and useful in making referral decisions, and most agreed or strongly agreed that it improves the referral process. Mixed opinions regarding implementing the AMI READMITS score in clinical practice, combined with qualitative findings, suggest that a lack of external validation of the AMI READMITS presents a barrier to its long-term adoption. All providers who participated in the survey agreed or strongly agreed that the risk assessment should be incorporated into electronic clinical notes. We have begun the process of working with the EHR vendor to automate the AMI risk-assessment within the referral work-flow, which will provide an opportunity for a follow-up quality improvement study.

This quality improvement project has several limitations. First, it implemented a small change in 2 inpatient units at 1 hospital using a simple pre- posttest design. Therefore, the findings are not generalizable to other settings. Prior to the intervention, some referrals may have been made without documentation. While the authors were able to trace undocumented referrals for patients who were referred to the post-MI clinic or to a cardiologist affiliated with the hospital, some patients may have been referred to cardiologists who were not affiliated with the hospital. Another limitation was that the self-created provider survey used was not tested in other clinical settings; thus, it cannot be determined whether the sensitivity and specificity of the survey questions are high. In addition, the clinical providers who participated in the study knew the study team, which may have influenced their behavior during the study period. Furthermore, the identified improvement in clinicians’ referral practices may not be sustainable due to the complexity and effort required to manually calculate the risk score. This limitation could be eliminated by integrating the risk score calculation into the EHR.

Conclusion

Early follow-up after discharge plays an important role in supporting patients’ self-management of some risk factors (ie, diet, weight, and smoking) and identifying gaps in postdischarge care which may lead to readmission. This project provides evidence that integrating the AMI READMITS risk assessment score into the referral process can help to guide discharge decision-making and increase timely, appropriate referrals for patients with MI. Integration of a specific risk assessment, such as the AMI READMITS, within the post-MI referral protocol may help clinicians make more efficient, educated referral decisions. Future studies should explore more specifically how and why the new protocol impacts clinicians’ decision-making and behavior related to post-MI referrals. In addition, future studies should investigate challenges associated with scheduling postdischarge appointments. It will be important to investigate how integration of the new protocol within the EHR may increase efficiency, consistency, and provider satisfaction with the new referral process. Additional research investigating the effects of the AMI READMITS score on readmissions reduction will be important to promote long-term adoption of the improved referral protocol in clinical practice.

Acknowledgments: The authors thank Shelly Conaway, ANP-BC, MSN, Angela Street, ANP-BC, MSN, Andrew Geis, ACNP-BC, MSN, Richard P. Jones II, MD, Eunice Young, MD, Joy Rothwell, MSN, RN-BC, Allison Olazo, MBA, MSN, RN-BC, Elizabeth Heck, RN-BC, and Matthew Trojanowski, MHA, MS, RRT, CSSBB for their support of this study.

Corresponding author: Nailya Muganlinskaya, DNP, MPH, ACNP-BC, MSN, The Johns Hopkins Hospital, 1800 Orleans St, Baltimore, MD 21287; [email protected].

Financial disclosures: None.

From The Johns Hopkins Hospital, Baltimore, MD (Dr. Muganlinskaya and Dr. Skojec, retired); The George Washington University, Washington, DC (Dr. Posey); and Johns Hopkins University, Baltimore, MD (Dr. Resar).

Abstract

Objective: Assessing the risk characteristics of patients with acute myocardial infarction (MI) can help providers make appropriate referral decisions. This quality improvement project sought to improve timely, appropriate referrals among patients with type I MI by adding a risk assessment, the AMI READMITS score, to the existing referral protocol.

Methods: Patients’ chart data were analyzed to assess changes in referrals and timely follow-up appointments from pre-intervention to intervention. A survey assessed providers’ satisfaction with the new referral protocol.

Results: Among 57 patients (n = 29 preintervention; n = 28 intervention), documented referrals increased significantly from 66% to 89% (χ2 = 4.571, df = 1, P = 0.033); and timely appointments increased by 10%, which was not significant (χ2 = 3.550, df = 2, P = 0.169). Most providers agreed that the new protocol was easy to use, useful in making referral decisions, and improved the referral process. All agreed the risk score should be incorporated into electronic clinical notes. Provider opinions related to implementing the risk score in clinical practice were mixed. Qualitative feedback suggests this was due to limited validation of the AMI READMITS score in reducing readmissions.

Conclusions: Our risk-based referral protocol helped to increase appropriate referrals among patients with type I MI. Provider adoption may be enhanced by incorporating the protocol into electronic clinical notes. Research to further validate the accuracy of the AMI READMITS score in predicting readmissions may support adoption of the protocol in clinical practice.

Keywords: quality improvement; type I myocardial infarction; referral process; readmission risk; risk assessment; chart review.

Early follow-up after discharge is an important strategy to reduce the risk of unplanned hospital readmissions among patients with various conditions.^1-3 While patient confounding factors, such as chronic health problems, environment, socioeconomic status, and literacy, make it difficult to avoid all unplanned readmissions, early follow-up may help providers identify and appropriately manage some health-related issues, and as such is a pivotal element of a readmission prevention strategy.⁴ There is evidence that patients with non-ST elevation myocardial infarction (NSTEMI) who have an outpatient appointment with a physician within 7 days after discharge have a lower risk of 30-day readmission.⁵

Our hospital’s postmyocardial infarction clinic was created to prevent unplanned readmissions within 30 days after discharge among patients with type I myocardial infarction (MI). Since inception, the number of referrals has been much lower than expected. In 2018, the total number of patients discharged from the hospital with type I MI and any troponin I level above 0.40 ng/mL was 313. Most of these patients were discharged from the hospital’s cardiac units; however, only 91 referrals were made. To increase referrals, the cardiology nurse practitioners (NPs) developed a post-MI referral protocol (Figure 1). However, this protocol was not consistently used and referrals to the clinic remained low.

Evidence-based risk assessment tools have the potential to increase effective patient management. For example, cardiology providers at the hospital utilize various scores, such as CHA₂DS₂-VASc⁶ and the Society of Thoracic Surgery risk score,⁷ to plan patient management. Among the scores used to predict unplanned readmissions for MI patients, the most promising is the AMI READMITS score.⁸ Unlike other nonspecific prediction models, the AMI READMITS score was developed based on variables extracted from the electronic health records (EHRs) of patients who were hospitalized for MI and readmitted within 30 days after discharge. Recognizing the potential to increase referrals by integrating an MI-specific risk assessment, this quality improvement study modified the existing referral protocol to include the patients’ AMI READMITS score and recommendations for follow-up.

Currently, there are no clear recommendations on how soon after discharge patients with MI should undergo follow-up. As research data vary, we selected 7 days follow-up for patients from high risk groups based on the “See you in 7” initiative for patients with heart failure (HF) and MI,^9,10 as well as evidence that patients with NSTEMI have a lower risk of 30-day readmission if they have follow-up within 7 days after discharge⁵; and we selected 14 days follow-up for patients from low-risk groups based on evidence that postdischarge follow-up within 14 days reduces risk of 30-day readmission in patients with acute myocardial infarction (AMI) and/or acutely decompensated HF.¹¹

Methods

This project was designed to answer the following question: For adult patients with type I MI, does implementation of a readmission risk assessment referral protocol increase the percentage of referrals and appointments scheduled within a recommended time? Anticipated outcomes included: (1) increased referrals to a cardiologist or the post-MI clinic; (2) increased scheduled follow-up appointments within 7 to 14 days; (3) provider satisfaction with the usability and usefulness of the new protocol; and (4) consistent provider adoption of the new risk assessment referral protocol.

To evaluate the degree to which these outcomes were achieved, we reviewed patient charts for 2 months prior and 2 months during implementation of the new referral protocol. As shown in Figure 2, the new protocol added the following process steps to the existing protocol: calculation of the AMI READMITS score, recommendations for follow-up based on patients’ risk score, and guidance to refer patients to the post-MI clinic if patients did not have an appointment with a cardiologist within 7 to 14 days after discharge. Patients’ risk assessment scores were obtained from forms completed by clinicians during the intervention. Clinician’s perceptions related to the usability and usefulness of the new protocol and feedback related to its long-term adoption were assessed using a descriptive survey.

Population

The project population included all adult patients (≥ 18 years old) with type I MI who were admitted or transferred to the hospital, had a percutaneous coronary intervention (PCI), or were managed without PCI and discharged from the hospital’s cardiac care unit (CCU) and progressive cardiac care unit (PCCU). The criteria for type I MI included the “detection of a rise and/or fall of cardiac troponin with at least 1 value above the 99th percentile and with at least 1 of the following: symptoms of acute myocardial ischemia; new ischemic electrocardiographic (ECG) changes; development of new pathological Q waves; imaging evidence of new loss of viable myocardium or new regional wall motion abnormality in a pattern consistent with an ischemic etiology; identification of a coronary thrombus by angiography including intracoronary imaging or by autopsy.”¹² The study excluded patients with type I MI who were referred for coronary bypass surgery.

Intervention

The revised risk assessment protocol was implemented within the CCU and PCCU. The lead investigator met with each provider to discuss the role of the post-MI clinic, current referral rates, the purpose of the project, and the new referral process to be completed during the project for each patient discharged with type I MI. Cardiology NPs, fellows, and residents were asked to use the risk-assessment form to calculate patients’ risk for readmission, and refer patients to the post-MI clinic if an appointment with a cardiologist was not available within 7 to 14 days after discharge. Every week during the intervention phase, the investigator sent reminder emails to ensure form completion. Providers were asked to calculate and write the score, the discharge and referral dates, where referrals were made (a cardiologist or the post-MI clinic), date of appointment, and reason for not scheduling an appointment or not referring on the risk assessment form, and to drop the completed forms in specific labeled boxes located at the CCU and PCCU work stations. The investigator collected the completed forms weekly. When the number of discharged patients did not match the number of completed forms, the investigator followed up with discharging providers to understand why.

Data and Data Collection

Data to determine whether the use of the new protocol increased discharge referrals among patients with type I MI within the recommended timeframes were collected by electronic chart review. Data included discharging unit, patients’ age, gender, admission and discharge date, diagnosis, referral to a cardiologist and the post-MI clinic, and appointment date. Clinical data needed to calculate the AMI READMITS score was also collected: PCI within 24 hours, serum creatinine, systolic blood pressure (SBP), brain natriuretic peptide (BNP), and diabetes status.

Data to assess provider satisfaction with the usability and usefulness of the new protocol were gathered through an online survey. The survey included 1 question related to the providers’ role, 1 question asking whether they used the risk assessment for each patient, and 5 Likert-items assessing the ease of usage. An additional open-ended question asked providers to share feedback related to integrating the AMI READMITS risk assessment score to the post-MI referral protocol long term.

To evaluate how consistently providers utilized the new referral protocol when discharging patients with type I MI, the number of completed forms was compared with the number of those patients who were discharged.

Statistical Analysis

Descriptive statistics were used to summarize patient demographics and to calculate the frequency of referrals before and during the intervention. Chi-square statistics were calculated to determine whether the change in percentage of referrals and timely referrals was significant. Descriptive statistics were used to determine the level of provider satisfaction related to each survey item. A content analysis method was used to synthesize themes from the open-ended question asking clinicians to share their feedback related to the new protocol.

Results

Fifty-seven patients met the study inclusion criteria: 29 patients during the preintervention phase and 28 patients during the intervention phase. There were 35 male (61.4%) and 22 female (38.6%) patients. Twenty-five patients (43.9%) were from age groups 41 through 60 years and 61 through 80 years, respectively, representing the majority of included patients. Seven patients (12.3%) were from the 81 years and older age group. There were no patients in the age group 18 through 40 years. Based on the AMI READMITS score calculation, 57.9% (n = 33) patients were from a low-risk group (includes extremely low and low risk for readmission) and 42.1% (n = 24) were from a high-risk group (includes moderate, high, and extremely high risk for readmission).

Provider adoption of the new protocol during the intervention was high. Referral forms were completed for 82% (n = 23) of the 28 patients during the intervention. Analysis findings showed a statistically significant increase in documented referrals after implementing the new referral protocol. During the preintervention phase, 66% (n = 19) of patients with type I MI were referred to see a cardiologist or an NP at a post-MI clinic and there was no documented referral for 34% (n = 10) of patients. During the intervention phase, 89% (n = 25) of patients were referred and there was no documented referral for 11% (n = 3) of patients. Chi-square results indicated that the increase in referrals was significant (χ2 = 4.571, df = 1, P = 0.033).

Data analysis examined whether patient referrals fell within the recommended timeframe of 7 days for the high-risk group (included moderate-to-extremely high risk) and 14 days for the low-risk group (included low-to-extremely low risk). During the preintervention phase, 31% (n = 9) of patient referrals were scheduled as recommended; 28% (n = 8) of patient referrals were scheduled but delayed; and there was no referral date documented for 41% (n = 12) of patients. During the intervention phase, referrals scheduled as recommended increased to 53% (n = 15); 25% (n = 7) of referrals were scheduled but delayed; and there was no referral date documented for 21.4% (n = 6) of patients. The change in appointments scheduled as recommended was not significant (χ2 = 3.550, df = 2, P = 0.169).

Surveys were emailed to 25 cardiology fellows and 3 cardiology NPs who participated in this study. Eighteen of the 28 clinicians (15 cardiology fellows and 3 cardiology NPs) responded for a response rate of 64%. One of several residents who rotated through the CCU and PCCU during the intervention also completed the survey, for a total of 19 participants. When asked if the protocol was easy to use, 79% agreed or strongly agreed. Eighteen of the 19 participants (95%) agreed or strongly agreed that the protocol was useful in making referral decisions. Sixty-eight percent agreed or strongly agreed that the AMI READMITS risk assessment score improves referral process. All participants agreed or strongly agreed that there should be an option to incorporate the AMI READMITS risk assessment score into electronic clinical notes. When asked whether the AMI READMITS risk score should be implemented in clinical practice, responses were mixed (Figure 3). A common theme among the 4 participants who responded with comments was the need for additional data to validate the usefulness of the AMI READMITS to reduce readmissions. In addition, 1 participant commented that “manual calculation [of the risk score] is not ideal.”

Discussion

This project demonstrated that implementing an evidence-based referral protocol integrating the AMI-READMITS score can increase timely postdischarge referrals among patients with type I MI. The percentage of appropriately scheduled appointments increased during the intervention phase; however, a relatively high number of appointments were scheduled outside of the recommended timeframe, similar to preintervention. Thus, while the new protocol increased referrals and provider documentation of these referrals, it appears that challenges in scheduling timely referral appointments remained. This project did not examine the reasons for delayed appointments.

The survey findings indicated that providers were generally satisfied with the usability and usefulness of the new risk assessment protocol. A large majority agreed or strongly agreed that it was easy to use and useful in making referral decisions, and most agreed or strongly agreed that it improves the referral process. Mixed opinions regarding implementing the AMI READMITS score in clinical practice, combined with qualitative findings, suggest that a lack of external validation of the AMI READMITS presents a barrier to its long-term adoption. All providers who participated in the survey agreed or strongly agreed that the risk assessment should be incorporated into electronic clinical notes. We have begun the process of working with the EHR vendor to automate the AMI risk-assessment within the referral work-flow, which will provide an opportunity for a follow-up quality improvement study.

This quality improvement project has several limitations. First, it implemented a small change in 2 inpatient units at 1 hospital using a simple pre- posttest design. Therefore, the findings are not generalizable to other settings. Prior to the intervention, some referrals may have been made without documentation. While the authors were able to trace undocumented referrals for patients who were referred to the post-MI clinic or to a cardiologist affiliated with the hospital, some patients may have been referred to cardiologists who were not affiliated with the hospital. Another limitation was that the self-created provider survey used was not tested in other clinical settings; thus, it cannot be determined whether the sensitivity and specificity of the survey questions are high. In addition, the clinical providers who participated in the study knew the study team, which may have influenced their behavior during the study period. Furthermore, the identified improvement in clinicians’ referral practices may not be sustainable due to the complexity and effort required to manually calculate the risk score. This limitation could be eliminated by integrating the risk score calculation into the EHR.

Conclusion

Early follow-up after discharge plays an important role in supporting patients’ self-management of some risk factors (ie, diet, weight, and smoking) and identifying gaps in postdischarge care which may lead to readmission. This project provides evidence that integrating the AMI READMITS risk assessment score into the referral process can help to guide discharge decision-making and increase timely, appropriate referrals for patients with MI. Integration of a specific risk assessment, such as the AMI READMITS, within the post-MI referral protocol may help clinicians make more efficient, educated referral decisions. Future studies should explore more specifically how and why the new protocol impacts clinicians’ decision-making and behavior related to post-MI referrals. In addition, future studies should investigate challenges associated with scheduling postdischarge appointments. It will be important to investigate how integration of the new protocol within the EHR may increase efficiency, consistency, and provider satisfaction with the new referral process. Additional research investigating the effects of the AMI READMITS score on readmissions reduction will be important to promote long-term adoption of the improved referral protocol in clinical practice.

Acknowledgments: The authors thank Shelly Conaway, ANP-BC, MSN, Angela Street, ANP-BC, MSN, Andrew Geis, ACNP-BC, MSN, Richard P. Jones II, MD, Eunice Young, MD, Joy Rothwell, MSN, RN-BC, Allison Olazo, MBA, MSN, RN-BC, Elizabeth Heck, RN-BC, and Matthew Trojanowski, MHA, MS, RRT, CSSBB for their support of this study.

Corresponding author: Nailya Muganlinskaya, DNP, MPH, ACNP-BC, MSN, The Johns Hopkins Hospital, 1800 Orleans St, Baltimore, MD 21287; [email protected].

Financial disclosures: None.

From The Johns Hopkins Hospital, Baltimore, MD (Dr. Muganlinskaya and Dr. Skojec, retired); The George Washington University, Washington, DC (Dr. Posey); and Johns Hopkins University, Baltimore, MD (Dr. Resar).

Abstract

Objective: Assessing the risk characteristics of patients with acute myocardial infarction (MI) can help providers make appropriate referral decisions. This quality improvement project sought to improve timely, appropriate referrals among patients with type I MI by adding a risk assessment, the AMI READMITS score, to the existing referral protocol.

Methods: Patients’ chart data were analyzed to assess changes in referrals and timely follow-up appointments from pre-intervention to intervention. A survey assessed providers’ satisfaction with the new referral protocol.

Results: Among 57 patients (n = 29 preintervention; n = 28 intervention), documented referrals increased significantly from 66% to 89% (χ2 = 4.571, df = 1, P = 0.033); and timely appointments increased by 10%, which was not significant (χ2 = 3.550, df = 2, P = 0.169). Most providers agreed that the new protocol was easy to use, useful in making referral decisions, and improved the referral process. All agreed the risk score should be incorporated into electronic clinical notes. Provider opinions related to implementing the risk score in clinical practice were mixed. Qualitative feedback suggests this was due to limited validation of the AMI READMITS score in reducing readmissions.

Conclusions: Our risk-based referral protocol helped to increase appropriate referrals among patients with type I MI. Provider adoption may be enhanced by incorporating the protocol into electronic clinical notes. Research to further validate the accuracy of the AMI READMITS score in predicting readmissions may support adoption of the protocol in clinical practice.

Keywords: quality improvement; type I myocardial infarction; referral process; readmission risk; risk assessment; chart review.

Early follow-up after discharge is an important strategy to reduce the risk of unplanned hospital readmissions among patients with various conditions.^1-3 While patient confounding factors, such as chronic health problems, environment, socioeconomic status, and literacy, make it difficult to avoid all unplanned readmissions, early follow-up may help providers identify and appropriately manage some health-related issues, and as such is a pivotal element of a readmission prevention strategy.⁴ There is evidence that patients with non-ST elevation myocardial infarction (NSTEMI) who have an outpatient appointment with a physician within 7 days after discharge have a lower risk of 30-day readmission.⁵

Our hospital’s postmyocardial infarction clinic was created to prevent unplanned readmissions within 30 days after discharge among patients with type I myocardial infarction (MI). Since inception, the number of referrals has been much lower than expected. In 2018, the total number of patients discharged from the hospital with type I MI and any troponin I level above 0.40 ng/mL was 313. Most of these patients were discharged from the hospital’s cardiac units; however, only 91 referrals were made. To increase referrals, the cardiology nurse practitioners (NPs) developed a post-MI referral protocol (Figure 1). However, this protocol was not consistently used and referrals to the clinic remained low.

Evidence-based risk assessment tools have the potential to increase effective patient management. For example, cardiology providers at the hospital utilize various scores, such as CHA₂DS₂-VASc⁶ and the Society of Thoracic Surgery risk score,⁷ to plan patient management. Among the scores used to predict unplanned readmissions for MI patients, the most promising is the AMI READMITS score.⁸ Unlike other nonspecific prediction models, the AMI READMITS score was developed based on variables extracted from the electronic health records (EHRs) of patients who were hospitalized for MI and readmitted within 30 days after discharge. Recognizing the potential to increase referrals by integrating an MI-specific risk assessment, this quality improvement study modified the existing referral protocol to include the patients’ AMI READMITS score and recommendations for follow-up.

Currently, there are no clear recommendations on how soon after discharge patients with MI should undergo follow-up. As research data vary, we selected 7 days follow-up for patients from high risk groups based on the “See you in 7” initiative for patients with heart failure (HF) and MI,^9,10 as well as evidence that patients with NSTEMI have a lower risk of 30-day readmission if they have follow-up within 7 days after discharge⁵; and we selected 14 days follow-up for patients from low-risk groups based on evidence that postdischarge follow-up within 14 days reduces risk of 30-day readmission in patients with acute myocardial infarction (AMI) and/or acutely decompensated HF.¹¹

Methods

This project was designed to answer the following question: For adult patients with type I MI, does implementation of a readmission risk assessment referral protocol increase the percentage of referrals and appointments scheduled within a recommended time? Anticipated outcomes included: (1) increased referrals to a cardiologist or the post-MI clinic; (2) increased scheduled follow-up appointments within 7 to 14 days; (3) provider satisfaction with the usability and usefulness of the new protocol; and (4) consistent provider adoption of the new risk assessment referral protocol.

