Transforming Healthcare

In 2001, the Institute of Medicine called for a major redesign of the US healthcare system, describing the chasm between the quality of care Americans receive and the quality of healthcare they deserve.[1] The healthcare community recognizes its ongoing quality and value gaps, but progress has been limited by outdated care models, fragmented organizational structures, and insufficient advances in system design.[2] Many healthcare organizations are searching for new care delivery models capable of producing greater value.

A major constraint in hospitals is the persistence of underperforming frontline clinical care teams.[3] Physicians typically travel from 1 unit or patient to the next in unpredictable patterns, resulting in missed opportunities to share perspectives and coordinate care with nurses, discharge planning personnel, pharmacists, therapists, and patients. This geographic fragmentation almost certainly contributes to interprofessional silos and hierarchies, nonspecific care plans, and failure to initiate or intensify therapy when indicated.[4] Modern hospital units could benefit from having a standard care model that synchronizes frontline professionals into teams routinely coordinating and progressing a shared plan of care.

EFFECTIVE CLINICAL MICROSYSTEMS REFLECTED IN THE DESIGN OF THE ACCOUNTABLE CARE UNIT

High‐value healthcare organizations deliberately design clinical microsystems.[5] An effective clinical microsystem combines several traits: (1) a small group of people who work together in a defined setting on a regular basis to provide care, (2) linked care processes and a shared information environment that includes individuals who receive that care, (3) performance outcomes, and (4) set service and care aims.[6] For the accountable care unit (ACU) to reflect the traits of an effective clinical microsystem, we designed it with analogous features: (1) unit‐based teams, (2) structured interdisciplinary bedside rounds (SIBR), (3) unit‐level performance reporting, and (4) unit‐level nurse and physician coleadership. We launched the ACU on September 1, 2010 in a high‐acuity 24‐bed medical unit at Emory University Hospital, a 579‐bed tertiary academic medical center. Herein we provide a brief report of our experience implementing and refining the ACU over a 4‐year period to help others gauge feasibility and sustainability.

FEATURES OF AN ACU

Unit‐Based Teams

Design

Geographic alignment fosters mutual respect, cohesiveness, communication, timeliness, and face‐to‐face problem solving,[7, 8] and has been linked to improved patient satisfaction, decreased length of stay, and reductions in morbidity and mortality.[9, 10, 11] At our hospital, though, patients newly admitted or transferred to the hospital medicine service traditionally had been distributed to physician teams without regard to geography, typically based on physician call schedules or traditions of balancing patient volumes across colleagues. These traditional practices geographically dispersed our teams. Physicians would be forced regularly to travel to 5 to 8 different units each day to see 10 to 18 patients. Nurses might perceive this as a parade of different physician teams coming and going off the unit at unpredictable times. To temporally and spatially align physicians with unit‐based staff, specific physician teams were assigned to the ACU.

Implementation

The first step in implementing unit‐based teams was to identify the smallest number of physician teams that could be assigned to the ACU. Two internal medicine resident teams are assigned to care for all medical patients in the unit. Each resident team consists of 1 hospital medicine attending physician, 1 internal medicine resident, 3 interns (2 covering the day shift and 1 overnight every other night), and up to 2 medical students. The 2 teams alternate a 24‐hour call cycle where the on‐call team admits every patient arriving to the unit. For patients arriving to the unit from 6 _pm to 7 _am, the on‐call overnight intern admits the patients and hands over care to the team in the morning. The on‐call team becomes aware of an incoming patient once the patient has been assigned a bed in the home unit. Several patients per day may arrive on the unit as transfers from a medical or surgical intensive care unit, but most patients arrive as emergency room or direct admissions. On any given day it is acceptable and typical for a team to have several patients off the ACU. No specific changes were made to nurse staffing, with the unit continuing to have 1 nurse unit manager, 1 charge nurse per shift, and a nurse‐to‐patient ratio of 1 to 4.

Results

Geographic patient assignment has been successful (Figure 1). Prior to implementing the ACU, more than 5 different hospital medicine physician teams cared for patients on the unit, with no single team caring for more than 25% of them. In the ACU, all medical patients are assigned to 1 of the 2 unit‐based physician teams (physician teams 1 and 2), which regularly represents more than 95% of all patients on the unit. Over the 4 years, these 2 ACU teams have had an average of 12.9 total patient encounters per day (compared to 11.8 in the year before the ACU when these teams were not unit based). The 2 unit‐based teams have over 90% of their patients on the ACU daily. In contrast, 3 attending‐only hospital medicine teams (physician teams 3, 4, and 5) are still dispersed over 6 to 8 units every day (Figure 2), primarily due to high hospital occupancy and a relative scarcity of units eligible to become dedicated hospital medicine units.

Figure 1

Figure 2

Effects of the Change

Through unit‐based teams, the ACU achieves the first trait of an effective clinical microsystem. Although an evaluation of the cultural gains are beyond the scope of this article, the logistical advantages are self‐evident; having the fewest necessary physician teams overseeing care for nearly all patients in 1 unit and where those physician teams simultaneously have nearly all of their patients on that 1 unit, makes it possible to schedule interdisciplinary teamwork activities, such as SIBR, not otherwise feasible.

Structured Interdisciplinary Bedside Rounds

Design

To reflect the second trait of an effective clinical microsystem, a hospital unit should routinely combine best practices for communication, including daily goals sheets,[12] safety checklists,[13] and multidisciplinary rounds.[14, 15] ACU design achieves this through SIBR, a patient‐ and family‐centered, team‐based approach to rounds that brings the nurse, physician, and available allied health professionals to the patient's bedside every day to exchange perspectives using a standard format to cross‐check information with the patient, family, and one another, and articulate a clear plan for the day. Before the SIBR hour starts, physicians and nurses have already performed independent patient assessments through usual activities such as handover, chart review, patient interviews, and physical examinations. Participants in SIBR are expected to give or receive inputs according to the standard SIBR communication protocol (Figure 3), review a quality‐safety checklist together, and ensure the plan of care is verbalized. Including the patient and family allows all parties to hear and be heard, cross‐check information for accuracy, and hold each person accountable for contributions.[16, 17]

Figure 3

Implementation

Each ACU staff member receives orientation to the SIBR communication protocol and is expected to be prepared and punctual for the midmorning start times. The charge nurse serves as the SIBR rounds manager, ensuring physicians waste no time searching for the next nurse and each team's eligible patients are seen in the SIBR hour. For each patient, SIBR begins when the nurse and physician are both present at the bedside. The intern begins SIBR by introducing team members before reviewing the patient's active problem list, response to treatment, and interval test results or consultant inputs. The nurse then relays the patient's goal for the day, overnight events, nursing concerns, and reviews the quality‐safety checklist. The intern then invites allied health professionals to share inputs that might impact medical decision making or discharge planning, before synthesizing all inputs into a shared plan for the day.

Throughout SIBR, the patient and family are encouraged to ask questions or correct misinformation. Although newcomers to SIBR often imagine that inviting patient inputs will disrupt efficiency, we have found teams readily learn to manage this risk, for instance discerning the core question among multiple seemingly disparate ones, or volunteering to return after the SIBR hour to explore a complex issue.

Results

Since the launch of the ACU on September 1, 2010, SIBR has been embedded as a routine on the unit with both physician teams and the nursing staff conducting it every day. Patients not considered eligible for SIBR are those whom the entire physician team has not yet evaluated, typically patients who arrived to the unit overnight. For patients who opt out due to personal preference, or for patients away from the unit for a procedure or a test, SIBR occurs without the patient so the rest of the team can still exchange inputs and formulate a plan of care. A visitor to the unit sees SIBR start punctually at 9 _am and 10 _am for successive teams, with each completing SIBR on eligible patients in under 60 minutes.

Effects of the Change

The second trait of an effective clinical microsystem is achieved through SIBR's routine forum for staff to share information with each other and the patient. By practicing SIBR every workday, staff are presented with multiple routine opportunities to experience an environment reflective of high‐performing frontline units.[18] We found that SIBR resembled other competencies, with a bell curve of performance. For this reason, by the start of the third year we added a SIBR certification program, a SIBR skills training program where permanent and rotating staff are evaluated through an in vivo observed structured clinical exam, typically with a charge nurse or physician as preceptor. When a nurse, medical student, intern, or resident demonstrates an ability to perform a series of specific high performance SIBR behaviors in 5 of 6 consecutive patients, they can achieve SIBR certification. In the first 2 years of this voluntary certification program, all daytime nursing staff and rotating interns have achieved this demonstration of interdisciplinary teamwork competence.

CHALLENGES

Challenges with implementing the ACU fell into 3 primary categories: (1) performing change management required for a successful launch, (2) solving logistics of maintaining unit‐based physician teams, and (3) training physicians and nurses to perform SIBR at a high level.

For change management, the leadership pair was able to explain the rationale of the model to all staff in sufficient detail to launch the ACU. To build momentum for ACU routines and relationships, the physician leader and the nurse unit manager were both present on the unit daily for the first 100 days. As ACU operations became routine and competencies formed among clinicians, the amount of time spent by these leaders was de‐escalated.

Creating and maintaining unit‐based physician teams required shared understanding and coordination between on‐call hospital medicine physicians and the bed control office so that new admissions or transfers could be consistently assigned to unit‐based teams without adversely affecting patient flow. We found this challenge to be manageable once stakeholders accepted the rationale for the care mode and figured out how to support it.

The challenge of building high‐performance SIBR across the unit, including competence of rotating trainees new to the model, requires individualized assessment and feedback necessary for SIBR certification. We addressed this challenge by creating a SIBR train‐the‐trainer programa list of observable high‐performance SIBR behaviors coupled with a short course about giving effective feedback to learnersand found that once the ACU had several nurse and physician SIBR trainers in the staffing mix every day, the required amount of SIBR coaching expertise was available when needed.

CONCLUSION

Improving value and reliability in hospital care may require new models of care. The ACU is a hospital care model specifically designed to organize physicians, nurses, and allied health professionals into high‐functioning, unit‐based teams. It converges standard workflow, patient‐centered communication, quality‐safety checklists, best‐practice protocols, performance measurement, and progressive leadership. Our experience with the ACU suggests that hospital units can be reorganized as effective clinical microsystems where consistent unit professionals can share time and space, a sense of purpose, code of conduct, shared mental model for teamwork, an interprofessional management structure, and an important level of accountability to each other and their patients.

Disclosures: Jason Stein, MD: grant support from the US Health & Resources Services Administration to support organizational implementation of the care model described; recipient of consulting fees and royalties for licensed intellectual property to support implementation of the care model described; founder and president of nonprofit Centripital, provider of consulting services to hospital systems implementing the care model described. The terms of this arrangement have been reviewed and approved by Emory University in accordance with its conflict of interest policies. Liam Chadwick, PhD, and Diaz Clark, MS, RN: recipients of consulting fees through Centripital to support implementation of the care model described. Bryan W. Castle, MBA, RN: grant support from the US Health & Resources Services Administration to support organizational implementation of the care model described; recipient of consulting fees through Centripital to support implementation of the care model described. The authors report no other conflicts of interest.

In 2001, the Institute of Medicine called for a major redesign of the US healthcare system, describing the chasm between the quality of care Americans receive and the quality of healthcare they deserve.[1] The healthcare community recognizes its ongoing quality and value gaps, but progress has been limited by outdated care models, fragmented organizational structures, and insufficient advances in system design.[2] Many healthcare organizations are searching for new care delivery models capable of producing greater value.

A major constraint in hospitals is the persistence of underperforming frontline clinical care teams.[3] Physicians typically travel from 1 unit or patient to the next in unpredictable patterns, resulting in missed opportunities to share perspectives and coordinate care with nurses, discharge planning personnel, pharmacists, therapists, and patients. This geographic fragmentation almost certainly contributes to interprofessional silos and hierarchies, nonspecific care plans, and failure to initiate or intensify therapy when indicated.[4] Modern hospital units could benefit from having a standard care model that synchronizes frontline professionals into teams routinely coordinating and progressing a shared plan of care.

EFFECTIVE CLINICAL MICROSYSTEMS REFLECTED IN THE DESIGN OF THE ACCOUNTABLE CARE UNIT

High‐value healthcare organizations deliberately design clinical microsystems.[5] An effective clinical microsystem combines several traits: (1) a small group of people who work together in a defined setting on a regular basis to provide care, (2) linked care processes and a shared information environment that includes individuals who receive that care, (3) performance outcomes, and (4) set service and care aims.[6] For the accountable care unit (ACU) to reflect the traits of an effective clinical microsystem, we designed it with analogous features: (1) unit‐based teams, (2) structured interdisciplinary bedside rounds (SIBR), (3) unit‐level performance reporting, and (4) unit‐level nurse and physician coleadership. We launched the ACU on September 1, 2010 in a high‐acuity 24‐bed medical unit at Emory University Hospital, a 579‐bed tertiary academic medical center. Herein we provide a brief report of our experience implementing and refining the ACU over a 4‐year period to help others gauge feasibility and sustainability.

FEATURES OF AN ACU

Unit‐Based Teams

Design

Geographic alignment fosters mutual respect, cohesiveness, communication, timeliness, and face‐to‐face problem solving,[7, 8] and has been linked to improved patient satisfaction, decreased length of stay, and reductions in morbidity and mortality.[9, 10, 11] At our hospital, though, patients newly admitted or transferred to the hospital medicine service traditionally had been distributed to physician teams without regard to geography, typically based on physician call schedules or traditions of balancing patient volumes across colleagues. These traditional practices geographically dispersed our teams. Physicians would be forced regularly to travel to 5 to 8 different units each day to see 10 to 18 patients. Nurses might perceive this as a parade of different physician teams coming and going off the unit at unpredictable times. To temporally and spatially align physicians with unit‐based staff, specific physician teams were assigned to the ACU.

Implementation

The first step in implementing unit‐based teams was to identify the smallest number of physician teams that could be assigned to the ACU. Two internal medicine resident teams are assigned to care for all medical patients in the unit. Each resident team consists of 1 hospital medicine attending physician, 1 internal medicine resident, 3 interns (2 covering the day shift and 1 overnight every other night), and up to 2 medical students. The 2 teams alternate a 24‐hour call cycle where the on‐call team admits every patient arriving to the unit. For patients arriving to the unit from 6 _pm to 7 _am, the on‐call overnight intern admits the patients and hands over care to the team in the morning. The on‐call team becomes aware of an incoming patient once the patient has been assigned a bed in the home unit. Several patients per day may arrive on the unit as transfers from a medical or surgical intensive care unit, but most patients arrive as emergency room or direct admissions. On any given day it is acceptable and typical for a team to have several patients off the ACU. No specific changes were made to nurse staffing, with the unit continuing to have 1 nurse unit manager, 1 charge nurse per shift, and a nurse‐to‐patient ratio of 1 to 4.

Results

Geographic patient assignment has been successful (Figure 1). Prior to implementing the ACU, more than 5 different hospital medicine physician teams cared for patients on the unit, with no single team caring for more than 25% of them. In the ACU, all medical patients are assigned to 1 of the 2 unit‐based physician teams (physician teams 1 and 2), which regularly represents more than 95% of all patients on the unit. Over the 4 years, these 2 ACU teams have had an average of 12.9 total patient encounters per day (compared to 11.8 in the year before the ACU when these teams were not unit based). The 2 unit‐based teams have over 90% of their patients on the ACU daily. In contrast, 3 attending‐only hospital medicine teams (physician teams 3, 4, and 5) are still dispersed over 6 to 8 units every day (Figure 2), primarily due to high hospital occupancy and a relative scarcity of units eligible to become dedicated hospital medicine units.

Figure 1

Figure 2

Effects of the Change

Through unit‐based teams, the ACU achieves the first trait of an effective clinical microsystem. Although an evaluation of the cultural gains are beyond the scope of this article, the logistical advantages are self‐evident; having the fewest necessary physician teams overseeing care for nearly all patients in 1 unit and where those physician teams simultaneously have nearly all of their patients on that 1 unit, makes it possible to schedule interdisciplinary teamwork activities, such as SIBR, not otherwise feasible.

Structured Interdisciplinary Bedside Rounds

Design

To reflect the second trait of an effective clinical microsystem, a hospital unit should routinely combine best practices for communication, including daily goals sheets,[12] safety checklists,[13] and multidisciplinary rounds.[14, 15] ACU design achieves this through SIBR, a patient‐ and family‐centered, team‐based approach to rounds that brings the nurse, physician, and available allied health professionals to the patient's bedside every day to exchange perspectives using a standard format to cross‐check information with the patient, family, and one another, and articulate a clear plan for the day. Before the SIBR hour starts, physicians and nurses have already performed independent patient assessments through usual activities such as handover, chart review, patient interviews, and physical examinations. Participants in SIBR are expected to give or receive inputs according to the standard SIBR communication protocol (Figure 3), review a quality‐safety checklist together, and ensure the plan of care is verbalized. Including the patient and family allows all parties to hear and be heard, cross‐check information for accuracy, and hold each person accountable for contributions.[16, 17]

Figure 3

Implementation

Each ACU staff member receives orientation to the SIBR communication protocol and is expected to be prepared and punctual for the midmorning start times. The charge nurse serves as the SIBR rounds manager, ensuring physicians waste no time searching for the next nurse and each team's eligible patients are seen in the SIBR hour. For each patient, SIBR begins when the nurse and physician are both present at the bedside. The intern begins SIBR by introducing team members before reviewing the patient's active problem list, response to treatment, and interval test results or consultant inputs. The nurse then relays the patient's goal for the day, overnight events, nursing concerns, and reviews the quality‐safety checklist. The intern then invites allied health professionals to share inputs that might impact medical decision making or discharge planning, before synthesizing all inputs into a shared plan for the day.

Throughout SIBR, the patient and family are encouraged to ask questions or correct misinformation. Although newcomers to SIBR often imagine that inviting patient inputs will disrupt efficiency, we have found teams readily learn to manage this risk, for instance discerning the core question among multiple seemingly disparate ones, or volunteering to return after the SIBR hour to explore a complex issue.

Results

Since the launch of the ACU on September 1, 2010, SIBR has been embedded as a routine on the unit with both physician teams and the nursing staff conducting it every day. Patients not considered eligible for SIBR are those whom the entire physician team has not yet evaluated, typically patients who arrived to the unit overnight. For patients who opt out due to personal preference, or for patients away from the unit for a procedure or a test, SIBR occurs without the patient so the rest of the team can still exchange inputs and formulate a plan of care. A visitor to the unit sees SIBR start punctually at 9 _am and 10 _am for successive teams, with each completing SIBR on eligible patients in under 60 minutes.

Effects of the Change

The second trait of an effective clinical microsystem is achieved through SIBR's routine forum for staff to share information with each other and the patient. By practicing SIBR every workday, staff are presented with multiple routine opportunities to experience an environment reflective of high‐performing frontline units.[18] We found that SIBR resembled other competencies, with a bell curve of performance. For this reason, by the start of the third year we added a SIBR certification program, a SIBR skills training program where permanent and rotating staff are evaluated through an in vivo observed structured clinical exam, typically with a charge nurse or physician as preceptor. When a nurse, medical student, intern, or resident demonstrates an ability to perform a series of specific high performance SIBR behaviors in 5 of 6 consecutive patients, they can achieve SIBR certification. In the first 2 years of this voluntary certification program, all daytime nursing staff and rotating interns have achieved this demonstration of interdisciplinary teamwork competence.

CHALLENGES

Challenges with implementing the ACU fell into 3 primary categories: (1) performing change management required for a successful launch, (2) solving logistics of maintaining unit‐based physician teams, and (3) training physicians and nurses to perform SIBR at a high level.

For change management, the leadership pair was able to explain the rationale of the model to all staff in sufficient detail to launch the ACU. To build momentum for ACU routines and relationships, the physician leader and the nurse unit manager were both present on the unit daily for the first 100 days. As ACU operations became routine and competencies formed among clinicians, the amount of time spent by these leaders was de‐escalated.

Creating and maintaining unit‐based physician teams required shared understanding and coordination between on‐call hospital medicine physicians and the bed control office so that new admissions or transfers could be consistently assigned to unit‐based teams without adversely affecting patient flow. We found this challenge to be manageable once stakeholders accepted the rationale for the care mode and figured out how to support it.

