The word housestaff came from an era when you moved into the hospital (house) for your internship (staff). In those days, call every other night meant you missed half the learning opportunities. In those days, supervision was as remote as it could get. My, have things changed.

In 2003, the Accreditation Council for Graduate Medical Education (ACGME) implemented duty hour standards that called for restricting resident duty hours to no more than 80 hours per week, and no more than 24 continuous hours with an additional 6 hours for transfer of care.1 The expectation was that limiting work hours would have positive effects on outcomes, such as resident fatigue, the educational experience, and patient safety. A recent systematic review commissioned by the ACGME reviewed three decades worth of research on duty hours and patient safety. It reviewed laboratory studies of resident performance on clinical simulations, and tests of cognitive and fine motor skills under conditions of sleep deprivation. The results of these studies imply that limiting residents' duty hours should have positive effects on patient safety. However, that assumption has not been borne out. Paradoxically, examination of studies of alternative duty hour schedules, as well as the follow‐up studies on the impact of the New York duty hour rules and ACGME duty hour regulations have documented few measurable effects on patient safety.2 Thus, the simplistic view that fewer hours worked equals better patient care deserves a second‐look. In fact, a cause‐and‐effect relationship between resident fatigue and medical errors has yet to be proven.

The hospitalist movement was in full swing by 2003 and had already transformed the care of hospitalized patients. The early studies of hospitalists focused on cost and comparing outcomes with those of other providers. As many studies showed the economic benefit of hospitalists, residency programs began redesigning their inpatient services with hospitalists at the helm. Institutions sponsoring graduate medical education programs saw hospitalists as a necessary resource to enable and maintain compliance with duty hour reform. Some institutions created hospitalist‐only services with no housestaff (nonteaching services). Others created hospitalist teaching services that used indirect supervision at night from a hospitalist on‐call from home, while others had on‐site hospitalists at night that provided direct supervision of housestaff. It is clear that hospitalists are a versatile and valuable resource.

In this issue of the Journal of Hospital Medicine, Khanna and colleagues3 describe a retrospective medical record review at a large, urban academic hospital that evaluates the potential association between night admissions by night‐shiftbased hospitalists and a resident night float system, and hospitalization‐relevant outcomes. The outcomes included length of stay, hospital charges, Intensive Care Unit (ICU) transfers, Emergency Department visits within 30 days of discharge, 30‐day readmission rates, and poor outcomes within the first 24 hours of admission. Their results showed that night or weekend admission was not associated with worse hospitalization‐relevant outcomes. The extension of on‐site faculty supervision to nights, as well as awards for nursing excellence, likely had a strong positive impact on these outcomes. In some cases, night or weekend admission was associated with better outcomes, particularly in terms of ICU transfer during hospitalization and hospital charges. Maybe off‐hours care is not so off after all.

I am not sure where we are going, but the duty hour standards have been revised again. The new duty hour standards limit continuous duty hours for 1st‐year residents to 16, and all others to 24. Upper‐level residents can stay an additional 4 hours for transfer of care.4 Functionally, this proposed standard eliminates overnight call for 1st‐year residents. Although I would never advocate returning to the days of yore where autonomy and independence were the rule, I mourn the loss of extended shifts for 1st‐year residents. The loss of the extended shift will come at the expense of overnight call in many programs. I worry about the loss of some unique teaching moments for all residents, and the increased number of transitions in a patients' care. The hospital is a different place at night. It is less chaotic with fewer visitors, and there are less interruptions in patient care. There is time for teaching and learning because the work intensity is lower. The bedside teaching among senior and junior residents, during an acute patient episode in the middle of the night, is an experience that many of us would recount as an invaluable part of our training. It is also a time when on‐site faculty supervision and nurse staffing ratios are typically reduced. In other words, the entire system of care is different at night. This dichotomy in day versus night medicine has spurred research into whether patient outcomes differed between the two. Various investigators have looked at the differences in care, and there is some data to support that patient care outcomes are worse at night than during the day.58

The results from Khanna and colleagues3 are welcome news, since graduate medical education is undergoing a paradigm shift with further reductions in duty hours. Hospital‐based medicine is a 24‐hour operation, 7 days a week. We don't get to pick when patients are admitted or need care. In fact, this study demonstrates that 58% of the admissions came after 7 _PM. It turns out that this pattern of admissions is not atypical.9 In light of the new standards, programs are left deciding where to deploy their 1st‐year resident resources. There are essentially two types of models to provide patient care off‐hours in the absence of an extended shift option: night float and night medicine. Night float implies that a trainee, usually a resident, comes from a different site or rotation to provide nocturnal care of a group of patients. That resident may do one isolated night of coverage or several nights in a row. The resident may perform cross‐cover functions, admit patients, or do both. Night medicine refers to having teams of housestaff at night along with an educational program as part of a rotation. The latter implies that trainees will get the full educational experience whether working days or nights. If one designs a model of care that puts 1st‐year residents where the volume is, they would come in late and leave early in the morning the next day. Both create more handoffs in patient care, which critics of these models have said lead to worse patient outcomes. This research challenges such earlier work and should make us all feel more comfortable that care at night does not need to be different if the proper systems are in place to manage it.

The graduate medical education community needs to take a step back and ask: Where are we going? As a long‐standing program director, I don't know anymore. As training has become competency‐based, and milestones are being developed for each level of training, we need to ask if reducing hours further is in the best interests of graduate medical education and, ultimately, the general public. I have long been interested in work intensity as a contributor to resident fatigue, burnout, and errors. A systematic review of resident workload and work activity looked at 21 studies examining resident time allocation and suggests that at least one‐third of resident time is spent on activities of limited educational value. Eight of the nine studies evaluating the impact of resident workload on patient care showed a negative impact on outcomes, including length of stay, mortality, patient satisfaction, medication errors, and lab utilization.10 If we are going to further restrict duty hours, then we need to transform the learning environment in a similar fashion. It does not seem logical to restrict duty hours further without a major redesign of the learning environment. I fear that all we have done is create more work compression because the system of care will be further strained in its weakest areas.

Hospitalists are uniquely poised to mitigate some of these concerns by being available for direct and indirect supervision, as well as being highly skilled at working in our complex systems. They are experts in systems‐based practice. The extension of direct faculty supervision to nights may be a major factor in eliminating the variances in care between nights and days. Not every program or institution will have the ability to employ night hospitalists, let alone have them in a supervisory and teaching role. Those that do can take great comfort in this research. Those that don't must challenge themselves to think about how to get there.

We all share the collective goal of producing the next generation of well‐trained physicians who will practice safe and effective care. As we begin to embrace competency‐based training, the prism through which we view patient safety must extend beyond duty hours and include learning environment reform. While the Khanna et al.3 study supports that off‐hours care is no longer off, we need to be careful that we don't drift too far towards shift work as the fundamental structure of medical training. In the beginning, residents lived in the hospital, thus the term. As we move through these stormy seas, we don't want to trade patient safety for transient trainees. There is a middle ground here that will be informed by future research. There is something to be said for living in your house.

The word housestaff came from an era when you moved into the hospital (house) for your internship (staff). In those days, call every other night meant you missed half the learning opportunities. In those days, supervision was as remote as it could get. My, have things changed.

In 2003, the Accreditation Council for Graduate Medical Education (ACGME) implemented duty hour standards that called for restricting resident duty hours to no more than 80 hours per week, and no more than 24 continuous hours with an additional 6 hours for transfer of care.1 The expectation was that limiting work hours would have positive effects on outcomes, such as resident fatigue, the educational experience, and patient safety. A recent systematic review commissioned by the ACGME reviewed three decades worth of research on duty hours and patient safety. It reviewed laboratory studies of resident performance on clinical simulations, and tests of cognitive and fine motor skills under conditions of sleep deprivation. The results of these studies imply that limiting residents' duty hours should have positive effects on patient safety. However, that assumption has not been borne out. Paradoxically, examination of studies of alternative duty hour schedules, as well as the follow‐up studies on the impact of the New York duty hour rules and ACGME duty hour regulations have documented few measurable effects on patient safety.2 Thus, the simplistic view that fewer hours worked equals better patient care deserves a second‐look. In fact, a cause‐and‐effect relationship between resident fatigue and medical errors has yet to be proven.

The hospitalist movement was in full swing by 2003 and had already transformed the care of hospitalized patients. The early studies of hospitalists focused on cost and comparing outcomes with those of other providers. As many studies showed the economic benefit of hospitalists, residency programs began redesigning their inpatient services with hospitalists at the helm. Institutions sponsoring graduate medical education programs saw hospitalists as a necessary resource to enable and maintain compliance with duty hour reform. Some institutions created hospitalist‐only services with no housestaff (nonteaching services). Others created hospitalist teaching services that used indirect supervision at night from a hospitalist on‐call from home, while others had on‐site hospitalists at night that provided direct supervision of housestaff. It is clear that hospitalists are a versatile and valuable resource.

In this issue of the Journal of Hospital Medicine, Khanna and colleagues3 describe a retrospective medical record review at a large, urban academic hospital that evaluates the potential association between night admissions by night‐shiftbased hospitalists and a resident night float system, and hospitalization‐relevant outcomes. The outcomes included length of stay, hospital charges, Intensive Care Unit (ICU) transfers, Emergency Department visits within 30 days of discharge, 30‐day readmission rates, and poor outcomes within the first 24 hours of admission. Their results showed that night or weekend admission was not associated with worse hospitalization‐relevant outcomes. The extension of on‐site faculty supervision to nights, as well as awards for nursing excellence, likely had a strong positive impact on these outcomes. In some cases, night or weekend admission was associated with better outcomes, particularly in terms of ICU transfer during hospitalization and hospital charges. Maybe off‐hours care is not so off after all.

I am not sure where we are going, but the duty hour standards have been revised again. The new duty hour standards limit continuous duty hours for 1st‐year residents to 16, and all others to 24. Upper‐level residents can stay an additional 4 hours for transfer of care.4 Functionally, this proposed standard eliminates overnight call for 1st‐year residents. Although I would never advocate returning to the days of yore where autonomy and independence were the rule, I mourn the loss of extended shifts for 1st‐year residents. The loss of the extended shift will come at the expense of overnight call in many programs. I worry about the loss of some unique teaching moments for all residents, and the increased number of transitions in a patients' care. The hospital is a different place at night. It is less chaotic with fewer visitors, and there are less interruptions in patient care. There is time for teaching and learning because the work intensity is lower. The bedside teaching among senior and junior residents, during an acute patient episode in the middle of the night, is an experience that many of us would recount as an invaluable part of our training. It is also a time when on‐site faculty supervision and nurse staffing ratios are typically reduced. In other words, the entire system of care is different at night. This dichotomy in day versus night medicine has spurred research into whether patient outcomes differed between the two. Various investigators have looked at the differences in care, and there is some data to support that patient care outcomes are worse at night than during the day.58

The results from Khanna and colleagues3 are welcome news, since graduate medical education is undergoing a paradigm shift with further reductions in duty hours. Hospital‐based medicine is a 24‐hour operation, 7 days a week. We don't get to pick when patients are admitted or need care. In fact, this study demonstrates that 58% of the admissions came after 7 _PM. It turns out that this pattern of admissions is not atypical.9 In light of the new standards, programs are left deciding where to deploy their 1st‐year resident resources. There are essentially two types of models to provide patient care off‐hours in the absence of an extended shift option: night float and night medicine. Night float implies that a trainee, usually a resident, comes from a different site or rotation to provide nocturnal care of a group of patients. That resident may do one isolated night of coverage or several nights in a row. The resident may perform cross‐cover functions, admit patients, or do both. Night medicine refers to having teams of housestaff at night along with an educational program as part of a rotation. The latter implies that trainees will get the full educational experience whether working days or nights. If one designs a model of care that puts 1st‐year residents where the volume is, they would come in late and leave early in the morning the next day. Both create more handoffs in patient care, which critics of these models have said lead to worse patient outcomes. This research challenges such earlier work and should make us all feel more comfortable that care at night does not need to be different if the proper systems are in place to manage it.

The graduate medical education community needs to take a step back and ask: Where are we going? As a long‐standing program director, I don't know anymore. As training has become competency‐based, and milestones are being developed for each level of training, we need to ask if reducing hours further is in the best interests of graduate medical education and, ultimately, the general public. I have long been interested in work intensity as a contributor to resident fatigue, burnout, and errors. A systematic review of resident workload and work activity looked at 21 studies examining resident time allocation and suggests that at least one‐third of resident time is spent on activities of limited educational value. Eight of the nine studies evaluating the impact of resident workload on patient care showed a negative impact on outcomes, including length of stay, mortality, patient satisfaction, medication errors, and lab utilization.10 If we are going to further restrict duty hours, then we need to transform the learning environment in a similar fashion. It does not seem logical to restrict duty hours further without a major redesign of the learning environment. I fear that all we have done is create more work compression because the system of care will be further strained in its weakest areas.

Hospitalists are uniquely poised to mitigate some of these concerns by being available for direct and indirect supervision, as well as being highly skilled at working in our complex systems. They are experts in systems‐based practice. The extension of direct faculty supervision to nights may be a major factor in eliminating the variances in care between nights and days. Not every program or institution will have the ability to employ night hospitalists, let alone have them in a supervisory and teaching role. Those that do can take great comfort in this research. Those that don't must challenge themselves to think about how to get there.

We all share the collective goal of producing the next generation of well‐trained physicians who will practice safe and effective care. As we begin to embrace competency‐based training, the prism through which we view patient safety must extend beyond duty hours and include learning environment reform. While the Khanna et al.3 study supports that off‐hours care is no longer off, we need to be careful that we don't drift too far towards shift work as the fundamental structure of medical training. In the beginning, residents lived in the hospital, thus the term. As we move through these stormy seas, we don't want to trade patient safety for transient trainees. There is a middle ground here that will be informed by future research. There is something to be said for living in your house.

The word housestaff came from an era when you moved into the hospital (house) for your internship (staff). In those days, call every other night meant you missed half the learning opportunities. In those days, supervision was as remote as it could get. My, have things changed.

In 2003, the Accreditation Council for Graduate Medical Education (ACGME) implemented duty hour standards that called for restricting resident duty hours to no more than 80 hours per week, and no more than 24 continuous hours with an additional 6 hours for transfer of care.1 The expectation was that limiting work hours would have positive effects on outcomes, such as resident fatigue, the educational experience, and patient safety. A recent systematic review commissioned by the ACGME reviewed three decades worth of research on duty hours and patient safety. It reviewed laboratory studies of resident performance on clinical simulations, and tests of cognitive and fine motor skills under conditions of sleep deprivation. The results of these studies imply that limiting residents' duty hours should have positive effects on patient safety. However, that assumption has not been borne out. Paradoxically, examination of studies of alternative duty hour schedules, as well as the follow‐up studies on the impact of the New York duty hour rules and ACGME duty hour regulations have documented few measurable effects on patient safety.2 Thus, the simplistic view that fewer hours worked equals better patient care deserves a second‐look. In fact, a cause‐and‐effect relationship between resident fatigue and medical errors has yet to be proven.

The hospitalist movement was in full swing by 2003 and had already transformed the care of hospitalized patients. The early studies of hospitalists focused on cost and comparing outcomes with those of other providers. As many studies showed the economic benefit of hospitalists, residency programs began redesigning their inpatient services with hospitalists at the helm. Institutions sponsoring graduate medical education programs saw hospitalists as a necessary resource to enable and maintain compliance with duty hour reform. Some institutions created hospitalist‐only services with no housestaff (nonteaching services). Others created hospitalist teaching services that used indirect supervision at night from a hospitalist on‐call from home, while others had on‐site hospitalists at night that provided direct supervision of housestaff. It is clear that hospitalists are a versatile and valuable resource.