To evaluate the degree to which these outcomes were achieved, we reviewed patient charts for 2 months prior and 2 months during implementation of the new referral protocol. As shown in Figure 2, the new protocol added the following process steps to the existing protocol: calculation of the AMI READMITS score, recommendations for follow-up based on patients’ risk score, and guidance to refer patients to the post-MI clinic if patients did not have an appointment with a cardiologist within 7 to 14 days after discharge. Patients’ risk assessment scores were obtained from forms completed by clinicians during the intervention. Clinician’s perceptions related to the usability and usefulness of the new protocol and feedback related to its long-term adoption were assessed using a descriptive survey.

Population

The project population included all adult patients (≥ 18 years old) with type I MI who were admitted or transferred to the hospital, had a percutaneous coronary intervention (PCI), or were managed without PCI and discharged from the hospital’s cardiac care unit (CCU) and progressive cardiac care unit (PCCU). The criteria for type I MI included the “detection of a rise and/or fall of cardiac troponin with at least 1 value above the 99th percentile and with at least 1 of the following: symptoms of acute myocardial ischemia; new ischemic electrocardiographic (ECG) changes; development of new pathological Q waves; imaging evidence of new loss of viable myocardium or new regional wall motion abnormality in a pattern consistent with an ischemic etiology; identification of a coronary thrombus by angiography including intracoronary imaging or by autopsy.”¹² The study excluded patients with type I MI who were referred for coronary bypass surgery.

Intervention

The revised risk assessment protocol was implemented within the CCU and PCCU. The lead investigator met with each provider to discuss the role of the post-MI clinic, current referral rates, the purpose of the project, and the new referral process to be completed during the project for each patient discharged with type I MI. Cardiology NPs, fellows, and residents were asked to use the risk-assessment form to calculate patients’ risk for readmission, and refer patients to the post-MI clinic if an appointment with a cardiologist was not available within 7 to 14 days after discharge. Every week during the intervention phase, the investigator sent reminder emails to ensure form completion. Providers were asked to calculate and write the score, the discharge and referral dates, where referrals were made (a cardiologist or the post-MI clinic), date of appointment, and reason for not scheduling an appointment or not referring on the risk assessment form, and to drop the completed forms in specific labeled boxes located at the CCU and PCCU work stations. The investigator collected the completed forms weekly. When the number of discharged patients did not match the number of completed forms, the investigator followed up with discharging providers to understand why.

Data and Data Collection

Data to determine whether the use of the new protocol increased discharge referrals among patients with type I MI within the recommended timeframes were collected by electronic chart review. Data included discharging unit, patients’ age, gender, admission and discharge date, diagnosis, referral to a cardiologist and the post-MI clinic, and appointment date. Clinical data needed to calculate the AMI READMITS score was also collected: PCI within 24 hours, serum creatinine, systolic blood pressure (SBP), brain natriuretic peptide (BNP), and diabetes status.

Data to assess provider satisfaction with the usability and usefulness of the new protocol were gathered through an online survey. The survey included 1 question related to the providers’ role, 1 question asking whether they used the risk assessment for each patient, and 5 Likert-items assessing the ease of usage. An additional open-ended question asked providers to share feedback related to integrating the AMI READMITS risk assessment score to the post-MI referral protocol long term.

To evaluate how consistently providers utilized the new referral protocol when discharging patients with type I MI, the number of completed forms was compared with the number of those patients who were discharged.

Statistical Analysis

Descriptive statistics were used to summarize patient demographics and to calculate the frequency of referrals before and during the intervention. Chi-square statistics were calculated to determine whether the change in percentage of referrals and timely referrals was significant. Descriptive statistics were used to determine the level of provider satisfaction related to each survey item. A content analysis method was used to synthesize themes from the open-ended question asking clinicians to share their feedback related to the new protocol.

Results

Fifty-seven patients met the study inclusion criteria: 29 patients during the preintervention phase and 28 patients during the intervention phase. There were 35 male (61.4%) and 22 female (38.6%) patients. Twenty-five patients (43.9%) were from age groups 41 through 60 years and 61 through 80 years, respectively, representing the majority of included patients. Seven patients (12.3%) were from the 81 years and older age group. There were no patients in the age group 18 through 40 years. Based on the AMI READMITS score calculation, 57.9% (n = 33) patients were from a low-risk group (includes extremely low and low risk for readmission) and 42.1% (n = 24) were from a high-risk group (includes moderate, high, and extremely high risk for readmission).

Provider adoption of the new protocol during the intervention was high. Referral forms were completed for 82% (n = 23) of the 28 patients during the intervention. Analysis findings showed a statistically significant increase in documented referrals after implementing the new referral protocol. During the preintervention phase, 66% (n = 19) of patients with type I MI were referred to see a cardiologist or an NP at a post-MI clinic and there was no documented referral for 34% (n = 10) of patients. During the intervention phase, 89% (n = 25) of patients were referred and there was no documented referral for 11% (n = 3) of patients. Chi-square results indicated that the increase in referrals was significant (χ2 = 4.571, df = 1, P = 0.033).

Data analysis examined whether patient referrals fell within the recommended timeframe of 7 days for the high-risk group (included moderate-to-extremely high risk) and 14 days for the low-risk group (included low-to-extremely low risk). During the preintervention phase, 31% (n = 9) of patient referrals were scheduled as recommended; 28% (n = 8) of patient referrals were scheduled but delayed; and there was no referral date documented for 41% (n = 12) of patients. During the intervention phase, referrals scheduled as recommended increased to 53% (n = 15); 25% (n = 7) of referrals were scheduled but delayed; and there was no referral date documented for 21.4% (n = 6) of patients. The change in appointments scheduled as recommended was not significant (χ2 = 3.550, df = 2, P = 0.169).

Surveys were emailed to 25 cardiology fellows and 3 cardiology NPs who participated in this study. Eighteen of the 28 clinicians (15 cardiology fellows and 3 cardiology NPs) responded for a response rate of 64%. One of several residents who rotated through the CCU and PCCU during the intervention also completed the survey, for a total of 19 participants. When asked if the protocol was easy to use, 79% agreed or strongly agreed. Eighteen of the 19 participants (95%) agreed or strongly agreed that the protocol was useful in making referral decisions. Sixty-eight percent agreed or strongly agreed that the AMI READMITS risk assessment score improves referral process. All participants agreed or strongly agreed that there should be an option to incorporate the AMI READMITS risk assessment score into electronic clinical notes. When asked whether the AMI READMITS risk score should be implemented in clinical practice, responses were mixed (Figure 3). A common theme among the 4 participants who responded with comments was the need for additional data to validate the usefulness of the AMI READMITS to reduce readmissions. In addition, 1 participant commented that “manual calculation [of the risk score] is not ideal.”

Discussion

This project demonstrated that implementing an evidence-based referral protocol integrating the AMI-READMITS score can increase timely postdischarge referrals among patients with type I MI. The percentage of appropriately scheduled appointments increased during the intervention phase; however, a relatively high number of appointments were scheduled outside of the recommended timeframe, similar to preintervention. Thus, while the new protocol increased referrals and provider documentation of these referrals, it appears that challenges in scheduling timely referral appointments remained. This project did not examine the reasons for delayed appointments.

The survey findings indicated that providers were generally satisfied with the usability and usefulness of the new risk assessment protocol. A large majority agreed or strongly agreed that it was easy to use and useful in making referral decisions, and most agreed or strongly agreed that it improves the referral process. Mixed opinions regarding implementing the AMI READMITS score in clinical practice, combined with qualitative findings, suggest that a lack of external validation of the AMI READMITS presents a barrier to its long-term adoption. All providers who participated in the survey agreed or strongly agreed that the risk assessment should be incorporated into electronic clinical notes. We have begun the process of working with the EHR vendor to automate the AMI risk-assessment within the referral work-flow, which will provide an opportunity for a follow-up quality improvement study.

This quality improvement project has several limitations. First, it implemented a small change in 2 inpatient units at 1 hospital using a simple pre- posttest design. Therefore, the findings are not generalizable to other settings. Prior to the intervention, some referrals may have been made without documentation. While the authors were able to trace undocumented referrals for patients who were referred to the post-MI clinic or to a cardiologist affiliated with the hospital, some patients may have been referred to cardiologists who were not affiliated with the hospital. Another limitation was that the self-created provider survey used was not tested in other clinical settings; thus, it cannot be determined whether the sensitivity and specificity of the survey questions are high. In addition, the clinical providers who participated in the study knew the study team, which may have influenced their behavior during the study period. Furthermore, the identified improvement in clinicians’ referral practices may not be sustainable due to the complexity and effort required to manually calculate the risk score. This limitation could be eliminated by integrating the risk score calculation into the EHR.

Conclusion

Early follow-up after discharge plays an important role in supporting patients’ self-management of some risk factors (ie, diet, weight, and smoking) and identifying gaps in postdischarge care which may lead to readmission. This project provides evidence that integrating the AMI READMITS risk assessment score into the referral process can help to guide discharge decision-making and increase timely, appropriate referrals for patients with MI. Integration of a specific risk assessment, such as the AMI READMITS, within the post-MI referral protocol may help clinicians make more efficient, educated referral decisions. Future studies should explore more specifically how and why the new protocol impacts clinicians’ decision-making and behavior related to post-MI referrals. In addition, future studies should investigate challenges associated with scheduling postdischarge appointments. It will be important to investigate how integration of the new protocol within the EHR may increase efficiency, consistency, and provider satisfaction with the new referral process. Additional research investigating the effects of the AMI READMITS score on readmissions reduction will be important to promote long-term adoption of the improved referral protocol in clinical practice.

Acknowledgments: The authors thank Shelly Conaway, ANP-BC, MSN, Angela Street, ANP-BC, MSN, Andrew Geis, ACNP-BC, MSN, Richard P. Jones II, MD, Eunice Young, MD, Joy Rothwell, MSN, RN-BC, Allison Olazo, MBA, MSN, RN-BC, Elizabeth Heck, RN-BC, and Matthew Trojanowski, MHA, MS, RRT, CSSBB for their support of this study.

Corresponding author: Nailya Muganlinskaya, DNP, MPH, ACNP-BC, MSN, The Johns Hopkins Hospital, 1800 Orleans St, Baltimore, MD 21287; [email protected].

Financial disclosures: None.

To benefit patients and the public health of their communities, children’s hospitals across the United States prepared for and responded to COVID-19 by conserving personal protective equipment, suspending noncritical in-person healthcare encounters (including outpatient visits and elective surgeries), and implementing socially distanced essential care.^1,2 These measures were promptly instituted during a time of both substantial uncertainty about the pandemic’s behavior in children—including its severity and duration—and extreme variation in local and state governments’ responses to the pandemic.

Congruent with other healthcare institutions, children’s hospitals calibrated their clinical operations to the evolving nature of the pandemic, prioritizing the safety of patients and staff while striving to maintain financial viability in the setting of increased costs and decreased revenue. In some cases, children’s hospitals aided adult hospitals and health systems by admitting young and middle-aged adult patients and by centralizing all pediatric patients requiring intensive care within a region. These efforts occurred while many children’s hospitals remained the sole source of specialized pediatric care, including care for rare complex health problems.

As the COVID-19 pandemic continues, there is a critical need to assess how the initial phase of the pandemic affected healthcare encounters and related finances in children’s hospitals. Understanding these trends will position children’s hospitals to project and prepare for subsequent COVID-19 surges, as well as future related public health crises that necessitate widespread social distancing. Therefore, we compared year-over-year trends in healthcare encounters and hospital charges across US children’s hospitals before and during the COVID-19 pandemic, focusing on the beginning of COVID-19 in the United States, which was defined as February through June 2020.

This is a retrospective analysis of 26 children’s hospitals (22 freestanding, 4 nonfreestanding) from all US regions (12 South, 7 Midwest, 5 West, 2 Northeast) contributing encounter and financial data to the PROSPECT database (Children’s Hospital Association, Lenexa, Kansas) from February 1 to June 30 in both 2019 (before COVID-19) and 2020 (during COVID-19). In response to COVID-19, hospitals participating in PROSPECT increased the efficiency of data centralization and reporting in 2020 during the period February 1 to June 30 to expedite analysis and dissemination of findings.

The main outcome measures were the percentage of change in weekly encounters (inpatient bed-days, emergency department [ED] visits, and surgeries) and inflation-adjusted charges (categorized as inpatient care and outpatient care, such as ambulatory surgery, clinics, and ED visits) before vs during COVID-19. Number of encounters and charges were compared using the Wilcoxon signed-rank test. The distribution of weekly change in outcome measures from 2019 vs 2020 across hospitals was reported with medians and interquartile ranges (IQRs). The threshold of statistical significance was set at P < .05. All analyses were performed with SAS version 9.4 (SAS Institute). This study was considered exempt from human subjects research by the Institutional Review Board of Children’s Mercy Hospital (Kansas City, Missouri).

All 26 children’s hospitals experienced similar trends in healthcare encounters and charges during the study period (Figure and Table). From February 1 to March 10, 2020, the volume of healthcare encounters in the children’s hospitals remained the same as that for the same period in 2019 (P > .1) (Figure).

Charges that accrued from February 1 to June 30 were lower in 2020 by a median 23.6% (IQR, –28.7% to –19.1%) per children’s hospital than they were in 2019, corresponding to a median decrease of $276.3 million (IQR, $404.0-$126.0 million) in charges per hospital (Table). Forty percent of this decrease was attributable to decreased charges resulting from fewer inpatient healthcare encounters.

During the initial phase of the COVID-19 pandemic in the United States, children’s hospitals experienced a substantial decrease in healthcare encounters and charges. Greater decreases were observed for ED visits and surgery encounters than for inpatient bed-days. Nonetheless, inpatient bed-days decreased by more than one-third, consistent with the decrease in inpatient resource use reported for adult hospitals.³ Remarkably, these trends were consistent across children’s hospitals, despite variation in the content and installation of and adherence with social distancing policies in their surrounding local areas.

These findings beg the question of how well children’s hospitals are positioned to weather a recurrent surge in COVID-19. Because the severity of illness of COVID-19 has been lower to date in the pediatric vs adult populations, an increase in COVID-19-related visits to EDs and admissions to offset the decreased resource use of other pediatric healthcare problems is not anticipated. Existing hospital financial reserves as well as federal aid from the Coronavirus Aid, Relief, and Economic Security Act that helped mitigate the initial encounter and financial losses during the beginning of COVID-19 may not be readily available over time.^4,5 Certainly, the findings from the current study support continued lobbying for additional state and federal funds allocated through future relief packages to children’s hospitals.

Additional approaches to financial solvency in children’s hospitals during the sustained COVID-19 pandemic include addressing surgical backlogs and sharing best practices for safe and sustained reopening of clinical operations and financial practices across institutions. Although the PROSPECT database does not contain information on the types of surgeries present within this backlog, our experiences suggest that both same-day and inpatient elective surgeries have been affected, especially lengthy procedures (eg, spinal fusion for neuromuscular scoliosis). Spread and scale of feasible and efficient solutions to reengineer and expand patient capacities and throughput for operating rooms, postanesthesia recovery areas, and intensive care and floor units are needed. Enhanced analytics that accurately predict postoperative length of hospital stay, coupled with early recovery after surgery clinical protocols, could help optimize hospital bed management. Effective ways to convert hospital rooms from single to double occupancy, to manage family visitation, and to proactively test asymptomatic patients, family, and hospital staff will mitigate continued COVID-19 penetration through children’s hospitals.

One important limitation of the current study is the measurement of hospitals’ charges. The charge data were not positioned to comprehensively measure each hospital’s financial state during the COVID-19 pandemic. However, the decrease in hospital charges reported by the children’s hospitals in the current study is comparable with the financial losses reported for many adult hospitals during the pandemic.^6,7 It is important to recognize that the amount of the charges may not be equivalent to the cost of care or revenue collected by the hospitals. PROSPECT does not contain information on cost, and current cost-to-charge ratios are based on historical (ie, pre-COVID-19) data; therefore, they do not account for increased cost of personal protective equipment and other related costs that occurred during the pandemic, which makes use of these ratios challenging. Nevertheless, it is possible that the relative difference in costs incurred and revenue collected before and during COVID-19 may have been similar to the differences observed in hospital charges.

Children’s hospitals’ ability to serve the nation’s pediatric patients depends on the success of the hospitals’ plans to manage current and future COVID-19 surges and to reopen and recover from the surges that have passed. Additional investigation is needed to identify best operational and financial practices among children’s hospitals that have enabled them to endure the COVID-19 pandemic.

To benefit patients and the public health of their communities, children’s hospitals across the United States prepared for and responded to COVID-19 by conserving personal protective equipment, suspending noncritical in-person healthcare encounters (including outpatient visits and elective surgeries), and implementing socially distanced essential care.^1,2 These measures were promptly instituted during a time of both substantial uncertainty about the pandemic’s behavior in children—including its severity and duration—and extreme variation in local and state governments’ responses to the pandemic.

Congruent with other healthcare institutions, children’s hospitals calibrated their clinical operations to the evolving nature of the pandemic, prioritizing the safety of patients and staff while striving to maintain financial viability in the setting of increased costs and decreased revenue. In some cases, children’s hospitals aided adult hospitals and health systems by admitting young and middle-aged adult patients and by centralizing all pediatric patients requiring intensive care within a region. These efforts occurred while many children’s hospitals remained the sole source of specialized pediatric care, including care for rare complex health problems.

As the COVID-19 pandemic continues, there is a critical need to assess how the initial phase of the pandemic affected healthcare encounters and related finances in children’s hospitals. Understanding these trends will position children’s hospitals to project and prepare for subsequent COVID-19 surges, as well as future related public health crises that necessitate widespread social distancing. Therefore, we compared year-over-year trends in healthcare encounters and hospital charges across US children’s hospitals before and during the COVID-19 pandemic, focusing on the beginning of COVID-19 in the United States, which was defined as February through June 2020.

This is a retrospective analysis of 26 children’s hospitals (22 freestanding, 4 nonfreestanding) from all US regions (12 South, 7 Midwest, 5 West, 2 Northeast) contributing encounter and financial data to the PROSPECT database (Children’s Hospital Association, Lenexa, Kansas) from February 1 to June 30 in both 2019 (before COVID-19) and 2020 (during COVID-19). In response to COVID-19, hospitals participating in PROSPECT increased the efficiency of data centralization and reporting in 2020 during the period February 1 to June 30 to expedite analysis and dissemination of findings.

The main outcome measures were the percentage of change in weekly encounters (inpatient bed-days, emergency department [ED] visits, and surgeries) and inflation-adjusted charges (categorized as inpatient care and outpatient care, such as ambulatory surgery, clinics, and ED visits) before vs during COVID-19. Number of encounters and charges were compared using the Wilcoxon signed-rank test. The distribution of weekly change in outcome measures from 2019 vs 2020 across hospitals was reported with medians and interquartile ranges (IQRs). The threshold of statistical significance was set at P < .05. All analyses were performed with SAS version 9.4 (SAS Institute). This study was considered exempt from human subjects research by the Institutional Review Board of Children’s Mercy Hospital (Kansas City, Missouri).

All 26 children’s hospitals experienced similar trends in healthcare encounters and charges during the study period (Figure and Table). From February 1 to March 10, 2020, the volume of healthcare encounters in the children’s hospitals remained the same as that for the same period in 2019 (P > .1) (Figure).

Charges that accrued from February 1 to June 30 were lower in 2020 by a median 23.6% (IQR, –28.7% to –19.1%) per children’s hospital than they were in 2019, corresponding to a median decrease of $276.3 million (IQR, $404.0-$126.0 million) in charges per hospital (Table). Forty percent of this decrease was attributable to decreased charges resulting from fewer inpatient healthcare encounters.

During the initial phase of the COVID-19 pandemic in the United States, children’s hospitals experienced a substantial decrease in healthcare encounters and charges. Greater decreases were observed for ED visits and surgery encounters than for inpatient bed-days. Nonetheless, inpatient bed-days decreased by more than one-third, consistent with the decrease in inpatient resource use reported for adult hospitals.³ Remarkably, these trends were consistent across children’s hospitals, despite variation in the content and installation of and adherence with social distancing policies in their surrounding local areas.

These findings beg the question of how well children’s hospitals are positioned to weather a recurrent surge in COVID-19. Because the severity of illness of COVID-19 has been lower to date in the pediatric vs adult populations, an increase in COVID-19-related visits to EDs and admissions to offset the decreased resource use of other pediatric healthcare problems is not anticipated. Existing hospital financial reserves as well as federal aid from the Coronavirus Aid, Relief, and Economic Security Act that helped mitigate the initial encounter and financial losses during the beginning of COVID-19 may not be readily available over time.^4,5 Certainly, the findings from the current study support continued lobbying for additional state and federal funds allocated through future relief packages to children’s hospitals.

Additional approaches to financial solvency in children’s hospitals during the sustained COVID-19 pandemic include addressing surgical backlogs and sharing best practices for safe and sustained reopening of clinical operations and financial practices across institutions. Although the PROSPECT database does not contain information on the types of surgeries present within this backlog, our experiences suggest that both same-day and inpatient elective surgeries have been affected, especially lengthy procedures (eg, spinal fusion for neuromuscular scoliosis). Spread and scale of feasible and efficient solutions to reengineer and expand patient capacities and throughput for operating rooms, postanesthesia recovery areas, and intensive care and floor units are needed. Enhanced analytics that accurately predict postoperative length of hospital stay, coupled with early recovery after surgery clinical protocols, could help optimize hospital bed management. Effective ways to convert hospital rooms from single to double occupancy, to manage family visitation, and to proactively test asymptomatic patients, family, and hospital staff will mitigate continued COVID-19 penetration through children’s hospitals.