The challenge of building high‐performance SIBR across the unit, including competence of rotating trainees new to the model, requires individualized assessment and feedback necessary for SIBR certification. We addressed this challenge by creating a SIBR train‐the‐trainer programa list of observable high‐performance SIBR behaviors coupled with a short course about giving effective feedback to learnersand found that once the ACU had several nurse and physician SIBR trainers in the staffing mix every day, the required amount of SIBR coaching expertise was available when needed.

CONCLUSION

Improving value and reliability in hospital care may require new models of care. The ACU is a hospital care model specifically designed to organize physicians, nurses, and allied health professionals into high‐functioning, unit‐based teams. It converges standard workflow, patient‐centered communication, quality‐safety checklists, best‐practice protocols, performance measurement, and progressive leadership. Our experience with the ACU suggests that hospital units can be reorganized as effective clinical microsystems where consistent unit professionals can share time and space, a sense of purpose, code of conduct, shared mental model for teamwork, an interprofessional management structure, and an important level of accountability to each other and their patients.

Disclosures: Jason Stein, MD: grant support from the US Health & Resources Services Administration to support organizational implementation of the care model described; recipient of consulting fees and royalties for licensed intellectual property to support implementation of the care model described; founder and president of nonprofit Centripital, provider of consulting services to hospital systems implementing the care model described. The terms of this arrangement have been reviewed and approved by Emory University in accordance with its conflict of interest policies. Liam Chadwick, PhD, and Diaz Clark, MS, RN: recipients of consulting fees through Centripital to support implementation of the care model described. Bryan W. Castle, MBA, RN: grant support from the US Health & Resources Services Administration to support organizational implementation of the care model described; recipient of consulting fees through Centripital to support implementation of the care model described. The authors report no other conflicts of interest.

In 2001, the Institute of Medicine called for a major redesign of the US healthcare system, describing the chasm between the quality of care Americans receive and the quality of healthcare they deserve.[1] The healthcare community recognizes its ongoing quality and value gaps, but progress has been limited by outdated care models, fragmented organizational structures, and insufficient advances in system design.[2] Many healthcare organizations are searching for new care delivery models capable of producing greater value.

A major constraint in hospitals is the persistence of underperforming frontline clinical care teams.[3] Physicians typically travel from 1 unit or patient to the next in unpredictable patterns, resulting in missed opportunities to share perspectives and coordinate care with nurses, discharge planning personnel, pharmacists, therapists, and patients. This geographic fragmentation almost certainly contributes to interprofessional silos and hierarchies, nonspecific care plans, and failure to initiate or intensify therapy when indicated.[4] Modern hospital units could benefit from having a standard care model that synchronizes frontline professionals into teams routinely coordinating and progressing a shared plan of care.

EFFECTIVE CLINICAL MICROSYSTEMS REFLECTED IN THE DESIGN OF THE ACCOUNTABLE CARE UNIT

High‐value healthcare organizations deliberately design clinical microsystems.[5] An effective clinical microsystem combines several traits: (1) a small group of people who work together in a defined setting on a regular basis to provide care, (2) linked care processes and a shared information environment that includes individuals who receive that care, (3) performance outcomes, and (4) set service and care aims.[6] For the accountable care unit (ACU) to reflect the traits of an effective clinical microsystem, we designed it with analogous features: (1) unit‐based teams, (2) structured interdisciplinary bedside rounds (SIBR), (3) unit‐level performance reporting, and (4) unit‐level nurse and physician coleadership. We launched the ACU on September 1, 2010 in a high‐acuity 24‐bed medical unit at Emory University Hospital, a 579‐bed tertiary academic medical center. Herein we provide a brief report of our experience implementing and refining the ACU over a 4‐year period to help others gauge feasibility and sustainability.

FEATURES OF AN ACU

Unit‐Based Teams

Design

Geographic alignment fosters mutual respect, cohesiveness, communication, timeliness, and face‐to‐face problem solving,[7, 8] and has been linked to improved patient satisfaction, decreased length of stay, and reductions in morbidity and mortality.[9, 10, 11] At our hospital, though, patients newly admitted or transferred to the hospital medicine service traditionally had been distributed to physician teams without regard to geography, typically based on physician call schedules or traditions of balancing patient volumes across colleagues. These traditional practices geographically dispersed our teams. Physicians would be forced regularly to travel to 5 to 8 different units each day to see 10 to 18 patients. Nurses might perceive this as a parade of different physician teams coming and going off the unit at unpredictable times. To temporally and spatially align physicians with unit‐based staff, specific physician teams were assigned to the ACU.

Implementation

The first step in implementing unit‐based teams was to identify the smallest number of physician teams that could be assigned to the ACU. Two internal medicine resident teams are assigned to care for all medical patients in the unit. Each resident team consists of 1 hospital medicine attending physician, 1 internal medicine resident, 3 interns (2 covering the day shift and 1 overnight every other night), and up to 2 medical students. The 2 teams alternate a 24‐hour call cycle where the on‐call team admits every patient arriving to the unit. For patients arriving to the unit from 6 _pm to 7 _am, the on‐call overnight intern admits the patients and hands over care to the team in the morning. The on‐call team becomes aware of an incoming patient once the patient has been assigned a bed in the home unit. Several patients per day may arrive on the unit as transfers from a medical or surgical intensive care unit, but most patients arrive as emergency room or direct admissions. On any given day it is acceptable and typical for a team to have several patients off the ACU. No specific changes were made to nurse staffing, with the unit continuing to have 1 nurse unit manager, 1 charge nurse per shift, and a nurse‐to‐patient ratio of 1 to 4.

Results

Geographic patient assignment has been successful (Figure 1). Prior to implementing the ACU, more than 5 different hospital medicine physician teams cared for patients on the unit, with no single team caring for more than 25% of them. In the ACU, all medical patients are assigned to 1 of the 2 unit‐based physician teams (physician teams 1 and 2), which regularly represents more than 95% of all patients on the unit. Over the 4 years, these 2 ACU teams have had an average of 12.9 total patient encounters per day (compared to 11.8 in the year before the ACU when these teams were not unit based). The 2 unit‐based teams have over 90% of their patients on the ACU daily. In contrast, 3 attending‐only hospital medicine teams (physician teams 3, 4, and 5) are still dispersed over 6 to 8 units every day (Figure 2), primarily due to high hospital occupancy and a relative scarcity of units eligible to become dedicated hospital medicine units.

Figure 1

Figure 2

Effects of the Change

Through unit‐based teams, the ACU achieves the first trait of an effective clinical microsystem. Although an evaluation of the cultural gains are beyond the scope of this article, the logistical advantages are self‐evident; having the fewest necessary physician teams overseeing care for nearly all patients in 1 unit and where those physician teams simultaneously have nearly all of their patients on that 1 unit, makes it possible to schedule interdisciplinary teamwork activities, such as SIBR, not otherwise feasible.

Structured Interdisciplinary Bedside Rounds

Design

To reflect the second trait of an effective clinical microsystem, a hospital unit should routinely combine best practices for communication, including daily goals sheets,[12] safety checklists,[13] and multidisciplinary rounds.[14, 15] ACU design achieves this through SIBR, a patient‐ and family‐centered, team‐based approach to rounds that brings the nurse, physician, and available allied health professionals to the patient's bedside every day to exchange perspectives using a standard format to cross‐check information with the patient, family, and one another, and articulate a clear plan for the day. Before the SIBR hour starts, physicians and nurses have already performed independent patient assessments through usual activities such as handover, chart review, patient interviews, and physical examinations. Participants in SIBR are expected to give or receive inputs according to the standard SIBR communication protocol (Figure 3), review a quality‐safety checklist together, and ensure the plan of care is verbalized. Including the patient and family allows all parties to hear and be heard, cross‐check information for accuracy, and hold each person accountable for contributions.[16, 17]

Figure 3

Implementation

Each ACU staff member receives orientation to the SIBR communication protocol and is expected to be prepared and punctual for the midmorning start times. The charge nurse serves as the SIBR rounds manager, ensuring physicians waste no time searching for the next nurse and each team's eligible patients are seen in the SIBR hour. For each patient, SIBR begins when the nurse and physician are both present at the bedside. The intern begins SIBR by introducing team members before reviewing the patient's active problem list, response to treatment, and interval test results or consultant inputs. The nurse then relays the patient's goal for the day, overnight events, nursing concerns, and reviews the quality‐safety checklist. The intern then invites allied health professionals to share inputs that might impact medical decision making or discharge planning, before synthesizing all inputs into a shared plan for the day.

Throughout SIBR, the patient and family are encouraged to ask questions or correct misinformation. Although newcomers to SIBR often imagine that inviting patient inputs will disrupt efficiency, we have found teams readily learn to manage this risk, for instance discerning the core question among multiple seemingly disparate ones, or volunteering to return after the SIBR hour to explore a complex issue.

Results

Since the launch of the ACU on September 1, 2010, SIBR has been embedded as a routine on the unit with both physician teams and the nursing staff conducting it every day. Patients not considered eligible for SIBR are those whom the entire physician team has not yet evaluated, typically patients who arrived to the unit overnight. For patients who opt out due to personal preference, or for patients away from the unit for a procedure or a test, SIBR occurs without the patient so the rest of the team can still exchange inputs and formulate a plan of care. A visitor to the unit sees SIBR start punctually at 9 _am and 10 _am for successive teams, with each completing SIBR on eligible patients in under 60 minutes.

Effects of the Change

The second trait of an effective clinical microsystem is achieved through SIBR's routine forum for staff to share information with each other and the patient. By practicing SIBR every workday, staff are presented with multiple routine opportunities to experience an environment reflective of high‐performing frontline units.[18] We found that SIBR resembled other competencies, with a bell curve of performance. For this reason, by the start of the third year we added a SIBR certification program, a SIBR skills training program where permanent and rotating staff are evaluated through an in vivo observed structured clinical exam, typically with a charge nurse or physician as preceptor. When a nurse, medical student, intern, or resident demonstrates an ability to perform a series of specific high performance SIBR behaviors in 5 of 6 consecutive patients, they can achieve SIBR certification. In the first 2 years of this voluntary certification program, all daytime nursing staff and rotating interns have achieved this demonstration of interdisciplinary teamwork competence.

CHALLENGES

Challenges with implementing the ACU fell into 3 primary categories: (1) performing change management required for a successful launch, (2) solving logistics of maintaining unit‐based physician teams, and (3) training physicians and nurses to perform SIBR at a high level.

For change management, the leadership pair was able to explain the rationale of the model to all staff in sufficient detail to launch the ACU. To build momentum for ACU routines and relationships, the physician leader and the nurse unit manager were both present on the unit daily for the first 100 days. As ACU operations became routine and competencies formed among clinicians, the amount of time spent by these leaders was de‐escalated.

Creating and maintaining unit‐based physician teams required shared understanding and coordination between on‐call hospital medicine physicians and the bed control office so that new admissions or transfers could be consistently assigned to unit‐based teams without adversely affecting patient flow. We found this challenge to be manageable once stakeholders accepted the rationale for the care mode and figured out how to support it.

The challenge of building high‐performance SIBR across the unit, including competence of rotating trainees new to the model, requires individualized assessment and feedback necessary for SIBR certification. We addressed this challenge by creating a SIBR train‐the‐trainer programa list of observable high‐performance SIBR behaviors coupled with a short course about giving effective feedback to learnersand found that once the ACU had several nurse and physician SIBR trainers in the staffing mix every day, the required amount of SIBR coaching expertise was available when needed.

CONCLUSION

Improving value and reliability in hospital care may require new models of care. The ACU is a hospital care model specifically designed to organize physicians, nurses, and allied health professionals into high‐functioning, unit‐based teams. It converges standard workflow, patient‐centered communication, quality‐safety checklists, best‐practice protocols, performance measurement, and progressive leadership. Our experience with the ACU suggests that hospital units can be reorganized as effective clinical microsystems where consistent unit professionals can share time and space, a sense of purpose, code of conduct, shared mental model for teamwork, an interprofessional management structure, and an important level of accountability to each other and their patients.

Disclosures: Jason Stein, MD: grant support from the US Health & Resources Services Administration to support organizational implementation of the care model described; recipient of consulting fees and royalties for licensed intellectual property to support implementation of the care model described; founder and president of nonprofit Centripital, provider of consulting services to hospital systems implementing the care model described. The terms of this arrangement have been reviewed and approved by Emory University in accordance with its conflict of interest policies. Liam Chadwick, PhD, and Diaz Clark, MS, RN: recipients of consulting fees through Centripital to support implementation of the care model described. Bryan W. Castle, MBA, RN: grant support from the US Health & Resources Services Administration to support organizational implementation of the care model described; recipient of consulting fees through Centripital to support implementation of the care model described. The authors report no other conflicts of interest.

With a United States medical system that spends as much as $750 billion each year on care that does not result in improved health outcomes,[1] many policy initiatives, including the Centers for Medicare and Medicaid Services' Value‐Based Purchasing program, seek to realign hospitals' financial incentives from a focus on production to one on value (quality divided by cost).[2, 3] Professional organizations have now deemed resource stewardship an ethical responsibility for professionalism,[4, 5] and campaigns such as the American Board of Internal Medicine (ABIM) Foundation's Choosing Wisely effort and the American College of Physicians' High‐Value Care platform are calling on frontline clinicians to address unnecessary and wasteful services.[6, 7]

Despite these pressures and initiatives, most physicians lack the knowledge and tools necessary to prioritize the delivery of their own healthcare services according to value.[8, 9, 10] Hospital medicine physicians are unaware of the costs associated with the interventions they order,[10] and the majority of medical training programs lack curricula focused on healthcare costs,[11] creating a large gap between physicians' perceived, desired, and actual knowledge related to costs.[12] Novel frameworks and frontline physician engagement are required if clinicians are to improve the value of the care they deliver.

We describe 1 of our first steps at the University of California, San Francisco (UCSF) to promote high‐value care (HVC) delivery: the creation of a HVC program led by clinicians and administrators focused on identifying and addressing wasteful practices within our hospitalist group. The program aims to (1) use financial and clinical data to identify areas with clear evidence of waste in the hospital, (2) promote evidence‐based interventions that improve both quality of care and value, and (3) pair interventions with evidence‐based cost awareness education to drive culture change. Our experience and inaugural projects provide a model of the key features, inherent challenges, and lessons learned, which may help inform similar efforts.

METHODS

In March 2012, we launched an HVC program within our Division of Hospital Medicine at UCSF Medical Center, a 600‐bed academic medical center in an urban setting. During the 2013 academic year, our division included 45 physicians. The medicine service, comprised of 8 teaching medical ward teams (1 attending, 1 resident, 2 interns, and variable number of medical students), and 1 nonteaching medical ward team (1 attending), admitted 4700 patients that year.

Framework for Identifying Target Projects

Robust Analysis of Costs

We created a framework for identifying, designing, and promoting projects specifically aimed at improving healthcare value (Figure 1). Financial data were used to identify areas with clear evidence of waste in the hospital, areas of high cost with no benefit in health outcomes. We focused particularly on obtaining cost and billing data for our medical service, which provided important insight into potential targets for improvements in value. For example, in 2011, the Division of Hospital Medicine spent more than $1 million annually in direct costs for the administration of nebulized bronchodilator therapies (nebs) to nonintensive care unit patients on the medical service.[13] These high costs, exposed by billing data, were believed to represent potential unnecessary testing and/or procedures. Not every area of high cost was deemed a target for intervention. For example, the use of recombinant factor VIII appeared a necessary expenditure (over $1 million per year) for our patients with hemophilia. Although our efforts focused on reducing waste, it is worth noting that healthcare value can also be increased by improving the delivery of high‐value services.

Figure 1

Recognized Benefits in Quality of Care

The program also evaluated the impact of cost reduction efforts on the quality of care, based on a high standard of current evidence. Though value can be improved by interventions that decrease costs while being quality neutral, our group chose to focus first on projects that would simultaneously improve quality while decreasing costs. We felt that this win‐win strategy would help obtain buy‐in from clinicians weary of prior cost‐cutting programs. For example, we pursued interventions aimed at reducing inappropriate gastric stress ulcer prophylaxis, which had the potential to both cut costs and minimize risks of hospital‐acquired pneumonia and Clostridium difficile infections.[14, 15] All proposed HVC targets were vetted through a review of the literature and published guidelines. In general, our initial projects had to be strongly supported by evidence, with high‐quality studies, preferably meta‐analyses or systematic reviews, that displayed the safety of our recommended changes. We reviewed the literature with experts. For example, we met with faculty pulmonologists to discuss the evidence supporting the use of inhalers instead of nebulizers in adults with obstructive pulmonary disease. The goals of our projects were chosen by our HVC committee, based on an analysis of our baseline data and the perceived potential effects of our proposed interventions.

Educational Intervention

Last, we paired interventions with evidence‐based cost awareness education to drive culture change. At UCSF we have an ongoing longitudinal cost‐awareness curriculum for residents, which has previously been described.[16] We took advantage of this educational forum to address gaps in clinician knowledge related to the targeted areas. When launching the initiative to decrease unnecessary inpatient nebulizer usage and improve transitions to inhalers, we utilized the chronic obstructive pulmonary disease case in the cost‐awareness series. Doing so allowed us to both review the evidence behind the effectiveness of inhalers, and introduce our Nebs No More After 24 campaign, which sought to transition adult inpatients with obstructive pulmonary symptoms from nebs to inhalers within 24 hours of admission.[13]

Intervention Strategy

Our general approach has been to design and implement multifaceted interventions, adapted from previous QI literature (Figure 1).[17] Given the importance of frontline clinician engagement to successful project implementation,[18, 19, 20] our interventions are physician‐driven and are vetted by a large group of clinicians prior to launch. The HVC program also explicitly seeks stakeholder input, perspective, and buy‐in prior to implementation. For example, we involved respiratory therapists (RTs) in the design of the Nebs No More After 24 project, thus ensuring that the interventions fit within their workflow and align with their care‐delivery goals.

Local publicity campaigns provide education and reminders for clinicians. Posters, such as the Nebs No More After 24 poster (Figure 2), were hung in physician, nursing, and RT work areas. Pens featuring the catchphrase Nebs No More After 24 were distributed to clinicians.

Figure 2

In addition to presentations to residents through the UCSF cost awareness curriculum, educational presentations were also delivered to attending physicians and to other allied members of the healthcare team (eg, nurses, RTs) during regularly scheduled staff meetings.

The metrics for each of the projects were regularly monitored, and targeted feedback was provided to clinicians. For the Nebs No More After 24 campaign, data for the number of nebs delivered on the target floor were provided to resident physicians during the cost awareness conference each month, and the data were presented to attending hospitalists in the monthly QI newsletter. This academic year, transfusion and telemetry data are presented via the same strategy.

Stakeholder recruitment, education, and promotional campaigns are important to program launches, but to sustain projects over the long‐term, system changes may be necessary. We have pursued changes in the computerized provider order entry (CPOE) system, such as removing nebs from the admission order set or putting a default duration for certain telemetry orders. Systems‐level interventions, although more difficult to achieve, play an important role in creating enduring changes when paired with educational interventions.

RESULTS

During our first 2 years we have initiated ongoing projects directed at 6 major targets (Table 1). Our flagship project, Nebs No More After 24, resulted in a decrease of nebulizer rates by more than 50% on a high‐acuity medical floor, as previously published.[13] We created a financial model that primarily accounted for RT time and pharmaceutical costs, and estimated a savings of approximately $250,000 annually on this single medical ward (see Supporting Information, Table 1, in the online version of this article).[13]

The HVC program also provided an arena for collaborating with and supporting value‐based projects launched by other groups, such as the UCSF Medication Outcomes Center's inappropriate gastric stress ulcer prophylaxis program.[21] Our group helped support the development and implementation of evidence‐based clinical practice guidelines, and we assisted educational interventions targeting clinicians. This program resulted in a decrease in inappropriate stress ulcer prophylaxis in intensive care unit patients from 19% to 6.6% within 1 month following implementation.[21]

DISCUSSION

Physicians are increasingly being asked to embrace and lead efforts to improve healthcare value and reduce costs. Our program provides a framework to guide physician‐led initiatives to identify and address areas of healthcare waste.

CONCLUSIONS

Our HVC program has been successful in promoting improved healthcare value and engaging clinicians in this effort. The program is guided by the use of financial data to identify areas with clear evidence of waste in the hospital, the creation of evidence‐based interventions that improve quality of care while cutting costs, and the pairing of interventions with evidence‐based cost awareness education to drive culture change.

