Attending rounds at academic medical centers are often disconnected from patients and non-physician care team members. Time spent bedside is consistently less than one third of total rounding time, with observational studies reporting a range of 9% to 33% over the past several decades.^1-8 Rounds are often conducted outside patient rooms, denying patients, families, and nurses the opportunity to participate and offer valuable insights. Lack of bedside rounds thus limits patient and family engagement, patient input into the care plan, teaching of the physical examination, and communication and collaboration with nurses. In one study, physicians and nurses on rounds engaged in interprofessional communication in only 12% of patient cases.¹ Studies have found interdisciplinary bedside rounds have several benefits, including subjectively improved communication and teamwork between physicians and nurses; increased patient satisfaction, including feeling more cared for by the medical team; and decreased length of stay and costs of care.^2-10

However, there are many barriers to conducting interdisciplinary bedside rounds at large academic medical centers. Patients cared for by a single medical team are often geographically dispersed to several nursing units, and nurses are unable to predict when physicians will round on their patients. This situation limits nursing involvement on rounds and keeps doctors and nurses isolated from each other.² Regionalization of care teams reduces this fragmentation by facilitating more interaction among doctors, patients, families, and nursing staff.

There are few data on how regionalized patients and interdisciplinary bedside rounds affect rounding time and the nature of rounds. This information is needed to understand how these structural changes mediate their effects, whether other steps are required to optimize outcomes, and how to maximize efficiency. We used time-motion analysis (TMA) to investigate how regionalization of medical teams, encouragement of bedside rounding, and systematic inclusion of nurses on ward rounds affect amount of time spent with patients, nursing presence on rounds, and total rounding time.

Setting

This prospective interventional study, approved by the Institutional Review Board of Partners HealthCare, was conducted on the general medical wards at Brigham and Women’s Hospital, an academic 793-bed tertiary-care center in Boston, Massachusetts. Housestaff teams consist of 1 attending, 1 resident, and 2 interns with or without a medical student. Before June 20, 2013, daily rounds on medical inpatients were conducted largely on the patient unit but outside patient rooms. After completing most of a rounding discussion outside a patient’s room, the team might walk in to examine or speak with the patient. A typical medical team had patients dispersed over 7 medical units on average, and over as many as 13. As nurses were unit based, they did not consistently participate in rounds.

Intervention

In June 2013, as part of a general medical service care redesign initiative, the general medical teams were regionalized to specific inpatient units. The goal was to have teams admit patients predominantly to the team’s designated unit and to have all patients on a unit be cared for by the unit’s assigned team as often as possible, with an 85% goal for both. Toward those ends, the admitting structure was changed from a traditional 4-day call cycle to daily admitting for all teams, based on each unit’s bed availability.¹¹

Teams were also expected to conduct rounds with nurses, and a system for facilitating these rounds was established. As physician and nurse care teams were now geographically co-located, it became possible for residents and nurses to check a rounding sheet for the planned patient rounding order, which had been set by the resident and nurse-in-charge before rounds. No more than about 5 minutes was needed to prepare each day’s order. The rounding sheet prioritized sick patients, newly admitted patients, and planned morning discharges, but patients were also always grouped by nurse. For example, the physician team rounded with the first nurse on all 3 of a nurse’s patients, and then proceeded to the next group of 3 patients with the next nurse, until all patients were seen.

Teams were encouraged to conduct patient- and family-centered rounds exclusively at bedside, except when bedside rounding was thought to be detrimental to a patient (eg, one with delirium). After an intern’s bedside presentation, which included a brief summary and details about overnight events and vital signs, the concerns of the patient, family, and nurse were shared, a focused physical examination performed, relevant data (eg, laboratory test results and imaging studies) reviewed, and the day’s plan formulated. The entire team, including the attending, was expected to have read new patients’ admission notes before rounds. Bedside rounds could thus be focused more on patient assessment and patient/family engagement and less on data transfer.

Several actions were taken to facilitate these changes. Residents, attendings, nurses, and other interdisciplinary team members participated in a series of focus groups and conferences to define workflows and share best practices for patient- and family-centered bedside rounds. Tips on bedside rounding were included in a general medicine rotation guidebook made available to residents and attendings. At the beginning of each post-intervention general medicine rotation, attendings and residents attended brief orientation sessions to review the new daily schedule, have interdisciplinary huddles, and share expectations for patient- and family-centered bedside rounds. On the general medicine units, new medical directors were hired to partner with existing nursing directors to support adoption of the workflows. Last, an interdisciplinary leadership team was formed to support the care redesign efforts. This team started meeting every 2 weeks.

Study Design

We used a pre–post analysis to study the effects of care redesign. Analysis was performed at the same time of year for 2 consecutive years to control for the stage of training and experience of the housestaff. TMA was performed by trained medical students using computer tablets linked to a customized Microsoft Access database form (Redmond, Washington). The form and the database were designed with specific buttons that, when pressed, recorded the time of particular events, such as the coming and going of each participant, the location of rounds, and the beginning and the end of rounding encounters with a patient. One research assistant using an Access entry form was able to dynamically track all events in real time, as they occurred. We collected data on 4 teams at baseline and 5 teams after the intervention. Each of the 4 baseline teams was followed for 4 consecutive weekdays—16 rounds total, April-June 2013—to capture the 4-day call cycle. Each of the 5 post-intervention teams was followed for 5 consecutive weekdays—25 rounds total, April–June 2014—to capture the 5-day cycle. (Because of technical difficulties, data from 1 rounding session were not captured.) For inclusion in the statistical analyses, TMA captured 166 on-service patients before the intervention and 304 afterward. Off-service patients, those with an attending other than the team attending, were excluded because their rounds were conducted separately.

We examined 2 primary outcomes, the proportion of time each clinical team member was present on rounds and the proportion of bedside rounding time. Secondary outcomes were round duration, rounding time per patient, and total non-patient time per rounding session (total rounding time minus total patient time).

Statistical Analysis

TMA data were organized in an Access database and analyzed with SAS Version 9.3 (SAS Institute, Cary, North Carolina). We analyzed the data by round session as well as by patient.

Data are presented as means with standard deviations, medians with interquartile ranges, and proportions, as appropriate. For analyses by round session, we used unadjusted linear regression; for patient-level analyses, we used general estimating equations to adjust for clustering of patients within each session; for nurse presence during any part of a round by patient, we used a χ² test. Total non-patient time per round session was compared with use of patient-clustered general estimating equations using a γ distribution to account for the non-normality of the data.

Patient and Care Team Characteristics

Over the first year of the initiative, 85% of a team’s patients were on their assigned unit, and 87% of a unit’s patients were with the assigned team. Census numbers were 10.4 patients per general medicine team in April-June 2013 and 12.7 patients per team in April-June 2014, a 22% increase after care redesign. There were no statistically significant differences in patient characteristics, including age, sex, race, language, admission source, and comorbidity measure (Elixhauser score), between the pre-intervention and post-intervention study periods, except for a slightly higher proportion of patients admitted from home and fewer patients admitted directly from clinic (Table 1).

Primary Outcomes

Mean proportion of time the nurse was present on rounds per round session increased significantly (P < 0.001), from 24.1% to 67.8% (Figure 1A, Table 2). For individual patient encounters, the increased overall nursing presence was attributable to having more nurses on rounds and having nurses present for a larger proportion of individual rounding encounters (Figure 1B, Table 2). Nurses were present for at least some part of rounds for 53% of patients before the intervention and 93% afterward (P < 0.001). Mean proportion of round time by each of the 2 interns on each team decreased from 59.6% to 49.6% (P = 0.007).

Total bedside rounding time increased significantly ( P < 0.001), from 39.9% before the intervention to 55.8% afterward (Table 2). Meanwhile, percentage of rounding time spent on the unit but outside patient rooms decreased significantly ( P = 0.004), from 55.2% to 42.2%, as did rounding time on a unit completely different from the patient’s (4.9% before intervention, 2.0% afterward; P = 0.03). Again, patient-level results were similar (Figure 2, Table 2), but the decreased time spent on the unit, outside the patient rooms, was not significant.

Secondary Outcomes

Total rounding time decreased significantly, from a mean of 182 minutes (3.0 hours) at baseline to a mean of 146 minutes (2.4 hours) after the intervention, despite the higher post-intervention census. (When adjusted for patient census, the difference increased from 35.5 to 53.8 minutes; Table 2.) Mean rounding time per patient decreased significantly, from 14.7 minutes at baseline to 10.5 minutes after the intervention. For newly admitted patients, mean rounding time per patient decreased from 30.0 minutes before implementation to 16.3 minutes afterward. Mean rounding time also decreased, though much less, for subsequent-day patients (Table 2). For both new and existing patients, the decrease in rounding time largely was a reduction in time spent rounding outside patient rooms, with minimal impact on bedside time (Table 2). Mean time nurses were present during a patient’s rounds increased significantly, from 4.5 to 8.0 minutes (Table 2). Total nurse rounding time increased from 45.1 minutes per session to 98.8 minutes. Rounding time not related to patient discussion or evaluation decreased from 22.7 minutes per session to 13.3 minutes ( P = 0.003).

TMA of our care redesign initiative showed that this multipronged intervention, which included team regionalization, encouragement of bedside rounding with nurses, call structure changes, and attendings’ reading of admission notes before rounds, resulted in an increased proportion of rounding time spent with patients and an increased proportion of time nurses were present on rounds. Secondarily, round duration decreased even as patient census increased.

Regionalized teams have been found to improve interdisciplinary communication.¹ The present study elaborates on that finding by demonstrating a dramatic increase in nursing presence on rounds, likely resulting from the unit’s use of rounding schedules and nurses’ prioritization of rounding orders, both of which were made possible by geographic co-localization. Other research has noted that one of the most significant barriers to interdisciplinary rounds is difficulty coordinating the start times of physician/nurse bedside rounding encounters. The system we have studied directly addresses this difficulty.⁹ Of note, nursing presence on rounds is necessary but not sufficient for true physician–nurse collaboration and effective communication,¹ as reflected in a separate study of the intervention showing no significant difference in the concordance of the patient care plan between nurses and physicians before and after regionalization.¹² Additional interventions may be needed to ensure that communication during bedside rounds is effective.

Our regionalized teams spent a significantly higher proportion of rounding time bedside, likely because of a cultural shift in expectations and the increased convenience of seeing patients on the team’s unit. Nevertheless, bedside time was not 100%. Structural barriers (eg, patients off-unit for dialysis) and cultural barriers likely contributed to the less than full adoption of bedside rounding. As described previously, cultural barriers to bedside rounding include trainees’ anxiety about being questioned in front of patients, the desire to freely exchange academic ideas in a conference room, and attendings’ doubts about their bedside teaching ability.^1,9,13 Bedside rounds provide an important opportunity to apply the principles of patient- and family-centered care, including promotion of dignity and respect, information sharing, and collaboration. Thus, overcoming the concerns of housestaff and attendings and helping them feel prepared for bedside rounds can benefit the patient experience. More attention should be given to these practices as these types of interventions are implemented at Brigham and Women’s Hospital and elsewhere.^1,13-15

Another primary concern about interdisciplinary bedside rounding is the perception that it takes more time.⁹ Therefore, it was important for us to measure round duration as a balancing measure to be considered for our intervention. Fortunately, we found round duration decreased with regionalization and encouragement of bedside rounding. This decrease was driven largely by a significant decrease in mean rounding time per new patient, which may be attributable at least in part to setting expectations that attendings and residents will read admission notes before rounds and that interns will summarize rather than recount information from admission notes. However, we also found rounding time decreases for subsequent-day patients, suggesting an underlying time savings. Spending a larger proportion of time bedside may therefore result in more efficient rounds. Bedside presentations can reduce redundancies, such as discussing a patient’s case outside his or her room and subsequently walking in and going over much of the same information with the patient. Our model de-emphasizes data transfer in favor of discussion of care plans. There was also a decrease in non-patient time, likely reflecting reduced transit time for regionalized teams. This decrease aligns with a recent finding that bedside rounding was at least as efficient as rounding outside the room.¹⁶

Of note, though a larger percentage of time was spent bedside after implementation of the care redesign, the absolute amount of bedside time did not change significantly. Our data showed that, even with shorter rounds, the same amount of absolute time can be spent bedside, face to face with the patient, by increasing the proportion of bedside rounding time. In other words, teams on average did not spend more time with patients, though the content and the structure of those encounters may have changed. This finding may be attributable to eliminating redundancy, forgoing the outside-the-room discussion, and thus the largest time reductions were realized there. In addition, teams incompletely adopted beside rounds, as reflected in the data. We expect that, with more complete adoption, an even larger proportion of time will be spent bedside, and absolute time bedside might increase as a result.

An unexpected result of the care redesign was that interns’ proportion of rounding time decreased after the intervention. This decrease most likely is attributable to interns’ being less likely to participate in rounds for a co-intern’s patient, and to their staying outside that patient’s room to give themselves more time to advance the care of their own patients. Before the intervention, when more rounding time was spent outside patient rooms, interns were more likely to join rounds for their co-intern’s patients because they could easily break away, as needed, to continue care of their own patients. The resident is now encouraged to use the morning huddle to identify which patients likely have the most educational value, and both interns are expected to join the bedside rounds for these patients.

This study had a few limitations. First, the pre–post design made it difficult to exclude the possibility that other temporal changes may have affected outcomes, though we did account for time-of-year effects by aligning our data-collection phases. In addition, the authors, including the director of the general medical service, are unaware of any co-interventions during the study period. Second, the multipronged intervention included care team regionalization, encouragement of bedside rounding with nurses, call structure changes (from 4 days to daily admitting), and attendings’ reading of admission notes before rounds. Thus, parsing which component(s) contributed to the results was difficult, though all the changes instituted likely were necessary for system redesign. For example, regionalization of clinicians to unit-based teams was made possible by switching to a daily admitting system.

Time that team members spent preparing for rounds was not recorded before or after the intervention. Thus, the decrease in total rounding time could have been accompanied by an increase in time spent preparing for rounds. However, admission notes were available in our electronic medical record before and after the intervention, and most residents and attendings were already reading them pre-intervention. After the intervention, pre-round note reading was more clearly defined as an expectation, and we were able to set the expectation that interns should use their presentations to summarize rather than recount information. In addition, in the post-intervention period, we did not include time spent preparing rounding orders; as already noted, however, preparation took only 5 minutes per day. Also, we did not analyze the content or the quality of the discussion on rounds, but simply recorded who was present where and when. Regarding the effect of the intervention on patient care, results were mixed. As reported in 2016, we saw no difference in frequency of adverse events with this intervention.¹² However, a more sensitive measure of adverse events—used in a study on handoffs—showed our regionalization efforts had an additive effect on reducing overnight adverse events.¹⁷Researchers should now focus on the effects of care redesign on clinical outcomes, interdisciplinary care team communication, patient engagement and satisfaction, provider opinions of communication, workflow, patient care, and housestaff education. Our methodology can be used as a model to link structure, process, and outcome related to rounds and thereby better understand how best to optimize patient care and efficiency. Additional studies are needed to analyze the content of rounds and their association with patient and educational outcomes. Last, it will be important to conduct a study to see if the effects we have identified can be sustained. Such a study is already under way.

In conclusion, creating regionalized care teams and encouraging focused bedside rounds increased the proportion of bedside time and the presence of nurses on rounds. Rounds were shorter despite higher patient census. TMA revealed that regionalized care teams and bedside rounding at a large academic hospital are feasible, and are useful in establishing the necessary structures for increasing physician–nurse and provider–patient interactions.

Acknowledgments

The authors acknowledge Dr. Stan Ashley, Dr. Jacqueline Somerville, and Sheila Harris for their support of the regionalization initiative.

Disclosures

Dr. Schnipper received funding from Sanofi-aventis to conduct an investigator-initiated study to implement and evaluate a multi-faceted intervention to improve transitions of care in patients discharged home on insulin. The study was also supported by funding from the Marshall A. Wolf Medical Education Fund, Brigham and Women’s Hospital, and Dr. Stan Ashley, Chief Medical Officer, Brigham and Women’s Hospital. Some of the content of this article was orally presented at the annual meeting of the Society of Hospital Medicine; March 29-April 1, 2015; National Harbor, MD.

Attending rounds at academic medical centers are often disconnected from patients and non-physician care team members. Time spent bedside is consistently less than one third of total rounding time, with observational studies reporting a range of 9% to 33% over the past several decades.^1-8 Rounds are often conducted outside patient rooms, denying patients, families, and nurses the opportunity to participate and offer valuable insights. Lack of bedside rounds thus limits patient and family engagement, patient input into the care plan, teaching of the physical examination, and communication and collaboration with nurses. In one study, physicians and nurses on rounds engaged in interprofessional communication in only 12% of patient cases.¹ Studies have found interdisciplinary bedside rounds have several benefits, including subjectively improved communication and teamwork between physicians and nurses; increased patient satisfaction, including feeling more cared for by the medical team; and decreased length of stay and costs of care.^2-10

However, there are many barriers to conducting interdisciplinary bedside rounds at large academic medical centers. Patients cared for by a single medical team are often geographically dispersed to several nursing units, and nurses are unable to predict when physicians will round on their patients. This situation limits nursing involvement on rounds and keeps doctors and nurses isolated from each other.² Regionalization of care teams reduces this fragmentation by facilitating more interaction among doctors, patients, families, and nursing staff.

There are few data on how regionalized patients and interdisciplinary bedside rounds affect rounding time and the nature of rounds. This information is needed to understand how these structural changes mediate their effects, whether other steps are required to optimize outcomes, and how to maximize efficiency. We used time-motion analysis (TMA) to investigate how regionalization of medical teams, encouragement of bedside rounding, and systematic inclusion of nurses on ward rounds affect amount of time spent with patients, nursing presence on rounds, and total rounding time.

Setting

This prospective interventional study, approved by the Institutional Review Board of Partners HealthCare, was conducted on the general medical wards at Brigham and Women’s Hospital, an academic 793-bed tertiary-care center in Boston, Massachusetts. Housestaff teams consist of 1 attending, 1 resident, and 2 interns with or without a medical student. Before June 20, 2013, daily rounds on medical inpatients were conducted largely on the patient unit but outside patient rooms. After completing most of a rounding discussion outside a patient’s room, the team might walk in to examine or speak with the patient. A typical medical team had patients dispersed over 7 medical units on average, and over as many as 13. As nurses were unit based, they did not consistently participate in rounds.

Intervention

In June 2013, as part of a general medical service care redesign initiative, the general medical teams were regionalized to specific inpatient units. The goal was to have teams admit patients predominantly to the team’s designated unit and to have all patients on a unit be cared for by the unit’s assigned team as often as possible, with an 85% goal for both. Toward those ends, the admitting structure was changed from a traditional 4-day call cycle to daily admitting for all teams, based on each unit’s bed availability.¹¹

Teams were also expected to conduct rounds with nurses, and a system for facilitating these rounds was established. As physician and nurse care teams were now geographically co-located, it became possible for residents and nurses to check a rounding sheet for the planned patient rounding order, which had been set by the resident and nurse-in-charge before rounds. No more than about 5 minutes was needed to prepare each day’s order. The rounding sheet prioritized sick patients, newly admitted patients, and planned morning discharges, but patients were also always grouped by nurse. For example, the physician team rounded with the first nurse on all 3 of a nurse’s patients, and then proceeded to the next group of 3 patients with the next nurse, until all patients were seen.

Teams were encouraged to conduct patient- and family-centered rounds exclusively at bedside, except when bedside rounding was thought to be detrimental to a patient (eg, one with delirium). After an intern’s bedside presentation, which included a brief summary and details about overnight events and vital signs, the concerns of the patient, family, and nurse were shared, a focused physical examination performed, relevant data (eg, laboratory test results and imaging studies) reviewed, and the day’s plan formulated. The entire team, including the attending, was expected to have read new patients’ admission notes before rounds. Bedside rounds could thus be focused more on patient assessment and patient/family engagement and less on data transfer.

Several actions were taken to facilitate these changes. Residents, attendings, nurses, and other interdisciplinary team members participated in a series of focus groups and conferences to define workflows and share best practices for patient- and family-centered bedside rounds. Tips on bedside rounding were included in a general medicine rotation guidebook made available to residents and attendings. At the beginning of each post-intervention general medicine rotation, attendings and residents attended brief orientation sessions to review the new daily schedule, have interdisciplinary huddles, and share expectations for patient- and family-centered bedside rounds. On the general medicine units, new medical directors were hired to partner with existing nursing directors to support adoption of the workflows. Last, an interdisciplinary leadership team was formed to support the care redesign efforts. This team started meeting every 2 weeks.

Study Design

We used a pre–post analysis to study the effects of care redesign. Analysis was performed at the same time of year for 2 consecutive years to control for the stage of training and experience of the housestaff. TMA was performed by trained medical students using computer tablets linked to a customized Microsoft Access database form (Redmond, Washington). The form and the database were designed with specific buttons that, when pressed, recorded the time of particular events, such as the coming and going of each participant, the location of rounds, and the beginning and the end of rounding encounters with a patient. One research assistant using an Access entry form was able to dynamically track all events in real time, as they occurred. We collected data on 4 teams at baseline and 5 teams after the intervention. Each of the 4 baseline teams was followed for 4 consecutive weekdays—16 rounds total, April-June 2013—to capture the 4-day call cycle. Each of the 5 post-intervention teams was followed for 5 consecutive weekdays—25 rounds total, April–June 2014—to capture the 5-day cycle. (Because of technical difficulties, data from 1 rounding session were not captured.) For inclusion in the statistical analyses, TMA captured 166 on-service patients before the intervention and 304 afterward. Off-service patients, those with an attending other than the team attending, were excluded because their rounds were conducted separately.

We examined 2 primary outcomes, the proportion of time each clinical team member was present on rounds and the proportion of bedside rounding time. Secondary outcomes were round duration, rounding time per patient, and total non-patient time per rounding session (total rounding time minus total patient time).

Statistical Analysis

TMA data were organized in an Access database and analyzed with SAS Version 9.3 (SAS Institute, Cary, North Carolina). We analyzed the data by round session as well as by patient.

Data are presented as means with standard deviations, medians with interquartile ranges, and proportions, as appropriate. For analyses by round session, we used unadjusted linear regression; for patient-level analyses, we used general estimating equations to adjust for clustering of patients within each session; for nurse presence during any part of a round by patient, we used a χ² test. Total non-patient time per round session was compared with use of patient-clustered general estimating equations using a γ distribution to account for the non-normality of the data.

Patient and Care Team Characteristics

Over the first year of the initiative, 85% of a team’s patients were on their assigned unit, and 87% of a unit’s patients were with the assigned team. Census numbers were 10.4 patients per general medicine team in April-June 2013 and 12.7 patients per team in April-June 2014, a 22% increase after care redesign. There were no statistically significant differences in patient characteristics, including age, sex, race, language, admission source, and comorbidity measure (Elixhauser score), between the pre-intervention and post-intervention study periods, except for a slightly higher proportion of patients admitted from home and fewer patients admitted directly from clinic (Table 1).

Primary Outcomes

Mean proportion of time the nurse was present on rounds per round session increased significantly (P < 0.001), from 24.1% to 67.8% (Figure 1A, Table 2). For individual patient encounters, the increased overall nursing presence was attributable to having more nurses on rounds and having nurses present for a larger proportion of individual rounding encounters (Figure 1B, Table 2). Nurses were present for at least some part of rounds for 53% of patients before the intervention and 93% afterward (P < 0.001). Mean proportion of round time by each of the 2 interns on each team decreased from 59.6% to 49.6% (P = 0.007).

Total bedside rounding time increased significantly ( P < 0.001), from 39.9% before the intervention to 55.8% afterward (Table 2). Meanwhile, percentage of rounding time spent on the unit but outside patient rooms decreased significantly ( P = 0.004), from 55.2% to 42.2%, as did rounding time on a unit completely different from the patient’s (4.9% before intervention, 2.0% afterward; P = 0.03). Again, patient-level results were similar (Figure 2, Table 2), but the decreased time spent on the unit, outside the patient rooms, was not significant.

Secondary Outcomes

Total rounding time decreased significantly, from a mean of 182 minutes (3.0 hours) at baseline to a mean of 146 minutes (2.4 hours) after the intervention, despite the higher post-intervention census. (When adjusted for patient census, the difference increased from 35.5 to 53.8 minutes; Table 2.) Mean rounding time per patient decreased significantly, from 14.7 minutes at baseline to 10.5 minutes after the intervention. For newly admitted patients, mean rounding time per patient decreased from 30.0 minutes before implementation to 16.3 minutes afterward. Mean rounding time also decreased, though much less, for subsequent-day patients (Table 2). For both new and existing patients, the decrease in rounding time largely was a reduction in time spent rounding outside patient rooms, with minimal impact on bedside time (Table 2). Mean time nurses were present during a patient’s rounds increased significantly, from 4.5 to 8.0 minutes (Table 2). Total nurse rounding time increased from 45.1 minutes per session to 98.8 minutes. Rounding time not related to patient discussion or evaluation decreased from 22.7 minutes per session to 13.3 minutes ( P = 0.003).

TMA of our care redesign initiative showed that this multipronged intervention, which included team regionalization, encouragement of bedside rounding with nurses, call structure changes, and attendings’ reading of admission notes before rounds, resulted in an increased proportion of rounding time spent with patients and an increased proportion of time nurses were present on rounds. Secondarily, round duration decreased even as patient census increased.

Regionalized teams have been found to improve interdisciplinary communication.¹ The present study elaborates on that finding by demonstrating a dramatic increase in nursing presence on rounds, likely resulting from the unit’s use of rounding schedules and nurses’ prioritization of rounding orders, both of which were made possible by geographic co-localization. Other research has noted that one of the most significant barriers to interdisciplinary rounds is difficulty coordinating the start times of physician/nurse bedside rounding encounters. The system we have studied directly addresses this difficulty.⁹ Of note, nursing presence on rounds is necessary but not sufficient for true physician–nurse collaboration and effective communication,¹ as reflected in a separate study of the intervention showing no significant difference in the concordance of the patient care plan between nurses and physicians before and after regionalization.¹² Additional interventions may be needed to ensure that communication during bedside rounds is effective.

Our regionalized teams spent a significantly higher proportion of rounding time bedside, likely because of a cultural shift in expectations and the increased convenience of seeing patients on the team’s unit. Nevertheless, bedside time was not 100%. Structural barriers (eg, patients off-unit for dialysis) and cultural barriers likely contributed to the less than full adoption of bedside rounding. As described previously, cultural barriers to bedside rounding include trainees’ anxiety about being questioned in front of patients, the desire to freely exchange academic ideas in a conference room, and attendings’ doubts about their bedside teaching ability.^1,9,13 Bedside rounds provide an important opportunity to apply the principles of patient- and family-centered care, including promotion of dignity and respect, information sharing, and collaboration. Thus, overcoming the concerns of housestaff and attendings and helping them feel prepared for bedside rounds can benefit the patient experience. More attention should be given to these practices as these types of interventions are implemented at Brigham and Women’s Hospital and elsewhere.^1,13-15

Another primary concern about interdisciplinary bedside rounding is the perception that it takes more time.⁹ Therefore, it was important for us to measure round duration as a balancing measure to be considered for our intervention. Fortunately, we found round duration decreased with regionalization and encouragement of bedside rounding. This decrease was driven largely by a significant decrease in mean rounding time per new patient, which may be attributable at least in part to setting expectations that attendings and residents will read admission notes before rounds and that interns will summarize rather than recount information from admission notes. However, we also found rounding time decreases for subsequent-day patients, suggesting an underlying time savings. Spending a larger proportion of time bedside may therefore result in more efficient rounds. Bedside presentations can reduce redundancies, such as discussing a patient’s case outside his or her room and subsequently walking in and going over much of the same information with the patient. Our model de-emphasizes data transfer in favor of discussion of care plans. There was also a decrease in non-patient time, likely reflecting reduced transit time for regionalized teams. This decrease aligns with a recent finding that bedside rounding was at least as efficient as rounding outside the room.¹⁶

Of note, though a larger percentage of time was spent bedside after implementation of the care redesign, the absolute amount of bedside time did not change significantly. Our data showed that, even with shorter rounds, the same amount of absolute time can be spent bedside, face to face with the patient, by increasing the proportion of bedside rounding time. In other words, teams on average did not spend more time with patients, though the content and the structure of those encounters may have changed. This finding may be attributable to eliminating redundancy, forgoing the outside-the-room discussion, and thus the largest time reductions were realized there. In addition, teams incompletely adopted beside rounds, as reflected in the data. We expect that, with more complete adoption, an even larger proportion of time will be spent bedside, and absolute time bedside might increase as a result.

An unexpected result of the care redesign was that interns’ proportion of rounding time decreased after the intervention. This decrease most likely is attributable to interns’ being less likely to participate in rounds for a co-intern’s patient, and to their staying outside that patient’s room to give themselves more time to advance the care of their own patients. Before the intervention, when more rounding time was spent outside patient rooms, interns were more likely to join rounds for their co-intern’s patients because they could easily break away, as needed, to continue care of their own patients. The resident is now encouraged to use the morning huddle to identify which patients likely have the most educational value, and both interns are expected to join the bedside rounds for these patients.

This study had a few limitations. First, the pre–post design made it difficult to exclude the possibility that other temporal changes may have affected outcomes, though we did account for time-of-year effects by aligning our data-collection phases. In addition, the authors, including the director of the general medical service, are unaware of any co-interventions during the study period. Second, the multipronged intervention included care team regionalization, encouragement of bedside rounding with nurses, call structure changes (from 4 days to daily admitting), and attendings’ reading of admission notes before rounds. Thus, parsing which component(s) contributed to the results was difficult, though all the changes instituted likely were necessary for system redesign. For example, regionalization of clinicians to unit-based teams was made possible by switching to a daily admitting system.

Time that team members spent preparing for rounds was not recorded before or after the intervention. Thus, the decrease in total rounding time could have been accompanied by an increase in time spent preparing for rounds. However, admission notes were available in our electronic medical record before and after the intervention, and most residents and attendings were already reading them pre-intervention. After the intervention, pre-round note reading was more clearly defined as an expectation, and we were able to set the expectation that interns should use their presentations to summarize rather than recount information. In addition, in the post-intervention period, we did not include time spent preparing rounding orders; as already noted, however, preparation took only 5 minutes per day. Also, we did not analyze the content or the quality of the discussion on rounds, but simply recorded who was present where and when. Regarding the effect of the intervention on patient care, results were mixed. As reported in 2016, we saw no difference in frequency of adverse events with this intervention.¹² However, a more sensitive measure of adverse events—used in a study on handoffs—showed our regionalization efforts had an additive effect on reducing overnight adverse events.¹⁷Researchers should now focus on the effects of care redesign on clinical outcomes, interdisciplinary care team communication, patient engagement and satisfaction, provider opinions of communication, workflow, patient care, and housestaff education. Our methodology can be used as a model to link structure, process, and outcome related to rounds and thereby better understand how best to optimize patient care and efficiency. Additional studies are needed to analyze the content of rounds and their association with patient and educational outcomes. Last, it will be important to conduct a study to see if the effects we have identified can be sustained. Such a study is already under way.

In conclusion, creating regionalized care teams and encouraging focused bedside rounds increased the proportion of bedside time and the presence of nurses on rounds. Rounds were shorter despite higher patient census. TMA revealed that regionalized care teams and bedside rounding at a large academic hospital are feasible, and are useful in establishing the necessary structures for increasing physician–nurse and provider–patient interactions.