One important limitation of the current study is the measurement of hospitals’ charges. The charge data were not positioned to comprehensively measure each hospital’s financial state during the COVID-19 pandemic. However, the decrease in hospital charges reported by the children’s hospitals in the current study is comparable with the financial losses reported for many adult hospitals during the pandemic.^6,7 It is important to recognize that the amount of the charges may not be equivalent to the cost of care or revenue collected by the hospitals. PROSPECT does not contain information on cost, and current cost-to-charge ratios are based on historical (ie, pre-COVID-19) data; therefore, they do not account for increased cost of personal protective equipment and other related costs that occurred during the pandemic, which makes use of these ratios challenging. Nevertheless, it is possible that the relative difference in costs incurred and revenue collected before and during COVID-19 may have been similar to the differences observed in hospital charges.

Children’s hospitals’ ability to serve the nation’s pediatric patients depends on the success of the hospitals’ plans to manage current and future COVID-19 surges and to reopen and recover from the surges that have passed. Additional investigation is needed to identify best operational and financial practices among children’s hospitals that have enabled them to endure the COVID-19 pandemic.

To benefit patients and the public health of their communities, children’s hospitals across the United States prepared for and responded to COVID-19 by conserving personal protective equipment, suspending noncritical in-person healthcare encounters (including outpatient visits and elective surgeries), and implementing socially distanced essential care.^1,2 These measures were promptly instituted during a time of both substantial uncertainty about the pandemic’s behavior in children—including its severity and duration—and extreme variation in local and state governments’ responses to the pandemic.

Congruent with other healthcare institutions, children’s hospitals calibrated their clinical operations to the evolving nature of the pandemic, prioritizing the safety of patients and staff while striving to maintain financial viability in the setting of increased costs and decreased revenue. In some cases, children’s hospitals aided adult hospitals and health systems by admitting young and middle-aged adult patients and by centralizing all pediatric patients requiring intensive care within a region. These efforts occurred while many children’s hospitals remained the sole source of specialized pediatric care, including care for rare complex health problems.

As the COVID-19 pandemic continues, there is a critical need to assess how the initial phase of the pandemic affected healthcare encounters and related finances in children’s hospitals. Understanding these trends will position children’s hospitals to project and prepare for subsequent COVID-19 surges, as well as future related public health crises that necessitate widespread social distancing. Therefore, we compared year-over-year trends in healthcare encounters and hospital charges across US children’s hospitals before and during the COVID-19 pandemic, focusing on the beginning of COVID-19 in the United States, which was defined as February through June 2020.

This is a retrospective analysis of 26 children’s hospitals (22 freestanding, 4 nonfreestanding) from all US regions (12 South, 7 Midwest, 5 West, 2 Northeast) contributing encounter and financial data to the PROSPECT database (Children’s Hospital Association, Lenexa, Kansas) from February 1 to June 30 in both 2019 (before COVID-19) and 2020 (during COVID-19). In response to COVID-19, hospitals participating in PROSPECT increased the efficiency of data centralization and reporting in 2020 during the period February 1 to June 30 to expedite analysis and dissemination of findings.

The main outcome measures were the percentage of change in weekly encounters (inpatient bed-days, emergency department [ED] visits, and surgeries) and inflation-adjusted charges (categorized as inpatient care and outpatient care, such as ambulatory surgery, clinics, and ED visits) before vs during COVID-19. Number of encounters and charges were compared using the Wilcoxon signed-rank test. The distribution of weekly change in outcome measures from 2019 vs 2020 across hospitals was reported with medians and interquartile ranges (IQRs). The threshold of statistical significance was set at P < .05. All analyses were performed with SAS version 9.4 (SAS Institute). This study was considered exempt from human subjects research by the Institutional Review Board of Children’s Mercy Hospital (Kansas City, Missouri).

All 26 children’s hospitals experienced similar trends in healthcare encounters and charges during the study period (Figure and Table). From February 1 to March 10, 2020, the volume of healthcare encounters in the children’s hospitals remained the same as that for the same period in 2019 (P > .1) (Figure).

Charges that accrued from February 1 to June 30 were lower in 2020 by a median 23.6% (IQR, –28.7% to –19.1%) per children’s hospital than they were in 2019, corresponding to a median decrease of $276.3 million (IQR, $404.0-$126.0 million) in charges per hospital (Table). Forty percent of this decrease was attributable to decreased charges resulting from fewer inpatient healthcare encounters.

During the initial phase of the COVID-19 pandemic in the United States, children’s hospitals experienced a substantial decrease in healthcare encounters and charges. Greater decreases were observed for ED visits and surgery encounters than for inpatient bed-days. Nonetheless, inpatient bed-days decreased by more than one-third, consistent with the decrease in inpatient resource use reported for adult hospitals.³ Remarkably, these trends were consistent across children’s hospitals, despite variation in the content and installation of and adherence with social distancing policies in their surrounding local areas.

These findings beg the question of how well children’s hospitals are positioned to weather a recurrent surge in COVID-19. Because the severity of illness of COVID-19 has been lower to date in the pediatric vs adult populations, an increase in COVID-19-related visits to EDs and admissions to offset the decreased resource use of other pediatric healthcare problems is not anticipated. Existing hospital financial reserves as well as federal aid from the Coronavirus Aid, Relief, and Economic Security Act that helped mitigate the initial encounter and financial losses during the beginning of COVID-19 may not be readily available over time.^4,5 Certainly, the findings from the current study support continued lobbying for additional state and federal funds allocated through future relief packages to children’s hospitals.

Additional approaches to financial solvency in children’s hospitals during the sustained COVID-19 pandemic include addressing surgical backlogs and sharing best practices for safe and sustained reopening of clinical operations and financial practices across institutions. Although the PROSPECT database does not contain information on the types of surgeries present within this backlog, our experiences suggest that both same-day and inpatient elective surgeries have been affected, especially lengthy procedures (eg, spinal fusion for neuromuscular scoliosis). Spread and scale of feasible and efficient solutions to reengineer and expand patient capacities and throughput for operating rooms, postanesthesia recovery areas, and intensive care and floor units are needed. Enhanced analytics that accurately predict postoperative length of hospital stay, coupled with early recovery after surgery clinical protocols, could help optimize hospital bed management. Effective ways to convert hospital rooms from single to double occupancy, to manage family visitation, and to proactively test asymptomatic patients, family, and hospital staff will mitigate continued COVID-19 penetration through children’s hospitals.

One important limitation of the current study is the measurement of hospitals’ charges. The charge data were not positioned to comprehensively measure each hospital’s financial state during the COVID-19 pandemic. However, the decrease in hospital charges reported by the children’s hospitals in the current study is comparable with the financial losses reported for many adult hospitals during the pandemic.^6,7 It is important to recognize that the amount of the charges may not be equivalent to the cost of care or revenue collected by the hospitals. PROSPECT does not contain information on cost, and current cost-to-charge ratios are based on historical (ie, pre-COVID-19) data; therefore, they do not account for increased cost of personal protective equipment and other related costs that occurred during the pandemic, which makes use of these ratios challenging. Nevertheless, it is possible that the relative difference in costs incurred and revenue collected before and during COVID-19 may have been similar to the differences observed in hospital charges.

Children’s hospitals’ ability to serve the nation’s pediatric patients depends on the success of the hospitals’ plans to manage current and future COVID-19 surges and to reopen and recover from the surges that have passed. Additional investigation is needed to identify best operational and financial practices among children’s hospitals that have enabled them to endure the COVID-19 pandemic.

It is clear that certain patient-level factors, such as age, sex, and comorbidities, predict outcomes of SARS-CoV-2 infection.^1,2 Less is known about whether hospital-level factors, including surges of patients with COVID-19, are associated with patient outcomes.

In a multicenter cohort study of 2,215 patients with COVID-19 in 65 intensive care units (ICU) across the United States, mortality rates varied widely (6.6%-80.8%), with improved survival for patients admitted to a hospital with more (>100) rather than fewer (<50) ICU beds.³ A different study found that at the state level, COVID-19 mortality increased with increasing COVID-19 admissions.⁴ Together, these studies suggest that surges in COVID-19 patient volume may be associated with excess mortality. However, the first study was restricted to the ICU population, limiting generalizability, and did not consider admission volume, only ICU bed count. Meanwhile, the second study considered both hospital capacity and patient volume, but it describes a relatively small sample, did not adjust for patient-level predictors of mortality, and does not report outcomes at the hospital level.

Here, we used a large dataset to compare in-hospital mortality rates for patients with COVID-19 across US hospitals, hypothesizing that mortality would be higher in hospitals with the highest burden of COVID-19 admissions. By adjusting for patient-level predictors of mortality and normalizing admission volume for hospital size, we are able to describe residual variability in mortality that may be attributable to differences in COVID-19 patient volume.

We included patients with an International Statistical Classification of Diseases, Tenth Revision (ICD)-10 diagnosis of COVID-19 (U07.1) who were admitted to a US hospital that contracts with CarePort Health.⁵ CarePort is a platform for discharge planning and care coordination that contracts with hospitals in all US regions and auto-extracts data using interface feeds.

We restricted the population to patients admitted between April 1 and April 30, 2020, after a new ICD-10 code for confirmed COVID-19 infection became available, and to hospitals that provided real-time ICD-10 data and pertinent demographic information and could be linked to Centers for Medicare & Medicaid Services (CMS) data by National Provider Identifier. We assumed that the 145 patients (1.0%) who remained hospitalized at 5 weeks all survived. For the 5.9% of patients with multiple admissions during the study period, we included only the first admission with a diagnosis code for COVID-19.

We adjusted for patient age, sex, and the 31 comorbidities in the Elixhauser index, defined by ICD-10 codes. This set of comorbidities includes those previously associated with COVID-19 survival.^1,2,6 Unfortunately, inconsistent reporting of vital signs and laboratory data precluded adjusting for acute illness severity. For those patients whose residence zip code was known, we report the racial breakdown (White vs non-White) and adjusted gross income (AGI), based on linked information from the 2018 American Community Survey.⁷

We defined COVID-19 burden as the quotient of COVID-19 admissions in April 2020 and each hospital’s certified bed count, as reported to the CMS.⁸ This allowed us to normalize COVID-19 patient volume for variation in hospital size, acknowledging that admitting 10 patients with COVID-19 to a 1,000-bed hospital is different from admitting 10 patients with COVID-19 to a 20-bed hospital. Certified bed count seemed the ideal denominator because it excludes beds not readily deployable to care for patients with COVID-19 (eg, radiology suites, labor and delivery rooms).

We computed hospital-specific adjusted mortality proportions and 95% confidence intervals based on hierarchical multivariable logistic regression, adjusting for age, sex, and comorbidities, and a random effect for each hospital.^9,10 Hypothesizing that there may be a threshold of burden beyond which mortality begins to rise, we compared the in-hospital mortality rate at hospitals in the highest quintile of COVID-19 burden to all other hospitals.

We conducted eight post-hoc sensitivity analyses: (1) restricting the study population to patients aged 75 years and older; (2) restricting study hospitals to those with at least 100 beds and 20 COVID-19 admissions; (3) assuming that all patients who remained hospitalized at 5 weeks had died; (4) using each patient’s last admission during the month of April rather than the first; sequentially incorporating (5) zip code–level information on race (limited to White, non-White) and (6) AGI (treated as a continuous variable) into our model; (7) computing two burdens for each hospital (one for each half of April) and using whichever was higher; and (8) treating COVID-19 burden as a continuous predictor. Analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc) using the GLIMMIX procedure. This study was deemed exempt by the University of California, San Francisco Institutional Review Board.

The study population included 14,226 patients with COVID-19 (median age, 66 years [range, 0-110 years]; 45.2% women) at 117 US hospitals. Based on patients’ zip code of residence, we estimate that 47.0% of patients were White and 29.1% Black, and that the mean household AGI was $61,956. Most hospitals were nonprofit (56%) or private (39%), with approximately one quarter coming from each US census region (range, 25 hospitals [21%] in Midwest to 33 hospitals [28%] in Northeast). Nine hospitals (8%) had more than 700 beds, 40 (34%) had 300 to 700 beds, and 68 (58%) had fewer than 300 beds. Thirty-six hospitals (30.8%) admitted fewer than 20 patients with COVID-19, while six hospitals (5.1%) admitted more than 500 such patients. COVID burden ranged from 0.004 to 2.03 admissions per bed.

As of June 5, 2020, 78.1% of patients had been discharged alive, 20.9% had died, and 1.0% remained hospitalized. At the hospital level, the observed mortality ranged from 0% to 44.4%, was 17.1% among hospitals in COVID-19 burden quintiles one through four, and was 22.7% in the highest burden quintile (Table).

Results were similar across multiple sensitivity analyses (see Appendix Table), although the relationship between COVID-19 burden and in-hospital mortality was attenuated and not significant when the sample was restricted to hospitals with at least 100 beds and 20 COVID-19 admissions, or in analyses adjusted for race and AGI.

In this study of 14,226 patients with COVID-19 across 117 US hospitals, those patients admitted to the most burdened hospitals had a higher odds of death. This relationship, which persisted after adjusting for age, sex, and comorbid conditions, suggests that a threshold exists at which patient surges may cause excess mortality.

Notably, in sensitivity analyses adjusting for race and AGI, COVID-19 burden was no longer associated with in-hospital mortality and the point estimate was attenuated. This raises the possibility that our primary results are confounded by these factors. However, prior studies of hospitalized patients have not found race to be predictive of mortality, after adjusting for other factors.^11,12

We also note that the relationship between COVID-19 burden and mortality was not significant (P = .07) when the sample was restricted to larger hospitals with more than 20 COVID-19 admissions; again, the point estimate was attenuated. This suggests that larger hospitals may be more resilient in the face of patient surges. Whether this is due to increased availability of staff who can be redeployed to patient care (as with researchers at academic centers), increased experience managing severe respiratory failure, or other factors is uncertain.

Interestingly, in-hospital mortality varied widely across study hospitals, even among the most-burdened hospitals. The reasons for this residual variability—after adjusting for age, sex, and comorbidities and stratifying by COVID-19 burden—are uncertain. To the extent that this variability reflects differences in patient management, hospital staffing, or use of investigational or advanced therapies, it will be critical to identify and disseminate any replicable best practices from high-burden hospitals with low mortality rates.

Whereas other reports have often described single-center or regional experiences,^13-15 leaving open the possibility that their results were highly influenced by the local nature of the pandemic in their respective settings, our report from a large sample of hospitals across the United States in high- and low-burden settings provides a more generalizable description of mortality rates for hospitalized patients. Additional study strengths include our adjustment for comorbidities known to be associated with COVID-19 survival, the reporting of definitive outcomes for 99% of patients, and the inclusion of multiple sensitivity analyses to assess the stability of findings.

Our principal limitation is the inability to adjust for severity of acute illness due to inconsistent reporting of laboratory and vital signs data from study hospitals and missing information on interhospital transfers. While our adjusted analyses clearly suggest an association between COVID-19 burden and patient outcomes, our results may still be confounded by differences in illness severity at study hospitals. Thus, our findings should be considered hypothesis-generating and will require confirmation in future studies that include adjustment for acute illness severity.

Other limitations of our study include overrepresentation of large urban hospitals in the Northeast, although this represents the geography of the US pandemic during the study period. Our adjustment for race/ethnicity and socioeconomic status was limited in that we only had zip code-of-residence level information, did not know the zip code of residence for one quarter of study patients, and had to bifurcate the population into White/non-White categories. Finally, our definition of burden does not account for hospital resources, including staffing, ICU capacity, and the availability of advanced or investigational therapies.

In this study of 14,226 patients with COVID-19 admitted to 1 of 117 US hospitals, we found that the odds of in-hospital mortality were higher in hospitals that had the highest burden of COVID-19 admissions. This relationship, which persisted after adjustment for age, sex, and comorbid conditions, suggests that patient surges may be an independent risk factor for in-hospital death among patients with COVID-19.

The authors thank Bocheng Jing, MS, Senior Statistician at the UCSF Pepper Center, for providing code to identify Elixhauser conditions from ICD-10 data; and Scott Kerber, BS, and Scott Magnoni, MS, both of CarePort Health, for assistance with data extraction. They were not compensated for this work beyond their regular salaries.

It is clear that certain patient-level factors, such as age, sex, and comorbidities, predict outcomes of SARS-CoV-2 infection.^1,2 Less is known about whether hospital-level factors, including surges of patients with COVID-19, are associated with patient outcomes.

In a multicenter cohort study of 2,215 patients with COVID-19 in 65 intensive care units (ICU) across the United States, mortality rates varied widely (6.6%-80.8%), with improved survival for patients admitted to a hospital with more (>100) rather than fewer (<50) ICU beds.³ A different study found that at the state level, COVID-19 mortality increased with increasing COVID-19 admissions.⁴ Together, these studies suggest that surges in COVID-19 patient volume may be associated with excess mortality. However, the first study was restricted to the ICU population, limiting generalizability, and did not consider admission volume, only ICU bed count. Meanwhile, the second study considered both hospital capacity and patient volume, but it describes a relatively small sample, did not adjust for patient-level predictors of mortality, and does not report outcomes at the hospital level.

Here, we used a large dataset to compare in-hospital mortality rates for patients with COVID-19 across US hospitals, hypothesizing that mortality would be higher in hospitals with the highest burden of COVID-19 admissions. By adjusting for patient-level predictors of mortality and normalizing admission volume for hospital size, we are able to describe residual variability in mortality that may be attributable to differences in COVID-19 patient volume.

We included patients with an International Statistical Classification of Diseases, Tenth Revision (ICD)-10 diagnosis of COVID-19 (U07.1) who were admitted to a US hospital that contracts with CarePort Health.⁵ CarePort is a platform for discharge planning and care coordination that contracts with hospitals in all US regions and auto-extracts data using interface feeds.

We restricted the population to patients admitted between April 1 and April 30, 2020, after a new ICD-10 code for confirmed COVID-19 infection became available, and to hospitals that provided real-time ICD-10 data and pertinent demographic information and could be linked to Centers for Medicare & Medicaid Services (CMS) data by National Provider Identifier. We assumed that the 145 patients (1.0%) who remained hospitalized at 5 weeks all survived. For the 5.9% of patients with multiple admissions during the study period, we included only the first admission with a diagnosis code for COVID-19.

We adjusted for patient age, sex, and the 31 comorbidities in the Elixhauser index, defined by ICD-10 codes. This set of comorbidities includes those previously associated with COVID-19 survival.^1,2,6 Unfortunately, inconsistent reporting of vital signs and laboratory data precluded adjusting for acute illness severity. For those patients whose residence zip code was known, we report the racial breakdown (White vs non-White) and adjusted gross income (AGI), based on linked information from the 2018 American Community Survey.⁷

We defined COVID-19 burden as the quotient of COVID-19 admissions in April 2020 and each hospital’s certified bed count, as reported to the CMS.⁸ This allowed us to normalize COVID-19 patient volume for variation in hospital size, acknowledging that admitting 10 patients with COVID-19 to a 1,000-bed hospital is different from admitting 10 patients with COVID-19 to a 20-bed hospital. Certified bed count seemed the ideal denominator because it excludes beds not readily deployable to care for patients with COVID-19 (eg, radiology suites, labor and delivery rooms).

We computed hospital-specific adjusted mortality proportions and 95% confidence intervals based on hierarchical multivariable logistic regression, adjusting for age, sex, and comorbidities, and a random effect for each hospital.^9,10 Hypothesizing that there may be a threshold of burden beyond which mortality begins to rise, we compared the in-hospital mortality rate at hospitals in the highest quintile of COVID-19 burden to all other hospitals.

We conducted eight post-hoc sensitivity analyses: (1) restricting the study population to patients aged 75 years and older; (2) restricting study hospitals to those with at least 100 beds and 20 COVID-19 admissions; (3) assuming that all patients who remained hospitalized at 5 weeks had died; (4) using each patient’s last admission during the month of April rather than the first; sequentially incorporating (5) zip code–level information on race (limited to White, non-White) and (6) AGI (treated as a continuous variable) into our model; (7) computing two burdens for each hospital (one for each half of April) and using whichever was higher; and (8) treating COVID-19 burden as a continuous predictor. Analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc) using the GLIMMIX procedure. This study was deemed exempt by the University of California, San Francisco Institutional Review Board.

The study population included 14,226 patients with COVID-19 (median age, 66 years [range, 0-110 years]; 45.2% women) at 117 US hospitals. Based on patients’ zip code of residence, we estimate that 47.0% of patients were White and 29.1% Black, and that the mean household AGI was $61,956. Most hospitals were nonprofit (56%) or private (39%), with approximately one quarter coming from each US census region (range, 25 hospitals [21%] in Midwest to 33 hospitals [28%] in Northeast). Nine hospitals (8%) had more than 700 beds, 40 (34%) had 300 to 700 beds, and 68 (58%) had fewer than 300 beds. Thirty-six hospitals (30.8%) admitted fewer than 20 patients with COVID-19, while six hospitals (5.1%) admitted more than 500 such patients. COVID burden ranged from 0.004 to 2.03 admissions per bed.

As of June 5, 2020, 78.1% of patients had been discharged alive, 20.9% had died, and 1.0% remained hospitalized. At the hospital level, the observed mortality ranged from 0% to 44.4%, was 17.1% among hospitals in COVID-19 burden quintiles one through four, and was 22.7% in the highest burden quintile (Table).

Results were similar across multiple sensitivity analyses (see Appendix Table), although the relationship between COVID-19 burden and in-hospital mortality was attenuated and not significant when the sample was restricted to hospitals with at least 100 beds and 20 COVID-19 admissions, or in analyses adjusted for race and AGI.