With a United States medical system that spends as much as $750 billion each year on care that does not result in improved health outcomes,[1] many policy initiatives, including the Centers for Medicare and Medicaid Services' Value‐Based Purchasing program, seek to realign hospitals' financial incentives from a focus on production to one on value (quality divided by cost).[2, 3] Professional organizations have now deemed resource stewardship an ethical responsibility for professionalism,[4, 5] and campaigns such as the American Board of Internal Medicine (ABIM) Foundation's Choosing Wisely effort and the American College of Physicians' High‐Value Care platform are calling on frontline clinicians to address unnecessary and wasteful services.[6, 7]

Despite these pressures and initiatives, most physicians lack the knowledge and tools necessary to prioritize the delivery of their own healthcare services according to value.[8, 9, 10] Hospital medicine physicians are unaware of the costs associated with the interventions they order,[10] and the majority of medical training programs lack curricula focused on healthcare costs,[11] creating a large gap between physicians' perceived, desired, and actual knowledge related to costs.[12] Novel frameworks and frontline physician engagement are required if clinicians are to improve the value of the care they deliver.

We describe 1 of our first steps at the University of California, San Francisco (UCSF) to promote high‐value care (HVC) delivery: the creation of a HVC program led by clinicians and administrators focused on identifying and addressing wasteful practices within our hospitalist group. The program aims to (1) use financial and clinical data to identify areas with clear evidence of waste in the hospital, (2) promote evidence‐based interventions that improve both quality of care and value, and (3) pair interventions with evidence‐based cost awareness education to drive culture change. Our experience and inaugural projects provide a model of the key features, inherent challenges, and lessons learned, which may help inform similar efforts.

METHODS

In March 2012, we launched an HVC program within our Division of Hospital Medicine at UCSF Medical Center, a 600‐bed academic medical center in an urban setting. During the 2013 academic year, our division included 45 physicians. The medicine service, comprised of 8 teaching medical ward teams (1 attending, 1 resident, 2 interns, and variable number of medical students), and 1 nonteaching medical ward team (1 attending), admitted 4700 patients that year.

Framework for Identifying Target Projects

Robust Analysis of Costs

We created a framework for identifying, designing, and promoting projects specifically aimed at improving healthcare value (Figure 1). Financial data were used to identify areas with clear evidence of waste in the hospital, areas of high cost with no benefit in health outcomes. We focused particularly on obtaining cost and billing data for our medical service, which provided important insight into potential targets for improvements in value. For example, in 2011, the Division of Hospital Medicine spent more than $1 million annually in direct costs for the administration of nebulized bronchodilator therapies (nebs) to nonintensive care unit patients on the medical service.[13] These high costs, exposed by billing data, were believed to represent potential unnecessary testing and/or procedures. Not every area of high cost was deemed a target for intervention. For example, the use of recombinant factor VIII appeared a necessary expenditure (over $1 million per year) for our patients with hemophilia. Although our efforts focused on reducing waste, it is worth noting that healthcare value can also be increased by improving the delivery of high‐value services.

Figure 1

Recognized Benefits in Quality of Care

The program also evaluated the impact of cost reduction efforts on the quality of care, based on a high standard of current evidence. Though value can be improved by interventions that decrease costs while being quality neutral, our group chose to focus first on projects that would simultaneously improve quality while decreasing costs. We felt that this win‐win strategy would help obtain buy‐in from clinicians weary of prior cost‐cutting programs. For example, we pursued interventions aimed at reducing inappropriate gastric stress ulcer prophylaxis, which had the potential to both cut costs and minimize risks of hospital‐acquired pneumonia and Clostridium difficile infections.[14, 15] All proposed HVC targets were vetted through a review of the literature and published guidelines. In general, our initial projects had to be strongly supported by evidence, with high‐quality studies, preferably meta‐analyses or systematic reviews, that displayed the safety of our recommended changes. We reviewed the literature with experts. For example, we met with faculty pulmonologists to discuss the evidence supporting the use of inhalers instead of nebulizers in adults with obstructive pulmonary disease. The goals of our projects were chosen by our HVC committee, based on an analysis of our baseline data and the perceived potential effects of our proposed interventions.

Educational Intervention

Last, we paired interventions with evidence‐based cost awareness education to drive culture change. At UCSF we have an ongoing longitudinal cost‐awareness curriculum for residents, which has previously been described.[16] We took advantage of this educational forum to address gaps in clinician knowledge related to the targeted areas. When launching the initiative to decrease unnecessary inpatient nebulizer usage and improve transitions to inhalers, we utilized the chronic obstructive pulmonary disease case in the cost‐awareness series. Doing so allowed us to both review the evidence behind the effectiveness of inhalers, and introduce our Nebs No More After 24 campaign, which sought to transition adult inpatients with obstructive pulmonary symptoms from nebs to inhalers within 24 hours of admission.[13]

Intervention Strategy

Our general approach has been to design and implement multifaceted interventions, adapted from previous QI literature (Figure 1).[17] Given the importance of frontline clinician engagement to successful project implementation,[18, 19, 20] our interventions are physician‐driven and are vetted by a large group of clinicians prior to launch. The HVC program also explicitly seeks stakeholder input, perspective, and buy‐in prior to implementation. For example, we involved respiratory therapists (RTs) in the design of the Nebs No More After 24 project, thus ensuring that the interventions fit within their workflow and align with their care‐delivery goals.

Local publicity campaigns provide education and reminders for clinicians. Posters, such as the Nebs No More After 24 poster (Figure 2), were hung in physician, nursing, and RT work areas. Pens featuring the catchphrase Nebs No More After 24 were distributed to clinicians.

Figure 2

In addition to presentations to residents through the UCSF cost awareness curriculum, educational presentations were also delivered to attending physicians and to other allied members of the healthcare team (eg, nurses, RTs) during regularly scheduled staff meetings.

The metrics for each of the projects were regularly monitored, and targeted feedback was provided to clinicians. For the Nebs No More After 24 campaign, data for the number of nebs delivered on the target floor were provided to resident physicians during the cost awareness conference each month, and the data were presented to attending hospitalists in the monthly QI newsletter. This academic year, transfusion and telemetry data are presented via the same strategy.

Stakeholder recruitment, education, and promotional campaigns are important to program launches, but to sustain projects over the long‐term, system changes may be necessary. We have pursued changes in the computerized provider order entry (CPOE) system, such as removing nebs from the admission order set or putting a default duration for certain telemetry orders. Systems‐level interventions, although more difficult to achieve, play an important role in creating enduring changes when paired with educational interventions.

RESULTS

During our first 2 years we have initiated ongoing projects directed at 6 major targets (Table 1). Our flagship project, Nebs No More After 24, resulted in a decrease of nebulizer rates by more than 50% on a high‐acuity medical floor, as previously published.[13] We created a financial model that primarily accounted for RT time and pharmaceutical costs, and estimated a savings of approximately $250,000 annually on this single medical ward (see Supporting Information, Table 1, in the online version of this article).[13]

The HVC program also provided an arena for collaborating with and supporting value‐based projects launched by other groups, such as the UCSF Medication Outcomes Center's inappropriate gastric stress ulcer prophylaxis program.[21] Our group helped support the development and implementation of evidence‐based clinical practice guidelines, and we assisted educational interventions targeting clinicians. This program resulted in a decrease in inappropriate stress ulcer prophylaxis in intensive care unit patients from 19% to 6.6% within 1 month following implementation.[21]

DISCUSSION

Physicians are increasingly being asked to embrace and lead efforts to improve healthcare value and reduce costs. Our program provides a framework to guide physician‐led initiatives to identify and address areas of healthcare waste.

CONCLUSIONS

Our HVC program has been successful in promoting improved healthcare value and engaging clinicians in this effort. The program is guided by the use of financial data to identify areas with clear evidence of waste in the hospital, the creation of evidence‐based interventions that improve quality of care while cutting costs, and the pairing of interventions with evidence‐based cost awareness education to drive culture change.

With a United States medical system that spends as much as $750 billion each year on care that does not result in improved health outcomes,[1] many policy initiatives, including the Centers for Medicare and Medicaid Services' Value‐Based Purchasing program, seek to realign hospitals' financial incentives from a focus on production to one on value (quality divided by cost).[2, 3] Professional organizations have now deemed resource stewardship an ethical responsibility for professionalism,[4, 5] and campaigns such as the American Board of Internal Medicine (ABIM) Foundation's Choosing Wisely effort and the American College of Physicians' High‐Value Care platform are calling on frontline clinicians to address unnecessary and wasteful services.[6, 7]

Despite these pressures and initiatives, most physicians lack the knowledge and tools necessary to prioritize the delivery of their own healthcare services according to value.[8, 9, 10] Hospital medicine physicians are unaware of the costs associated with the interventions they order,[10] and the majority of medical training programs lack curricula focused on healthcare costs,[11] creating a large gap between physicians' perceived, desired, and actual knowledge related to costs.[12] Novel frameworks and frontline physician engagement are required if clinicians are to improve the value of the care they deliver.

We describe 1 of our first steps at the University of California, San Francisco (UCSF) to promote high‐value care (HVC) delivery: the creation of a HVC program led by clinicians and administrators focused on identifying and addressing wasteful practices within our hospitalist group. The program aims to (1) use financial and clinical data to identify areas with clear evidence of waste in the hospital, (2) promote evidence‐based interventions that improve both quality of care and value, and (3) pair interventions with evidence‐based cost awareness education to drive culture change. Our experience and inaugural projects provide a model of the key features, inherent challenges, and lessons learned, which may help inform similar efforts.

METHODS

In March 2012, we launched an HVC program within our Division of Hospital Medicine at UCSF Medical Center, a 600‐bed academic medical center in an urban setting. During the 2013 academic year, our division included 45 physicians. The medicine service, comprised of 8 teaching medical ward teams (1 attending, 1 resident, 2 interns, and variable number of medical students), and 1 nonteaching medical ward team (1 attending), admitted 4700 patients that year.

Framework for Identifying Target Projects

Robust Analysis of Costs

We created a framework for identifying, designing, and promoting projects specifically aimed at improving healthcare value (Figure 1). Financial data were used to identify areas with clear evidence of waste in the hospital, areas of high cost with no benefit in health outcomes. We focused particularly on obtaining cost and billing data for our medical service, which provided important insight into potential targets for improvements in value. For example, in 2011, the Division of Hospital Medicine spent more than $1 million annually in direct costs for the administration of nebulized bronchodilator therapies (nebs) to nonintensive care unit patients on the medical service.[13] These high costs, exposed by billing data, were believed to represent potential unnecessary testing and/or procedures. Not every area of high cost was deemed a target for intervention. For example, the use of recombinant factor VIII appeared a necessary expenditure (over $1 million per year) for our patients with hemophilia. Although our efforts focused on reducing waste, it is worth noting that healthcare value can also be increased by improving the delivery of high‐value services.

Figure 1

Recognized Benefits in Quality of Care

The program also evaluated the impact of cost reduction efforts on the quality of care, based on a high standard of current evidence. Though value can be improved by interventions that decrease costs while being quality neutral, our group chose to focus first on projects that would simultaneously improve quality while decreasing costs. We felt that this win‐win strategy would help obtain buy‐in from clinicians weary of prior cost‐cutting programs. For example, we pursued interventions aimed at reducing inappropriate gastric stress ulcer prophylaxis, which had the potential to both cut costs and minimize risks of hospital‐acquired pneumonia and Clostridium difficile infections.[14, 15] All proposed HVC targets were vetted through a review of the literature and published guidelines. In general, our initial projects had to be strongly supported by evidence, with high‐quality studies, preferably meta‐analyses or systematic reviews, that displayed the safety of our recommended changes. We reviewed the literature with experts. For example, we met with faculty pulmonologists to discuss the evidence supporting the use of inhalers instead of nebulizers in adults with obstructive pulmonary disease. The goals of our projects were chosen by our HVC committee, based on an analysis of our baseline data and the perceived potential effects of our proposed interventions.

Educational Intervention

Last, we paired interventions with evidence‐based cost awareness education to drive culture change. At UCSF we have an ongoing longitudinal cost‐awareness curriculum for residents, which has previously been described.[16] We took advantage of this educational forum to address gaps in clinician knowledge related to the targeted areas. When launching the initiative to decrease unnecessary inpatient nebulizer usage and improve transitions to inhalers, we utilized the chronic obstructive pulmonary disease case in the cost‐awareness series. Doing so allowed us to both review the evidence behind the effectiveness of inhalers, and introduce our Nebs No More After 24 campaign, which sought to transition adult inpatients with obstructive pulmonary symptoms from nebs to inhalers within 24 hours of admission.[13]

Intervention Strategy

Our general approach has been to design and implement multifaceted interventions, adapted from previous QI literature (Figure 1).[17] Given the importance of frontline clinician engagement to successful project implementation,[18, 19, 20] our interventions are physician‐driven and are vetted by a large group of clinicians prior to launch. The HVC program also explicitly seeks stakeholder input, perspective, and buy‐in prior to implementation. For example, we involved respiratory therapists (RTs) in the design of the Nebs No More After 24 project, thus ensuring that the interventions fit within their workflow and align with their care‐delivery goals.

Local publicity campaigns provide education and reminders for clinicians. Posters, such as the Nebs No More After 24 poster (Figure 2), were hung in physician, nursing, and RT work areas. Pens featuring the catchphrase Nebs No More After 24 were distributed to clinicians.

Figure 2

In addition to presentations to residents through the UCSF cost awareness curriculum, educational presentations were also delivered to attending physicians and to other allied members of the healthcare team (eg, nurses, RTs) during regularly scheduled staff meetings.

The metrics for each of the projects were regularly monitored, and targeted feedback was provided to clinicians. For the Nebs No More After 24 campaign, data for the number of nebs delivered on the target floor were provided to resident physicians during the cost awareness conference each month, and the data were presented to attending hospitalists in the monthly QI newsletter. This academic year, transfusion and telemetry data are presented via the same strategy.

Stakeholder recruitment, education, and promotional campaigns are important to program launches, but to sustain projects over the long‐term, system changes may be necessary. We have pursued changes in the computerized provider order entry (CPOE) system, such as removing nebs from the admission order set or putting a default duration for certain telemetry orders. Systems‐level interventions, although more difficult to achieve, play an important role in creating enduring changes when paired with educational interventions.

RESULTS

During our first 2 years we have initiated ongoing projects directed at 6 major targets (Table 1). Our flagship project, Nebs No More After 24, resulted in a decrease of nebulizer rates by more than 50% on a high‐acuity medical floor, as previously published.[13] We created a financial model that primarily accounted for RT time and pharmaceutical costs, and estimated a savings of approximately $250,000 annually on this single medical ward (see Supporting Information, Table 1, in the online version of this article).[13]

The HVC program also provided an arena for collaborating with and supporting value‐based projects launched by other groups, such as the UCSF Medication Outcomes Center's inappropriate gastric stress ulcer prophylaxis program.[21] Our group helped support the development and implementation of evidence‐based clinical practice guidelines, and we assisted educational interventions targeting clinicians. This program resulted in a decrease in inappropriate stress ulcer prophylaxis in intensive care unit patients from 19% to 6.6% within 1 month following implementation.[21]

DISCUSSION

Physicians are increasingly being asked to embrace and lead efforts to improve healthcare value and reduce costs. Our program provides a framework to guide physician‐led initiatives to identify and address areas of healthcare waste.

CONCLUSIONS

Our HVC program has been successful in promoting improved healthcare value and engaging clinicians in this effort. The program is guided by the use of financial data to identify areas with clear evidence of waste in the hospital, the creation of evidence‐based interventions that improve quality of care while cutting costs, and the pairing of interventions with evidence‐based cost awareness education to drive culture change.

STAT is an abbreviation of the Latin word statim, meaning immediately,[1] and has been a part of healthcare's lexicon for almost as long as there have been hospitals. STAT conveys a sense of urgency, compelling those who hear STAT to act quickly. Unfortunately, given the lack of a consistent understanding of STAT, the term in reality often has an alternate use: to hurry up or to complete sooner than routine, and is sometimes used to circumvent a system that is perceived to be too slow to accomplish a routine task in a timely manner.

As part of a larger systems redesign effort to improve patient safety and quality of care, an institutional review board (IRB)‐approved qualitative study was conducted on 2 medical‐surgical units in a US Department of Veterans Affairs (VA) hospital to explore communication patterns between physicians and nurses.[2] The study revealed wide variation in understanding between physicians and nurses on the ordering and administration of STAT medication. Physicians were unaware that when they placed a STAT order into the computerized patient record system (CPRS), nurses were not automatically alerted about the order. At this facility, nurses did not carry pagers. Although each unit had a supply of wireless telephones, they were often unreliable and therefore not used consistently. Nurses were required by policy to check the CPRS for new orders every 2 hours. This was an inefficient and possibly dangerous process,[3] because if a nurse was not expecting a STAT order, 2 hours could elapse before she or he saw the order in the CPRS and began to look for the medication. A follow‐up survey completed by physicians, nurses, pharmacists, and pharmacy technicians demonstrated stark differences on the definition of STAT and overlap with similar terms such as NOW and ASAP. Interviews with ordering providers indicated that 36% of the time a STAT was ordered it was not clinically urgent, but instead ordered STAT to speed up the process.

The STAT medication process was clearly in need of improvement, but previous quality improvement projects in our organization had varying degrees of success. For example, we used Lean methodology in an attempt to improve our discharge process. We conducted a modified rapid process discharge improvement workshop[4] structured in phases over 4 weeks. During the workshops, a strong emphasis remained on the solutions to the problem, and we were unable to help the team move from a mindset of fix it to create it. This limited the buy‐in of team members, the creativity of their ideas for improvement, and ultimately the momentum to improve the process.

In this article we describe our adaptation of A3 Thinking,[5, 6] a structure for guiding quality improvement based in Lean methodology, to improve the STAT medication process. We chose A3 Thinking for several reasons. A3 Thinking focuses on process improvement and thus aligned well with our interest in improving the STAT medication process. A3 Thinking also reveals otherwise hidden nonvalue‐added activities that should be eliminated.[7] Finally A3 Thinking reinforces a deeper understanding of the way the work is currently being done, providing critical information needed before making a change. This provides a tremendous opportunity to look at work differently and see opportunities for improvement.[8] Given these strengths as well as the lack of congruence between what the STAT process should consist of and how the STAT process was actually being used in our organization, A3 Thinking offered the best fit between an improvement process and the problem to be solved.

METHODS

A search of healthcare literature yielded very few studies on the STAT process.[9, 10] Only 1 intervention to improve the process was found, and this focused on a specific procedure.[10] An informal survey of local VA and non‐VA hospitals regarding their experiences with the STAT medication process revealed insufficient information to aid our efforts. We next searched the business and manufacturing literature and found examples of how the Lean methodology was successfully applied to other problems in healthcare, including improving pediatric surgery workflow and decreasing ventilator‐associated pneumonia.[11, 12]

Therefore, the STAT project was structured to adapt a problem‐solving process commonly used in Lean organizationsA3 Thinkingwhich challenges team members to work through a discovery phase to develop a shared understanding of the process, an envisioning phase to conceptualize an ideal process experience, and finally an experimentation phase to identify and trial possible solutions through prioritization, iterative testing, structured reflection, and adjustment on resulting changes. Our application of the term experimentation in this context is distinct from that of controlled experimentation in clinical research; the term is intended to convey iterative learning as changes are tested, evaluated, and modified during this quality improvement project. Figure 1 displays a conceptual model of our adaptation of A3 Thinking. As this was a quality‐improvement project, it was exempt from IRB review.

Figure 1

DISCOVERY

To begin the discovery phase, a workgroup consisting of representatives of all groups that had a role in the STAT process (ie, physician, pharmacist, nurse, pharmacy technician, clerk) gathered to identify the opportunity we are looking to address and learn from each other's individual experiences with the STAT medication process. The group was facilitated by an industrial engineer familiar with the A3 Thinking process. The team completed a mapping exercise to lay out, step‐by‐step, the current STAT medication process. This activity allowed the team to build shared empathy with others' experiences and to appreciate the challenges experienced by others through their individual responsibilities in the process. The current process was found to consist of 4 overarching components: a provider entered the STAT order into the CPRS; the order was verified by a pharmacist; a pharmacy technician delivered the medication to the unit (or a nurse retrieved the medication from the Omnicell (Omnicell Inc., Mountain View, CA), a proprietary automated medication dispensing system); and finally the nurse administered the medication to a patient.