Acknowledgments

The authors acknowledge Dr. Stan Ashley, Dr. Jacqueline Somerville, and Sheila Harris for their support of the regionalization initiative.

Disclosures

Dr. Schnipper received funding from Sanofi-aventis to conduct an investigator-initiated study to implement and evaluate a multi-faceted intervention to improve transitions of care in patients discharged home on insulin. The study was also supported by funding from the Marshall A. Wolf Medical Education Fund, Brigham and Women’s Hospital, and Dr. Stan Ashley, Chief Medical Officer, Brigham and Women’s Hospital. Some of the content of this article was orally presented at the annual meeting of the Society of Hospital Medicine; March 29-April 1, 2015; National Harbor, MD.

Attending rounds at academic medical centers are often disconnected from patients and non-physician care team members. Time spent bedside is consistently less than one third of total rounding time, with observational studies reporting a range of 9% to 33% over the past several decades.^1-8 Rounds are often conducted outside patient rooms, denying patients, families, and nurses the opportunity to participate and offer valuable insights. Lack of bedside rounds thus limits patient and family engagement, patient input into the care plan, teaching of the physical examination, and communication and collaboration with nurses. In one study, physicians and nurses on rounds engaged in interprofessional communication in only 12% of patient cases.¹ Studies have found interdisciplinary bedside rounds have several benefits, including subjectively improved communication and teamwork between physicians and nurses; increased patient satisfaction, including feeling more cared for by the medical team; and decreased length of stay and costs of care.^2-10

However, there are many barriers to conducting interdisciplinary bedside rounds at large academic medical centers. Patients cared for by a single medical team are often geographically dispersed to several nursing units, and nurses are unable to predict when physicians will round on their patients. This situation limits nursing involvement on rounds and keeps doctors and nurses isolated from each other.² Regionalization of care teams reduces this fragmentation by facilitating more interaction among doctors, patients, families, and nursing staff.

There are few data on how regionalized patients and interdisciplinary bedside rounds affect rounding time and the nature of rounds. This information is needed to understand how these structural changes mediate their effects, whether other steps are required to optimize outcomes, and how to maximize efficiency. We used time-motion analysis (TMA) to investigate how regionalization of medical teams, encouragement of bedside rounding, and systematic inclusion of nurses on ward rounds affect amount of time spent with patients, nursing presence on rounds, and total rounding time.

Setting

This prospective interventional study, approved by the Institutional Review Board of Partners HealthCare, was conducted on the general medical wards at Brigham and Women’s Hospital, an academic 793-bed tertiary-care center in Boston, Massachusetts. Housestaff teams consist of 1 attending, 1 resident, and 2 interns with or without a medical student. Before June 20, 2013, daily rounds on medical inpatients were conducted largely on the patient unit but outside patient rooms. After completing most of a rounding discussion outside a patient’s room, the team might walk in to examine or speak with the patient. A typical medical team had patients dispersed over 7 medical units on average, and over as many as 13. As nurses were unit based, they did not consistently participate in rounds.

Intervention

In June 2013, as part of a general medical service care redesign initiative, the general medical teams were regionalized to specific inpatient units. The goal was to have teams admit patients predominantly to the team’s designated unit and to have all patients on a unit be cared for by the unit’s assigned team as often as possible, with an 85% goal for both. Toward those ends, the admitting structure was changed from a traditional 4-day call cycle to daily admitting for all teams, based on each unit’s bed availability.¹¹

Teams were also expected to conduct rounds with nurses, and a system for facilitating these rounds was established. As physician and nurse care teams were now geographically co-located, it became possible for residents and nurses to check a rounding sheet for the planned patient rounding order, which had been set by the resident and nurse-in-charge before rounds. No more than about 5 minutes was needed to prepare each day’s order. The rounding sheet prioritized sick patients, newly admitted patients, and planned morning discharges, but patients were also always grouped by nurse. For example, the physician team rounded with the first nurse on all 3 of a nurse’s patients, and then proceeded to the next group of 3 patients with the next nurse, until all patients were seen.

Teams were encouraged to conduct patient- and family-centered rounds exclusively at bedside, except when bedside rounding was thought to be detrimental to a patient (eg, one with delirium). After an intern’s bedside presentation, which included a brief summary and details about overnight events and vital signs, the concerns of the patient, family, and nurse were shared, a focused physical examination performed, relevant data (eg, laboratory test results and imaging studies) reviewed, and the day’s plan formulated. The entire team, including the attending, was expected to have read new patients’ admission notes before rounds. Bedside rounds could thus be focused more on patient assessment and patient/family engagement and less on data transfer.

Several actions were taken to facilitate these changes. Residents, attendings, nurses, and other interdisciplinary team members participated in a series of focus groups and conferences to define workflows and share best practices for patient- and family-centered bedside rounds. Tips on bedside rounding were included in a general medicine rotation guidebook made available to residents and attendings. At the beginning of each post-intervention general medicine rotation, attendings and residents attended brief orientation sessions to review the new daily schedule, have interdisciplinary huddles, and share expectations for patient- and family-centered bedside rounds. On the general medicine units, new medical directors were hired to partner with existing nursing directors to support adoption of the workflows. Last, an interdisciplinary leadership team was formed to support the care redesign efforts. This team started meeting every 2 weeks.

Study Design

We used a pre–post analysis to study the effects of care redesign. Analysis was performed at the same time of year for 2 consecutive years to control for the stage of training and experience of the housestaff. TMA was performed by trained medical students using computer tablets linked to a customized Microsoft Access database form (Redmond, Washington). The form and the database were designed with specific buttons that, when pressed, recorded the time of particular events, such as the coming and going of each participant, the location of rounds, and the beginning and the end of rounding encounters with a patient. One research assistant using an Access entry form was able to dynamically track all events in real time, as they occurred. We collected data on 4 teams at baseline and 5 teams after the intervention. Each of the 4 baseline teams was followed for 4 consecutive weekdays—16 rounds total, April-June 2013—to capture the 4-day call cycle. Each of the 5 post-intervention teams was followed for 5 consecutive weekdays—25 rounds total, April–June 2014—to capture the 5-day cycle. (Because of technical difficulties, data from 1 rounding session were not captured.) For inclusion in the statistical analyses, TMA captured 166 on-service patients before the intervention and 304 afterward. Off-service patients, those with an attending other than the team attending, were excluded because their rounds were conducted separately.

We examined 2 primary outcomes, the proportion of time each clinical team member was present on rounds and the proportion of bedside rounding time. Secondary outcomes were round duration, rounding time per patient, and total non-patient time per rounding session (total rounding time minus total patient time).

Statistical Analysis

TMA data were organized in an Access database and analyzed with SAS Version 9.3 (SAS Institute, Cary, North Carolina). We analyzed the data by round session as well as by patient.

Data are presented as means with standard deviations, medians with interquartile ranges, and proportions, as appropriate. For analyses by round session, we used unadjusted linear regression; for patient-level analyses, we used general estimating equations to adjust for clustering of patients within each session; for nurse presence during any part of a round by patient, we used a χ² test. Total non-patient time per round session was compared with use of patient-clustered general estimating equations using a γ distribution to account for the non-normality of the data.

Patient and Care Team Characteristics

Over the first year of the initiative, 85% of a team’s patients were on their assigned unit, and 87% of a unit’s patients were with the assigned team. Census numbers were 10.4 patients per general medicine team in April-June 2013 and 12.7 patients per team in April-June 2014, a 22% increase after care redesign. There were no statistically significant differences in patient characteristics, including age, sex, race, language, admission source, and comorbidity measure (Elixhauser score), between the pre-intervention and post-intervention study periods, except for a slightly higher proportion of patients admitted from home and fewer patients admitted directly from clinic (Table 1).

Primary Outcomes

Mean proportion of time the nurse was present on rounds per round session increased significantly (P < 0.001), from 24.1% to 67.8% (Figure 1A, Table 2). For individual patient encounters, the increased overall nursing presence was attributable to having more nurses on rounds and having nurses present for a larger proportion of individual rounding encounters (Figure 1B, Table 2). Nurses were present for at least some part of rounds for 53% of patients before the intervention and 93% afterward (P < 0.001). Mean proportion of round time by each of the 2 interns on each team decreased from 59.6% to 49.6% (P = 0.007).

Total bedside rounding time increased significantly ( P < 0.001), from 39.9% before the intervention to 55.8% afterward (Table 2). Meanwhile, percentage of rounding time spent on the unit but outside patient rooms decreased significantly ( P = 0.004), from 55.2% to 42.2%, as did rounding time on a unit completely different from the patient’s (4.9% before intervention, 2.0% afterward; P = 0.03). Again, patient-level results were similar (Figure 2, Table 2), but the decreased time spent on the unit, outside the patient rooms, was not significant.

Secondary Outcomes

Total rounding time decreased significantly, from a mean of 182 minutes (3.0 hours) at baseline to a mean of 146 minutes (2.4 hours) after the intervention, despite the higher post-intervention census. (When adjusted for patient census, the difference increased from 35.5 to 53.8 minutes; Table 2.) Mean rounding time per patient decreased significantly, from 14.7 minutes at baseline to 10.5 minutes after the intervention. For newly admitted patients, mean rounding time per patient decreased from 30.0 minutes before implementation to 16.3 minutes afterward. Mean rounding time also decreased, though much less, for subsequent-day patients (Table 2). For both new and existing patients, the decrease in rounding time largely was a reduction in time spent rounding outside patient rooms, with minimal impact on bedside time (Table 2). Mean time nurses were present during a patient’s rounds increased significantly, from 4.5 to 8.0 minutes (Table 2). Total nurse rounding time increased from 45.1 minutes per session to 98.8 minutes. Rounding time not related to patient discussion or evaluation decreased from 22.7 minutes per session to 13.3 minutes ( P = 0.003).

TMA of our care redesign initiative showed that this multipronged intervention, which included team regionalization, encouragement of bedside rounding with nurses, call structure changes, and attendings’ reading of admission notes before rounds, resulted in an increased proportion of rounding time spent with patients and an increased proportion of time nurses were present on rounds. Secondarily, round duration decreased even as patient census increased.

Regionalized teams have been found to improve interdisciplinary communication.¹ The present study elaborates on that finding by demonstrating a dramatic increase in nursing presence on rounds, likely resulting from the unit’s use of rounding schedules and nurses’ prioritization of rounding orders, both of which were made possible by geographic co-localization. Other research has noted that one of the most significant barriers to interdisciplinary rounds is difficulty coordinating the start times of physician/nurse bedside rounding encounters. The system we have studied directly addresses this difficulty.⁹ Of note, nursing presence on rounds is necessary but not sufficient for true physician–nurse collaboration and effective communication,¹ as reflected in a separate study of the intervention showing no significant difference in the concordance of the patient care plan between nurses and physicians before and after regionalization.¹² Additional interventions may be needed to ensure that communication during bedside rounds is effective.

Our regionalized teams spent a significantly higher proportion of rounding time bedside, likely because of a cultural shift in expectations and the increased convenience of seeing patients on the team’s unit. Nevertheless, bedside time was not 100%. Structural barriers (eg, patients off-unit for dialysis) and cultural barriers likely contributed to the less than full adoption of bedside rounding. As described previously, cultural barriers to bedside rounding include trainees’ anxiety about being questioned in front of patients, the desire to freely exchange academic ideas in a conference room, and attendings’ doubts about their bedside teaching ability.^1,9,13 Bedside rounds provide an important opportunity to apply the principles of patient- and family-centered care, including promotion of dignity and respect, information sharing, and collaboration. Thus, overcoming the concerns of housestaff and attendings and helping them feel prepared for bedside rounds can benefit the patient experience. More attention should be given to these practices as these types of interventions are implemented at Brigham and Women’s Hospital and elsewhere.^1,13-15

Another primary concern about interdisciplinary bedside rounding is the perception that it takes more time.⁹ Therefore, it was important for us to measure round duration as a balancing measure to be considered for our intervention. Fortunately, we found round duration decreased with regionalization and encouragement of bedside rounding. This decrease was driven largely by a significant decrease in mean rounding time per new patient, which may be attributable at least in part to setting expectations that attendings and residents will read admission notes before rounds and that interns will summarize rather than recount information from admission notes. However, we also found rounding time decreases for subsequent-day patients, suggesting an underlying time savings. Spending a larger proportion of time bedside may therefore result in more efficient rounds. Bedside presentations can reduce redundancies, such as discussing a patient’s case outside his or her room and subsequently walking in and going over much of the same information with the patient. Our model de-emphasizes data transfer in favor of discussion of care plans. There was also a decrease in non-patient time, likely reflecting reduced transit time for regionalized teams. This decrease aligns with a recent finding that bedside rounding was at least as efficient as rounding outside the room.¹⁶

Of note, though a larger percentage of time was spent bedside after implementation of the care redesign, the absolute amount of bedside time did not change significantly. Our data showed that, even with shorter rounds, the same amount of absolute time can be spent bedside, face to face with the patient, by increasing the proportion of bedside rounding time. In other words, teams on average did not spend more time with patients, though the content and the structure of those encounters may have changed. This finding may be attributable to eliminating redundancy, forgoing the outside-the-room discussion, and thus the largest time reductions were realized there. In addition, teams incompletely adopted beside rounds, as reflected in the data. We expect that, with more complete adoption, an even larger proportion of time will be spent bedside, and absolute time bedside might increase as a result.

An unexpected result of the care redesign was that interns’ proportion of rounding time decreased after the intervention. This decrease most likely is attributable to interns’ being less likely to participate in rounds for a co-intern’s patient, and to their staying outside that patient’s room to give themselves more time to advance the care of their own patients. Before the intervention, when more rounding time was spent outside patient rooms, interns were more likely to join rounds for their co-intern’s patients because they could easily break away, as needed, to continue care of their own patients. The resident is now encouraged to use the morning huddle to identify which patients likely have the most educational value, and both interns are expected to join the bedside rounds for these patients.

This study had a few limitations. First, the pre–post design made it difficult to exclude the possibility that other temporal changes may have affected outcomes, though we did account for time-of-year effects by aligning our data-collection phases. In addition, the authors, including the director of the general medical service, are unaware of any co-interventions during the study period. Second, the multipronged intervention included care team regionalization, encouragement of bedside rounding with nurses, call structure changes (from 4 days to daily admitting), and attendings’ reading of admission notes before rounds. Thus, parsing which component(s) contributed to the results was difficult, though all the changes instituted likely were necessary for system redesign. For example, regionalization of clinicians to unit-based teams was made possible by switching to a daily admitting system.

Time that team members spent preparing for rounds was not recorded before or after the intervention. Thus, the decrease in total rounding time could have been accompanied by an increase in time spent preparing for rounds. However, admission notes were available in our electronic medical record before and after the intervention, and most residents and attendings were already reading them pre-intervention. After the intervention, pre-round note reading was more clearly defined as an expectation, and we were able to set the expectation that interns should use their presentations to summarize rather than recount information. In addition, in the post-intervention period, we did not include time spent preparing rounding orders; as already noted, however, preparation took only 5 minutes per day. Also, we did not analyze the content or the quality of the discussion on rounds, but simply recorded who was present where and when. Regarding the effect of the intervention on patient care, results were mixed. As reported in 2016, we saw no difference in frequency of adverse events with this intervention.¹² However, a more sensitive measure of adverse events—used in a study on handoffs—showed our regionalization efforts had an additive effect on reducing overnight adverse events.¹⁷Researchers should now focus on the effects of care redesign on clinical outcomes, interdisciplinary care team communication, patient engagement and satisfaction, provider opinions of communication, workflow, patient care, and housestaff education. Our methodology can be used as a model to link structure, process, and outcome related to rounds and thereby better understand how best to optimize patient care and efficiency. Additional studies are needed to analyze the content of rounds and their association with patient and educational outcomes. Last, it will be important to conduct a study to see if the effects we have identified can be sustained. Such a study is already under way.

In conclusion, creating regionalized care teams and encouraging focused bedside rounds increased the proportion of bedside time and the presence of nurses on rounds. Rounds were shorter despite higher patient census. TMA revealed that regionalized care teams and bedside rounding at a large academic hospital are feasible, and are useful in establishing the necessary structures for increasing physician–nurse and provider–patient interactions.

Acknowledgments

The authors acknowledge Dr. Stan Ashley, Dr. Jacqueline Somerville, and Sheila Harris for their support of the regionalization initiative.

Disclosures

Dr. Schnipper received funding from Sanofi-aventis to conduct an investigator-initiated study to implement and evaluate a multi-faceted intervention to improve transitions of care in patients discharged home on insulin. The study was also supported by funding from the Marshall A. Wolf Medical Education Fund, Brigham and Women’s Hospital, and Dr. Stan Ashley, Chief Medical Officer, Brigham and Women’s Hospital. Some of the content of this article was orally presented at the annual meeting of the Society of Hospital Medicine; March 29-April 1, 2015; National Harbor, MD.

In recent years, rapid response teams (RRTs) have been widely implemented to improve patient safety and quality of care. RRTs traditionally are activated by providers to address a clinically deteriorating patient; trained nurses, respiratory care specialists, and physicians are brought bedside to assist in triage and management. After the Joint Commission¹ endorsed patient engagement as a strategy for enhancing patient safety, new initiatives were developed to meet the challenge. Programs designed to enhance patient engagement have taken a variety of forms, including educational campaigns encouraging patients to report adverse events, requests for handwashing by providers, and the institution of patient- and family-activated RRTs.² Patient involvement is viewed favorably and has been shown to increase patients’ perception of health care quality.³ Although these initiatives are presumed helpful in encouraging communication, there is limited evidence that more communication leads to safety improvements. Despite the increasing prevalence of patient-activated RRTs in the United States, they have gone largely unevaluated in the adult population, and their efficacy remains unclear.

Condition Help (CH) is a patient- and family-initiated RRT designed to prevent medical errors and communication problems and improve patient safety. Patients and families are encouraged to call the CH hotline if they believe that there has been a breakdown in care or that their health is in imminent danger. This RRT was inspired by the case of Josie King, an 18-month-old girl who died of preventable causes at a large children’s hospital.⁴ After her daughter’s death, Sorrel King started the Josie King Foundation, an organization committed to preventing medical errors and creating a culture of patient safety. With the support of this foundation, CH was launched in 2005 at the Children’s Hospital of Pittsburgh at the University of Pittsburgh Medical Center (UPMC). Later it was implemented at the UPMC adult tertiary-care center, and now it is available in all UPMC facilities.

On admission, patients receive a brochure that details the purpose of CH and provides examples of when and how to call the CH hotline. In this brochure, patients are instructed to call CH in 3 situations: “1) There is an emergency and you cannot get the attention of hospital staff, 2) You see a change in the patient’s condition and the health care team is not recognizing the concern, or 3) There is breakdown in how care is given or uncertainty over what needs to be done.” These instructions are printed on bulletins placed in elevators and hallways throughout the hospital. Patients and families may activate the system at any time and can even do so from home.

When a patient or family member calls the hotline, an operator notifies the CH team. This team, which consists of a patient care liaison (or an on-duty administrator) and the unit charge nurse, convenes bedside to address the patient’s concern. The team was designed without a physician to ensure that the primary team remains in charge of the care plan. CH is kept separate from our traditional RRT and does not compete for resources (personnel, equipment, time) with the RRT, which is designed to address a clinically deteriorating patient.

In this article, we describe the characteristics of patients for whom CH was activated at our adult hospital. We also describe reasons for calls, whether changes in care were implemented, and outcomes, including traditional RRT activation, transfer to intensive care unit (ICU), and inpatient mortality. As CH was designed with patient safety as a goal, we tracked 2 types of calls, those involving safety issues and those involving nonsafety issues.

This study was approved by the quality improvement committee at the University of Pittsburgh and was considered exempt from review by the university’s Institutional Review Board.

Our integrated health system consists of more than 20 hospitals serving a tristate region. UPMC Presbyterian and UPMC Montefiore are adult tertiary-care referral hospitals with more than 750 medical/surgical beds and 150 critical care beds and more than 30,000 annual inpatient admissions. These hospitals are physically connected and function as a single large medical center. We reviewed all CH events that occurred at this combined hospital during the period January 2012 through June 2015. The dates coincided with CH data acquisition.

CH was available 24 hours a day 7 days a week. A patient care liaison (or an on-duty administrator) and the unit charge nurse responded to CH calls. Data from all calls included date and time of call, day of week, primary service, patient location, unique patient identifiers, call initiator (patient or family), whether a call led to changes in care, and primary reason for call. Each call reason was sorted into 1 of 10 categories: pain control, staff problem, lack of communication between patient/family and care team, questions about patient management, care delays, delays in a particular service, questions about discharge, administrative issues, acute psychiatric needs, and unknown/other. In addition, after a call, we reviewed all charts to determine if a safety issue was involved; Dr. Eden and Dr. Bump independently reviewed calls for safety issues and discussed any differences until they reached consensus. We also recorded outcomes, including activation of a traditional RRT or transfer to ICU within 24 hours of CH call, inpatient mortality, and against medical advice (AMA) discharges. Given that many calls were made by patients who called more than once (during a single admission or over multiple admissions), we also sorted patients into one-time callers and repeat callers for comparison. Patient satisfaction data were unavailable for review.

Patient demographic data are presented as means, standard deviations, and percentages, and call characteristics as percentages. Chi-square tests and t tests were used for analyses except for comparisons having few observations. For those, Fisher exact test was used. All analyses were performed with SAS Version 9.4 (SAS Institute, Cary, North Carolina).

From January 2012 through June 2015, 367 CH calls were made, about 105 annually. During this period, there were about 33,000 admissions, 800 combined grievances and complaints, 170 AMA discharges, 155 cardiac arrests, 2300 traditional RRT activations, and 1200 inpatient deaths per year. The 367 CH calls were made by 240 patients (Table 1). Of these 240 patients, 43 (18%) activated the CH team with multiple calls; their calls accounted for (46.3%) of all calls (170/367). The majority of calls were made by patients (76.8%) rather than family members (21.8%). Mean (SD) patient age was 45.8 (17.4) years. Mean (SD) number of admissions per patient per year was 2.7 (3.5). More events were activated for patients admitted to medical services (66%) than surgical services (34%). Calls were evenly distributed between time of day and day of week.

The most common reason for CH calls was inadequate pain control (48.2%), followed by dissatisfaction with staff (12.5%); the remaining calls were evenly distributed among the other categories. The majority of calls involved nonsafety issues (83.4%) rather than safety issues (11.4%); in 5.2% of calls, the distinction could not be made because of lack of information (Table 2). In 152 (41.4%) of the 367 total calls, a change in care or alteration in management was made. Of these 152 calls, 99 (65.1%) involved distinct changes in the care plan, such as medication changes, imaging or additional testing, or consultation with other physicians; the other 53 calls (34.9%) involved additional patient counseling or nonmedical changes. Our traditional RRT was activated within 24 hours of CH in 19 cases (5.2%); of the 19 patients, 6 were transferred to ICU. Seven patients (2.9%) died during admission. Twelve (3.3%) were discharged AMA. We compared outcomes of patients who made safety-issue calls with those of patients who made nonsafety-issue calls. The composite outcome of RRT activation, ICU transfer, and mortality was found for 6 (14.3%) of the 42 safety-issue calls and 15 (4.9%) of the 306 nonsafety-issue calls (P = 0.0291).

Patient- and family-activated RRTs provide unique opportunities for patient and family engagement during inpatient hospital stays. Our study described the results obtained with use of a well-established patient-activated RRT over several years, one of the longer observation periods reported in the literature. We found that, with use of patient-activated RRTs, patient safety issues were identified, though these were far outnumbered by nonsafety issues.

Almost half of all CH events were related to pain. Pain as the primary driver for RRT activation may be attributable to several factors, including degree of illness, poor communication about pain management expectations, positive reinforcement of narcotic-seeking behavior as a result of CH activation, and high rate of opiate use in the catchment area. A striking finding of our analysis was repeat calling; only 43 (18%) of the 240 callers were repeat callers, but they made almost half of all the calls. In some cases, during a single admission, multiple calls were made because the first had no effect on care or management; more typically, though, multiple calls were made over several admissions. Repeat callers were admitted more often per year, and they used hospital services more. They should be further studied with a goal of designing programs that better meet their needs and that prospectively address expectations of pain control.

Our study was unique in describing several outcomes related to CH events. We found that traditional RRTs were seldom activated, level of care was seldom escalated, and mortality was rare, though these outcomes occurred more often for safety-issue calls than nonsafety-issue calls. We also found that activation of CH teams often led to changes in medical management, though we could not determine whether these changes in care led to different patient outcomes.

Patient-initiated RRTs are described in a limited number of pediatric and adult studies, all with findings differing from ours. In the pediatric models, most calls were initiated by family members, were less frequent, and tended to signal higher patient acuity.^5,6 For example, in a pediatric RRT model,⁵ family members activated the RRT only twice within the study year, but both calls resulted in ICU transfer. Most descriptions of patient-activated RRTs in adult hospitals are from pilot studies, which similarly identified infrequent RRT calls but often did not identify call reasons or specific outcomes.⁷ A single-center study concluded that, after implementation of a mixed-model RRT⁸—a traditional practitioner-activated RRT later enhanced with a patient/family activation mechanism—non-ICU codes decreased, and there was a statistically significant drop in hospital-wide mortality rates. However, this RRT was patient-activated only 25 times over 2 years, and the specific outcomes of those events were not described.

Other initiatives have been designed to enhance patient care and communication. Purposeful rounding systems⁹ involve hourly rounding by bedside nurses and daily rounding by nurse leaders to improve timely patient care and provide proactive service. Such systems ideally preempt calls involving dissatisfaction and nonsafety issues. Although they would reduce the number of patient-dissatisfaction calls made in the CH system, they may not be any better than the CH system is in its main purpose, identifying safety issues. In addition, whether patient-activated RRTs or purposeful rounding systems are better at addressing patient dissatisfaction is unclear.

This study had its limitations. First, like other studies, it was a single-center observational study without a concurrent control group. Second, because CH was first implemented 10 years ago, we could not compare patient outcomes or traditional RRT use before and after program initiation. Third, our study cohort consisted of patients hospitalized at one academic tertiary-care center in one region, and the hospital is a training site for multiple residencies and fellowships. These factors likely affect the generalizability of our data to smaller or community-based centers. Fourth, some determinations were subjective (eg, whether calls involved safety or nonsafety issues). We tried to minimize bias by having 2 authors independently review cases, but the process did not reflect patient experience or perspective. Fifth, our hospital adopted its traditional RRT years before its CH system. The criteria used by hospital personnel for traditional RRT activation are designed to encourage staff to call for help at early signs of patient deterioration. Consequently, traditional RRT activations substantially outnumber CH calls. Whether this resulted in fewer CH safety calls is unclear. Sixth, we did not capture the financial implications of using CH teams.

Although patient-activated RRTs identified patient safety issues, questions about the utility or necessity of these RRTs remain. In our era of limited hospital resources, the case has not been definitively made that these teams are practical, based on patient outcomes, though other studies have found improved patient satisfaction.⁷ Most of the RRT calls in our study involved patient dissatisfaction and communication issues. CH may not be the ideal approach for managing these issues, but it represents the last line of patient advocacy once other systems have failed.

We think patient-activated RRTs have the potential to effect patient engagement in safe care. Given the importance of establishing a culture of patient safety and engagement, and increased detection of safety-related events, CH remains active throughout our hospital system. Newer iterations of CH may benefit from stricter language in defining appropriate occasions for calling RRTs, and from descriptions of other resources for patient advocacy within the hospital. These modifications could end up restricting RRT activations to patient complaints and preserving CH resources for patients with safety concerns. Our study lays the groundwork for other institutions that are considering similar interventions. Studies should now start evaluating how well patient- and family-activated RRTs improve patient satisfaction, staff satisfaction, and patient outcomes.

Patient- and family-activated RRTs were designed to engage patients and families in safe care. Although CH detects patient safety issues, these are far outnumbered by nonsafety issues. CH demonstrates a commitment to patient engagement and a culture that emphasizes patient safety.

Acknowledgements

This work was presented as a poster at the annual meeting of the Society of Hospital Medicine; March 6-9, 2016; San Diego, CA.

Disclosure

Nothing to report.

In recent years, rapid response teams (RRTs) have been widely implemented to improve patient safety and quality of care. RRTs traditionally are activated by providers to address a clinically deteriorating patient; trained nurses, respiratory care specialists, and physicians are brought bedside to assist in triage and management. After the Joint Commission¹ endorsed patient engagement as a strategy for enhancing patient safety, new initiatives were developed to meet the challenge. Programs designed to enhance patient engagement have taken a variety of forms, including educational campaigns encouraging patients to report adverse events, requests for handwashing by providers, and the institution of patient- and family-activated RRTs.² Patient involvement is viewed favorably and has been shown to increase patients’ perception of health care quality.³ Although these initiatives are presumed helpful in encouraging communication, there is limited evidence that more communication leads to safety improvements. Despite the increasing prevalence of patient-activated RRTs in the United States, they have gone largely unevaluated in the adult population, and their efficacy remains unclear.

Condition Help (CH) is a patient- and family-initiated RRT designed to prevent medical errors and communication problems and improve patient safety. Patients and families are encouraged to call the CH hotline if they believe that there has been a breakdown in care or that their health is in imminent danger. This RRT was inspired by the case of Josie King, an 18-month-old girl who died of preventable causes at a large children’s hospital.⁴ After her daughter’s death, Sorrel King started the Josie King Foundation, an organization committed to preventing medical errors and creating a culture of patient safety. With the support of this foundation, CH was launched in 2005 at the Children’s Hospital of Pittsburgh at the University of Pittsburgh Medical Center (UPMC). Later it was implemented at the UPMC adult tertiary-care center, and now it is available in all UPMC facilities.

On admission, patients receive a brochure that details the purpose of CH and provides examples of when and how to call the CH hotline. In this brochure, patients are instructed to call CH in 3 situations: “1) There is an emergency and you cannot get the attention of hospital staff, 2) You see a change in the patient’s condition and the health care team is not recognizing the concern, or 3) There is breakdown in how care is given or uncertainty over what needs to be done.” These instructions are printed on bulletins placed in elevators and hallways throughout the hospital. Patients and families may activate the system at any time and can even do so from home.

When a patient or family member calls the hotline, an operator notifies the CH team. This team, which consists of a patient care liaison (or an on-duty administrator) and the unit charge nurse, convenes bedside to address the patient’s concern. The team was designed without a physician to ensure that the primary team remains in charge of the care plan. CH is kept separate from our traditional RRT and does not compete for resources (personnel, equipment, time) with the RRT, which is designed to address a clinically deteriorating patient.

In this article, we describe the characteristics of patients for whom CH was activated at our adult hospital. We also describe reasons for calls, whether changes in care were implemented, and outcomes, including traditional RRT activation, transfer to intensive care unit (ICU), and inpatient mortality. As CH was designed with patient safety as a goal, we tracked 2 types of calls, those involving safety issues and those involving nonsafety issues.

This study was approved by the quality improvement committee at the University of Pittsburgh and was considered exempt from review by the university’s Institutional Review Board.

Our integrated health system consists of more than 20 hospitals serving a tristate region. UPMC Presbyterian and UPMC Montefiore are adult tertiary-care referral hospitals with more than 750 medical/surgical beds and 150 critical care beds and more than 30,000 annual inpatient admissions. These hospitals are physically connected and function as a single large medical center. We reviewed all CH events that occurred at this combined hospital during the period January 2012 through June 2015. The dates coincided with CH data acquisition.