In this study of 14,226 patients with COVID-19 across 117 US hospitals, those patients admitted to the most burdened hospitals had a higher odds of death. This relationship, which persisted after adjusting for age, sex, and comorbid conditions, suggests that a threshold exists at which patient surges may cause excess mortality.

Notably, in sensitivity analyses adjusting for race and AGI, COVID-19 burden was no longer associated with in-hospital mortality and the point estimate was attenuated. This raises the possibility that our primary results are confounded by these factors. However, prior studies of hospitalized patients have not found race to be predictive of mortality, after adjusting for other factors.^11,12

We also note that the relationship between COVID-19 burden and mortality was not significant (P = .07) when the sample was restricted to larger hospitals with more than 20 COVID-19 admissions; again, the point estimate was attenuated. This suggests that larger hospitals may be more resilient in the face of patient surges. Whether this is due to increased availability of staff who can be redeployed to patient care (as with researchers at academic centers), increased experience managing severe respiratory failure, or other factors is uncertain.

Interestingly, in-hospital mortality varied widely across study hospitals, even among the most-burdened hospitals. The reasons for this residual variability—after adjusting for age, sex, and comorbidities and stratifying by COVID-19 burden—are uncertain. To the extent that this variability reflects differences in patient management, hospital staffing, or use of investigational or advanced therapies, it will be critical to identify and disseminate any replicable best practices from high-burden hospitals with low mortality rates.

Whereas other reports have often described single-center or regional experiences,^13-15 leaving open the possibility that their results were highly influenced by the local nature of the pandemic in their respective settings, our report from a large sample of hospitals across the United States in high- and low-burden settings provides a more generalizable description of mortality rates for hospitalized patients. Additional study strengths include our adjustment for comorbidities known to be associated with COVID-19 survival, the reporting of definitive outcomes for 99% of patients, and the inclusion of multiple sensitivity analyses to assess the stability of findings.

Our principal limitation is the inability to adjust for severity of acute illness due to inconsistent reporting of laboratory and vital signs data from study hospitals and missing information on interhospital transfers. While our adjusted analyses clearly suggest an association between COVID-19 burden and patient outcomes, our results may still be confounded by differences in illness severity at study hospitals. Thus, our findings should be considered hypothesis-generating and will require confirmation in future studies that include adjustment for acute illness severity.

Other limitations of our study include overrepresentation of large urban hospitals in the Northeast, although this represents the geography of the US pandemic during the study period. Our adjustment for race/ethnicity and socioeconomic status was limited in that we only had zip code-of-residence level information, did not know the zip code of residence for one quarter of study patients, and had to bifurcate the population into White/non-White categories. Finally, our definition of burden does not account for hospital resources, including staffing, ICU capacity, and the availability of advanced or investigational therapies.

In this study of 14,226 patients with COVID-19 admitted to 1 of 117 US hospitals, we found that the odds of in-hospital mortality were higher in hospitals that had the highest burden of COVID-19 admissions. This relationship, which persisted after adjustment for age, sex, and comorbid conditions, suggests that patient surges may be an independent risk factor for in-hospital death among patients with COVID-19.

The authors thank Bocheng Jing, MS, Senior Statistician at the UCSF Pepper Center, for providing code to identify Elixhauser conditions from ICD-10 data; and Scott Kerber, BS, and Scott Magnoni, MS, both of CarePort Health, for assistance with data extraction. They were not compensated for this work beyond their regular salaries.

It is clear that certain patient-level factors, such as age, sex, and comorbidities, predict outcomes of SARS-CoV-2 infection.^1,2 Less is known about whether hospital-level factors, including surges of patients with COVID-19, are associated with patient outcomes.

In a multicenter cohort study of 2,215 patients with COVID-19 in 65 intensive care units (ICU) across the United States, mortality rates varied widely (6.6%-80.8%), with improved survival for patients admitted to a hospital with more (>100) rather than fewer (<50) ICU beds.³ A different study found that at the state level, COVID-19 mortality increased with increasing COVID-19 admissions.⁴ Together, these studies suggest that surges in COVID-19 patient volume may be associated with excess mortality. However, the first study was restricted to the ICU population, limiting generalizability, and did not consider admission volume, only ICU bed count. Meanwhile, the second study considered both hospital capacity and patient volume, but it describes a relatively small sample, did not adjust for patient-level predictors of mortality, and does not report outcomes at the hospital level.

Here, we used a large dataset to compare in-hospital mortality rates for patients with COVID-19 across US hospitals, hypothesizing that mortality would be higher in hospitals with the highest burden of COVID-19 admissions. By adjusting for patient-level predictors of mortality and normalizing admission volume for hospital size, we are able to describe residual variability in mortality that may be attributable to differences in COVID-19 patient volume.

We included patients with an International Statistical Classification of Diseases, Tenth Revision (ICD)-10 diagnosis of COVID-19 (U07.1) who were admitted to a US hospital that contracts with CarePort Health.⁵ CarePort is a platform for discharge planning and care coordination that contracts with hospitals in all US regions and auto-extracts data using interface feeds.

We restricted the population to patients admitted between April 1 and April 30, 2020, after a new ICD-10 code for confirmed COVID-19 infection became available, and to hospitals that provided real-time ICD-10 data and pertinent demographic information and could be linked to Centers for Medicare & Medicaid Services (CMS) data by National Provider Identifier. We assumed that the 145 patients (1.0%) who remained hospitalized at 5 weeks all survived. For the 5.9% of patients with multiple admissions during the study period, we included only the first admission with a diagnosis code for COVID-19.

We adjusted for patient age, sex, and the 31 comorbidities in the Elixhauser index, defined by ICD-10 codes. This set of comorbidities includes those previously associated with COVID-19 survival.^1,2,6 Unfortunately, inconsistent reporting of vital signs and laboratory data precluded adjusting for acute illness severity. For those patients whose residence zip code was known, we report the racial breakdown (White vs non-White) and adjusted gross income (AGI), based on linked information from the 2018 American Community Survey.⁷

We defined COVID-19 burden as the quotient of COVID-19 admissions in April 2020 and each hospital’s certified bed count, as reported to the CMS.⁸ This allowed us to normalize COVID-19 patient volume for variation in hospital size, acknowledging that admitting 10 patients with COVID-19 to a 1,000-bed hospital is different from admitting 10 patients with COVID-19 to a 20-bed hospital. Certified bed count seemed the ideal denominator because it excludes beds not readily deployable to care for patients with COVID-19 (eg, radiology suites, labor and delivery rooms).

We computed hospital-specific adjusted mortality proportions and 95% confidence intervals based on hierarchical multivariable logistic regression, adjusting for age, sex, and comorbidities, and a random effect for each hospital.^9,10 Hypothesizing that there may be a threshold of burden beyond which mortality begins to rise, we compared the in-hospital mortality rate at hospitals in the highest quintile of COVID-19 burden to all other hospitals.

We conducted eight post-hoc sensitivity analyses: (1) restricting the study population to patients aged 75 years and older; (2) restricting study hospitals to those with at least 100 beds and 20 COVID-19 admissions; (3) assuming that all patients who remained hospitalized at 5 weeks had died; (4) using each patient’s last admission during the month of April rather than the first; sequentially incorporating (5) zip code–level information on race (limited to White, non-White) and (6) AGI (treated as a continuous variable) into our model; (7) computing two burdens for each hospital (one for each half of April) and using whichever was higher; and (8) treating COVID-19 burden as a continuous predictor. Analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc) using the GLIMMIX procedure. This study was deemed exempt by the University of California, San Francisco Institutional Review Board.

The study population included 14,226 patients with COVID-19 (median age, 66 years [range, 0-110 years]; 45.2% women) at 117 US hospitals. Based on patients’ zip code of residence, we estimate that 47.0% of patients were White and 29.1% Black, and that the mean household AGI was $61,956. Most hospitals were nonprofit (56%) or private (39%), with approximately one quarter coming from each US census region (range, 25 hospitals [21%] in Midwest to 33 hospitals [28%] in Northeast). Nine hospitals (8%) had more than 700 beds, 40 (34%) had 300 to 700 beds, and 68 (58%) had fewer than 300 beds. Thirty-six hospitals (30.8%) admitted fewer than 20 patients with COVID-19, while six hospitals (5.1%) admitted more than 500 such patients. COVID burden ranged from 0.004 to 2.03 admissions per bed.

As of June 5, 2020, 78.1% of patients had been discharged alive, 20.9% had died, and 1.0% remained hospitalized. At the hospital level, the observed mortality ranged from 0% to 44.4%, was 17.1% among hospitals in COVID-19 burden quintiles one through four, and was 22.7% in the highest burden quintile (Table).

Results were similar across multiple sensitivity analyses (see Appendix Table), although the relationship between COVID-19 burden and in-hospital mortality was attenuated and not significant when the sample was restricted to hospitals with at least 100 beds and 20 COVID-19 admissions, or in analyses adjusted for race and AGI.

In this study of 14,226 patients with COVID-19 across 117 US hospitals, those patients admitted to the most burdened hospitals had a higher odds of death. This relationship, which persisted after adjusting for age, sex, and comorbid conditions, suggests that a threshold exists at which patient surges may cause excess mortality.

Notably, in sensitivity analyses adjusting for race and AGI, COVID-19 burden was no longer associated with in-hospital mortality and the point estimate was attenuated. This raises the possibility that our primary results are confounded by these factors. However, prior studies of hospitalized patients have not found race to be predictive of mortality, after adjusting for other factors.^11,12

We also note that the relationship between COVID-19 burden and mortality was not significant (P = .07) when the sample was restricted to larger hospitals with more than 20 COVID-19 admissions; again, the point estimate was attenuated. This suggests that larger hospitals may be more resilient in the face of patient surges. Whether this is due to increased availability of staff who can be redeployed to patient care (as with researchers at academic centers), increased experience managing severe respiratory failure, or other factors is uncertain.

Interestingly, in-hospital mortality varied widely across study hospitals, even among the most-burdened hospitals. The reasons for this residual variability—after adjusting for age, sex, and comorbidities and stratifying by COVID-19 burden—are uncertain. To the extent that this variability reflects differences in patient management, hospital staffing, or use of investigational or advanced therapies, it will be critical to identify and disseminate any replicable best practices from high-burden hospitals with low mortality rates.

Whereas other reports have often described single-center or regional experiences,^13-15 leaving open the possibility that their results were highly influenced by the local nature of the pandemic in their respective settings, our report from a large sample of hospitals across the United States in high- and low-burden settings provides a more generalizable description of mortality rates for hospitalized patients. Additional study strengths include our adjustment for comorbidities known to be associated with COVID-19 survival, the reporting of definitive outcomes for 99% of patients, and the inclusion of multiple sensitivity analyses to assess the stability of findings.

Our principal limitation is the inability to adjust for severity of acute illness due to inconsistent reporting of laboratory and vital signs data from study hospitals and missing information on interhospital transfers. While our adjusted analyses clearly suggest an association between COVID-19 burden and patient outcomes, our results may still be confounded by differences in illness severity at study hospitals. Thus, our findings should be considered hypothesis-generating and will require confirmation in future studies that include adjustment for acute illness severity.

Other limitations of our study include overrepresentation of large urban hospitals in the Northeast, although this represents the geography of the US pandemic during the study period. Our adjustment for race/ethnicity and socioeconomic status was limited in that we only had zip code-of-residence level information, did not know the zip code of residence for one quarter of study patients, and had to bifurcate the population into White/non-White categories. Finally, our definition of burden does not account for hospital resources, including staffing, ICU capacity, and the availability of advanced or investigational therapies.

In this study of 14,226 patients with COVID-19 admitted to 1 of 117 US hospitals, we found that the odds of in-hospital mortality were higher in hospitals that had the highest burden of COVID-19 admissions. This relationship, which persisted after adjustment for age, sex, and comorbid conditions, suggests that patient surges may be an independent risk factor for in-hospital death among patients with COVID-19.

The authors thank Bocheng Jing, MS, Senior Statistician at the UCSF Pepper Center, for providing code to identify Elixhauser conditions from ICD-10 data; and Scott Kerber, BS, and Scott Magnoni, MS, both of CarePort Health, for assistance with data extraction. They were not compensated for this work beyond their regular salaries.

The World Health Organization (WHO) defines anemia as a hemoglobin value less than 12 g/dL in women and less than 13 g/dL in men.¹ Hospital-acquired anemia is loosely defined as normal hemoglobin levels on admission that, at their nadir during hospitalization or on discharge, are less than WHO sex-defined cutoffs. Hospital-acquired anemia or significant decreases in hemoglobin are often identified during hospitalization.^2-6 Potential causes include blood loss from phlebotomy, occult gastrointestinal bleeding, hemolysis, anemia of inflammation, and hemodilution due to fluid resuscitation. Of these causes, some are dangerous to patients, some are iatrogenic, and some are due to laboratory error.⁷ Physicians often evaluate decreases in hemoglobin, which could otherwise be explained by laboratory error, hemodilution, or expected decrease in hemoglobin due to hospitalization, to identify causes that may lead to potential harm.

Jacob et al⁸ demonstrated the effect of posture on hemoglobin concentrations in healthy volunteers, showing an average 11% relative increase in hemoglobin when going from lying to standing. This increase was attributed to shifts in plasma volume to the vascular space with recumbence. They hypothesized that the initial hemoglobin on admission is measured when patients are upright or recently upright, whereas after admission, patients are more likely to be supine, resulting in lower hemoglobin concentrations. Others have also demonstrated similar effects of patient posture on hemoglobin concentration.^9-13 However, these prior results are not readily generalizable to hospitalized patients. These prior studies enrolled healthy volunteers, and most examined postural changes from the supine and standing positions; blood is rarely obtained from hospitalized patients when they are standing.

The aim of this study was to investigate whether postural changes in hemoglobin can be demonstrated in positions that patients routinely encountered during in-hospital phlebotomy: upright in a chair or recumbent in a bed. Patient position, which is not standardized during blood draws, may contribute to lower measured hemoglobin concentrations in some patients, especially sicker individuals who are recumbent more frequently. We hypothesized that going from supine to upright in a chair would result in a relative increase in hemoglobin concentration of 5% to 6%, approximately half the value of going from supine to standing.⁸ To investigate this, we conducted a quasi-experimental study exploring the effect of position (supine or sitting in chair) on hemoglobin concentrations in medical inpatients.

Participants

Patients were enrolled in this single-center study between October 2017 and August 2018. Patients aged 18 years or older who were hospitalized on the general internal medicine wards were screened to determine if they met the following inclusion criteria: hospitalized for <5 days, had blood work scheduled as part of routine care (in order to decrease phlebotomy required by this study), had baseline hemoglobin >8 g/dL, and were able to remain supine without interruption overnight and able to sit in a chair for at least 1 hour the following morning. Patients were excluded from the study if they had a hematologic malignancy, were at risk of >100 mL of blood loss (eg, admitted for gastrointestinal bleeding, planned surgery), had a transfusion requirement, or received intravascular modifiers such as fluid (>100 cc/h) or intravenous diuretics. The Johns Hopkins Institutional Review Board approved this study, and all patients provided written informed consent.

Study Design

Patients enrolled in this quasi-experimental study were asked to remain supine for at least 6 hours overnight. Adherence to the recumbent position was tracked by patient self-report and by corroboration with the patient’s nurse overnight. Any interruptions to supine positioning resulted in exclusion from the study. The following morning, a member of the study team performed phlebotomy while the patient remained supine. Patients were then asked to sit comfortably in a chair for at least 1 hour with their feet on the ground; the blood draw was then repeated. All blood samples were acquired by venipuncture. Prior to each blood draw, a tourniquet was placed over the upper arm below the axilla. An antecubital vein on either arm was visualized under ultrasound guidance, and a 23-G × 3/4” butterfly needle was used for venipuncture. The vials of blood were immediately inverted after blood collection. Hemoglobin assays were processed and analyzed using Sysmex XN-10 analyzer (Sysmex Corporation). The reference range for hemoglobin in our facility was 12.0 to 15.0 g/dL for women and 13.9 to 16.3 g/dL for men. Laboratory technicians were blinded to and uninvolved in the study.

We determined, a priori, that 33 enrolled patients would provide 80% power (alpha 0.05) to detect an average hemoglobin change of 4.1%, assuming that the standard deviation of the hemoglobin change was twice the mean (ie, SD = 8.2%). The Wilcoxon signed-rank test was used to test the significance of postural hemoglobin changes. Analyses were conducted using JMP Pro 13.0 (SAS) and GraphPad Prism 8 (GraphPad Software). Significance was defined at P < .05 for all analyses.

Thirty-nine patients were consented and enrolled in the study; four patients were excluded prior to blood draw (two patients because of interruption of supine time, two patients because of refusal in the morning). Of the 35 patients who completed the study, 13 were women (37%); median age was 49 years (range, 25-83 years). Median supine hemoglobin concentration in our sample was 11.7 g/dL (range, 9.3-18.1 g/dL), and median baseline creatinine level was 0.70 mg/dL (range, 0.5-2.5 mg/dL). Median supine hemoglobin levels were 11.7 g/dL (range, 9.6-13.2 g/dL) in women and 11.8 g/dL (range, 9.3-18.1 g/dL) in men. In aggregate, patients had a median increase in hemoglobin concentration of 0.60 g/dL (range, –0.6 to 1.4 g/dL) with sitting, a 5.2% (range, –4.5% to 15.1%) relative change (P < .001) (Figure 1).

International blood collection guidelines acknowledge postural changes in laboratory values and recommend standardization of patient position to either sitting in a chair or lying flat in a bed, without changes in position for 15 minutes prior to blood draw.¹⁴ When these positional accommodations cannot be met, documenting positional disruptions is recommended so that laboratory values can be interpreted accordingly. To the best of our knowledge, no hospital in the United States has standardized patient position as part of phlebotomy procedure such that patient position is documented and can be made available to interpreting providers.

Relative increases in hemoglobin or hematocrit range from 7% to 12% when patients go from supine to standing.^8,9,11 The reverse relationship has also been shown, where upright-to-supine position results in decreases in hemoglobin concentrations.^10,13 We found that going from supine to a seated position resulted in significant increases in hemoglobin of 0.6 g/dL and in a more than 1 g/dL increase in 29% of the patients. Although four of the 35 patients experienced either no change or a slight decrease in their hemoglobin concentration when going from supine to upright and not all patients saw a uniform effect, providers should be aware that the patient’s position can contribute to changes in hemoglobin concentration in the hospitalized setting. Providers may be able to use this information to avoid an extensive diagnostic workup when anemia is identified in hospitalized patients, although more research is needed to identify patient subsets who are at higher risk for this effect.

Until hospitals implement protocols that require phlebotomists to report patient position during phlebotomy in a standardized fashion, providers should be alert to the fact that supine positioning may result in a hemoglobin level that is significantly lower than that when drawn in a sitting position, and in almost one-third of patients, this difference may be 1.0 g/dL or greater.

Given our study criteria requiring supine positions of at least 6 hours and a baseline hemoglobin concentration >8 g/dL, our sample of patients may have been younger and healthier than the average hospitalized patient on general internal medicine wards. Since greater relative changes in plasma volume shifts and hemoglobin might be seen in patients with lower baseline hemoglobin and lower baseline plasma protein, this selection bias may underestimate the effects of position on hemoglobin changes for the average inpatient population. Additionally, we intentionally sought to obtain sitting hemoglobin levels after the supine samples to avoid the possibility of incorrectly attributing dropping hemoglobin levels to progressive hospital-acquired anemia from phlebotomy or illness. Any concomitant trend of falling hemoglobin levels in our patients would be expected to lead to a systematic underestimation of the positional change in hemoglobin we observed. We did not objectively observe adherence to supine and upright position and instead relied on patient self-reporting, which is one possible contributor to the variable effects of position on hemoglobin concentration, with some patients having no change or decreases in hemoglobin concentrations.

Posture can significantly influence hemoglobin levels in hospitalized patients on general medicine wards. Further research can determine whether it would be cost and time effective to standardize patient positions prior to phlebotomy, or at least to report patient positioning with the laboratory testing results.

The World Health Organization (WHO) defines anemia as a hemoglobin value less than 12 g/dL in women and less than 13 g/dL in men.¹ Hospital-acquired anemia is loosely defined as normal hemoglobin levels on admission that, at their nadir during hospitalization or on discharge, are less than WHO sex-defined cutoffs. Hospital-acquired anemia or significant decreases in hemoglobin are often identified during hospitalization.^2-6 Potential causes include blood loss from phlebotomy, occult gastrointestinal bleeding, hemolysis, anemia of inflammation, and hemodilution due to fluid resuscitation. Of these causes, some are dangerous to patients, some are iatrogenic, and some are due to laboratory error.⁷ Physicians often evaluate decreases in hemoglobin, which could otherwise be explained by laboratory error, hemodilution, or expected decrease in hemoglobin due to hospitalization, to identify causes that may lead to potential harm.

Jacob et al⁸ demonstrated the effect of posture on hemoglobin concentrations in healthy volunteers, showing an average 11% relative increase in hemoglobin when going from lying to standing. This increase was attributed to shifts in plasma volume to the vascular space with recumbence. They hypothesized that the initial hemoglobin on admission is measured when patients are upright or recently upright, whereas after admission, patients are more likely to be supine, resulting in lower hemoglobin concentrations. Others have also demonstrated similar effects of patient posture on hemoglobin concentration.^9-13 However, these prior results are not readily generalizable to hospitalized patients. These prior studies enrolled healthy volunteers, and most examined postural changes from the supine and standing positions; blood is rarely obtained from hospitalized patients when they are standing.

The aim of this study was to investigate whether postural changes in hemoglobin can be demonstrated in positions that patients routinely encountered during in-hospital phlebotomy: upright in a chair or recumbent in a bed. Patient position, which is not standardized during blood draws, may contribute to lower measured hemoglobin concentrations in some patients, especially sicker individuals who are recumbent more frequently. We hypothesized that going from supine to upright in a chair would result in a relative increase in hemoglobin concentration of 5% to 6%, approximately half the value of going from supine to standing.⁸ To investigate this, we conducted a quasi-experimental study exploring the effect of position (supine or sitting in chair) on hemoglobin concentrations in medical inpatients.