A large, color‐coded flow map of the STAT medication process was constructed over several meetings to capture all perspectives and allow team members to gather feedback from their peers. To further our understanding of the current process, the team participated in a modified Go to the Gemba (ie, go to where the work is done)[13] on a real‐time STAT order. Once all workgroup members were satisfied that the flow map represented the current state of the STAT medication process, we came to a consensus on the goals needed to meet our main objective.

We agreed that our main objective was that STAT medication orders should be recognized, verified, and administered to patients in a timely and appropriate manner to ensure quality care. We identified 3 goals to meet this objective: (1) STAT should be consistently defined and understood by everyone; (2) an easy, intuitive STAT process should be available for all stakeholders; and (3) the STAT process should be transparent and ideally visual so that everyone involved can understand at which point in the process a specific STAT order is currently situated. We also identified additional information we would need to reach the goals.

Shortly after the process‐mapping sessions, 2 workgroup members conducted real‐time STAT order time studies to track medications from order to administration. Three time periods in the STAT process were identified for observation and measurement: the time from physician order entry in the CPRS to the time a pharmacist verified the medication, the time from verification to when the medication arrived on the nursing unit, and the time from arrival on the nursing unit to when that medication was administered. Using a data‐collection template, each time period was recorded, and 28 time studies were collected over 1 month. To monitor the progress of our initiatives, the time study was repeated 3 months into the project.

ENVISIONING

Following the discovery phase, the team was better equipped to identify the specific changes needed to achieve an improved process. The envisioning phase allowed the team freedom to imagine an ideal process barring any preconceived notion of constraints within the current process.

In 2 meetings we brainstormed as many improvement ideas as possible. To prioritize and focus our ideas, we developed a matrix (see Supporting Information, Appendix A, in the online version of this article), placing our ideas in 1 of 4 quadrants based on the anticipated effort to implement the change (x‐axis) and impact of making the change (y‐axis). The matrix helped us see that some ideas would be relatively simple to implement (eg, color‐coded bags for STAT medication delivery), whereas others would require more sophisticated efforts and involvement of other people (eg, monthly education sessions to resident physicians).

EXPERIMENTING

Experiments were conducted to meet each of the 3 goals identified above. The team used the outcomes of the prioritization exercise to identify initial experiments to test. To build momentum by showing progress and improvement with a few quick wins, the team began with low‐effort/high‐impact opportunities. Each experiment followed a standard Plan‐Do‐Study‐Act (PDSA) cycle to encourage reflection, learning, adaptation, and adjustment as a result of the experiential learning process.[5]

RESULTS

At the start of our project, the average time from STAT order to medication administration was 1 hour and 7 minutes (range, 6 minutes 2 hours and 22 minutes). As a result of the 2 sets of countermeasures outlined in Tables 1 and 2, the average total time from STAT order entry to administration decreased by 21% to an average of 53 minutes. The total time from medication delivery to administration decreased by 26% from 34 minutes to 25 minutes postimplementation. On average, 391 STAT medications were ordered per month during the project period, which represents a decrease of 9.5% from the 432 orders per month for the same time period the previous year. After implementing the countermeasures in Table 2, we followed another 26 STAT medications through the process to evaluate our efforts. Of 15 STAT medications requiring delivery, only 1 nurse (7%) was not notified of the delivery of a STAT medication, and 1 pharmacy technician (7%) was not informed the medication was STAT. The 151% increase in notification of nurses to delivery of a STAT medication suggests that use of the STAT bags, STAT magnets on patient doors, and whenever possible direct delivery of STAT medications to the nurse has improved communication between the technicians and nurses. Similarly, the 27% increase in technician awareness of a STAT designation suggests STAT is being better communicated to them. The improvement in awareness and notification of a STAT medication is summarized in Figure 2.

Figure 2

Due to time and financial constraints, the following limitations may have affected our findings. First, resident physicians were not directly represented in our discussions. Attending medicine hospitalists provided the physician perspective, which provides a biased view given their intimate knowledge of the CPRS and additional years of experience. Similarly, nurse perspectives were limited to staff and clinical nurse leaders. Last, our low‐cost approach was mandated by limited resources; a more resource‐rich environment may have devised alternative approaches.

CONCLUSIONS

Adapting A3 Thinking for process improvement was a low‐cost/low‐tech option for a VA facility. Having buy‐in from all levels was crucial to the success of the project. The size and diversity of the group was also very important, as different opinions and aspects of the process were represented. Cross‐discipline relationships and respect were formed, which will be valuable for collaboration in future projects. Although we focused on the STAT medication process, other quality‐improvement projects could also benefit from A3 Thinking. Moreover, there were enough people to serve as ambassadors, taking the project back to their work areas to share with their peers, gather consensus, and elicit additional feedback. The collaboration led to comprehensive understanding of the process, the nature of the problems within the process, and the complexity of solving the problem. For example, although the number of STAT orders did not decrease dramatically, we have learned from these experiments that we may need to change how we approach structuring additional experiments. Future work will focus on increasing communication between physicians and nurses when placing STAT medication orders, enhancing resident education to ensure appropriate use of the STAT designation, and continuing our efforts to improve the delivery process of STAT medications.

Other quality‐improvement methodologies we could have used include: total quality management (TQM), continuous quality improvement (CQI), business process redesign, Lean, Six Sigma, and others.[14] Differences between these can be broadly classified as putting an emphasis on people (eg, inclusion of front line staff in CQI or leadership in TQM) or on process (eg, understanding process function to reduce waste in Lean or statistical process control in Six Sigma).[14] Using A3 Thinking methodology was more useful than these others for the STAT medication process for some very important reasons. The A3 process not only led to a better understanding of the meaning of STAT across disciplines, increasing the intuitive nature, transparency and visual aspects of the whole process, but also promoted a collaborative, multidisciplinary, integrative culture, in which other hospital‐wide problems may be addressed in the future.

STAT is an abbreviation of the Latin word statim, meaning immediately,[1] and has been a part of healthcare's lexicon for almost as long as there have been hospitals. STAT conveys a sense of urgency, compelling those who hear STAT to act quickly. Unfortunately, given the lack of a consistent understanding of STAT, the term in reality often has an alternate use: to hurry up or to complete sooner than routine, and is sometimes used to circumvent a system that is perceived to be too slow to accomplish a routine task in a timely manner.

As part of a larger systems redesign effort to improve patient safety and quality of care, an institutional review board (IRB)‐approved qualitative study was conducted on 2 medical‐surgical units in a US Department of Veterans Affairs (VA) hospital to explore communication patterns between physicians and nurses.[2] The study revealed wide variation in understanding between physicians and nurses on the ordering and administration of STAT medication. Physicians were unaware that when they placed a STAT order into the computerized patient record system (CPRS), nurses were not automatically alerted about the order. At this facility, nurses did not carry pagers. Although each unit had a supply of wireless telephones, they were often unreliable and therefore not used consistently. Nurses were required by policy to check the CPRS for new orders every 2 hours. This was an inefficient and possibly dangerous process,[3] because if a nurse was not expecting a STAT order, 2 hours could elapse before she or he saw the order in the CPRS and began to look for the medication. A follow‐up survey completed by physicians, nurses, pharmacists, and pharmacy technicians demonstrated stark differences on the definition of STAT and overlap with similar terms such as NOW and ASAP. Interviews with ordering providers indicated that 36% of the time a STAT was ordered it was not clinically urgent, but instead ordered STAT to speed up the process.

The STAT medication process was clearly in need of improvement, but previous quality improvement projects in our organization had varying degrees of success. For example, we used Lean methodology in an attempt to improve our discharge process. We conducted a modified rapid process discharge improvement workshop[4] structured in phases over 4 weeks. During the workshops, a strong emphasis remained on the solutions to the problem, and we were unable to help the team move from a mindset of fix it to create it. This limited the buy‐in of team members, the creativity of their ideas for improvement, and ultimately the momentum to improve the process.

In this article we describe our adaptation of A3 Thinking,[5, 6] a structure for guiding quality improvement based in Lean methodology, to improve the STAT medication process. We chose A3 Thinking for several reasons. A3 Thinking focuses on process improvement and thus aligned well with our interest in improving the STAT medication process. A3 Thinking also reveals otherwise hidden nonvalue‐added activities that should be eliminated.[7] Finally A3 Thinking reinforces a deeper understanding of the way the work is currently being done, providing critical information needed before making a change. This provides a tremendous opportunity to look at work differently and see opportunities for improvement.[8] Given these strengths as well as the lack of congruence between what the STAT process should consist of and how the STAT process was actually being used in our organization, A3 Thinking offered the best fit between an improvement process and the problem to be solved.

METHODS

A search of healthcare literature yielded very few studies on the STAT process.[9, 10] Only 1 intervention to improve the process was found, and this focused on a specific procedure.[10] An informal survey of local VA and non‐VA hospitals regarding their experiences with the STAT medication process revealed insufficient information to aid our efforts. We next searched the business and manufacturing literature and found examples of how the Lean methodology was successfully applied to other problems in healthcare, including improving pediatric surgery workflow and decreasing ventilator‐associated pneumonia.[11, 12]

Therefore, the STAT project was structured to adapt a problem‐solving process commonly used in Lean organizationsA3 Thinkingwhich challenges team members to work through a discovery phase to develop a shared understanding of the process, an envisioning phase to conceptualize an ideal process experience, and finally an experimentation phase to identify and trial possible solutions through prioritization, iterative testing, structured reflection, and adjustment on resulting changes. Our application of the term experimentation in this context is distinct from that of controlled experimentation in clinical research; the term is intended to convey iterative learning as changes are tested, evaluated, and modified during this quality improvement project. Figure 1 displays a conceptual model of our adaptation of A3 Thinking. As this was a quality‐improvement project, it was exempt from IRB review.

Figure 1

DISCOVERY

To begin the discovery phase, a workgroup consisting of representatives of all groups that had a role in the STAT process (ie, physician, pharmacist, nurse, pharmacy technician, clerk) gathered to identify the opportunity we are looking to address and learn from each other's individual experiences with the STAT medication process. The group was facilitated by an industrial engineer familiar with the A3 Thinking process. The team completed a mapping exercise to lay out, step‐by‐step, the current STAT medication process. This activity allowed the team to build shared empathy with others' experiences and to appreciate the challenges experienced by others through their individual responsibilities in the process. The current process was found to consist of 4 overarching components: a provider entered the STAT order into the CPRS; the order was verified by a pharmacist; a pharmacy technician delivered the medication to the unit (or a nurse retrieved the medication from the Omnicell (Omnicell Inc., Mountain View, CA), a proprietary automated medication dispensing system); and finally the nurse administered the medication to a patient.

A large, color‐coded flow map of the STAT medication process was constructed over several meetings to capture all perspectives and allow team members to gather feedback from their peers. To further our understanding of the current process, the team participated in a modified Go to the Gemba (ie, go to where the work is done)[13] on a real‐time STAT order. Once all workgroup members were satisfied that the flow map represented the current state of the STAT medication process, we came to a consensus on the goals needed to meet our main objective.

We agreed that our main objective was that STAT medication orders should be recognized, verified, and administered to patients in a timely and appropriate manner to ensure quality care. We identified 3 goals to meet this objective: (1) STAT should be consistently defined and understood by everyone; (2) an easy, intuitive STAT process should be available for all stakeholders; and (3) the STAT process should be transparent and ideally visual so that everyone involved can understand at which point in the process a specific STAT order is currently situated. We also identified additional information we would need to reach the goals.

Shortly after the process‐mapping sessions, 2 workgroup members conducted real‐time STAT order time studies to track medications from order to administration. Three time periods in the STAT process were identified for observation and measurement: the time from physician order entry in the CPRS to the time a pharmacist verified the medication, the time from verification to when the medication arrived on the nursing unit, and the time from arrival on the nursing unit to when that medication was administered. Using a data‐collection template, each time period was recorded, and 28 time studies were collected over 1 month. To monitor the progress of our initiatives, the time study was repeated 3 months into the project.

ENVISIONING

Following the discovery phase, the team was better equipped to identify the specific changes needed to achieve an improved process. The envisioning phase allowed the team freedom to imagine an ideal process barring any preconceived notion of constraints within the current process.

In 2 meetings we brainstormed as many improvement ideas as possible. To prioritize and focus our ideas, we developed a matrix (see Supporting Information, Appendix A, in the online version of this article), placing our ideas in 1 of 4 quadrants based on the anticipated effort to implement the change (x‐axis) and impact of making the change (y‐axis). The matrix helped us see that some ideas would be relatively simple to implement (eg, color‐coded bags for STAT medication delivery), whereas others would require more sophisticated efforts and involvement of other people (eg, monthly education sessions to resident physicians).

EXPERIMENTING

Experiments were conducted to meet each of the 3 goals identified above. The team used the outcomes of the prioritization exercise to identify initial experiments to test. To build momentum by showing progress and improvement with a few quick wins, the team began with low‐effort/high‐impact opportunities. Each experiment followed a standard Plan‐Do‐Study‐Act (PDSA) cycle to encourage reflection, learning, adaptation, and adjustment as a result of the experiential learning process.[5]

RESULTS

At the start of our project, the average time from STAT order to medication administration was 1 hour and 7 minutes (range, 6 minutes 2 hours and 22 minutes). As a result of the 2 sets of countermeasures outlined in Tables 1 and 2, the average total time from STAT order entry to administration decreased by 21% to an average of 53 minutes. The total time from medication delivery to administration decreased by 26% from 34 minutes to 25 minutes postimplementation. On average, 391 STAT medications were ordered per month during the project period, which represents a decrease of 9.5% from the 432 orders per month for the same time period the previous year. After implementing the countermeasures in Table 2, we followed another 26 STAT medications through the process to evaluate our efforts. Of 15 STAT medications requiring delivery, only 1 nurse (7%) was not notified of the delivery of a STAT medication, and 1 pharmacy technician (7%) was not informed the medication was STAT. The 151% increase in notification of nurses to delivery of a STAT medication suggests that use of the STAT bags, STAT magnets on patient doors, and whenever possible direct delivery of STAT medications to the nurse has improved communication between the technicians and nurses. Similarly, the 27% increase in technician awareness of a STAT designation suggests STAT is being better communicated to them. The improvement in awareness and notification of a STAT medication is summarized in Figure 2.

Figure 2

Due to time and financial constraints, the following limitations may have affected our findings. First, resident physicians were not directly represented in our discussions. Attending medicine hospitalists provided the physician perspective, which provides a biased view given their intimate knowledge of the CPRS and additional years of experience. Similarly, nurse perspectives were limited to staff and clinical nurse leaders. Last, our low‐cost approach was mandated by limited resources; a more resource‐rich environment may have devised alternative approaches.

CONCLUSIONS

Adapting A3 Thinking for process improvement was a low‐cost/low‐tech option for a VA facility. Having buy‐in from all levels was crucial to the success of the project. The size and diversity of the group was also very important, as different opinions and aspects of the process were represented. Cross‐discipline relationships and respect were formed, which will be valuable for collaboration in future projects. Although we focused on the STAT medication process, other quality‐improvement projects could also benefit from A3 Thinking. Moreover, there were enough people to serve as ambassadors, taking the project back to their work areas to share with their peers, gather consensus, and elicit additional feedback. The collaboration led to comprehensive understanding of the process, the nature of the problems within the process, and the complexity of solving the problem. For example, although the number of STAT orders did not decrease dramatically, we have learned from these experiments that we may need to change how we approach structuring additional experiments. Future work will focus on increasing communication between physicians and nurses when placing STAT medication orders, enhancing resident education to ensure appropriate use of the STAT designation, and continuing our efforts to improve the delivery process of STAT medications.

Other quality‐improvement methodologies we could have used include: total quality management (TQM), continuous quality improvement (CQI), business process redesign, Lean, Six Sigma, and others.[14] Differences between these can be broadly classified as putting an emphasis on people (eg, inclusion of front line staff in CQI or leadership in TQM) or on process (eg, understanding process function to reduce waste in Lean or statistical process control in Six Sigma).[14] Using A3 Thinking methodology was more useful than these others for the STAT medication process for some very important reasons. The A3 process not only led to a better understanding of the meaning of STAT across disciplines, increasing the intuitive nature, transparency and visual aspects of the whole process, but also promoted a collaborative, multidisciplinary, integrative culture, in which other hospital‐wide problems may be addressed in the future.

STAT is an abbreviation of the Latin word statim, meaning immediately,[1] and has been a part of healthcare's lexicon for almost as long as there have been hospitals. STAT conveys a sense of urgency, compelling those who hear STAT to act quickly. Unfortunately, given the lack of a consistent understanding of STAT, the term in reality often has an alternate use: to hurry up or to complete sooner than routine, and is sometimes used to circumvent a system that is perceived to be too slow to accomplish a routine task in a timely manner.

As part of a larger systems redesign effort to improve patient safety and quality of care, an institutional review board (IRB)‐approved qualitative study was conducted on 2 medical‐surgical units in a US Department of Veterans Affairs (VA) hospital to explore communication patterns between physicians and nurses.[2] The study revealed wide variation in understanding between physicians and nurses on the ordering and administration of STAT medication. Physicians were unaware that when they placed a STAT order into the computerized patient record system (CPRS), nurses were not automatically alerted about the order. At this facility, nurses did not carry pagers. Although each unit had a supply of wireless telephones, they were often unreliable and therefore not used consistently. Nurses were required by policy to check the CPRS for new orders every 2 hours. This was an inefficient and possibly dangerous process,[3] because if a nurse was not expecting a STAT order, 2 hours could elapse before she or he saw the order in the CPRS and began to look for the medication. A follow‐up survey completed by physicians, nurses, pharmacists, and pharmacy technicians demonstrated stark differences on the definition of STAT and overlap with similar terms such as NOW and ASAP. Interviews with ordering providers indicated that 36% of the time a STAT was ordered it was not clinically urgent, but instead ordered STAT to speed up the process.

The STAT medication process was clearly in need of improvement, but previous quality improvement projects in our organization had varying degrees of success. For example, we used Lean methodology in an attempt to improve our discharge process. We conducted a modified rapid process discharge improvement workshop[4] structured in phases over 4 weeks. During the workshops, a strong emphasis remained on the solutions to the problem, and we were unable to help the team move from a mindset of fix it to create it. This limited the buy‐in of team members, the creativity of their ideas for improvement, and ultimately the momentum to improve the process.

In this article we describe our adaptation of A3 Thinking,[5, 6] a structure for guiding quality improvement based in Lean methodology, to improve the STAT medication process. We chose A3 Thinking for several reasons. A3 Thinking focuses on process improvement and thus aligned well with our interest in improving the STAT medication process. A3 Thinking also reveals otherwise hidden nonvalue‐added activities that should be eliminated.[7] Finally A3 Thinking reinforces a deeper understanding of the way the work is currently being done, providing critical information needed before making a change. This provides a tremendous opportunity to look at work differently and see opportunities for improvement.[8] Given these strengths as well as the lack of congruence between what the STAT process should consist of and how the STAT process was actually being used in our organization, A3 Thinking offered the best fit between an improvement process and the problem to be solved.

METHODS

A search of healthcare literature yielded very few studies on the STAT process.[9, 10] Only 1 intervention to improve the process was found, and this focused on a specific procedure.[10] An informal survey of local VA and non‐VA hospitals regarding their experiences with the STAT medication process revealed insufficient information to aid our efforts. We next searched the business and manufacturing literature and found examples of how the Lean methodology was successfully applied to other problems in healthcare, including improving pediatric surgery workflow and decreasing ventilator‐associated pneumonia.[11, 12]

Therefore, the STAT project was structured to adapt a problem‐solving process commonly used in Lean organizationsA3 Thinkingwhich challenges team members to work through a discovery phase to develop a shared understanding of the process, an envisioning phase to conceptualize an ideal process experience, and finally an experimentation phase to identify and trial possible solutions through prioritization, iterative testing, structured reflection, and adjustment on resulting changes. Our application of the term experimentation in this context is distinct from that of controlled experimentation in clinical research; the term is intended to convey iterative learning as changes are tested, evaluated, and modified during this quality improvement project. Figure 1 displays a conceptual model of our adaptation of A3 Thinking. As this was a quality‐improvement project, it was exempt from IRB review.