CH was available 24 hours a day 7 days a week. A patient care liaison (or an on-duty administrator) and the unit charge nurse responded to CH calls. Data from all calls included date and time of call, day of week, primary service, patient location, unique patient identifiers, call initiator (patient or family), whether a call led to changes in care, and primary reason for call. Each call reason was sorted into 1 of 10 categories: pain control, staff problem, lack of communication between patient/family and care team, questions about patient management, care delays, delays in a particular service, questions about discharge, administrative issues, acute psychiatric needs, and unknown/other. In addition, after a call, we reviewed all charts to determine if a safety issue was involved; Dr. Eden and Dr. Bump independently reviewed calls for safety issues and discussed any differences until they reached consensus. We also recorded outcomes, including activation of a traditional RRT or transfer to ICU within 24 hours of CH call, inpatient mortality, and against medical advice (AMA) discharges. Given that many calls were made by patients who called more than once (during a single admission or over multiple admissions), we also sorted patients into one-time callers and repeat callers for comparison. Patient satisfaction data were unavailable for review.

Patient demographic data are presented as means, standard deviations, and percentages, and call characteristics as percentages. Chi-square tests and t tests were used for analyses except for comparisons having few observations. For those, Fisher exact test was used. All analyses were performed with SAS Version 9.4 (SAS Institute, Cary, North Carolina).

From January 2012 through June 2015, 367 CH calls were made, about 105 annually. During this period, there were about 33,000 admissions, 800 combined grievances and complaints, 170 AMA discharges, 155 cardiac arrests, 2300 traditional RRT activations, and 1200 inpatient deaths per year. The 367 CH calls were made by 240 patients (Table 1). Of these 240 patients, 43 (18%) activated the CH team with multiple calls; their calls accounted for (46.3%) of all calls (170/367). The majority of calls were made by patients (76.8%) rather than family members (21.8%). Mean (SD) patient age was 45.8 (17.4) years. Mean (SD) number of admissions per patient per year was 2.7 (3.5). More events were activated for patients admitted to medical services (66%) than surgical services (34%). Calls were evenly distributed between time of day and day of week.

The most common reason for CH calls was inadequate pain control (48.2%), followed by dissatisfaction with staff (12.5%); the remaining calls were evenly distributed among the other categories. The majority of calls involved nonsafety issues (83.4%) rather than safety issues (11.4%); in 5.2% of calls, the distinction could not be made because of lack of information (Table 2). In 152 (41.4%) of the 367 total calls, a change in care or alteration in management was made. Of these 152 calls, 99 (65.1%) involved distinct changes in the care plan, such as medication changes, imaging or additional testing, or consultation with other physicians; the other 53 calls (34.9%) involved additional patient counseling or nonmedical changes. Our traditional RRT was activated within 24 hours of CH in 19 cases (5.2%); of the 19 patients, 6 were transferred to ICU. Seven patients (2.9%) died during admission. Twelve (3.3%) were discharged AMA. We compared outcomes of patients who made safety-issue calls with those of patients who made nonsafety-issue calls. The composite outcome of RRT activation, ICU transfer, and mortality was found for 6 (14.3%) of the 42 safety-issue calls and 15 (4.9%) of the 306 nonsafety-issue calls (P = 0.0291).

Patient- and family-activated RRTs provide unique opportunities for patient and family engagement during inpatient hospital stays. Our study described the results obtained with use of a well-established patient-activated RRT over several years, one of the longer observation periods reported in the literature. We found that, with use of patient-activated RRTs, patient safety issues were identified, though these were far outnumbered by nonsafety issues.

Almost half of all CH events were related to pain. Pain as the primary driver for RRT activation may be attributable to several factors, including degree of illness, poor communication about pain management expectations, positive reinforcement of narcotic-seeking behavior as a result of CH activation, and high rate of opiate use in the catchment area. A striking finding of our analysis was repeat calling; only 43 (18%) of the 240 callers were repeat callers, but they made almost half of all the calls. In some cases, during a single admission, multiple calls were made because the first had no effect on care or management; more typically, though, multiple calls were made over several admissions. Repeat callers were admitted more often per year, and they used hospital services more. They should be further studied with a goal of designing programs that better meet their needs and that prospectively address expectations of pain control.

Our study was unique in describing several outcomes related to CH events. We found that traditional RRTs were seldom activated, level of care was seldom escalated, and mortality was rare, though these outcomes occurred more often for safety-issue calls than nonsafety-issue calls. We also found that activation of CH teams often led to changes in medical management, though we could not determine whether these changes in care led to different patient outcomes.

Patient-initiated RRTs are described in a limited number of pediatric and adult studies, all with findings differing from ours. In the pediatric models, most calls were initiated by family members, were less frequent, and tended to signal higher patient acuity.^5,6 For example, in a pediatric RRT model,⁵ family members activated the RRT only twice within the study year, but both calls resulted in ICU transfer. Most descriptions of patient-activated RRTs in adult hospitals are from pilot studies, which similarly identified infrequent RRT calls but often did not identify call reasons or specific outcomes.⁷ A single-center study concluded that, after implementation of a mixed-model RRT⁸—a traditional practitioner-activated RRT later enhanced with a patient/family activation mechanism—non-ICU codes decreased, and there was a statistically significant drop in hospital-wide mortality rates. However, this RRT was patient-activated only 25 times over 2 years, and the specific outcomes of those events were not described.

Other initiatives have been designed to enhance patient care and communication. Purposeful rounding systems⁹ involve hourly rounding by bedside nurses and daily rounding by nurse leaders to improve timely patient care and provide proactive service. Such systems ideally preempt calls involving dissatisfaction and nonsafety issues. Although they would reduce the number of patient-dissatisfaction calls made in the CH system, they may not be any better than the CH system is in its main purpose, identifying safety issues. In addition, whether patient-activated RRTs or purposeful rounding systems are better at addressing patient dissatisfaction is unclear.

This study had its limitations. First, like other studies, it was a single-center observational study without a concurrent control group. Second, because CH was first implemented 10 years ago, we could not compare patient outcomes or traditional RRT use before and after program initiation. Third, our study cohort consisted of patients hospitalized at one academic tertiary-care center in one region, and the hospital is a training site for multiple residencies and fellowships. These factors likely affect the generalizability of our data to smaller or community-based centers. Fourth, some determinations were subjective (eg, whether calls involved safety or nonsafety issues). We tried to minimize bias by having 2 authors independently review cases, but the process did not reflect patient experience or perspective. Fifth, our hospital adopted its traditional RRT years before its CH system. The criteria used by hospital personnel for traditional RRT activation are designed to encourage staff to call for help at early signs of patient deterioration. Consequently, traditional RRT activations substantially outnumber CH calls. Whether this resulted in fewer CH safety calls is unclear. Sixth, we did not capture the financial implications of using CH teams.

Although patient-activated RRTs identified patient safety issues, questions about the utility or necessity of these RRTs remain. In our era of limited hospital resources, the case has not been definitively made that these teams are practical, based on patient outcomes, though other studies have found improved patient satisfaction.⁷ Most of the RRT calls in our study involved patient dissatisfaction and communication issues. CH may not be the ideal approach for managing these issues, but it represents the last line of patient advocacy once other systems have failed.

We think patient-activated RRTs have the potential to effect patient engagement in safe care. Given the importance of establishing a culture of patient safety and engagement, and increased detection of safety-related events, CH remains active throughout our hospital system. Newer iterations of CH may benefit from stricter language in defining appropriate occasions for calling RRTs, and from descriptions of other resources for patient advocacy within the hospital. These modifications could end up restricting RRT activations to patient complaints and preserving CH resources for patients with safety concerns. Our study lays the groundwork for other institutions that are considering similar interventions. Studies should now start evaluating how well patient- and family-activated RRTs improve patient satisfaction, staff satisfaction, and patient outcomes.

Patient- and family-activated RRTs were designed to engage patients and families in safe care. Although CH detects patient safety issues, these are far outnumbered by nonsafety issues. CH demonstrates a commitment to patient engagement and a culture that emphasizes patient safety.

Acknowledgements

This work was presented as a poster at the annual meeting of the Society of Hospital Medicine; March 6-9, 2016; San Diego, CA.

Disclosure

Nothing to report.

In recent years, rapid response teams (RRTs) have been widely implemented to improve patient safety and quality of care. RRTs traditionally are activated by providers to address a clinically deteriorating patient; trained nurses, respiratory care specialists, and physicians are brought bedside to assist in triage and management. After the Joint Commission¹ endorsed patient engagement as a strategy for enhancing patient safety, new initiatives were developed to meet the challenge. Programs designed to enhance patient engagement have taken a variety of forms, including educational campaigns encouraging patients to report adverse events, requests for handwashing by providers, and the institution of patient- and family-activated RRTs.² Patient involvement is viewed favorably and has been shown to increase patients’ perception of health care quality.³ Although these initiatives are presumed helpful in encouraging communication, there is limited evidence that more communication leads to safety improvements. Despite the increasing prevalence of patient-activated RRTs in the United States, they have gone largely unevaluated in the adult population, and their efficacy remains unclear.

Condition Help (CH) is a patient- and family-initiated RRT designed to prevent medical errors and communication problems and improve patient safety. Patients and families are encouraged to call the CH hotline if they believe that there has been a breakdown in care or that their health is in imminent danger. This RRT was inspired by the case of Josie King, an 18-month-old girl who died of preventable causes at a large children’s hospital.⁴ After her daughter’s death, Sorrel King started the Josie King Foundation, an organization committed to preventing medical errors and creating a culture of patient safety. With the support of this foundation, CH was launched in 2005 at the Children’s Hospital of Pittsburgh at the University of Pittsburgh Medical Center (UPMC). Later it was implemented at the UPMC adult tertiary-care center, and now it is available in all UPMC facilities.

On admission, patients receive a brochure that details the purpose of CH and provides examples of when and how to call the CH hotline. In this brochure, patients are instructed to call CH in 3 situations: “1) There is an emergency and you cannot get the attention of hospital staff, 2) You see a change in the patient’s condition and the health care team is not recognizing the concern, or 3) There is breakdown in how care is given or uncertainty over what needs to be done.” These instructions are printed on bulletins placed in elevators and hallways throughout the hospital. Patients and families may activate the system at any time and can even do so from home.

When a patient or family member calls the hotline, an operator notifies the CH team. This team, which consists of a patient care liaison (or an on-duty administrator) and the unit charge nurse, convenes bedside to address the patient’s concern. The team was designed without a physician to ensure that the primary team remains in charge of the care plan. CH is kept separate from our traditional RRT and does not compete for resources (personnel, equipment, time) with the RRT, which is designed to address a clinically deteriorating patient.

In this article, we describe the characteristics of patients for whom CH was activated at our adult hospital. We also describe reasons for calls, whether changes in care were implemented, and outcomes, including traditional RRT activation, transfer to intensive care unit (ICU), and inpatient mortality. As CH was designed with patient safety as a goal, we tracked 2 types of calls, those involving safety issues and those involving nonsafety issues.

This study was approved by the quality improvement committee at the University of Pittsburgh and was considered exempt from review by the university’s Institutional Review Board.

Our integrated health system consists of more than 20 hospitals serving a tristate region. UPMC Presbyterian and UPMC Montefiore are adult tertiary-care referral hospitals with more than 750 medical/surgical beds and 150 critical care beds and more than 30,000 annual inpatient admissions. These hospitals are physically connected and function as a single large medical center. We reviewed all CH events that occurred at this combined hospital during the period January 2012 through June 2015. The dates coincided with CH data acquisition.

CH was available 24 hours a day 7 days a week. A patient care liaison (or an on-duty administrator) and the unit charge nurse responded to CH calls. Data from all calls included date and time of call, day of week, primary service, patient location, unique patient identifiers, call initiator (patient or family), whether a call led to changes in care, and primary reason for call. Each call reason was sorted into 1 of 10 categories: pain control, staff problem, lack of communication between patient/family and care team, questions about patient management, care delays, delays in a particular service, questions about discharge, administrative issues, acute psychiatric needs, and unknown/other. In addition, after a call, we reviewed all charts to determine if a safety issue was involved; Dr. Eden and Dr. Bump independently reviewed calls for safety issues and discussed any differences until they reached consensus. We also recorded outcomes, including activation of a traditional RRT or transfer to ICU within 24 hours of CH call, inpatient mortality, and against medical advice (AMA) discharges. Given that many calls were made by patients who called more than once (during a single admission or over multiple admissions), we also sorted patients into one-time callers and repeat callers for comparison. Patient satisfaction data were unavailable for review.

Patient demographic data are presented as means, standard deviations, and percentages, and call characteristics as percentages. Chi-square tests and t tests were used for analyses except for comparisons having few observations. For those, Fisher exact test was used. All analyses were performed with SAS Version 9.4 (SAS Institute, Cary, North Carolina).

From January 2012 through June 2015, 367 CH calls were made, about 105 annually. During this period, there were about 33,000 admissions, 800 combined grievances and complaints, 170 AMA discharges, 155 cardiac arrests, 2300 traditional RRT activations, and 1200 inpatient deaths per year. The 367 CH calls were made by 240 patients (Table 1). Of these 240 patients, 43 (18%) activated the CH team with multiple calls; their calls accounted for (46.3%) of all calls (170/367). The majority of calls were made by patients (76.8%) rather than family members (21.8%). Mean (SD) patient age was 45.8 (17.4) years. Mean (SD) number of admissions per patient per year was 2.7 (3.5). More events were activated for patients admitted to medical services (66%) than surgical services (34%). Calls were evenly distributed between time of day and day of week.

The most common reason for CH calls was inadequate pain control (48.2%), followed by dissatisfaction with staff (12.5%); the remaining calls were evenly distributed among the other categories. The majority of calls involved nonsafety issues (83.4%) rather than safety issues (11.4%); in 5.2% of calls, the distinction could not be made because of lack of information (Table 2). In 152 (41.4%) of the 367 total calls, a change in care or alteration in management was made. Of these 152 calls, 99 (65.1%) involved distinct changes in the care plan, such as medication changes, imaging or additional testing, or consultation with other physicians; the other 53 calls (34.9%) involved additional patient counseling or nonmedical changes. Our traditional RRT was activated within 24 hours of CH in 19 cases (5.2%); of the 19 patients, 6 were transferred to ICU. Seven patients (2.9%) died during admission. Twelve (3.3%) were discharged AMA. We compared outcomes of patients who made safety-issue calls with those of patients who made nonsafety-issue calls. The composite outcome of RRT activation, ICU transfer, and mortality was found for 6 (14.3%) of the 42 safety-issue calls and 15 (4.9%) of the 306 nonsafety-issue calls (P = 0.0291).

Patient- and family-activated RRTs provide unique opportunities for patient and family engagement during inpatient hospital stays. Our study described the results obtained with use of a well-established patient-activated RRT over several years, one of the longer observation periods reported in the literature. We found that, with use of patient-activated RRTs, patient safety issues were identified, though these were far outnumbered by nonsafety issues.

Almost half of all CH events were related to pain. Pain as the primary driver for RRT activation may be attributable to several factors, including degree of illness, poor communication about pain management expectations, positive reinforcement of narcotic-seeking behavior as a result of CH activation, and high rate of opiate use in the catchment area. A striking finding of our analysis was repeat calling; only 43 (18%) of the 240 callers were repeat callers, but they made almost half of all the calls. In some cases, during a single admission, multiple calls were made because the first had no effect on care or management; more typically, though, multiple calls were made over several admissions. Repeat callers were admitted more often per year, and they used hospital services more. They should be further studied with a goal of designing programs that better meet their needs and that prospectively address expectations of pain control.

Our study was unique in describing several outcomes related to CH events. We found that traditional RRTs were seldom activated, level of care was seldom escalated, and mortality was rare, though these outcomes occurred more often for safety-issue calls than nonsafety-issue calls. We also found that activation of CH teams often led to changes in medical management, though we could not determine whether these changes in care led to different patient outcomes.

Patient-initiated RRTs are described in a limited number of pediatric and adult studies, all with findings differing from ours. In the pediatric models, most calls were initiated by family members, were less frequent, and tended to signal higher patient acuity.^5,6 For example, in a pediatric RRT model,⁵ family members activated the RRT only twice within the study year, but both calls resulted in ICU transfer. Most descriptions of patient-activated RRTs in adult hospitals are from pilot studies, which similarly identified infrequent RRT calls but often did not identify call reasons or specific outcomes.⁷ A single-center study concluded that, after implementation of a mixed-model RRT⁸—a traditional practitioner-activated RRT later enhanced with a patient/family activation mechanism—non-ICU codes decreased, and there was a statistically significant drop in hospital-wide mortality rates. However, this RRT was patient-activated only 25 times over 2 years, and the specific outcomes of those events were not described.

Other initiatives have been designed to enhance patient care and communication. Purposeful rounding systems⁹ involve hourly rounding by bedside nurses and daily rounding by nurse leaders to improve timely patient care and provide proactive service. Such systems ideally preempt calls involving dissatisfaction and nonsafety issues. Although they would reduce the number of patient-dissatisfaction calls made in the CH system, they may not be any better than the CH system is in its main purpose, identifying safety issues. In addition, whether patient-activated RRTs or purposeful rounding systems are better at addressing patient dissatisfaction is unclear.

This study had its limitations. First, like other studies, it was a single-center observational study without a concurrent control group. Second, because CH was first implemented 10 years ago, we could not compare patient outcomes or traditional RRT use before and after program initiation. Third, our study cohort consisted of patients hospitalized at one academic tertiary-care center in one region, and the hospital is a training site for multiple residencies and fellowships. These factors likely affect the generalizability of our data to smaller or community-based centers. Fourth, some determinations were subjective (eg, whether calls involved safety or nonsafety issues). We tried to minimize bias by having 2 authors independently review cases, but the process did not reflect patient experience or perspective. Fifth, our hospital adopted its traditional RRT years before its CH system. The criteria used by hospital personnel for traditional RRT activation are designed to encourage staff to call for help at early signs of patient deterioration. Consequently, traditional RRT activations substantially outnumber CH calls. Whether this resulted in fewer CH safety calls is unclear. Sixth, we did not capture the financial implications of using CH teams.

Although patient-activated RRTs identified patient safety issues, questions about the utility or necessity of these RRTs remain. In our era of limited hospital resources, the case has not been definitively made that these teams are practical, based on patient outcomes, though other studies have found improved patient satisfaction.⁷ Most of the RRT calls in our study involved patient dissatisfaction and communication issues. CH may not be the ideal approach for managing these issues, but it represents the last line of patient advocacy once other systems have failed.

We think patient-activated RRTs have the potential to effect patient engagement in safe care. Given the importance of establishing a culture of patient safety and engagement, and increased detection of safety-related events, CH remains active throughout our hospital system. Newer iterations of CH may benefit from stricter language in defining appropriate occasions for calling RRTs, and from descriptions of other resources for patient advocacy within the hospital. These modifications could end up restricting RRT activations to patient complaints and preserving CH resources for patients with safety concerns. Our study lays the groundwork for other institutions that are considering similar interventions. Studies should now start evaluating how well patient- and family-activated RRTs improve patient satisfaction, staff satisfaction, and patient outcomes.

Patient- and family-activated RRTs were designed to engage patients and families in safe care. Although CH detects patient safety issues, these are far outnumbered by nonsafety issues. CH demonstrates a commitment to patient engagement and a culture that emphasizes patient safety.

Acknowledgements

This work was presented as a poster at the annual meeting of the Society of Hospital Medicine; March 6-9, 2016; San Diego, CA.

Disclosure

Nothing to report.

Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,¹ and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.² While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.³

When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;^4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.⁵ However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.⁷ One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.⁸ This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.⁸ However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.

The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.

In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger study to implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.

We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.

We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.

All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.

We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;⁹ we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.

A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,¹⁰ which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.

Qualitative data were analyzed using a constant comparative analytic technique.¹¹ The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.¹² All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.⁹ Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).

We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.

We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.

Perceived Impact on Patient Outcomes

Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:

“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”

Hospital Structural Characteristics

For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”

Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.

Provider Knowledge and Self-Efficacy

Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.

For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.

A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”

Hospital Priorities

Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”

Continuity and Communication

According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.

Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”

Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”

Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,¹³ clinical inertia,¹⁴ difficulty with close monitoring of patients,¹⁵ and care fragmentation.¹⁶ Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,⁷ providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.

Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,⁷ our study and others⁸ have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.¹⁷ The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.^4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.

For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.¹⁹

Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.

The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.

In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.

Disclosures

This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.

Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,¹ and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.² While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.³

When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;^4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.⁵ However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.⁷ One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.⁸ This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.⁸ However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.

The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.

In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger study to implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.

We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.

We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.

All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.

We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;⁹ we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.

A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,¹⁰ which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.

Qualitative data were analyzed using a constant comparative analytic technique.¹¹ The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.¹² All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.⁹ Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).

We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.

We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.

Perceived Impact on Patient Outcomes

Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:

“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”

Hospital Structural Characteristics

For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”

Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.

Provider Knowledge and Self-Efficacy

Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.

For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.

A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”

Hospital Priorities

Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”

Continuity and Communication

According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.

Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”

Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”

Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,¹³ clinical inertia,¹⁴ difficulty with close monitoring of patients,¹⁵ and care fragmentation.¹⁶ Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,⁷ providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.

Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,⁷ our study and others⁸ have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.¹⁷ The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.^4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.

For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.¹⁹

Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.

The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.

In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.

Disclosures

This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.

Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,¹ and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.² While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.³

When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;^4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.⁵ However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.⁷ One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.⁸ This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.⁸ However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.

The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.

In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger study to implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.

We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.

We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.

All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.

We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;⁹ we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.

A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,¹⁰ which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.

Qualitative data were analyzed using a constant comparative analytic technique.¹¹ The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.¹² All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.⁹ Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).

We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.

We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.

Perceived Impact on Patient Outcomes

Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:

“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”

Hospital Structural Characteristics

For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”

Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.

Provider Knowledge and Self-Efficacy

Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.

For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.

A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”

Hospital Priorities

Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”

Continuity and Communication

According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.

Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”

Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”

Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,¹³ clinical inertia,¹⁴ difficulty with close monitoring of patients,¹⁵ and care fragmentation.¹⁶ Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,⁷ providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.

Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,⁷ our study and others⁸ have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.¹⁷ The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.^4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.

For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.¹⁹

Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.

The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.

In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.

Disclosures

This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.

As the US population ages and becomes increasingly frail, the need for rehabilitation rises. By 2030, an estimated 20% of the population will be 65 years old or older, and almost 10% will be over 75.¹ About 20% of hospitalized Medicare patients receive subsequent care in postacute inpatient rehabilitation (PAIR) facilities, accounting for $31 billion in Medicare expenditures in 2014.² Although the need for rehabilitation will continue to rise, Medicare policy restricts access to it.

Under Medicare policy, PAIR services are covered for certain hospitalized patients but not others. Hospitalized patients are either inpatients, who are billed under Medicare Part A, or outpatients, billed under Part B. When hospital length of stay (LOS) is anticipated to be less than 2 midnights, patients are admitted as outpatients under the term observation status; when longer stays are expected, patients are admitted as inpatients.³ This recently implemented time-based distinction has been criticized as arbitrary, and as potentially shifting many patients from inpatient to outpatient (observation) status.⁴

The distinction between inpatient and observation status has significant consequences for posthospital care. Medicare Part A covers care in skilled nursing facilities (SNFs) and acute inpatient rehabilitation facilities (IRFs); after hospitalization, inpatients have access to either, without copay. As observation patients are covered under Medicare Part B, they are technically not covered for either service after their hospital stay. IRFs sometimes accept patients from ambulatory and nonacute settings; observation patients may be accepted in rare circumstances, but they pay the Part A deductible ($1288 in 2016) to have the services covered by Medicare. SNF services are never covered for observation patients, and access to this care requires an average out-of-pocket payment of more than $10,503 per beneficiary for a typical SNF stay.⁵ Given that about 70% of Medicare patients fall below 300% of the federal poverty line,⁶ the out-of-pocket costs for PAIR services for observation patients can be prohibitive.

Although only 0.75% of community-dwelling Medicare observation patients are discharged to PAIR facilities,⁷ it is unclear if the need for this care is higher but remains unmet secondary to cost concerns of Medicare beneficiaries. Also unclear is whether observation patients who would benefit from this care but do not receive it end up with poorer health outcomes and therefore use more healthcare services.

The purpose of this study was to estimate the proportion of Medicare observation patients who are admitted from home and receive a recommendation for placement in a PAIR facility, and to determine the ultimate disposition of such patients. We also sought to evaluate the association between recommendation for PAIR placement, LOS, and 30-day hospital revisit rate.

The Institutional Review Board of Christiana Care Health System (CCHS) approved this study.

Sample and Design

This was an observational study of community-dwelling Medicare patients admitted under observation status to Delaware’s CCHS, which consists of a 907-bed regional tertiary-care facility in Newark and a 241-bed community hospital in Wilmington. The study period was January 1 to December 31, 2013. We limited our sample to patients treated by hospitalists on hospital wards, as this care constitutes 80% of the care provided to observation patients at CCHS and the majority of care nationally.⁸ As neither SNF care nor IRF care is covered under Medicare Part B, and both would result in high out-of-pocket costs for Medicare observation patients, we combined them into a single variable, PAIR.

All data were obtained from institutional electronic medical record and administrative data systems. Study inclusion criteria were Medicare as primary insurance, admission to hospital from home, and care received at either CCHS facility. Exclusion criteria were admission from PAIR facility, long-term care facility, assisted-living facility, or inpatient psychiatric facility; death; discharge against medical advice (AMA) or to hospice, non-SNF, or inpatient psychiatric facility; and discovery (during review of case management [CM] notes) of erroneous listing of Medicare as primary insurance, or of inpatient admission (within 30 days before index observation stay) that qualified for PAIR coverage under Medicare Part A.

We reviewed the medical charts of a representative (~30%) sample of the cohort and examined physical therapy (PT) and CM notes to determine the proportions of patients with recommendations for home with no services, home-based PT, possible PAIR, and PAIR. Charts were sorted by medical record number and were reviewed in consecutive order. We coded a patient as having a recommendation for possible PAIR if the PT notes indicated the patient may benefit from PAIR but could have home PT if PAIR placement was not possible. CM notes were also reviewed for evidence of patient or family preference regarding PAIR placement. All questions about PT and CM recommendations were resolved by consensus.

Measures

For the total study sample, we calculated descriptive statistics and frequencies for demographic and administrative variables, including age, sex, race (Caucasian, African American, other), ethnicity (Hispanic/non-Hispanic), ICD-9 (International Classification of Diseases, Ninth Revision) primary diagnosis code, LOS (in hours) for index observation admission, discharge disposition (home with no services, home PT, possible PAIR, PAIR), and 30-day hospital revisit (emergency department, observation, inpatient admission). We used χ² test, Student t test, and analysis of variance (ANOVA) to test for statistically significant differences in characteristics between the chart review subgroup and the rest of the sample and between the groups with different disposition recommendations from PT notes.

For the chart review subgroup, we used ANOVA to calculate the unadjusted association between PT recommendation and LOS. We then adjusted for potential confounders, using multivariable linear regression with PT recommendation as a predictor and LOS as the outcome, controlling for variables previously associated with increased LOS among observation patients (primary diagnosis category, age, sex).⁶ We also adjusted for hospitalist group to account for potential variability in care delivery. As LOS was not normally distributed, we calculated the fourth root of LOS, which resulted in a more normal distribution, and used the transformed values in the regression model. We then calculated predicted values from the regression and back-transformed these to obtain adjusted mean values for LOS.

Of the 1417 unique patients who had Medicare as primary insurance and were admitted under observation status to a hospitalist service during the study period (2013), 94 were excluded (Figure). Of the remaining 1323 patients, the majority were 65 years old or older, female, white, and non-Hispanic. The most common ICD-9 diagnoses were syncope and chest pain. Mean LOS was 46.7 hours (range, 0-519 h). Less than 1% of patients were discharged to PAIR. Almost 25% of patients returned to the hospital, either for an emergency department visit or for observation or inpatient stay, within 30 days (Table).

Of the 419 charts reviewed to determine the proportion of patients evaluated by PT, and their subsequent recommendations, 33 were excluded, leaving 386 (92%) for analysis (Figure). There were no significant demographic differences between the patients in the chart review subgroup and the rest of the patients (Appendix). Of the 386 patients whose charts were analyzed, 181 (46.9%) had a PT evaluation, and 17 (4.4%) received a PAIR recommendation (Figure). Of the 17 patients recommended for PAIR, 12 (70.5%) were 65 years old or older, and 1 was discharged to a PAIR facility. Of the 46 patients recommended for home PT, 29 (63%) were discharged home with no services (Table).

PT-evaluated patients had unadjusted mean LOS of 52.2 hours (discharged home with no services), 64.1 hours (home PT or possible PAIR), and 83.1 hours (PAIR) (P = 0.001). With adjustment made for variables previously associated with increased LOS for observation patients, mean LOS for patients recommended for PAIR remained higher than that for patients in the other 2 categories (Table). Patients recommended for PAIR were more likely to return to hospital within 30 days than patients recommended for home PT or possible PAIR and patients discharged home with no services (Table).

Review of CM notes revealed that, of the 17 patients recommended for PAIR, 7 would have accepted PAIR services had they been covered by Medicare, 4 preferred discharge with home health services, and 6 did not provide clear details of patient or family preference.

To our knowledge, this is the first study to use chart review to examine the proportion of observation patients who would benefit from PAIR and the relationships among these patients’ rehabilitation needs, dispositions, and outcomes. We tried to be conservative in our estimates by limiting the study population to patients admitted from home. Nevertheless, the potential need for PAIR significantly outweighed the actual use of PAIR on discharge. The study sample was consistent with nationally representative samples of observation patients in terms of proportion of patients admitted from and discharged to facilities⁷ and the most common ICD-9 diagnoses.⁹

Physical Therapy Consultations and Observation

Of the 386 patients whose charts were reviewed and analyzed, 17 (4.4%) were evaluated as medically qualifying for and potentially benefiting from PAIR. Although the rate represents a minority of patients, it is 5- to 6-fold higher than the rate of discharge to PAIR, both in our study population and in previous national samples that used administrative data.⁷ In some cases, the decision not to discharge the patient to PAIR reflected patient and family preference. However, in other cases, patients clearly could have benefited from PAIR and would have gone had it been covered by Medicare. The gap suggests an unmet need for PAIR among a substantial proportion of Medicare beneficiaries for whom the therapy is recommended and wanted.

Efforts to expand coverage for PAIR have been resisted. According to Medicare regulations, beneficiaries qualify for PAIR coverage if they are hospitalized as inpatients for 3 midnights or longer. Days under observation status do not count toward this requirement, even if this status is changed to inpatient.¹⁰ The Medicare Payment Advisory Commission (MedPAC) recommendation that time under observation status count toward the Medicare requirement¹¹ has not been accepted,¹² in large part because further expansion of PAIR services likely would be unaffordable to Medicare under its payment structure.¹³ Given our finding that the need for PAIR likely is much higher than previously anticipated, Medicare policy makers should consider broadening access to PAIR while efforts are made to rein in expenditures through payment reform.

One potential area of cost savings is more judicious use of PT evaluation for observation patients, particularly given our finding that the majority of PT consultations resulted in no further recommendations. Efforts to triage PT consultations for appropriateness have had some success, though the literature is scant.¹⁴ To improve value for Medicare, healthcare systems, and patients, researchers should rigorously evaluate approaches that maximize appropriate use of PT services.