Participants

Patients were enrolled in this single-center study between October 2017 and August 2018. Patients aged 18 years or older who were hospitalized on the general internal medicine wards were screened to determine if they met the following inclusion criteria: hospitalized for <5 days, had blood work scheduled as part of routine care (in order to decrease phlebotomy required by this study), had baseline hemoglobin >8 g/dL, and were able to remain supine without interruption overnight and able to sit in a chair for at least 1 hour the following morning. Patients were excluded from the study if they had a hematologic malignancy, were at risk of >100 mL of blood loss (eg, admitted for gastrointestinal bleeding, planned surgery), had a transfusion requirement, or received intravascular modifiers such as fluid (>100 cc/h) or intravenous diuretics. The Johns Hopkins Institutional Review Board approved this study, and all patients provided written informed consent.

Study Design

Patients enrolled in this quasi-experimental study were asked to remain supine for at least 6 hours overnight. Adherence to the recumbent position was tracked by patient self-report and by corroboration with the patient’s nurse overnight. Any interruptions to supine positioning resulted in exclusion from the study. The following morning, a member of the study team performed phlebotomy while the patient remained supine. Patients were then asked to sit comfortably in a chair for at least 1 hour with their feet on the ground; the blood draw was then repeated. All blood samples were acquired by venipuncture. Prior to each blood draw, a tourniquet was placed over the upper arm below the axilla. An antecubital vein on either arm was visualized under ultrasound guidance, and a 23-G × 3/4” butterfly needle was used for venipuncture. The vials of blood were immediately inverted after blood collection. Hemoglobin assays were processed and analyzed using Sysmex XN-10 analyzer (Sysmex Corporation). The reference range for hemoglobin in our facility was 12.0 to 15.0 g/dL for women and 13.9 to 16.3 g/dL for men. Laboratory technicians were blinded to and uninvolved in the study.

We determined, a priori, that 33 enrolled patients would provide 80% power (alpha 0.05) to detect an average hemoglobin change of 4.1%, assuming that the standard deviation of the hemoglobin change was twice the mean (ie, SD = 8.2%). The Wilcoxon signed-rank test was used to test the significance of postural hemoglobin changes. Analyses were conducted using JMP Pro 13.0 (SAS) and GraphPad Prism 8 (GraphPad Software). Significance was defined at P < .05 for all analyses.

Thirty-nine patients were consented and enrolled in the study; four patients were excluded prior to blood draw (two patients because of interruption of supine time, two patients because of refusal in the morning). Of the 35 patients who completed the study, 13 were women (37%); median age was 49 years (range, 25-83 years). Median supine hemoglobin concentration in our sample was 11.7 g/dL (range, 9.3-18.1 g/dL), and median baseline creatinine level was 0.70 mg/dL (range, 0.5-2.5 mg/dL). Median supine hemoglobin levels were 11.7 g/dL (range, 9.6-13.2 g/dL) in women and 11.8 g/dL (range, 9.3-18.1 g/dL) in men. In aggregate, patients had a median increase in hemoglobin concentration of 0.60 g/dL (range, –0.6 to 1.4 g/dL) with sitting, a 5.2% (range, –4.5% to 15.1%) relative change (P < .001) (Figure 1).

International blood collection guidelines acknowledge postural changes in laboratory values and recommend standardization of patient position to either sitting in a chair or lying flat in a bed, without changes in position for 15 minutes prior to blood draw.¹⁴ When these positional accommodations cannot be met, documenting positional disruptions is recommended so that laboratory values can be interpreted accordingly. To the best of our knowledge, no hospital in the United States has standardized patient position as part of phlebotomy procedure such that patient position is documented and can be made available to interpreting providers.

Relative increases in hemoglobin or hematocrit range from 7% to 12% when patients go from supine to standing.^8,9,11 The reverse relationship has also been shown, where upright-to-supine position results in decreases in hemoglobin concentrations.^10,13 We found that going from supine to a seated position resulted in significant increases in hemoglobin of 0.6 g/dL and in a more than 1 g/dL increase in 29% of the patients. Although four of the 35 patients experienced either no change or a slight decrease in their hemoglobin concentration when going from supine to upright and not all patients saw a uniform effect, providers should be aware that the patient’s position can contribute to changes in hemoglobin concentration in the hospitalized setting. Providers may be able to use this information to avoid an extensive diagnostic workup when anemia is identified in hospitalized patients, although more research is needed to identify patient subsets who are at higher risk for this effect.

Until hospitals implement protocols that require phlebotomists to report patient position during phlebotomy in a standardized fashion, providers should be alert to the fact that supine positioning may result in a hemoglobin level that is significantly lower than that when drawn in a sitting position, and in almost one-third of patients, this difference may be 1.0 g/dL or greater.

Given our study criteria requiring supine positions of at least 6 hours and a baseline hemoglobin concentration >8 g/dL, our sample of patients may have been younger and healthier than the average hospitalized patient on general internal medicine wards. Since greater relative changes in plasma volume shifts and hemoglobin might be seen in patients with lower baseline hemoglobin and lower baseline plasma protein, this selection bias may underestimate the effects of position on hemoglobin changes for the average inpatient population. Additionally, we intentionally sought to obtain sitting hemoglobin levels after the supine samples to avoid the possibility of incorrectly attributing dropping hemoglobin levels to progressive hospital-acquired anemia from phlebotomy or illness. Any concomitant trend of falling hemoglobin levels in our patients would be expected to lead to a systematic underestimation of the positional change in hemoglobin we observed. We did not objectively observe adherence to supine and upright position and instead relied on patient self-reporting, which is one possible contributor to the variable effects of position on hemoglobin concentration, with some patients having no change or decreases in hemoglobin concentrations.

Posture can significantly influence hemoglobin levels in hospitalized patients on general medicine wards. Further research can determine whether it would be cost and time effective to standardize patient positions prior to phlebotomy, or at least to report patient positioning with the laboratory testing results.

The World Health Organization (WHO) defines anemia as a hemoglobin value less than 12 g/dL in women and less than 13 g/dL in men.¹ Hospital-acquired anemia is loosely defined as normal hemoglobin levels on admission that, at their nadir during hospitalization or on discharge, are less than WHO sex-defined cutoffs. Hospital-acquired anemia or significant decreases in hemoglobin are often identified during hospitalization.^2-6 Potential causes include blood loss from phlebotomy, occult gastrointestinal bleeding, hemolysis, anemia of inflammation, and hemodilution due to fluid resuscitation. Of these causes, some are dangerous to patients, some are iatrogenic, and some are due to laboratory error.⁷ Physicians often evaluate decreases in hemoglobin, which could otherwise be explained by laboratory error, hemodilution, or expected decrease in hemoglobin due to hospitalization, to identify causes that may lead to potential harm.

Jacob et al⁸ demonstrated the effect of posture on hemoglobin concentrations in healthy volunteers, showing an average 11% relative increase in hemoglobin when going from lying to standing. This increase was attributed to shifts in plasma volume to the vascular space with recumbence. They hypothesized that the initial hemoglobin on admission is measured when patients are upright or recently upright, whereas after admission, patients are more likely to be supine, resulting in lower hemoglobin concentrations. Others have also demonstrated similar effects of patient posture on hemoglobin concentration.^9-13 However, these prior results are not readily generalizable to hospitalized patients. These prior studies enrolled healthy volunteers, and most examined postural changes from the supine and standing positions; blood is rarely obtained from hospitalized patients when they are standing.

The aim of this study was to investigate whether postural changes in hemoglobin can be demonstrated in positions that patients routinely encountered during in-hospital phlebotomy: upright in a chair or recumbent in a bed. Patient position, which is not standardized during blood draws, may contribute to lower measured hemoglobin concentrations in some patients, especially sicker individuals who are recumbent more frequently. We hypothesized that going from supine to upright in a chair would result in a relative increase in hemoglobin concentration of 5% to 6%, approximately half the value of going from supine to standing.⁸ To investigate this, we conducted a quasi-experimental study exploring the effect of position (supine or sitting in chair) on hemoglobin concentrations in medical inpatients.

Participants

Patients were enrolled in this single-center study between October 2017 and August 2018. Patients aged 18 years or older who were hospitalized on the general internal medicine wards were screened to determine if they met the following inclusion criteria: hospitalized for <5 days, had blood work scheduled as part of routine care (in order to decrease phlebotomy required by this study), had baseline hemoglobin >8 g/dL, and were able to remain supine without interruption overnight and able to sit in a chair for at least 1 hour the following morning. Patients were excluded from the study if they had a hematologic malignancy, were at risk of >100 mL of blood loss (eg, admitted for gastrointestinal bleeding, planned surgery), had a transfusion requirement, or received intravascular modifiers such as fluid (>100 cc/h) or intravenous diuretics. The Johns Hopkins Institutional Review Board approved this study, and all patients provided written informed consent.

Study Design

Patients enrolled in this quasi-experimental study were asked to remain supine for at least 6 hours overnight. Adherence to the recumbent position was tracked by patient self-report and by corroboration with the patient’s nurse overnight. Any interruptions to supine positioning resulted in exclusion from the study. The following morning, a member of the study team performed phlebotomy while the patient remained supine. Patients were then asked to sit comfortably in a chair for at least 1 hour with their feet on the ground; the blood draw was then repeated. All blood samples were acquired by venipuncture. Prior to each blood draw, a tourniquet was placed over the upper arm below the axilla. An antecubital vein on either arm was visualized under ultrasound guidance, and a 23-G × 3/4” butterfly needle was used for venipuncture. The vials of blood were immediately inverted after blood collection. Hemoglobin assays were processed and analyzed using Sysmex XN-10 analyzer (Sysmex Corporation). The reference range for hemoglobin in our facility was 12.0 to 15.0 g/dL for women and 13.9 to 16.3 g/dL for men. Laboratory technicians were blinded to and uninvolved in the study.

We determined, a priori, that 33 enrolled patients would provide 80% power (alpha 0.05) to detect an average hemoglobin change of 4.1%, assuming that the standard deviation of the hemoglobin change was twice the mean (ie, SD = 8.2%). The Wilcoxon signed-rank test was used to test the significance of postural hemoglobin changes. Analyses were conducted using JMP Pro 13.0 (SAS) and GraphPad Prism 8 (GraphPad Software). Significance was defined at P < .05 for all analyses.

Thirty-nine patients were consented and enrolled in the study; four patients were excluded prior to blood draw (two patients because of interruption of supine time, two patients because of refusal in the morning). Of the 35 patients who completed the study, 13 were women (37%); median age was 49 years (range, 25-83 years). Median supine hemoglobin concentration in our sample was 11.7 g/dL (range, 9.3-18.1 g/dL), and median baseline creatinine level was 0.70 mg/dL (range, 0.5-2.5 mg/dL). Median supine hemoglobin levels were 11.7 g/dL (range, 9.6-13.2 g/dL) in women and 11.8 g/dL (range, 9.3-18.1 g/dL) in men. In aggregate, patients had a median increase in hemoglobin concentration of 0.60 g/dL (range, –0.6 to 1.4 g/dL) with sitting, a 5.2% (range, –4.5% to 15.1%) relative change (P < .001) (Figure 1).

International blood collection guidelines acknowledge postural changes in laboratory values and recommend standardization of patient position to either sitting in a chair or lying flat in a bed, without changes in position for 15 minutes prior to blood draw.¹⁴ When these positional accommodations cannot be met, documenting positional disruptions is recommended so that laboratory values can be interpreted accordingly. To the best of our knowledge, no hospital in the United States has standardized patient position as part of phlebotomy procedure such that patient position is documented and can be made available to interpreting providers.

Relative increases in hemoglobin or hematocrit range from 7% to 12% when patients go from supine to standing.^8,9,11 The reverse relationship has also been shown, where upright-to-supine position results in decreases in hemoglobin concentrations.^10,13 We found that going from supine to a seated position resulted in significant increases in hemoglobin of 0.6 g/dL and in a more than 1 g/dL increase in 29% of the patients. Although four of the 35 patients experienced either no change or a slight decrease in their hemoglobin concentration when going from supine to upright and not all patients saw a uniform effect, providers should be aware that the patient’s position can contribute to changes in hemoglobin concentration in the hospitalized setting. Providers may be able to use this information to avoid an extensive diagnostic workup when anemia is identified in hospitalized patients, although more research is needed to identify patient subsets who are at higher risk for this effect.

Until hospitals implement protocols that require phlebotomists to report patient position during phlebotomy in a standardized fashion, providers should be alert to the fact that supine positioning may result in a hemoglobin level that is significantly lower than that when drawn in a sitting position, and in almost one-third of patients, this difference may be 1.0 g/dL or greater.

Given our study criteria requiring supine positions of at least 6 hours and a baseline hemoglobin concentration >8 g/dL, our sample of patients may have been younger and healthier than the average hospitalized patient on general internal medicine wards. Since greater relative changes in plasma volume shifts and hemoglobin might be seen in patients with lower baseline hemoglobin and lower baseline plasma protein, this selection bias may underestimate the effects of position on hemoglobin changes for the average inpatient population. Additionally, we intentionally sought to obtain sitting hemoglobin levels after the supine samples to avoid the possibility of incorrectly attributing dropping hemoglobin levels to progressive hospital-acquired anemia from phlebotomy or illness. Any concomitant trend of falling hemoglobin levels in our patients would be expected to lead to a systematic underestimation of the positional change in hemoglobin we observed. We did not objectively observe adherence to supine and upright position and instead relied on patient self-reporting, which is one possible contributor to the variable effects of position on hemoglobin concentration, with some patients having no change or decreases in hemoglobin concentrations.

Posture can significantly influence hemoglobin levels in hospitalized patients on general medicine wards. Further research can determine whether it would be cost and time effective to standardize patient positions prior to phlebotomy, or at least to report patient positioning with the laboratory testing results.

Sickle cell disease (SCD) affects an estimated 100,000 people in the United States.¹ Pain is the most common complication of SCD and the primary reason patients with SCD seek medical attention.² In 2016, three-fourths of the approximately 130,000 SCD-related hospitalizations in the United States involved pain crises.³ When managing patients with SCD and chronic pain, an individualized and interdisciplinary approach is crucial. In 2020, the American Society of Hematology (ASH) developed guidelines reflecting the latest evidence in managing acute and chronic pain in adult and pediatric patients with SCD. The ASH guidelines provide 18 recommendations; here, we highlight the 8 recommendations most pertinent to the hospitalist.

Acute Pain

Acute pain in the guideline is defined as pain that results in an unplanned visit to an acute care center for treatment.

Recommendation 1. For adult and pediatric patients presenting to an acute care setting with SCD-related acute pain, the ASH guideline panel recommends rapid (ie, within 1 hour of arrival at the emergency department [ED]) assessment and administration of analgesia, with reassessments every 30-60 minutes to optimize pain control (Strong recommendation; low certainty in the evidence about effects).

Although the perceived benefits are unclear due to insufficient evidence, the panel agrees that delaying pain management results in undeniable harm to patients. Hence, this recommendation was deemed both acceptable and ethical. Rapid evaluation also allows for earlier identification and treatment of other potential SCD-related complications.

Recommendation 2. For adult and pediatric patients presenting to an acute care setting with SCD-associated pain for whom opioid therapy is indicated, the ASH guideline panel suggests tailored opioid dosing based on consideration of baseline opioid therapy and prior effective therapy. (For adults: conditional recommendation; moderate certainty in the evidence about effects. For children: conditional recommendation; low certainty in the evidence about effects).

One randomized controlled trial examined patient-specific opioid dosing (based on current chronic opioid therapy [COT] and previously known effective acute pain management) vs weight-based dosing in the ED and found that participants randomized into the patient-specific protocol had a greater reduction in pain and decreased rate of hospital admission.⁴

The panel acknowledges that intravenous patient-controlled opioid analgesia is generally the standard of care at most institutions. However, no clear data address whether continuous opioid infusion in addition to on-demand dosing is beneficial.

Recommendation 3. For adult and pediatric patients with acute pain related to SCD, the ASH guideline panel suggests a short course (5 to 7 days) of nonsteroidal anti-inflammatory drugs (NSAIDs) in addition to opioids (Conditional recommendation; very low certainty in the evidence about effects).

The use of NSAIDs for managing pain in hospitalized patients with SCD has been associated with a reduction in the use of opioids in the inpatient setting and decreased lengths of stay.⁵ The potential harms of NSAIDs, including renal and gastrointestinal toxicity, however, should be factored into the decision-making as the risks may outweigh the potential benefits.

Recommendation 4. For adult and pediatric patients with SCD hospitalized for acute pain, the ASH guideline panel suggests a subanesthetic (analgesic) infusion of ketamine as adjunctive treatment of pain refractory or not effectively treated with opioids alone (Conditional recommendation; very low certainty in the evidence about effects). The guideline panel also suggests regional anesthesia for localized pain refractory or not effectively treated with opioids alone (Conditional recommendation; very low certainty in the evidence about effects).

Studies have demonstrated reduced pain and opioid utilization in individuals who received adjuvant ketamine infusions⁶ or regional anesthesia (ie, epidural).⁷ Feasibility, however, is limited to centers that have the appropriate experience and expertise with these interventions.

Recommendation 5. For adult and pediatric patients who have recurrent acute pain associated with SCD, the ASH guideline panel suggests against chronic monthly transfusion therapy as a first-line strategy to prevent or reduce recurrent acute pain episodes (Conditional recommendation; low certainty in the evidence about effects). The evidence for monthly transfusions in preventing recurrent pain is limited. There is, however, a moderate risk of harm, including iron overload and transfusion reactions, in addition to substantial burden and costs.

Chronic Pain

Chronic pain in the guideline is defined as ongoing pain present on most days over the past 6 months.

Recommendation 6. For adult patients with SCD who have chronic pain from the SCD-related identifiable cause avascular necrosis (AVN) of the bone, the ASH guideline panel suggests the use of serotonin-norepinephrine reuptake inhibitors (SNRIs) or NSAIDs in the context of a comprehensive disease and pain management plan (Conditional recommendation; very low certainty in the evidence about effects). For patients with no identifiable cause beyond SCD, the guideline panel suggests SNRIs, tricyclic antidepressants, or gabapentinoids for pain management (Conditional recommendation; very low certainty in the evidence about effects). Given the lack of direct evidence, indirect evidence was used to formulate these recommendations. For pain associated with AVN, data were extrapolated from literature on osteoarthritis, a form of degenerative arthropathy. For pain without an identifiable cause, evidence was taken from studies on fibromyalgia, a condition the panel felt most closely aligned with chronic pain related to SCD.

No recommendations were made for pediatric patients as the indirect evidence base only addressed adult patients.

Recommendation 7. For adult and pediatric patients with SCD and emerging and/or recently developed chronic pain, the ASH guideline panel does not recommend initiating COT unless pain is refractory to multiple other treatment modalities (Conditional recommendation; very low certainty in the evidence about effects). For patients receiving COT who are functioning well and have perceived benefit, the ASH guideline panel suggests shared decision-making for continuation of COT (Conditional recommendation; very low certainty in the evidence about effects).

High-quality data on the benefit of long-term COT in individuals with chronic noncancer pain are lacking. The panel maintains that the decision to initiate or continue COT should be individualized after weighing appropriate risks and benefits.

Recommendation 8. For adult and pediatric patients with chronic pain related to SCD, the panel suggests cognitive and behavioral pain management strategies in the context of a comprehensive disease and pain management plan (Conditional recommendation; very low certainty in the evidence about effects). Cognitive behavioral therapy may decrease overall pain intensity and improve coping skills.⁸ The panel agrees that medications alone may not be effective in reducing the burden of chronic pain in adult and pediatric patients with SCD.

The guidelines were created by a multidisciplinary panel that included physicians from hematology, pain medicine, psychiatry, and emergency medicine, a doctoral nurse practitioner, and two patient representatives. The Mayo Evidence-Based Practice Research Program supported the guideline-development process. The GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach was used to assess evidence and make recommendations.

High-quality data in treating acute and chronic pain in both adult and pediatric patients with SCD are limited. As such, the majority of recommendations in these guidelines are conditional. The panel included studies that were indirectly related to SCD based on consensus (eg, inferred data from disease processes thought to be similar to SCD). One panelist disclosed receiving direct payments from a company that could be affected by these guidelines; however, it was deemed that the conflict was unlikely to have influenced any recommendations.

The panel acknowledges that further investigation is needed for both nonpharmacologic and pharmacologic modalities in treating acute and chronic pain related to SCD. Examples include evaluating the comparative-effectiveness of COT vs nonopioid pharmacotherapy, the benefits and harms of continuous opioid infusions in acute pain crises, and the impact of chronic transfusions on acute and chronic pain.

Sickle cell disease (SCD) affects an estimated 100,000 people in the United States.¹ Pain is the most common complication of SCD and the primary reason patients with SCD seek medical attention.² In 2016, three-fourths of the approximately 130,000 SCD-related hospitalizations in the United States involved pain crises.³ When managing patients with SCD and chronic pain, an individualized and interdisciplinary approach is crucial. In 2020, the American Society of Hematology (ASH) developed guidelines reflecting the latest evidence in managing acute and chronic pain in adult and pediatric patients with SCD. The ASH guidelines provide 18 recommendations; here, we highlight the 8 recommendations most pertinent to the hospitalist.

Acute Pain

Acute pain in the guideline is defined as pain that results in an unplanned visit to an acute care center for treatment.

Recommendation 1. For adult and pediatric patients presenting to an acute care setting with SCD-related acute pain, the ASH guideline panel recommends rapid (ie, within 1 hour of arrival at the emergency department [ED]) assessment and administration of analgesia, with reassessments every 30-60 minutes to optimize pain control (Strong recommendation; low certainty in the evidence about effects).