In this issue of the Journal of Hospital Medicine, Khanna and colleagues3 describe a retrospective medical record review at a large, urban academic hospital that evaluates the potential association between night admissions by night‐shiftbased hospitalists and a resident night float system, and hospitalization‐relevant outcomes. The outcomes included length of stay, hospital charges, Intensive Care Unit (ICU) transfers, Emergency Department visits within 30 days of discharge, 30‐day readmission rates, and poor outcomes within the first 24 hours of admission. Their results showed that night or weekend admission was not associated with worse hospitalization‐relevant outcomes. The extension of on‐site faculty supervision to nights, as well as awards for nursing excellence, likely had a strong positive impact on these outcomes. In some cases, night or weekend admission was associated with better outcomes, particularly in terms of ICU transfer during hospitalization and hospital charges. Maybe off‐hours care is not so off after all.

I am not sure where we are going, but the duty hour standards have been revised again. The new duty hour standards limit continuous duty hours for 1st‐year residents to 16, and all others to 24. Upper‐level residents can stay an additional 4 hours for transfer of care.4 Functionally, this proposed standard eliminates overnight call for 1st‐year residents. Although I would never advocate returning to the days of yore where autonomy and independence were the rule, I mourn the loss of extended shifts for 1st‐year residents. The loss of the extended shift will come at the expense of overnight call in many programs. I worry about the loss of some unique teaching moments for all residents, and the increased number of transitions in a patients' care. The hospital is a different place at night. It is less chaotic with fewer visitors, and there are less interruptions in patient care. There is time for teaching and learning because the work intensity is lower. The bedside teaching among senior and junior residents, during an acute patient episode in the middle of the night, is an experience that many of us would recount as an invaluable part of our training. It is also a time when on‐site faculty supervision and nurse staffing ratios are typically reduced. In other words, the entire system of care is different at night. This dichotomy in day versus night medicine has spurred research into whether patient outcomes differed between the two. Various investigators have looked at the differences in care, and there is some data to support that patient care outcomes are worse at night than during the day.58

The results from Khanna and colleagues3 are welcome news, since graduate medical education is undergoing a paradigm shift with further reductions in duty hours. Hospital‐based medicine is a 24‐hour operation, 7 days a week. We don't get to pick when patients are admitted or need care. In fact, this study demonstrates that 58% of the admissions came after 7 _PM. It turns out that this pattern of admissions is not atypical.9 In light of the new standards, programs are left deciding where to deploy their 1st‐year resident resources. There are essentially two types of models to provide patient care off‐hours in the absence of an extended shift option: night float and night medicine. Night float implies that a trainee, usually a resident, comes from a different site or rotation to provide nocturnal care of a group of patients. That resident may do one isolated night of coverage or several nights in a row. The resident may perform cross‐cover functions, admit patients, or do both. Night medicine refers to having teams of housestaff at night along with an educational program as part of a rotation. The latter implies that trainees will get the full educational experience whether working days or nights. If one designs a model of care that puts 1st‐year residents where the volume is, they would come in late and leave early in the morning the next day. Both create more handoffs in patient care, which critics of these models have said lead to worse patient outcomes. This research challenges such earlier work and should make us all feel more comfortable that care at night does not need to be different if the proper systems are in place to manage it.

The graduate medical education community needs to take a step back and ask: Where are we going? As a long‐standing program director, I don't know anymore. As training has become competency‐based, and milestones are being developed for each level of training, we need to ask if reducing hours further is in the best interests of graduate medical education and, ultimately, the general public. I have long been interested in work intensity as a contributor to resident fatigue, burnout, and errors. A systematic review of resident workload and work activity looked at 21 studies examining resident time allocation and suggests that at least one‐third of resident time is spent on activities of limited educational value. Eight of the nine studies evaluating the impact of resident workload on patient care showed a negative impact on outcomes, including length of stay, mortality, patient satisfaction, medication errors, and lab utilization.10 If we are going to further restrict duty hours, then we need to transform the learning environment in a similar fashion. It does not seem logical to restrict duty hours further without a major redesign of the learning environment. I fear that all we have done is create more work compression because the system of care will be further strained in its weakest areas.

Hospitalists are uniquely poised to mitigate some of these concerns by being available for direct and indirect supervision, as well as being highly skilled at working in our complex systems. They are experts in systems‐based practice. The extension of direct faculty supervision to nights may be a major factor in eliminating the variances in care between nights and days. Not every program or institution will have the ability to employ night hospitalists, let alone have them in a supervisory and teaching role. Those that do can take great comfort in this research. Those that don't must challenge themselves to think about how to get there.

We all share the collective goal of producing the next generation of well‐trained physicians who will practice safe and effective care. As we begin to embrace competency‐based training, the prism through which we view patient safety must extend beyond duty hours and include learning environment reform. While the Khanna et al.3 study supports that off‐hours care is no longer off, we need to be careful that we don't drift too far towards shift work as the fundamental structure of medical training. In the beginning, residents lived in the hospital, thus the term. As we move through these stormy seas, we don't want to trade patient safety for transient trainees. There is a middle ground here that will be informed by future research. There is something to be said for living in your house.

Discharge summaries (DS) correlate with rates of rehospitalization1, 2 and adverse events after discharge.3 The Joint Commission on the Accreditation of Healthcare Organizations acknowledges their importance and mandates that certain elements be included.4 Thus far, however, DS are not standardized across institutions and there is no expectation that they be available at postdischarge visits. There have been numerous attempts to improve the quality of DS by using more structured formats or computer generated summaries with positive results in term of comprehensiveness, clarity, and practitioner satisfaction58 but with persistence of serious errors and omissions.9

Postgraduate training is often the first opportunity for physicians to learn information transfer management skills. Unfortunately, DS are created by house staff who have minimal training in this area¹¹ and feel like they have to learn by osmosis,12 resulting in poor quality DS and lack of availability at the point of care.1315

Previous research suggested that individualized feedback sessions for Internal Medicine residents improved the quality of certain aspects of their completed DS.10 We postulated that an audit and feedback educational intervention on DS for first year geriatric medicine fellows would also improve their quality. This technique involves chart or case review of clinical practice behaviors for a specific task followed by recommendation of new behaviors when applicable.16 Audit and feedback incorporates adult learning theory,1719 an essential part of continuous quality improvement that fits within the Accreditation Council for Graduate Medical Education (ACGME) competency of practice based learning and improvement,20 as an educational activity.

Methods

Intervention

Audit #1

All available DS for each fellow's first month of inpatient service were audited for completeness of the 21 item discharge summary template by 1 author (AD). The 21 items were focused on 4 distinct periods of the hospitalization: admission, hospital course, discharge planning, and postdischarge care (Figure 1).

Figure 1

Content under each of the 21 items was classified as complete, partially complete, or absent. An item was considered complete if most information was present and appropriate medical terms were used, partially complete if information was unclear, and absent if no information was present for that area of the DS. To ensure investigator reliability, a random sample of 25% of each fellow's DS was scored by 2 additional investigators (RK and HF) and all disagreements were reviewed and resolved by consensus.

Feedback

Between December 2006 and January 2007, one‐on‐one formative feedback sessions were scheduled. The sessions were approximately 30 minutes long, confidential, performed by 1 of the authors (AD) and followed a written format. During these sessions, each fellow received the results of their discharge summary audit, each partially complete or absent item was discussed, and the importance of DS was emphasized.

Audit #2

All available DS for each fellow's second month of inpatient service were audited for completeness, using the same 21 item assessment tool and the same scoring system.

Results

Five fellows participated, 4 of whom were women; 2 were in postgraduate year 4, 3 in year 5. A total of 158 DS were audited, 89 prefeedback and 79 postfeedback. Each fellow dictated an average of 17 DS during each inpatient month.

During Audit #1, the 21 item DS were complete among 71%, incomplete among 18%, absent among 11%. Admission items, hospital course items, and discharge planning items were complete among 70%, 78%, and 77% of DS respectively, but postdischarge items were complete among only 57%. Examining individual items, the lowest completion rates were found for test result follow‐up (42%), caregiver information (10%), and home services (64%), as well for assessment at admission and discharge of cognitive and mental status (56% and 53% respectively) and functional status (57% and 40%). Of note, all these items are of particular importance to geriatric care.

After receiving the audit and feedback intervention, fellows were more likely to complete all required discharge summary data when compared to prior‐to‐feedback (91% vs. 71%, P 0.001). Discharge summary completeness improved for all composite outcomes examining the four domains of care: admission (93% vs. 70%, P 0.001), hospital course (93% vs. 78%, P 0.001), discharge planning (93% vs. 77%, P 0.02), and postdischarge care (83% vs. 57%., P 0.001) (Table 1).

Discussion

Our study found that audit and feedback sessions significantly improved the completeness of DS dictated by geriatric medicine fellows at 1 academic medical center. Before feedback, completeness was high in most traditional areas of the DS including admission data, hospital course, and discharge planning, but was low in other areas critical for safe transitions of older adults such as postdischarge care, test follow‐up, caregiver information, and cognitive and functional status changes. These findings were surprising, as using a template should render a completion rate close to 100%. Notably, during feedback sessions, fellows suggested low completion rates were due to lack of awareness regarding the importance of completing all 21 items of the template and missing documentation in patient medical records.

Feedback sessions dramatically improved overall completeness of subsequent DS and in most of areas of specific importance for geriatric care, although we remain uncertain why all areas did not show improvement (for example, caregiver information completion remained low). One possible explanation is the lack of accurate documentation for all necessary items in the hospital medical record. Moreover, we did not observe completion improvement for other items, ie, diet and activity. Overall, we believe that drawing attention to areas of particular importance to geriatric care transitions and providing learners with individual reports on their performance increased their awareness and motivated changes to their practice, improving discharge summary completion.

Our study has limitations. This study was a pilot intervention without a control group, because of time and budgetary constraints. Also, we were unable to assess for sustainability because the fellows studied for this project graduated after the second audit. Third, we studied discharge summary completion; further research should focus on accuracy of discharge summary content. Finally, while we did not use any advanced technologies or materials, faculty time required to conduct the audit and feedback in this study was estimated at 45 hours. In our opinion this estimate would classify our audit and feedback intervention as a low external cost and moderately‐high human cost intervention, which may represent a potential barrier to generalizability. On the other hand, we believe that even an audit of a small sample of DS done by a physician could provide valuable data for feedback and would involve less faculty time.

Our finding that audit and feedback sessions improved the completeness of DS among house‐staff is important for 2 reasons. First, we were able to demonstrate that focused feedback targeted to areas of particular importance to the transition of older adults changed subsequent behavior and resulted in improved documentation of these areas. Second, our study provides evidence of a programmatic approach to address the ACGME competency of practice‐based learning and improvement. We believe that our intervention can be reproduced by training programs across the country and are hopeful that such interventions will result in improved patient outcomes during critical care transitions such as hospital discharge.

Discharge summaries (DS) correlate with rates of rehospitalization1, 2 and adverse events after discharge.3 The Joint Commission on the Accreditation of Healthcare Organizations acknowledges their importance and mandates that certain elements be included.4 Thus far, however, DS are not standardized across institutions and there is no expectation that they be available at postdischarge visits. There have been numerous attempts to improve the quality of DS by using more structured formats or computer generated summaries with positive results in term of comprehensiveness, clarity, and practitioner satisfaction58 but with persistence of serious errors and omissions.9

Postgraduate training is often the first opportunity for physicians to learn information transfer management skills. Unfortunately, DS are created by house staff who have minimal training in this area¹¹ and feel like they have to learn by osmosis,12 resulting in poor quality DS and lack of availability at the point of care.1315

Previous research suggested that individualized feedback sessions for Internal Medicine residents improved the quality of certain aspects of their completed DS.10 We postulated that an audit and feedback educational intervention on DS for first year geriatric medicine fellows would also improve their quality. This technique involves chart or case review of clinical practice behaviors for a specific task followed by recommendation of new behaviors when applicable.16 Audit and feedback incorporates adult learning theory,1719 an essential part of continuous quality improvement that fits within the Accreditation Council for Graduate Medical Education (ACGME) competency of practice based learning and improvement,20 as an educational activity.

Methods

Intervention

Audit #1

All available DS for each fellow's first month of inpatient service were audited for completeness of the 21 item discharge summary template by 1 author (AD). The 21 items were focused on 4 distinct periods of the hospitalization: admission, hospital course, discharge planning, and postdischarge care (Figure 1).

Figure 1

Content under each of the 21 items was classified as complete, partially complete, or absent. An item was considered complete if most information was present and appropriate medical terms were used, partially complete if information was unclear, and absent if no information was present for that area of the DS. To ensure investigator reliability, a random sample of 25% of each fellow's DS was scored by 2 additional investigators (RK and HF) and all disagreements were reviewed and resolved by consensus.

Feedback

Between December 2006 and January 2007, one‐on‐one formative feedback sessions were scheduled. The sessions were approximately 30 minutes long, confidential, performed by 1 of the authors (AD) and followed a written format. During these sessions, each fellow received the results of their discharge summary audit, each partially complete or absent item was discussed, and the importance of DS was emphasized.

Audit #2

All available DS for each fellow's second month of inpatient service were audited for completeness, using the same 21 item assessment tool and the same scoring system.

Results

Five fellows participated, 4 of whom were women; 2 were in postgraduate year 4, 3 in year 5. A total of 158 DS were audited, 89 prefeedback and 79 postfeedback. Each fellow dictated an average of 17 DS during each inpatient month.

During Audit #1, the 21 item DS were complete among 71%, incomplete among 18%, absent among 11%. Admission items, hospital course items, and discharge planning items were complete among 70%, 78%, and 77% of DS respectively, but postdischarge items were complete among only 57%. Examining individual items, the lowest completion rates were found for test result follow‐up (42%), caregiver information (10%), and home services (64%), as well for assessment at admission and discharge of cognitive and mental status (56% and 53% respectively) and functional status (57% and 40%). Of note, all these items are of particular importance to geriatric care.

After receiving the audit and feedback intervention, fellows were more likely to complete all required discharge summary data when compared to prior‐to‐feedback (91% vs. 71%, P 0.001). Discharge summary completeness improved for all composite outcomes examining the four domains of care: admission (93% vs. 70%, P 0.001), hospital course (93% vs. 78%, P 0.001), discharge planning (93% vs. 77%, P 0.02), and postdischarge care (83% vs. 57%., P 0.001) (Table 1).

Discussion

Our study found that audit and feedback sessions significantly improved the completeness of DS dictated by geriatric medicine fellows at 1 academic medical center. Before feedback, completeness was high in most traditional areas of the DS including admission data, hospital course, and discharge planning, but was low in other areas critical for safe transitions of older adults such as postdischarge care, test follow‐up, caregiver information, and cognitive and functional status changes. These findings were surprising, as using a template should render a completion rate close to 100%. Notably, during feedback sessions, fellows suggested low completion rates were due to lack of awareness regarding the importance of completing all 21 items of the template and missing documentation in patient medical records.

Feedback sessions dramatically improved overall completeness of subsequent DS and in most of areas of specific importance for geriatric care, although we remain uncertain why all areas did not show improvement (for example, caregiver information completion remained low). One possible explanation is the lack of accurate documentation for all necessary items in the hospital medical record. Moreover, we did not observe completion improvement for other items, ie, diet and activity. Overall, we believe that drawing attention to areas of particular importance to geriatric care transitions and providing learners with individual reports on their performance increased their awareness and motivated changes to their practice, improving discharge summary completion.