Hospital Length of Stay

Our cohort’s mean hospital stay was longer than averages reported elsewhere,⁹ likely reflecting our selection of Medicare patients rather than a general medicine population.⁶ However, our cohort’s adjusted mean hospital stay was significantly longer for patients recommended for PAIR than for patients without PT needs. That out-of-pocket costs for observation patients increase dramatically as LOS goes past 48 hours⁶ could have significant financial implications for Medicare beneficiaries.

Return Visits

Almost 25% of our observation patients returned to hospital within 30 days. There was a significant trend toward increased rehospitalization among patients recommended for PAIR than among patients with no PT needs.

Policies related to PAIR for observation patients are rooted in the concern that expanded access to services will contribute to overuse of services and higher healthcare costs.¹⁵ However, patients who could have benefited from PAIR but were not covered also were at risk for increased healthcare use and costs. A recent study found that more than one fourth of observation patients with repeat observation stays accrued excessive financial liability.¹⁶ Researchers should determine more precisely how the cost of coverage for PAIR placement on an index observation admission compares with the cost of subsequent healthcare use potentially related to insufficient supportive care at home.

Study Limitations

Our results must be interpreted within the context of study limitations. First is the small sample size, particularly the subset of patients selected for detailed manual chart review. We were limited in our ability to calculate sample size prospectively because we were unaware of prior work that described the association between PT recommendation and outcomes among observation patients. However, post hoc analysis estimated that a sample size of 181 patients would have been needed to determine a statistically significant difference in 30-day hospital revisit between patients recommended for PAIR and patients with no PT needs with 80% power, which we achieved. Although there are significant limitations to post hoc sample size estimation, we consider our work hypothesis-generating and hope it will lead to larger studies.

We could not account for the potential bias of the physical therapists, whose evaluations could have been influenced by knowledge of patients’ observation status. Our findings could have underestimated the proportion of patients who otherwise would have been recommended for PAIR. Alternatively, therapists could have inaccurately assessed and overstated the need for PAIR. Although we could not account for the therapists’ accuracy and biases, their assessments provided crucial information beyond what was previously obtained from administrative data alone.^7,9

Hospital revisits were only accounted for within our hospital system—another potential source of underestimated findings. A significant proportion of patients recommended for home PT were discharged without services, which is counterintuitive, as Medicare covers home nursing services for observation patients. This finding most likely reflects administrative error but probably merits further evaluation.

Last, causality cannot be inferred from the results of a retrospective observational study.

As our study results suggest, there is an unmet need for PAIR services for Medicare observation patients, and LOS and subsequent use may be increased among patients recommended for PAIR. Our estimates are conservative and may underestimate the true need for services within this population. Our findings bolster MedPAC recommendations to amend the policies for Medicare coverage of PAIR services for observation patients.

Acknowledgment

The authors thank Paul Kolm, PhD, for statistical support.

Disclosures

Dr. Schwartz reports receiving personal fees from the Agency for Health Research and Quality, Bayer, the Blue Cross Blue Shield Association, Pfizer, and Takeda, all outside the submitted work. Dr. Hicks is supported by an Institutional Development Award from the National Institute of General Medical Sciences of the National Institutes of Health (grant U54-GM104941; principal investigator Stuart Binder-Macleod, PT, PhD, FAPTA). The other authors have nothing to report.

As the US population ages and becomes increasingly frail, the need for rehabilitation rises. By 2030, an estimated 20% of the population will be 65 years old or older, and almost 10% will be over 75.¹ About 20% of hospitalized Medicare patients receive subsequent care in postacute inpatient rehabilitation (PAIR) facilities, accounting for $31 billion in Medicare expenditures in 2014.² Although the need for rehabilitation will continue to rise, Medicare policy restricts access to it.

Under Medicare policy, PAIR services are covered for certain hospitalized patients but not others. Hospitalized patients are either inpatients, who are billed under Medicare Part A, or outpatients, billed under Part B. When hospital length of stay (LOS) is anticipated to be less than 2 midnights, patients are admitted as outpatients under the term observation status; when longer stays are expected, patients are admitted as inpatients.³ This recently implemented time-based distinction has been criticized as arbitrary, and as potentially shifting many patients from inpatient to outpatient (observation) status.⁴

The distinction between inpatient and observation status has significant consequences for posthospital care. Medicare Part A covers care in skilled nursing facilities (SNFs) and acute inpatient rehabilitation facilities (IRFs); after hospitalization, inpatients have access to either, without copay. As observation patients are covered under Medicare Part B, they are technically not covered for either service after their hospital stay. IRFs sometimes accept patients from ambulatory and nonacute settings; observation patients may be accepted in rare circumstances, but they pay the Part A deductible ($1288 in 2016) to have the services covered by Medicare. SNF services are never covered for observation patients, and access to this care requires an average out-of-pocket payment of more than $10,503 per beneficiary for a typical SNF stay.⁵ Given that about 70% of Medicare patients fall below 300% of the federal poverty line,⁶ the out-of-pocket costs for PAIR services for observation patients can be prohibitive.

Although only 0.75% of community-dwelling Medicare observation patients are discharged to PAIR facilities,⁷ it is unclear if the need for this care is higher but remains unmet secondary to cost concerns of Medicare beneficiaries. Also unclear is whether observation patients who would benefit from this care but do not receive it end up with poorer health outcomes and therefore use more healthcare services.

The purpose of this study was to estimate the proportion of Medicare observation patients who are admitted from home and receive a recommendation for placement in a PAIR facility, and to determine the ultimate disposition of such patients. We also sought to evaluate the association between recommendation for PAIR placement, LOS, and 30-day hospital revisit rate.

The Institutional Review Board of Christiana Care Health System (CCHS) approved this study.

Sample and Design

This was an observational study of community-dwelling Medicare patients admitted under observation status to Delaware’s CCHS, which consists of a 907-bed regional tertiary-care facility in Newark and a 241-bed community hospital in Wilmington. The study period was January 1 to December 31, 2013. We limited our sample to patients treated by hospitalists on hospital wards, as this care constitutes 80% of the care provided to observation patients at CCHS and the majority of care nationally.⁸ As neither SNF care nor IRF care is covered under Medicare Part B, and both would result in high out-of-pocket costs for Medicare observation patients, we combined them into a single variable, PAIR.

All data were obtained from institutional electronic medical record and administrative data systems. Study inclusion criteria were Medicare as primary insurance, admission to hospital from home, and care received at either CCHS facility. Exclusion criteria were admission from PAIR facility, long-term care facility, assisted-living facility, or inpatient psychiatric facility; death; discharge against medical advice (AMA) or to hospice, non-SNF, or inpatient psychiatric facility; and discovery (during review of case management [CM] notes) of erroneous listing of Medicare as primary insurance, or of inpatient admission (within 30 days before index observation stay) that qualified for PAIR coverage under Medicare Part A.

We reviewed the medical charts of a representative (~30%) sample of the cohort and examined physical therapy (PT) and CM notes to determine the proportions of patients with recommendations for home with no services, home-based PT, possible PAIR, and PAIR. Charts were sorted by medical record number and were reviewed in consecutive order. We coded a patient as having a recommendation for possible PAIR if the PT notes indicated the patient may benefit from PAIR but could have home PT if PAIR placement was not possible. CM notes were also reviewed for evidence of patient or family preference regarding PAIR placement. All questions about PT and CM recommendations were resolved by consensus.

Measures

For the total study sample, we calculated descriptive statistics and frequencies for demographic and administrative variables, including age, sex, race (Caucasian, African American, other), ethnicity (Hispanic/non-Hispanic), ICD-9 (International Classification of Diseases, Ninth Revision) primary diagnosis code, LOS (in hours) for index observation admission, discharge disposition (home with no services, home PT, possible PAIR, PAIR), and 30-day hospital revisit (emergency department, observation, inpatient admission). We used χ² test, Student t test, and analysis of variance (ANOVA) to test for statistically significant differences in characteristics between the chart review subgroup and the rest of the sample and between the groups with different disposition recommendations from PT notes.

For the chart review subgroup, we used ANOVA to calculate the unadjusted association between PT recommendation and LOS. We then adjusted for potential confounders, using multivariable linear regression with PT recommendation as a predictor and LOS as the outcome, controlling for variables previously associated with increased LOS among observation patients (primary diagnosis category, age, sex).⁶ We also adjusted for hospitalist group to account for potential variability in care delivery. As LOS was not normally distributed, we calculated the fourth root of LOS, which resulted in a more normal distribution, and used the transformed values in the regression model. We then calculated predicted values from the regression and back-transformed these to obtain adjusted mean values for LOS.

Of the 1417 unique patients who had Medicare as primary insurance and were admitted under observation status to a hospitalist service during the study period (2013), 94 were excluded (Figure). Of the remaining 1323 patients, the majority were 65 years old or older, female, white, and non-Hispanic. The most common ICD-9 diagnoses were syncope and chest pain. Mean LOS was 46.7 hours (range, 0-519 h). Less than 1% of patients were discharged to PAIR. Almost 25% of patients returned to the hospital, either for an emergency department visit or for observation or inpatient stay, within 30 days (Table).

Of the 419 charts reviewed to determine the proportion of patients evaluated by PT, and their subsequent recommendations, 33 were excluded, leaving 386 (92%) for analysis (Figure). There were no significant demographic differences between the patients in the chart review subgroup and the rest of the patients (Appendix). Of the 386 patients whose charts were analyzed, 181 (46.9%) had a PT evaluation, and 17 (4.4%) received a PAIR recommendation (Figure). Of the 17 patients recommended for PAIR, 12 (70.5%) were 65 years old or older, and 1 was discharged to a PAIR facility. Of the 46 patients recommended for home PT, 29 (63%) were discharged home with no services (Table).

PT-evaluated patients had unadjusted mean LOS of 52.2 hours (discharged home with no services), 64.1 hours (home PT or possible PAIR), and 83.1 hours (PAIR) (P = 0.001). With adjustment made for variables previously associated with increased LOS for observation patients, mean LOS for patients recommended for PAIR remained higher than that for patients in the other 2 categories (Table). Patients recommended for PAIR were more likely to return to hospital within 30 days than patients recommended for home PT or possible PAIR and patients discharged home with no services (Table).

Review of CM notes revealed that, of the 17 patients recommended for PAIR, 7 would have accepted PAIR services had they been covered by Medicare, 4 preferred discharge with home health services, and 6 did not provide clear details of patient or family preference.

To our knowledge, this is the first study to use chart review to examine the proportion of observation patients who would benefit from PAIR and the relationships among these patients’ rehabilitation needs, dispositions, and outcomes. We tried to be conservative in our estimates by limiting the study population to patients admitted from home. Nevertheless, the potential need for PAIR significantly outweighed the actual use of PAIR on discharge. The study sample was consistent with nationally representative samples of observation patients in terms of proportion of patients admitted from and discharged to facilities⁷ and the most common ICD-9 diagnoses.⁹

Physical Therapy Consultations and Observation

Of the 386 patients whose charts were reviewed and analyzed, 17 (4.4%) were evaluated as medically qualifying for and potentially benefiting from PAIR. Although the rate represents a minority of patients, it is 5- to 6-fold higher than the rate of discharge to PAIR, both in our study population and in previous national samples that used administrative data.⁷ In some cases, the decision not to discharge the patient to PAIR reflected patient and family preference. However, in other cases, patients clearly could have benefited from PAIR and would have gone had it been covered by Medicare. The gap suggests an unmet need for PAIR among a substantial proportion of Medicare beneficiaries for whom the therapy is recommended and wanted.

Efforts to expand coverage for PAIR have been resisted. According to Medicare regulations, beneficiaries qualify for PAIR coverage if they are hospitalized as inpatients for 3 midnights or longer. Days under observation status do not count toward this requirement, even if this status is changed to inpatient.¹⁰ The Medicare Payment Advisory Commission (MedPAC) recommendation that time under observation status count toward the Medicare requirement¹¹ has not been accepted,¹² in large part because further expansion of PAIR services likely would be unaffordable to Medicare under its payment structure.¹³ Given our finding that the need for PAIR likely is much higher than previously anticipated, Medicare policy makers should consider broadening access to PAIR while efforts are made to rein in expenditures through payment reform.

One potential area of cost savings is more judicious use of PT evaluation for observation patients, particularly given our finding that the majority of PT consultations resulted in no further recommendations. Efforts to triage PT consultations for appropriateness have had some success, though the literature is scant.¹⁴ To improve value for Medicare, healthcare systems, and patients, researchers should rigorously evaluate approaches that maximize appropriate use of PT services.

Hospital Length of Stay

Our cohort’s mean hospital stay was longer than averages reported elsewhere,⁹ likely reflecting our selection of Medicare patients rather than a general medicine population.⁶ However, our cohort’s adjusted mean hospital stay was significantly longer for patients recommended for PAIR than for patients without PT needs. That out-of-pocket costs for observation patients increase dramatically as LOS goes past 48 hours⁶ could have significant financial implications for Medicare beneficiaries.

Return Visits

Almost 25% of our observation patients returned to hospital within 30 days. There was a significant trend toward increased rehospitalization among patients recommended for PAIR than among patients with no PT needs.

Policies related to PAIR for observation patients are rooted in the concern that expanded access to services will contribute to overuse of services and higher healthcare costs.¹⁵ However, patients who could have benefited from PAIR but were not covered also were at risk for increased healthcare use and costs. A recent study found that more than one fourth of observation patients with repeat observation stays accrued excessive financial liability.¹⁶ Researchers should determine more precisely how the cost of coverage for PAIR placement on an index observation admission compares with the cost of subsequent healthcare use potentially related to insufficient supportive care at home.

Study Limitations

Our results must be interpreted within the context of study limitations. First is the small sample size, particularly the subset of patients selected for detailed manual chart review. We were limited in our ability to calculate sample size prospectively because we were unaware of prior work that described the association between PT recommendation and outcomes among observation patients. However, post hoc analysis estimated that a sample size of 181 patients would have been needed to determine a statistically significant difference in 30-day hospital revisit between patients recommended for PAIR and patients with no PT needs with 80% power, which we achieved. Although there are significant limitations to post hoc sample size estimation, we consider our work hypothesis-generating and hope it will lead to larger studies.

We could not account for the potential bias of the physical therapists, whose evaluations could have been influenced by knowledge of patients’ observation status. Our findings could have underestimated the proportion of patients who otherwise would have been recommended for PAIR. Alternatively, therapists could have inaccurately assessed and overstated the need for PAIR. Although we could not account for the therapists’ accuracy and biases, their assessments provided crucial information beyond what was previously obtained from administrative data alone.^7,9

Hospital revisits were only accounted for within our hospital system—another potential source of underestimated findings. A significant proportion of patients recommended for home PT were discharged without services, which is counterintuitive, as Medicare covers home nursing services for observation patients. This finding most likely reflects administrative error but probably merits further evaluation.

Last, causality cannot be inferred from the results of a retrospective observational study.

As our study results suggest, there is an unmet need for PAIR services for Medicare observation patients, and LOS and subsequent use may be increased among patients recommended for PAIR. Our estimates are conservative and may underestimate the true need for services within this population. Our findings bolster MedPAC recommendations to amend the policies for Medicare coverage of PAIR services for observation patients.

Acknowledgment

The authors thank Paul Kolm, PhD, for statistical support.

Disclosures

Dr. Schwartz reports receiving personal fees from the Agency for Health Research and Quality, Bayer, the Blue Cross Blue Shield Association, Pfizer, and Takeda, all outside the submitted work. Dr. Hicks is supported by an Institutional Development Award from the National Institute of General Medical Sciences of the National Institutes of Health (grant U54-GM104941; principal investigator Stuart Binder-Macleod, PT, PhD, FAPTA). The other authors have nothing to report.

As the US population ages and becomes increasingly frail, the need for rehabilitation rises. By 2030, an estimated 20% of the population will be 65 years old or older, and almost 10% will be over 75.¹ About 20% of hospitalized Medicare patients receive subsequent care in postacute inpatient rehabilitation (PAIR) facilities, accounting for $31 billion in Medicare expenditures in 2014.² Although the need for rehabilitation will continue to rise, Medicare policy restricts access to it.

Under Medicare policy, PAIR services are covered for certain hospitalized patients but not others. Hospitalized patients are either inpatients, who are billed under Medicare Part A, or outpatients, billed under Part B. When hospital length of stay (LOS) is anticipated to be less than 2 midnights, patients are admitted as outpatients under the term observation status; when longer stays are expected, patients are admitted as inpatients.³ This recently implemented time-based distinction has been criticized as arbitrary, and as potentially shifting many patients from inpatient to outpatient (observation) status.⁴

The distinction between inpatient and observation status has significant consequences for posthospital care. Medicare Part A covers care in skilled nursing facilities (SNFs) and acute inpatient rehabilitation facilities (IRFs); after hospitalization, inpatients have access to either, without copay. As observation patients are covered under Medicare Part B, they are technically not covered for either service after their hospital stay. IRFs sometimes accept patients from ambulatory and nonacute settings; observation patients may be accepted in rare circumstances, but they pay the Part A deductible ($1288 in 2016) to have the services covered by Medicare. SNF services are never covered for observation patients, and access to this care requires an average out-of-pocket payment of more than $10,503 per beneficiary for a typical SNF stay.⁵ Given that about 70% of Medicare patients fall below 300% of the federal poverty line,⁶ the out-of-pocket costs for PAIR services for observation patients can be prohibitive.

Although only 0.75% of community-dwelling Medicare observation patients are discharged to PAIR facilities,⁷ it is unclear if the need for this care is higher but remains unmet secondary to cost concerns of Medicare beneficiaries. Also unclear is whether observation patients who would benefit from this care but do not receive it end up with poorer health outcomes and therefore use more healthcare services.

The purpose of this study was to estimate the proportion of Medicare observation patients who are admitted from home and receive a recommendation for placement in a PAIR facility, and to determine the ultimate disposition of such patients. We also sought to evaluate the association between recommendation for PAIR placement, LOS, and 30-day hospital revisit rate.

The Institutional Review Board of Christiana Care Health System (CCHS) approved this study.

Sample and Design

This was an observational study of community-dwelling Medicare patients admitted under observation status to Delaware’s CCHS, which consists of a 907-bed regional tertiary-care facility in Newark and a 241-bed community hospital in Wilmington. The study period was January 1 to December 31, 2013. We limited our sample to patients treated by hospitalists on hospital wards, as this care constitutes 80% of the care provided to observation patients at CCHS and the majority of care nationally.⁸ As neither SNF care nor IRF care is covered under Medicare Part B, and both would result in high out-of-pocket costs for Medicare observation patients, we combined them into a single variable, PAIR.

All data were obtained from institutional electronic medical record and administrative data systems. Study inclusion criteria were Medicare as primary insurance, admission to hospital from home, and care received at either CCHS facility. Exclusion criteria were admission from PAIR facility, long-term care facility, assisted-living facility, or inpatient psychiatric facility; death; discharge against medical advice (AMA) or to hospice, non-SNF, or inpatient psychiatric facility; and discovery (during review of case management [CM] notes) of erroneous listing of Medicare as primary insurance, or of inpatient admission (within 30 days before index observation stay) that qualified for PAIR coverage under Medicare Part A.

We reviewed the medical charts of a representative (~30%) sample of the cohort and examined physical therapy (PT) and CM notes to determine the proportions of patients with recommendations for home with no services, home-based PT, possible PAIR, and PAIR. Charts were sorted by medical record number and were reviewed in consecutive order. We coded a patient as having a recommendation for possible PAIR if the PT notes indicated the patient may benefit from PAIR but could have home PT if PAIR placement was not possible. CM notes were also reviewed for evidence of patient or family preference regarding PAIR placement. All questions about PT and CM recommendations were resolved by consensus.

Measures

For the total study sample, we calculated descriptive statistics and frequencies for demographic and administrative variables, including age, sex, race (Caucasian, African American, other), ethnicity (Hispanic/non-Hispanic), ICD-9 (International Classification of Diseases, Ninth Revision) primary diagnosis code, LOS (in hours) for index observation admission, discharge disposition (home with no services, home PT, possible PAIR, PAIR), and 30-day hospital revisit (emergency department, observation, inpatient admission). We used χ² test, Student t test, and analysis of variance (ANOVA) to test for statistically significant differences in characteristics between the chart review subgroup and the rest of the sample and between the groups with different disposition recommendations from PT notes.

For the chart review subgroup, we used ANOVA to calculate the unadjusted association between PT recommendation and LOS. We then adjusted for potential confounders, using multivariable linear regression with PT recommendation as a predictor and LOS as the outcome, controlling for variables previously associated with increased LOS among observation patients (primary diagnosis category, age, sex).⁶ We also adjusted for hospitalist group to account for potential variability in care delivery. As LOS was not normally distributed, we calculated the fourth root of LOS, which resulted in a more normal distribution, and used the transformed values in the regression model. We then calculated predicted values from the regression and back-transformed these to obtain adjusted mean values for LOS.

Of the 1417 unique patients who had Medicare as primary insurance and were admitted under observation status to a hospitalist service during the study period (2013), 94 were excluded (Figure). Of the remaining 1323 patients, the majority were 65 years old or older, female, white, and non-Hispanic. The most common ICD-9 diagnoses were syncope and chest pain. Mean LOS was 46.7 hours (range, 0-519 h). Less than 1% of patients were discharged to PAIR. Almost 25% of patients returned to the hospital, either for an emergency department visit or for observation or inpatient stay, within 30 days (Table).

Of the 419 charts reviewed to determine the proportion of patients evaluated by PT, and their subsequent recommendations, 33 were excluded, leaving 386 (92%) for analysis (Figure). There were no significant demographic differences between the patients in the chart review subgroup and the rest of the patients (Appendix). Of the 386 patients whose charts were analyzed, 181 (46.9%) had a PT evaluation, and 17 (4.4%) received a PAIR recommendation (Figure). Of the 17 patients recommended for PAIR, 12 (70.5%) were 65 years old or older, and 1 was discharged to a PAIR facility. Of the 46 patients recommended for home PT, 29 (63%) were discharged home with no services (Table).

PT-evaluated patients had unadjusted mean LOS of 52.2 hours (discharged home with no services), 64.1 hours (home PT or possible PAIR), and 83.1 hours (PAIR) (P = 0.001). With adjustment made for variables previously associated with increased LOS for observation patients, mean LOS for patients recommended for PAIR remained higher than that for patients in the other 2 categories (Table). Patients recommended for PAIR were more likely to return to hospital within 30 days than patients recommended for home PT or possible PAIR and patients discharged home with no services (Table).

Review of CM notes revealed that, of the 17 patients recommended for PAIR, 7 would have accepted PAIR services had they been covered by Medicare, 4 preferred discharge with home health services, and 6 did not provide clear details of patient or family preference.

To our knowledge, this is the first study to use chart review to examine the proportion of observation patients who would benefit from PAIR and the relationships among these patients’ rehabilitation needs, dispositions, and outcomes. We tried to be conservative in our estimates by limiting the study population to patients admitted from home. Nevertheless, the potential need for PAIR significantly outweighed the actual use of PAIR on discharge. The study sample was consistent with nationally representative samples of observation patients in terms of proportion of patients admitted from and discharged to facilities⁷ and the most common ICD-9 diagnoses.⁹

Physical Therapy Consultations and Observation

Of the 386 patients whose charts were reviewed and analyzed, 17 (4.4%) were evaluated as medically qualifying for and potentially benefiting from PAIR. Although the rate represents a minority of patients, it is 5- to 6-fold higher than the rate of discharge to PAIR, both in our study population and in previous national samples that used administrative data.⁷ In some cases, the decision not to discharge the patient to PAIR reflected patient and family preference. However, in other cases, patients clearly could have benefited from PAIR and would have gone had it been covered by Medicare. The gap suggests an unmet need for PAIR among a substantial proportion of Medicare beneficiaries for whom the therapy is recommended and wanted.

Efforts to expand coverage for PAIR have been resisted. According to Medicare regulations, beneficiaries qualify for PAIR coverage if they are hospitalized as inpatients for 3 midnights or longer. Days under observation status do not count toward this requirement, even if this status is changed to inpatient.¹⁰ The Medicare Payment Advisory Commission (MedPAC) recommendation that time under observation status count toward the Medicare requirement¹¹ has not been accepted,¹² in large part because further expansion of PAIR services likely would be unaffordable to Medicare under its payment structure.¹³ Given our finding that the need for PAIR likely is much higher than previously anticipated, Medicare policy makers should consider broadening access to PAIR while efforts are made to rein in expenditures through payment reform.

One potential area of cost savings is more judicious use of PT evaluation for observation patients, particularly given our finding that the majority of PT consultations resulted in no further recommendations. Efforts to triage PT consultations for appropriateness have had some success, though the literature is scant.¹⁴ To improve value for Medicare, healthcare systems, and patients, researchers should rigorously evaluate approaches that maximize appropriate use of PT services.

Hospital Length of Stay

Our cohort’s mean hospital stay was longer than averages reported elsewhere,⁹ likely reflecting our selection of Medicare patients rather than a general medicine population.⁶ However, our cohort’s adjusted mean hospital stay was significantly longer for patients recommended for PAIR than for patients without PT needs. That out-of-pocket costs for observation patients increase dramatically as LOS goes past 48 hours⁶ could have significant financial implications for Medicare beneficiaries.

Return Visits

Almost 25% of our observation patients returned to hospital within 30 days. There was a significant trend toward increased rehospitalization among patients recommended for PAIR than among patients with no PT needs.

Policies related to PAIR for observation patients are rooted in the concern that expanded access to services will contribute to overuse of services and higher healthcare costs.¹⁵ However, patients who could have benefited from PAIR but were not covered also were at risk for increased healthcare use and costs. A recent study found that more than one fourth of observation patients with repeat observation stays accrued excessive financial liability.¹⁶ Researchers should determine more precisely how the cost of coverage for PAIR placement on an index observation admission compares with the cost of subsequent healthcare use potentially related to insufficient supportive care at home.

Study Limitations

Our results must be interpreted within the context of study limitations. First is the small sample size, particularly the subset of patients selected for detailed manual chart review. We were limited in our ability to calculate sample size prospectively because we were unaware of prior work that described the association between PT recommendation and outcomes among observation patients. However, post hoc analysis estimated that a sample size of 181 patients would have been needed to determine a statistically significant difference in 30-day hospital revisit between patients recommended for PAIR and patients with no PT needs with 80% power, which we achieved. Although there are significant limitations to post hoc sample size estimation, we consider our work hypothesis-generating and hope it will lead to larger studies.

We could not account for the potential bias of the physical therapists, whose evaluations could have been influenced by knowledge of patients’ observation status. Our findings could have underestimated the proportion of patients who otherwise would have been recommended for PAIR. Alternatively, therapists could have inaccurately assessed and overstated the need for PAIR. Although we could not account for the therapists’ accuracy and biases, their assessments provided crucial information beyond what was previously obtained from administrative data alone.^7,9

Hospital revisits were only accounted for within our hospital system—another potential source of underestimated findings. A significant proportion of patients recommended for home PT were discharged without services, which is counterintuitive, as Medicare covers home nursing services for observation patients. This finding most likely reflects administrative error but probably merits further evaluation.

Last, causality cannot be inferred from the results of a retrospective observational study.

As our study results suggest, there is an unmet need for PAIR services for Medicare observation patients, and LOS and subsequent use may be increased among patients recommended for PAIR. Our estimates are conservative and may underestimate the true need for services within this population. Our findings bolster MedPAC recommendations to amend the policies for Medicare coverage of PAIR services for observation patients.

Acknowledgment

The authors thank Paul Kolm, PhD, for statistical support.

Disclosures

Dr. Schwartz reports receiving personal fees from the Agency for Health Research and Quality, Bayer, the Blue Cross Blue Shield Association, Pfizer, and Takeda, all outside the submitted work. Dr. Hicks is supported by an Institutional Development Award from the National Institute of General Medical Sciences of the National Institutes of Health (grant U54-GM104941; principal investigator Stuart Binder-Macleod, PT, PhD, FAPTA). The other authors have nothing to report.

The field of hospital medicine (HM) is rapidly expanding in the areas of clinical medicine, administration, and quality improvement (QI).¹ Emerging with this growth is a gap in the traditional internal medicine (IM) training and skills needed to be effective in HM.^1,2 These skills include clinical and nonclinical aptitudes, such as process improvement, health care economics, and leadership.^1-3 However, resident education on these topics must compete with other required curricular content in IM residency training.^2,4 Few IM residencies offer focused HM training that emphasizes key components of successful HM careers.^3,5

Within the past decade, designated HM tracks within IM residency programs have been proposed as a potential solution. Initially, calls for such tracks focused on gaps in the clinical competencies required of hospitalists.¹ Tracks have since evolved to also include skills required to drive high-value care, process improvement, and scholarship. Designated HM tracks address these areas through greater breadth of curricula, additional time for reflection, participation in group projects, and active application to clinical care.⁴ We conducted a study to identify themes that could inform the ongoing evolution of dedicated HM tracks.

Programs were initially identified through communication among professional networks. The phrases hospital medicine residency track and internal medicine residency hospitalist track were used in broader Google searches, as there is no database of such tracks. Searches were performed quarterly during the 2015–2016 academic year. The top 20 hits were manually filtered to identify tracks affiliated with major academic centers. IM residency program websites provided basic information for programs with tracks. We excluded tracks focused entirely on QI⁶ because, though a crucial part of HM, QI training alone is probably insufficient for preparing residents for success as hospitalists on residency completion. Similarly, IM residencies with stand-alone HM clinical rotations without longitudinal HM curricula were excluded.

Semistructured interviews with track directors were conducted by e-mail or telephone for all tracks except one, the details of which are published.⁷ We tabulated data and reviewed qualitative information to identify themes among the different tracks. As this study did not involve human participants, Institutional Review Board approval was not needed.

We identified 11 HM residency training programs at major academic centers across the United States: Cleveland Clinic, Stanford University, Tulane University, University of California Davis, University of California Irvine, University of Colorado, University of Kentucky, University of Minnesota, University of New Mexico, Virginia Commonwealth University, and Wake Forest University (Table 1). We reviewed the websites of about 10 other programs, but none suggested existence of a track. Additional programs contacted reported no current track.

Track Participants and Structure

HM tracks mainly target third-year residents (Table 1). Some extend into the second year of residency, and 4 have opportunities for intern involvement, including a separate match number at Colorado. Tracks accept up to 12 residents per class. Two programs, at Colorado and Virginia, are part of IM programs in which all residents belong to a track (eg, HM, primary care, research).

HM track structures vary widely and are heavily influenced by the content delivery platforms of their IM residency programs. Several HM track directors emphasized the importance of fitting into existing educational frameworks to ensure access to residents and to minimize the burden of participation. Four programs deliver the bulk of their nonclinical content in dedicated blocks; 6 others use brief recurring sessions to deliver smaller aliquots longitudinally (Table 1). The number of protected hours for content delivery ranges from 10 to more than 40 annually. All tracks use multiple content delivery modes, including didactic sessions and journal clubs. Four tracks employ panel discussions to explore career options within HM. Several also use online platforms, including discussions, readings, and modules.

Quality Improvement

The vast majority of curricula prominently feature experiential QI project involvement (Table 2). These mentored longitudinal projects allow applied delivery of content, such as QI methods and management skills. Four tracks use material from the Institute for Healthcare Improvement.⁸ Several also offer dedicated QI rotations that immerse residents in ongoing QI efforts.