Although the perceived benefits are unclear due to insufficient evidence, the panel agrees that delaying pain management results in undeniable harm to patients. Hence, this recommendation was deemed both acceptable and ethical. Rapid evaluation also allows for earlier identification and treatment of other potential SCD-related complications.

Recommendation 2. For adult and pediatric patients presenting to an acute care setting with SCD-associated pain for whom opioid therapy is indicated, the ASH guideline panel suggests tailored opioid dosing based on consideration of baseline opioid therapy and prior effective therapy. (For adults: conditional recommendation; moderate certainty in the evidence about effects. For children: conditional recommendation; low certainty in the evidence about effects).

One randomized controlled trial examined patient-specific opioid dosing (based on current chronic opioid therapy [COT] and previously known effective acute pain management) vs weight-based dosing in the ED and found that participants randomized into the patient-specific protocol had a greater reduction in pain and decreased rate of hospital admission.⁴

The panel acknowledges that intravenous patient-controlled opioid analgesia is generally the standard of care at most institutions. However, no clear data address whether continuous opioid infusion in addition to on-demand dosing is beneficial.

Recommendation 3. For adult and pediatric patients with acute pain related to SCD, the ASH guideline panel suggests a short course (5 to 7 days) of nonsteroidal anti-inflammatory drugs (NSAIDs) in addition to opioids (Conditional recommendation; very low certainty in the evidence about effects).

The use of NSAIDs for managing pain in hospitalized patients with SCD has been associated with a reduction in the use of opioids in the inpatient setting and decreased lengths of stay.⁵ The potential harms of NSAIDs, including renal and gastrointestinal toxicity, however, should be factored into the decision-making as the risks may outweigh the potential benefits.

Recommendation 4. For adult and pediatric patients with SCD hospitalized for acute pain, the ASH guideline panel suggests a subanesthetic (analgesic) infusion of ketamine as adjunctive treatment of pain refractory or not effectively treated with opioids alone (Conditional recommendation; very low certainty in the evidence about effects). The guideline panel also suggests regional anesthesia for localized pain refractory or not effectively treated with opioids alone (Conditional recommendation; very low certainty in the evidence about effects).

Studies have demonstrated reduced pain and opioid utilization in individuals who received adjuvant ketamine infusions⁶ or regional anesthesia (ie, epidural).⁷ Feasibility, however, is limited to centers that have the appropriate experience and expertise with these interventions.

Recommendation 5. For adult and pediatric patients who have recurrent acute pain associated with SCD, the ASH guideline panel suggests against chronic monthly transfusion therapy as a first-line strategy to prevent or reduce recurrent acute pain episodes (Conditional recommendation; low certainty in the evidence about effects). The evidence for monthly transfusions in preventing recurrent pain is limited. There is, however, a moderate risk of harm, including iron overload and transfusion reactions, in addition to substantial burden and costs.

Chronic Pain

Chronic pain in the guideline is defined as ongoing pain present on most days over the past 6 months.

Recommendation 6. For adult patients with SCD who have chronic pain from the SCD-related identifiable cause avascular necrosis (AVN) of the bone, the ASH guideline panel suggests the use of serotonin-norepinephrine reuptake inhibitors (SNRIs) or NSAIDs in the context of a comprehensive disease and pain management plan (Conditional recommendation; very low certainty in the evidence about effects). For patients with no identifiable cause beyond SCD, the guideline panel suggests SNRIs, tricyclic antidepressants, or gabapentinoids for pain management (Conditional recommendation; very low certainty in the evidence about effects). Given the lack of direct evidence, indirect evidence was used to formulate these recommendations. For pain associated with AVN, data were extrapolated from literature on osteoarthritis, a form of degenerative arthropathy. For pain without an identifiable cause, evidence was taken from studies on fibromyalgia, a condition the panel felt most closely aligned with chronic pain related to SCD.

No recommendations were made for pediatric patients as the indirect evidence base only addressed adult patients.

Recommendation 7. For adult and pediatric patients with SCD and emerging and/or recently developed chronic pain, the ASH guideline panel does not recommend initiating COT unless pain is refractory to multiple other treatment modalities (Conditional recommendation; very low certainty in the evidence about effects). For patients receiving COT who are functioning well and have perceived benefit, the ASH guideline panel suggests shared decision-making for continuation of COT (Conditional recommendation; very low certainty in the evidence about effects).

High-quality data on the benefit of long-term COT in individuals with chronic noncancer pain are lacking. The panel maintains that the decision to initiate or continue COT should be individualized after weighing appropriate risks and benefits.

Recommendation 8. For adult and pediatric patients with chronic pain related to SCD, the panel suggests cognitive and behavioral pain management strategies in the context of a comprehensive disease and pain management plan (Conditional recommendation; very low certainty in the evidence about effects). Cognitive behavioral therapy may decrease overall pain intensity and improve coping skills.⁸ The panel agrees that medications alone may not be effective in reducing the burden of chronic pain in adult and pediatric patients with SCD.

The guidelines were created by a multidisciplinary panel that included physicians from hematology, pain medicine, psychiatry, and emergency medicine, a doctoral nurse practitioner, and two patient representatives. The Mayo Evidence-Based Practice Research Program supported the guideline-development process. The GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach was used to assess evidence and make recommendations.

High-quality data in treating acute and chronic pain in both adult and pediatric patients with SCD are limited. As such, the majority of recommendations in these guidelines are conditional. The panel included studies that were indirectly related to SCD based on consensus (eg, inferred data from disease processes thought to be similar to SCD). One panelist disclosed receiving direct payments from a company that could be affected by these guidelines; however, it was deemed that the conflict was unlikely to have influenced any recommendations.

The panel acknowledges that further investigation is needed for both nonpharmacologic and pharmacologic modalities in treating acute and chronic pain related to SCD. Examples include evaluating the comparative-effectiveness of COT vs nonopioid pharmacotherapy, the benefits and harms of continuous opioid infusions in acute pain crises, and the impact of chronic transfusions on acute and chronic pain.

Sickle cell disease (SCD) affects an estimated 100,000 people in the United States.¹ Pain is the most common complication of SCD and the primary reason patients with SCD seek medical attention.² In 2016, three-fourths of the approximately 130,000 SCD-related hospitalizations in the United States involved pain crises.³ When managing patients with SCD and chronic pain, an individualized and interdisciplinary approach is crucial. In 2020, the American Society of Hematology (ASH) developed guidelines reflecting the latest evidence in managing acute and chronic pain in adult and pediatric patients with SCD. The ASH guidelines provide 18 recommendations; here, we highlight the 8 recommendations most pertinent to the hospitalist.

Acute Pain

Acute pain in the guideline is defined as pain that results in an unplanned visit to an acute care center for treatment.

Recommendation 1. For adult and pediatric patients presenting to an acute care setting with SCD-related acute pain, the ASH guideline panel recommends rapid (ie, within 1 hour of arrival at the emergency department [ED]) assessment and administration of analgesia, with reassessments every 30-60 minutes to optimize pain control (Strong recommendation; low certainty in the evidence about effects).

Although the perceived benefits are unclear due to insufficient evidence, the panel agrees that delaying pain management results in undeniable harm to patients. Hence, this recommendation was deemed both acceptable and ethical. Rapid evaluation also allows for earlier identification and treatment of other potential SCD-related complications.

Recommendation 2. For adult and pediatric patients presenting to an acute care setting with SCD-associated pain for whom opioid therapy is indicated, the ASH guideline panel suggests tailored opioid dosing based on consideration of baseline opioid therapy and prior effective therapy. (For adults: conditional recommendation; moderate certainty in the evidence about effects. For children: conditional recommendation; low certainty in the evidence about effects).

One randomized controlled trial examined patient-specific opioid dosing (based on current chronic opioid therapy [COT] and previously known effective acute pain management) vs weight-based dosing in the ED and found that participants randomized into the patient-specific protocol had a greater reduction in pain and decreased rate of hospital admission.⁴

The panel acknowledges that intravenous patient-controlled opioid analgesia is generally the standard of care at most institutions. However, no clear data address whether continuous opioid infusion in addition to on-demand dosing is beneficial.

Recommendation 3. For adult and pediatric patients with acute pain related to SCD, the ASH guideline panel suggests a short course (5 to 7 days) of nonsteroidal anti-inflammatory drugs (NSAIDs) in addition to opioids (Conditional recommendation; very low certainty in the evidence about effects).

The use of NSAIDs for managing pain in hospitalized patients with SCD has been associated with a reduction in the use of opioids in the inpatient setting and decreased lengths of stay.⁵ The potential harms of NSAIDs, including renal and gastrointestinal toxicity, however, should be factored into the decision-making as the risks may outweigh the potential benefits.

Recommendation 4. For adult and pediatric patients with SCD hospitalized for acute pain, the ASH guideline panel suggests a subanesthetic (analgesic) infusion of ketamine as adjunctive treatment of pain refractory or not effectively treated with opioids alone (Conditional recommendation; very low certainty in the evidence about effects). The guideline panel also suggests regional anesthesia for localized pain refractory or not effectively treated with opioids alone (Conditional recommendation; very low certainty in the evidence about effects).

Studies have demonstrated reduced pain and opioid utilization in individuals who received adjuvant ketamine infusions⁶ or regional anesthesia (ie, epidural).⁷ Feasibility, however, is limited to centers that have the appropriate experience and expertise with these interventions.

Recommendation 5. For adult and pediatric patients who have recurrent acute pain associated with SCD, the ASH guideline panel suggests against chronic monthly transfusion therapy as a first-line strategy to prevent or reduce recurrent acute pain episodes (Conditional recommendation; low certainty in the evidence about effects). The evidence for monthly transfusions in preventing recurrent pain is limited. There is, however, a moderate risk of harm, including iron overload and transfusion reactions, in addition to substantial burden and costs.

Chronic Pain

Chronic pain in the guideline is defined as ongoing pain present on most days over the past 6 months.

Recommendation 6. For adult patients with SCD who have chronic pain from the SCD-related identifiable cause avascular necrosis (AVN) of the bone, the ASH guideline panel suggests the use of serotonin-norepinephrine reuptake inhibitors (SNRIs) or NSAIDs in the context of a comprehensive disease and pain management plan (Conditional recommendation; very low certainty in the evidence about effects). For patients with no identifiable cause beyond SCD, the guideline panel suggests SNRIs, tricyclic antidepressants, or gabapentinoids for pain management (Conditional recommendation; very low certainty in the evidence about effects). Given the lack of direct evidence, indirect evidence was used to formulate these recommendations. For pain associated with AVN, data were extrapolated from literature on osteoarthritis, a form of degenerative arthropathy. For pain without an identifiable cause, evidence was taken from studies on fibromyalgia, a condition the panel felt most closely aligned with chronic pain related to SCD.

No recommendations were made for pediatric patients as the indirect evidence base only addressed adult patients.

Recommendation 7. For adult and pediatric patients with SCD and emerging and/or recently developed chronic pain, the ASH guideline panel does not recommend initiating COT unless pain is refractory to multiple other treatment modalities (Conditional recommendation; very low certainty in the evidence about effects). For patients receiving COT who are functioning well and have perceived benefit, the ASH guideline panel suggests shared decision-making for continuation of COT (Conditional recommendation; very low certainty in the evidence about effects).

High-quality data on the benefit of long-term COT in individuals with chronic noncancer pain are lacking. The panel maintains that the decision to initiate or continue COT should be individualized after weighing appropriate risks and benefits.

Recommendation 8. For adult and pediatric patients with chronic pain related to SCD, the panel suggests cognitive and behavioral pain management strategies in the context of a comprehensive disease and pain management plan (Conditional recommendation; very low certainty in the evidence about effects). Cognitive behavioral therapy may decrease overall pain intensity and improve coping skills.⁸ The panel agrees that medications alone may not be effective in reducing the burden of chronic pain in adult and pediatric patients with SCD.

The guidelines were created by a multidisciplinary panel that included physicians from hematology, pain medicine, psychiatry, and emergency medicine, a doctoral nurse practitioner, and two patient representatives. The Mayo Evidence-Based Practice Research Program supported the guideline-development process. The GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach was used to assess evidence and make recommendations.

High-quality data in treating acute and chronic pain in both adult and pediatric patients with SCD are limited. As such, the majority of recommendations in these guidelines are conditional. The panel included studies that were indirectly related to SCD based on consensus (eg, inferred data from disease processes thought to be similar to SCD). One panelist disclosed receiving direct payments from a company that could be affected by these guidelines; however, it was deemed that the conflict was unlikely to have influenced any recommendations.

The panel acknowledges that further investigation is needed for both nonpharmacologic and pharmacologic modalities in treating acute and chronic pain related to SCD. Examples include evaluating the comparative-effectiveness of COT vs nonopioid pharmacotherapy, the benefits and harms of continuous opioid infusions in acute pain crises, and the impact of chronic transfusions on acute and chronic pain.

Over the past 20 years, hospitalists have served as the primary workforce for the clinical care of medical inpatients in the United States.^1,2 Core competencies¹ state that hospitalists should be able to perform the following bedside procedures: lumbar puncture, paracentesis, thoracentesis, arthrocentesis, and central venous catheter placement. More recently, standard of care has dictated that these procedures be performed under ultrasound guidance,^3-6 and thus hospitalists are also expected to be adept at point-of-care ultrasound (POCUS).⁷

However, no current national standard exists for ensuring basic competency among hospitalists performing bedside procedures. In addition, hospitalists’ procedural volumes are declining,^8,9 and standards for procedural training during internal medicine residency have been reduced.¹⁰ As a result, many residents who intend to become hospitalists are no longer prepared to perform these procedures.

The ramifications of the loss of procedural competency for hospitalists are manifold. Technical errors are a significant source of patient morbidity and mortality,^11-15 and complications arising specifically from nonoperative procedures range from 0 to 19%,¹⁶ although these data do not distinguish technical errors from unpreventable adverse events nor the degree to which hospitalists contributed to these complications. Second, hospitalists in academic medical centers might be ill equipped to function as supervisors of trainees performing procedures, which could perpetuate a cycle of suboptimal technical skills.¹⁷ Finally, the discrepancy between consensus guidelines for hospitalists and their scope of practice represents a significant area of risk management for institutions that base their credentialing policies on published competencies.

There are many compelling reasons for why hospitalists should maintain—in fact reclaim—a primary role in bedside procedures.¹⁸ Hospitalists in community and rural settings might not have easy access to procedural specialists. In academic institutions, hospitalists are the primary instructors and supervisors of procedures performed by internal medicine residents. The increased availability of POCUS allows formally trained hospitalists to perform procedures more safely under imaging guidance.¹⁶

The literature on procedures performed by hospitalists, although limited, has focused on POCUS, systems innovations such as medical procedure services (MPS), and policy recommendations for procedural credentialing. Most studies on effective procedural instructional approaches have been conducted among trainees, who are procedural novices. This research does not sufficiently address the dilemma that hospitalists face as independent physicians for whom procedures are not a significant component of their practice, yet are expected to perform invasive procedures occasionally. The purpose of our literature review is to synthesize the available research to characterize contributors to hospitalists’ procedural competency. We conclude with considerations for hospital medicine practice.

We performed a structured literature search for peer-reviewed articles related to hospitalists conducting procedures, being trained in procedures, or related to hospitalist-run MPS. We focused our search on the core hospitalist procedures with the highest potential morbidity (ie, lumbar puncture, abdominal paracentesis, thoracentesis, and central venous catheterization). We searched PubMed and Google Scholar for articles published since 1996 (when the term “hospitalists” was first coined) using keyword searches for [hospitalist OR hospital medicine] AND [procedur* OR medical procedur* OR medical procedure service] OR [(procedur* AND (train* OR educat* OR teach OR instruct*)] OR abdominal paracentes* OR thoracentes* OR lumbar puncture OR central venous catheter* OR ultrasound OR point-of-care. We included original research, brief research reports, perspectives, guidelines, and consensus statements. Exclusion criteria were articles that focused on nonhospitalists and conference abstracts. We used pearling to identify secondary sources from included articles’ bibliographies, without limits on year of publication.

Trends Towards Specialist Referrals

Between 1986 and 2007, the number and variety of procedures performed by internists decreased by half.¹⁹ Hospitalists still completed procedures in greater volume and variety than nonhospitalists,⁸ with approximately 50% of hospitalists performing lumbar punctures (50%), abdominal paracenteses (49%), and thoracenteses (44%) compared with less than 25% for all three procedures for nonhospitalists. Additionally, only 11% of surveyed hospitalists⁸ performed all nine core procedures, although these included procedures that are largely cognitive in nature (eg, electrocardiogram interpretation, chest X-ray interpretation) or procedures that have been relegated to other specialists (eg, endotracheal intubation, ventilator management, or joint injection/aspiration).

Surveys showed that, especially in larger cities and academic centers, procedural specialists have taken over a disproportionate share of procedures even as the number of procedures performed continued to rise.²⁰ Between 1993 and 2008, the number of paracenteses and thoracenteses increased by 133% and decreased by 14%, respectively, but the share of procedures performed by radiologists increased by 964% and 358%, respectively, as evident in an analysis of Medicare billing data.²⁰ A more recent study of Medicare claims from 2004 to 2016 similarly revealed that the percentage of paracenteses performed by radiologists compared with nonradiologists rose from 70% to 80% and thoracenteses from 47% to 66%, respectively.²¹ Comparable trends were apparent in claims data for lumbar punctures; between 1991 and 2011, the share of lumbar punctures performed by radiologists rose from 11% to 48%.²²

In academic medical centers, hospitalists might have the opportunity to pursue other activities (eg, education, administration, research) as they progress in their careers, resulting in less clinical activity. Although hospitalists who are more clinically active in hospital care tended to perform more procedures,⁸ those with smaller clinical footprints reported lower levels of comfort with performing procedures⁸ and might have less available time to maintain procedural competency or train in new technologies such as POCUS.¹⁷

Additionally, hospitalists in both academic and community settings cited efficiency as a major reason for procedural referral. Hospitalists tended to perform more procedures if they had fixed salaries or if less than 50% of their income was based on clinical productivity, although this trend was not significant.⁸ Further, they also might be motivated by competing opportunity costs such as time lost caring for other patients or length of shift, which influences the amount of time spent at work.²³

Notably, speculation that hospitalists referred more complex cases to specialists was not borne out by studies examining referral patterns.^21,24,25

Procedural Outcomes for Hospitalists vs Nonhospitalists

No convincing data exist that procedures performed by specialists have better outcomes than those completed at the bedside by well-trained generalists, although studies were limited to the inpatient setting, to generalists who have some exposure to procedures, and to internal medicine residents on inpatient rotations. In one retrospective review, interventional radiology (IR) referrals were associated with more platelet or plasma transfusions and intensive care unit transfers than those performed at the bedside by internal medicine residents, findings that remained significant after accounting for complexity (eg, Model for End-stage Liver Disease score, need for dialysis, and platelet count).²⁴ Similarly, a prospective audit of 529 bedside procedures did not show any differences in complication rates between generalists and pulmonologists, once generalists underwent standardized training and used pleural safety checklists and ultrasound guidance.²⁶ An administrative database review of 130,000 inpatient thoracenteses across several university hospitals between 2010 and 2013 found that the risk of iatrogenic pneumothorax was similar among operators from IR, medicine, and pulmonary (2.8%, 2.9%, and 3.1%, respectively)²⁷; these findings have been reproduced in other studies.²⁸ Finally, the increasing adoption of procedural ultrasound permits procedures to be conducted more safely at the bedside, without the need to refer to radiology for imaging guidance.^3-5

IR procedures also are associated with increased healthcare costs compared with bedside procedures. One study showed that hospital costs for admissions when paracenteses were performed by radiologists were higher than those in which the procedure was completed at the bedside by gastroenterologists or hepatologists.²⁵ A chart review examining 399 paracenteses, thoracenteses, and lumbar punctures found that the average procedure cost increased by 38% for referred procedures and 56% for radiology-performed procedures, as compared with bedside procedures.²⁹ Needing ancillary staffing in dedicated suites to perform procedures contributed to the excess cost.⁹ Moreover, referred procedures resulted in increased length of stay, which can incur additional costs. However, the data were conflicting; two studies did not show a statistical difference,^25,28 while others found an increased length of stay,^24,27,29 which might be due to the unavailability of specialists during off hours, thereby delaying nonemergent procedures.²¹ Detailed cost analyses have controlled for the use of procedural facilities and blood transfusions among IR specialists and simulation training among generalists, showing that total costs were $663 per patient undergoing IR procedures compared with $134 per patient undergoing bedside procedures.³⁰

Lack of Standardized Procedural Training or Assessment

A robust body of primary studies and systematic reviews supports the use of simulation for procedural training to improve comfort and skill as well as reduce complication rates and costs.^31,32 A systematic review that investigated the impact of four paradigms of procedural training found that MPS and quality improvement/patient safety approaches led to the most active learning compared with apprenticeship (ie, “see one, do one”) and approaches based on educational theories.³³ Nevertheless, the vast majority of the research has been conducted in trainees,^32,34 with sparse evidence among practicing physicians. One cohort study of attending physicians’ central venous catheter insertion skills on simulators found low and variable short-term performance, showing overall poor adherence to checklists.³⁵ One article suggested that hospitalists’ procedural skills were below established thresholds of competency at baseline and that simulation-based training did not result in sustained skills, but the small sample size and high attrition limited meaningful conclusions.³⁶ Although continuing medical education courses are available to hospitalists, there is no published evidence supporting their effectiveness.

Proxies for procedural skill have included comfort and experience, yet these markers have broadly been shown to be inadequate.^34,36,37 Additionally, the natural decline of skill over time has invoked the need for periodic reassessment of proficiency.^36,38 Credentialing has been equally inconstant; a survey of the Society of Hospital Medicine’s (SHM) POCUS task force revealed that only half of respondents reported their hospitals required a minimum number of procedures for initial credentialing and recredentialing.³⁹ In short, periodic assessment of procedural skills among hospitalists has not been a routine process at many institutions.