Our study has limitations. This study was a pilot intervention without a control group, because of time and budgetary constraints. Also, we were unable to assess for sustainability because the fellows studied for this project graduated after the second audit. Third, we studied discharge summary completion; further research should focus on accuracy of discharge summary content. Finally, while we did not use any advanced technologies or materials, faculty time required to conduct the audit and feedback in this study was estimated at 45 hours. In our opinion this estimate would classify our audit and feedback intervention as a low external cost and moderately‐high human cost intervention, which may represent a potential barrier to generalizability. On the other hand, we believe that even an audit of a small sample of DS done by a physician could provide valuable data for feedback and would involve less faculty time.

Our finding that audit and feedback sessions improved the completeness of DS among house‐staff is important for 2 reasons. First, we were able to demonstrate that focused feedback targeted to areas of particular importance to the transition of older adults changed subsequent behavior and resulted in improved documentation of these areas. Second, our study provides evidence of a programmatic approach to address the ACGME competency of practice‐based learning and improvement. We believe that our intervention can be reproduced by training programs across the country and are hopeful that such interventions will result in improved patient outcomes during critical care transitions such as hospital discharge.

Discharge summaries (DS) correlate with rates of rehospitalization1, 2 and adverse events after discharge.3 The Joint Commission on the Accreditation of Healthcare Organizations acknowledges their importance and mandates that certain elements be included.4 Thus far, however, DS are not standardized across institutions and there is no expectation that they be available at postdischarge visits. There have been numerous attempts to improve the quality of DS by using more structured formats or computer generated summaries with positive results in term of comprehensiveness, clarity, and practitioner satisfaction58 but with persistence of serious errors and omissions.9

Postgraduate training is often the first opportunity for physicians to learn information transfer management skills. Unfortunately, DS are created by house staff who have minimal training in this area¹¹ and feel like they have to learn by osmosis,12 resulting in poor quality DS and lack of availability at the point of care.1315

Previous research suggested that individualized feedback sessions for Internal Medicine residents improved the quality of certain aspects of their completed DS.10 We postulated that an audit and feedback educational intervention on DS for first year geriatric medicine fellows would also improve their quality. This technique involves chart or case review of clinical practice behaviors for a specific task followed by recommendation of new behaviors when applicable.16 Audit and feedback incorporates adult learning theory,1719 an essential part of continuous quality improvement that fits within the Accreditation Council for Graduate Medical Education (ACGME) competency of practice based learning and improvement,20 as an educational activity.

Methods

Intervention

Audit #1

All available DS for each fellow's first month of inpatient service were audited for completeness of the 21 item discharge summary template by 1 author (AD). The 21 items were focused on 4 distinct periods of the hospitalization: admission, hospital course, discharge planning, and postdischarge care (Figure 1).

Figure 1

Content under each of the 21 items was classified as complete, partially complete, or absent. An item was considered complete if most information was present and appropriate medical terms were used, partially complete if information was unclear, and absent if no information was present for that area of the DS. To ensure investigator reliability, a random sample of 25% of each fellow's DS was scored by 2 additional investigators (RK and HF) and all disagreements were reviewed and resolved by consensus.

Feedback

Between December 2006 and January 2007, one‐on‐one formative feedback sessions were scheduled. The sessions were approximately 30 minutes long, confidential, performed by 1 of the authors (AD) and followed a written format. During these sessions, each fellow received the results of their discharge summary audit, each partially complete or absent item was discussed, and the importance of DS was emphasized.

Audit #2

All available DS for each fellow's second month of inpatient service were audited for completeness, using the same 21 item assessment tool and the same scoring system.

Results

Five fellows participated, 4 of whom were women; 2 were in postgraduate year 4, 3 in year 5. A total of 158 DS were audited, 89 prefeedback and 79 postfeedback. Each fellow dictated an average of 17 DS during each inpatient month.

During Audit #1, the 21 item DS were complete among 71%, incomplete among 18%, absent among 11%. Admission items, hospital course items, and discharge planning items were complete among 70%, 78%, and 77% of DS respectively, but postdischarge items were complete among only 57%. Examining individual items, the lowest completion rates were found for test result follow‐up (42%), caregiver information (10%), and home services (64%), as well for assessment at admission and discharge of cognitive and mental status (56% and 53% respectively) and functional status (57% and 40%). Of note, all these items are of particular importance to geriatric care.

After receiving the audit and feedback intervention, fellows were more likely to complete all required discharge summary data when compared to prior‐to‐feedback (91% vs. 71%, P 0.001). Discharge summary completeness improved for all composite outcomes examining the four domains of care: admission (93% vs. 70%, P 0.001), hospital course (93% vs. 78%, P 0.001), discharge planning (93% vs. 77%, P 0.02), and postdischarge care (83% vs. 57%., P 0.001) (Table 1).

Discussion

Our study found that audit and feedback sessions significantly improved the completeness of DS dictated by geriatric medicine fellows at 1 academic medical center. Before feedback, completeness was high in most traditional areas of the DS including admission data, hospital course, and discharge planning, but was low in other areas critical for safe transitions of older adults such as postdischarge care, test follow‐up, caregiver information, and cognitive and functional status changes. These findings were surprising, as using a template should render a completion rate close to 100%. Notably, during feedback sessions, fellows suggested low completion rates were due to lack of awareness regarding the importance of completing all 21 items of the template and missing documentation in patient medical records.

Feedback sessions dramatically improved overall completeness of subsequent DS and in most of areas of specific importance for geriatric care, although we remain uncertain why all areas did not show improvement (for example, caregiver information completion remained low). One possible explanation is the lack of accurate documentation for all necessary items in the hospital medical record. Moreover, we did not observe completion improvement for other items, ie, diet and activity. Overall, we believe that drawing attention to areas of particular importance to geriatric care transitions and providing learners with individual reports on their performance increased their awareness and motivated changes to their practice, improving discharge summary completion.

Our study has limitations. This study was a pilot intervention without a control group, because of time and budgetary constraints. Also, we were unable to assess for sustainability because the fellows studied for this project graduated after the second audit. Third, we studied discharge summary completion; further research should focus on accuracy of discharge summary content. Finally, while we did not use any advanced technologies or materials, faculty time required to conduct the audit and feedback in this study was estimated at 45 hours. In our opinion this estimate would classify our audit and feedback intervention as a low external cost and moderately‐high human cost intervention, which may represent a potential barrier to generalizability. On the other hand, we believe that even an audit of a small sample of DS done by a physician could provide valuable data for feedback and would involve less faculty time.

Our finding that audit and feedback sessions improved the completeness of DS among house‐staff is important for 2 reasons. First, we were able to demonstrate that focused feedback targeted to areas of particular importance to the transition of older adults changed subsequent behavior and resulted in improved documentation of these areas. Second, our study provides evidence of a programmatic approach to address the ACGME competency of practice‐based learning and improvement. We believe that our intervention can be reproduced by training programs across the country and are hopeful that such interventions will result in improved patient outcomes during critical care transitions such as hospital discharge.

If you wish to receive credit for this activity, which beginson the next page, please refer to the website: www.blackwellpublishing.com/cme.

Accreditation and Designation Statement

Blackwell Futura Media Services designates this educational activity for a 1 AMA PRA Category 1 Credit. Physicians should only claim credit commensurate with the extent of their participation in the activity.

Blackwell Futura Media Services is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.

Educational Objectives

Continuous participation in the Journal of Hospital Medicine CME program will enable learners to be better able to:

• Interpret clinical guidelines and their applications for higher quality and more efficient care for all hospitalized patients.

• Describe the standard of care for common illnesses and conditions treated in the hospital; such as pneumonia, COPD exacerbation, acute coronary syndrome, HF exacerbation, glycemic control, venous thromboembolic disease, stroke, etc.

• Discuss evidence‐based recommendations involving transitions of care, including the hospital discharge process.

• Gain insights into the roles of hospitalists as medical educators, researchers, medical ethicists, palliative care providers, and hospital‐based geriatricians.

• Incorporate best practices for hospitalist administration, including quality improvement, patient safety, practice management, leadership, and demonstrating hospitalist value.

• Identify evidence‐based best practices and trends for both adult and pediatric hospital medicine.

Instructions on Receiving Credit

For information on applicability and acceptance of continuing medical education credit for this activity, please consult your professional licensing board.

This activity is designed to be completed within the time designated on the title page; physicians should claim only those credits that reflect the time actually spent in the activity. To successfully earn credit, participants must complete the activity during the valid credit period that is noted on the title page.

Follow these steps to earn credit:

• Log on to www.blackwellpublishing.com/cme.

• Read the target audience, learning objectives, and author disclosures.

• Read the article in print or online format.

• Reflect on the article.

• Access the CME Exam, and choose the best answer to each question.

• Complete the required evaluation component of the activity.

If you wish to receive credit for this activity, which beginson the next page, please refer to the website: www.blackwellpublishing.com/cme.

Accreditation and Designation Statement

Blackwell Futura Media Services designates this educational activity for a 1 AMA PRA Category 1 Credit. Physicians should only claim credit commensurate with the extent of their participation in the activity.

Blackwell Futura Media Services is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.

Educational Objectives

Continuous participation in the Journal of Hospital Medicine CME program will enable learners to be better able to:

• Interpret clinical guidelines and their applications for higher quality and more efficient care for all hospitalized patients.

• Describe the standard of care for common illnesses and conditions treated in the hospital; such as pneumonia, COPD exacerbation, acute coronary syndrome, HF exacerbation, glycemic control, venous thromboembolic disease, stroke, etc.

• Discuss evidence‐based recommendations involving transitions of care, including the hospital discharge process.

• Gain insights into the roles of hospitalists as medical educators, researchers, medical ethicists, palliative care providers, and hospital‐based geriatricians.

• Incorporate best practices for hospitalist administration, including quality improvement, patient safety, practice management, leadership, and demonstrating hospitalist value.

• Identify evidence‐based best practices and trends for both adult and pediatric hospital medicine.

Instructions on Receiving Credit

For information on applicability and acceptance of continuing medical education credit for this activity, please consult your professional licensing board.

This activity is designed to be completed within the time designated on the title page; physicians should claim only those credits that reflect the time actually spent in the activity. To successfully earn credit, participants must complete the activity during the valid credit period that is noted on the title page.

Follow these steps to earn credit:

• Log on to www.blackwellpublishing.com/cme.

• Read the target audience, learning objectives, and author disclosures.

• Read the article in print or online format.

• Reflect on the article.

• Access the CME Exam, and choose the best answer to each question.

• Complete the required evaluation component of the activity.

If you wish to receive credit for this activity, which beginson the next page, please refer to the website: www.blackwellpublishing.com/cme.

Accreditation and Designation Statement

Blackwell Futura Media Services designates this educational activity for a 1 AMA PRA Category 1 Credit. Physicians should only claim credit commensurate with the extent of their participation in the activity.

Blackwell Futura Media Services is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.

Educational Objectives

Continuous participation in the Journal of Hospital Medicine CME program will enable learners to be better able to:

• Interpret clinical guidelines and their applications for higher quality and more efficient care for all hospitalized patients.

• Describe the standard of care for common illnesses and conditions treated in the hospital; such as pneumonia, COPD exacerbation, acute coronary syndrome, HF exacerbation, glycemic control, venous thromboembolic disease, stroke, etc.

• Discuss evidence‐based recommendations involving transitions of care, including the hospital discharge process.

• Gain insights into the roles of hospitalists as medical educators, researchers, medical ethicists, palliative care providers, and hospital‐based geriatricians.

• Incorporate best practices for hospitalist administration, including quality improvement, patient safety, practice management, leadership, and demonstrating hospitalist value.

• Identify evidence‐based best practices and trends for both adult and pediatric hospital medicine.

Instructions on Receiving Credit

For information on applicability and acceptance of continuing medical education credit for this activity, please consult your professional licensing board.

This activity is designed to be completed within the time designated on the title page; physicians should claim only those credits that reflect the time actually spent in the activity. To successfully earn credit, participants must complete the activity during the valid credit period that is noted on the title page.

Follow these steps to earn credit:

• Log on to www.blackwellpublishing.com/cme.

• Read the target audience, learning objectives, and author disclosures.

• Read the article in print or online format.

• Reflect on the article.

• Access the CME Exam, and choose the best answer to each question.

• Complete the required evaluation component of the activity.

This issue of the Journal of Hospital Medicine features an important contribution concerning the current state of academic hospital medicine. The survey of 57 hospitalists revealed what many of us already suspected: the state of mentorship in academic hospitalist groups is unsatisfactory.1

While the conclusion is alarming, it is also not surprising. Over the past decade academic medical centers enthusiastically hired hospitalists to improve efficiency for inpatient services and to lessen the effect of Accreditation Council for Graduate Medical Education (ACGME) regulations on duty hours and patient caps. Few departments of internal medicine, however, hired academic hospitalists with the intent of creating academic divisions. Thus many institutions appear to view hospitalists primarily as hospital employees ignoring their potential academic contributions, and as a result it should not be a surprise that many hospitalist groups lack the mentorship infrastructure of other divisions within a typical Department of Medicine. Compounding the hospital employee problem, the new field of academic hospital medicine has emerged only in the last decade, a time frame that has resulted in very few hospitalists qualified to serve as senior mentors.

We cannot easily remove these limitations: the past is the past, and over time, hospital medicine will mature and develop more senior mentors. But what should we do until that maturation occurs? We believe that the academic work of hospitalists, both in education and research (Quality and Patient Safety) are important endeavors too valuable to be left to chance. With 30,000 hospitalists delivering care, it is critical that research in the optimal delivery of this care be performed, targeting systems improvements to enact anticipated outcomes in quality and patient safety. The physicians who are regularly and intimately involved in this system of inpatient care delivery, the hospitalists, are best suited for identifying the unique features of the inpatient care system needing improvement. Mentorship is essential in ensuring the advancement of both areas, and the sustainability of hospital medicine in medical academe. The article by Harrison et al.1 both establishes the depth of these issues and provides important insights into potential solutions for closing this mentorship gap while the field matures.

Utilizing Other Mentors

No measure of systems change will make young hospitalists immediately experienced, such that they have the sophistication to be senior mentors for younger hospitalists. But we can compensate for this temporary gap in mentorship experience. First, in the next 5 to 10 years, young academic hospitalists need explicit direction from those within Departments of Medicine who do have this mentorship experience, even if these mentors do not work in hospital medicine. Mentors within General Internal Medicine or the subspecialties can still provide the guidance and support to ensure that academic hospitalists are engaging in the appropriate endeavors toward promotion and intellectual growth. Second, academic hospitalists have to seek out mentorship from afar through their participation in the organizations primarily devoted to the academic welfare of hospitalists: The Society of Hospital Medicine (SHM) and the Society of General Internal Medicine (SGIM). Both organizations sponsor mentorship programs, and regular attendance at regional and/or national meetings (followed by email correspondence) can greatly improve an academic hospitalist's career trajectory. Finally, midlevel and senior‐level hospitalists have to learn mentorship skills; mere experience in the field does not ensure acquisition of the necessary mentorship skills, anymore than experience in medicine ensures teaching skills. Mentorship is its own skill set, and receiving appropriate training via the SHM or SGIM national meetings or the Academic Hospitalist Academy (referenced below) is critical.