Figure 1

DISCOVERY

To begin the discovery phase, a workgroup consisting of representatives of all groups that had a role in the STAT process (ie, physician, pharmacist, nurse, pharmacy technician, clerk) gathered to identify the opportunity we are looking to address and learn from each other's individual experiences with the STAT medication process. The group was facilitated by an industrial engineer familiar with the A3 Thinking process. The team completed a mapping exercise to lay out, step‐by‐step, the current STAT medication process. This activity allowed the team to build shared empathy with others' experiences and to appreciate the challenges experienced by others through their individual responsibilities in the process. The current process was found to consist of 4 overarching components: a provider entered the STAT order into the CPRS; the order was verified by a pharmacist; a pharmacy technician delivered the medication to the unit (or a nurse retrieved the medication from the Omnicell (Omnicell Inc., Mountain View, CA), a proprietary automated medication dispensing system); and finally the nurse administered the medication to a patient.

A large, color‐coded flow map of the STAT medication process was constructed over several meetings to capture all perspectives and allow team members to gather feedback from their peers. To further our understanding of the current process, the team participated in a modified Go to the Gemba (ie, go to where the work is done)[13] on a real‐time STAT order. Once all workgroup members were satisfied that the flow map represented the current state of the STAT medication process, we came to a consensus on the goals needed to meet our main objective.

We agreed that our main objective was that STAT medication orders should be recognized, verified, and administered to patients in a timely and appropriate manner to ensure quality care. We identified 3 goals to meet this objective: (1) STAT should be consistently defined and understood by everyone; (2) an easy, intuitive STAT process should be available for all stakeholders; and (3) the STAT process should be transparent and ideally visual so that everyone involved can understand at which point in the process a specific STAT order is currently situated. We also identified additional information we would need to reach the goals.

Shortly after the process‐mapping sessions, 2 workgroup members conducted real‐time STAT order time studies to track medications from order to administration. Three time periods in the STAT process were identified for observation and measurement: the time from physician order entry in the CPRS to the time a pharmacist verified the medication, the time from verification to when the medication arrived on the nursing unit, and the time from arrival on the nursing unit to when that medication was administered. Using a data‐collection template, each time period was recorded, and 28 time studies were collected over 1 month. To monitor the progress of our initiatives, the time study was repeated 3 months into the project.

ENVISIONING

Following the discovery phase, the team was better equipped to identify the specific changes needed to achieve an improved process. The envisioning phase allowed the team freedom to imagine an ideal process barring any preconceived notion of constraints within the current process.

In 2 meetings we brainstormed as many improvement ideas as possible. To prioritize and focus our ideas, we developed a matrix (see Supporting Information, Appendix A, in the online version of this article), placing our ideas in 1 of 4 quadrants based on the anticipated effort to implement the change (x‐axis) and impact of making the change (y‐axis). The matrix helped us see that some ideas would be relatively simple to implement (eg, color‐coded bags for STAT medication delivery), whereas others would require more sophisticated efforts and involvement of other people (eg, monthly education sessions to resident physicians).

EXPERIMENTING

Experiments were conducted to meet each of the 3 goals identified above. The team used the outcomes of the prioritization exercise to identify initial experiments to test. To build momentum by showing progress and improvement with a few quick wins, the team began with low‐effort/high‐impact opportunities. Each experiment followed a standard Plan‐Do‐Study‐Act (PDSA) cycle to encourage reflection, learning, adaptation, and adjustment as a result of the experiential learning process.[5]

RESULTS

At the start of our project, the average time from STAT order to medication administration was 1 hour and 7 minutes (range, 6 minutes 2 hours and 22 minutes). As a result of the 2 sets of countermeasures outlined in Tables 1 and 2, the average total time from STAT order entry to administration decreased by 21% to an average of 53 minutes. The total time from medication delivery to administration decreased by 26% from 34 minutes to 25 minutes postimplementation. On average, 391 STAT medications were ordered per month during the project period, which represents a decrease of 9.5% from the 432 orders per month for the same time period the previous year. After implementing the countermeasures in Table 2, we followed another 26 STAT medications through the process to evaluate our efforts. Of 15 STAT medications requiring delivery, only 1 nurse (7%) was not notified of the delivery of a STAT medication, and 1 pharmacy technician (7%) was not informed the medication was STAT. The 151% increase in notification of nurses to delivery of a STAT medication suggests that use of the STAT bags, STAT magnets on patient doors, and whenever possible direct delivery of STAT medications to the nurse has improved communication between the technicians and nurses. Similarly, the 27% increase in technician awareness of a STAT designation suggests STAT is being better communicated to them. The improvement in awareness and notification of a STAT medication is summarized in Figure 2.

Figure 2

Due to time and financial constraints, the following limitations may have affected our findings. First, resident physicians were not directly represented in our discussions. Attending medicine hospitalists provided the physician perspective, which provides a biased view given their intimate knowledge of the CPRS and additional years of experience. Similarly, nurse perspectives were limited to staff and clinical nurse leaders. Last, our low‐cost approach was mandated by limited resources; a more resource‐rich environment may have devised alternative approaches.

CONCLUSIONS

Adapting A3 Thinking for process improvement was a low‐cost/low‐tech option for a VA facility. Having buy‐in from all levels was crucial to the success of the project. The size and diversity of the group was also very important, as different opinions and aspects of the process were represented. Cross‐discipline relationships and respect were formed, which will be valuable for collaboration in future projects. Although we focused on the STAT medication process, other quality‐improvement projects could also benefit from A3 Thinking. Moreover, there were enough people to serve as ambassadors, taking the project back to their work areas to share with their peers, gather consensus, and elicit additional feedback. The collaboration led to comprehensive understanding of the process, the nature of the problems within the process, and the complexity of solving the problem. For example, although the number of STAT orders did not decrease dramatically, we have learned from these experiments that we may need to change how we approach structuring additional experiments. Future work will focus on increasing communication between physicians and nurses when placing STAT medication orders, enhancing resident education to ensure appropriate use of the STAT designation, and continuing our efforts to improve the delivery process of STAT medications.

Other quality‐improvement methodologies we could have used include: total quality management (TQM), continuous quality improvement (CQI), business process redesign, Lean, Six Sigma, and others.[14] Differences between these can be broadly classified as putting an emphasis on people (eg, inclusion of front line staff in CQI or leadership in TQM) or on process (eg, understanding process function to reduce waste in Lean or statistical process control in Six Sigma).[14] Using A3 Thinking methodology was more useful than these others for the STAT medication process for some very important reasons. The A3 process not only led to a better understanding of the meaning of STAT across disciplines, increasing the intuitive nature, transparency and visual aspects of the whole process, but also promoted a collaborative, multidisciplinary, integrative culture, in which other hospital‐wide problems may be addressed in the future.

Decision‐making capacity is a dynamic, integrative cognitive function necessary for informed consent. Capacity is assessed relative to a specific choice about medical care (eg, Does this patient with mild Alzheimer's disease have the capacity to decide whether to undergo valvuloplasty for severe aortic stenosis?), Capacity may be impaired by acute illnesses (eg, toxidromes and withdrawal states, medical illness‐related delirium, decompensated psychiatric episodes), as well as chronic conditions (eg, dementia, developmental disability, traumatic brain injuries, central nervous system (CNS) degenerative disorders). Given the proper training, clinicians from any specialty can assess a patient's decision‐making capacity.[1] A patient must satisfy 4 principles to have the capacity for a given decision: understanding of the condition, ability to communicate a choice, conception of the risks and benefits of the decision, and a rational approach to decision making.[2, 3, 4] Management of incapacitated persons may require consideration of the individual's stated or demonstrated preferences, medical ethics principles (eg, to consider the balance between autonomy, beneficence, and nonmaleficence during shared decision making), and institutional and situational norms and standards. Management may include immediate or long‐term medical and safety planning, and the selection of a surrogate decision maker or public guardian.[1, 2, 3, 4, 5, 6, 7, 8] A related term, competency, describes a legal judgment regarding a person's ability to make decisions, and persons deemed incompetent require an appointed guardian to make 1 or more types of decision (eg, medical, financial, and long‐term care planning).[1, 8]

Over one‐quarter of general medical inpatients display impaired decision‐making capacity based on a recent review of multiple studies.[2] Nursing home residents, persons with Alzheimer's dementia, and persons with developmental disabilitygroups commonly encountered in the inpatient settingdemonstrate impaired capacity in greater than 40% to 60% of cases.[2] Capacity impairment is present in three‐quarters of inpatients with life‐threatening illnesses.[5] The frequency of capacity impairment is complicated by the fact that physicians fail to recognize impaired capacity in as much as 60% of cases.[1, 2] Misunderstanding of the laws and medical and ethical principles related to capacity is common, even among specialists who commonly care for incapacitated patients, such as consult liaison psychiatrists, geriatricians, and psychologists.[1]

Loss of decision‐making capacity may be associated with negative consequences to the patient and to the provider‐patient dyad. Patients with capacity impairment have been shown to have an increased risk of mortality in a community setting.[6] Potential ethical pitfalls between provider and incapacitated patient have been described.[5] The high cost of long‐term management of subsets of incapacitated patients has also been noted.[7]

Improved identification and management of incapacitated patients has potential benefit to medical outcomes, patient safety, and cost containment.[6, 7, 9] The importance of education in this regard, especially to early career clinicians and to providers in specialties other than mental health, has been noted.[9] This article describes a clinical quality improvement project at San Francisco General Hospital and Trauma Center (SFGH) to improve provider identification and management of patients with impaired decision‐making capacity via a clinical decision algorithm.

METHODS

In 2012, the Department of Risk Management at SFGH created a multidisciplinary workgroup, including attending physicians, nurses, administrators, and hospital safety officers to improve the institutional process for identification and management of inpatients with impaired decision‐making capacity. The workgroup reviewed prior experience with incapacitated patients and data from multiple sources, including unusual occurrence reports, hospital root cause analyses, and hospital policies regarding patients with cognitive impairment. Expert opinion was solicited from attending psychiatry and neuropsychology providers.

SFGHan urban, academic, safety‐net hospitalcares for a diverse, underserved, and medically vulnerable patient population with high rates of cognitive and capacity impairment. A publication currently under review from SFGH shows that among a cohort of roughly 700 general medical inpatients 50 years and older, greater than 54% have mild or greater degrees of cognitive impairment based on the Telephone Interview for Cognitive Status test (unpublished data).[10] Among SFGH medical inpatients with extended lengths of stay, roughly one‐third have impaired capacity, require a family surrogate decision maker, or have an established public guardian (unpublished data). Among incapacitated patients, a particularly challenging subset have impaired decision making but significant physical capacity, creating risk of harm to self or others (eg, during the 18 months preintervention, an average of 9 incapacitated but physically capable inpatients per month attempted to leave SFGH prior to discharge) (unpublished data).

The majority of incapacitated patients at SFGH are cared for by 5 inpatient medical services staffed by resident and attending physicians from the University of California San Francisco: cardiology, family medicine, internal medicine, neurology, and psychiatry (unpublished data). Despite the commonality of capacity impairment on these services, education about capacity impairment and management was consistently reviewed only in the Department of Psychiatry.

Challenges common to prior experience with incapacitated patients were considered, including inefficient navigation of a complex, multistep identification and management process; difficulty addressing the high‐risk subset of incapacitated, able‐bodied patients who may pose an immediate safety risk; and incomplete understanding of the timing and indications for consultants (including psychiatry, neuropsychology, and medical ethics). To improve clinical outcome and patient safety through clinician identification and management, the workgroup created a clinical decision algorithm in a visual process map format for ease of use at the point of care.

Using MEDLINE and PubMed, the workgroup conducted a brief review of existing tools for incapacitated patients with relevant search terms and Medical Subjects Headings, including capacity, inpatient, shared decision making, mental competency, guideline, and algorithm. Publications reviewed included tools for capacity assessment (Addenbrooke's Cognitive Examination, MacArthur Competence Assessment Tool for Treatment)[2, 3, 4, 11] delineation of the basic process of capacity evaluation and subsequent management,[12, 13, 14, 15, 16] and explanation of the role of specialty consultation.[3, 9, 17] Specific attention was given to finding published visual algorithms; here, search results tended to focus on specialty consultation (eg, neuropsychology testing),[17] highly specific clinical situations (eg, sexual assault),[18] or to systems outside the United States.[19, 20, 21, 22] Byatt et al.'s work (2006) contains a useful visual algorithm about management of incapacitated patients, but it operates from the perspective of consult liaison psychiatrists, and the algorithm does not include principles of capacity assessment.[23] Derse ([16]) provides a text‐based algorithm relevant to primary inpatient providers, but does not have a visual illustration.[16] In our review, we were unable to find a visual algorithm that consolidates the process of identification, evaluation, and management of hospital inpatients with impaired decision‐making capacity.

Based on the described needs assessment, the workgroup created a draft algorithm for review by the SFGH medical executive committee, nursing quality council, and ethics committee.

RESULTS

The Clinical Decision Algorithm for Hospital Inpatients With Impaired Decision‐Making Capacity (adapted version, Figure 1) consolidates identification and management into a 1‐page visual process map, emphasizes safety planning for high‐risk patients, and explains indication and timing for multidisciplinary consultation, thereby addressing the 3 most prominent challenges based on our data and case review. Following hospital executive approval, the algorithm and a set of illustrative cases were disseminated to clinicians via email from service leadership, laminated copies were posted in housestaff workrooms, an electronic copy was posted on the website of the SFGH Department of Risk Management, and the algorithm was incorporated into hospital policy. Workgroup members conducted trainings with housestaff from the services identified as most frequently caring for incapacitated inpatients.

Figure 1

During trainings, housestaff participants expressed an improved sense of understanding and decreased anxiety about identification and management of incapacitated patients. During subsequent discussions, inpatient housestaff noted improvement in teamwork with safety officers, including cases involving agitated or threatening patients prior to capacity assessment.

An unexpected benefit of the algorithm was recognition of the need for associated resources, including a surrogate decision‐maker documentation form, off‐hours attending physician oversight for medical inpatients with capacity‐related emergencies, and a formal agreement with hospital safety officers regarding the care of high‐risk incapacitated patients not previously on a legal hold or surrogate guardianship. These were created in parallel with the algorithm and have become an integral part of management of incapacitated patients.

CLINICAL DECISION ALGORITHM APPLICATION TO PATIENT SCENARIOS

The following 3 scenarios exemplify common challenges in caring for inpatients with compromised decision‐making capacity. Assessment and multidisciplinary management are explained in relation to the clinical decision algorithm (Figure 1.)

DISCUSSION

Impaired decision‐making capacity is a common and challenging condition among hospitalized patients, including at our institution. Prior studies show that physicians frequently fail to recognize capacity impairment, and also demonstrate common misunderstandings about the medicolegal framework that governs capacity determination and subsequent care. Patients with impaired decision‐making capacity are vulnerable to adverse outcomes, and there is potential for negative effects on healthcare systems. The management of patients with impaired capacity may involve multiple disciplines and a complex intersection of medical, legal, ethical, and neuropsychological principles.

To promote safety of this vulnerable population at SFGH, our workgroup created a visual algorithm to guide clinicians. The algorithm may improve on existing tools by consolidating the steps from identification through management into a 1‐page visual tool, by emphasizing safety planning for high‐risk incapacitated patients and by elucidating roles and timing for other members of the multidisciplinary management team. Creation of the algorithm facilitated intervention for other practical issues, including institutional and departmental agreements and documentation regarding surrogate decision makers for incapacitated patients.

Although based on a multispecialty institutional review and previously published tools, there are potential limitations to this tool. It seems reasonable to assume that a tool to organize a complex process, such as identification and management of incapacitated patients, should improve patient care versus a non‐standardized process. Although the algorithm is posted in resident workrooms, on the hospital's risk management website, and incorporated as part of hospital policy, we have not yet had the opportunity to study the frequency of its use and impact in patient care. Patient safety and clinical outcome of patients managed with this algorithm could be assessed; however, the impact of the algorithm at SFGH may be confounded by a separate intervention addressing nursing and safety officers that was initiated shortly after the algorithm was produced.

To assess health‐system effects of incapacitated patients, future studies might compare patients with capacity impairment versus those with intact decision making relative to demographic background and payer mix, rates of adverse events during inpatient stay (eg, hospital‐acquired injury), rates of morbidity and mortality, rate of provider identification and documentation of surrogates, patient and surrogate satisfaction data, length of stay and cost of hospitalization, and rates of successful discharge to a community‐based setting. We present this algorithm as an example for diverse settings to address the common challenge of caring for acutely ill patients with impaired decision‐making capacity.

Disclosures

Nothing to report.

Decision‐making capacity is a dynamic, integrative cognitive function necessary for informed consent. Capacity is assessed relative to a specific choice about medical care (eg, Does this patient with mild Alzheimer's disease have the capacity to decide whether to undergo valvuloplasty for severe aortic stenosis?), Capacity may be impaired by acute illnesses (eg, toxidromes and withdrawal states, medical illness‐related delirium, decompensated psychiatric episodes), as well as chronic conditions (eg, dementia, developmental disability, traumatic brain injuries, central nervous system (CNS) degenerative disorders). Given the proper training, clinicians from any specialty can assess a patient's decision‐making capacity.[1] A patient must satisfy 4 principles to have the capacity for a given decision: understanding of the condition, ability to communicate a choice, conception of the risks and benefits of the decision, and a rational approach to decision making.[2, 3, 4] Management of incapacitated persons may require consideration of the individual's stated or demonstrated preferences, medical ethics principles (eg, to consider the balance between autonomy, beneficence, and nonmaleficence during shared decision making), and institutional and situational norms and standards. Management may include immediate or long‐term medical and safety planning, and the selection of a surrogate decision maker or public guardian.[1, 2, 3, 4, 5, 6, 7, 8] A related term, competency, describes a legal judgment regarding a person's ability to make decisions, and persons deemed incompetent require an appointed guardian to make 1 or more types of decision (eg, medical, financial, and long‐term care planning).[1, 8]

Over one‐quarter of general medical inpatients display impaired decision‐making capacity based on a recent review of multiple studies.[2] Nursing home residents, persons with Alzheimer's dementia, and persons with developmental disabilitygroups commonly encountered in the inpatient settingdemonstrate impaired capacity in greater than 40% to 60% of cases.[2] Capacity impairment is present in three‐quarters of inpatients with life‐threatening illnesses.[5] The frequency of capacity impairment is complicated by the fact that physicians fail to recognize impaired capacity in as much as 60% of cases.[1, 2] Misunderstanding of the laws and medical and ethical principles related to capacity is common, even among specialists who commonly care for incapacitated patients, such as consult liaison psychiatrists, geriatricians, and psychologists.[1]

Loss of decision‐making capacity may be associated with negative consequences to the patient and to the provider‐patient dyad. Patients with capacity impairment have been shown to have an increased risk of mortality in a community setting.[6] Potential ethical pitfalls between provider and incapacitated patient have been described.[5] The high cost of long‐term management of subsets of incapacitated patients has also been noted.[7]

Improved identification and management of incapacitated patients has potential benefit to medical outcomes, patient safety, and cost containment.[6, 7, 9] The importance of education in this regard, especially to early career clinicians and to providers in specialties other than mental health, has been noted.[9] This article describes a clinical quality improvement project at San Francisco General Hospital and Trauma Center (SFGH) to improve provider identification and management of patients with impaired decision‐making capacity via a clinical decision algorithm.

METHODS

In 2012, the Department of Risk Management at SFGH created a multidisciplinary workgroup, including attending physicians, nurses, administrators, and hospital safety officers to improve the institutional process for identification and management of inpatients with impaired decision‐making capacity. The workgroup reviewed prior experience with incapacitated patients and data from multiple sources, including unusual occurrence reports, hospital root cause analyses, and hospital policies regarding patients with cognitive impairment. Expert opinion was solicited from attending psychiatry and neuropsychology providers.