Role of Hospitalist-Run Medical Procedure Services

It might not be necessary for all hospitalists to be proficient and credentialed in a given procedure,¹ and a trend has emerged in the creation of MPS staffed by hospitalists as proceduralists. The primary aim of these MPS has been to recapture the procedures—and associated revenue—that would otherwise be referred to specialists. Moreover, concentrating procedures among a core group of hospitalists endeavors to support patient safety through several principles: (1) to increase technical proficiency through higher procedural volumes, (2) to facilitate rigorous training and assessment among dedicated individuals, and (3) to systematize best practices of operator performance, communication, and documentation.

MPS have been implemented around the country and have demonstrated several advantages. In one institution, medical firms that were offered the use of an MPS had 48% more procedural attempts by nonspecialists, without significant differences in the proportions of successful attempts or complications compared with the firms who more often referred to specialists.⁴⁰ A retrospective study analyzed outcomes of 1,707 bedside procedures, of which 548 were performed by an MPS, and found that procedures done by the MPS were more likely to result in lower rates of unsuccessful procedures and to use best-practice safety processes (ie, to involve attending physicians, to use ultrasound guidance, and to avoid femoral sites for catheterization).¹² Satisfaction was high among patients who underwent bedside procedures performed by a hospitalist-supervised, intern-based procedure service with a focus on bedside communication.⁴¹ From a workforce perspective, MPS have also allowed surgical or radiological subspecialties to focus on more complex cases with higher reimbursement rates,^18,42 for proceduralists to expand beyond core procedures (eg, bone marrow biopsies⁴³), and to train advanced practice providers.⁴⁴ Although studies have not shown that the outcomes of procedures completed by an MPS are better than the outcomes of procedures performed by other specialists,⁴⁵ one can potentially extrapolate from earlier data that procedures done at the bedside by nonradiologists would have comparable outcomes.

A myriad of factors is influencing hospitalists’ scope of practice with respect to bedside procedures. Some evidence suggests that procedures performed by specialists are not superior to those done by generalists and might be associated with increased costs. The most promising developments in the past few decades include simulation-based training, which has demonstrated effectiveness across an array of clinical outcomes but has not been sufficiently evaluated in hospitalists to draw conclusions, and hospitalist-led MPS, which promote safe and productive procedural clinical practices. However, decreasing procedural volume, increasing referrals to specialists, dwindling hospitalist interest and/or confidence, time constraints, limited training opportunities, nonuniform credentialing policies, and lack of standardized assessment are cumulatively contributing to a loss of procedural competency among hospitalists.

Taken together, these forces should compel hospital medicine groups that expect their hospitalists to perform their own procedures to identify necessary steps for ensuring the safety of hospitalized patients under their care. The following considerations derive from the available—albeit modest—evidence on procedural performance in hospital medicine (Table).

1. Create MPS to establish a core set of hospitalists to perform procedures and train them using evidence-based practices. Creation of an MPS places the responsibility of core bedside procedures in the hands of a select group of proceduralists. This strategy streamlines training and assessment of individual procedural competency to meet standards set by SHM^36,46 and improves educational outcomes.^47-49 MPS could improve clinical outcomes,^12,42,50-52 including length of stay and cost, while maintaining patient satisfaction,⁴¹ as well as recoup lost revenue from referrals by increasing the volume of procedures done by generalists,^40,49 although no robust data supporting the latter point exists. Implementing an MPS requires full-time equivalent (FTE) support for proceduralists and administrative support for data collection and tracking complications. Furthermore, a well-functioning MPS will require investment in portable ultrasound machines and training in POCUS, which has been shown to decrease complications and increase success of invasive bedside procedures.^3-7 Hospital medicine groups should be aware that staffing an MPS can divert hospitalist labor and resources from other needed clinical areas, especially during the initial, low-volume phases of implementation. Strategies to offset relative value unit (RVU) loss include combining the MPS with existing clinical roles such as medical consults, code triage, and rapid response teams; or with services with lower patient caps, which might work particularly well in community hospitals. In many institutions, hospitalists can bill for procedural consults in addition to the procedures when the consult involves nonmedical patients, which is relevant when the procedure ultimately cannot be performed (eg, too little ascites to safely perform a paracentesis). Further research should establish best practices of MPS to ensure maximum procedural productivity and safety, because there are no rigorous prospective studies that evaluate strategies to create this service. Such strategies include determining the optimal ratio of proceduralists to general hospitalists, hospital characteristics that benefit most from MPS (eg, referral centers, urban-based settings), volume and type of procedures performed, and the proportion and type of referrals that are most cost-effective.

2. Establish policies with procedural specialists to arrange coverage for off-hours procedures and delineate thresholds for procedures that specialists should perform. Expanding hospitalists’ capabilities in performing procedures should trigger reconsideration of the medical center’s approach to procedural safety. A goal would be to have hospital medicine groups work collaboratively with specialists and other disciplines (eg, surgery, emergency medicine, anesthesia, or radiology) to ensure 24-hour, 7-day a week coverage of urgent bedside procedures. The potential to decrease length of stay and improve off-hour procedural quality might be a compelling rationale for hospital administration, whether or not an MPS is used. That said, we recognize that other services might be unable or unwilling to provide such coverage and that specialist off-hour coverage would incur increased costs and could reduce exposure opportunities for internal medicine residents.

A hospital-level procedures committee might be required to support an institutional imperative for procedural safety and to oversee the implementation of approaches that are practical, financially sustainable, and equitable for all service lines, especially because hospitalist groups might bear the early costs of training and retraining.

3. Hospitalist–proceduralists should collaborate with internal medicine residency programs to offer intensive procedural training experiences to residents who want these skills to be part of their future practice. Robust procedural training for trainees promotes better outcomes for the current workforce and helps to populate the future workforce with procedurally competent practitioners. Simulation-based training is a well-established procedural instruction method that is safe, authentic, and effective in terms of clinical outcomes.³⁴ As the primary teachers of residents in many institutions, hospitalists often are the ones who impart procedural skills to residents, despite uneven skill sets. It is in the interest of internal medicine residency program directors to advocate for a core group of hospitalist–proceduralists, as MPS offer an infrastructure for training that has been shown to increase procedural volume and improve skills.^47,48,50 Program directors could therefore be incentivized to sponsor some of these procedural roles with teaching and administration funds, as a trade-off for securing higher-quality procedural training and closer supervision for their trainees. The dual necessity of teaching procedural skills to residents and attending physicians alike offers economies of scale for the use of facilities, personnel, and equipment, and gives faculty an opportunity to extend their clinical teaching skills into the domain of procedural supervision.

4. Hospital medicine groups should re-evaluate credentialing and privileging to ensure procedural competency. Given the lack of published data that characterizes how many hospital medicine groups credential hospitalists to perform procedures and what practices they use to assess competency, hospital medicine groups might be signing off on procedures without verifying hospitalists’ proficiency in core procedures. SHM’s position statement on credentialing for ultrasound-guided procedures⁴⁶ describes standards that could be applied to other procedures. It proposes that credentialing processes should be grounded in simulation- and patient-based assessments of cognitive and psychomotor skills, using published checklists and global ratings for feedback. Simulation training could support provisional certification, but hospitalists should reach minimum thresholds of supervised patient-based experience before initial credentialing, with continuous reassessment of competency to mitigate skill decay. Prospectively tracking procedural metrics, such as procedural volume and complication rates, also will support systematic skill assessment. Finally, similar to any other medical error, near misses and complications should trigger periprocedural safety reviews.

Limitations

The modest body of research on hospitalists and procedures is the central limitation of our synthesis. Much of the literature consisted of consensus statements, retrospective studies, and small prospective educational studies. As a result, we did not adopt all strategies considered standard in a scoping or systematic review. The literature on MPS specifically was insufficient to draw conclusions about their operational and financial impact or effects on procedure quality. Our primary recommendation to implement MPS requires significant fiscal investment and infrastructure. It also entails risks that must be proactively addressed, including the potential for net financial loss and decreased educational opportunities for residents.

Hospitalists regularly face the predicament of being expected to independently perform procedures, with little access to training, minimal experience, and no ongoing assessment to ensure their proficiency or the safety of their patients. Past assumptions about hospitalists’ responsibility do not reflect realities in practice patterns and have not translated to widespread adoption of procedural training, monitoring, and assessment mechanisms. Our work summarizes a body of literature that, although limited in empiric studies of hospitalists themselves, offers insights with recommendations for hospital medicine groups wishing to uphold procedural skills as part of their providers’ professional identity.

Over the past 20 years, hospitalists have served as the primary workforce for the clinical care of medical inpatients in the United States.^1,2 Core competencies¹ state that hospitalists should be able to perform the following bedside procedures: lumbar puncture, paracentesis, thoracentesis, arthrocentesis, and central venous catheter placement. More recently, standard of care has dictated that these procedures be performed under ultrasound guidance,^3-6 and thus hospitalists are also expected to be adept at point-of-care ultrasound (POCUS).⁷

However, no current national standard exists for ensuring basic competency among hospitalists performing bedside procedures. In addition, hospitalists’ procedural volumes are declining,^8,9 and standards for procedural training during internal medicine residency have been reduced.¹⁰ As a result, many residents who intend to become hospitalists are no longer prepared to perform these procedures.

The ramifications of the loss of procedural competency for hospitalists are manifold. Technical errors are a significant source of patient morbidity and mortality,^11-15 and complications arising specifically from nonoperative procedures range from 0 to 19%,¹⁶ although these data do not distinguish technical errors from unpreventable adverse events nor the degree to which hospitalists contributed to these complications. Second, hospitalists in academic medical centers might be ill equipped to function as supervisors of trainees performing procedures, which could perpetuate a cycle of suboptimal technical skills.¹⁷ Finally, the discrepancy between consensus guidelines for hospitalists and their scope of practice represents a significant area of risk management for institutions that base their credentialing policies on published competencies.

There are many compelling reasons for why hospitalists should maintain—in fact reclaim—a primary role in bedside procedures.¹⁸ Hospitalists in community and rural settings might not have easy access to procedural specialists. In academic institutions, hospitalists are the primary instructors and supervisors of procedures performed by internal medicine residents. The increased availability of POCUS allows formally trained hospitalists to perform procedures more safely under imaging guidance.¹⁶

The literature on procedures performed by hospitalists, although limited, has focused on POCUS, systems innovations such as medical procedure services (MPS), and policy recommendations for procedural credentialing. Most studies on effective procedural instructional approaches have been conducted among trainees, who are procedural novices. This research does not sufficiently address the dilemma that hospitalists face as independent physicians for whom procedures are not a significant component of their practice, yet are expected to perform invasive procedures occasionally. The purpose of our literature review is to synthesize the available research to characterize contributors to hospitalists’ procedural competency. We conclude with considerations for hospital medicine practice.

We performed a structured literature search for peer-reviewed articles related to hospitalists conducting procedures, being trained in procedures, or related to hospitalist-run MPS. We focused our search on the core hospitalist procedures with the highest potential morbidity (ie, lumbar puncture, abdominal paracentesis, thoracentesis, and central venous catheterization). We searched PubMed and Google Scholar for articles published since 1996 (when the term “hospitalists” was first coined) using keyword searches for [hospitalist OR hospital medicine] AND [procedur* OR medical procedur* OR medical procedure service] OR [(procedur* AND (train* OR educat* OR teach OR instruct*)] OR abdominal paracentes* OR thoracentes* OR lumbar puncture OR central venous catheter* OR ultrasound OR point-of-care. We included original research, brief research reports, perspectives, guidelines, and consensus statements. Exclusion criteria were articles that focused on nonhospitalists and conference abstracts. We used pearling to identify secondary sources from included articles’ bibliographies, without limits on year of publication.

Trends Towards Specialist Referrals

Between 1986 and 2007, the number and variety of procedures performed by internists decreased by half.¹⁹ Hospitalists still completed procedures in greater volume and variety than nonhospitalists,⁸ with approximately 50% of hospitalists performing lumbar punctures (50%), abdominal paracenteses (49%), and thoracenteses (44%) compared with less than 25% for all three procedures for nonhospitalists. Additionally, only 11% of surveyed hospitalists⁸ performed all nine core procedures, although these included procedures that are largely cognitive in nature (eg, electrocardiogram interpretation, chest X-ray interpretation) or procedures that have been relegated to other specialists (eg, endotracheal intubation, ventilator management, or joint injection/aspiration).

Surveys showed that, especially in larger cities and academic centers, procedural specialists have taken over a disproportionate share of procedures even as the number of procedures performed continued to rise.²⁰ Between 1993 and 2008, the number of paracenteses and thoracenteses increased by 133% and decreased by 14%, respectively, but the share of procedures performed by radiologists increased by 964% and 358%, respectively, as evident in an analysis of Medicare billing data.²⁰ A more recent study of Medicare claims from 2004 to 2016 similarly revealed that the percentage of paracenteses performed by radiologists compared with nonradiologists rose from 70% to 80% and thoracenteses from 47% to 66%, respectively.²¹ Comparable trends were apparent in claims data for lumbar punctures; between 1991 and 2011, the share of lumbar punctures performed by radiologists rose from 11% to 48%.²²

In academic medical centers, hospitalists might have the opportunity to pursue other activities (eg, education, administration, research) as they progress in their careers, resulting in less clinical activity. Although hospitalists who are more clinically active in hospital care tended to perform more procedures,⁸ those with smaller clinical footprints reported lower levels of comfort with performing procedures⁸ and might have less available time to maintain procedural competency or train in new technologies such as POCUS.¹⁷

Additionally, hospitalists in both academic and community settings cited efficiency as a major reason for procedural referral. Hospitalists tended to perform more procedures if they had fixed salaries or if less than 50% of their income was based on clinical productivity, although this trend was not significant.⁸ Further, they also might be motivated by competing opportunity costs such as time lost caring for other patients or length of shift, which influences the amount of time spent at work.²³

Notably, speculation that hospitalists referred more complex cases to specialists was not borne out by studies examining referral patterns.^21,24,25

Procedural Outcomes for Hospitalists vs Nonhospitalists

No convincing data exist that procedures performed by specialists have better outcomes than those completed at the bedside by well-trained generalists, although studies were limited to the inpatient setting, to generalists who have some exposure to procedures, and to internal medicine residents on inpatient rotations. In one retrospective review, interventional radiology (IR) referrals were associated with more platelet or plasma transfusions and intensive care unit transfers than those performed at the bedside by internal medicine residents, findings that remained significant after accounting for complexity (eg, Model for End-stage Liver Disease score, need for dialysis, and platelet count).²⁴ Similarly, a prospective audit of 529 bedside procedures did not show any differences in complication rates between generalists and pulmonologists, once generalists underwent standardized training and used pleural safety checklists and ultrasound guidance.²⁶ An administrative database review of 130,000 inpatient thoracenteses across several university hospitals between 2010 and 2013 found that the risk of iatrogenic pneumothorax was similar among operators from IR, medicine, and pulmonary (2.8%, 2.9%, and 3.1%, respectively)²⁷; these findings have been reproduced in other studies.²⁸ Finally, the increasing adoption of procedural ultrasound permits procedures to be conducted more safely at the bedside, without the need to refer to radiology for imaging guidance.^3-5

IR procedures also are associated with increased healthcare costs compared with bedside procedures. One study showed that hospital costs for admissions when paracenteses were performed by radiologists were higher than those in which the procedure was completed at the bedside by gastroenterologists or hepatologists.²⁵ A chart review examining 399 paracenteses, thoracenteses, and lumbar punctures found that the average procedure cost increased by 38% for referred procedures and 56% for radiology-performed procedures, as compared with bedside procedures.²⁹ Needing ancillary staffing in dedicated suites to perform procedures contributed to the excess cost.⁹ Moreover, referred procedures resulted in increased length of stay, which can incur additional costs. However, the data were conflicting; two studies did not show a statistical difference,^25,28 while others found an increased length of stay,^24,27,29 which might be due to the unavailability of specialists during off hours, thereby delaying nonemergent procedures.²¹ Detailed cost analyses have controlled for the use of procedural facilities and blood transfusions among IR specialists and simulation training among generalists, showing that total costs were $663 per patient undergoing IR procedures compared with $134 per patient undergoing bedside procedures.³⁰

Lack of Standardized Procedural Training or Assessment

A robust body of primary studies and systematic reviews supports the use of simulation for procedural training to improve comfort and skill as well as reduce complication rates and costs.^31,32 A systematic review that investigated the impact of four paradigms of procedural training found that MPS and quality improvement/patient safety approaches led to the most active learning compared with apprenticeship (ie, “see one, do one”) and approaches based on educational theories.³³ Nevertheless, the vast majority of the research has been conducted in trainees,^32,34 with sparse evidence among practicing physicians. One cohort study of attending physicians’ central venous catheter insertion skills on simulators found low and variable short-term performance, showing overall poor adherence to checklists.³⁵ One article suggested that hospitalists’ procedural skills were below established thresholds of competency at baseline and that simulation-based training did not result in sustained skills, but the small sample size and high attrition limited meaningful conclusions.³⁶ Although continuing medical education courses are available to hospitalists, there is no published evidence supporting their effectiveness.

Proxies for procedural skill have included comfort and experience, yet these markers have broadly been shown to be inadequate.^34,36,37 Additionally, the natural decline of skill over time has invoked the need for periodic reassessment of proficiency.^36,38 Credentialing has been equally inconstant; a survey of the Society of Hospital Medicine’s (SHM) POCUS task force revealed that only half of respondents reported their hospitals required a minimum number of procedures for initial credentialing and recredentialing.³⁹ In short, periodic assessment of procedural skills among hospitalists has not been a routine process at many institutions.

Role of Hospitalist-Run Medical Procedure Services

It might not be necessary for all hospitalists to be proficient and credentialed in a given procedure,¹ and a trend has emerged in the creation of MPS staffed by hospitalists as proceduralists. The primary aim of these MPS has been to recapture the procedures—and associated revenue—that would otherwise be referred to specialists. Moreover, concentrating procedures among a core group of hospitalists endeavors to support patient safety through several principles: (1) to increase technical proficiency through higher procedural volumes, (2) to facilitate rigorous training and assessment among dedicated individuals, and (3) to systematize best practices of operator performance, communication, and documentation.

MPS have been implemented around the country and have demonstrated several advantages. In one institution, medical firms that were offered the use of an MPS had 48% more procedural attempts by nonspecialists, without significant differences in the proportions of successful attempts or complications compared with the firms who more often referred to specialists.⁴⁰ A retrospective study analyzed outcomes of 1,707 bedside procedures, of which 548 were performed by an MPS, and found that procedures done by the MPS were more likely to result in lower rates of unsuccessful procedures and to use best-practice safety processes (ie, to involve attending physicians, to use ultrasound guidance, and to avoid femoral sites for catheterization).¹² Satisfaction was high among patients who underwent bedside procedures performed by a hospitalist-supervised, intern-based procedure service with a focus on bedside communication.⁴¹ From a workforce perspective, MPS have also allowed surgical or radiological subspecialties to focus on more complex cases with higher reimbursement rates,^18,42 for proceduralists to expand beyond core procedures (eg, bone marrow biopsies⁴³), and to train advanced practice providers.⁴⁴ Although studies have not shown that the outcomes of procedures completed by an MPS are better than the outcomes of procedures performed by other specialists,⁴⁵ one can potentially extrapolate from earlier data that procedures done at the bedside by nonradiologists would have comparable outcomes.

A myriad of factors is influencing hospitalists’ scope of practice with respect to bedside procedures. Some evidence suggests that procedures performed by specialists are not superior to those done by generalists and might be associated with increased costs. The most promising developments in the past few decades include simulation-based training, which has demonstrated effectiveness across an array of clinical outcomes but has not been sufficiently evaluated in hospitalists to draw conclusions, and hospitalist-led MPS, which promote safe and productive procedural clinical practices. However, decreasing procedural volume, increasing referrals to specialists, dwindling hospitalist interest and/or confidence, time constraints, limited training opportunities, nonuniform credentialing policies, and lack of standardized assessment are cumulatively contributing to a loss of procedural competency among hospitalists.

Taken together, these forces should compel hospital medicine groups that expect their hospitalists to perform their own procedures to identify necessary steps for ensuring the safety of hospitalized patients under their care. The following considerations derive from the available—albeit modest—evidence on procedural performance in hospital medicine (Table).

1. Create MPS to establish a core set of hospitalists to perform procedures and train them using evidence-based practices. Creation of an MPS places the responsibility of core bedside procedures in the hands of a select group of proceduralists. This strategy streamlines training and assessment of individual procedural competency to meet standards set by SHM^36,46 and improves educational outcomes.^47-49 MPS could improve clinical outcomes,^12,42,50-52 including length of stay and cost, while maintaining patient satisfaction,⁴¹ as well as recoup lost revenue from referrals by increasing the volume of procedures done by generalists,^40,49 although no robust data supporting the latter point exists. Implementing an MPS requires full-time equivalent (FTE) support for proceduralists and administrative support for data collection and tracking complications. Furthermore, a well-functioning MPS will require investment in portable ultrasound machines and training in POCUS, which has been shown to decrease complications and increase success of invasive bedside procedures.^3-7 Hospital medicine groups should be aware that staffing an MPS can divert hospitalist labor and resources from other needed clinical areas, especially during the initial, low-volume phases of implementation. Strategies to offset relative value unit (RVU) loss include combining the MPS with existing clinical roles such as medical consults, code triage, and rapid response teams; or with services with lower patient caps, which might work particularly well in community hospitals. In many institutions, hospitalists can bill for procedural consults in addition to the procedures when the consult involves nonmedical patients, which is relevant when the procedure ultimately cannot be performed (eg, too little ascites to safely perform a paracentesis). Further research should establish best practices of MPS to ensure maximum procedural productivity and safety, because there are no rigorous prospective studies that evaluate strategies to create this service. Such strategies include determining the optimal ratio of proceduralists to general hospitalists, hospital characteristics that benefit most from MPS (eg, referral centers, urban-based settings), volume and type of procedures performed, and the proportion and type of referrals that are most cost-effective.