Defining Academic Expectations for Hospitalists

Harrison et al.1 note that academic hospitalists felt there was a lack of respect for the scholarly work that hospitalists do as part of their job, raising the proposition that the mentorship dearth for academic hospitalists might result from currently available mentors not knowing what to say. Even if mentors were plentiful today, we still must ask the question, What would the mentor advise the young hospitalist to do? The academic hospitalist offers extraordinary value to the Department, but in a way that is different from the standard R0RO‐1 Grant paradigm. Even if hospitalists acquire extramural funding, it will likely come from sources different from the National Institutes of Health (NIH): Agency for Healthcare Research and Quality (AHRQ), foundations (eg, The Robert Wood Johnson Foundation or The John A. Hartford Foundation), intramural hospital‐originating funding, etc. And while extramural funding may be a measure of a hospitalist's contribution to the Department, it should not be the only measure of the hospitalist's career development. There are 2 ways to get rich: acquire more money, or spend less money. Academic hospitalists, unlike other specialties in Medicine, are likely to fall into the latter category, by offering decreased hospital costs (ie, decreased length of stay, decreased never events, etc.). Further, hospitalists may save in opportunity costs: the hospitalist staffing a ward service is less costly than a subspecialist who could be performing procedures, or a basic science researcher who could be acquiring grants. The problem today is that there is no way to quantify this decreased loss, and having this sort of metric will greatly enable mentors to provide hospitalists with ways of showing value to the department outside of the standard NIH grant paradigm. The Quality Portfolio developed by the SGIM and the forthcoming Benchmarks for Academic Hospitalists Promotion white paper (as developed by the SHM's Academic Practice and Promotion Committee) will greatly improve the substance of mentorship for academic hospitalists.2 Leaders of academic hospital medicine must learn to educate chairs of medicine and medical school deans as to the value‐added services intrinsic in the integration of hospitalists into the academic environment.

Having an Academic Plan

Mentorship is a 2‐way relationship: the mentor has responsibilities, but so too, does the mentee. As we wait for the hospitalist field to further develop, new academic hospitalists must become proactive in seeking guidance in career development. The Academic Hospitalist Academy, cosponsored by SHM, SGIM, and ACLGIM, is an example of this type of training.3 As a part of this course, participants learn of the rules and the opportunities for success in academic hospital medicine. Success for academic hospitalist groups will likely follow from understanding what success looks like. The Academy provides an excellent program for distributing that knowledge.

Research Training in Hospital Medicine

Many traditionalists would insist that Hospital Medicine could evoke the same training paradigm as other subspecialties in medicine (ie, fellowships). Unfortunately there are not a sufficient number of GME‐funded positions to handle the number of hospitalists required to advance the mission of academic hospital medicine. Moreover, fellowship training for every academic hospitalist would be unlikely to produce the desired results of improving the delivery of inpatient care. The academic agenda for the hospitalist depends on understanding the hospital system, and then executing improvements that lead to safer, more efficient and effective care. In this way, the academic hospitalist academic training is much more akin to a Master of Business Administration (MBA) than it is to a Bachelor of Science (BS) degree: namely, via job immersion, the hospitalist develops a greater systems understanding that should inform his or her academic career. Thus, a fellowship right out of residency may not have the same urgency for the hospitalist as it does for the subspecialist. Nevertheless, those hospitalists seeking an academic scholarly career will experience major benefits from fellowship training. Academic hospitalists need not focus only on the few existing hospitalist fellowships; they can obtain the necessary training in research skills via a general medicine fellowship, of which there are many. For this cohort of hospitalists, we strongly encourage training in a general medicine, health services, or outcomes research fellowship, with an emphasis on research techniques as they apply to the measurement of quality, patient safety, and/or clinical education.

With respect to academic hospitalists, it is likely that nothing is as important as the question of mentorship. Even the hardest working hospitalist can lose their way without guidance and a roadmap; the mentor is central to both. But the lost opportunity is not borne by the individual physician alone; the academic department loses too. Because the hospitalist's value depends on sufficient familiarity with a specific system prior to leveraging improvements, the department accrues maximal benefits in efficiency and effectiveness only if it can maintain retention for at least 2 years.4 The turnover carries major costs; recruitment costs money, and every new hospitalist engenders major start‐up costs. Faculty members who become completely integrated into the department have higher retention rates than those who consider themselves outside the main stream. Mentorship will greatly increase the probability that hospitalists will progress and feel the importance to the department.

Academic hospital medicine must strive over the next 5 to 10 years to become totally integrated in the academic culture of every institution. This task will take great leadership both at the local level and at a national level. We agree with the authors that the SHM and the SGIM can both provide important assistance to young hospital medicine groups. We applaud the authors of this article for making explicit this next major challenge for the field.

This issue of the Journal of Hospital Medicine features an important contribution concerning the current state of academic hospital medicine. The survey of 57 hospitalists revealed what many of us already suspected: the state of mentorship in academic hospitalist groups is unsatisfactory.1

While the conclusion is alarming, it is also not surprising. Over the past decade academic medical centers enthusiastically hired hospitalists to improve efficiency for inpatient services and to lessen the effect of Accreditation Council for Graduate Medical Education (ACGME) regulations on duty hours and patient caps. Few departments of internal medicine, however, hired academic hospitalists with the intent of creating academic divisions. Thus many institutions appear to view hospitalists primarily as hospital employees ignoring their potential academic contributions, and as a result it should not be a surprise that many hospitalist groups lack the mentorship infrastructure of other divisions within a typical Department of Medicine. Compounding the hospital employee problem, the new field of academic hospital medicine has emerged only in the last decade, a time frame that has resulted in very few hospitalists qualified to serve as senior mentors.

We cannot easily remove these limitations: the past is the past, and over time, hospital medicine will mature and develop more senior mentors. But what should we do until that maturation occurs? We believe that the academic work of hospitalists, both in education and research (Quality and Patient Safety) are important endeavors too valuable to be left to chance. With 30,000 hospitalists delivering care, it is critical that research in the optimal delivery of this care be performed, targeting systems improvements to enact anticipated outcomes in quality and patient safety. The physicians who are regularly and intimately involved in this system of inpatient care delivery, the hospitalists, are best suited for identifying the unique features of the inpatient care system needing improvement. Mentorship is essential in ensuring the advancement of both areas, and the sustainability of hospital medicine in medical academe. The article by Harrison et al.1 both establishes the depth of these issues and provides important insights into potential solutions for closing this mentorship gap while the field matures.

Utilizing Other Mentors

No measure of systems change will make young hospitalists immediately experienced, such that they have the sophistication to be senior mentors for younger hospitalists. But we can compensate for this temporary gap in mentorship experience. First, in the next 5 to 10 years, young academic hospitalists need explicit direction from those within Departments of Medicine who do have this mentorship experience, even if these mentors do not work in hospital medicine. Mentors within General Internal Medicine or the subspecialties can still provide the guidance and support to ensure that academic hospitalists are engaging in the appropriate endeavors toward promotion and intellectual growth. Second, academic hospitalists have to seek out mentorship from afar through their participation in the organizations primarily devoted to the academic welfare of hospitalists: The Society of Hospital Medicine (SHM) and the Society of General Internal Medicine (SGIM). Both organizations sponsor mentorship programs, and regular attendance at regional and/or national meetings (followed by email correspondence) can greatly improve an academic hospitalist's career trajectory. Finally, midlevel and senior‐level hospitalists have to learn mentorship skills; mere experience in the field does not ensure acquisition of the necessary mentorship skills, anymore than experience in medicine ensures teaching skills. Mentorship is its own skill set, and receiving appropriate training via the SHM or SGIM national meetings or the Academic Hospitalist Academy (referenced below) is critical.

Defining Academic Expectations for Hospitalists

Harrison et al.1 note that academic hospitalists felt there was a lack of respect for the scholarly work that hospitalists do as part of their job, raising the proposition that the mentorship dearth for academic hospitalists might result from currently available mentors not knowing what to say. Even if mentors were plentiful today, we still must ask the question, What would the mentor advise the young hospitalist to do? The academic hospitalist offers extraordinary value to the Department, but in a way that is different from the standard R0RO‐1 Grant paradigm. Even if hospitalists acquire extramural funding, it will likely come from sources different from the National Institutes of Health (NIH): Agency for Healthcare Research and Quality (AHRQ), foundations (eg, The Robert Wood Johnson Foundation or The John A. Hartford Foundation), intramural hospital‐originating funding, etc. And while extramural funding may be a measure of a hospitalist's contribution to the Department, it should not be the only measure of the hospitalist's career development. There are 2 ways to get rich: acquire more money, or spend less money. Academic hospitalists, unlike other specialties in Medicine, are likely to fall into the latter category, by offering decreased hospital costs (ie, decreased length of stay, decreased never events, etc.). Further, hospitalists may save in opportunity costs: the hospitalist staffing a ward service is less costly than a subspecialist who could be performing procedures, or a basic science researcher who could be acquiring grants. The problem today is that there is no way to quantify this decreased loss, and having this sort of metric will greatly enable mentors to provide hospitalists with ways of showing value to the department outside of the standard NIH grant paradigm. The Quality Portfolio developed by the SGIM and the forthcoming Benchmarks for Academic Hospitalists Promotion white paper (as developed by the SHM's Academic Practice and Promotion Committee) will greatly improve the substance of mentorship for academic hospitalists.2 Leaders of academic hospital medicine must learn to educate chairs of medicine and medical school deans as to the value‐added services intrinsic in the integration of hospitalists into the academic environment.

Having an Academic Plan

Mentorship is a 2‐way relationship: the mentor has responsibilities, but so too, does the mentee. As we wait for the hospitalist field to further develop, new academic hospitalists must become proactive in seeking guidance in career development. The Academic Hospitalist Academy, cosponsored by SHM, SGIM, and ACLGIM, is an example of this type of training.3 As a part of this course, participants learn of the rules and the opportunities for success in academic hospital medicine. Success for academic hospitalist groups will likely follow from understanding what success looks like. The Academy provides an excellent program for distributing that knowledge.

Research Training in Hospital Medicine

Many traditionalists would insist that Hospital Medicine could evoke the same training paradigm as other subspecialties in medicine (ie, fellowships). Unfortunately there are not a sufficient number of GME‐funded positions to handle the number of hospitalists required to advance the mission of academic hospital medicine. Moreover, fellowship training for every academic hospitalist would be unlikely to produce the desired results of improving the delivery of inpatient care. The academic agenda for the hospitalist depends on understanding the hospital system, and then executing improvements that lead to safer, more efficient and effective care. In this way, the academic hospitalist academic training is much more akin to a Master of Business Administration (MBA) than it is to a Bachelor of Science (BS) degree: namely, via job immersion, the hospitalist develops a greater systems understanding that should inform his or her academic career. Thus, a fellowship right out of residency may not have the same urgency for the hospitalist as it does for the subspecialist. Nevertheless, those hospitalists seeking an academic scholarly career will experience major benefits from fellowship training. Academic hospitalists need not focus only on the few existing hospitalist fellowships; they can obtain the necessary training in research skills via a general medicine fellowship, of which there are many. For this cohort of hospitalists, we strongly encourage training in a general medicine, health services, or outcomes research fellowship, with an emphasis on research techniques as they apply to the measurement of quality, patient safety, and/or clinical education.

With respect to academic hospitalists, it is likely that nothing is as important as the question of mentorship. Even the hardest working hospitalist can lose their way without guidance and a roadmap; the mentor is central to both. But the lost opportunity is not borne by the individual physician alone; the academic department loses too. Because the hospitalist's value depends on sufficient familiarity with a specific system prior to leveraging improvements, the department accrues maximal benefits in efficiency and effectiveness only if it can maintain retention for at least 2 years.4 The turnover carries major costs; recruitment costs money, and every new hospitalist engenders major start‐up costs. Faculty members who become completely integrated into the department have higher retention rates than those who consider themselves outside the main stream. Mentorship will greatly increase the probability that hospitalists will progress and feel the importance to the department.

Academic hospital medicine must strive over the next 5 to 10 years to become totally integrated in the academic culture of every institution. This task will take great leadership both at the local level and at a national level. We agree with the authors that the SHM and the SGIM can both provide important assistance to young hospital medicine groups. We applaud the authors of this article for making explicit this next major challenge for the field.

This issue of the Journal of Hospital Medicine features an important contribution concerning the current state of academic hospital medicine. The survey of 57 hospitalists revealed what many of us already suspected: the state of mentorship in academic hospitalist groups is unsatisfactory.1

While the conclusion is alarming, it is also not surprising. Over the past decade academic medical centers enthusiastically hired hospitalists to improve efficiency for inpatient services and to lessen the effect of Accreditation Council for Graduate Medical Education (ACGME) regulations on duty hours and patient caps. Few departments of internal medicine, however, hired academic hospitalists with the intent of creating academic divisions. Thus many institutions appear to view hospitalists primarily as hospital employees ignoring their potential academic contributions, and as a result it should not be a surprise that many hospitalist groups lack the mentorship infrastructure of other divisions within a typical Department of Medicine. Compounding the hospital employee problem, the new field of academic hospital medicine has emerged only in the last decade, a time frame that has resulted in very few hospitalists qualified to serve as senior mentors.

We cannot easily remove these limitations: the past is the past, and over time, hospital medicine will mature and develop more senior mentors. But what should we do until that maturation occurs? We believe that the academic work of hospitalists, both in education and research (Quality and Patient Safety) are important endeavors too valuable to be left to chance. With 30,000 hospitalists delivering care, it is critical that research in the optimal delivery of this care be performed, targeting systems improvements to enact anticipated outcomes in quality and patient safety. The physicians who are regularly and intimately involved in this system of inpatient care delivery, the hospitalists, are best suited for identifying the unique features of the inpatient care system needing improvement. Mentorship is essential in ensuring the advancement of both areas, and the sustainability of hospital medicine in medical academe. The article by Harrison et al.1 both establishes the depth of these issues and provides important insights into potential solutions for closing this mentorship gap while the field matures.

Utilizing Other Mentors

No measure of systems change will make young hospitalists immediately experienced, such that they have the sophistication to be senior mentors for younger hospitalists. But we can compensate for this temporary gap in mentorship experience. First, in the next 5 to 10 years, young academic hospitalists need explicit direction from those within Departments of Medicine who do have this mentorship experience, even if these mentors do not work in hospital medicine. Mentors within General Internal Medicine or the subspecialties can still provide the guidance and support to ensure that academic hospitalists are engaging in the appropriate endeavors toward promotion and intellectual growth. Second, academic hospitalists have to seek out mentorship from afar through their participation in the organizations primarily devoted to the academic welfare of hospitalists: The Society of Hospital Medicine (SHM) and the Society of General Internal Medicine (SGIM). Both organizations sponsor mentorship programs, and regular attendance at regional and/or national meetings (followed by email correspondence) can greatly improve an academic hospitalist's career trajectory. Finally, midlevel and senior‐level hospitalists have to learn mentorship skills; mere experience in the field does not ensure acquisition of the necessary mentorship skills, anymore than experience in medicine ensures teaching skills. Mentorship is its own skill set, and receiving appropriate training via the SHM or SGIM national meetings or the Academic Hospitalist Academy (referenced below) is critical.

Defining Academic Expectations for Hospitalists

Harrison et al.1 note that academic hospitalists felt there was a lack of respect for the scholarly work that hospitalists do as part of their job, raising the proposition that the mentorship dearth for academic hospitalists might result from currently available mentors not knowing what to say. Even if mentors were plentiful today, we still must ask the question, What would the mentor advise the young hospitalist to do? The academic hospitalist offers extraordinary value to the Department, but in a way that is different from the standard R0RO‐1 Grant paradigm. Even if hospitalists acquire extramural funding, it will likely come from sources different from the National Institutes of Health (NIH): Agency for Healthcare Research and Quality (AHRQ), foundations (eg, The Robert Wood Johnson Foundation or The John A. Hartford Foundation), intramural hospital‐originating funding, etc. And while extramural funding may be a measure of a hospitalist's contribution to the Department, it should not be the only measure of the hospitalist's career development. There are 2 ways to get rich: acquire more money, or spend less money. Academic hospitalists, unlike other specialties in Medicine, are likely to fall into the latter category, by offering decreased hospital costs (ie, decreased length of stay, decreased never events, etc.). Further, hospitalists may save in opportunity costs: the hospitalist staffing a ward service is less costly than a subspecialist who could be performing procedures, or a basic science researcher who could be acquiring grants. The problem today is that there is no way to quantify this decreased loss, and having this sort of metric will greatly enable mentors to provide hospitalists with ways of showing value to the department outside of the standard NIH grant paradigm. The Quality Portfolio developed by the SGIM and the forthcoming Benchmarks for Academic Hospitalists Promotion white paper (as developed by the SHM's Academic Practice and Promotion Committee) will greatly improve the substance of mentorship for academic hospitalists.2 Leaders of academic hospital medicine must learn to educate chairs of medicine and medical school deans as to the value‐added services intrinsic in the integration of hospitalists into the academic environment.