SFGHan urban, academic, safety‐net hospitalcares for a diverse, underserved, and medically vulnerable patient population with high rates of cognitive and capacity impairment. A publication currently under review from SFGH shows that among a cohort of roughly 700 general medical inpatients 50 years and older, greater than 54% have mild or greater degrees of cognitive impairment based on the Telephone Interview for Cognitive Status test (unpublished data).[10] Among SFGH medical inpatients with extended lengths of stay, roughly one‐third have impaired capacity, require a family surrogate decision maker, or have an established public guardian (unpublished data). Among incapacitated patients, a particularly challenging subset have impaired decision making but significant physical capacity, creating risk of harm to self or others (eg, during the 18 months preintervention, an average of 9 incapacitated but physically capable inpatients per month attempted to leave SFGH prior to discharge) (unpublished data).

The majority of incapacitated patients at SFGH are cared for by 5 inpatient medical services staffed by resident and attending physicians from the University of California San Francisco: cardiology, family medicine, internal medicine, neurology, and psychiatry (unpublished data). Despite the commonality of capacity impairment on these services, education about capacity impairment and management was consistently reviewed only in the Department of Psychiatry.

Challenges common to prior experience with incapacitated patients were considered, including inefficient navigation of a complex, multistep identification and management process; difficulty addressing the high‐risk subset of incapacitated, able‐bodied patients who may pose an immediate safety risk; and incomplete understanding of the timing and indications for consultants (including psychiatry, neuropsychology, and medical ethics). To improve clinical outcome and patient safety through clinician identification and management, the workgroup created a clinical decision algorithm in a visual process map format for ease of use at the point of care.

Using MEDLINE and PubMed, the workgroup conducted a brief review of existing tools for incapacitated patients with relevant search terms and Medical Subjects Headings, including capacity, inpatient, shared decision making, mental competency, guideline, and algorithm. Publications reviewed included tools for capacity assessment (Addenbrooke's Cognitive Examination, MacArthur Competence Assessment Tool for Treatment)[2, 3, 4, 11] delineation of the basic process of capacity evaluation and subsequent management,[12, 13, 14, 15, 16] and explanation of the role of specialty consultation.[3, 9, 17] Specific attention was given to finding published visual algorithms; here, search results tended to focus on specialty consultation (eg, neuropsychology testing),[17] highly specific clinical situations (eg, sexual assault),[18] or to systems outside the United States.[19, 20, 21, 22] Byatt et al.'s work (2006) contains a useful visual algorithm about management of incapacitated patients, but it operates from the perspective of consult liaison psychiatrists, and the algorithm does not include principles of capacity assessment.[23] Derse ([16]) provides a text‐based algorithm relevant to primary inpatient providers, but does not have a visual illustration.[16] In our review, we were unable to find a visual algorithm that consolidates the process of identification, evaluation, and management of hospital inpatients with impaired decision‐making capacity.

Based on the described needs assessment, the workgroup created a draft algorithm for review by the SFGH medical executive committee, nursing quality council, and ethics committee.

RESULTS

The Clinical Decision Algorithm for Hospital Inpatients With Impaired Decision‐Making Capacity (adapted version, Figure 1) consolidates identification and management into a 1‐page visual process map, emphasizes safety planning for high‐risk patients, and explains indication and timing for multidisciplinary consultation, thereby addressing the 3 most prominent challenges based on our data and case review. Following hospital executive approval, the algorithm and a set of illustrative cases were disseminated to clinicians via email from service leadership, laminated copies were posted in housestaff workrooms, an electronic copy was posted on the website of the SFGH Department of Risk Management, and the algorithm was incorporated into hospital policy. Workgroup members conducted trainings with housestaff from the services identified as most frequently caring for incapacitated inpatients.

Figure 1

During trainings, housestaff participants expressed an improved sense of understanding and decreased anxiety about identification and management of incapacitated patients. During subsequent discussions, inpatient housestaff noted improvement in teamwork with safety officers, including cases involving agitated or threatening patients prior to capacity assessment.

An unexpected benefit of the algorithm was recognition of the need for associated resources, including a surrogate decision‐maker documentation form, off‐hours attending physician oversight for medical inpatients with capacity‐related emergencies, and a formal agreement with hospital safety officers regarding the care of high‐risk incapacitated patients not previously on a legal hold or surrogate guardianship. These were created in parallel with the algorithm and have become an integral part of management of incapacitated patients.

CLINICAL DECISION ALGORITHM APPLICATION TO PATIENT SCENARIOS

The following 3 scenarios exemplify common challenges in caring for inpatients with compromised decision‐making capacity. Assessment and multidisciplinary management are explained in relation to the clinical decision algorithm (Figure 1.)

DISCUSSION

Impaired decision‐making capacity is a common and challenging condition among hospitalized patients, including at our institution. Prior studies show that physicians frequently fail to recognize capacity impairment, and also demonstrate common misunderstandings about the medicolegal framework that governs capacity determination and subsequent care. Patients with impaired decision‐making capacity are vulnerable to adverse outcomes, and there is potential for negative effects on healthcare systems. The management of patients with impaired capacity may involve multiple disciplines and a complex intersection of medical, legal, ethical, and neuropsychological principles.

To promote safety of this vulnerable population at SFGH, our workgroup created a visual algorithm to guide clinicians. The algorithm may improve on existing tools by consolidating the steps from identification through management into a 1‐page visual tool, by emphasizing safety planning for high‐risk incapacitated patients and by elucidating roles and timing for other members of the multidisciplinary management team. Creation of the algorithm facilitated intervention for other practical issues, including institutional and departmental agreements and documentation regarding surrogate decision makers for incapacitated patients.

Although based on a multispecialty institutional review and previously published tools, there are potential limitations to this tool. It seems reasonable to assume that a tool to organize a complex process, such as identification and management of incapacitated patients, should improve patient care versus a non‐standardized process. Although the algorithm is posted in resident workrooms, on the hospital's risk management website, and incorporated as part of hospital policy, we have not yet had the opportunity to study the frequency of its use and impact in patient care. Patient safety and clinical outcome of patients managed with this algorithm could be assessed; however, the impact of the algorithm at SFGH may be confounded by a separate intervention addressing nursing and safety officers that was initiated shortly after the algorithm was produced.

To assess health‐system effects of incapacitated patients, future studies might compare patients with capacity impairment versus those with intact decision making relative to demographic background and payer mix, rates of adverse events during inpatient stay (eg, hospital‐acquired injury), rates of morbidity and mortality, rate of provider identification and documentation of surrogates, patient and surrogate satisfaction data, length of stay and cost of hospitalization, and rates of successful discharge to a community‐based setting. We present this algorithm as an example for diverse settings to address the common challenge of caring for acutely ill patients with impaired decision‐making capacity.

Disclosures

Nothing to report.

Decision‐making capacity is a dynamic, integrative cognitive function necessary for informed consent. Capacity is assessed relative to a specific choice about medical care (eg, Does this patient with mild Alzheimer's disease have the capacity to decide whether to undergo valvuloplasty for severe aortic stenosis?), Capacity may be impaired by acute illnesses (eg, toxidromes and withdrawal states, medical illness‐related delirium, decompensated psychiatric episodes), as well as chronic conditions (eg, dementia, developmental disability, traumatic brain injuries, central nervous system (CNS) degenerative disorders). Given the proper training, clinicians from any specialty can assess a patient's decision‐making capacity.[1] A patient must satisfy 4 principles to have the capacity for a given decision: understanding of the condition, ability to communicate a choice, conception of the risks and benefits of the decision, and a rational approach to decision making.[2, 3, 4] Management of incapacitated persons may require consideration of the individual's stated or demonstrated preferences, medical ethics principles (eg, to consider the balance between autonomy, beneficence, and nonmaleficence during shared decision making), and institutional and situational norms and standards. Management may include immediate or long‐term medical and safety planning, and the selection of a surrogate decision maker or public guardian.[1, 2, 3, 4, 5, 6, 7, 8] A related term, competency, describes a legal judgment regarding a person's ability to make decisions, and persons deemed incompetent require an appointed guardian to make 1 or more types of decision (eg, medical, financial, and long‐term care planning).[1, 8]

Over one‐quarter of general medical inpatients display impaired decision‐making capacity based on a recent review of multiple studies.[2] Nursing home residents, persons with Alzheimer's dementia, and persons with developmental disabilitygroups commonly encountered in the inpatient settingdemonstrate impaired capacity in greater than 40% to 60% of cases.[2] Capacity impairment is present in three‐quarters of inpatients with life‐threatening illnesses.[5] The frequency of capacity impairment is complicated by the fact that physicians fail to recognize impaired capacity in as much as 60% of cases.[1, 2] Misunderstanding of the laws and medical and ethical principles related to capacity is common, even among specialists who commonly care for incapacitated patients, such as consult liaison psychiatrists, geriatricians, and psychologists.[1]

Loss of decision‐making capacity may be associated with negative consequences to the patient and to the provider‐patient dyad. Patients with capacity impairment have been shown to have an increased risk of mortality in a community setting.[6] Potential ethical pitfalls between provider and incapacitated patient have been described.[5] The high cost of long‐term management of subsets of incapacitated patients has also been noted.[7]

Improved identification and management of incapacitated patients has potential benefit to medical outcomes, patient safety, and cost containment.[6, 7, 9] The importance of education in this regard, especially to early career clinicians and to providers in specialties other than mental health, has been noted.[9] This article describes a clinical quality improvement project at San Francisco General Hospital and Trauma Center (SFGH) to improve provider identification and management of patients with impaired decision‐making capacity via a clinical decision algorithm.

METHODS

In 2012, the Department of Risk Management at SFGH created a multidisciplinary workgroup, including attending physicians, nurses, administrators, and hospital safety officers to improve the institutional process for identification and management of inpatients with impaired decision‐making capacity. The workgroup reviewed prior experience with incapacitated patients and data from multiple sources, including unusual occurrence reports, hospital root cause analyses, and hospital policies regarding patients with cognitive impairment. Expert opinion was solicited from attending psychiatry and neuropsychology providers.

SFGHan urban, academic, safety‐net hospitalcares for a diverse, underserved, and medically vulnerable patient population with high rates of cognitive and capacity impairment. A publication currently under review from SFGH shows that among a cohort of roughly 700 general medical inpatients 50 years and older, greater than 54% have mild or greater degrees of cognitive impairment based on the Telephone Interview for Cognitive Status test (unpublished data).[10] Among SFGH medical inpatients with extended lengths of stay, roughly one‐third have impaired capacity, require a family surrogate decision maker, or have an established public guardian (unpublished data). Among incapacitated patients, a particularly challenging subset have impaired decision making but significant physical capacity, creating risk of harm to self or others (eg, during the 18 months preintervention, an average of 9 incapacitated but physically capable inpatients per month attempted to leave SFGH prior to discharge) (unpublished data).

The majority of incapacitated patients at SFGH are cared for by 5 inpatient medical services staffed by resident and attending physicians from the University of California San Francisco: cardiology, family medicine, internal medicine, neurology, and psychiatry (unpublished data). Despite the commonality of capacity impairment on these services, education about capacity impairment and management was consistently reviewed only in the Department of Psychiatry.

Challenges common to prior experience with incapacitated patients were considered, including inefficient navigation of a complex, multistep identification and management process; difficulty addressing the high‐risk subset of incapacitated, able‐bodied patients who may pose an immediate safety risk; and incomplete understanding of the timing and indications for consultants (including psychiatry, neuropsychology, and medical ethics). To improve clinical outcome and patient safety through clinician identification and management, the workgroup created a clinical decision algorithm in a visual process map format for ease of use at the point of care.

Using MEDLINE and PubMed, the workgroup conducted a brief review of existing tools for incapacitated patients with relevant search terms and Medical Subjects Headings, including capacity, inpatient, shared decision making, mental competency, guideline, and algorithm. Publications reviewed included tools for capacity assessment (Addenbrooke's Cognitive Examination, MacArthur Competence Assessment Tool for Treatment)[2, 3, 4, 11] delineation of the basic process of capacity evaluation and subsequent management,[12, 13, 14, 15, 16] and explanation of the role of specialty consultation.[3, 9, 17] Specific attention was given to finding published visual algorithms; here, search results tended to focus on specialty consultation (eg, neuropsychology testing),[17] highly specific clinical situations (eg, sexual assault),[18] or to systems outside the United States.[19, 20, 21, 22] Byatt et al.'s work (2006) contains a useful visual algorithm about management of incapacitated patients, but it operates from the perspective of consult liaison psychiatrists, and the algorithm does not include principles of capacity assessment.[23] Derse ([16]) provides a text‐based algorithm relevant to primary inpatient providers, but does not have a visual illustration.[16] In our review, we were unable to find a visual algorithm that consolidates the process of identification, evaluation, and management of hospital inpatients with impaired decision‐making capacity.

Based on the described needs assessment, the workgroup created a draft algorithm for review by the SFGH medical executive committee, nursing quality council, and ethics committee.

RESULTS

The Clinical Decision Algorithm for Hospital Inpatients With Impaired Decision‐Making Capacity (adapted version, Figure 1) consolidates identification and management into a 1‐page visual process map, emphasizes safety planning for high‐risk patients, and explains indication and timing for multidisciplinary consultation, thereby addressing the 3 most prominent challenges based on our data and case review. Following hospital executive approval, the algorithm and a set of illustrative cases were disseminated to clinicians via email from service leadership, laminated copies were posted in housestaff workrooms, an electronic copy was posted on the website of the SFGH Department of Risk Management, and the algorithm was incorporated into hospital policy. Workgroup members conducted trainings with housestaff from the services identified as most frequently caring for incapacitated inpatients.

Figure 1

During trainings, housestaff participants expressed an improved sense of understanding and decreased anxiety about identification and management of incapacitated patients. During subsequent discussions, inpatient housestaff noted improvement in teamwork with safety officers, including cases involving agitated or threatening patients prior to capacity assessment.

An unexpected benefit of the algorithm was recognition of the need for associated resources, including a surrogate decision‐maker documentation form, off‐hours attending physician oversight for medical inpatients with capacity‐related emergencies, and a formal agreement with hospital safety officers regarding the care of high‐risk incapacitated patients not previously on a legal hold or surrogate guardianship. These were created in parallel with the algorithm and have become an integral part of management of incapacitated patients.

CLINICAL DECISION ALGORITHM APPLICATION TO PATIENT SCENARIOS

The following 3 scenarios exemplify common challenges in caring for inpatients with compromised decision‐making capacity. Assessment and multidisciplinary management are explained in relation to the clinical decision algorithm (Figure 1.)

DISCUSSION

Impaired decision‐making capacity is a common and challenging condition among hospitalized patients, including at our institution. Prior studies show that physicians frequently fail to recognize capacity impairment, and also demonstrate common misunderstandings about the medicolegal framework that governs capacity determination and subsequent care. Patients with impaired decision‐making capacity are vulnerable to adverse outcomes, and there is potential for negative effects on healthcare systems. The management of patients with impaired capacity may involve multiple disciplines and a complex intersection of medical, legal, ethical, and neuropsychological principles.

To promote safety of this vulnerable population at SFGH, our workgroup created a visual algorithm to guide clinicians. The algorithm may improve on existing tools by consolidating the steps from identification through management into a 1‐page visual tool, by emphasizing safety planning for high‐risk incapacitated patients and by elucidating roles and timing for other members of the multidisciplinary management team. Creation of the algorithm facilitated intervention for other practical issues, including institutional and departmental agreements and documentation regarding surrogate decision makers for incapacitated patients.

Although based on a multispecialty institutional review and previously published tools, there are potential limitations to this tool. It seems reasonable to assume that a tool to organize a complex process, such as identification and management of incapacitated patients, should improve patient care versus a non‐standardized process. Although the algorithm is posted in resident workrooms, on the hospital's risk management website, and incorporated as part of hospital policy, we have not yet had the opportunity to study the frequency of its use and impact in patient care. Patient safety and clinical outcome of patients managed with this algorithm could be assessed; however, the impact of the algorithm at SFGH may be confounded by a separate intervention addressing nursing and safety officers that was initiated shortly after the algorithm was produced.

To assess health‐system effects of incapacitated patients, future studies might compare patients with capacity impairment versus those with intact decision making relative to demographic background and payer mix, rates of adverse events during inpatient stay (eg, hospital‐acquired injury), rates of morbidity and mortality, rate of provider identification and documentation of surrogates, patient and surrogate satisfaction data, length of stay and cost of hospitalization, and rates of successful discharge to a community‐based setting. We present this algorithm as an example for diverse settings to address the common challenge of caring for acutely ill patients with impaired decision‐making capacity.

Disclosures

Nothing to report.

According to the American Academy of Pediatrics clinical report on physicians' roles in coordinating care of hospitalized children, there are several important components of hospital discharge planning.[1] Foremost is that discharge planning should begin, and discharge criteria should be set, at the time of hospital admission. This allows for optimal engagement of parents and providers in the effort to adequately prepare patients for the transition to home.

As pediatric inpatients become increasingly complex,[2] adequately preparing families for the transition to home becomes more challenging.[3] There are a myriad of issues to address and the burden of this preparation effort falls on multiple individuals other than the bedside nurse and physician. Large multidisciplinary teams often play a significant role in the discharge of medically complex children.[4] Several challenges may hinder the team's ability to effectively navigate the discharge process such as financial or insurance‐related issues, language differences, or geographic barriers. Patient and family anxieties may also complicate the transition to home.[5]

The challenges of a multidisciplinary approach to discharge planning are further magnified by the limitations of the electronic health record (EHR). The EHR is well designed to record individual encounters, but poorly designed to coordinate longitudinal care across settings.[6] Although multidisciplinary providers may spend significant and well‐intentioned energy to facilitate hospital discharge, their efforts may go unseen or be duplicative.

We developed a discharge readiness report (DRR) for the EHR, an integrated summary of discharge‐related issues, organized into a highly visible and easily accessible report. The development of the discharge planning tool was the first step in a larger quality improvement (QI) initiative aimed at improving the efficiency, effectiveness, and safety of hospital discharge. Our team recognized that improving the flow and visibility of information between disciplines was the first step toward accomplishing this larger aim. Health information technology offers an important opportunity for the improvement of patient safety and care transitions⁷; therefore, we leveraged the EHR to create an integrated discharge report. We used QI methods to understand our hospital's discharge processes, examined potential pitfalls in interdisciplinary communication, determined relevant information to include in the report, and optimized ways to display the data. To our knowledge, this use of the EHR is novel. The objectives of this article were to describe our team's development and implementation strategies, as well as challenges encountered, in the design of this electronic discharge planning tool.

METHODS

Planning the Intervention

Based on our learning, we developed a key driver diagram (Figure 1). Our aim was to create a DRR that organized important discharge‐related information into 1 easily accessible report. Key drivers that were identified as relevant to the content of the DRR included: barriers to discharge, discharge criteria, home care, postdischarge care, and last minute delays. We also identified secondary drivers related to the design of the DRR. We hypothesized that addressing the secondary drivers would be essential to end user adoption of the tool. The secondary drivers included: accessibility, relevance, ease of updating, automation, and readability.

Figure 1

With the swim lane diagram as well as our primary and secondary drivers in mind, we created a mock DRR on paper. We conducted multiple patient discharge simulations with representatives from all disciplines, walking through each step of a patient hospitalization from registration to discharge. This allowed us to map out how preexisting, yet disparate, EHR data could be channeled into 1 report. A few changes were made to processes involving data collection and documentation to facilitate timely transfer of information to the report. For example, questions addressing potential barriers to discharge and whether a school/work excuse was needed were added to the admission nursing assessment.

We then moved the paper DRR to the electronic environment. Data elements that were pulled automatically into the report included: potential barriers to discharge collected during nursing intake, case management information on home care needs, discharge criteria entered by resident and attending physicians, PCP, home pharmacy, follow‐up appointments, school/work excuse information gathered by resident liaisons, and active patient problems drawn from the problem list section. These data were organized into 4 distinct domains within the final DRR: potential barriers, transitional care, home care, and discharge criteria (Table 1).