Institutional partnerships support these initiatives at several sites. The Minnesota track is a joint venture of the university and Regions Hospital, a nonprofit community hospital. The Virginia track positions HM residents to lead university-wide interdisciplinary QI teams. For project support, the Colorado and Kentucky tracks partner with local QI resources—the Institute for Healthcare Quality, Safety, and Efficiency at Colorado and the Office of Value and Innovation in Healthcare Delivery at Kentucky.

Health Care Economics and Value

Many programs leverage the rapidly growing emphasis on health care “value” as an opportunity for synergy between IM programs and HM tracks. Examples include involving residents in efforts to improve documentation or didactic instruction on topics such as health care finance. The New Mexico and Wake Forest tracks offer elective rotations on health care economics. Several track directors mentioned successfully expanding curricula on health care value from the HM track into IM residency programs at large, providing a measurable service to the residency programs while ensuring content delivery and freeing up additional time for track activities.

Scholarship and Career Development

Most programs provide targeted career development for residents. Six tracks provide sessions on job procurement skills, such as curriculum vitae preparation and interviewing (Table 2). Many also provide content on venues for disseminating scholarly activity. The Colorado, Kentucky, New Mexico, and Tulane programs feature content on abstract and poster creation. Leadership development is addressed in several tracks through dedicated track activities or participation in discrete, outside-track events. Specifically, Colorado offers a leadership track for residents interested in hospital administration, Cleveland has a leadership journal club, Wake Forest enrolls HM residents in leadership training available through the university, and Minnesota sends residents to the Society of Hospital Medicine’s Leadership Academy (Table 2).

Clinical Rotations

Almost all tracks include a clinical rotation, typically pairing residents directly with hospitalist attendings to encourage autonomy and mentorship. Several also offer elective rotations in various disciplines within HM (Table 2). The Kentucky and Virginia tracks incorporate working with advanced practice providers into their practicums. The Cleveland, Minnesota, Tulane, and Virginia tracks offer HM rotations in community hospitals or postacute settings.

HM rotations also pair clinical experiences with didactic education on relevant topics (eg, billing and coding). The Cleveland, Minnesota, and Virginia tracks developed clinical rotations reflecting the common 7-on and 7-off schedule with nonclinical obligations, such as seminars linking specific content to clinical experiences, during nonclinical time.

Our investigation into the current state of HM training found that HM track curricula focus largely on QI, health care economics, and professional development. This focus likely developed in response to hospitalists’ increasing engagement in related endeavors. HM tracks have dynamic and variable structures, reflecting an evolving field and the need to fit into existing IM residency program structures. Similarly, the content covered in HM tracks is tightly linked to perceived opportunities within IM residency curricula. The heterogeneity of content suggests the breadth and ambiguity of necessary competencies for aspiring hospitalists. One of the 11 tracks has not had any residents enroll within the past few years—a testament to the continued effort necessary to sustain such tracks, including curricular updates and recruiting. Conversely, many programs now share track content with the larger IM residency program, suggesting HM tracks may be near the forefront of medical education in some areas.

Our study had several limitations. As we are unaware of any databases of HM tracks, we discussed tracks with professional contacts, performed Internet searches, and reviewed IM residency program websites. Our search, however, was not exhaustive; despite our best efforts, we may have missed or mischaracterized some track offerings. Nevertheless, we think that our analysis represents the first thorough compilation of HM tracks and that it will be useful to institutions seeking to create or enhance HM-specific training.

As the field continues to evolve, we are optimistic about the future of HM training. We suspect that HM residency training tracks will continue to expand. More work is needed so these tracks can adjust to the changing HM and IM residency program landscapes and supply well-trained physicians for the HM workforce.

Acknowledgment

The authors thank track directors Alpesh Amin, David Gugliotti, Rick Hilger, Karnjit Johl, Nasir Majeed, Georgia McIntosh, Charles Pizanis, and Jeff Wiese for making this study possible.

Disclosure

Nothing to report.

The field of hospital medicine (HM) is rapidly expanding in the areas of clinical medicine, administration, and quality improvement (QI).¹ Emerging with this growth is a gap in the traditional internal medicine (IM) training and skills needed to be effective in HM.^1,2 These skills include clinical and nonclinical aptitudes, such as process improvement, health care economics, and leadership.^1-3 However, resident education on these topics must compete with other required curricular content in IM residency training.^2,4 Few IM residencies offer focused HM training that emphasizes key components of successful HM careers.^3,5

Within the past decade, designated HM tracks within IM residency programs have been proposed as a potential solution. Initially, calls for such tracks focused on gaps in the clinical competencies required of hospitalists.¹ Tracks have since evolved to also include skills required to drive high-value care, process improvement, and scholarship. Designated HM tracks address these areas through greater breadth of curricula, additional time for reflection, participation in group projects, and active application to clinical care.⁴ We conducted a study to identify themes that could inform the ongoing evolution of dedicated HM tracks.

Programs were initially identified through communication among professional networks. The phrases hospital medicine residency track and internal medicine residency hospitalist track were used in broader Google searches, as there is no database of such tracks. Searches were performed quarterly during the 2015–2016 academic year. The top 20 hits were manually filtered to identify tracks affiliated with major academic centers. IM residency program websites provided basic information for programs with tracks. We excluded tracks focused entirely on QI⁶ because, though a crucial part of HM, QI training alone is probably insufficient for preparing residents for success as hospitalists on residency completion. Similarly, IM residencies with stand-alone HM clinical rotations without longitudinal HM curricula were excluded.

Semistructured interviews with track directors were conducted by e-mail or telephone for all tracks except one, the details of which are published.⁷ We tabulated data and reviewed qualitative information to identify themes among the different tracks. As this study did not involve human participants, Institutional Review Board approval was not needed.

We identified 11 HM residency training programs at major academic centers across the United States: Cleveland Clinic, Stanford University, Tulane University, University of California Davis, University of California Irvine, University of Colorado, University of Kentucky, University of Minnesota, University of New Mexico, Virginia Commonwealth University, and Wake Forest University (Table 1). We reviewed the websites of about 10 other programs, but none suggested existence of a track. Additional programs contacted reported no current track.

Track Participants and Structure

HM tracks mainly target third-year residents (Table 1). Some extend into the second year of residency, and 4 have opportunities for intern involvement, including a separate match number at Colorado. Tracks accept up to 12 residents per class. Two programs, at Colorado and Virginia, are part of IM programs in which all residents belong to a track (eg, HM, primary care, research).

HM track structures vary widely and are heavily influenced by the content delivery platforms of their IM residency programs. Several HM track directors emphasized the importance of fitting into existing educational frameworks to ensure access to residents and to minimize the burden of participation. Four programs deliver the bulk of their nonclinical content in dedicated blocks; 6 others use brief recurring sessions to deliver smaller aliquots longitudinally (Table 1). The number of protected hours for content delivery ranges from 10 to more than 40 annually. All tracks use multiple content delivery modes, including didactic sessions and journal clubs. Four tracks employ panel discussions to explore career options within HM. Several also use online platforms, including discussions, readings, and modules.

Quality Improvement

The vast majority of curricula prominently feature experiential QI project involvement (Table 2). These mentored longitudinal projects allow applied delivery of content, such as QI methods and management skills. Four tracks use material from the Institute for Healthcare Improvement.⁸ Several also offer dedicated QI rotations that immerse residents in ongoing QI efforts.

Institutional partnerships support these initiatives at several sites. The Minnesota track is a joint venture of the university and Regions Hospital, a nonprofit community hospital. The Virginia track positions HM residents to lead university-wide interdisciplinary QI teams. For project support, the Colorado and Kentucky tracks partner with local QI resources—the Institute for Healthcare Quality, Safety, and Efficiency at Colorado and the Office of Value and Innovation in Healthcare Delivery at Kentucky.

Health Care Economics and Value

Many programs leverage the rapidly growing emphasis on health care “value” as an opportunity for synergy between IM programs and HM tracks. Examples include involving residents in efforts to improve documentation or didactic instruction on topics such as health care finance. The New Mexico and Wake Forest tracks offer elective rotations on health care economics. Several track directors mentioned successfully expanding curricula on health care value from the HM track into IM residency programs at large, providing a measurable service to the residency programs while ensuring content delivery and freeing up additional time for track activities.

Scholarship and Career Development

Most programs provide targeted career development for residents. Six tracks provide sessions on job procurement skills, such as curriculum vitae preparation and interviewing (Table 2). Many also provide content on venues for disseminating scholarly activity. The Colorado, Kentucky, New Mexico, and Tulane programs feature content on abstract and poster creation. Leadership development is addressed in several tracks through dedicated track activities or participation in discrete, outside-track events. Specifically, Colorado offers a leadership track for residents interested in hospital administration, Cleveland has a leadership journal club, Wake Forest enrolls HM residents in leadership training available through the university, and Minnesota sends residents to the Society of Hospital Medicine’s Leadership Academy (Table 2).

Clinical Rotations

Almost all tracks include a clinical rotation, typically pairing residents directly with hospitalist attendings to encourage autonomy and mentorship. Several also offer elective rotations in various disciplines within HM (Table 2). The Kentucky and Virginia tracks incorporate working with advanced practice providers into their practicums. The Cleveland, Minnesota, Tulane, and Virginia tracks offer HM rotations in community hospitals or postacute settings.

HM rotations also pair clinical experiences with didactic education on relevant topics (eg, billing and coding). The Cleveland, Minnesota, and Virginia tracks developed clinical rotations reflecting the common 7-on and 7-off schedule with nonclinical obligations, such as seminars linking specific content to clinical experiences, during nonclinical time.

Our investigation into the current state of HM training found that HM track curricula focus largely on QI, health care economics, and professional development. This focus likely developed in response to hospitalists’ increasing engagement in related endeavors. HM tracks have dynamic and variable structures, reflecting an evolving field and the need to fit into existing IM residency program structures. Similarly, the content covered in HM tracks is tightly linked to perceived opportunities within IM residency curricula. The heterogeneity of content suggests the breadth and ambiguity of necessary competencies for aspiring hospitalists. One of the 11 tracks has not had any residents enroll within the past few years—a testament to the continued effort necessary to sustain such tracks, including curricular updates and recruiting. Conversely, many programs now share track content with the larger IM residency program, suggesting HM tracks may be near the forefront of medical education in some areas.

Our study had several limitations. As we are unaware of any databases of HM tracks, we discussed tracks with professional contacts, performed Internet searches, and reviewed IM residency program websites. Our search, however, was not exhaustive; despite our best efforts, we may have missed or mischaracterized some track offerings. Nevertheless, we think that our analysis represents the first thorough compilation of HM tracks and that it will be useful to institutions seeking to create or enhance HM-specific training.

As the field continues to evolve, we are optimistic about the future of HM training. We suspect that HM residency training tracks will continue to expand. More work is needed so these tracks can adjust to the changing HM and IM residency program landscapes and supply well-trained physicians for the HM workforce.

Acknowledgment

The authors thank track directors Alpesh Amin, David Gugliotti, Rick Hilger, Karnjit Johl, Nasir Majeed, Georgia McIntosh, Charles Pizanis, and Jeff Wiese for making this study possible.

Disclosure

Nothing to report.

The field of hospital medicine (HM) is rapidly expanding in the areas of clinical medicine, administration, and quality improvement (QI).¹ Emerging with this growth is a gap in the traditional internal medicine (IM) training and skills needed to be effective in HM.^1,2 These skills include clinical and nonclinical aptitudes, such as process improvement, health care economics, and leadership.^1-3 However, resident education on these topics must compete with other required curricular content in IM residency training.^2,4 Few IM residencies offer focused HM training that emphasizes key components of successful HM careers.^3,5

Within the past decade, designated HM tracks within IM residency programs have been proposed as a potential solution. Initially, calls for such tracks focused on gaps in the clinical competencies required of hospitalists.¹ Tracks have since evolved to also include skills required to drive high-value care, process improvement, and scholarship. Designated HM tracks address these areas through greater breadth of curricula, additional time for reflection, participation in group projects, and active application to clinical care.⁴ We conducted a study to identify themes that could inform the ongoing evolution of dedicated HM tracks.

Programs were initially identified through communication among professional networks. The phrases hospital medicine residency track and internal medicine residency hospitalist track were used in broader Google searches, as there is no database of such tracks. Searches were performed quarterly during the 2015–2016 academic year. The top 20 hits were manually filtered to identify tracks affiliated with major academic centers. IM residency program websites provided basic information for programs with tracks. We excluded tracks focused entirely on QI⁶ because, though a crucial part of HM, QI training alone is probably insufficient for preparing residents for success as hospitalists on residency completion. Similarly, IM residencies with stand-alone HM clinical rotations without longitudinal HM curricula were excluded.

Semistructured interviews with track directors were conducted by e-mail or telephone for all tracks except one, the details of which are published.⁷ We tabulated data and reviewed qualitative information to identify themes among the different tracks. As this study did not involve human participants, Institutional Review Board approval was not needed.

We identified 11 HM residency training programs at major academic centers across the United States: Cleveland Clinic, Stanford University, Tulane University, University of California Davis, University of California Irvine, University of Colorado, University of Kentucky, University of Minnesota, University of New Mexico, Virginia Commonwealth University, and Wake Forest University (Table 1). We reviewed the websites of about 10 other programs, but none suggested existence of a track. Additional programs contacted reported no current track.

Track Participants and Structure

HM tracks mainly target third-year residents (Table 1). Some extend into the second year of residency, and 4 have opportunities for intern involvement, including a separate match number at Colorado. Tracks accept up to 12 residents per class. Two programs, at Colorado and Virginia, are part of IM programs in which all residents belong to a track (eg, HM, primary care, research).

HM track structures vary widely and are heavily influenced by the content delivery platforms of their IM residency programs. Several HM track directors emphasized the importance of fitting into existing educational frameworks to ensure access to residents and to minimize the burden of participation. Four programs deliver the bulk of their nonclinical content in dedicated blocks; 6 others use brief recurring sessions to deliver smaller aliquots longitudinally (Table 1). The number of protected hours for content delivery ranges from 10 to more than 40 annually. All tracks use multiple content delivery modes, including didactic sessions and journal clubs. Four tracks employ panel discussions to explore career options within HM. Several also use online platforms, including discussions, readings, and modules.

Quality Improvement

The vast majority of curricula prominently feature experiential QI project involvement (Table 2). These mentored longitudinal projects allow applied delivery of content, such as QI methods and management skills. Four tracks use material from the Institute for Healthcare Improvement.⁸ Several also offer dedicated QI rotations that immerse residents in ongoing QI efforts.

Institutional partnerships support these initiatives at several sites. The Minnesota track is a joint venture of the university and Regions Hospital, a nonprofit community hospital. The Virginia track positions HM residents to lead university-wide interdisciplinary QI teams. For project support, the Colorado and Kentucky tracks partner with local QI resources—the Institute for Healthcare Quality, Safety, and Efficiency at Colorado and the Office of Value and Innovation in Healthcare Delivery at Kentucky.

Health Care Economics and Value

Many programs leverage the rapidly growing emphasis on health care “value” as an opportunity for synergy between IM programs and HM tracks. Examples include involving residents in efforts to improve documentation or didactic instruction on topics such as health care finance. The New Mexico and Wake Forest tracks offer elective rotations on health care economics. Several track directors mentioned successfully expanding curricula on health care value from the HM track into IM residency programs at large, providing a measurable service to the residency programs while ensuring content delivery and freeing up additional time for track activities.

Scholarship and Career Development

Most programs provide targeted career development for residents. Six tracks provide sessions on job procurement skills, such as curriculum vitae preparation and interviewing (Table 2). Many also provide content on venues for disseminating scholarly activity. The Colorado, Kentucky, New Mexico, and Tulane programs feature content on abstract and poster creation. Leadership development is addressed in several tracks through dedicated track activities or participation in discrete, outside-track events. Specifically, Colorado offers a leadership track for residents interested in hospital administration, Cleveland has a leadership journal club, Wake Forest enrolls HM residents in leadership training available through the university, and Minnesota sends residents to the Society of Hospital Medicine’s Leadership Academy (Table 2).

Clinical Rotations

Almost all tracks include a clinical rotation, typically pairing residents directly with hospitalist attendings to encourage autonomy and mentorship. Several also offer elective rotations in various disciplines within HM (Table 2). The Kentucky and Virginia tracks incorporate working with advanced practice providers into their practicums. The Cleveland, Minnesota, Tulane, and Virginia tracks offer HM rotations in community hospitals or postacute settings.

HM rotations also pair clinical experiences with didactic education on relevant topics (eg, billing and coding). The Cleveland, Minnesota, and Virginia tracks developed clinical rotations reflecting the common 7-on and 7-off schedule with nonclinical obligations, such as seminars linking specific content to clinical experiences, during nonclinical time.

Our investigation into the current state of HM training found that HM track curricula focus largely on QI, health care economics, and professional development. This focus likely developed in response to hospitalists’ increasing engagement in related endeavors. HM tracks have dynamic and variable structures, reflecting an evolving field and the need to fit into existing IM residency program structures. Similarly, the content covered in HM tracks is tightly linked to perceived opportunities within IM residency curricula. The heterogeneity of content suggests the breadth and ambiguity of necessary competencies for aspiring hospitalists. One of the 11 tracks has not had any residents enroll within the past few years—a testament to the continued effort necessary to sustain such tracks, including curricular updates and recruiting. Conversely, many programs now share track content with the larger IM residency program, suggesting HM tracks may be near the forefront of medical education in some areas.

Our study had several limitations. As we are unaware of any databases of HM tracks, we discussed tracks with professional contacts, performed Internet searches, and reviewed IM residency program websites. Our search, however, was not exhaustive; despite our best efforts, we may have missed or mischaracterized some track offerings. Nevertheless, we think that our analysis represents the first thorough compilation of HM tracks and that it will be useful to institutions seeking to create or enhance HM-specific training.

As the field continues to evolve, we are optimistic about the future of HM training. We suspect that HM residency training tracks will continue to expand. More work is needed so these tracks can adjust to the changing HM and IM residency program landscapes and supply well-trained physicians for the HM workforce.

Acknowledgment

The authors thank track directors Alpesh Amin, David Gugliotti, Rick Hilger, Karnjit Johl, Nasir Majeed, Georgia McIntosh, Charles Pizanis, and Jeff Wiese for making this study possible.

Disclosure

Nothing to report.

Psychiatric conditions affect 1 in 5 children,^1,2 and having a comorbid psychiatric condition is associated with worse outcomes in children hospitalized for medical or surgical indications.^3-7 Although little is known about interventions for improving outcomes for hospitalized children with psychiatric conditions,⁸ several interventions that integrate medical and psychiatric care are known to improve ambulatory patient outcomes.^9-14 The success of initiatives that test whether integrated medical and psychiatric care models can improve pediatric hospital outcomes depends on reliable identification of comorbid psychiatric conditions and family and clinician having a shared understanding of a patient’s psychiatric diagnoses.

Mental health care system fragmentation, stigma, and privacy issues^15-20 may contribute to clinical teams and families having disparate views of psychiatric comorbidities. Evidence suggests that hospital clinicians caring for pediatric medical and surgical inpatients are often unaware of a psychiatric condition that has been diagnosed or managed in the ambulatory setting,^3,6 even in cases in which the patient and family are aware of the diagnosis. Conversely, for other patients, clinicians may be aware of a psychiatric diagnosis, but patient and family may not share that understanding or reliably report a psychiatric diagnosis.^21-23 Although hospitalization may not be the ideal setting for identifying a new psychiatric diagnosis, given the short-term relationship between patient and clinical care team, addressing and managing a psychiatric comorbidity that is known to family or clinician are important elements of patient-centered hospital care.

The success of interventions in improving hospital outcomes for hospitalized children with psychiatric comorbidity depends on patients, families, and clinicians having a shared understanding of which patients have psychiatric conditions, and on accurate estimates of the scope of the population in need of psychiatric care during pediatric hospitalization.

We conducted a study to compare estimates of point prevalence of psychiatric comorbidity identified by family report (FR) or clinician documentation (CD) and to determine the degree of FR–CD agreement regarding the presence of psychiatric comorbidity in hospitalized children.

We estimated point prevalence and determined FR–CD agreement regarding diagnosed psychiatric comorbidities in a cross-sectional sample of pediatric medical and surgical hospitalizations at Children’s Hospital of Philadelphia (CHOP). CHOP is a free-standing 535-bed children’s hospital that serves as a community hospital for the city of Philadelphia; a regional referral center for eastern Pennsylvania, Delaware, and southern New Jersey; and a national and international quaternary referral center. This study was approved by CHOP’s institutional review board.

Patients eligible for inclusion in the study were 4 to 21 years old and hospitalized for a medical or surgical indication. Patients were ineligible if they were hospitalized for a primary psychiatric indication, were medically unstable (eg, received end-of-life care or escalating interventions for a life-threatening condition), had significant cognitive impairment precluding communication (eg, history of severe hypoxic-ischemic encephalopathy), or did not speak English (pertains to consenting parent, guardian, or patient).

The cross-sectional patient sample was selected using a point prevalence recruitment strategy. All eligible patients on each of CHOP’s 20 inpatient medical, surgical, and critical care units were approached for study participation on 2 dates between July 2015 and March 2016. To avoid enrolling the same patient multiple times for a single hospitalization, we separated recruitment dates on each unit by at least 3 months. A goal sample size of 100 to 150 patients was selected to provide precision sufficient to achieve a confidence interval (CI) of 10% around an estimate of the point prevalence of any mental health condition.

To obtain family report of prior psychiatric diagnoses, we interviewed patients and/or their parents during the hospitalization. For 18- to 21-year-old patients, the adolescent patient completed the interview. For patients under 18 years old, parents completed the interview, and for 14- to 17-year-old adolescents,either the parent, the patient, or both could complete the interview. Adolescents were asked to complete the interview confidentially without a parent present. The structured interview included questions derived from the National Survey of Children’s Health²⁴ and the Services Assessment for Children and Adolescents²² to report the patient’s active psychiatric conditions. Interviewees reported whether the patient had ever been diagnosed with any psychiatric disorder, whether the condition was ongoing in the year prior to hospitalization, and whether the patient received any mental health services in clinical settings or school in the 12 months prior to hospitalization.

For CD, we identified a psychiatric diagnosis associated with the index hospitalization if a psychiatric diagnosis was noted in the patient’s admission note, discharge summary, or hospital problem list, or if an International Classification of Diseases (ICD) code for a psychiatric diagnosis was submitted for billing for the index hospitalization. The Healthcare Cost and Utilization Project condition classification system was used to sort psychiatric condition codes^25-27 into 5 categories: attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, depression, disruptive behavior disorders, and autism spectrum disorders. A residual category of other, less common psychiatric conditions included eating disorders, attachment disorders, and bipolar disorder.

For each condition category, we determined the point prevalence of having a psychiatric diagnosis identified by FR and having a diagnosis identified by CD. We used McNemar tests to compare point prevalence estimates, the Clopper-Pearson method to calculate CIs around the estimates,²⁸ and Cohen κ statistics to estimate FR–CD agreement regarding psychiatric diagnoses, grouping patients by type of psychiatric diagnosis and by clinical and demographic characteristics. All statistical tests were 2-sided, and P < 0.05 was used for statistical significance. All statistical analyses were performed with Stata Version 13.1 (StataCorp, College Station, Texas).

Of 640 patients hospitalized on study recruitment dates, 411 were ineligible for the study (282 were <4 or >21 years old, 42 were not English speakers, 37 had cognitive impairment, 30 were not medically stable, and 20 were admitted for a primary psychiatric diagnosis). Of the 229 eligible patients, 119 (52%) enrolled. Included patients were 57% female; 9% Hispanic; and 35% black, 55% white, and 15% other race. Forty-eight percent of the enrollees had Medicaid (48%), and 52% had private health insurance. Mean age was 12.3 years. Of enrolled patients, 38% were admitted to subspecialty medical services. Enrollee demographics were representative of hospital-level demographics for the study-eligible population; there were no significant differences in age, sex, race, ethnicity, payer type, or hospital service admission type between enrollees and patients who declined to participate (all Ps > 0.05). Table 1 lists demographic and clinical characteristics of the complete study sample and of the groups with FR- or CD-identified psychiatric diagnosis.

Of 119 enrollees, 26 (22%; 95% CI, 15%-30%) had at least 1 FR-identified comorbid psychiatric diagnosis, and 30 (25%; 95% CI, 17%-33%) had at least 1 CD-identified diagnosis. In 13 cases, adolescents (age, 14-17 years) and their parents both completed the structured interview; there were no discrepancies between interview results.

In total, 39 of 119 patients (33%, 95% CI: 24-42%) had either a family-reported or clinician-documented psychiatric diagnosis at the time of hospitalization. For 17 of 119 patients (14%; 95% CI: 9-22%), family-report and clinician-documentation both identified the patient as having a comorbid psychiatric diagnosis. For 9 of 119 patients (8%; 95% CI: 4-14%) families reported a psychiatric diagnosis, but clinicians did not document one. Conversely, for another 13 of 119 patients (11%; 95% CI: 6-18%), a clinician documented a psychiatric diagnosis but the family did not report one. The Figure shows the point prevalence of family-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses for 5 common psychiatric condition categories.

At a tertiary-care children’s hospital, we found high point prevalence of comorbid psychiatric conditions and low agreement between FR- and CD-identified psychiatric conditions. Estimates of the prevalence of psychiatric comorbidity among pediatric medical and surgical inpatients were similar for FR- and CD-identified psychiatric conditions, though each method missed about one third of the cases identified by the other method. FR only and CD only each identified about 1 in 4 or 5 hospitalized children and adolescents with a psychiatric comorbidity. When FR and CD were combined, a comorbid psychiatric diagnosis was identified in about 1 in 3 medical and surgical inpatients aged 4 to 21 years. FR–CD agreement was substantial only for ADHD and was fair to slight for most other psychiatric conditions, including autism, depression, anxiety, and disruptive behavior disorders (eg, conduct disorder, oppositional defiant disorder).

Our finding that psychiatric conditions were more commonly reported by families and documented by clinicians for white patients is consistent with a large body of evidence showing that racial or ethnic minority patients experience more stigma related to mental health diagnoses and use mental health services less.^29-33 Families were more likely to report use of mental health services than a known mental health diagnosis. This finding may reflect families’ willingness to use services even if they do not understand or experience stigma related to psychiatric diagnoses. Alternatively, use of mental health services without a diagnosis may reflect clinicians’ willingness to refer a child for services when the child is perceived to have an impairment even in the absence of a clear psychiatric diagnosis.

The low FR–CD agreement regarding psychiatric conditions in hospitalized children and adolescents raises 3 issues for pediatric hospital care. First, earlier studies likely underestimated the prevalence of these conditions. A 2014 study of a national sample found that 13% of children hospitalized for a physical health condition had psychiatric comorbidity.²⁵ That study and other large-scale studies showing a high and increasing prevalence of primary psychiatric conditions in hospitalized children and adolescents have relied on administrative data derived from clinician-documented diagnoses.^25-27 Our study findings suggest that reliance on administrative data could result in underestimation of the prevalence of psychiatric comorbidity in hospitalized children by as much as 40%. Pediatric hospitals are reporting a shortage of pediatric mental health specialists.³⁴ Augmenting estimates of the prevalence of psychiatric comorbidity in hospitalized children with reports from other sources, including families or outpatient administrative records, may aid health systems in allocating mental health resources for pediatric inpatients.

The second issue is that the present data suggest that families and clinicians do not share the same information about a child’s psychiatric diagnoses when the child is hospitalized for a medical condition or surgical procedure. Low FR–CD agreement regarding psychiatric diagnoses suggests families and clinical teams are not always “on the same page” about psychiatric needs during hospitalization. Implications of this finding are relevant to inpatient and ambulatory care settings. In cases in which a clinician recognizes a psychiatric condition but the family does not, the family may not seek outpatient treatment. In the present study, one third of patients with a psychiatric diagnosis identified by CD but not FR were not engaged in ambulatory treatment for the condition. Conversely, a psychiatric diagnosis identified by FR but not CD suggests clinical teams lack the skills and knowledge needed to elicit information about psychiatric conditions and their potential relevance to inpatient care. As a result, clinicians may miss opportunities to provide interventions that may improve physical or mental health outcomes. For example, clinical teams with information about a patient’s anxiety disorder may be better able to provide brief interventions to prevent medical treatments from triggering anxiety symptoms and to mitigate the risk for traumatic stress symptoms related to the hospitalization.

The third issue is that anxiety disorders were most likely to be the subject of FR–CD disagreement. This finding identifies children with anxiety disorders as a priority population for research into differences between families and clinicians in understanding patients’ psychiatric diagnoses. Our findings suggest families and clinicians have different views of patients’ anxiety symptoms. Anxiety disorders are a risk factor for worse outcomes in children with chronic physical conditions,^3,35-37 and acute hospitalization is associated with posthospital anxiety symptoms.^38,39 Thus, anxiety disorders are particularly relevant to hospital care and are a priority for research on the differences between families’ and clinicians’ perspectives on children’s psychiatric diagnoses.

Our findings should be interpreted in the context of study limitations. First, because of resource limitations, we did not obtain psychiatric diagnostic evaluations or records to confirm FR- and CD-identified psychiatric diagnoses. Although this lack of clinical confirmation could have resulted in misclassification bias, the risk of bias was no higher than in many other studies that have successfully used hospital records^21,25 and family reports to identify psychiatric comorbidity.⁴⁰ Second, because the study included only English-speaking patients and families, results cannot be generalized to non-English-speaking populations. Third, this was a single-center study, conducted in a free-standing tertiary-care children’s hospital. Sample size was small, particularly for estimating the prevalence of individual psychiatric conditions. Patient characteristics and clinical practice patterns may differ at other types of hospitals. Larger multicenter studies are warranted. Despite these limitations, our results provide important new information that can further our understanding of the epidemiology of psychiatric conditions in hospitalized children. This information should interest clinical teams caring for children with comorbid physical and mental health conditions.

Low FR–CD agreement regarding hospitalized children’s psychiatric comorbidities suggests that patients and their families and clinicians do not always share the same information about these comorbidities, and that the prevalence of psychiatric comorbidity in hospitalized children is likely underestimated. To allocate adequate resources for these children, health systems may need to obtain information from multiple sources. Furthermore, we need to better our understanding of strategies for communicating about hospitalized children’s psychiatric conditions so that we can develop interventions to improve hospital outcomes for this vulnerable population.

Disclosures

The direct costs of this project were funded by an internal pilot grant from the Center for Pediatric Health Disparities at Children’s Hospital of Philadelphia. Dr. Doupnik was supported by Ruth L. Kirschstein National Research Service Award T32-HP010026-11, funded by the National Institutes of Health. The sponsors had no role in study design; collection, analysis, or interpretation of data; manuscript writing; or deciding to submit this article for publication.

Psychiatric conditions affect 1 in 5 children,^1,2 and having a comorbid psychiatric condition is associated with worse outcomes in children hospitalized for medical or surgical indications.^3-7 Although little is known about interventions for improving outcomes for hospitalized children with psychiatric conditions,⁸ several interventions that integrate medical and psychiatric care are known to improve ambulatory patient outcomes.^9-14 The success of initiatives that test whether integrated medical and psychiatric care models can improve pediatric hospital outcomes depends on reliable identification of comorbid psychiatric conditions and family and clinician having a shared understanding of a patient’s psychiatric diagnoses.