Having an Academic Plan

Mentorship is a 2‐way relationship: the mentor has responsibilities, but so too, does the mentee. As we wait for the hospitalist field to further develop, new academic hospitalists must become proactive in seeking guidance in career development. The Academic Hospitalist Academy, cosponsored by SHM, SGIM, and ACLGIM, is an example of this type of training.3 As a part of this course, participants learn of the rules and the opportunities for success in academic hospital medicine. Success for academic hospitalist groups will likely follow from understanding what success looks like. The Academy provides an excellent program for distributing that knowledge.

Research Training in Hospital Medicine

Many traditionalists would insist that Hospital Medicine could evoke the same training paradigm as other subspecialties in medicine (ie, fellowships). Unfortunately there are not a sufficient number of GME‐funded positions to handle the number of hospitalists required to advance the mission of academic hospital medicine. Moreover, fellowship training for every academic hospitalist would be unlikely to produce the desired results of improving the delivery of inpatient care. The academic agenda for the hospitalist depends on understanding the hospital system, and then executing improvements that lead to safer, more efficient and effective care. In this way, the academic hospitalist academic training is much more akin to a Master of Business Administration (MBA) than it is to a Bachelor of Science (BS) degree: namely, via job immersion, the hospitalist develops a greater systems understanding that should inform his or her academic career. Thus, a fellowship right out of residency may not have the same urgency for the hospitalist as it does for the subspecialist. Nevertheless, those hospitalists seeking an academic scholarly career will experience major benefits from fellowship training. Academic hospitalists need not focus only on the few existing hospitalist fellowships; they can obtain the necessary training in research skills via a general medicine fellowship, of which there are many. For this cohort of hospitalists, we strongly encourage training in a general medicine, health services, or outcomes research fellowship, with an emphasis on research techniques as they apply to the measurement of quality, patient safety, and/or clinical education.

With respect to academic hospitalists, it is likely that nothing is as important as the question of mentorship. Even the hardest working hospitalist can lose their way without guidance and a roadmap; the mentor is central to both. But the lost opportunity is not borne by the individual physician alone; the academic department loses too. Because the hospitalist's value depends on sufficient familiarity with a specific system prior to leveraging improvements, the department accrues maximal benefits in efficiency and effectiveness only if it can maintain retention for at least 2 years.4 The turnover carries major costs; recruitment costs money, and every new hospitalist engenders major start‐up costs. Faculty members who become completely integrated into the department have higher retention rates than those who consider themselves outside the main stream. Mentorship will greatly increase the probability that hospitalists will progress and feel the importance to the department.

Academic hospital medicine must strive over the next 5 to 10 years to become totally integrated in the academic culture of every institution. This task will take great leadership both at the local level and at a national level. We agree with the authors that the SHM and the SGIM can both provide important assistance to young hospital medicine groups. We applaud the authors of this article for making explicit this next major challenge for the field.

Nosocomial, or hospital‐acquired, infections (HAIs) are a major cause of patient morbidity and mortality in the United States and other countries.15 In 2002, approximately 1.7 million HAIs occurred in US hospitals and were associated with an estimated 98,987 deaths.1 Of particular note, increasing percentages of HAIs are now caused by antimicrobial‐resistant pathogens, which have been linked with increases in morbidity, mortality, length of hospital stay, and healthcare costs.6

The 2004 data summary from the United States National Nosocomial Infections Surveillance (NNIS) System Report highlighted substantial increases for year 2003 versus 1998 through 2002 in vancomycin‐resistant enterococci (VRE); methicillin‐resistant Staphylococcus aureus; Klebsiella pneumoniae resistant to third‐generation cephalosporins; and Pseudomonas aeruginosa resistant to imipenem, quinolones, or third‐generation cephalosporins.7 Other gram‐negative bacteria of concern include Escherichia coli and Acinetobacter baumannii, as well as Enterobacter cloacae and E. aerogenes.8, 9

The increasing number of multidrug‐resistant (MDR) gram‐negative bacteria within the healthcare setting is particularly concerning.1013 Too frequently, clinicians in the United States now encounter gram‐negative bacteria species that are resistant to many, and occasionally all, currently available antibiotics. For many of these MDR gram‐negative pathogens the antimicrobials that potentially remain active (eg, aminoglycosides and polymyxins) are often more toxic and less efficacious for some infections.14 Particularly problematic is that the pharmaceutical industry's developmental pipeline for new antibiotics, with novel mechanisms of action that might be used against MDR gram‐negative pathogens, has virtually come to a standstill.15, 16 Even if an investigational drug was in phase 2 or 3 trials right now or entered the US Food and Drug Administration (FDA) Fast Track Development Program, it would be at least 10 or 15 years before that drug would be available on the US market.

What this means is that the clinician's current antibiotic armamentarium is all they can expect in the foreseeable future. It also means that special care needs to be taken to optimally use currently available agents to ensure continued activity against the pathogens encountered in the hospital (and community) setting, now and in the future. Maximizing clinical outcomes, while minimizing the emergence and spread of antimicrobial resistance (and other adverse effects associated with suboptimal antimicrobial drug use), falls under the purview of antimicrobial stewardship, the focus of this paper.

Antimicrobial StewardshipWhy Is It Needed, What Is It, and What Are Its Goals?

Constructing an Antimicrobial Stewardship Program

A Case Study: the Wake Forest University Baptist Medical Center (WFUBMC) Program

Figure 1A,B provides an overview of the general structure of the antimicrobial stewardship program at WFUBMC implemented in 2000. To establish and provide the information needed for day‐to‐day operations of a stewardship program at WFUBMC, data are needed on how, where, and by whom antibiotics are used within the institution. In addition, microbiology data, including the frequency and susceptibility of infecting pathogens, are essential. Obtaining these data often requires the help and cooperation of the information technology (IT) staff at the institution. Considerable time and effort may be required at the outset, but once information system programs are established, ongoing data mining is much easier. At the time of program initiation, it was decided to assess antibiotic use or density (amount of drug per inpatient geographic unit) using the defined‐daily‐dose (DDD) method. This entails assigning a predetermined weight of administered antibiotic as a dose and dividing by a denominator of 1000 patient days. Subsequently, days of therapy per drug has been found to be a more accurate measure of antimicrobial consumption. When developing a program, it is vital to first obtain baseline usage data. Such data should include, if possible, a detailed inventory of usage within different units of the hospital or for particular services, or sometimes even for a particular provider with a history of high antimicrobial usage. Ongoing measurement over time allows the impact of new stewardship interventions and guidelines to be measured, as well as identifying potential new problem areas in usage.

Figure 1

Good microbiology data are also essential to determine problem pathogens at the institution and where they are located. Such data are useful not only to define areas of resistance (potentially warranting changes in antimicrobial policies to alter selection pressures), but also for gathering information necessary for defining local guidelines for antimicrobial use. For example, if the local antibiograms show that a particular pathogen in the hospital ICU has a particular resistance pattern, then initial empiric therapy for patients at risk of infection with those organisms should be chosen to cover the problematic resistant pathogen. Once subsequent microbiology data become available, patients not infected with the pathogen can be de‐escalated to a more narrow‐spectrum antibiotic.

As illustrated in Figure 1A, at WFUBMC, all the collected data are integrated to provide information concerning antibiotic density, usage, and patterns of antimicrobial resistance. This information is received by the staff of the antimicrobial stewardship program, which at WFUBMC is called the Center for Antimicrobial Utilization, Stewardship and Epidemiology (CAUSE). The CAUSE staff works with the day‐to‐day elements of program administration and operations and includes 2 infectious diseases physicians and 2 infectious disease PharmDs. The CAUSE staff works very closely with the microbiology laboratory, hospital pharmacy, and the medical director of hospital epidemiology and infection control.

The CAUSE program at WFUBMC administratively functions through an advisory board committee that includes thought leaders from different medical specialties and patient units of the hospitalparticularly those with high antimicrobial usage, such as hematology/oncology, pulmonary, critical care, and transplantation. The CAUSE staff and advisory board exist to exchange ideas concerning what is working or not working and where problem areas may be, and to propose possible changes in antimicrobial practices at the institution. In addition, thought leaders on the advisory board also receive and evaluate information from various sources about new antimicrobial agents and national guidelines, and, in turn, help disseminate this information to the hospital personnel who will be involved in program implementation.

At WFUBMC, it is the advisory board committee, working in conjunction with the CAUSE staff/medical director, that presents antimicrobials for formulary consideration to the Pharmacy and Therapeutics (P&T) Committee, in addition to any major interventions CAUSE and its advisory board feel are indicated. The P&T committee then reports to the medical staff executive committee and hospital leadership. As should be evident, the approach to stewardship at WFUBMC is medical staff‐driven, rather than a function of administrative constituents.

Finally, no matter how well‐organized an antimicrobial stewardship program is, it will not be fully successful if the entire medical staff does not buy into the process and agree with the need for the proposed changes and interventions involving the practice of antimicrobial therapy. It is important to spend some time early in program development to ensure that the need for an antimicrobial stewardship program, the process, and the outcomes (both in terms of patient care and clinical outcomes at the institution) are clearly communicated to the medical staff, and that their full commitment and cooperation are enlisted. In cases where hospital‐wide infection or resistance rates are known and antimicrobial utilization data are available, it is important to present such information in an understandable and convincing manner that makes the case for a proposed change or intervention, not only at the hospital level but also at the level of the patient.

Elements of a Successful Program: Basic Strategies

Potential strategies or elements of an antimicrobial stewardship program are listed in Table 2. Two evidence‐based fundamental or core strategies have been recommended by the IDSA/SHEA guidelines24 and implemented at numerous institutions with various levels of success. The first is a so‐called back‐end approach to modifying antimicrobial therapy on the basis of prospective audit of antimicrobial use, with intervention and feedback to the provider. The second is a front‐end approach using formulary restriction and preauthorization requirements for specific antimicrobial agents. Various supplemental strategies, including large group and patient case‐based education, guidelines and clinical pathways, antimicrobial order forms, and computerized clinical decision support, are also recommended.

Prospective Audit With Intervention and Feedback

This approach usually involves the use of an antimicrobial support team that reviews initial or ongoing therapy and then intervenes to provide feedback and suggested modifications to the medical care provider to improve therapy. This can be done by an infectious diseases physician, a clinical pharmacist, or a hospitalist or internist with expertise in antimicrobial therapy. The aim is to provide patient‐specific education and/or suggest changes to antimicrobial utilization (when needed) to improve and streamline therapy. Suggested interventions could include discontinuing or changing 1 or more drugs, switching intravenous to oral drug administration, and suggesting a short‐course duration of therapy. Occasionally, suggestions are made when appropriate to actually escalate or intensify therapy to increase efficacy. Identification of patients for targeting or focusing prospective audit and feedback efforts typically involves using computer surveillance to single out problem antimicrobials or problematic usage, given local resistance patterns or patient characteristics.24 Examples could include a focus on asymptomatic bacteruria, excessive duration of therapy for ventilator‐associated pneumonia, or overzealous use of certain classes of antimicrobials. Another potential activity for a prospective audit and feedback team is to review reports of patient‐specific blood and sterile body fluid culture results matched to the patient's current antimicrobial therapy. This allows for daily review of the appropriateness of therapy for potentially serious infections. Some patients seen by the antibiotic support team may be referred for infectious diseases or other expert consultative opinion if their infections or therapy are felt to be too complicated for routine prospective audit and feedback recommendations.

A number of studies have demonstrated that strategies involving prospective audit with intervention and feedback can improve antibiotic stewardship, as measured by reductions in inappropriate antibiotic use,30 reduced antibiotic costs or overall consumption,3135 greater compliance with hospital treatment guidelines or policies33, 36, 37 and, in some cases, reduced number of infections due to C difficile infection32, 37 or resistant pathogens.31, 32, 37 Prospective audit with feedback is probably the best and most effective core strategy for a community hospital program where other interventions are cumbersome or not as well tolerated by the medical staff. One potential disadvantage of the prospective audit with intervention and feedback approach is that medical provider adherence is largely voluntary. The team can make suggestions, but if the provider disagrees or is unobtainable, the suggestion is never implemented. Also, this strategy can also be resource‐intensive from a personnel perspective.

Evaluating Antimicrobial Stewardship Programs

Two general types of measures are used to evaluate the effectiveness of antimicrobial stewardship: process and outcome. As with most things done in the hospital, process measures are easier. They measure surrogate impacts of a program, accountability, resource use, and cost effectiveness. In essence, process measures evaluate whether the program accomplished what it set out to do in terms of changing certain processes or prescriber behaviors. It is important to measure resource use, as this helps to continue funding and to keep workers involved in the project. Good programs will save money; this can easily be measured, even if it is just as simple as going to the hospital pharmacy and looking at the cost of antimicrobials provided per patient day.

Outcomes like decreases in particular infections, less emergence of antimicrobial resistance, or other patient‐specific measures are likely more important in the big picture, but they are also much more difficult to measure. For example, where one would like to measure changes in pathogen resistance after making some changes in antimicrobial stewardship, it often takes years before the benefits of a particular intervention or change materialize in terms of less resistance or reduced emergence of resistance. If that type of change is to be measured, then one needs to be persistent and continue measurements over a long period of time. In addition, given the protracted amount of time before these outcomes may be observed, a number of other changes are likely to happen that coincide with the antimicrobial stewardship interventions and make assessment of causality difficult and biased.

Having said that, a number of studies have demonstrated a relationship between antibiotic restriction48, 8285 or other antimicrobial stewardship policies32, 86 and decreases in nosocomial C difficile infections or disease. Figure 2 illustrates the impact of a nonrestrictive antimicrobial stewardship program at a secondary/tertiary‐care hospital in Quebec, Canada, on an epidemic of C difficileassociated disease (CDAD) that occurred at the institution during the latter portion of 2003.86 Following program implementation, and the major drop in targeted antibiotic consumption, the incidence of CDAD also significantly decreased. Earlier implementation of infection control measures had no effect on CDAD incidence.

Figure 2

A smaller number of studies have reported decreases in resistant gram‐negative bacteria following implementation of antimicrobial stewardship programs. For example, Meyer and colleagues reported a marked reduction in ceftazidime‐resistant K pneumoniae at a 487‐bed general hospital in New York City after implementation of enhanced ceftazidime restriction and barrier precautions following an outbreak of infections caused by the resistant K pneumoniae.87 Similarly, Carling and coworkers reported a significant decrease in nosocomial infections caused by resistant Enterobacteriaceae following implementation of a multidisciplinary antibiotic stewardship program to minimize inappropriate use of third‐generation cephalosporins (Figure 3).32 More recently, a retrospective, longitudinal, multicenter analysis of a consortium of 22 academic health centers in the United States showed that incidence rates of carbapenem‐resistant P aeruginosa were lower at hospitals that restricted carbapenems than those that did not (P = .01).88

Figure 3

Evidence suggesting a beneficial impact of antimicrobial stewardship programs on resistance in gram‐positive organisms is limited. More specifically, the study by Carling and colleagues reported an apparent decrease in VRE rates following implementation of their program to reduce inappropriate use of third‐generation cephalosporins.32 The hospital had VRE rates similar to other NNIS System hospitals prior to beginning the program, but after antibiotic stewardship measures were implemented, the VRE rate began to drop, falling to 6% by 1999. This should be compared with a VRE rate of 24% for similar NNIS System hospitals in 1999.