Table 1.Discharge Readiness Report Domains

Discharge Readiness Report Domain	Example Content
NOTE: Abbreviations: PCP, primary care provider.
Potential barriers to discharge	Geographic location of the family, whether patient lives in more than 1 household, primary spoken language, financial or insurance concern, and need for work/subhool excuses
Transitional care	PCP and home pharmacy information, follow‐up ambulatory and imaging appointments, and care team communications with the PCP
Home care	Planned discharge date/time and home care needs assessments such as needs for special equipment or skilled home nursing
Discharge criteria	Clinical, social, or other care coordination conditions for discharge

Additional features potentially important to discharge planning were also incorporated into the report based on end user feedback. These included hyperlinks to discharge orders, home oxygen prescriptions, and the after‐visit summary for families, and the patient's home care company (if present). To facilitate discharge and transitional care related communication between the primary team and subspecialty teams, consults involved during the hospitalization were included on the report. As home care arrangements often involve care for active lines and drains, they were added to the report (Figure 2).

Figure 2

Implementation

The report was activated within the EHR in June 2012. The team focused initial promotion and education efforts on medical floors. Education was widely disseminated via email and in‐person presentations.

The DRR was incorporated into daily CCRs for medical patients in July 2012. These multidisciplinary rounds occurred after medical‐team bedside rounds, focusing on care coordination and discharge planning. For each patient discussed, the DRR was projected onto a large screen, allowing all team members to view and discuss relevant discharge information. A process improvement (PI) specialist attended CCRs daily for several months, educating participants and monitoring use of the DRR. The PI specialist solicited feedback on ways to improve the DRR, and timed rounds to measure whether use of the DRR prolonged CCRs.

In the first weeks postimplementation, the use of the DRR prolonged rounds by as much as 1 minute per patient. Based on direct observation, the team focused interventions on barriers to the efficient use of the report during CCRs including: the need to scroll through the report, which was not visible on 1 screen; the need to navigate between patients; the need to quickly update the report based on discussion; and the need to update discharge criteria (Figure 3).

Figure 3

RESULTS

Creation of the final DRR required significant time and effort and was the culmination of a uniquely collaborative effort between clinicians, ancillary staff, and information technology specialists (Figure 4). The report is used consistently for all general medical and medical subspecialty patients during CCRs. After interventions were implemented to improve the efficiency of using the DRR during CCRs, the use of the DRR did not prolong CCRs. Members of the care team acknowledge that all sections of the report are populated and accurate. Though end users have commented on their use of the report outside of CCRs, we have not been able to formally measure this.

Figure 4

We have noticed a shift in the focus of discussion since implementation of the DRR. Prior to this initiative, care teams at our institution did not regularly discuss discharge criteria during bedside or CCRs. The phrase discharge criteria has now become part of our shared language.

Informally, the DRR appears to have reduced inefficiency and the potential for communication error. The practice of writing notes on printed patient lists to be used to sign‐out or communicate to other team members not in attendance at CCRs has largely disappeared.

The DRR has proven to be adaptable across patient units, and can be tailored to the specific transitional care needs of a given patient population. At discharge institution, the DRR has been modified for, and has taken on a prominent role in, the discharge planning of highly complex populations such as rehabilitation and ventilated patients.

DISCUSSION

Discharge planning is a multifaceted, multidisciplinary process that should begin at the time of hospital admission. Safe patient transition depends on efficient discharge processes and effective communication across settings.[8] Although not well studied in the inpatient setting, care process variability can result in inefficient patient flow and increased stress among staff.[9] Patients and families may experience confusion, coping difficulties, and increased readmission due to ineffective discharge planning.[10] These potential pitfalls highlight the need for healthcare providers to develop patient‐centered, systematic approaches to improving the discharge process.[11]

To our knowledge, this is the first description of a discharge planning tool for the EHR in the pediatric setting. Our discharge report is centralized, easily accessible by all members of the care team, and includes important patient‐specific discharge‐related information that be used to focus discussion and streamline multidisciplinary discharge planning rounds.

We anticipate that the report will allow the entire healthcare team to function more efficiently, decrease discharge‐related delays and failures based on communication roadblocks, and improve family and caregiver satisfaction with the discharge process. We are currently testing these hypotheses and evaluating several implementation strategies in an ongoing research study. Assuming positive impact, we plan to spread the use of the DRR to all inpatient care areas at our hospital, and potentially to other hospitals.

The limitations of this QI project are consistent with other initiatives to improve care. The challenges we encounter at our freestanding tertiary care teaching hospital with regard to effective discharge planning and multidisciplinary communication may not be generalizable to other nonteaching or community hospitals, and the DRR may not be useful in other settings. Though the report is now a part of our EHR, the most impactful implementation strategies remain to be determined. The report and related changes represent significant diversion from years of deeply ingrained workflows for some providers, and we encountered some resistance from staff during the early stages of implementation. The most important of which was that some team members are uncomfortable with technology and prefer to use paper. Most of this initial resistance was overcome by implementing changes to improve the ease of use of the report (Figure 3). Though input from end users and key stakeholders has been incorporated throughout this initiative, more work is needed to measure end user adoption and satisfaction with the report.

CONCLUSION

High‐quality hospital discharge planning requires an increasingly multidisciplinary approach. The EHR can be leveraged to improve transparency and interdisciplinary communication around the discharge process. An integrated summary of discharge‐related issues, organized into 1 highly visible and easily accessible report in the EHR has the potential to improve care transitions.

According to the American Academy of Pediatrics clinical report on physicians' roles in coordinating care of hospitalized children, there are several important components of hospital discharge planning.[1] Foremost is that discharge planning should begin, and discharge criteria should be set, at the time of hospital admission. This allows for optimal engagement of parents and providers in the effort to adequately prepare patients for the transition to home.

As pediatric inpatients become increasingly complex,[2] adequately preparing families for the transition to home becomes more challenging.[3] There are a myriad of issues to address and the burden of this preparation effort falls on multiple individuals other than the bedside nurse and physician. Large multidisciplinary teams often play a significant role in the discharge of medically complex children.[4] Several challenges may hinder the team's ability to effectively navigate the discharge process such as financial or insurance‐related issues, language differences, or geographic barriers. Patient and family anxieties may also complicate the transition to home.[5]

The challenges of a multidisciplinary approach to discharge planning are further magnified by the limitations of the electronic health record (EHR). The EHR is well designed to record individual encounters, but poorly designed to coordinate longitudinal care across settings.[6] Although multidisciplinary providers may spend significant and well‐intentioned energy to facilitate hospital discharge, their efforts may go unseen or be duplicative.

We developed a discharge readiness report (DRR) for the EHR, an integrated summary of discharge‐related issues, organized into a highly visible and easily accessible report. The development of the discharge planning tool was the first step in a larger quality improvement (QI) initiative aimed at improving the efficiency, effectiveness, and safety of hospital discharge. Our team recognized that improving the flow and visibility of information between disciplines was the first step toward accomplishing this larger aim. Health information technology offers an important opportunity for the improvement of patient safety and care transitions⁷; therefore, we leveraged the EHR to create an integrated discharge report. We used QI methods to understand our hospital's discharge processes, examined potential pitfalls in interdisciplinary communication, determined relevant information to include in the report, and optimized ways to display the data. To our knowledge, this use of the EHR is novel. The objectives of this article were to describe our team's development and implementation strategies, as well as challenges encountered, in the design of this electronic discharge planning tool.

METHODS

Planning the Intervention

Based on our learning, we developed a key driver diagram (Figure 1). Our aim was to create a DRR that organized important discharge‐related information into 1 easily accessible report. Key drivers that were identified as relevant to the content of the DRR included: barriers to discharge, discharge criteria, home care, postdischarge care, and last minute delays. We also identified secondary drivers related to the design of the DRR. We hypothesized that addressing the secondary drivers would be essential to end user adoption of the tool. The secondary drivers included: accessibility, relevance, ease of updating, automation, and readability.

Figure 1

With the swim lane diagram as well as our primary and secondary drivers in mind, we created a mock DRR on paper. We conducted multiple patient discharge simulations with representatives from all disciplines, walking through each step of a patient hospitalization from registration to discharge. This allowed us to map out how preexisting, yet disparate, EHR data could be channeled into 1 report. A few changes were made to processes involving data collection and documentation to facilitate timely transfer of information to the report. For example, questions addressing potential barriers to discharge and whether a school/work excuse was needed were added to the admission nursing assessment.

We then moved the paper DRR to the electronic environment. Data elements that were pulled automatically into the report included: potential barriers to discharge collected during nursing intake, case management information on home care needs, discharge criteria entered by resident and attending physicians, PCP, home pharmacy, follow‐up appointments, school/work excuse information gathered by resident liaisons, and active patient problems drawn from the problem list section. These data were organized into 4 distinct domains within the final DRR: potential barriers, transitional care, home care, and discharge criteria (Table 1).

Table 1.Discharge Readiness Report Domains

Discharge Readiness Report Domain	Example Content
NOTE: Abbreviations: PCP, primary care provider.
Potential barriers to discharge	Geographic location of the family, whether patient lives in more than 1 household, primary spoken language, financial or insurance concern, and need for work/subhool excuses
Transitional care	PCP and home pharmacy information, follow‐up ambulatory and imaging appointments, and care team communications with the PCP
Home care	Planned discharge date/time and home care needs assessments such as needs for special equipment or skilled home nursing
Discharge criteria	Clinical, social, or other care coordination conditions for discharge

Additional features potentially important to discharge planning were also incorporated into the report based on end user feedback. These included hyperlinks to discharge orders, home oxygen prescriptions, and the after‐visit summary for families, and the patient's home care company (if present). To facilitate discharge and transitional care related communication between the primary team and subspecialty teams, consults involved during the hospitalization were included on the report. As home care arrangements often involve care for active lines and drains, they were added to the report (Figure 2).

Figure 2

Implementation

The report was activated within the EHR in June 2012. The team focused initial promotion and education efforts on medical floors. Education was widely disseminated via email and in‐person presentations.

The DRR was incorporated into daily CCRs for medical patients in July 2012. These multidisciplinary rounds occurred after medical‐team bedside rounds, focusing on care coordination and discharge planning. For each patient discussed, the DRR was projected onto a large screen, allowing all team members to view and discuss relevant discharge information. A process improvement (PI) specialist attended CCRs daily for several months, educating participants and monitoring use of the DRR. The PI specialist solicited feedback on ways to improve the DRR, and timed rounds to measure whether use of the DRR prolonged CCRs.

In the first weeks postimplementation, the use of the DRR prolonged rounds by as much as 1 minute per patient. Based on direct observation, the team focused interventions on barriers to the efficient use of the report during CCRs including: the need to scroll through the report, which was not visible on 1 screen; the need to navigate between patients; the need to quickly update the report based on discussion; and the need to update discharge criteria (Figure 3).

Figure 3

RESULTS

Creation of the final DRR required significant time and effort and was the culmination of a uniquely collaborative effort between clinicians, ancillary staff, and information technology specialists (Figure 4). The report is used consistently for all general medical and medical subspecialty patients during CCRs. After interventions were implemented to improve the efficiency of using the DRR during CCRs, the use of the DRR did not prolong CCRs. Members of the care team acknowledge that all sections of the report are populated and accurate. Though end users have commented on their use of the report outside of CCRs, we have not been able to formally measure this.

Figure 4

We have noticed a shift in the focus of discussion since implementation of the DRR. Prior to this initiative, care teams at our institution did not regularly discuss discharge criteria during bedside or CCRs. The phrase discharge criteria has now become part of our shared language.

Informally, the DRR appears to have reduced inefficiency and the potential for communication error. The practice of writing notes on printed patient lists to be used to sign‐out or communicate to other team members not in attendance at CCRs has largely disappeared.

The DRR has proven to be adaptable across patient units, and can be tailored to the specific transitional care needs of a given patient population. At discharge institution, the DRR has been modified for, and has taken on a prominent role in, the discharge planning of highly complex populations such as rehabilitation and ventilated patients.

DISCUSSION

Discharge planning is a multifaceted, multidisciplinary process that should begin at the time of hospital admission. Safe patient transition depends on efficient discharge processes and effective communication across settings.[8] Although not well studied in the inpatient setting, care process variability can result in inefficient patient flow and increased stress among staff.[9] Patients and families may experience confusion, coping difficulties, and increased readmission due to ineffective discharge planning.[10] These potential pitfalls highlight the need for healthcare providers to develop patient‐centered, systematic approaches to improving the discharge process.[11]

To our knowledge, this is the first description of a discharge planning tool for the EHR in the pediatric setting. Our discharge report is centralized, easily accessible by all members of the care team, and includes important patient‐specific discharge‐related information that be used to focus discussion and streamline multidisciplinary discharge planning rounds.

We anticipate that the report will allow the entire healthcare team to function more efficiently, decrease discharge‐related delays and failures based on communication roadblocks, and improve family and caregiver satisfaction with the discharge process. We are currently testing these hypotheses and evaluating several implementation strategies in an ongoing research study. Assuming positive impact, we plan to spread the use of the DRR to all inpatient care areas at our hospital, and potentially to other hospitals.

The limitations of this QI project are consistent with other initiatives to improve care. The challenges we encounter at our freestanding tertiary care teaching hospital with regard to effective discharge planning and multidisciplinary communication may not be generalizable to other nonteaching or community hospitals, and the DRR may not be useful in other settings. Though the report is now a part of our EHR, the most impactful implementation strategies remain to be determined. The report and related changes represent significant diversion from years of deeply ingrained workflows for some providers, and we encountered some resistance from staff during the early stages of implementation. The most important of which was that some team members are uncomfortable with technology and prefer to use paper. Most of this initial resistance was overcome by implementing changes to improve the ease of use of the report (Figure 3). Though input from end users and key stakeholders has been incorporated throughout this initiative, more work is needed to measure end user adoption and satisfaction with the report.

CONCLUSION

High‐quality hospital discharge planning requires an increasingly multidisciplinary approach. The EHR can be leveraged to improve transparency and interdisciplinary communication around the discharge process. An integrated summary of discharge‐related issues, organized into 1 highly visible and easily accessible report in the EHR has the potential to improve care transitions.

According to the American Academy of Pediatrics clinical report on physicians' roles in coordinating care of hospitalized children, there are several important components of hospital discharge planning.[1] Foremost is that discharge planning should begin, and discharge criteria should be set, at the time of hospital admission. This allows for optimal engagement of parents and providers in the effort to adequately prepare patients for the transition to home.

As pediatric inpatients become increasingly complex,[2] adequately preparing families for the transition to home becomes more challenging.[3] There are a myriad of issues to address and the burden of this preparation effort falls on multiple individuals other than the bedside nurse and physician. Large multidisciplinary teams often play a significant role in the discharge of medically complex children.[4] Several challenges may hinder the team's ability to effectively navigate the discharge process such as financial or insurance‐related issues, language differences, or geographic barriers. Patient and family anxieties may also complicate the transition to home.[5]

The challenges of a multidisciplinary approach to discharge planning are further magnified by the limitations of the electronic health record (EHR). The EHR is well designed to record individual encounters, but poorly designed to coordinate longitudinal care across settings.[6] Although multidisciplinary providers may spend significant and well‐intentioned energy to facilitate hospital discharge, their efforts may go unseen or be duplicative.

We developed a discharge readiness report (DRR) for the EHR, an integrated summary of discharge‐related issues, organized into a highly visible and easily accessible report. The development of the discharge planning tool was the first step in a larger quality improvement (QI) initiative aimed at improving the efficiency, effectiveness, and safety of hospital discharge. Our team recognized that improving the flow and visibility of information between disciplines was the first step toward accomplishing this larger aim. Health information technology offers an important opportunity for the improvement of patient safety and care transitions⁷; therefore, we leveraged the EHR to create an integrated discharge report. We used QI methods to understand our hospital's discharge processes, examined potential pitfalls in interdisciplinary communication, determined relevant information to include in the report, and optimized ways to display the data. To our knowledge, this use of the EHR is novel. The objectives of this article were to describe our team's development and implementation strategies, as well as challenges encountered, in the design of this electronic discharge planning tool.

METHODS

Planning the Intervention

Based on our learning, we developed a key driver diagram (Figure 1). Our aim was to create a DRR that organized important discharge‐related information into 1 easily accessible report. Key drivers that were identified as relevant to the content of the DRR included: barriers to discharge, discharge criteria, home care, postdischarge care, and last minute delays. We also identified secondary drivers related to the design of the DRR. We hypothesized that addressing the secondary drivers would be essential to end user adoption of the tool. The secondary drivers included: accessibility, relevance, ease of updating, automation, and readability.

Figure 1

With the swim lane diagram as well as our primary and secondary drivers in mind, we created a mock DRR on paper. We conducted multiple patient discharge simulations with representatives from all disciplines, walking through each step of a patient hospitalization from registration to discharge. This allowed us to map out how preexisting, yet disparate, EHR data could be channeled into 1 report. A few changes were made to processes involving data collection and documentation to facilitate timely transfer of information to the report. For example, questions addressing potential barriers to discharge and whether a school/work excuse was needed were added to the admission nursing assessment.

We then moved the paper DRR to the electronic environment. Data elements that were pulled automatically into the report included: potential barriers to discharge collected during nursing intake, case management information on home care needs, discharge criteria entered by resident and attending physicians, PCP, home pharmacy, follow‐up appointments, school/work excuse information gathered by resident liaisons, and active patient problems drawn from the problem list section. These data were organized into 4 distinct domains within the final DRR: potential barriers, transitional care, home care, and discharge criteria (Table 1).

Table 1.Discharge Readiness Report Domains

Discharge Readiness Report Domain	Example Content
NOTE: Abbreviations: PCP, primary care provider.
Potential barriers to discharge	Geographic location of the family, whether patient lives in more than 1 household, primary spoken language, financial or insurance concern, and need for work/subhool excuses
Transitional care	PCP and home pharmacy information, follow‐up ambulatory and imaging appointments, and care team communications with the PCP
Home care	Planned discharge date/time and home care needs assessments such as needs for special equipment or skilled home nursing
Discharge criteria	Clinical, social, or other care coordination conditions for discharge

Additional features potentially important to discharge planning were also incorporated into the report based on end user feedback. These included hyperlinks to discharge orders, home oxygen prescriptions, and the after‐visit summary for families, and the patient's home care company (if present). To facilitate discharge and transitional care related communication between the primary team and subspecialty teams, consults involved during the hospitalization were included on the report. As home care arrangements often involve care for active lines and drains, they were added to the report (Figure 2).

Figure 2

Implementation

The report was activated within the EHR in June 2012. The team focused initial promotion and education efforts on medical floors. Education was widely disseminated via email and in‐person presentations.

The DRR was incorporated into daily CCRs for medical patients in July 2012. These multidisciplinary rounds occurred after medical‐team bedside rounds, focusing on care coordination and discharge planning. For each patient discussed, the DRR was projected onto a large screen, allowing all team members to view and discuss relevant discharge information. A process improvement (PI) specialist attended CCRs daily for several months, educating participants and monitoring use of the DRR. The PI specialist solicited feedback on ways to improve the DRR, and timed rounds to measure whether use of the DRR prolonged CCRs.

In the first weeks postimplementation, the use of the DRR prolonged rounds by as much as 1 minute per patient. Based on direct observation, the team focused interventions on barriers to the efficient use of the report during CCRs including: the need to scroll through the report, which was not visible on 1 screen; the need to navigate between patients; the need to quickly update the report based on discussion; and the need to update discharge criteria (Figure 3).

Figure 3

RESULTS

Creation of the final DRR required significant time and effort and was the culmination of a uniquely collaborative effort between clinicians, ancillary staff, and information technology specialists (Figure 4). The report is used consistently for all general medical and medical subspecialty patients during CCRs. After interventions were implemented to improve the efficiency of using the DRR during CCRs, the use of the DRR did not prolong CCRs. Members of the care team acknowledge that all sections of the report are populated and accurate. Though end users have commented on their use of the report outside of CCRs, we have not been able to formally measure this.

Figure 4

We have noticed a shift in the focus of discussion since implementation of the DRR. Prior to this initiative, care teams at our institution did not regularly discuss discharge criteria during bedside or CCRs. The phrase discharge criteria has now become part of our shared language.

Informally, the DRR appears to have reduced inefficiency and the potential for communication error. The practice of writing notes on printed patient lists to be used to sign‐out or communicate to other team members not in attendance at CCRs has largely disappeared.

The DRR has proven to be adaptable across patient units, and can be tailored to the specific transitional care needs of a given patient population. At discharge institution, the DRR has been modified for, and has taken on a prominent role in, the discharge planning of highly complex populations such as rehabilitation and ventilated patients.