Mental health care system fragmentation, stigma, and privacy issues^15-20 may contribute to clinical teams and families having disparate views of psychiatric comorbidities. Evidence suggests that hospital clinicians caring for pediatric medical and surgical inpatients are often unaware of a psychiatric condition that has been diagnosed or managed in the ambulatory setting,^3,6 even in cases in which the patient and family are aware of the diagnosis. Conversely, for other patients, clinicians may be aware of a psychiatric diagnosis, but patient and family may not share that understanding or reliably report a psychiatric diagnosis.^21-23 Although hospitalization may not be the ideal setting for identifying a new psychiatric diagnosis, given the short-term relationship between patient and clinical care team, addressing and managing a psychiatric comorbidity that is known to family or clinician are important elements of patient-centered hospital care.

The success of interventions in improving hospital outcomes for hospitalized children with psychiatric comorbidity depends on patients, families, and clinicians having a shared understanding of which patients have psychiatric conditions, and on accurate estimates of the scope of the population in need of psychiatric care during pediatric hospitalization.

We conducted a study to compare estimates of point prevalence of psychiatric comorbidity identified by family report (FR) or clinician documentation (CD) and to determine the degree of FR–CD agreement regarding the presence of psychiatric comorbidity in hospitalized children.

We estimated point prevalence and determined FR–CD agreement regarding diagnosed psychiatric comorbidities in a cross-sectional sample of pediatric medical and surgical hospitalizations at Children’s Hospital of Philadelphia (CHOP). CHOP is a free-standing 535-bed children’s hospital that serves as a community hospital for the city of Philadelphia; a regional referral center for eastern Pennsylvania, Delaware, and southern New Jersey; and a national and international quaternary referral center. This study was approved by CHOP’s institutional review board.

Patients eligible for inclusion in the study were 4 to 21 years old and hospitalized for a medical or surgical indication. Patients were ineligible if they were hospitalized for a primary psychiatric indication, were medically unstable (eg, received end-of-life care or escalating interventions for a life-threatening condition), had significant cognitive impairment precluding communication (eg, history of severe hypoxic-ischemic encephalopathy), or did not speak English (pertains to consenting parent, guardian, or patient).

The cross-sectional patient sample was selected using a point prevalence recruitment strategy. All eligible patients on each of CHOP’s 20 inpatient medical, surgical, and critical care units were approached for study participation on 2 dates between July 2015 and March 2016. To avoid enrolling the same patient multiple times for a single hospitalization, we separated recruitment dates on each unit by at least 3 months. A goal sample size of 100 to 150 patients was selected to provide precision sufficient to achieve a confidence interval (CI) of 10% around an estimate of the point prevalence of any mental health condition.

To obtain family report of prior psychiatric diagnoses, we interviewed patients and/or their parents during the hospitalization. For 18- to 21-year-old patients, the adolescent patient completed the interview. For patients under 18 years old, parents completed the interview, and for 14- to 17-year-old adolescents,either the parent, the patient, or both could complete the interview. Adolescents were asked to complete the interview confidentially without a parent present. The structured interview included questions derived from the National Survey of Children’s Health²⁴ and the Services Assessment for Children and Adolescents²² to report the patient’s active psychiatric conditions. Interviewees reported whether the patient had ever been diagnosed with any psychiatric disorder, whether the condition was ongoing in the year prior to hospitalization, and whether the patient received any mental health services in clinical settings or school in the 12 months prior to hospitalization.

For CD, we identified a psychiatric diagnosis associated with the index hospitalization if a psychiatric diagnosis was noted in the patient’s admission note, discharge summary, or hospital problem list, or if an International Classification of Diseases (ICD) code for a psychiatric diagnosis was submitted for billing for the index hospitalization. The Healthcare Cost and Utilization Project condition classification system was used to sort psychiatric condition codes^25-27 into 5 categories: attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, depression, disruptive behavior disorders, and autism spectrum disorders. A residual category of other, less common psychiatric conditions included eating disorders, attachment disorders, and bipolar disorder.

For each condition category, we determined the point prevalence of having a psychiatric diagnosis identified by FR and having a diagnosis identified by CD. We used McNemar tests to compare point prevalence estimates, the Clopper-Pearson method to calculate CIs around the estimates,²⁸ and Cohen κ statistics to estimate FR–CD agreement regarding psychiatric diagnoses, grouping patients by type of psychiatric diagnosis and by clinical and demographic characteristics. All statistical tests were 2-sided, and P < 0.05 was used for statistical significance. All statistical analyses were performed with Stata Version 13.1 (StataCorp, College Station, Texas).

Of 640 patients hospitalized on study recruitment dates, 411 were ineligible for the study (282 were <4 or >21 years old, 42 were not English speakers, 37 had cognitive impairment, 30 were not medically stable, and 20 were admitted for a primary psychiatric diagnosis). Of the 229 eligible patients, 119 (52%) enrolled. Included patients were 57% female; 9% Hispanic; and 35% black, 55% white, and 15% other race. Forty-eight percent of the enrollees had Medicaid (48%), and 52% had private health insurance. Mean age was 12.3 years. Of enrolled patients, 38% were admitted to subspecialty medical services. Enrollee demographics were representative of hospital-level demographics for the study-eligible population; there were no significant differences in age, sex, race, ethnicity, payer type, or hospital service admission type between enrollees and patients who declined to participate (all Ps > 0.05). Table 1 lists demographic and clinical characteristics of the complete study sample and of the groups with FR- or CD-identified psychiatric diagnosis.

Of 119 enrollees, 26 (22%; 95% CI, 15%-30%) had at least 1 FR-identified comorbid psychiatric diagnosis, and 30 (25%; 95% CI, 17%-33%) had at least 1 CD-identified diagnosis. In 13 cases, adolescents (age, 14-17 years) and their parents both completed the structured interview; there were no discrepancies between interview results.

In total, 39 of 119 patients (33%, 95% CI: 24-42%) had either a family-reported or clinician-documented psychiatric diagnosis at the time of hospitalization. For 17 of 119 patients (14%; 95% CI: 9-22%), family-report and clinician-documentation both identified the patient as having a comorbid psychiatric diagnosis. For 9 of 119 patients (8%; 95% CI: 4-14%) families reported a psychiatric diagnosis, but clinicians did not document one. Conversely, for another 13 of 119 patients (11%; 95% CI: 6-18%), a clinician documented a psychiatric diagnosis but the family did not report one. The Figure shows the point prevalence of family-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses for 5 common psychiatric condition categories.

At a tertiary-care children’s hospital, we found high point prevalence of comorbid psychiatric conditions and low agreement between FR- and CD-identified psychiatric conditions. Estimates of the prevalence of psychiatric comorbidity among pediatric medical and surgical inpatients were similar for FR- and CD-identified psychiatric conditions, though each method missed about one third of the cases identified by the other method. FR only and CD only each identified about 1 in 4 or 5 hospitalized children and adolescents with a psychiatric comorbidity. When FR and CD were combined, a comorbid psychiatric diagnosis was identified in about 1 in 3 medical and surgical inpatients aged 4 to 21 years. FR–CD agreement was substantial only for ADHD and was fair to slight for most other psychiatric conditions, including autism, depression, anxiety, and disruptive behavior disorders (eg, conduct disorder, oppositional defiant disorder).

Our finding that psychiatric conditions were more commonly reported by families and documented by clinicians for white patients is consistent with a large body of evidence showing that racial or ethnic minority patients experience more stigma related to mental health diagnoses and use mental health services less.^29-33 Families were more likely to report use of mental health services than a known mental health diagnosis. This finding may reflect families’ willingness to use services even if they do not understand or experience stigma related to psychiatric diagnoses. Alternatively, use of mental health services without a diagnosis may reflect clinicians’ willingness to refer a child for services when the child is perceived to have an impairment even in the absence of a clear psychiatric diagnosis.

The low FR–CD agreement regarding psychiatric conditions in hospitalized children and adolescents raises 3 issues for pediatric hospital care. First, earlier studies likely underestimated the prevalence of these conditions. A 2014 study of a national sample found that 13% of children hospitalized for a physical health condition had psychiatric comorbidity.²⁵ That study and other large-scale studies showing a high and increasing prevalence of primary psychiatric conditions in hospitalized children and adolescents have relied on administrative data derived from clinician-documented diagnoses.^25-27 Our study findings suggest that reliance on administrative data could result in underestimation of the prevalence of psychiatric comorbidity in hospitalized children by as much as 40%. Pediatric hospitals are reporting a shortage of pediatric mental health specialists.³⁴ Augmenting estimates of the prevalence of psychiatric comorbidity in hospitalized children with reports from other sources, including families or outpatient administrative records, may aid health systems in allocating mental health resources for pediatric inpatients.

The second issue is that the present data suggest that families and clinicians do not share the same information about a child’s psychiatric diagnoses when the child is hospitalized for a medical condition or surgical procedure. Low FR–CD agreement regarding psychiatric diagnoses suggests families and clinical teams are not always “on the same page” about psychiatric needs during hospitalization. Implications of this finding are relevant to inpatient and ambulatory care settings. In cases in which a clinician recognizes a psychiatric condition but the family does not, the family may not seek outpatient treatment. In the present study, one third of patients with a psychiatric diagnosis identified by CD but not FR were not engaged in ambulatory treatment for the condition. Conversely, a psychiatric diagnosis identified by FR but not CD suggests clinical teams lack the skills and knowledge needed to elicit information about psychiatric conditions and their potential relevance to inpatient care. As a result, clinicians may miss opportunities to provide interventions that may improve physical or mental health outcomes. For example, clinical teams with information about a patient’s anxiety disorder may be better able to provide brief interventions to prevent medical treatments from triggering anxiety symptoms and to mitigate the risk for traumatic stress symptoms related to the hospitalization.

The third issue is that anxiety disorders were most likely to be the subject of FR–CD disagreement. This finding identifies children with anxiety disorders as a priority population for research into differences between families and clinicians in understanding patients’ psychiatric diagnoses. Our findings suggest families and clinicians have different views of patients’ anxiety symptoms. Anxiety disorders are a risk factor for worse outcomes in children with chronic physical conditions,^3,35-37 and acute hospitalization is associated with posthospital anxiety symptoms.^38,39 Thus, anxiety disorders are particularly relevant to hospital care and are a priority for research on the differences between families’ and clinicians’ perspectives on children’s psychiatric diagnoses.

Our findings should be interpreted in the context of study limitations. First, because of resource limitations, we did not obtain psychiatric diagnostic evaluations or records to confirm FR- and CD-identified psychiatric diagnoses. Although this lack of clinical confirmation could have resulted in misclassification bias, the risk of bias was no higher than in many other studies that have successfully used hospital records^21,25 and family reports to identify psychiatric comorbidity.⁴⁰ Second, because the study included only English-speaking patients and families, results cannot be generalized to non-English-speaking populations. Third, this was a single-center study, conducted in a free-standing tertiary-care children’s hospital. Sample size was small, particularly for estimating the prevalence of individual psychiatric conditions. Patient characteristics and clinical practice patterns may differ at other types of hospitals. Larger multicenter studies are warranted. Despite these limitations, our results provide important new information that can further our understanding of the epidemiology of psychiatric conditions in hospitalized children. This information should interest clinical teams caring for children with comorbid physical and mental health conditions.

Low FR–CD agreement regarding hospitalized children’s psychiatric comorbidities suggests that patients and their families and clinicians do not always share the same information about these comorbidities, and that the prevalence of psychiatric comorbidity in hospitalized children is likely underestimated. To allocate adequate resources for these children, health systems may need to obtain information from multiple sources. Furthermore, we need to better our understanding of strategies for communicating about hospitalized children’s psychiatric conditions so that we can develop interventions to improve hospital outcomes for this vulnerable population.

Disclosures

The direct costs of this project were funded by an internal pilot grant from the Center for Pediatric Health Disparities at Children’s Hospital of Philadelphia. Dr. Doupnik was supported by Ruth L. Kirschstein National Research Service Award T32-HP010026-11, funded by the National Institutes of Health. The sponsors had no role in study design; collection, analysis, or interpretation of data; manuscript writing; or deciding to submit this article for publication.

Psychiatric conditions affect 1 in 5 children,^1,2 and having a comorbid psychiatric condition is associated with worse outcomes in children hospitalized for medical or surgical indications.^3-7 Although little is known about interventions for improving outcomes for hospitalized children with psychiatric conditions,⁸ several interventions that integrate medical and psychiatric care are known to improve ambulatory patient outcomes.^9-14 The success of initiatives that test whether integrated medical and psychiatric care models can improve pediatric hospital outcomes depends on reliable identification of comorbid psychiatric conditions and family and clinician having a shared understanding of a patient’s psychiatric diagnoses.

Mental health care system fragmentation, stigma, and privacy issues^15-20 may contribute to clinical teams and families having disparate views of psychiatric comorbidities. Evidence suggests that hospital clinicians caring for pediatric medical and surgical inpatients are often unaware of a psychiatric condition that has been diagnosed or managed in the ambulatory setting,^3,6 even in cases in which the patient and family are aware of the diagnosis. Conversely, for other patients, clinicians may be aware of a psychiatric diagnosis, but patient and family may not share that understanding or reliably report a psychiatric diagnosis.^21-23 Although hospitalization may not be the ideal setting for identifying a new psychiatric diagnosis, given the short-term relationship between patient and clinical care team, addressing and managing a psychiatric comorbidity that is known to family or clinician are important elements of patient-centered hospital care.

The success of interventions in improving hospital outcomes for hospitalized children with psychiatric comorbidity depends on patients, families, and clinicians having a shared understanding of which patients have psychiatric conditions, and on accurate estimates of the scope of the population in need of psychiatric care during pediatric hospitalization.

We conducted a study to compare estimates of point prevalence of psychiatric comorbidity identified by family report (FR) or clinician documentation (CD) and to determine the degree of FR–CD agreement regarding the presence of psychiatric comorbidity in hospitalized children.

We estimated point prevalence and determined FR–CD agreement regarding diagnosed psychiatric comorbidities in a cross-sectional sample of pediatric medical and surgical hospitalizations at Children’s Hospital of Philadelphia (CHOP). CHOP is a free-standing 535-bed children’s hospital that serves as a community hospital for the city of Philadelphia; a regional referral center for eastern Pennsylvania, Delaware, and southern New Jersey; and a national and international quaternary referral center. This study was approved by CHOP’s institutional review board.

Patients eligible for inclusion in the study were 4 to 21 years old and hospitalized for a medical or surgical indication. Patients were ineligible if they were hospitalized for a primary psychiatric indication, were medically unstable (eg, received end-of-life care or escalating interventions for a life-threatening condition), had significant cognitive impairment precluding communication (eg, history of severe hypoxic-ischemic encephalopathy), or did not speak English (pertains to consenting parent, guardian, or patient).

The cross-sectional patient sample was selected using a point prevalence recruitment strategy. All eligible patients on each of CHOP’s 20 inpatient medical, surgical, and critical care units were approached for study participation on 2 dates between July 2015 and March 2016. To avoid enrolling the same patient multiple times for a single hospitalization, we separated recruitment dates on each unit by at least 3 months. A goal sample size of 100 to 150 patients was selected to provide precision sufficient to achieve a confidence interval (CI) of 10% around an estimate of the point prevalence of any mental health condition.

To obtain family report of prior psychiatric diagnoses, we interviewed patients and/or their parents during the hospitalization. For 18- to 21-year-old patients, the adolescent patient completed the interview. For patients under 18 years old, parents completed the interview, and for 14- to 17-year-old adolescents,either the parent, the patient, or both could complete the interview. Adolescents were asked to complete the interview confidentially without a parent present. The structured interview included questions derived from the National Survey of Children’s Health²⁴ and the Services Assessment for Children and Adolescents²² to report the patient’s active psychiatric conditions. Interviewees reported whether the patient had ever been diagnosed with any psychiatric disorder, whether the condition was ongoing in the year prior to hospitalization, and whether the patient received any mental health services in clinical settings or school in the 12 months prior to hospitalization.

For CD, we identified a psychiatric diagnosis associated with the index hospitalization if a psychiatric diagnosis was noted in the patient’s admission note, discharge summary, or hospital problem list, or if an International Classification of Diseases (ICD) code for a psychiatric diagnosis was submitted for billing for the index hospitalization. The Healthcare Cost and Utilization Project condition classification system was used to sort psychiatric condition codes^25-27 into 5 categories: attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, depression, disruptive behavior disorders, and autism spectrum disorders. A residual category of other, less common psychiatric conditions included eating disorders, attachment disorders, and bipolar disorder.

For each condition category, we determined the point prevalence of having a psychiatric diagnosis identified by FR and having a diagnosis identified by CD. We used McNemar tests to compare point prevalence estimates, the Clopper-Pearson method to calculate CIs around the estimates,²⁸ and Cohen κ statistics to estimate FR–CD agreement regarding psychiatric diagnoses, grouping patients by type of psychiatric diagnosis and by clinical and demographic characteristics. All statistical tests were 2-sided, and P < 0.05 was used for statistical significance. All statistical analyses were performed with Stata Version 13.1 (StataCorp, College Station, Texas).

Of 640 patients hospitalized on study recruitment dates, 411 were ineligible for the study (282 were <4 or >21 years old, 42 were not English speakers, 37 had cognitive impairment, 30 were not medically stable, and 20 were admitted for a primary psychiatric diagnosis). Of the 229 eligible patients, 119 (52%) enrolled. Included patients were 57% female; 9% Hispanic; and 35% black, 55% white, and 15% other race. Forty-eight percent of the enrollees had Medicaid (48%), and 52% had private health insurance. Mean age was 12.3 years. Of enrolled patients, 38% were admitted to subspecialty medical services. Enrollee demographics were representative of hospital-level demographics for the study-eligible population; there were no significant differences in age, sex, race, ethnicity, payer type, or hospital service admission type between enrollees and patients who declined to participate (all Ps > 0.05). Table 1 lists demographic and clinical characteristics of the complete study sample and of the groups with FR- or CD-identified psychiatric diagnosis.

Of 119 enrollees, 26 (22%; 95% CI, 15%-30%) had at least 1 FR-identified comorbid psychiatric diagnosis, and 30 (25%; 95% CI, 17%-33%) had at least 1 CD-identified diagnosis. In 13 cases, adolescents (age, 14-17 years) and their parents both completed the structured interview; there were no discrepancies between interview results.

In total, 39 of 119 patients (33%, 95% CI: 24-42%) had either a family-reported or clinician-documented psychiatric diagnosis at the time of hospitalization. For 17 of 119 patients (14%; 95% CI: 9-22%), family-report and clinician-documentation both identified the patient as having a comorbid psychiatric diagnosis. For 9 of 119 patients (8%; 95% CI: 4-14%) families reported a psychiatric diagnosis, but clinicians did not document one. Conversely, for another 13 of 119 patients (11%; 95% CI: 6-18%), a clinician documented a psychiatric diagnosis but the family did not report one. The Figure shows the point prevalence of family-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses for 5 common psychiatric condition categories.

At a tertiary-care children’s hospital, we found high point prevalence of comorbid psychiatric conditions and low agreement between FR- and CD-identified psychiatric conditions. Estimates of the prevalence of psychiatric comorbidity among pediatric medical and surgical inpatients were similar for FR- and CD-identified psychiatric conditions, though each method missed about one third of the cases identified by the other method. FR only and CD only each identified about 1 in 4 or 5 hospitalized children and adolescents with a psychiatric comorbidity. When FR and CD were combined, a comorbid psychiatric diagnosis was identified in about 1 in 3 medical and surgical inpatients aged 4 to 21 years. FR–CD agreement was substantial only for ADHD and was fair to slight for most other psychiatric conditions, including autism, depression, anxiety, and disruptive behavior disorders (eg, conduct disorder, oppositional defiant disorder).

Our finding that psychiatric conditions were more commonly reported by families and documented by clinicians for white patients is consistent with a large body of evidence showing that racial or ethnic minority patients experience more stigma related to mental health diagnoses and use mental health services less.^29-33 Families were more likely to report use of mental health services than a known mental health diagnosis. This finding may reflect families’ willingness to use services even if they do not understand or experience stigma related to psychiatric diagnoses. Alternatively, use of mental health services without a diagnosis may reflect clinicians’ willingness to refer a child for services when the child is perceived to have an impairment even in the absence of a clear psychiatric diagnosis.

The low FR–CD agreement regarding psychiatric conditions in hospitalized children and adolescents raises 3 issues for pediatric hospital care. First, earlier studies likely underestimated the prevalence of these conditions. A 2014 study of a national sample found that 13% of children hospitalized for a physical health condition had psychiatric comorbidity.²⁵ That study and other large-scale studies showing a high and increasing prevalence of primary psychiatric conditions in hospitalized children and adolescents have relied on administrative data derived from clinician-documented diagnoses.^25-27 Our study findings suggest that reliance on administrative data could result in underestimation of the prevalence of psychiatric comorbidity in hospitalized children by as much as 40%. Pediatric hospitals are reporting a shortage of pediatric mental health specialists.³⁴ Augmenting estimates of the prevalence of psychiatric comorbidity in hospitalized children with reports from other sources, including families or outpatient administrative records, may aid health systems in allocating mental health resources for pediatric inpatients.

The second issue is that the present data suggest that families and clinicians do not share the same information about a child’s psychiatric diagnoses when the child is hospitalized for a medical condition or surgical procedure. Low FR–CD agreement regarding psychiatric diagnoses suggests families and clinical teams are not always “on the same page” about psychiatric needs during hospitalization. Implications of this finding are relevant to inpatient and ambulatory care settings. In cases in which a clinician recognizes a psychiatric condition but the family does not, the family may not seek outpatient treatment. In the present study, one third of patients with a psychiatric diagnosis identified by CD but not FR were not engaged in ambulatory treatment for the condition. Conversely, a psychiatric diagnosis identified by FR but not CD suggests clinical teams lack the skills and knowledge needed to elicit information about psychiatric conditions and their potential relevance to inpatient care. As a result, clinicians may miss opportunities to provide interventions that may improve physical or mental health outcomes. For example, clinical teams with information about a patient’s anxiety disorder may be better able to provide brief interventions to prevent medical treatments from triggering anxiety symptoms and to mitigate the risk for traumatic stress symptoms related to the hospitalization.

The third issue is that anxiety disorders were most likely to be the subject of FR–CD disagreement. This finding identifies children with anxiety disorders as a priority population for research into differences between families and clinicians in understanding patients’ psychiatric diagnoses. Our findings suggest families and clinicians have different views of patients’ anxiety symptoms. Anxiety disorders are a risk factor for worse outcomes in children with chronic physical conditions,^3,35-37 and acute hospitalization is associated with posthospital anxiety symptoms.^38,39 Thus, anxiety disorders are particularly relevant to hospital care and are a priority for research on the differences between families’ and clinicians’ perspectives on children’s psychiatric diagnoses.

Our findings should be interpreted in the context of study limitations. First, because of resource limitations, we did not obtain psychiatric diagnostic evaluations or records to confirm FR- and CD-identified psychiatric diagnoses. Although this lack of clinical confirmation could have resulted in misclassification bias, the risk of bias was no higher than in many other studies that have successfully used hospital records^21,25 and family reports to identify psychiatric comorbidity.⁴⁰ Second, because the study included only English-speaking patients and families, results cannot be generalized to non-English-speaking populations. Third, this was a single-center study, conducted in a free-standing tertiary-care children’s hospital. Sample size was small, particularly for estimating the prevalence of individual psychiatric conditions. Patient characteristics and clinical practice patterns may differ at other types of hospitals. Larger multicenter studies are warranted. Despite these limitations, our results provide important new information that can further our understanding of the epidemiology of psychiatric conditions in hospitalized children. This information should interest clinical teams caring for children with comorbid physical and mental health conditions.

Low FR–CD agreement regarding hospitalized children’s psychiatric comorbidities suggests that patients and their families and clinicians do not always share the same information about these comorbidities, and that the prevalence of psychiatric comorbidity in hospitalized children is likely underestimated. To allocate adequate resources for these children, health systems may need to obtain information from multiple sources. Furthermore, we need to better our understanding of strategies for communicating about hospitalized children’s psychiatric conditions so that we can develop interventions to improve hospital outcomes for this vulnerable population.

Disclosures

The direct costs of this project were funded by an internal pilot grant from the Center for Pediatric Health Disparities at Children’s Hospital of Philadelphia. Dr. Doupnik was supported by Ruth L. Kirschstein National Research Service Award T32-HP010026-11, funded by the National Institutes of Health. The sponsors had no role in study design; collection, analysis, or interpretation of data; manuscript writing; or deciding to submit this article for publication.

Recent publications in the medical literature and lay press have stirred controversy regarding the use of inpatient ‘very important person’ (VIP) services.^1-3 The term “VIP services” often refers to select conveniences offered in addition to the assumed basic level of care and services provided by a hospital. Examples include additional space, enhanced facilities, specific comforts, or personal support. In some instances, these amenities may only be provided to patients who have close financial, social, or professional relationships with the hospital.

How VIP patients interact with their health system to obtain VIP services has raised unique concerns. Some have speculated that the presence of a VIP patient may be disruptive to the care of non-VIP patients, while others have cautioned physicians about potential dangers to the VIP patients themselves.^4-6 Despite much being written on the topics of VIP patients and services in both the lay and academic press, our literature review identified only 1 study on the topic, which cataloged the preferential treatment of VIP patients in the emergency department.⁶ We are unaware of any investigations of VIP-service use in the inpatient setting. Through a multisite survey of hospital medicine physicians, we assessed physician viewpoints and behavior regarding VIP services.

The Hospital Medicine Reengineering Network (HOMERuN) is a nation-wide learning organization focused on measuring and improving the outcomes of hospitalized patients.⁷ We surveyed hospitalists from 8 HOMERuN hospitals (Appendix 1). The survey instrument contained 4 sections: nonidentifying respondent demographics, local use of VIP services, reported physician perceptions of VIP services, and case-based assessments (Appendix 2). Survey questions and individual cases were developed by study authors and based on real scenarios and concerns provided by front-line clinical providers. Content, length, and reliability of physician understanding were assessed by a 5-person focus group consisting of physicians not included in the survey population.

Subjects were identified via administrative rosters from each HOMERuN site. Surveys were administered via SurveyMonkey, and results were analyzed descriptively. Populations were compared via the Fisher exact test. “VIP services” were defined as conveniences provided in addition to the assumed basic level of care and services (eg, private or luxury-style rooms, access to a special menu, better views, dedicated personal care attendants, hospital liaisons). VIP patients were defined as those patients receiving VIP services. A hospital was identified as providing VIP services if 50% or more of respondents from that site reported the presence of VIP services.

Of 366 hospitalists contacted, 160 completed the survey (44%). Respondent characteristics and reported prevalence of VIP services are demonstrated in Table 1. In total, 78 respondents (45%) reported the presence of VIP services at their hospital. Of the 8 sites surveyed, a majority of physicians at 4 sites (50%) reported presence of VIP services.

Of respondents reporting the presence of VIP services at their hospital, a majority felt that, from a patient safety perspective, the care received by VIP patients was the same as care received by non-VIP patients (Table 2). A majority reported they had felt pressured by a VIP patient or a family member to order additional tests or treatments that the physician believed were medically unnecessary and that they would be more likely to comply with VIP patient’s requests for tests or treatments they felt were unnecessary. More than one-third (36%) felt pressured by other hospital employees or representatives to comply with VIP services patient’s requests for additional tests or treatments that the physicians believed were medically unnecessary.

In our study, a majority of physicians who reported the presence of VIP services at their hospital felt pressured by VIP patients or their family members to perform unnecessary testing or treatment. While this study was not designed to quantify the burden of unnecessary care for VIP patients, our results have implications for individual patients and public health, including potential effects on resource availability, the identification of clinically irrelevant incidental findings, and short- and long-term medical complications of procedures, testing and radiation exposure.

Prior publications have advocated that physicians and hospitals should not allow VIP status to influence management decisions.^3,5 We found that more than one-third of physicians who reported the presence of VIP services at their hospital also reported receiving pressure from hospital representatives to provide care to VIP patients that was not medically indicated. These findings highlight an example of the tension faced by physicians who are caught between patient requests and the delivery of value-based care. This potential conflict may be amplified particularly for those patients with close financial, social, or professional ties to the hospitals (and physicians) providing their care. These results suggest the need for physicians, administrators, and patients to work together to address the potential blurring of ethical boundaries created by VIP relationships. Prevention of harm and avoidance of placing physicians in morally distressing situations are common goals for all involved parties.

Efforts to reduce unnecessary care have predominantly focused on structural and knowledge-based drivers.^4,8,9 Our results highlight the presence of additional forces. A majority of physician respondents who reported the presence of VIP services at their hospital also reported that they would be more likely to comply with requests for unnecessary care for a VIP patient as compared to a non-VIP patient. Furthermore, in case-based questions about the requests of a VIP patient and their family for additional unnecessary care, a significant portion of physicians who reported they would comply with these requests listed the VIP status of the patient or family as a factor underlying this decision. Only a minority of physicians reported their decision to provide additional care was the result of their own medically-based concerns. Because these cases were hypothetical and we did not include comparator cases involving non-VIP patients, it remains uncertain whether the observed perceptions accurately reflect real-world differences in the care of VIP and non-VIP patients. Nonetheless, our findings emphasize the importance of better understanding the social drivers of overuse and physician communication strategies related to medically inappropriate tests.^10,11

Demand for unnecessary testing may be driven by the mentality that “more is better.”¹² Contrary to this belief, provision of unnecessary care can increase the risk of patient harm.¹³ Despite physician respondents reporting that VIP patients requested and/or received additional unnecessary care, a majority of respondents felt that patient safety for VIP patients was equivalent to that for non-VIP patients. As we assessed only physician perceptions of safety, which may not necessarily correlate with actual safety, further research in this area is needed.

Our study was limited by several factors. While our study population included hospitalists from 8 geographically broad hospitals, including university, safety net, and community hospitals, study responses may not be reflective of nationwide trends. Our response rate may limit our ability to generalize conclusions beyond respondents. Second, our study captured physician perceptions of behavior and safety rather than actually measuring practice and outcomes. Studies comparing physician practice patterns and outcomes between VIP and non-VIP patients would be informative. Additionally, despite our inclusive survey design process, our survey was not validated, and it is possible that our questions were not interpreted as intended. Lastly, despite the anonymous nature of our survey, physicians may have felt compelled to respond in a particular way due to conflicting professional, financial, or social factors.

Our findings provide initial insight into how care for the VIP patient may present unique challenges for physicians, hospitals, and society by systematizing care inequities, as well as potentially incentivizing low-value care practices. Whether these imbalances produce clinical harms or benefits remains worthy of future studies.

Disclosure

Nothing to report.

Recent publications in the medical literature and lay press have stirred controversy regarding the use of inpatient ‘very important person’ (VIP) services.^1-3 The term “VIP services” often refers to select conveniences offered in addition to the assumed basic level of care and services provided by a hospital. Examples include additional space, enhanced facilities, specific comforts, or personal support. In some instances, these amenities may only be provided to patients who have close financial, social, or professional relationships with the hospital.