As far as reducing healthcare costs, Figure 4A illustrates the direct antimicrobial cost savings at WFUBMC after implementation of the CAUSE antimicrobial stewardship program, and Figure 4B after supplemental interventions were implemented. Although decreasing antimicrobial cost is important, one would like to show decreases in overall healthcare costs associated with an antimicrobial stewardship program. Unfortunately, this is often a little more difficult to demonstrate. Extrapolations, however, may be possible. Because antimicrobial resistance, adverse drug effects, and secondary unintended infections such as C difficile colitis have been linked with increased patient morbidity and mortality, longer hospital stays, and increased healthcare costs,6, 89, 90 improved antimicrobial stewardship is expected to optimize patient care and lower overall healthcare costs. A study in a large tertiary care academic medical center estimated more than $4.25 million in total healthcare savings over 1 year with a stewardship program using both preauthorization and, to a lesser extent, prospective audit and feedback.91 Despite the fact cost saving should not be a primary goal of an antimicrobial stewardship program, lower antimicrobial costs can help keep a program funded and buttress a proposal for an antimicrobial stewardship program to hospital leadership.

Figure 4

Many hospitals recognize other areas where an antimicrobial stewardship program can demonstrate its value. This includes implementation of a rapid change in drug utilization during antimicrobial supply shortages and assistance with regulatory mandates and surgical infection prophylaxis. Stewardship teams often assist microbiology with protocols for microbiology reporting, antibiograms, planning for susceptibility testing, and communicating changes in microbiology tests or protocols to clinicians.

Conclusions

Overuse or misuse of antibiotics and other antimicrobials for hospital inpatients is relatively common, and can be associated with several unintended negative consequences. Improving medical care necessarily includes better use of antimicrobials to optimize outcomes and preserve the effectiveness of currently available agents. Further, an important additional consequence of effective antimicrobial stewardship and improved patient care is typically a lowering of overall healthcare costs. The recent 2007 IDSA/SHEA guidelines provide recommendations for developing an institutional program to enhance antimicrobial stewardship. However, individual institutions need to look closely at their own systems and patients to develop an antimicrobial stewardship program that best serves the needs of their hospital and the people it serves.

Nosocomial, or hospital‐acquired, infections (HAIs) are a major cause of patient morbidity and mortality in the United States and other countries.15 In 2002, approximately 1.7 million HAIs occurred in US hospitals and were associated with an estimated 98,987 deaths.1 Of particular note, increasing percentages of HAIs are now caused by antimicrobial‐resistant pathogens, which have been linked with increases in morbidity, mortality, length of hospital stay, and healthcare costs.6

The 2004 data summary from the United States National Nosocomial Infections Surveillance (NNIS) System Report highlighted substantial increases for year 2003 versus 1998 through 2002 in vancomycin‐resistant enterococci (VRE); methicillin‐resistant Staphylococcus aureus; Klebsiella pneumoniae resistant to third‐generation cephalosporins; and Pseudomonas aeruginosa resistant to imipenem, quinolones, or third‐generation cephalosporins.7 Other gram‐negative bacteria of concern include Escherichia coli and Acinetobacter baumannii, as well as Enterobacter cloacae and E. aerogenes.8, 9

The increasing number of multidrug‐resistant (MDR) gram‐negative bacteria within the healthcare setting is particularly concerning.1013 Too frequently, clinicians in the United States now encounter gram‐negative bacteria species that are resistant to many, and occasionally all, currently available antibiotics. For many of these MDR gram‐negative pathogens the antimicrobials that potentially remain active (eg, aminoglycosides and polymyxins) are often more toxic and less efficacious for some infections.14 Particularly problematic is that the pharmaceutical industry's developmental pipeline for new antibiotics, with novel mechanisms of action that might be used against MDR gram‐negative pathogens, has virtually come to a standstill.15, 16 Even if an investigational drug was in phase 2 or 3 trials right now or entered the US Food and Drug Administration (FDA) Fast Track Development Program, it would be at least 10 or 15 years before that drug would be available on the US market.

What this means is that the clinician's current antibiotic armamentarium is all they can expect in the foreseeable future. It also means that special care needs to be taken to optimally use currently available agents to ensure continued activity against the pathogens encountered in the hospital (and community) setting, now and in the future. Maximizing clinical outcomes, while minimizing the emergence and spread of antimicrobial resistance (and other adverse effects associated with suboptimal antimicrobial drug use), falls under the purview of antimicrobial stewardship, the focus of this paper.

Antimicrobial StewardshipWhy Is It Needed, What Is It, and What Are Its Goals?

Constructing an Antimicrobial Stewardship Program

A Case Study: the Wake Forest University Baptist Medical Center (WFUBMC) Program

Figure 1A,B provides an overview of the general structure of the antimicrobial stewardship program at WFUBMC implemented in 2000. To establish and provide the information needed for day‐to‐day operations of a stewardship program at WFUBMC, data are needed on how, where, and by whom antibiotics are used within the institution. In addition, microbiology data, including the frequency and susceptibility of infecting pathogens, are essential. Obtaining these data often requires the help and cooperation of the information technology (IT) staff at the institution. Considerable time and effort may be required at the outset, but once information system programs are established, ongoing data mining is much easier. At the time of program initiation, it was decided to assess antibiotic use or density (amount of drug per inpatient geographic unit) using the defined‐daily‐dose (DDD) method. This entails assigning a predetermined weight of administered antibiotic as a dose and dividing by a denominator of 1000 patient days. Subsequently, days of therapy per drug has been found to be a more accurate measure of antimicrobial consumption. When developing a program, it is vital to first obtain baseline usage data. Such data should include, if possible, a detailed inventory of usage within different units of the hospital or for particular services, or sometimes even for a particular provider with a history of high antimicrobial usage. Ongoing measurement over time allows the impact of new stewardship interventions and guidelines to be measured, as well as identifying potential new problem areas in usage.

Figure 1

Good microbiology data are also essential to determine problem pathogens at the institution and where they are located. Such data are useful not only to define areas of resistance (potentially warranting changes in antimicrobial policies to alter selection pressures), but also for gathering information necessary for defining local guidelines for antimicrobial use. For example, if the local antibiograms show that a particular pathogen in the hospital ICU has a particular resistance pattern, then initial empiric therapy for patients at risk of infection with those organisms should be chosen to cover the problematic resistant pathogen. Once subsequent microbiology data become available, patients not infected with the pathogen can be de‐escalated to a more narrow‐spectrum antibiotic.

As illustrated in Figure 1A, at WFUBMC, all the collected data are integrated to provide information concerning antibiotic density, usage, and patterns of antimicrobial resistance. This information is received by the staff of the antimicrobial stewardship program, which at WFUBMC is called the Center for Antimicrobial Utilization, Stewardship and Epidemiology (CAUSE). The CAUSE staff works with the day‐to‐day elements of program administration and operations and includes 2 infectious diseases physicians and 2 infectious disease PharmDs. The CAUSE staff works very closely with the microbiology laboratory, hospital pharmacy, and the medical director of hospital epidemiology and infection control.

The CAUSE program at WFUBMC administratively functions through an advisory board committee that includes thought leaders from different medical specialties and patient units of the hospitalparticularly those with high antimicrobial usage, such as hematology/oncology, pulmonary, critical care, and transplantation. The CAUSE staff and advisory board exist to exchange ideas concerning what is working or not working and where problem areas may be, and to propose possible changes in antimicrobial practices at the institution. In addition, thought leaders on the advisory board also receive and evaluate information from various sources about new antimicrobial agents and national guidelines, and, in turn, help disseminate this information to the hospital personnel who will be involved in program implementation.

At WFUBMC, it is the advisory board committee, working in conjunction with the CAUSE staff/medical director, that presents antimicrobials for formulary consideration to the Pharmacy and Therapeutics (P&T) Committee, in addition to any major interventions CAUSE and its advisory board feel are indicated. The P&T committee then reports to the medical staff executive committee and hospital leadership. As should be evident, the approach to stewardship at WFUBMC is medical staff‐driven, rather than a function of administrative constituents.

Finally, no matter how well‐organized an antimicrobial stewardship program is, it will not be fully successful if the entire medical staff does not buy into the process and agree with the need for the proposed changes and interventions involving the practice of antimicrobial therapy. It is important to spend some time early in program development to ensure that the need for an antimicrobial stewardship program, the process, and the outcomes (both in terms of patient care and clinical outcomes at the institution) are clearly communicated to the medical staff, and that their full commitment and cooperation are enlisted. In cases where hospital‐wide infection or resistance rates are known and antimicrobial utilization data are available, it is important to present such information in an understandable and convincing manner that makes the case for a proposed change or intervention, not only at the hospital level but also at the level of the patient.

Elements of a Successful Program: Basic Strategies

Potential strategies or elements of an antimicrobial stewardship program are listed in Table 2. Two evidence‐based fundamental or core strategies have been recommended by the IDSA/SHEA guidelines24 and implemented at numerous institutions with various levels of success. The first is a so‐called back‐end approach to modifying antimicrobial therapy on the basis of prospective audit of antimicrobial use, with intervention and feedback to the provider. The second is a front‐end approach using formulary restriction and preauthorization requirements for specific antimicrobial agents. Various supplemental strategies, including large group and patient case‐based education, guidelines and clinical pathways, antimicrobial order forms, and computerized clinical decision support, are also recommended.

Prospective Audit With Intervention and Feedback

This approach usually involves the use of an antimicrobial support team that reviews initial or ongoing therapy and then intervenes to provide feedback and suggested modifications to the medical care provider to improve therapy. This can be done by an infectious diseases physician, a clinical pharmacist, or a hospitalist or internist with expertise in antimicrobial therapy. The aim is to provide patient‐specific education and/or suggest changes to antimicrobial utilization (when needed) to improve and streamline therapy. Suggested interventions could include discontinuing or changing 1 or more drugs, switching intravenous to oral drug administration, and suggesting a short‐course duration of therapy. Occasionally, suggestions are made when appropriate to actually escalate or intensify therapy to increase efficacy. Identification of patients for targeting or focusing prospective audit and feedback efforts typically involves using computer surveillance to single out problem antimicrobials or problematic usage, given local resistance patterns or patient characteristics.24 Examples could include a focus on asymptomatic bacteruria, excessive duration of therapy for ventilator‐associated pneumonia, or overzealous use of certain classes of antimicrobials. Another potential activity for a prospective audit and feedback team is to review reports of patient‐specific blood and sterile body fluid culture results matched to the patient's current antimicrobial therapy. This allows for daily review of the appropriateness of therapy for potentially serious infections. Some patients seen by the antibiotic support team may be referred for infectious diseases or other expert consultative opinion if their infections or therapy are felt to be too complicated for routine prospective audit and feedback recommendations.

A number of studies have demonstrated that strategies involving prospective audit with intervention and feedback can improve antibiotic stewardship, as measured by reductions in inappropriate antibiotic use,30 reduced antibiotic costs or overall consumption,3135 greater compliance with hospital treatment guidelines or policies33, 36, 37 and, in some cases, reduced number of infections due to C difficile infection32, 37 or resistant pathogens.31, 32, 37 Prospective audit with feedback is probably the best and most effective core strategy for a community hospital program where other interventions are cumbersome or not as well tolerated by the medical staff. One potential disadvantage of the prospective audit with intervention and feedback approach is that medical provider adherence is largely voluntary. The team can make suggestions, but if the provider disagrees or is unobtainable, the suggestion is never implemented. Also, this strategy can also be resource‐intensive from a personnel perspective.

Evaluating Antimicrobial Stewardship Programs

Two general types of measures are used to evaluate the effectiveness of antimicrobial stewardship: process and outcome. As with most things done in the hospital, process measures are easier. They measure surrogate impacts of a program, accountability, resource use, and cost effectiveness. In essence, process measures evaluate whether the program accomplished what it set out to do in terms of changing certain processes or prescriber behaviors. It is important to measure resource use, as this helps to continue funding and to keep workers involved in the project. Good programs will save money; this can easily be measured, even if it is just as simple as going to the hospital pharmacy and looking at the cost of antimicrobials provided per patient day.

Outcomes like decreases in particular infections, less emergence of antimicrobial resistance, or other patient‐specific measures are likely more important in the big picture, but they are also much more difficult to measure. For example, where one would like to measure changes in pathogen resistance after making some changes in antimicrobial stewardship, it often takes years before the benefits of a particular intervention or change materialize in terms of less resistance or reduced emergence of resistance. If that type of change is to be measured, then one needs to be persistent and continue measurements over a long period of time. In addition, given the protracted amount of time before these outcomes may be observed, a number of other changes are likely to happen that coincide with the antimicrobial stewardship interventions and make assessment of causality difficult and biased.

Having said that, a number of studies have demonstrated a relationship between antibiotic restriction48, 8285 or other antimicrobial stewardship policies32, 86 and decreases in nosocomial C difficile infections or disease. Figure 2 illustrates the impact of a nonrestrictive antimicrobial stewardship program at a secondary/tertiary‐care hospital in Quebec, Canada, on an epidemic of C difficileassociated disease (CDAD) that occurred at the institution during the latter portion of 2003.86 Following program implementation, and the major drop in targeted antibiotic consumption, the incidence of CDAD also significantly decreased. Earlier implementation of infection control measures had no effect on CDAD incidence.

Figure 2

A smaller number of studies have reported decreases in resistant gram‐negative bacteria following implementation of antimicrobial stewardship programs. For example, Meyer and colleagues reported a marked reduction in ceftazidime‐resistant K pneumoniae at a 487‐bed general hospital in New York City after implementation of enhanced ceftazidime restriction and barrier precautions following an outbreak of infections caused by the resistant K pneumoniae.87 Similarly, Carling and coworkers reported a significant decrease in nosocomial infections caused by resistant Enterobacteriaceae following implementation of a multidisciplinary antibiotic stewardship program to minimize inappropriate use of third‐generation cephalosporins (Figure 3).32 More recently, a retrospective, longitudinal, multicenter analysis of a consortium of 22 academic health centers in the United States showed that incidence rates of carbapenem‐resistant P aeruginosa were lower at hospitals that restricted carbapenems than those that did not (P = .01).88

Figure 3

Evidence suggesting a beneficial impact of antimicrobial stewardship programs on resistance in gram‐positive organisms is limited. More specifically, the study by Carling and colleagues reported an apparent decrease in VRE rates following implementation of their program to reduce inappropriate use of third‐generation cephalosporins.32 The hospital had VRE rates similar to other NNIS System hospitals prior to beginning the program, but after antibiotic stewardship measures were implemented, the VRE rate began to drop, falling to 6% by 1999. This should be compared with a VRE rate of 24% for similar NNIS System hospitals in 1999.