DISCUSSION

Discharge planning is a multifaceted, multidisciplinary process that should begin at the time of hospital admission. Safe patient transition depends on efficient discharge processes and effective communication across settings.[8] Although not well studied in the inpatient setting, care process variability can result in inefficient patient flow and increased stress among staff.[9] Patients and families may experience confusion, coping difficulties, and increased readmission due to ineffective discharge planning.[10] These potential pitfalls highlight the need for healthcare providers to develop patient‐centered, systematic approaches to improving the discharge process.[11]

To our knowledge, this is the first description of a discharge planning tool for the EHR in the pediatric setting. Our discharge report is centralized, easily accessible by all members of the care team, and includes important patient‐specific discharge‐related information that be used to focus discussion and streamline multidisciplinary discharge planning rounds.

We anticipate that the report will allow the entire healthcare team to function more efficiently, decrease discharge‐related delays and failures based on communication roadblocks, and improve family and caregiver satisfaction with the discharge process. We are currently testing these hypotheses and evaluating several implementation strategies in an ongoing research study. Assuming positive impact, we plan to spread the use of the DRR to all inpatient care areas at our hospital, and potentially to other hospitals.

The limitations of this QI project are consistent with other initiatives to improve care. The challenges we encounter at our freestanding tertiary care teaching hospital with regard to effective discharge planning and multidisciplinary communication may not be generalizable to other nonteaching or community hospitals, and the DRR may not be useful in other settings. Though the report is now a part of our EHR, the most impactful implementation strategies remain to be determined. The report and related changes represent significant diversion from years of deeply ingrained workflows for some providers, and we encountered some resistance from staff during the early stages of implementation. The most important of which was that some team members are uncomfortable with technology and prefer to use paper. Most of this initial resistance was overcome by implementing changes to improve the ease of use of the report (Figure 3). Though input from end users and key stakeholders has been incorporated throughout this initiative, more work is needed to measure end user adoption and satisfaction with the report.

CONCLUSION

High‐quality hospital discharge planning requires an increasingly multidisciplinary approach. The EHR can be leveraged to improve transparency and interdisciplinary communication around the discharge process. An integrated summary of discharge‐related issues, organized into 1 highly visible and easily accessible report in the EHR has the potential to improve care transitions.

Patient flow refers to the management and movement of patients in a healthcare facility. Healthcare institutions utilize patient flow analyses to evaluate and improve aspects of the patient experience including safety, effectiveness, efficiency, timeliness, patient centeredness, and equity.[1, 2, 3, 4, 5, 6, 7, 8] Hospitals can evaluate patient flow using specific metrics, such as time in emergency department (ED) or percent of discharges completed by a certain time of day. However, no single metric can represent the full spectrum of processes inherent to patient flow. For example, ED length of stay (LOS) is dependent on inpatient occupancy, which is dependent on discharge timeliness. Each of these activities depends on various smaller activities, such as cleaning rooms or identifying available beds.

Evaluating the quality that healthcare organizations deliver is growing in importance.[9] Composite scores are being used increasingly to assess clinical processes and outcomes for professionals and institutions.[10, 11] Where various aspects of performance coexist, composite measures can incorporate multiple metrics into a comprehensive summary.[12, 13, 14, 15, 16] They also allow organizations to track a range of metrics for more holistic, comprehensive evaluations.[9, 13]

This article describes a balanced scorecard with composite scoring used at a large urban children's hospital to evaluate patient flow and direct improvement resources where they are needed most.

METHODS

The Children's Hospital of Philadelphia identified patient flow improvement as an operating plan initiative. Previously, performance was measured with a series of independent measures including time from ED arrival to transfer to the inpatient floor, and time from discharge order to room vacancy. These metrics were dismissed as sole measures of flow because they did not reflect the complexity and interdependence of processes or improvement efforts. There were also concerns that efforts to improve a measure caused unintended consequences for others, which at best lead to little overall improvement, and at worst reduced performance elsewhere in the value chain. For example, to meet a goal time for entering discharge orders, physicians could enter orders earlier. But, if patients were not actually ready to leave, their beds were not made available any earlier. Similarly, bed management staff could rush to meet a goal for speed of unit assignment, but this could cause an increase in patients admitted to the wrong specialty floor.

To address these concerns, a group of physicians, nurses, quality improvement specialists, and researchers designed a patient flow scorecard with composite measurement. Five domains of patient flow were identified: (1) ED and ED‐to‐inpatient transition, (2) bed management, (3) discharge process, (4) room turnover and environmental services department (ESD) activities, and (5) scheduling and utilization. Component measures for each domain were selected for 1 of 3 purposes: (1) to correspond to processes of importance to flow and improvement work, (2) to act as adjusters for factors that affect performance, or (3) to act as balancing measures so that progress in a measure would not result in the degradation of another. Each domain was assigned 20 points, which were distributed across the domain's components based on a consensus of the component's relative importance to overall domain performance (Figure 1). Data from the previous year were used as guidelines for setting performance percentile goals. For example, a goal of 80% in 60 minutes for arrival to physician evaluation meant that 80% of patients should see a physician within 1 hour of arriving at the ED.

Figure 1

Scores were also categorized to correspond to commonly used color descriptors.[17] For each component measure, performance meeting or exceeding the goal fell into the green category. Performances <10 percentage points below the goal fell into the yellow category, and performances below that level fell into the red category. Domain‐level scores and overall composite scores were also assigned colors. Performance at or above 80% (16 on the 20‐point domain scale, or 80 on the 100‐point overall scale) were designated green, scores between 70% and 79% were yellow, and scores below 70% were red.

DOMAINS OF THE PATIENT FLOW COMPOSITE SCORE

RESULTS

The balanced scorecard with composite measures provided improvement teams and administrators with a picture of patient flow (Figure 2). The overall score provided a global perspective on patient flow over time and captured trends in performance during various states of hospital occupancy. One trend that it captured was an association between high volume and poor composite scores (Figure 3). Notably, the H1N1 influenza pandemic in the fall of 2009 and the turnover of computer systems in January 2011 can be linked to dips in performance. The changes between fiscal years reflect a shift in baseline metrics.

Figure 2

Figure 3

In addition to the overall composite score, the domain level and individual component scores allowed for more specific evaluation of variables affecting quality of care and enabled targeted improvement activities (Figure 4). For example, in December 2010 and January 2011, room turnover and ESD domain scores dropped, especially in the total discharge to clean time component. In response, the ESD made staffing adjustments, and starting in February 2011, component scores and the domain score improved. Feedback from the scheduling and utilization domain scores also initiated positive change. In August 2010, the CV: scheduled occupancy component score started to drop. In response, certain elective admissions were shifted to weekends to distribute hospital occupancy more evenly throughout the week. By February 2011, the component returned to its goal level. This continual evaluation of performance motivates continual improvement.

Figure 4

DISCUSSION

The use of a patient flow balanced scorecard with composite measurement overcomes pitfalls associated with a single or unaggregated measure. Aggregate scores alone mask important differences and relationships among components.[13] For example, 2 domains may be inversely related, or a provider with an overall average score might score above average in 1 domain but below in another. The composite scorecard, however, shows individual component and domain scores in addition to an aggregate score. The individual component and domain level scores highlight specific areas that need improvement and allow attention to be directed to those areas.

Additionally, a composite score is more likely to engage the range of staff involved in patient flow. Scaling out of 100 points and the red‐yellow‐green model are familiar for operations performance and can be easily understood.[17] Moreover, a composite score allows for dynamic performance goals while maintaining a stable measurement structure. For example, standardized LOS ratios, readmission rates, and denied hospital days can be added to the scorecard to provide more information and balancing measures.

Although balanced scorecards with composites can make holistic performance visible across multiple operational domains, they have some disadvantages. First, because there is a degree of complexity associated with a measure that incorporates multiple aspects of flow, certain elements, such as the relationship between a metric and its balancing measure, may not be readily apparent. Second, composite measures may not provide actionable information if the measure is not clearly related to a process that can be improved.[13, 14] Third, individual metrics may not be replicable between locations, so composites may need to be individualized to each setting.[10, 20]

Improving patient flow is a goal at many hospitals. Although measurement is crucial to identifying and mitigating variations, measuring the multidimensional aspects of flow and their impact on quality is difficult. Our scorecard, with composite measurement, addresses the need for an improved method to assess patient flow and improve quality by tracking care processes simultaneously.

Patient flow refers to the management and movement of patients in a healthcare facility. Healthcare institutions utilize patient flow analyses to evaluate and improve aspects of the patient experience including safety, effectiveness, efficiency, timeliness, patient centeredness, and equity.[1, 2, 3, 4, 5, 6, 7, 8] Hospitals can evaluate patient flow using specific metrics, such as time in emergency department (ED) or percent of discharges completed by a certain time of day. However, no single metric can represent the full spectrum of processes inherent to patient flow. For example, ED length of stay (LOS) is dependent on inpatient occupancy, which is dependent on discharge timeliness. Each of these activities depends on various smaller activities, such as cleaning rooms or identifying available beds.

Evaluating the quality that healthcare organizations deliver is growing in importance.[9] Composite scores are being used increasingly to assess clinical processes and outcomes for professionals and institutions.[10, 11] Where various aspects of performance coexist, composite measures can incorporate multiple metrics into a comprehensive summary.[12, 13, 14, 15, 16] They also allow organizations to track a range of metrics for more holistic, comprehensive evaluations.[9, 13]

This article describes a balanced scorecard with composite scoring used at a large urban children's hospital to evaluate patient flow and direct improvement resources where they are needed most.

METHODS

The Children's Hospital of Philadelphia identified patient flow improvement as an operating plan initiative. Previously, performance was measured with a series of independent measures including time from ED arrival to transfer to the inpatient floor, and time from discharge order to room vacancy. These metrics were dismissed as sole measures of flow because they did not reflect the complexity and interdependence of processes or improvement efforts. There were also concerns that efforts to improve a measure caused unintended consequences for others, which at best lead to little overall improvement, and at worst reduced performance elsewhere in the value chain. For example, to meet a goal time for entering discharge orders, physicians could enter orders earlier. But, if patients were not actually ready to leave, their beds were not made available any earlier. Similarly, bed management staff could rush to meet a goal for speed of unit assignment, but this could cause an increase in patients admitted to the wrong specialty floor.

To address these concerns, a group of physicians, nurses, quality improvement specialists, and researchers designed a patient flow scorecard with composite measurement. Five domains of patient flow were identified: (1) ED and ED‐to‐inpatient transition, (2) bed management, (3) discharge process, (4) room turnover and environmental services department (ESD) activities, and (5) scheduling and utilization. Component measures for each domain were selected for 1 of 3 purposes: (1) to correspond to processes of importance to flow and improvement work, (2) to act as adjusters for factors that affect performance, or (3) to act as balancing measures so that progress in a measure would not result in the degradation of another. Each domain was assigned 20 points, which were distributed across the domain's components based on a consensus of the component's relative importance to overall domain performance (Figure 1). Data from the previous year were used as guidelines for setting performance percentile goals. For example, a goal of 80% in 60 minutes for arrival to physician evaluation meant that 80% of patients should see a physician within 1 hour of arriving at the ED.

Figure 1

Scores were also categorized to correspond to commonly used color descriptors.[17] For each component measure, performance meeting or exceeding the goal fell into the green category. Performances <10 percentage points below the goal fell into the yellow category, and performances below that level fell into the red category. Domain‐level scores and overall composite scores were also assigned colors. Performance at or above 80% (16 on the 20‐point domain scale, or 80 on the 100‐point overall scale) were designated green, scores between 70% and 79% were yellow, and scores below 70% were red.

DOMAINS OF THE PATIENT FLOW COMPOSITE SCORE

RESULTS

The balanced scorecard with composite measures provided improvement teams and administrators with a picture of patient flow (Figure 2). The overall score provided a global perspective on patient flow over time and captured trends in performance during various states of hospital occupancy. One trend that it captured was an association between high volume and poor composite scores (Figure 3). Notably, the H1N1 influenza pandemic in the fall of 2009 and the turnover of computer systems in January 2011 can be linked to dips in performance. The changes between fiscal years reflect a shift in baseline metrics.

Figure 2

Figure 3

In addition to the overall composite score, the domain level and individual component scores allowed for more specific evaluation of variables affecting quality of care and enabled targeted improvement activities (Figure 4). For example, in December 2010 and January 2011, room turnover and ESD domain scores dropped, especially in the total discharge to clean time component. In response, the ESD made staffing adjustments, and starting in February 2011, component scores and the domain score improved. Feedback from the scheduling and utilization domain scores also initiated positive change. In August 2010, the CV: scheduled occupancy component score started to drop. In response, certain elective admissions were shifted to weekends to distribute hospital occupancy more evenly throughout the week. By February 2011, the component returned to its goal level. This continual evaluation of performance motivates continual improvement.

Figure 4

DISCUSSION

The use of a patient flow balanced scorecard with composite measurement overcomes pitfalls associated with a single or unaggregated measure. Aggregate scores alone mask important differences and relationships among components.[13] For example, 2 domains may be inversely related, or a provider with an overall average score might score above average in 1 domain but below in another. The composite scorecard, however, shows individual component and domain scores in addition to an aggregate score. The individual component and domain level scores highlight specific areas that need improvement and allow attention to be directed to those areas.

Additionally, a composite score is more likely to engage the range of staff involved in patient flow. Scaling out of 100 points and the red‐yellow‐green model are familiar for operations performance and can be easily understood.[17] Moreover, a composite score allows for dynamic performance goals while maintaining a stable measurement structure. For example, standardized LOS ratios, readmission rates, and denied hospital days can be added to the scorecard to provide more information and balancing measures.

Although balanced scorecards with composites can make holistic performance visible across multiple operational domains, they have some disadvantages. First, because there is a degree of complexity associated with a measure that incorporates multiple aspects of flow, certain elements, such as the relationship between a metric and its balancing measure, may not be readily apparent. Second, composite measures may not provide actionable information if the measure is not clearly related to a process that can be improved.[13, 14] Third, individual metrics may not be replicable between locations, so composites may need to be individualized to each setting.[10, 20]

Improving patient flow is a goal at many hospitals. Although measurement is crucial to identifying and mitigating variations, measuring the multidimensional aspects of flow and their impact on quality is difficult. Our scorecard, with composite measurement, addresses the need for an improved method to assess patient flow and improve quality by tracking care processes simultaneously.

Patient flow refers to the management and movement of patients in a healthcare facility. Healthcare institutions utilize patient flow analyses to evaluate and improve aspects of the patient experience including safety, effectiveness, efficiency, timeliness, patient centeredness, and equity.[1, 2, 3, 4, 5, 6, 7, 8] Hospitals can evaluate patient flow using specific metrics, such as time in emergency department (ED) or percent of discharges completed by a certain time of day. However, no single metric can represent the full spectrum of processes inherent to patient flow. For example, ED length of stay (LOS) is dependent on inpatient occupancy, which is dependent on discharge timeliness. Each of these activities depends on various smaller activities, such as cleaning rooms or identifying available beds.

Evaluating the quality that healthcare organizations deliver is growing in importance.[9] Composite scores are being used increasingly to assess clinical processes and outcomes for professionals and institutions.[10, 11] Where various aspects of performance coexist, composite measures can incorporate multiple metrics into a comprehensive summary.[12, 13, 14, 15, 16] They also allow organizations to track a range of metrics for more holistic, comprehensive evaluations.[9, 13]

This article describes a balanced scorecard with composite scoring used at a large urban children's hospital to evaluate patient flow and direct improvement resources where they are needed most.

METHODS

The Children's Hospital of Philadelphia identified patient flow improvement as an operating plan initiative. Previously, performance was measured with a series of independent measures including time from ED arrival to transfer to the inpatient floor, and time from discharge order to room vacancy. These metrics were dismissed as sole measures of flow because they did not reflect the complexity and interdependence of processes or improvement efforts. There were also concerns that efforts to improve a measure caused unintended consequences for others, which at best lead to little overall improvement, and at worst reduced performance elsewhere in the value chain. For example, to meet a goal time for entering discharge orders, physicians could enter orders earlier. But, if patients were not actually ready to leave, their beds were not made available any earlier. Similarly, bed management staff could rush to meet a goal for speed of unit assignment, but this could cause an increase in patients admitted to the wrong specialty floor.

To address these concerns, a group of physicians, nurses, quality improvement specialists, and researchers designed a patient flow scorecard with composite measurement. Five domains of patient flow were identified: (1) ED and ED‐to‐inpatient transition, (2) bed management, (3) discharge process, (4) room turnover and environmental services department (ESD) activities, and (5) scheduling and utilization. Component measures for each domain were selected for 1 of 3 purposes: (1) to correspond to processes of importance to flow and improvement work, (2) to act as adjusters for factors that affect performance, or (3) to act as balancing measures so that progress in a measure would not result in the degradation of another. Each domain was assigned 20 points, which were distributed across the domain's components based on a consensus of the component's relative importance to overall domain performance (Figure 1). Data from the previous year were used as guidelines for setting performance percentile goals. For example, a goal of 80% in 60 minutes for arrival to physician evaluation meant that 80% of patients should see a physician within 1 hour of arriving at the ED.

Figure 1

Scores were also categorized to correspond to commonly used color descriptors.[17] For each component measure, performance meeting or exceeding the goal fell into the green category. Performances <10 percentage points below the goal fell into the yellow category, and performances below that level fell into the red category. Domain‐level scores and overall composite scores were also assigned colors. Performance at or above 80% (16 on the 20‐point domain scale, or 80 on the 100‐point overall scale) were designated green, scores between 70% and 79% were yellow, and scores below 70% were red.

DOMAINS OF THE PATIENT FLOW COMPOSITE SCORE

RESULTS

The balanced scorecard with composite measures provided improvement teams and administrators with a picture of patient flow (Figure 2). The overall score provided a global perspective on patient flow over time and captured trends in performance during various states of hospital occupancy. One trend that it captured was an association between high volume and poor composite scores (Figure 3). Notably, the H1N1 influenza pandemic in the fall of 2009 and the turnover of computer systems in January 2011 can be linked to dips in performance. The changes between fiscal years reflect a shift in baseline metrics.

Figure 2

Figure 3

In addition to the overall composite score, the domain level and individual component scores allowed for more specific evaluation of variables affecting quality of care and enabled targeted improvement activities (Figure 4). For example, in December 2010 and January 2011, room turnover and ESD domain scores dropped, especially in the total discharge to clean time component. In response, the ESD made staffing adjustments, and starting in February 2011, component scores and the domain score improved. Feedback from the scheduling and utilization domain scores also initiated positive change. In August 2010, the CV: scheduled occupancy component score started to drop. In response, certain elective admissions were shifted to weekends to distribute hospital occupancy more evenly throughout the week. By February 2011, the component returned to its goal level. This continual evaluation of performance motivates continual improvement.

Figure 4

DISCUSSION

The use of a patient flow balanced scorecard with composite measurement overcomes pitfalls associated with a single or unaggregated measure. Aggregate scores alone mask important differences and relationships among components.[13] For example, 2 domains may be inversely related, or a provider with an overall average score might score above average in 1 domain but below in another. The composite scorecard, however, shows individual component and domain scores in addition to an aggregate score. The individual component and domain level scores highlight specific areas that need improvement and allow attention to be directed to those areas.

Additionally, a composite score is more likely to engage the range of staff involved in patient flow. Scaling out of 100 points and the red‐yellow‐green model are familiar for operations performance and can be easily understood.[17] Moreover, a composite score allows for dynamic performance goals while maintaining a stable measurement structure. For example, standardized LOS ratios, readmission rates, and denied hospital days can be added to the scorecard to provide more information and balancing measures.

Although balanced scorecards with composites can make holistic performance visible across multiple operational domains, they have some disadvantages. First, because there is a degree of complexity associated with a measure that incorporates multiple aspects of flow, certain elements, such as the relationship between a metric and its balancing measure, may not be readily apparent. Second, composite measures may not provide actionable information if the measure is not clearly related to a process that can be improved.[13, 14] Third, individual metrics may not be replicable between locations, so composites may need to be individualized to each setting.[10, 20]

Improving patient flow is a goal at many hospitals. Although measurement is crucial to identifying and mitigating variations, measuring the multidimensional aspects of flow and their impact on quality is difficult. Our scorecard, with composite measurement, addresses the need for an improved method to assess patient flow and improve quality by tracking care processes simultaneously.