How VIP patients interact with their health system to obtain VIP services has raised unique concerns. Some have speculated that the presence of a VIP patient may be disruptive to the care of non-VIP patients, while others have cautioned physicians about potential dangers to the VIP patients themselves.^4-6 Despite much being written on the topics of VIP patients and services in both the lay and academic press, our literature review identified only 1 study on the topic, which cataloged the preferential treatment of VIP patients in the emergency department.⁶ We are unaware of any investigations of VIP-service use in the inpatient setting. Through a multisite survey of hospital medicine physicians, we assessed physician viewpoints and behavior regarding VIP services.

The Hospital Medicine Reengineering Network (HOMERuN) is a nation-wide learning organization focused on measuring and improving the outcomes of hospitalized patients.⁷ We surveyed hospitalists from 8 HOMERuN hospitals (Appendix 1). The survey instrument contained 4 sections: nonidentifying respondent demographics, local use of VIP services, reported physician perceptions of VIP services, and case-based assessments (Appendix 2). Survey questions and individual cases were developed by study authors and based on real scenarios and concerns provided by front-line clinical providers. Content, length, and reliability of physician understanding were assessed by a 5-person focus group consisting of physicians not included in the survey population.

Subjects were identified via administrative rosters from each HOMERuN site. Surveys were administered via SurveyMonkey, and results were analyzed descriptively. Populations were compared via the Fisher exact test. “VIP services” were defined as conveniences provided in addition to the assumed basic level of care and services (eg, private or luxury-style rooms, access to a special menu, better views, dedicated personal care attendants, hospital liaisons). VIP patients were defined as those patients receiving VIP services. A hospital was identified as providing VIP services if 50% or more of respondents from that site reported the presence of VIP services.

Of 366 hospitalists contacted, 160 completed the survey (44%). Respondent characteristics and reported prevalence of VIP services are demonstrated in Table 1. In total, 78 respondents (45%) reported the presence of VIP services at their hospital. Of the 8 sites surveyed, a majority of physicians at 4 sites (50%) reported presence of VIP services.

Of respondents reporting the presence of VIP services at their hospital, a majority felt that, from a patient safety perspective, the care received by VIP patients was the same as care received by non-VIP patients (Table 2). A majority reported they had felt pressured by a VIP patient or a family member to order additional tests or treatments that the physician believed were medically unnecessary and that they would be more likely to comply with VIP patient’s requests for tests or treatments they felt were unnecessary. More than one-third (36%) felt pressured by other hospital employees or representatives to comply with VIP services patient’s requests for additional tests or treatments that the physicians believed were medically unnecessary.

In our study, a majority of physicians who reported the presence of VIP services at their hospital felt pressured by VIP patients or their family members to perform unnecessary testing or treatment. While this study was not designed to quantify the burden of unnecessary care for VIP patients, our results have implications for individual patients and public health, including potential effects on resource availability, the identification of clinically irrelevant incidental findings, and short- and long-term medical complications of procedures, testing and radiation exposure.

Prior publications have advocated that physicians and hospitals should not allow VIP status to influence management decisions.^3,5 We found that more than one-third of physicians who reported the presence of VIP services at their hospital also reported receiving pressure from hospital representatives to provide care to VIP patients that was not medically indicated. These findings highlight an example of the tension faced by physicians who are caught between patient requests and the delivery of value-based care. This potential conflict may be amplified particularly for those patients with close financial, social, or professional ties to the hospitals (and physicians) providing their care. These results suggest the need for physicians, administrators, and patients to work together to address the potential blurring of ethical boundaries created by VIP relationships. Prevention of harm and avoidance of placing physicians in morally distressing situations are common goals for all involved parties.

Efforts to reduce unnecessary care have predominantly focused on structural and knowledge-based drivers.^4,8,9 Our results highlight the presence of additional forces. A majority of physician respondents who reported the presence of VIP services at their hospital also reported that they would be more likely to comply with requests for unnecessary care for a VIP patient as compared to a non-VIP patient. Furthermore, in case-based questions about the requests of a VIP patient and their family for additional unnecessary care, a significant portion of physicians who reported they would comply with these requests listed the VIP status of the patient or family as a factor underlying this decision. Only a minority of physicians reported their decision to provide additional care was the result of their own medically-based concerns. Because these cases were hypothetical and we did not include comparator cases involving non-VIP patients, it remains uncertain whether the observed perceptions accurately reflect real-world differences in the care of VIP and non-VIP patients. Nonetheless, our findings emphasize the importance of better understanding the social drivers of overuse and physician communication strategies related to medically inappropriate tests.^10,11

Demand for unnecessary testing may be driven by the mentality that “more is better.”¹² Contrary to this belief, provision of unnecessary care can increase the risk of patient harm.¹³ Despite physician respondents reporting that VIP patients requested and/or received additional unnecessary care, a majority of respondents felt that patient safety for VIP patients was equivalent to that for non-VIP patients. As we assessed only physician perceptions of safety, which may not necessarily correlate with actual safety, further research in this area is needed.

Our study was limited by several factors. While our study population included hospitalists from 8 geographically broad hospitals, including university, safety net, and community hospitals, study responses may not be reflective of nationwide trends. Our response rate may limit our ability to generalize conclusions beyond respondents. Second, our study captured physician perceptions of behavior and safety rather than actually measuring practice and outcomes. Studies comparing physician practice patterns and outcomes between VIP and non-VIP patients would be informative. Additionally, despite our inclusive survey design process, our survey was not validated, and it is possible that our questions were not interpreted as intended. Lastly, despite the anonymous nature of our survey, physicians may have felt compelled to respond in a particular way due to conflicting professional, financial, or social factors.

Our findings provide initial insight into how care for the VIP patient may present unique challenges for physicians, hospitals, and society by systematizing care inequities, as well as potentially incentivizing low-value care practices. Whether these imbalances produce clinical harms or benefits remains worthy of future studies.

Disclosure

Nothing to report.

Recent publications in the medical literature and lay press have stirred controversy regarding the use of inpatient ‘very important person’ (VIP) services.^1-3 The term “VIP services” often refers to select conveniences offered in addition to the assumed basic level of care and services provided by a hospital. Examples include additional space, enhanced facilities, specific comforts, or personal support. In some instances, these amenities may only be provided to patients who have close financial, social, or professional relationships with the hospital.

How VIP patients interact with their health system to obtain VIP services has raised unique concerns. Some have speculated that the presence of a VIP patient may be disruptive to the care of non-VIP patients, while others have cautioned physicians about potential dangers to the VIP patients themselves.^4-6 Despite much being written on the topics of VIP patients and services in both the lay and academic press, our literature review identified only 1 study on the topic, which cataloged the preferential treatment of VIP patients in the emergency department.⁶ We are unaware of any investigations of VIP-service use in the inpatient setting. Through a multisite survey of hospital medicine physicians, we assessed physician viewpoints and behavior regarding VIP services.

The Hospital Medicine Reengineering Network (HOMERuN) is a nation-wide learning organization focused on measuring and improving the outcomes of hospitalized patients.⁷ We surveyed hospitalists from 8 HOMERuN hospitals (Appendix 1). The survey instrument contained 4 sections: nonidentifying respondent demographics, local use of VIP services, reported physician perceptions of VIP services, and case-based assessments (Appendix 2). Survey questions and individual cases were developed by study authors and based on real scenarios and concerns provided by front-line clinical providers. Content, length, and reliability of physician understanding were assessed by a 5-person focus group consisting of physicians not included in the survey population.

Subjects were identified via administrative rosters from each HOMERuN site. Surveys were administered via SurveyMonkey, and results were analyzed descriptively. Populations were compared via the Fisher exact test. “VIP services” were defined as conveniences provided in addition to the assumed basic level of care and services (eg, private or luxury-style rooms, access to a special menu, better views, dedicated personal care attendants, hospital liaisons). VIP patients were defined as those patients receiving VIP services. A hospital was identified as providing VIP services if 50% or more of respondents from that site reported the presence of VIP services.

Of 366 hospitalists contacted, 160 completed the survey (44%). Respondent characteristics and reported prevalence of VIP services are demonstrated in Table 1. In total, 78 respondents (45%) reported the presence of VIP services at their hospital. Of the 8 sites surveyed, a majority of physicians at 4 sites (50%) reported presence of VIP services.

Of respondents reporting the presence of VIP services at their hospital, a majority felt that, from a patient safety perspective, the care received by VIP patients was the same as care received by non-VIP patients (Table 2). A majority reported they had felt pressured by a VIP patient or a family member to order additional tests or treatments that the physician believed were medically unnecessary and that they would be more likely to comply with VIP patient’s requests for tests or treatments they felt were unnecessary. More than one-third (36%) felt pressured by other hospital employees or representatives to comply with VIP services patient’s requests for additional tests or treatments that the physicians believed were medically unnecessary.

In our study, a majority of physicians who reported the presence of VIP services at their hospital felt pressured by VIP patients or their family members to perform unnecessary testing or treatment. While this study was not designed to quantify the burden of unnecessary care for VIP patients, our results have implications for individual patients and public health, including potential effects on resource availability, the identification of clinically irrelevant incidental findings, and short- and long-term medical complications of procedures, testing and radiation exposure.

Prior publications have advocated that physicians and hospitals should not allow VIP status to influence management decisions.^3,5 We found that more than one-third of physicians who reported the presence of VIP services at their hospital also reported receiving pressure from hospital representatives to provide care to VIP patients that was not medically indicated. These findings highlight an example of the tension faced by physicians who are caught between patient requests and the delivery of value-based care. This potential conflict may be amplified particularly for those patients with close financial, social, or professional ties to the hospitals (and physicians) providing their care. These results suggest the need for physicians, administrators, and patients to work together to address the potential blurring of ethical boundaries created by VIP relationships. Prevention of harm and avoidance of placing physicians in morally distressing situations are common goals for all involved parties.

Efforts to reduce unnecessary care have predominantly focused on structural and knowledge-based drivers.^4,8,9 Our results highlight the presence of additional forces. A majority of physician respondents who reported the presence of VIP services at their hospital also reported that they would be more likely to comply with requests for unnecessary care for a VIP patient as compared to a non-VIP patient. Furthermore, in case-based questions about the requests of a VIP patient and their family for additional unnecessary care, a significant portion of physicians who reported they would comply with these requests listed the VIP status of the patient or family as a factor underlying this decision. Only a minority of physicians reported their decision to provide additional care was the result of their own medically-based concerns. Because these cases were hypothetical and we did not include comparator cases involving non-VIP patients, it remains uncertain whether the observed perceptions accurately reflect real-world differences in the care of VIP and non-VIP patients. Nonetheless, our findings emphasize the importance of better understanding the social drivers of overuse and physician communication strategies related to medically inappropriate tests.^10,11

Demand for unnecessary testing may be driven by the mentality that “more is better.”¹² Contrary to this belief, provision of unnecessary care can increase the risk of patient harm.¹³ Despite physician respondents reporting that VIP patients requested and/or received additional unnecessary care, a majority of respondents felt that patient safety for VIP patients was equivalent to that for non-VIP patients. As we assessed only physician perceptions of safety, which may not necessarily correlate with actual safety, further research in this area is needed.

Our study was limited by several factors. While our study population included hospitalists from 8 geographically broad hospitals, including university, safety net, and community hospitals, study responses may not be reflective of nationwide trends. Our response rate may limit our ability to generalize conclusions beyond respondents. Second, our study captured physician perceptions of behavior and safety rather than actually measuring practice and outcomes. Studies comparing physician practice patterns and outcomes between VIP and non-VIP patients would be informative. Additionally, despite our inclusive survey design process, our survey was not validated, and it is possible that our questions were not interpreted as intended. Lastly, despite the anonymous nature of our survey, physicians may have felt compelled to respond in a particular way due to conflicting professional, financial, or social factors.

Our findings provide initial insight into how care for the VIP patient may present unique challenges for physicians, hospitals, and society by systematizing care inequities, as well as potentially incentivizing low-value care practices. Whether these imbalances produce clinical harms or benefits remains worthy of future studies.

Disclosure

Nothing to report.

Using pharmacists to obtain admission medication histories (AMHs) reduces medication errors by 70% to 83% and resultant adverse drug events (ADEs) by 15%.^1-3 Dissemination of this practice has been limited by several factors, including clinician practice models, staff availability, confusion in provider roles and accountability, and absence of standardized best practices.^4-5 This paper assesses one of these barriers: the high cost of utilizing pharmacists. Third-person observer time and motion analysis shows that pharmacists require 46 and 92 minutes to obtain AMHs from medical and geriatric patients,⁶ respectively, resulting in pharmacist costs of $44 to $88 per patient, based on 2015 US Bureau of Labor Statistics (BLS) hourly wage data for pharmacists ($57.34).⁷

Pharmacist-supervised pharmacy technicians (PSPTs) achieve AMH accuracy comparable to pharmacists,^8-9 but their hourly wages are only 26% of pharmacists’.⁷ We conducted a third-person observer time and motion study¹⁰ to compare the amount of time and labor cost necessary for pharmacists and PSPTs to obtain AMHs for patients at high risk for ADEs.

This study originated as part of a randomized, controlled trial conducted during January-February 2014 at Cedars-Sinai Medical Center (CSMC), an 896-bed, university-affiliated, not-for-profit hospital.⁹ Pharmacy staff included pharmacists, PGY-1 pharmacy residents, and pharmacy technicians, each of whom received standardized didactic and experiential training (Appendix 1).

The pharmacists’ AMH and general pharmacy experience ranged from <1 to 3 years and <1 to 5 years, respectively. For PSPTs, AMH and general pharmacy experience ranged from <1 to 2 years and 1 to 17 years, respectively. Three additional pharmacists were involved in supervising PSPTs, and their experience fell within the aforementioned ranges, except for one pharmacist with general pharmacy experience of 16 years. The CSMC Institutional Review Board approved this study with oral consent from pharmacy staff.

For the trial, pharmacists and PSPTs obtained AMHs from 185 patients identified as high-risk for ADEs in the CSMC Emergency Department (ED). Patients were randomized into each arm using RANDI2 software¹¹ if they met one of the trial inclusion criteria, accessed via electronic health record (EHR) (Appendix 2). For several days during this trial, a trained research nurse shadowed pharmacists and PSPTs to record tasks performed, as well as the actual time, including start and end times, dedicated to each task.

After excluding AMHs with incomplete data, we calculated mean AMH times and component task times (Table). We compared mean times for pharmacists and PSPTs using two sample t tests (Table). We calculated mean times of tasks across only AMHs that required the task, mean times of tasks across all AMHs studied, regardless of whether the AMH required the task or not (assigning 0 minutes for the task if it was not required), and percent mean time of task per patient for providers combined (Table).

Of the 37 observed AMHs, 30 had complete data. Seven AMHs were excluded because not all task times were recorded, due to the schedule restraints of the research nurse. Pharmacists and PSPTs obtained 12 and 18 AMHs, respectively. Mean patient ages were 83.3 (95% confidence interval [CI], 77.3-89.2) and 79.8 (95% CI, 71.5-88.0), for pharmacists and PSPTs, respectively (P = 0.55). Patient’s EHRs contained a mean of 14.3 (95% CI, 11.2-17.5) and 16.3 (95% CI, 13.2-19.5) medications, prior to pharmacists and PSPTs obtaining an AMH, respectively (P = 0.41).

The mean time pharmacists and PSPTs needed to obtain an AMH was 58.5 (95% CI, 46.9-70.1) and 79.4 (95% CI, 59.1-99.8) minutes, respectively (P = 0.14). Summary time data per provider is reported in the Figure. The mean time for pharmacist supervision of technicians was 26 (95% CI, 14.9-37.1) minutes. Mean times of tasks and comparisons of these means times between providers are reported in the Table. The percent mean time for each task per patient for providers combined is also reported in the Table, in which utilizing the EHR was associated with the greatest percentage of time spent at 42.8% (95% CI, 37.4-48.2).

Although limited by sample size, we observed no difference in time or costs of obtaining AMHs between pharmacists and PSPTs. Several prior studies reported that pharmacists and technicians needed less time to obtain AMHs (20-40 minutes), as compared with our findings.^12-14 However, most prior studies used younger, healthier patients. Additionally, they used clinician self-reporting instead of third-person observer time and motion methodology. Indeed, the pharmacist times we observed in this study were consistent with prior findings⁶ that used accepted third-person observer time and motion methodology.¹⁰

We observed more variation in time to obtain AMHs among PSPTs than among pharmacists. While variation may be at least in part to the greater number of technicians studied, variation also points to the need for training and oversight of PSPTs. Selection of PSPTs with prior experience interacting with patients and functioning with higher levels of autonomy, standardized training of PSPTs, and consistent dedication of trained PSPTs to AMH functions to maintain their skills, may help to minimize such variation.

Limitations include the use of a single center and a small sample size. As such, the study may be underpowered to demonstrate statistically significant differences between providers. Furthermore, 7 AMHs (19%) had to be excluded because complete task times were missing. This was exclusively because the workday of the research nurse ended before the AMH had been completed. Another limitation was that the tasks observed could have been dissected further to identify even more specific factors that could be targeted to decrease AMH times. We recommend that future studies be larger, investigate in more depth various factors associated with time needed to obtain AMHs, consider which patients would most likely benefit from PSPTs, and use a measure of value (eg, number of history errors prevented/dollar spent).

In summary, we found that PSPTs can obtain AMHs for similar cost to pharmacists. It will be especially important to know whether PSPTs maintain the accuracy documented in prior studies.^8-9 If that continues to be the case, we expect our findings to allow many hospitals to implement programs using PSPTs to obtain accurate AMHs.

Acknowledgment

The authors thank Katherine M. Abdel-Razek for her role in data collection.

Disclosure

This research was supported by NIH/National Center for Advancing Translational Science UCLA CTSI Grant Number KL2TR000122 and National Institute on Aging Grant Number K23 AG049181-01 (Pevnick). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The investigators retained full independence in the conduct of this research.

Using pharmacists to obtain admission medication histories (AMHs) reduces medication errors by 70% to 83% and resultant adverse drug events (ADEs) by 15%.^1-3 Dissemination of this practice has been limited by several factors, including clinician practice models, staff availability, confusion in provider roles and accountability, and absence of standardized best practices.^4-5 This paper assesses one of these barriers: the high cost of utilizing pharmacists. Third-person observer time and motion analysis shows that pharmacists require 46 and 92 minutes to obtain AMHs from medical and geriatric patients,⁶ respectively, resulting in pharmacist costs of $44 to $88 per patient, based on 2015 US Bureau of Labor Statistics (BLS) hourly wage data for pharmacists ($57.34).⁷

Pharmacist-supervised pharmacy technicians (PSPTs) achieve AMH accuracy comparable to pharmacists,^8-9 but their hourly wages are only 26% of pharmacists’.⁷ We conducted a third-person observer time and motion study¹⁰ to compare the amount of time and labor cost necessary for pharmacists and PSPTs to obtain AMHs for patients at high risk for ADEs.

This study originated as part of a randomized, controlled trial conducted during January-February 2014 at Cedars-Sinai Medical Center (CSMC), an 896-bed, university-affiliated, not-for-profit hospital.⁹ Pharmacy staff included pharmacists, PGY-1 pharmacy residents, and pharmacy technicians, each of whom received standardized didactic and experiential training (Appendix 1).

The pharmacists’ AMH and general pharmacy experience ranged from <1 to 3 years and <1 to 5 years, respectively. For PSPTs, AMH and general pharmacy experience ranged from <1 to 2 years and 1 to 17 years, respectively. Three additional pharmacists were involved in supervising PSPTs, and their experience fell within the aforementioned ranges, except for one pharmacist with general pharmacy experience of 16 years. The CSMC Institutional Review Board approved this study with oral consent from pharmacy staff.

For the trial, pharmacists and PSPTs obtained AMHs from 185 patients identified as high-risk for ADEs in the CSMC Emergency Department (ED). Patients were randomized into each arm using RANDI2 software¹¹ if they met one of the trial inclusion criteria, accessed via electronic health record (EHR) (Appendix 2). For several days during this trial, a trained research nurse shadowed pharmacists and PSPTs to record tasks performed, as well as the actual time, including start and end times, dedicated to each task.

After excluding AMHs with incomplete data, we calculated mean AMH times and component task times (Table). We compared mean times for pharmacists and PSPTs using two sample t tests (Table). We calculated mean times of tasks across only AMHs that required the task, mean times of tasks across all AMHs studied, regardless of whether the AMH required the task or not (assigning 0 minutes for the task if it was not required), and percent mean time of task per patient for providers combined (Table).

Of the 37 observed AMHs, 30 had complete data. Seven AMHs were excluded because not all task times were recorded, due to the schedule restraints of the research nurse. Pharmacists and PSPTs obtained 12 and 18 AMHs, respectively. Mean patient ages were 83.3 (95% confidence interval [CI], 77.3-89.2) and 79.8 (95% CI, 71.5-88.0), for pharmacists and PSPTs, respectively (P = 0.55). Patient’s EHRs contained a mean of 14.3 (95% CI, 11.2-17.5) and 16.3 (95% CI, 13.2-19.5) medications, prior to pharmacists and PSPTs obtaining an AMH, respectively (P = 0.41).

The mean time pharmacists and PSPTs needed to obtain an AMH was 58.5 (95% CI, 46.9-70.1) and 79.4 (95% CI, 59.1-99.8) minutes, respectively (P = 0.14). Summary time data per provider is reported in the Figure. The mean time for pharmacist supervision of technicians was 26 (95% CI, 14.9-37.1) minutes. Mean times of tasks and comparisons of these means times between providers are reported in the Table. The percent mean time for each task per patient for providers combined is also reported in the Table, in which utilizing the EHR was associated with the greatest percentage of time spent at 42.8% (95% CI, 37.4-48.2).

Although limited by sample size, we observed no difference in time or costs of obtaining AMHs between pharmacists and PSPTs. Several prior studies reported that pharmacists and technicians needed less time to obtain AMHs (20-40 minutes), as compared with our findings.^12-14 However, most prior studies used younger, healthier patients. Additionally, they used clinician self-reporting instead of third-person observer time and motion methodology. Indeed, the pharmacist times we observed in this study were consistent with prior findings⁶ that used accepted third-person observer time and motion methodology.¹⁰

We observed more variation in time to obtain AMHs among PSPTs than among pharmacists. While variation may be at least in part to the greater number of technicians studied, variation also points to the need for training and oversight of PSPTs. Selection of PSPTs with prior experience interacting with patients and functioning with higher levels of autonomy, standardized training of PSPTs, and consistent dedication of trained PSPTs to AMH functions to maintain their skills, may help to minimize such variation.

Limitations include the use of a single center and a small sample size. As such, the study may be underpowered to demonstrate statistically significant differences between providers. Furthermore, 7 AMHs (19%) had to be excluded because complete task times were missing. This was exclusively because the workday of the research nurse ended before the AMH had been completed. Another limitation was that the tasks observed could have been dissected further to identify even more specific factors that could be targeted to decrease AMH times. We recommend that future studies be larger, investigate in more depth various factors associated with time needed to obtain AMHs, consider which patients would most likely benefit from PSPTs, and use a measure of value (eg, number of history errors prevented/dollar spent).

In summary, we found that PSPTs can obtain AMHs for similar cost to pharmacists. It will be especially important to know whether PSPTs maintain the accuracy documented in prior studies.^8-9 If that continues to be the case, we expect our findings to allow many hospitals to implement programs using PSPTs to obtain accurate AMHs.

Acknowledgment

The authors thank Katherine M. Abdel-Razek for her role in data collection.

Disclosure

This research was supported by NIH/National Center for Advancing Translational Science UCLA CTSI Grant Number KL2TR000122 and National Institute on Aging Grant Number K23 AG049181-01 (Pevnick). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The investigators retained full independence in the conduct of this research.

Using pharmacists to obtain admission medication histories (AMHs) reduces medication errors by 70% to 83% and resultant adverse drug events (ADEs) by 15%.^1-3 Dissemination of this practice has been limited by several factors, including clinician practice models, staff availability, confusion in provider roles and accountability, and absence of standardized best practices.^4-5 This paper assesses one of these barriers: the high cost of utilizing pharmacists. Third-person observer time and motion analysis shows that pharmacists require 46 and 92 minutes to obtain AMHs from medical and geriatric patients,⁶ respectively, resulting in pharmacist costs of $44 to $88 per patient, based on 2015 US Bureau of Labor Statistics (BLS) hourly wage data for pharmacists ($57.34).⁷

Pharmacist-supervised pharmacy technicians (PSPTs) achieve AMH accuracy comparable to pharmacists,^8-9 but their hourly wages are only 26% of pharmacists’.⁷ We conducted a third-person observer time and motion study¹⁰ to compare the amount of time and labor cost necessary for pharmacists and PSPTs to obtain AMHs for patients at high risk for ADEs.

This study originated as part of a randomized, controlled trial conducted during January-February 2014 at Cedars-Sinai Medical Center (CSMC), an 896-bed, university-affiliated, not-for-profit hospital.⁹ Pharmacy staff included pharmacists, PGY-1 pharmacy residents, and pharmacy technicians, each of whom received standardized didactic and experiential training (Appendix 1).

The pharmacists’ AMH and general pharmacy experience ranged from <1 to 3 years and <1 to 5 years, respectively. For PSPTs, AMH and general pharmacy experience ranged from <1 to 2 years and 1 to 17 years, respectively. Three additional pharmacists were involved in supervising PSPTs, and their experience fell within the aforementioned ranges, except for one pharmacist with general pharmacy experience of 16 years. The CSMC Institutional Review Board approved this study with oral consent from pharmacy staff.

For the trial, pharmacists and PSPTs obtained AMHs from 185 patients identified as high-risk for ADEs in the CSMC Emergency Department (ED). Patients were randomized into each arm using RANDI2 software¹¹ if they met one of the trial inclusion criteria, accessed via electronic health record (EHR) (Appendix 2). For several days during this trial, a trained research nurse shadowed pharmacists and PSPTs to record tasks performed, as well as the actual time, including start and end times, dedicated to each task.

After excluding AMHs with incomplete data, we calculated mean AMH times and component task times (Table). We compared mean times for pharmacists and PSPTs using two sample t tests (Table). We calculated mean times of tasks across only AMHs that required the task, mean times of tasks across all AMHs studied, regardless of whether the AMH required the task or not (assigning 0 minutes for the task if it was not required), and percent mean time of task per patient for providers combined (Table).

Of the 37 observed AMHs, 30 had complete data. Seven AMHs were excluded because not all task times were recorded, due to the schedule restraints of the research nurse. Pharmacists and PSPTs obtained 12 and 18 AMHs, respectively. Mean patient ages were 83.3 (95% confidence interval [CI], 77.3-89.2) and 79.8 (95% CI, 71.5-88.0), for pharmacists and PSPTs, respectively (P = 0.55). Patient’s EHRs contained a mean of 14.3 (95% CI, 11.2-17.5) and 16.3 (95% CI, 13.2-19.5) medications, prior to pharmacists and PSPTs obtaining an AMH, respectively (P = 0.41).

The mean time pharmacists and PSPTs needed to obtain an AMH was 58.5 (95% CI, 46.9-70.1) and 79.4 (95% CI, 59.1-99.8) minutes, respectively (P = 0.14). Summary time data per provider is reported in the Figure. The mean time for pharmacist supervision of technicians was 26 (95% CI, 14.9-37.1) minutes. Mean times of tasks and comparisons of these means times between providers are reported in the Table. The percent mean time for each task per patient for providers combined is also reported in the Table, in which utilizing the EHR was associated with the greatest percentage of time spent at 42.8% (95% CI, 37.4-48.2).

Although limited by sample size, we observed no difference in time or costs of obtaining AMHs between pharmacists and PSPTs. Several prior studies reported that pharmacists and technicians needed less time to obtain AMHs (20-40 minutes), as compared with our findings.^12-14 However, most prior studies used younger, healthier patients. Additionally, they used clinician self-reporting instead of third-person observer time and motion methodology. Indeed, the pharmacist times we observed in this study were consistent with prior findings⁶ that used accepted third-person observer time and motion methodology.¹⁰

We observed more variation in time to obtain AMHs among PSPTs than among pharmacists. While variation may be at least in part to the greater number of technicians studied, variation also points to the need for training and oversight of PSPTs. Selection of PSPTs with prior experience interacting with patients and functioning with higher levels of autonomy, standardized training of PSPTs, and consistent dedication of trained PSPTs to AMH functions to maintain their skills, may help to minimize such variation.

Limitations include the use of a single center and a small sample size. As such, the study may be underpowered to demonstrate statistically significant differences between providers. Furthermore, 7 AMHs (19%) had to be excluded because complete task times were missing. This was exclusively because the workday of the research nurse ended before the AMH had been completed. Another limitation was that the tasks observed could have been dissected further to identify even more specific factors that could be targeted to decrease AMH times. We recommend that future studies be larger, investigate in more depth various factors associated with time needed to obtain AMHs, consider which patients would most likely benefit from PSPTs, and use a measure of value (eg, number of history errors prevented/dollar spent).

In summary, we found that PSPTs can obtain AMHs for similar cost to pharmacists. It will be especially important to know whether PSPTs maintain the accuracy documented in prior studies.^8-9 If that continues to be the case, we expect our findings to allow many hospitals to implement programs using PSPTs to obtain accurate AMHs.

Acknowledgment

The authors thank Katherine M. Abdel-Razek for her role in data collection.

Disclosure

This research was supported by NIH/National Center for Advancing Translational Science UCLA CTSI Grant Number KL2TR000122 and National Institute on Aging Grant Number K23 AG049181-01 (Pevnick). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The investigators retained full independence in the conduct of this research.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

CASE REPORT

A 68-year-old woman with ischemic cardiomyopathy was admitted with abdominal cramping, diarrhea, and nausea, which had left her unable to keep food and liquids down for 2 days. She had been taking diuretics and had a remote history of intravenous drug use. On admission, she was afebrile and had blood pressure of 100/60 mm Hg and a heart rate of 100 bpm. Her extremities were cool and clammy. Blood test results showed an alanine aminotransferase (ALT) level of 1510 IU/L and an aspartate aminotransferase (AST) level of 1643 IU/L. The patient’s clinician did not know her baseline ALT and AST levels and thought the best approach was to identify the cause of the transaminase elevation.

Severe acute liver injury (liver enzymes, >10 × upper limit of normal [ULN], usually 40 IU/L) is a common presentation among hospitalized patients. Between 1997 and 2015, 1.5% of patients admitted to our hospital had severe liver injury. In another large cohort of hospitalized patients,¹ 0.6% had an ALT level higher than 1000 IU/L (~20 × ULN). A precise diagnosis is often needed to direct appropriate therapy, and serologic tests are available for many conditions, both common and rare (Table). Given the relative ease of bundled blood testing, nondirected testing has emerged as a popular, if reflexive, strategy.^2-5 In this approach, clinicians evaluate each patient for the set of testable diseases all at once—in contrast to taking a directed, stepwise testing approach guided by the patient’s history.

Use of nondirected testing is common in patients with severe acute liver injury. Of the 5795 such patients treated at our hospital between 2000 and 2015, within the same day of service 53% were tested for hepatitis C virus antibody, 38% for hemochromatosis (ferritin test), 28% for autoimmune hepatitis (antinuclear antibody test), and 15% for primary biliary cholangitis (antimitochondrial antibody test) by our clinical laboratory. Of the 5023 patients who had send-out tests performed for Wilson disease (ceruloplasmin), 81% were queried for hepatitis B virus infection, 76% for hepatitis C virus infection, 75% for autoimmune hepatitis, and 73.1% for hemochromatosis.² Similar trends were found for patients with severe liver injury tested for α₁-antitrypsin (AAT) deficiency.³ In sum, these data showed that each patient with severe liver injury was tested out of concern about diseases with markedly different epidemiology and clinical presentations (Table).