As far as reducing healthcare costs, Figure 4A illustrates the direct antimicrobial cost savings at WFUBMC after implementation of the CAUSE antimicrobial stewardship program, and Figure 4B after supplemental interventions were implemented. Although decreasing antimicrobial cost is important, one would like to show decreases in overall healthcare costs associated with an antimicrobial stewardship program. Unfortunately, this is often a little more difficult to demonstrate. Extrapolations, however, may be possible. Because antimicrobial resistance, adverse drug effects, and secondary unintended infections such as C difficile colitis have been linked with increased patient morbidity and mortality, longer hospital stays, and increased healthcare costs,6, 89, 90 improved antimicrobial stewardship is expected to optimize patient care and lower overall healthcare costs. A study in a large tertiary care academic medical center estimated more than $4.25 million in total healthcare savings over 1 year with a stewardship program using both preauthorization and, to a lesser extent, prospective audit and feedback.91 Despite the fact cost saving should not be a primary goal of an antimicrobial stewardship program, lower antimicrobial costs can help keep a program funded and buttress a proposal for an antimicrobial stewardship program to hospital leadership.

Figure 4

Many hospitals recognize other areas where an antimicrobial stewardship program can demonstrate its value. This includes implementation of a rapid change in drug utilization during antimicrobial supply shortages and assistance with regulatory mandates and surgical infection prophylaxis. Stewardship teams often assist microbiology with protocols for microbiology reporting, antibiograms, planning for susceptibility testing, and communicating changes in microbiology tests or protocols to clinicians.

Conclusions

Overuse or misuse of antibiotics and other antimicrobials for hospital inpatients is relatively common, and can be associated with several unintended negative consequences. Improving medical care necessarily includes better use of antimicrobials to optimize outcomes and preserve the effectiveness of currently available agents. Further, an important additional consequence of effective antimicrobial stewardship and improved patient care is typically a lowering of overall healthcare costs. The recent 2007 IDSA/SHEA guidelines provide recommendations for developing an institutional program to enhance antimicrobial stewardship. However, individual institutions need to look closely at their own systems and patients to develop an antimicrobial stewardship program that best serves the needs of their hospital and the people it serves.

Nosocomial, or hospital‐acquired, infections (HAIs) are a major cause of patient morbidity and mortality in the United States and other countries.15 In 2002, approximately 1.7 million HAIs occurred in US hospitals and were associated with an estimated 98,987 deaths.1 Of particular note, increasing percentages of HAIs are now caused by antimicrobial‐resistant pathogens, which have been linked with increases in morbidity, mortality, length of hospital stay, and healthcare costs.6

The 2004 data summary from the United States National Nosocomial Infections Surveillance (NNIS) System Report highlighted substantial increases for year 2003 versus 1998 through 2002 in vancomycin‐resistant enterococci (VRE); methicillin‐resistant Staphylococcus aureus; Klebsiella pneumoniae resistant to third‐generation cephalosporins; and Pseudomonas aeruginosa resistant to imipenem, quinolones, or third‐generation cephalosporins.7 Other gram‐negative bacteria of concern include Escherichia coli and Acinetobacter baumannii, as well as Enterobacter cloacae and E. aerogenes.8, 9

The increasing number of multidrug‐resistant (MDR) gram‐negative bacteria within the healthcare setting is particularly concerning.1013 Too frequently, clinicians in the United States now encounter gram‐negative bacteria species that are resistant to many, and occasionally all, currently available antibiotics. For many of these MDR gram‐negative pathogens the antimicrobials that potentially remain active (eg, aminoglycosides and polymyxins) are often more toxic and less efficacious for some infections.14 Particularly problematic is that the pharmaceutical industry's developmental pipeline for new antibiotics, with novel mechanisms of action that might be used against MDR gram‐negative pathogens, has virtually come to a standstill.15, 16 Even if an investigational drug was in phase 2 or 3 trials right now or entered the US Food and Drug Administration (FDA) Fast Track Development Program, it would be at least 10 or 15 years before that drug would be available on the US market.

What this means is that the clinician's current antibiotic armamentarium is all they can expect in the foreseeable future. It also means that special care needs to be taken to optimally use currently available agents to ensure continued activity against the pathogens encountered in the hospital (and community) setting, now and in the future. Maximizing clinical outcomes, while minimizing the emergence and spread of antimicrobial resistance (and other adverse effects associated with suboptimal antimicrobial drug use), falls under the purview of antimicrobial stewardship, the focus of this paper.

Antimicrobial StewardshipWhy Is It Needed, What Is It, and What Are Its Goals?

Constructing an Antimicrobial Stewardship Program

A Case Study: the Wake Forest University Baptist Medical Center (WFUBMC) Program

Figure 1A,B provides an overview of the general structure of the antimicrobial stewardship program at WFUBMC implemented in 2000. To establish and provide the information needed for day‐to‐day operations of a stewardship program at WFUBMC, data are needed on how, where, and by whom antibiotics are used within the institution. In addition, microbiology data, including the frequency and susceptibility of infecting pathogens, are essential. Obtaining these data often requires the help and cooperation of the information technology (IT) staff at the institution. Considerable time and effort may be required at the outset, but once information system programs are established, ongoing data mining is much easier. At the time of program initiation, it was decided to assess antibiotic use or density (amount of drug per inpatient geographic unit) using the defined‐daily‐dose (DDD) method. This entails assigning a predetermined weight of administered antibiotic as a dose and dividing by a denominator of 1000 patient days. Subsequently, days of therapy per drug has been found to be a more accurate measure of antimicrobial consumption. When developing a program, it is vital to first obtain baseline usage data. Such data should include, if possible, a detailed inventory of usage within different units of the hospital or for particular services, or sometimes even for a particular provider with a history of high antimicrobial usage. Ongoing measurement over time allows the impact of new stewardship interventions and guidelines to be measured, as well as identifying potential new problem areas in usage.

Figure 1

Good microbiology data are also essential to determine problem pathogens at the institution and where they are located. Such data are useful not only to define areas of resistance (potentially warranting changes in antimicrobial policies to alter selection pressures), but also for gathering information necessary for defining local guidelines for antimicrobial use. For example, if the local antibiograms show that a particular pathogen in the hospital ICU has a particular resistance pattern, then initial empiric therapy for patients at risk of infection with those organisms should be chosen to cover the problematic resistant pathogen. Once subsequent microbiology data become available, patients not infected with the pathogen can be de‐escalated to a more narrow‐spectrum antibiotic.

As illustrated in Figure 1A, at WFUBMC, all the collected data are integrated to provide information concerning antibiotic density, usage, and patterns of antimicrobial resistance. This information is received by the staff of the antimicrobial stewardship program, which at WFUBMC is called the Center for Antimicrobial Utilization, Stewardship and Epidemiology (CAUSE). The CAUSE staff works with the day‐to‐day elements of program administration and operations and includes 2 infectious diseases physicians and 2 infectious disease PharmDs. The CAUSE staff works very closely with the microbiology laboratory, hospital pharmacy, and the medical director of hospital epidemiology and infection control.

The CAUSE program at WFUBMC administratively functions through an advisory board committee that includes thought leaders from different medical specialties and patient units of the hospitalparticularly those with high antimicrobial usage, such as hematology/oncology, pulmonary, critical care, and transplantation. The CAUSE staff and advisory board exist to exchange ideas concerning what is working or not working and where problem areas may be, and to propose possible changes in antimicrobial practices at the institution. In addition, thought leaders on the advisory board also receive and evaluate information from various sources about new antimicrobial agents and national guidelines, and, in turn, help disseminate this information to the hospital personnel who will be involved in program implementation.

At WFUBMC, it is the advisory board committee, working in conjunction with the CAUSE staff/medical director, that presents antimicrobials for formulary consideration to the Pharmacy and Therapeutics (P&T) Committee, in addition to any major interventions CAUSE and its advisory board feel are indicated. The P&T committee then reports to the medical staff executive committee and hospital leadership. As should be evident, the approach to stewardship at WFUBMC is medical staff‐driven, rather than a function of administrative constituents.

Finally, no matter how well‐organized an antimicrobial stewardship program is, it will not be fully successful if the entire medical staff does not buy into the process and agree with the need for the proposed changes and interventions involving the practice of antimicrobial therapy. It is important to spend some time early in program development to ensure that the need for an antimicrobial stewardship program, the process, and the outcomes (both in terms of patient care and clinical outcomes at the institution) are clearly communicated to the medical staff, and that their full commitment and cooperation are enlisted. In cases where hospital‐wide infection or resistance rates are known and antimicrobial utilization data are available, it is important to present such information in an understandable and convincing manner that makes the case for a proposed change or intervention, not only at the hospital level but also at the level of the patient.

Elements of a Successful Program: Basic Strategies

Potential strategies or elements of an antimicrobial stewardship program are listed in Table 2. Two evidence‐based fundamental or core strategies have been recommended by the IDSA/SHEA guidelines24 and implemented at numerous institutions with various levels of success. The first is a so‐called back‐end approach to modifying antimicrobial therapy on the basis of prospective audit of antimicrobial use, with intervention and feedback to the provider. The second is a front‐end approach using formulary restriction and preauthorization requirements for specific antimicrobial agents. Various supplemental strategies, including large group and patient case‐based education, guidelines and clinical pathways, antimicrobial order forms, and computerized clinical decision support, are also recommended.

Prospective Audit With Intervention and Feedback

This approach usually involves the use of an antimicrobial support team that reviews initial or ongoing therapy and then intervenes to provide feedback and suggested modifications to the medical care provider to improve therapy. This can be done by an infectious diseases physician, a clinical pharmacist, or a hospitalist or internist with expertise in antimicrobial therapy. The aim is to provide patient‐specific education and/or suggest changes to antimicrobial utilization (when needed) to improve and streamline therapy. Suggested interventions could include discontinuing or changing 1 or more drugs, switching intravenous to oral drug administration, and suggesting a short‐course duration of therapy. Occasionally, suggestions are made when appropriate to actually escalate or intensify therapy to increase efficacy. Identification of patients for targeting or focusing prospective audit and feedback efforts typically involves using computer surveillance to single out problem antimicrobials or problematic usage, given local resistance patterns or patient characteristics.24 Examples could include a focus on asymptomatic bacteruria, excessive duration of therapy for ventilator‐associated pneumonia, or overzealous use of certain classes of antimicrobials. Another potential activity for a prospective audit and feedback team is to review reports of patient‐specific blood and sterile body fluid culture results matched to the patient's current antimicrobial therapy. This allows for daily review of the appropriateness of therapy for potentially serious infections. Some patients seen by the antibiotic support team may be referred for infectious diseases or other expert consultative opinion if their infections or therapy are felt to be too complicated for routine prospective audit and feedback recommendations.

A number of studies have demonstrated that strategies involving prospective audit with intervention and feedback can improve antibiotic stewardship, as measured by reductions in inappropriate antibiotic use,30 reduced antibiotic costs or overall consumption,3135 greater compliance with hospital treatment guidelines or policies33, 36, 37 and, in some cases, reduced number of infections due to C difficile infection32, 37 or resistant pathogens.31, 32, 37 Prospective audit with feedback is probably the best and most effective core strategy for a community hospital program where other interventions are cumbersome or not as well tolerated by the medical staff. One potential disadvantage of the prospective audit with intervention and feedback approach is that medical provider adherence is largely voluntary. The team can make suggestions, but if the provider disagrees or is unobtainable, the suggestion is never implemented. Also, this strategy can also be resource‐intensive from a personnel perspective.

Evaluating Antimicrobial Stewardship Programs

Two general types of measures are used to evaluate the effectiveness of antimicrobial stewardship: process and outcome. As with most things done in the hospital, process measures are easier. They measure surrogate impacts of a program, accountability, resource use, and cost effectiveness. In essence, process measures evaluate whether the program accomplished what it set out to do in terms of changing certain processes or prescriber behaviors. It is important to measure resource use, as this helps to continue funding and to keep workers involved in the project. Good programs will save money; this can easily be measured, even if it is just as simple as going to the hospital pharmacy and looking at the cost of antimicrobials provided per patient day.

Outcomes like decreases in particular infections, less emergence of antimicrobial resistance, or other patient‐specific measures are likely more important in the big picture, but they are also much more difficult to measure. For example, where one would like to measure changes in pathogen resistance after making some changes in antimicrobial stewardship, it often takes years before the benefits of a particular intervention or change materialize in terms of less resistance or reduced emergence of resistance. If that type of change is to be measured, then one needs to be persistent and continue measurements over a long period of time. In addition, given the protracted amount of time before these outcomes may be observed, a number of other changes are likely to happen that coincide with the antimicrobial stewardship interventions and make assessment of causality difficult and biased.

Having said that, a number of studies have demonstrated a relationship between antibiotic restriction48, 8285 or other antimicrobial stewardship policies32, 86 and decreases in nosocomial C difficile infections or disease. Figure 2 illustrates the impact of a nonrestrictive antimicrobial stewardship program at a secondary/tertiary‐care hospital in Quebec, Canada, on an epidemic of C difficileassociated disease (CDAD) that occurred at the institution during the latter portion of 2003.86 Following program implementation, and the major drop in targeted antibiotic consumption, the incidence of CDAD also significantly decreased. Earlier implementation of infection control measures had no effect on CDAD incidence.

Figure 2

A smaller number of studies have reported decreases in resistant gram‐negative bacteria following implementation of antimicrobial stewardship programs. For example, Meyer and colleagues reported a marked reduction in ceftazidime‐resistant K pneumoniae at a 487‐bed general hospital in New York City after implementation of enhanced ceftazidime restriction and barrier precautions following an outbreak of infections caused by the resistant K pneumoniae.87 Similarly, Carling and coworkers reported a significant decrease in nosocomial infections caused by resistant Enterobacteriaceae following implementation of a multidisciplinary antibiotic stewardship program to minimize inappropriate use of third‐generation cephalosporins (Figure 3).32 More recently, a retrospective, longitudinal, multicenter analysis of a consortium of 22 academic health centers in the United States showed that incidence rates of carbapenem‐resistant P aeruginosa were lower at hospitals that restricted carbapenems than those that did not (P = .01).88

Figure 3

Evidence suggesting a beneficial impact of antimicrobial stewardship programs on resistance in gram‐positive organisms is limited. More specifically, the study by Carling and colleagues reported an apparent decrease in VRE rates following implementation of their program to reduce inappropriate use of third‐generation cephalosporins.32 The hospital had VRE rates similar to other NNIS System hospitals prior to beginning the program, but after antibiotic stewardship measures were implemented, the VRE rate began to drop, falling to 6% by 1999. This should be compared with a VRE rate of 24% for similar NNIS System hospitals in 1999.

As far as reducing healthcare costs, Figure 4A illustrates the direct antimicrobial cost savings at WFUBMC after implementation of the CAUSE antimicrobial stewardship program, and Figure 4B after supplemental interventions were implemented. Although decreasing antimicrobial cost is important, one would like to show decreases in overall healthcare costs associated with an antimicrobial stewardship program. Unfortunately, this is often a little more difficult to demonstrate. Extrapolations, however, may be possible. Because antimicrobial resistance, adverse drug effects, and secondary unintended infections such as C difficile colitis have been linked with increased patient morbidity and mortality, longer hospital stays, and increased healthcare costs,6, 89, 90 improved antimicrobial stewardship is expected to optimize patient care and lower overall healthcare costs. A study in a large tertiary care academic medical center estimated more than $4.25 million in total healthcare savings over 1 year with a stewardship program using both preauthorization and, to a lesser extent, prospective audit and feedback.91 Despite the fact cost saving should not be a primary goal of an antimicrobial stewardship program, lower antimicrobial costs can help keep a program funded and buttress a proposal for an antimicrobial stewardship program to hospital leadership.

Figure 4

Many hospitals recognize other areas where an antimicrobial stewardship program can demonstrate its value. This includes implementation of a rapid change in drug utilization during antimicrobial supply shortages and assistance with regulatory mandates and surgical infection prophylaxis. Stewardship teams often assist microbiology with protocols for microbiology reporting, antibiograms, planning for susceptibility testing, and communicating changes in microbiology tests or protocols to clinicians.