2. Establish policies with procedural specialists to arrange coverage for off-hours procedures and delineate thresholds for procedures that specialists should perform. Expanding hospitalists’ capabilities in performing procedures should trigger reconsideration of the medical center’s approach to procedural safety. A goal would be to have hospital medicine groups work collaboratively with specialists and other disciplines (eg, surgery, emergency medicine, anesthesia, or radiology) to ensure 24-hour, 7-day a week coverage of urgent bedside procedures. The potential to decrease length of stay and improve off-hour procedural quality might be a compelling rationale for hospital administration, whether or not an MPS is used. That said, we recognize that other services might be unable or unwilling to provide such coverage and that specialist off-hour coverage would incur increased costs and could reduce exposure opportunities for internal medicine residents.

A hospital-level procedures committee might be required to support an institutional imperative for procedural safety and to oversee the implementation of approaches that are practical, financially sustainable, and equitable for all service lines, especially because hospitalist groups might bear the early costs of training and retraining.

3. Hospitalist–proceduralists should collaborate with internal medicine residency programs to offer intensive procedural training experiences to residents who want these skills to be part of their future practice. Robust procedural training for trainees promotes better outcomes for the current workforce and helps to populate the future workforce with procedurally competent practitioners. Simulation-based training is a well-established procedural instruction method that is safe, authentic, and effective in terms of clinical outcomes.³⁴ As the primary teachers of residents in many institutions, hospitalists often are the ones who impart procedural skills to residents, despite uneven skill sets. It is in the interest of internal medicine residency program directors to advocate for a core group of hospitalist–proceduralists, as MPS offer an infrastructure for training that has been shown to increase procedural volume and improve skills.^47,48,50 Program directors could therefore be incentivized to sponsor some of these procedural roles with teaching and administration funds, as a trade-off for securing higher-quality procedural training and closer supervision for their trainees. The dual necessity of teaching procedural skills to residents and attending physicians alike offers economies of scale for the use of facilities, personnel, and equipment, and gives faculty an opportunity to extend their clinical teaching skills into the domain of procedural supervision.

4. Hospital medicine groups should re-evaluate credentialing and privileging to ensure procedural competency. Given the lack of published data that characterizes how many hospital medicine groups credential hospitalists to perform procedures and what practices they use to assess competency, hospital medicine groups might be signing off on procedures without verifying hospitalists’ proficiency in core procedures. SHM’s position statement on credentialing for ultrasound-guided procedures⁴⁶ describes standards that could be applied to other procedures. It proposes that credentialing processes should be grounded in simulation- and patient-based assessments of cognitive and psychomotor skills, using published checklists and global ratings for feedback. Simulation training could support provisional certification, but hospitalists should reach minimum thresholds of supervised patient-based experience before initial credentialing, with continuous reassessment of competency to mitigate skill decay. Prospectively tracking procedural metrics, such as procedural volume and complication rates, also will support systematic skill assessment. Finally, similar to any other medical error, near misses and complications should trigger periprocedural safety reviews.

Limitations

The modest body of research on hospitalists and procedures is the central limitation of our synthesis. Much of the literature consisted of consensus statements, retrospective studies, and small prospective educational studies. As a result, we did not adopt all strategies considered standard in a scoping or systematic review. The literature on MPS specifically was insufficient to draw conclusions about their operational and financial impact or effects on procedure quality. Our primary recommendation to implement MPS requires significant fiscal investment and infrastructure. It also entails risks that must be proactively addressed, including the potential for net financial loss and decreased educational opportunities for residents.

Hospitalists regularly face the predicament of being expected to independently perform procedures, with little access to training, minimal experience, and no ongoing assessment to ensure their proficiency or the safety of their patients. Past assumptions about hospitalists’ responsibility do not reflect realities in practice patterns and have not translated to widespread adoption of procedural training, monitoring, and assessment mechanisms. Our work summarizes a body of literature that, although limited in empiric studies of hospitalists themselves, offers insights with recommendations for hospital medicine groups wishing to uphold procedural skills as part of their providers’ professional identity.

Over the past 20 years, hospitalists have served as the primary workforce for the clinical care of medical inpatients in the United States.^1,2 Core competencies¹ state that hospitalists should be able to perform the following bedside procedures: lumbar puncture, paracentesis, thoracentesis, arthrocentesis, and central venous catheter placement. More recently, standard of care has dictated that these procedures be performed under ultrasound guidance,^3-6 and thus hospitalists are also expected to be adept at point-of-care ultrasound (POCUS).⁷

However, no current national standard exists for ensuring basic competency among hospitalists performing bedside procedures. In addition, hospitalists’ procedural volumes are declining,^8,9 and standards for procedural training during internal medicine residency have been reduced.¹⁰ As a result, many residents who intend to become hospitalists are no longer prepared to perform these procedures.

The ramifications of the loss of procedural competency for hospitalists are manifold. Technical errors are a significant source of patient morbidity and mortality,^11-15 and complications arising specifically from nonoperative procedures range from 0 to 19%,¹⁶ although these data do not distinguish technical errors from unpreventable adverse events nor the degree to which hospitalists contributed to these complications. Second, hospitalists in academic medical centers might be ill equipped to function as supervisors of trainees performing procedures, which could perpetuate a cycle of suboptimal technical skills.¹⁷ Finally, the discrepancy between consensus guidelines for hospitalists and their scope of practice represents a significant area of risk management for institutions that base their credentialing policies on published competencies.

There are many compelling reasons for why hospitalists should maintain—in fact reclaim—a primary role in bedside procedures.¹⁸ Hospitalists in community and rural settings might not have easy access to procedural specialists. In academic institutions, hospitalists are the primary instructors and supervisors of procedures performed by internal medicine residents. The increased availability of POCUS allows formally trained hospitalists to perform procedures more safely under imaging guidance.¹⁶

The literature on procedures performed by hospitalists, although limited, has focused on POCUS, systems innovations such as medical procedure services (MPS), and policy recommendations for procedural credentialing. Most studies on effective procedural instructional approaches have been conducted among trainees, who are procedural novices. This research does not sufficiently address the dilemma that hospitalists face as independent physicians for whom procedures are not a significant component of their practice, yet are expected to perform invasive procedures occasionally. The purpose of our literature review is to synthesize the available research to characterize contributors to hospitalists’ procedural competency. We conclude with considerations for hospital medicine practice.

We performed a structured literature search for peer-reviewed articles related to hospitalists conducting procedures, being trained in procedures, or related to hospitalist-run MPS. We focused our search on the core hospitalist procedures with the highest potential morbidity (ie, lumbar puncture, abdominal paracentesis, thoracentesis, and central venous catheterization). We searched PubMed and Google Scholar for articles published since 1996 (when the term “hospitalists” was first coined) using keyword searches for [hospitalist OR hospital medicine] AND [procedur* OR medical procedur* OR medical procedure service] OR [(procedur* AND (train* OR educat* OR teach OR instruct*)] OR abdominal paracentes* OR thoracentes* OR lumbar puncture OR central venous catheter* OR ultrasound OR point-of-care. We included original research, brief research reports, perspectives, guidelines, and consensus statements. Exclusion criteria were articles that focused on nonhospitalists and conference abstracts. We used pearling to identify secondary sources from included articles’ bibliographies, without limits on year of publication.

Trends Towards Specialist Referrals

Between 1986 and 2007, the number and variety of procedures performed by internists decreased by half.¹⁹ Hospitalists still completed procedures in greater volume and variety than nonhospitalists,⁸ with approximately 50% of hospitalists performing lumbar punctures (50%), abdominal paracenteses (49%), and thoracenteses (44%) compared with less than 25% for all three procedures for nonhospitalists. Additionally, only 11% of surveyed hospitalists⁸ performed all nine core procedures, although these included procedures that are largely cognitive in nature (eg, electrocardiogram interpretation, chest X-ray interpretation) or procedures that have been relegated to other specialists (eg, endotracheal intubation, ventilator management, or joint injection/aspiration).

Surveys showed that, especially in larger cities and academic centers, procedural specialists have taken over a disproportionate share of procedures even as the number of procedures performed continued to rise.²⁰ Between 1993 and 2008, the number of paracenteses and thoracenteses increased by 133% and decreased by 14%, respectively, but the share of procedures performed by radiologists increased by 964% and 358%, respectively, as evident in an analysis of Medicare billing data.²⁰ A more recent study of Medicare claims from 2004 to 2016 similarly revealed that the percentage of paracenteses performed by radiologists compared with nonradiologists rose from 70% to 80% and thoracenteses from 47% to 66%, respectively.²¹ Comparable trends were apparent in claims data for lumbar punctures; between 1991 and 2011, the share of lumbar punctures performed by radiologists rose from 11% to 48%.²²

In academic medical centers, hospitalists might have the opportunity to pursue other activities (eg, education, administration, research) as they progress in their careers, resulting in less clinical activity. Although hospitalists who are more clinically active in hospital care tended to perform more procedures,⁸ those with smaller clinical footprints reported lower levels of comfort with performing procedures⁸ and might have less available time to maintain procedural competency or train in new technologies such as POCUS.¹⁷

Additionally, hospitalists in both academic and community settings cited efficiency as a major reason for procedural referral. Hospitalists tended to perform more procedures if they had fixed salaries or if less than 50% of their income was based on clinical productivity, although this trend was not significant.⁸ Further, they also might be motivated by competing opportunity costs such as time lost caring for other patients or length of shift, which influences the amount of time spent at work.²³

Notably, speculation that hospitalists referred more complex cases to specialists was not borne out by studies examining referral patterns.^21,24,25

Procedural Outcomes for Hospitalists vs Nonhospitalists

No convincing data exist that procedures performed by specialists have better outcomes than those completed at the bedside by well-trained generalists, although studies were limited to the inpatient setting, to generalists who have some exposure to procedures, and to internal medicine residents on inpatient rotations. In one retrospective review, interventional radiology (IR) referrals were associated with more platelet or plasma transfusions and intensive care unit transfers than those performed at the bedside by internal medicine residents, findings that remained significant after accounting for complexity (eg, Model for End-stage Liver Disease score, need for dialysis, and platelet count).²⁴ Similarly, a prospective audit of 529 bedside procedures did not show any differences in complication rates between generalists and pulmonologists, once generalists underwent standardized training and used pleural safety checklists and ultrasound guidance.²⁶ An administrative database review of 130,000 inpatient thoracenteses across several university hospitals between 2010 and 2013 found that the risk of iatrogenic pneumothorax was similar among operators from IR, medicine, and pulmonary (2.8%, 2.9%, and 3.1%, respectively)²⁷; these findings have been reproduced in other studies.²⁸ Finally, the increasing adoption of procedural ultrasound permits procedures to be conducted more safely at the bedside, without the need to refer to radiology for imaging guidance.^3-5

IR procedures also are associated with increased healthcare costs compared with bedside procedures. One study showed that hospital costs for admissions when paracenteses were performed by radiologists were higher than those in which the procedure was completed at the bedside by gastroenterologists or hepatologists.²⁵ A chart review examining 399 paracenteses, thoracenteses, and lumbar punctures found that the average procedure cost increased by 38% for referred procedures and 56% for radiology-performed procedures, as compared with bedside procedures.²⁹ Needing ancillary staffing in dedicated suites to perform procedures contributed to the excess cost.⁹ Moreover, referred procedures resulted in increased length of stay, which can incur additional costs. However, the data were conflicting; two studies did not show a statistical difference,^25,28 while others found an increased length of stay,^24,27,29 which might be due to the unavailability of specialists during off hours, thereby delaying nonemergent procedures.²¹ Detailed cost analyses have controlled for the use of procedural facilities and blood transfusions among IR specialists and simulation training among generalists, showing that total costs were $663 per patient undergoing IR procedures compared with $134 per patient undergoing bedside procedures.³⁰

Lack of Standardized Procedural Training or Assessment

A robust body of primary studies and systematic reviews supports the use of simulation for procedural training to improve comfort and skill as well as reduce complication rates and costs.^31,32 A systematic review that investigated the impact of four paradigms of procedural training found that MPS and quality improvement/patient safety approaches led to the most active learning compared with apprenticeship (ie, “see one, do one”) and approaches based on educational theories.³³ Nevertheless, the vast majority of the research has been conducted in trainees,^32,34 with sparse evidence among practicing physicians. One cohort study of attending physicians’ central venous catheter insertion skills on simulators found low and variable short-term performance, showing overall poor adherence to checklists.³⁵ One article suggested that hospitalists’ procedural skills were below established thresholds of competency at baseline and that simulation-based training did not result in sustained skills, but the small sample size and high attrition limited meaningful conclusions.³⁶ Although continuing medical education courses are available to hospitalists, there is no published evidence supporting their effectiveness.

Proxies for procedural skill have included comfort and experience, yet these markers have broadly been shown to be inadequate.^34,36,37 Additionally, the natural decline of skill over time has invoked the need for periodic reassessment of proficiency.^36,38 Credentialing has been equally inconstant; a survey of the Society of Hospital Medicine’s (SHM) POCUS task force revealed that only half of respondents reported their hospitals required a minimum number of procedures for initial credentialing and recredentialing.³⁹ In short, periodic assessment of procedural skills among hospitalists has not been a routine process at many institutions.

Role of Hospitalist-Run Medical Procedure Services

It might not be necessary for all hospitalists to be proficient and credentialed in a given procedure,¹ and a trend has emerged in the creation of MPS staffed by hospitalists as proceduralists. The primary aim of these MPS has been to recapture the procedures—and associated revenue—that would otherwise be referred to specialists. Moreover, concentrating procedures among a core group of hospitalists endeavors to support patient safety through several principles: (1) to increase technical proficiency through higher procedural volumes, (2) to facilitate rigorous training and assessment among dedicated individuals, and (3) to systematize best practices of operator performance, communication, and documentation.

MPS have been implemented around the country and have demonstrated several advantages. In one institution, medical firms that were offered the use of an MPS had 48% more procedural attempts by nonspecialists, without significant differences in the proportions of successful attempts or complications compared with the firms who more often referred to specialists.⁴⁰ A retrospective study analyzed outcomes of 1,707 bedside procedures, of which 548 were performed by an MPS, and found that procedures done by the MPS were more likely to result in lower rates of unsuccessful procedures and to use best-practice safety processes (ie, to involve attending physicians, to use ultrasound guidance, and to avoid femoral sites for catheterization).¹² Satisfaction was high among patients who underwent bedside procedures performed by a hospitalist-supervised, intern-based procedure service with a focus on bedside communication.⁴¹ From a workforce perspective, MPS have also allowed surgical or radiological subspecialties to focus on more complex cases with higher reimbursement rates,^18,42 for proceduralists to expand beyond core procedures (eg, bone marrow biopsies⁴³), and to train advanced practice providers.⁴⁴ Although studies have not shown that the outcomes of procedures completed by an MPS are better than the outcomes of procedures performed by other specialists,⁴⁵ one can potentially extrapolate from earlier data that procedures done at the bedside by nonradiologists would have comparable outcomes.

A myriad of factors is influencing hospitalists’ scope of practice with respect to bedside procedures. Some evidence suggests that procedures performed by specialists are not superior to those done by generalists and might be associated with increased costs. The most promising developments in the past few decades include simulation-based training, which has demonstrated effectiveness across an array of clinical outcomes but has not been sufficiently evaluated in hospitalists to draw conclusions, and hospitalist-led MPS, which promote safe and productive procedural clinical practices. However, decreasing procedural volume, increasing referrals to specialists, dwindling hospitalist interest and/or confidence, time constraints, limited training opportunities, nonuniform credentialing policies, and lack of standardized assessment are cumulatively contributing to a loss of procedural competency among hospitalists.

Taken together, these forces should compel hospital medicine groups that expect their hospitalists to perform their own procedures to identify necessary steps for ensuring the safety of hospitalized patients under their care. The following considerations derive from the available—albeit modest—evidence on procedural performance in hospital medicine (Table).

1. Create MPS to establish a core set of hospitalists to perform procedures and train them using evidence-based practices. Creation of an MPS places the responsibility of core bedside procedures in the hands of a select group of proceduralists. This strategy streamlines training and assessment of individual procedural competency to meet standards set by SHM^36,46 and improves educational outcomes.^47-49 MPS could improve clinical outcomes,^12,42,50-52 including length of stay and cost, while maintaining patient satisfaction,⁴¹ as well as recoup lost revenue from referrals by increasing the volume of procedures done by generalists,^40,49 although no robust data supporting the latter point exists. Implementing an MPS requires full-time equivalent (FTE) support for proceduralists and administrative support for data collection and tracking complications. Furthermore, a well-functioning MPS will require investment in portable ultrasound machines and training in POCUS, which has been shown to decrease complications and increase success of invasive bedside procedures.^3-7 Hospital medicine groups should be aware that staffing an MPS can divert hospitalist labor and resources from other needed clinical areas, especially during the initial, low-volume phases of implementation. Strategies to offset relative value unit (RVU) loss include combining the MPS with existing clinical roles such as medical consults, code triage, and rapid response teams; or with services with lower patient caps, which might work particularly well in community hospitals. In many institutions, hospitalists can bill for procedural consults in addition to the procedures when the consult involves nonmedical patients, which is relevant when the procedure ultimately cannot be performed (eg, too little ascites to safely perform a paracentesis). Further research should establish best practices of MPS to ensure maximum procedural productivity and safety, because there are no rigorous prospective studies that evaluate strategies to create this service. Such strategies include determining the optimal ratio of proceduralists to general hospitalists, hospital characteristics that benefit most from MPS (eg, referral centers, urban-based settings), volume and type of procedures performed, and the proportion and type of referrals that are most cost-effective.

2. Establish policies with procedural specialists to arrange coverage for off-hours procedures and delineate thresholds for procedures that specialists should perform. Expanding hospitalists’ capabilities in performing procedures should trigger reconsideration of the medical center’s approach to procedural safety. A goal would be to have hospital medicine groups work collaboratively with specialists and other disciplines (eg, surgery, emergency medicine, anesthesia, or radiology) to ensure 24-hour, 7-day a week coverage of urgent bedside procedures. The potential to decrease length of stay and improve off-hour procedural quality might be a compelling rationale for hospital administration, whether or not an MPS is used. That said, we recognize that other services might be unable or unwilling to provide such coverage and that specialist off-hour coverage would incur increased costs and could reduce exposure opportunities for internal medicine residents.

A hospital-level procedures committee might be required to support an institutional imperative for procedural safety and to oversee the implementation of approaches that are practical, financially sustainable, and equitable for all service lines, especially because hospitalist groups might bear the early costs of training and retraining.

3. Hospitalist–proceduralists should collaborate with internal medicine residency programs to offer intensive procedural training experiences to residents who want these skills to be part of their future practice. Robust procedural training for trainees promotes better outcomes for the current workforce and helps to populate the future workforce with procedurally competent practitioners. Simulation-based training is a well-established procedural instruction method that is safe, authentic, and effective in terms of clinical outcomes.³⁴ As the primary teachers of residents in many institutions, hospitalists often are the ones who impart procedural skills to residents, despite uneven skill sets. It is in the interest of internal medicine residency program directors to advocate for a core group of hospitalist–proceduralists, as MPS offer an infrastructure for training that has been shown to increase procedural volume and improve skills.^47,48,50 Program directors could therefore be incentivized to sponsor some of these procedural roles with teaching and administration funds, as a trade-off for securing higher-quality procedural training and closer supervision for their trainees. The dual necessity of teaching procedural skills to residents and attending physicians alike offers economies of scale for the use of facilities, personnel, and equipment, and gives faculty an opportunity to extend their clinical teaching skills into the domain of procedural supervision.

4. Hospital medicine groups should re-evaluate credentialing and privileging to ensure procedural competency. Given the lack of published data that characterizes how many hospital medicine groups credential hospitalists to perform procedures and what practices they use to assess competency, hospital medicine groups might be signing off on procedures without verifying hospitalists’ proficiency in core procedures. SHM’s position statement on credentialing for ultrasound-guided procedures⁴⁶ describes standards that could be applied to other procedures. It proposes that credentialing processes should be grounded in simulation- and patient-based assessments of cognitive and psychomotor skills, using published checklists and global ratings for feedback. Simulation training could support provisional certification, but hospitalists should reach minimum thresholds of supervised patient-based experience before initial credentialing, with continuous reassessment of competency to mitigate skill decay. Prospectively tracking procedural metrics, such as procedural volume and complication rates, also will support systematic skill assessment. Finally, similar to any other medical error, near misses and complications should trigger periprocedural safety reviews.

Limitations

The modest body of research on hospitalists and procedures is the central limitation of our synthesis. Much of the literature consisted of consensus statements, retrospective studies, and small prospective educational studies. As a result, we did not adopt all strategies considered standard in a scoping or systematic review. The literature on MPS specifically was insufficient to draw conclusions about their operational and financial impact or effects on procedure quality. Our primary recommendation to implement MPS requires significant fiscal investment and infrastructure. It also entails risks that must be proactively addressed, including the potential for net financial loss and decreased educational opportunities for residents.

Hospitalists regularly face the predicament of being expected to independently perform procedures, with little access to training, minimal experience, and no ongoing assessment to ensure their proficiency or the safety of their patients. Past assumptions about hospitalists’ responsibility do not reflect realities in practice patterns and have not translated to widespread adoption of procedural training, monitoring, and assessment mechanisms. Our work summarizes a body of literature that, although limited in empiric studies of hospitalists themselves, offers insights with recommendations for hospital medicine groups wishing to uphold procedural skills as part of their providers’ professional identity.