WHY YOU MIGHT THINK NONDIRECTED TESTING IS HELPFUL

Use of nondirected testing may reflect perceived urgency, convenience, and thoroughness.^2-6 Alternatively, it may simply involve following a consultant’s recommendations.⁴ As severe acute liver injury is often associated with tremendous morbidity, clinicians seeking answers may perceive directed, stepwise testing as inappropriately slow given the urgency of the presentation; they may think that nondirected testing can reduce hospital length of stay.

WHY NONDIRECTED TESTING IS NOT HELPFUL

Nondirected testing is a problem for at least 4 reasons: limited benefit of reflexive testing for rare diseases, no meaningful impact on outcomes, false positives, and financial cost.

First, immediately testing for rare causes of liver disease is unlikely to benefit patients with severe liver injury. The underlying etiologies of severe liver injury are relatively well circumscribed (Table). Overall, 42% of patients with severe liver injury and 57% of those with an ALT level higher than 1000 IU/L have ischemic hepatitis.⁷ Accounting for a significant percentage of severe liver injury cases are acute biliary obstruction (24%), drug-induced injury (10%-13%), and viral hepatitis (4%-7%).^1,8 Of the small subset of patients with severe liver injury that progresses to acute liver failure (ALF; encephalopathy, coagulopathy), 0.5% have autoimmune hepatitis and 0.1% have Wilson disease.⁹ Furthermore, many patients are tested for AAT deficiency, hemochromatosis, and primary biliary cholangitis, but these are never causes of severe acute liver injury (Table).

Second, diagnosing a rarer cause of acute liver injury modestly earlier has no meaningful impact on outcome. Work-up for more common etiologies can usually be completed within 2 or 3 days. This is true even for patients with ALF. Specific therapies generally are lacking for ALF, save for use of N-acetylcysteine for acetaminophen overdose and antiviral therapy for hepatitis B virus infection.^9,10 Furthermore, although effective therapies are available for both autoimmune hepatitis and Wilson disease, the potential benefit stems from altering the longer term course of disease. Initial management, even for these rare conditions, is no different from that for other etiologies. Conversely, acute liver injury caused by ischemic hepatitis, biliary disease, or drug-induced liver injury requires swift corrective action. Even if normotensive, patients with ischemic hepatitis are often in cardiogenic shock and benefit from careful monitoring and critical care.⁷ Patients with acute biliary obstruction may need therapeutic endoscopy. Last, patients with drug-induced liver injury benefit from immediate discontinuation of the offending drug.

Third, in the testing of patients with low pretest probabilities, false positives are common. For example, at our institution and at an institution in Austria, severe liver injury patients with a low ceruloplasmin level have a 95.1% to 98.1% chance of a false-positive result (they have a low ceruloplasmin level but do not have Wilson disease).^3,4 Furthermore, 91% of positive tests are never confirmed,³ indicating either that clinicians never valued the initial test or that other diagnoses were much more likely. Even worse, as was the case in 65% of patients with low AAT levels,^2,3 genetic diagnoses were based on unconfirmed, potentially false-positive serologic tests.

Fourth, although the financial cost for each individual test is small, at the population level the cost of nondirected testing is significant. For example, although each reflects testing for conditions that do not cause acute liver injury, the costs of ferritin, AAT, and antimitochondrial antibody tests are $13, $16, and $37, respectively (Medicare/Medicaid reimbursements in 2016 $US).¹¹ About 1.5% of admitted patients are found to have severe liver injury. If this proportion holds true for the roughly 40 million discharges from US hospitals each year, then there would be an annual cost of about $40 million if all 3 tests were performed for each patient with severe liver injury. In addition, although nondirected testing may seem clinically expedient, there are no data suggesting it reduces length of stay. In fact, ceruloplasmin, AAT, and many other tests are sent to external laboratories and are unlikely to be returned before discharge. If clinicians delay discharge for results, then nondirected testing would increase rather than decrease length of stay.

WHAT YOU SHOULD DO INSTEAD

In this era of increasing cost-consciousness, nondirected testing has escaped relatively unscathed. Indeed, nondirected testing is prevalent, yet has pitfalls similar to those of serologic testing (eg, vasculitis or arthritis,⁶ acute renal injury, infectious disease¹²). The alternative is deliberate, empirical, patient-centered testing that is attentive to the patient’s presentation and the harms of false positives. The idea is to select tests for each patient with acute liver injury according to presentation and the most likely corresponding diagnoses (Table, Figure).

• For each patient with severe acute liver injury, select tests on the basis of the presentation (Figure). Testing for rare diseases should be performed only after common diseases have been excluded.
• Avoid testing for hemochromatosis (iron indices, genetic tests), AAT deficiency (AAT levels or phenotypes), and primary biliary cholangitis (antimitochondrial antibodies) in patients with severe acute liver injury.
• Consider implementing decision supports that can curb nondirected testing in areas in which it is common.

Nondirected testing is associated with false positives and increased costs in the evaluation and management of severe acute liver injury. The alternative is deliberate, epidemiologically and clinically driven directed testing. Electronic ordering system decision supports can be useful in curtailing nondirected testing.

Disclosure

Nothing to report.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

CASE REPORT

A 68-year-old woman with ischemic cardiomyopathy was admitted with abdominal cramping, diarrhea, and nausea, which had left her unable to keep food and liquids down for 2 days. She had been taking diuretics and had a remote history of intravenous drug use. On admission, she was afebrile and had blood pressure of 100/60 mm Hg and a heart rate of 100 bpm. Her extremities were cool and clammy. Blood test results showed an alanine aminotransferase (ALT) level of 1510 IU/L and an aspartate aminotransferase (AST) level of 1643 IU/L. The patient’s clinician did not know her baseline ALT and AST levels and thought the best approach was to identify the cause of the transaminase elevation.

Severe acute liver injury (liver enzymes, >10 × upper limit of normal [ULN], usually 40 IU/L) is a common presentation among hospitalized patients. Between 1997 and 2015, 1.5% of patients admitted to our hospital had severe liver injury. In another large cohort of hospitalized patients,¹ 0.6% had an ALT level higher than 1000 IU/L (~20 × ULN). A precise diagnosis is often needed to direct appropriate therapy, and serologic tests are available for many conditions, both common and rare (Table). Given the relative ease of bundled blood testing, nondirected testing has emerged as a popular, if reflexive, strategy.^2-5 In this approach, clinicians evaluate each patient for the set of testable diseases all at once—in contrast to taking a directed, stepwise testing approach guided by the patient’s history.

Use of nondirected testing is common in patients with severe acute liver injury. Of the 5795 such patients treated at our hospital between 2000 and 2015, within the same day of service 53% were tested for hepatitis C virus antibody, 38% for hemochromatosis (ferritin test), 28% for autoimmune hepatitis (antinuclear antibody test), and 15% for primary biliary cholangitis (antimitochondrial antibody test) by our clinical laboratory. Of the 5023 patients who had send-out tests performed for Wilson disease (ceruloplasmin), 81% were queried for hepatitis B virus infection, 76% for hepatitis C virus infection, 75% for autoimmune hepatitis, and 73.1% for hemochromatosis.² Similar trends were found for patients with severe liver injury tested for α₁-antitrypsin (AAT) deficiency.³ In sum, these data showed that each patient with severe liver injury was tested out of concern about diseases with markedly different epidemiology and clinical presentations (Table).

WHY YOU MIGHT THINK NONDIRECTED TESTING IS HELPFUL

Use of nondirected testing may reflect perceived urgency, convenience, and thoroughness.^2-6 Alternatively, it may simply involve following a consultant’s recommendations.⁴ As severe acute liver injury is often associated with tremendous morbidity, clinicians seeking answers may perceive directed, stepwise testing as inappropriately slow given the urgency of the presentation; they may think that nondirected testing can reduce hospital length of stay.

WHY NONDIRECTED TESTING IS NOT HELPFUL

Nondirected testing is a problem for at least 4 reasons: limited benefit of reflexive testing for rare diseases, no meaningful impact on outcomes, false positives, and financial cost.

First, immediately testing for rare causes of liver disease is unlikely to benefit patients with severe liver injury. The underlying etiologies of severe liver injury are relatively well circumscribed (Table). Overall, 42% of patients with severe liver injury and 57% of those with an ALT level higher than 1000 IU/L have ischemic hepatitis.⁷ Accounting for a significant percentage of severe liver injury cases are acute biliary obstruction (24%), drug-induced injury (10%-13%), and viral hepatitis (4%-7%).^1,8 Of the small subset of patients with severe liver injury that progresses to acute liver failure (ALF; encephalopathy, coagulopathy), 0.5% have autoimmune hepatitis and 0.1% have Wilson disease.⁹ Furthermore, many patients are tested for AAT deficiency, hemochromatosis, and primary biliary cholangitis, but these are never causes of severe acute liver injury (Table).

Second, diagnosing a rarer cause of acute liver injury modestly earlier has no meaningful impact on outcome. Work-up for more common etiologies can usually be completed within 2 or 3 days. This is true even for patients with ALF. Specific therapies generally are lacking for ALF, save for use of N-acetylcysteine for acetaminophen overdose and antiviral therapy for hepatitis B virus infection.^9,10 Furthermore, although effective therapies are available for both autoimmune hepatitis and Wilson disease, the potential benefit stems from altering the longer term course of disease. Initial management, even for these rare conditions, is no different from that for other etiologies. Conversely, acute liver injury caused by ischemic hepatitis, biliary disease, or drug-induced liver injury requires swift corrective action. Even if normotensive, patients with ischemic hepatitis are often in cardiogenic shock and benefit from careful monitoring and critical care.⁷ Patients with acute biliary obstruction may need therapeutic endoscopy. Last, patients with drug-induced liver injury benefit from immediate discontinuation of the offending drug.

Third, in the testing of patients with low pretest probabilities, false positives are common. For example, at our institution and at an institution in Austria, severe liver injury patients with a low ceruloplasmin level have a 95.1% to 98.1% chance of a false-positive result (they have a low ceruloplasmin level but do not have Wilson disease).^3,4 Furthermore, 91% of positive tests are never confirmed,³ indicating either that clinicians never valued the initial test or that other diagnoses were much more likely. Even worse, as was the case in 65% of patients with low AAT levels,^2,3 genetic diagnoses were based on unconfirmed, potentially false-positive serologic tests.

Fourth, although the financial cost for each individual test is small, at the population level the cost of nondirected testing is significant. For example, although each reflects testing for conditions that do not cause acute liver injury, the costs of ferritin, AAT, and antimitochondrial antibody tests are $13, $16, and $37, respectively (Medicare/Medicaid reimbursements in 2016 $US).¹¹ About 1.5% of admitted patients are found to have severe liver injury. If this proportion holds true for the roughly 40 million discharges from US hospitals each year, then there would be an annual cost of about $40 million if all 3 tests were performed for each patient with severe liver injury. In addition, although nondirected testing may seem clinically expedient, there are no data suggesting it reduces length of stay. In fact, ceruloplasmin, AAT, and many other tests are sent to external laboratories and are unlikely to be returned before discharge. If clinicians delay discharge for results, then nondirected testing would increase rather than decrease length of stay.

WHAT YOU SHOULD DO INSTEAD

In this era of increasing cost-consciousness, nondirected testing has escaped relatively unscathed. Indeed, nondirected testing is prevalent, yet has pitfalls similar to those of serologic testing (eg, vasculitis or arthritis,⁶ acute renal injury, infectious disease¹²). The alternative is deliberate, empirical, patient-centered testing that is attentive to the patient’s presentation and the harms of false positives. The idea is to select tests for each patient with acute liver injury according to presentation and the most likely corresponding diagnoses (Table, Figure).

• For each patient with severe acute liver injury, select tests on the basis of the presentation (Figure). Testing for rare diseases should be performed only after common diseases have been excluded.
• Avoid testing for hemochromatosis (iron indices, genetic tests), AAT deficiency (AAT levels or phenotypes), and primary biliary cholangitis (antimitochondrial antibodies) in patients with severe acute liver injury.
• Consider implementing decision supports that can curb nondirected testing in areas in which it is common.

Nondirected testing is associated with false positives and increased costs in the evaluation and management of severe acute liver injury. The alternative is deliberate, epidemiologically and clinically driven directed testing. Electronic ordering system decision supports can be useful in curtailing nondirected testing.

Disclosure

Nothing to report.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

CASE REPORT

A 68-year-old woman with ischemic cardiomyopathy was admitted with abdominal cramping, diarrhea, and nausea, which had left her unable to keep food and liquids down for 2 days. She had been taking diuretics and had a remote history of intravenous drug use. On admission, she was afebrile and had blood pressure of 100/60 mm Hg and a heart rate of 100 bpm. Her extremities were cool and clammy. Blood test results showed an alanine aminotransferase (ALT) level of 1510 IU/L and an aspartate aminotransferase (AST) level of 1643 IU/L. The patient’s clinician did not know her baseline ALT and AST levels and thought the best approach was to identify the cause of the transaminase elevation.

Severe acute liver injury (liver enzymes, >10 × upper limit of normal [ULN], usually 40 IU/L) is a common presentation among hospitalized patients. Between 1997 and 2015, 1.5% of patients admitted to our hospital had severe liver injury. In another large cohort of hospitalized patients,¹ 0.6% had an ALT level higher than 1000 IU/L (~20 × ULN). A precise diagnosis is often needed to direct appropriate therapy, and serologic tests are available for many conditions, both common and rare (Table). Given the relative ease of bundled blood testing, nondirected testing has emerged as a popular, if reflexive, strategy.^2-5 In this approach, clinicians evaluate each patient for the set of testable diseases all at once—in contrast to taking a directed, stepwise testing approach guided by the patient’s history.

Use of nondirected testing is common in patients with severe acute liver injury. Of the 5795 such patients treated at our hospital between 2000 and 2015, within the same day of service 53% were tested for hepatitis C virus antibody, 38% for hemochromatosis (ferritin test), 28% for autoimmune hepatitis (antinuclear antibody test), and 15% for primary biliary cholangitis (antimitochondrial antibody test) by our clinical laboratory. Of the 5023 patients who had send-out tests performed for Wilson disease (ceruloplasmin), 81% were queried for hepatitis B virus infection, 76% for hepatitis C virus infection, 75% for autoimmune hepatitis, and 73.1% for hemochromatosis.² Similar trends were found for patients with severe liver injury tested for α₁-antitrypsin (AAT) deficiency.³ In sum, these data showed that each patient with severe liver injury was tested out of concern about diseases with markedly different epidemiology and clinical presentations (Table).

WHY YOU MIGHT THINK NONDIRECTED TESTING IS HELPFUL

Use of nondirected testing may reflect perceived urgency, convenience, and thoroughness.^2-6 Alternatively, it may simply involve following a consultant’s recommendations.⁴ As severe acute liver injury is often associated with tremendous morbidity, clinicians seeking answers may perceive directed, stepwise testing as inappropriately slow given the urgency of the presentation; they may think that nondirected testing can reduce hospital length of stay.

WHY NONDIRECTED TESTING IS NOT HELPFUL

Nondirected testing is a problem for at least 4 reasons: limited benefit of reflexive testing for rare diseases, no meaningful impact on outcomes, false positives, and financial cost.

First, immediately testing for rare causes of liver disease is unlikely to benefit patients with severe liver injury. The underlying etiologies of severe liver injury are relatively well circumscribed (Table). Overall, 42% of patients with severe liver injury and 57% of those with an ALT level higher than 1000 IU/L have ischemic hepatitis.⁷ Accounting for a significant percentage of severe liver injury cases are acute biliary obstruction (24%), drug-induced injury (10%-13%), and viral hepatitis (4%-7%).^1,8 Of the small subset of patients with severe liver injury that progresses to acute liver failure (ALF; encephalopathy, coagulopathy), 0.5% have autoimmune hepatitis and 0.1% have Wilson disease.⁹ Furthermore, many patients are tested for AAT deficiency, hemochromatosis, and primary biliary cholangitis, but these are never causes of severe acute liver injury (Table).

Second, diagnosing a rarer cause of acute liver injury modestly earlier has no meaningful impact on outcome. Work-up for more common etiologies can usually be completed within 2 or 3 days. This is true even for patients with ALF. Specific therapies generally are lacking for ALF, save for use of N-acetylcysteine for acetaminophen overdose and antiviral therapy for hepatitis B virus infection.^9,10 Furthermore, although effective therapies are available for both autoimmune hepatitis and Wilson disease, the potential benefit stems from altering the longer term course of disease. Initial management, even for these rare conditions, is no different from that for other etiologies. Conversely, acute liver injury caused by ischemic hepatitis, biliary disease, or drug-induced liver injury requires swift corrective action. Even if normotensive, patients with ischemic hepatitis are often in cardiogenic shock and benefit from careful monitoring and critical care.⁷ Patients with acute biliary obstruction may need therapeutic endoscopy. Last, patients with drug-induced liver injury benefit from immediate discontinuation of the offending drug.

Third, in the testing of patients with low pretest probabilities, false positives are common. For example, at our institution and at an institution in Austria, severe liver injury patients with a low ceruloplasmin level have a 95.1% to 98.1% chance of a false-positive result (they have a low ceruloplasmin level but do not have Wilson disease).^3,4 Furthermore, 91% of positive tests are never confirmed,³ indicating either that clinicians never valued the initial test or that other diagnoses were much more likely. Even worse, as was the case in 65% of patients with low AAT levels,^2,3 genetic diagnoses were based on unconfirmed, potentially false-positive serologic tests.

Fourth, although the financial cost for each individual test is small, at the population level the cost of nondirected testing is significant. For example, although each reflects testing for conditions that do not cause acute liver injury, the costs of ferritin, AAT, and antimitochondrial antibody tests are $13, $16, and $37, respectively (Medicare/Medicaid reimbursements in 2016 $US).¹¹ About 1.5% of admitted patients are found to have severe liver injury. If this proportion holds true for the roughly 40 million discharges from US hospitals each year, then there would be an annual cost of about $40 million if all 3 tests were performed for each patient with severe liver injury. In addition, although nondirected testing may seem clinically expedient, there are no data suggesting it reduces length of stay. In fact, ceruloplasmin, AAT, and many other tests are sent to external laboratories and are unlikely to be returned before discharge. If clinicians delay discharge for results, then nondirected testing would increase rather than decrease length of stay.

WHAT YOU SHOULD DO INSTEAD

In this era of increasing cost-consciousness, nondirected testing has escaped relatively unscathed. Indeed, nondirected testing is prevalent, yet has pitfalls similar to those of serologic testing (eg, vasculitis or arthritis,⁶ acute renal injury, infectious disease¹²). The alternative is deliberate, empirical, patient-centered testing that is attentive to the patient’s presentation and the harms of false positives. The idea is to select tests for each patient with acute liver injury according to presentation and the most likely corresponding diagnoses (Table, Figure).

• For each patient with severe acute liver injury, select tests on the basis of the presentation (Figure). Testing for rare diseases should be performed only after common diseases have been excluded.
• Avoid testing for hemochromatosis (iron indices, genetic tests), AAT deficiency (AAT levels or phenotypes), and primary biliary cholangitis (antimitochondrial antibodies) in patients with severe acute liver injury.
• Consider implementing decision supports that can curb nondirected testing in areas in which it is common.

Nondirected testing is associated with false positives and increased costs in the evaluation and management of severe acute liver injury. The alternative is deliberate, epidemiologically and clinically driven directed testing. Electronic ordering system decision supports can be useful in curtailing nondirected testing.

Disclosure

Nothing to report.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter.

A federal judge has blocked Texas from withholding Medicaid funds from Planned Parenthood centers in the state, ruling that state officials had no cause to terminate funding for the providers.

In his Feb. 21 decision, Judge Sam Sparks of the U.S. District Court for the Western District of Texas temporarily barred Texas from taking funds from Planned Parenthood while the case proceeds. Restricting the money would deprive Medicaid patients of their right to obtain health care from their chosen providers and would potentially disrupt care for 12,500 Medicaid patients, Judge Sparks wrote.

Texas Attorney General Ken Paxton (R) pledged to appeal the ruling.

“[The] decision is disappointing and flies in the face of basic human decency,” Mr. Paxton said in a statement. “The raw, unedited footage from undercover videos exposed a brazen willingness by Planned Parenthood officials to traffic in fetal body parts, as well as manipulate the timing and method of an abortion ... No taxpayer in Texas should have to subsidize this repugnant and illegal conduct. We should never lose sight of the fact that, as long as abortion is legal in the United States, the potential for these types of horrors will continue.”

Planned Parenthood representatives did not return messages seeking comment on the ruling.

The Texas Health and Human Services Commission (HHSC) terminated funding for its Planned Parenthood providers in the state following a 2015 video by two anti-abortion advocates that purported to show representatives of Planned Parenthood Gulf Coast (Houston) contracting to sell aborted human fetal tissue and body parts. State authorities investigated the facility and found no wrongdoing, but a grand jury indicted the two anti-abortion activists for using fake names and identities.

Separate investigations by the Texas Attorney General’s Office, the Texas Department of State Health Services, and HHSC also found no wrongdoing, according to court documents. In December 2016 however, HHSC sent termination letters to five Texas Planned Parenthood centers, stating that the facilities were not qualified to provide medical services in a “competent, safe, legal and ethical manner” under state and federal law, according to court records. The providers and several patients sued.

In his ruling, Judge Sparks said he found no evidence indicating that an actual program violation occurred warranting termination of funding for the providers.

“After reviewing the evidence currently in the record, the court finds the Inspector General, and thus HHSC, likely acted to disenroll qualified health care providers from Medicaid without cause,” Judge Sparks wrote. “The individual plaintiffs have met their burden to establish a likelihood of success on the merits. The Inspector General did not have prima facie of evidence, or even a scintilla of evidence, to conclude the basis of termination set forth in the final notice merited finding [the plaintiffs] were not qualified.”

Similar efforts were overturned in Virgina after Governor Terry McAuliffe (D) vetoed a bill that sought to restrict state and federal funding from Planned Parenthood providers in the state. In a statement, Gov. McAuliffe said the bill would have harmed Virginians who rely on health care services and programs provided by Planned Parenthood health centers by denying them access to affordable care.

Similar legislation to defund Planned Parenthood providers was introduced by Michigan lawmakers in February.

In addition, efforts are underway at the federal level. On Feb. 16, the House passed H.J.Res.43, which would allow states to withhold Title X family planning funds from providers that offer abortion services, overturning a rule put in place at the end of the Obama administration. The resolution passed 230-188, largely along party lines. It would strike down the Obama-era rule via the 1996 Congressional Review Act, which allows Congress to overturn new regulations within 60 days of their passage. H.J.Res.43 is currently before the Senate.

The American Congress of Obstetricians and Gynecologists expressed disappointment with the House resolution.

“The resolution allows states to discriminate against women’s health care providers for reasons unrelated to qualifications or best practices,” according to an ACOG statement. “Under this resolution, states could disqualify health centers, including Planned Parenthood, from providing Title X contraceptive and preventive care to over 4 million individuals. The Title X program is the only federal grant program exclusively dedicated to providing low-income patients with access to effective family planning and related preventive health services, including contraceptive care. Contraceptive access is essential to helping women achieve greater educational, financial, and professional success and stability. It’s critical to the economic success of this population.”

House Speaker Paul Ryan (R-Wis.) has promised that the bill to repeal the Affordable Care Act will include a measure stripping funds from Planned Parenthood.

[email protected]
On Twitter @legal_med

A federal judge has blocked Texas from withholding Medicaid funds from Planned Parenthood centers in the state, ruling that state officials had no cause to terminate funding for the providers.

In his Feb. 21 decision, Judge Sam Sparks of the U.S. District Court for the Western District of Texas temporarily barred Texas from taking funds from Planned Parenthood while the case proceeds. Restricting the money would deprive Medicaid patients of their right to obtain health care from their chosen providers and would potentially disrupt care for 12,500 Medicaid patients, Judge Sparks wrote.

Texas Attorney General Ken Paxton (R) pledged to appeal the ruling.

“[The] decision is disappointing and flies in the face of basic human decency,” Mr. Paxton said in a statement. “The raw, unedited footage from undercover videos exposed a brazen willingness by Planned Parenthood officials to traffic in fetal body parts, as well as manipulate the timing and method of an abortion ... No taxpayer in Texas should have to subsidize this repugnant and illegal conduct. We should never lose sight of the fact that, as long as abortion is legal in the United States, the potential for these types of horrors will continue.”

Planned Parenthood representatives did not return messages seeking comment on the ruling.

The Texas Health and Human Services Commission (HHSC) terminated funding for its Planned Parenthood providers in the state following a 2015 video by two anti-abortion advocates that purported to show representatives of Planned Parenthood Gulf Coast (Houston) contracting to sell aborted human fetal tissue and body parts. State authorities investigated the facility and found no wrongdoing, but a grand jury indicted the two anti-abortion activists for using fake names and identities.

Separate investigations by the Texas Attorney General’s Office, the Texas Department of State Health Services, and HHSC also found no wrongdoing, according to court documents. In December 2016 however, HHSC sent termination letters to five Texas Planned Parenthood centers, stating that the facilities were not qualified to provide medical services in a “competent, safe, legal and ethical manner” under state and federal law, according to court records. The providers and several patients sued.

In his ruling, Judge Sparks said he found no evidence indicating that an actual program violation occurred warranting termination of funding for the providers.

“After reviewing the evidence currently in the record, the court finds the Inspector General, and thus HHSC, likely acted to disenroll qualified health care providers from Medicaid without cause,” Judge Sparks wrote. “The individual plaintiffs have met their burden to establish a likelihood of success on the merits. The Inspector General did not have prima facie of evidence, or even a scintilla of evidence, to conclude the basis of termination set forth in the final notice merited finding [the plaintiffs] were not qualified.”

Similar efforts were overturned in Virgina after Governor Terry McAuliffe (D) vetoed a bill that sought to restrict state and federal funding from Planned Parenthood providers in the state. In a statement, Gov. McAuliffe said the bill would have harmed Virginians who rely on health care services and programs provided by Planned Parenthood health centers by denying them access to affordable care.

Similar legislation to defund Planned Parenthood providers was introduced by Michigan lawmakers in February.

In addition, efforts are underway at the federal level. On Feb. 16, the House passed H.J.Res.43, which would allow states to withhold Title X family planning funds from providers that offer abortion services, overturning a rule put in place at the end of the Obama administration. The resolution passed 230-188, largely along party lines. It would strike down the Obama-era rule via the 1996 Congressional Review Act, which allows Congress to overturn new regulations within 60 days of their passage. H.J.Res.43 is currently before the Senate.

The American Congress of Obstetricians and Gynecologists expressed disappointment with the House resolution.

“The resolution allows states to discriminate against women’s health care providers for reasons unrelated to qualifications or best practices,” according to an ACOG statement. “Under this resolution, states could disqualify health centers, including Planned Parenthood, from providing Title X contraceptive and preventive care to over 4 million individuals. The Title X program is the only federal grant program exclusively dedicated to providing low-income patients with access to effective family planning and related preventive health services, including contraceptive care. Contraceptive access is essential to helping women achieve greater educational, financial, and professional success and stability. It’s critical to the economic success of this population.”

House Speaker Paul Ryan (R-Wis.) has promised that the bill to repeal the Affordable Care Act will include a measure stripping funds from Planned Parenthood.

[email protected]
On Twitter @legal_med

A federal judge has blocked Texas from withholding Medicaid funds from Planned Parenthood centers in the state, ruling that state officials had no cause to terminate funding for the providers.

In his Feb. 21 decision, Judge Sam Sparks of the U.S. District Court for the Western District of Texas temporarily barred Texas from taking funds from Planned Parenthood while the case proceeds. Restricting the money would deprive Medicaid patients of their right to obtain health care from their chosen providers and would potentially disrupt care for 12,500 Medicaid patients, Judge Sparks wrote.

Texas Attorney General Ken Paxton (R) pledged to appeal the ruling.

“[The] decision is disappointing and flies in the face of basic human decency,” Mr. Paxton said in a statement. “The raw, unedited footage from undercover videos exposed a brazen willingness by Planned Parenthood officials to traffic in fetal body parts, as well as manipulate the timing and method of an abortion ... No taxpayer in Texas should have to subsidize this repugnant and illegal conduct. We should never lose sight of the fact that, as long as abortion is legal in the United States, the potential for these types of horrors will continue.”

Planned Parenthood representatives did not return messages seeking comment on the ruling.

The Texas Health and Human Services Commission (HHSC) terminated funding for its Planned Parenthood providers in the state following a 2015 video by two anti-abortion advocates that purported to show representatives of Planned Parenthood Gulf Coast (Houston) contracting to sell aborted human fetal tissue and body parts. State authorities investigated the facility and found no wrongdoing, but a grand jury indicted the two anti-abortion activists for using fake names and identities.

Separate investigations by the Texas Attorney General’s Office, the Texas Department of State Health Services, and HHSC also found no wrongdoing, according to court documents. In December 2016 however, HHSC sent termination letters to five Texas Planned Parenthood centers, stating that the facilities were not qualified to provide medical services in a “competent, safe, legal and ethical manner” under state and federal law, according to court records. The providers and several patients sued.

In his ruling, Judge Sparks said he found no evidence indicating that an actual program violation occurred warranting termination of funding for the providers.

“After reviewing the evidence currently in the record, the court finds the Inspector General, and thus HHSC, likely acted to disenroll qualified health care providers from Medicaid without cause,” Judge Sparks wrote. “The individual plaintiffs have met their burden to establish a likelihood of success on the merits. The Inspector General did not have prima facie of evidence, or even a scintilla of evidence, to conclude the basis of termination set forth in the final notice merited finding [the plaintiffs] were not qualified.”

Similar efforts were overturned in Virgina after Governor Terry McAuliffe (D) vetoed a bill that sought to restrict state and federal funding from Planned Parenthood providers in the state. In a statement, Gov. McAuliffe said the bill would have harmed Virginians who rely on health care services and programs provided by Planned Parenthood health centers by denying them access to affordable care.

Similar legislation to defund Planned Parenthood providers was introduced by Michigan lawmakers in February.

In addition, efforts are underway at the federal level. On Feb. 16, the House passed H.J.Res.43, which would allow states to withhold Title X family planning funds from providers that offer abortion services, overturning a rule put in place at the end of the Obama administration. The resolution passed 230-188, largely along party lines. It would strike down the Obama-era rule via the 1996 Congressional Review Act, which allows Congress to overturn new regulations within 60 days of their passage. H.J.Res.43 is currently before the Senate.

The American Congress of Obstetricians and Gynecologists expressed disappointment with the House resolution.

“The resolution allows states to discriminate against women’s health care providers for reasons unrelated to qualifications or best practices,” according to an ACOG statement. “Under this resolution, states could disqualify health centers, including Planned Parenthood, from providing Title X contraceptive and preventive care to over 4 million individuals. The Title X program is the only federal grant program exclusively dedicated to providing low-income patients with access to effective family planning and related preventive health services, including contraceptive care. Contraceptive access is essential to helping women achieve greater educational, financial, and professional success and stability. It’s critical to the economic success of this population.”

House Speaker Paul Ryan (R-Wis.) has promised that the bill to repeal the Affordable Care Act will include a measure stripping funds from Planned Parenthood.

[email protected]
On Twitter @legal_med