Conclusions

Overuse or misuse of antibiotics and other antimicrobials for hospital inpatients is relatively common, and can be associated with several unintended negative consequences. Improving medical care necessarily includes better use of antimicrobials to optimize outcomes and preserve the effectiveness of currently available agents. Further, an important additional consequence of effective antimicrobial stewardship and improved patient care is typically a lowering of overall healthcare costs. The recent 2007 IDSA/SHEA guidelines provide recommendations for developing an institutional program to enhance antimicrobial stewardship. However, individual institutions need to look closely at their own systems and patients to develop an antimicrobial stewardship program that best serves the needs of their hospital and the people it serves.

Collaboration between members of a multidisciplinary team is a key component of an effective institutional antimicrobial stewardship program, which itself is a key component of optimizing the care of hospitalized patients being treated with antimicrobial agents for proven or suspected infectious diseases. However, patient care does not and should not end once the patient is discharged from the hospital. In fact, high‐quality, value‐based health care across the full range of a medical condition depends on planning for optimization of care within the hospital as well as transitions of care to the outpatient setting. This extended care plan includes collaboration with multiple members of the health care community, both inside and outside the institution. The current review examines 3 aspects of patient care across the full cycle of an infectious disease condition: (1) value‐based health care, (2) stewardship of antimicrobials, and (3) community‐based parenteral anti‐infective therapy (CoPAT) as a model for antimicrobial stewardship outside the institutional setting.

Value‐Based Health Care

Patients first want to know that the health care professionals treating them actually care about them as individuals, and only then are patients concerned about how much the medical team knows. Patient‐centered care is a critical component of value‐based health care, a term that was bandied about quite a bit during the recent and ongoing health care debate in the United States. But what exactly does it mean? First, value in health care is defined by health care outcomes as a function of or divided by the cost of delivery of care. As Dr. Michael Porter and Dr. Elizabeth Olmsted Teisberg delineated in their 2007 article in the Journal of the American Medical Association,2 as well as in the 2006 book Redefining Health Care,3 The purpose of the healthcare system is not to minimize costs but to deliver value to patients, that is, better health per dollar spent. As they discuss value, it is a patient‐centric measure, and is focused on individual patient (not just diagnosis‐related group) outcomes and the cost of care across the full cycle. In this way of looking at things, an episode of care goes beyond the treatment provided during the acute admission to also include the transition of care to the outpatient or posthospital setting.

The reforms proposed by Porter and Teisberg are best achieved when the participating health care institutions have developed an information technology platform able to integrate and fully measure care across the full cycle of a medical condition. Furthermore, there is strong evidence that patient value increases with physician and team experience and volume for a particular condition.2 High volumes tend to correlate with the development of better information technology, as well as the formation of dedicated teams with tailored facilities, and with a greater capacity for constructive feedback to improve patient outcomes. The more experience a physician and team have with the management of a particular medical condition, the greater is the opportunity to learn and refine practices to provide greater value to the patient.

The Institute of Medicine has recommended that all healthcare professionals should be educated to deliver patient‐centered care as members of an interdisciplinary team emphasizing evidence‐based practice, quality improvement, and informatics.4 As has been demonstrated for patients with congestive heart failure59 and other conditions,10, 11 outcomes improve when components of care are integrated (often by nurse‐directed teams), preparing for the transition of care from the hospital to the home.12 This concept is the basis for the community‐based parenteral anti‐infective therapy program (CoPAT) at the Cleveland Clinic as a model for antimicrobial stewardship for patients requiring parenteral antimicrobial therapy at the time of discharge from the inpatient setting.

Community‐Based Parenteral Anti‐Infective Therapy as a Systems‐Based Approach to Antimicrobial Stewardship

A systems‐based approach for antimicrobial stewardship, CoPAT has been in operation at the Cleveland Clinic, a 1200‐bed hospital in downtown Cleveland, Ohio, since November 1979. The experiences of the authors and their colleagues demonstrate it to be a value‐based proposition for the patient that uses an antimicrobial stewardship platform. Also known as outpatient parenteral antimicrobial therapy (OPAT), CoPAT refers to the practice of administering antimicrobial therapy in the home or other outpatient settings, first introduced by Rucker and Harrison in 1974 in the context of outpatient management of cystic fibrosis.16 In the United States, CoPAT is a common practice today, and the IDSA has created practice guidelines for it.17

In 1983, Rehm and Weinstein coauthored an article describing their experiences at the Cleveland Clinic, in which selected patients were trained for home‐based antimicrobial therapy.12 Figure 1 illustrates the astronomical growth that has occurred over the years at the Cleveland Clinic in the number of patients discharged from the acute care center undergoing CoPAT (Gordon, unpublished data). It is anticipated that this growth will continue and in large part reflects the complexity of patients being seen and the desire to reduce length of stay. Evaluating the quality of any medical care is difficult, but there are 3 general approaches to assessing or measuring the quality of medical care: assessing the structure of care, assessing processes of care, and assessing outcomes.18 The quality of the CoPAT program at the Cleveland Clinic can be examined in the context of these 3 areas of assessment.

Figure 1

Settings or the Structure of Care

In a 1966 publication on quality of medical care evaluations, Donabedian described assessment of the structure of care as one of the primary approaches to measuring the quality of care.18 By structure, Donabedian meant the settings in which medical care takes place, including the adequacy of facilities and equipment, qualifications or expertise of medical staff and their organization, the administrative structure of the institution or institutional program of interest, and other administrative and related processes supporting and directing the delivery of care. Although the structure of care has the advantage of being concrete and relatively easy to assess, to be most meaningful, it ultimately needs to be related to the processes and outcomes of care.

With respect to the CoPAT program at the Cleveland Clinic main hospital, infectious diseases consultation is required for every patient being considered for discharge with parenteral antibiotics, whether the patient is going home or to another facility, including the clinic's own skilled nursing facility (SNF). Arrangements are then made for the delivery of antibiotics at home or in SNFs or long‐term acute care (LTAC) centers. The Cleveland Clinic CoPAT program does not use an outpatient infusion center.

The Cleveland Clinic uses a mandatory infectious diseases consultation for CoPAT because there are a number of important issues that need to be addressed before the patient is discharged, and for our system this is best accomplished by an infectious diseases specialist.12 For example, is antimicrobial therapy actually required in the first place? If it is, what is the optimal type, route, and duration of therapy? Are there other medical issues that need to be addressed? Decisions also need to be made about optimal vascular access and antimicrobial selection and administration, as well as arrangements being made for monitoring clinical and laboratory aspects. It is important that there is a smooth transition of care and prescheduled follow‐up in the outpatient clinic. The identification and use of an infectious diseases clinician directing the process leads to accountability. Notably, mandatory infectious disease consultation for outpatient parenteral antibiotic therapy has been used at Baystate Medical Center with improvement in reducing costs.19

The Process of Care

Assessments of the process of care involve examination of the particulars of medical care delivery, or whether what is recognized or accepted as good medical care has been applied. As discussed by Donabedian, process of care deals with issues such as the appropriateness and completeness of information obtained through clinical history, physical examination, and diagnostic tests; justification of diagnosis and therapy; technical competence in the performance of diagnostic and therapeutic procedures; and coordination and continuity of care.18

The CoPAT initiation process at the Cleveland Clinic is illustrated in Figure 2. It is a bundled process. As already mentioned, an infectious diseases consultation and evaluation is scheduled for all patients considered for CoPAT, after which a CoPAT form is completed and a follow‐up appointment made before the patient is discharged. In addition, the vascular access team is consulted and an appropriate vascular access device is placed in the patient prior to discharge. Likewise, a case manager is enlisted to identify a health care agency or SNF for patient placement or to determine whether the patient will receive home treatment. Once the appropriate setting is identified, the case manager transmits a completed CoPAT form to the health care agency or SNF, while forwarding a copy to the CoPAT nurse coordinator in the infectious disease department.

Figure 2

An electronic health record system is used at the Cleveland Clinic to provide real‐time information relevant for patient management. In 2007, a structured data form for CoPAT start‐of‐care was created within the Cleveland Clinic hospital electronic health record (EHR). This form contains a number of elements relevant for setting up patients for transition to CoPAT. In particular, the electronic CoPAT form contains information about the infection(s) and microorganism(s) being treated, intravenous antibiotic(s) prescribed (including treatment stop date), concurrent oral antibiotics, premedication recommendations (if appropriate), and recommended monitoring of laboratory tests. In addition, the form contains the telephone and fax numbers of the CoPAT coordinator and the name of the responsible physician, including a scheduled appointment for follow‐up (Fig. 3). The staff physician is responsible for completing the electronic CoPAT form or prescription. This CoPAT prescription then becomes part of the patient's electronic record and is transmissible and viewable by anyone with access to the EHR. This is important in terms of follow‐up and care accountability: an infectious disease staff clinician is identified as the contact person for clinical issues when a patient is on CoPAT.

Figure 3

After the patient is discharged, the CoPAT coordinator in the infectious disease department becomes responsible, together with the clinic's outpatient pharmacy, for reviewing laboratory results and notifying clinicians of potential problems that need to be addressed. These issues can pertain to laboratory findings, vascular access, or new symptoms or signs observed by the home nurse or patient. All this information is communicated via electronic health record messaging and/or through direct calls to the physician, when needed.

The CoPAT program has been widely accepted by internal customers of the Cleveland Clinic, which include hospitalists. This is probably because there is autonomy and accountability with the infectious diseases staff, the program or team is available 7 days per week, and the EHR facilitates communication. In addition, the use of infectious disease‐specific subspecialty groups (eg, bone marrow and solid‐organ transplant, bone and joint, and infective endocarditis groups) increases clinical credibility, as well as value received by patients of the clinic. Furthermore, the electronic CoPAT script facilitates discharge planning. CoPATs constitute approximately 25% of all ID consultation requests at the Cleveland Clinic and help to justify the 20 clinical ID clinical FTEs.

Outcomes of Medical Care

Assessment of medical care outcomes is another frequently used approach for measuring the quality of medical care.18 Medical care outcomes that have been examined as measures of quality of care include survival, number of hospital readmissions, time between discharge and readmissions, length of initial hospital stay and subsequent readmissions, quality of life, and health care costs. As has often been said, If you cannot measure it, you cannot manage it. The CoPAT program using the EHR has facilitated retrieval of structured reports in a format that provides clinicians with real‐time data enabling assessment of outcomes. By examining this data, the CoPAT team is in a better position to contemplate potential interventions for improving outpatient care and the value patients receive.

A 36‐month review of Cleveland Clinic CoPAT patient demographics from July 2007 to June 2010 demonstrated 6287 patients (56% male) had been prescribed 9471 courses of CoPAT (Gordon, unpublished data). Seventy‐nine percent of the patients were white, 16% African American, and 5% of other races. Most patients received 1 antibiotic per CoPAT course (79.1%), whereas 18.2%, 2.5%, and 0.2% received 2, 3, and 4 antibiotics per CoPAT course, respectively. Figure 4 highlights CoPAT distribution by source for anatomic site of infection. Bone and joint infections were the most common diagnoses associated with CoPAT at the Cleveland Clinic, followed by abdominal, cardiovascular, primary disseminated disease (eg, catheter‐associated bloodstream infections), and skin and soft‐tissue infection.

Figure 4

Figure 5 highlights the top‐10 pathogenic microorganisms in patients being discharged from the Cleveland Clinic with CoPAT, and the top‐10 antimicrobials prescribed for these patients. As can be seen, Staphylococcus aureus (methicillin susceptible and methicillin resistant) was the number one pathogen identified for patients undergoing CoPAT, followed by coagulase‐negative Staphylococcus and Enterococcus species. The most commonly identified gram‐negative bacteria among discharged patients was Pseudomonas aeruginosa. Only 2 of the top 10 pathogens were nonbacterial: Candida species and cytomegalovirus (CMV), the latter being the result of the high volume of transplantations performed at the clinic. With respect to the intravenous antimicrobials prescribed for patients undergoing CoPAT, the most commonly prescribed agent was vancomycin, followed by piperacillin/tazobactam. Of the 10 agents, only micafungin and ganciclovir were not antibacterial agents, indicating that the vast majority of patients discharged from the Cleveland Clinic with CoPAT had had bacterial, rather than fungal or viral, infections.

Figure 5

Of particular note, data collected from July 2007 through December 2008 demonstrated that more than 80% of patients discharged from the hospital with CoPAT did so with a prescheduled follow‐up visit. This patient‐centric measure is important because patients may not follow through with establishing appointments for follow‐up visits once discharge has already occurred. The Cleveland Clinic prides itself on making sure that a follow‐up appointment is actually made before the time of discharge for the vast majority of patients. The process also facilitates continuity of care with a specific infectious disease physician.

The various outcomes data collected by the Cleveland Clinic CoPAT Registry puts it in the position of being able to use the data to identify areas for improvement. Some of the projects made possible by the CoPAT Registry include analysis of: (1) outcomes of CoPAT in patients with bone and joint infections, (2) intensity of care in patients with cardiac and cardiac device infections while undergoing CoPAT, (3) C. difficile infections in patients undergoing CoPAT, and (4) emergency department (ED) visits or unanticipated readmissions in patients undergoing CoPAT. With respect to the last point, a 2009 article by Jencks and colleagues reported that 19.6% of the approximately 12 million Medicare beneficiaries who had been discharged from a hospital were rehospitalized within 30 days.20 Moreover, more than a third (34%) were rehospitalized within 90 days of discharge. It was estimated that no more than 10% of these readmissions were scheduled. More than 50% of patients with a medical condition who were rehospitalized within 30 days of discharge had not been billed for a physician visit between the time of discharge and hospitalization.20 This suggests that scheduling a follow‐up visit at the time of discharge might have reduced the need for many of these rehospitalizations. Unplanned rehospitalizations among the Medicare patients examined were not only relatively common but were also costly, resulting in an estimated $17.4 billion in additional Medicare costs.20 A New York Times editorial accompanying publication of the Jencks article noted that rehospitalizations and accompanying costs might be reduced by better discharge planning and closer cooperation between hospitals and physicians to ensure follow‐up care.21

At the Cleveland Clinic, data have recently been collected on the reasons for ED visits or hospital readmissions for patients receiving CoPAT at home through the Cleveland Clinic home care program. As illustrated in Figure 6, 24% of ED visits22 and 41% of hospital readmissions (Gordon, unpublished data) were for the infection being treated. Vascular access complications accounted for 23% of ED visits but only 2% of hospital readmissions. Nearly 50% of ED visits and 60% of hospital readmissions were for a reason unrelated to the infection being treated or CoPAT. It is hoped that closer examination of the data and perhaps additional analyses will suggest interventions to further reduce preventable readmissions or ED visits among patients discharged from the Cleveland Clinic on CoPAT.

Figure 6

Conclusions

Attention to antimicrobial stewardship and patient care should not end once the patient is discharged from the hospital or other institutional setting. Patients expect and should receive value‐based health care across the full cycle of their medical condition, and it is the responsibility of those caring for them to prepare for and provide such care during as well as after hospital discharge. The CoPAT program at the Cleveland Clinic provides a model for the extension of antimicrobial stewardship into the outpatient setting. The effectiveness of the program depends on a patient‐centric approach involving coordination and use of the expertise of multiple members of a team dedicated to patient value and facilitated by hospital‐based EHRs specialized for optimizing the transition of care into the outpatient setting for all patients scheduled to receive CoPAT. The quality of medical care provided by the Cleveland Clinic or other hospitals can be accessed through measurements of the structure, processes, and outcomes of care provided by the respective institutions. The data obtained can then be used to further refine care to optimize outcomes and provide high value for the patients treated at the institution. Achieving and then maintaining high‐quality medical care that provides value to patients is an ongoing process that should never be taken for granted.