Originally championed in the form of large multibed wards by Florence Nightingale in the nineteenth century, multioccupancy hospital rooms have recently been criticized for concerns about their cost, safety, lack of privacy, and unpopularity among patients.13 Specifically, they have been linked to longer hospital stays and the morbidity that those stays produceinfections, falls, and medical errors.13 In 2006, the Health Guidelines Revision Committee, the body that establishes guidelines for the construction of healthcare facilities in the United States, moved to a position recommending private rooms as the minimum standard for medical/surgical and postpartum hospital beds. The evidence supporting their position, however, has remained limited. Given the substantial cost and effort involved in converting hospital rooms, it is important to understand the impact of such policies on patients.

In a recent literature review by van de Glind et al.,4 the authors were able to identify only 4 randomized controlled trials comparing private and shared rooms. While they found that private rooms had a moderately positive impact on patient satisfaction and privacy, data on infection control was mixed, and data on patient safety was lacking.48 Though suggested, the association between private rooms and shortened hospital stays has not clearly been shown.13, 9 Certain advantages of shared rooms, such as decreased loneliness and increased patient social interaction, have also been identified.10 In addition, specialized multioccupancy wards for dementia and delirium have been demonstrated to be useful in the management of inpatients with those conditions, partly because of increased nursing presence.11 Finally, an additional theoretical benefit of shared rooms is the possibility of assistance between roommates in emergency situations.

For the sake of providing high‐quality patient‐centered care, patient concerns and preferences are also important to consider when establishing health policy regarding rooms. Previously, patient surveys and interviews have been helpful in identifying issues of concern to patients and understanding their preferences. Kirk12 identified privacy, quiet, improved sleep, and ability to have a family member stay in the room as reasons why 18 of 24 hospice patients preferred private hospice rooms.12 Conversely, in a 2002 Welsh survey of palliative patients, 68% wanted to be in an open area (4‐bed bay), identifying companionship as the major reason.13 Most recently, palliative care patients in the United Kingdom expressed a preference for shared rooms while well enough to interact with others but preferred private rooms when very ill or dying.14 These findings, given the lack of solid medical justification for 1 room type, raise multiple concerns about whether universal adherence to 1 room type will lead to the best patient care in the United States, and more specifically the veteran population. First, nearly all of the preference data described has been collected outside of the United States and veteran's populations. Second, studies have targeted highly‐specialized patient populations, and their applicability in a general medical/surgical population is unknown.10, 1217 Finally, the diversity of preferences exhibited in these studies underscores the difficulty in applying preference data to outside populations, such as the veteran population.

With 7.8 million patients enrolled in its healthcare system and operating 153 medical centers, the U.S. Department of Veterans Affairs has a large interest in the ongoing debate about shared vs. private hospital rooms. Because veterans share a common military background, they may place a higher value on companionship relative to privacy, and so they are an important group to study separately from the general U.S. population. To date, we have been unable to identify any studies that have looked at veterans' experiences and preferences in regards to room type. Through an anonymous survey, we sought to learn about these experiences and preferences and assess whether previously found advantages of each room type (eg, companionship in shared rooms, privacy in private rooms) could be confirmed in the veteran population. By doing so, we also aimed to better inform public policy about a general continued role for shared hospital rooms in the United States.

Methods

Our study used a point prevalence survey to examine patient preferences and experiences with shared (2‐person or 3‐person occupancy) vs. private rooms.

Results

A total of 162 surveys were completed with a participation rate of approximately 73%. Eighty three patients (51%) reported a shared room stay, while 70 patients (43%) reported a private room stay (17 of which were isolation). Nine did not report room type (Figure 1). Excluding isolation patients and those respondents who did not indicate room type, 5% of respondents reported age less than 45 years, 56% between 45 and 65 years, and 39% greater than 65 years. Sixty percent of patients identified themselves as African‐American, 28% as white, and 5% as other (7% did not answer). The majority (56%) reported having some formal higher education after high school. Median total household income was between $11,000 and $20,000.

Figure 1

Discussion

In this observational study of U.S. veterans, we found that patients in private rooms were more satisfied with privacy than their shared‐room counterparts, a finding consistent with previous studies.4, 5 We also found that private room patients were more likely to feel that nursing was available, a finding that had not previously been reported. Although increased loneliness and fear of death had been noted in private room patients before, our study did not support such findings.10 We found a strong overall preference for private rooms, adding to previously mixed evidence about patient room preferences.10, 1217 Although we had hypothesized a substantial demand for shared rooms among veterans based upon their common military experience, we found the demand modest at best.

The increased satisfaction with privacy among private room patients was expected.4, 5 That nearly one‐half of the patients in shared rooms found the privacy to be inadequate, however, was a strong finding that points to a fundamental problem with these rooms. That private room patient also felt more that nurses were available was surprising, since the higher number of patients in shared rooms might have been expected to generate a more visible nursing presence. Whether this finding is reproducible and whether it is based upon an objective difference in nursing behavior is something that should be studied further. It may simply have been a marker of increased satisfaction among the private room patients.

Although loneliness and increased fear of death had previously been identified as possible drawbacks of private rooms, our survey found no such suggestion.10 This may have been because our survey instrument simply asked patients to answer yes or no whether they had experienced the particular feelings; the previous study had used interviews. A more sensitive survey instrument would be needed to investigate questions of loneliness and fear of death further.

Except for the problem of privacy, the shared room experience generally appeared to be a positive one for patients, particularly in the exchange of conversation and help. A genuine value of these interactions to the patients was suggested by the finding that patients in shared rooms were more likely to prefer shared rooms than patients in private rooms. It is possible that once in them, patients found shared rooms to be better than expected, or that they warmed to them as a matter of resolving cognitive dissonance between their preconceptions and their placement. Although peace and quiet was a frequently cited reason for preferring private rooms, relatively few patients reported being bothered by their roommates. The high rate of assistance among roommates was perhaps the most surprising positive about the shared room experience, even if patients recalled helping others more than they remembered being helped themselves.

Preference for private rooms was greater than that found in previous studies, and appears to have been based largely upon concerns for privacy.1217 We suspect it may reflect changing societal values that have placed an increased priority on personal privacy; this is suggested by the greater willingness of elderly patients to stay in shared rooms. Whether the preference for private rooms is even higher in the general nonveteran population is something that should be assessed with future studies.

Our study was limited by a number of factors. While room assignments were based on bed availability, it is possible that patients requested particular rooms or were moved due to other factors. By selecting a study population of male veterans of the armed services, we limited the extent to which our results can be generalized. However, this particular population is an important one to study due to the size of the VA healthcare system, the prevalence of shared rooms in VA hospitals, and the unique cultural values of veterans. We were unable to obtain information from approximately 27% of our target population because they were unable or unwilling to participate, thus excluding a population subset from analysis and potentially biasing our results. Although a high prevalence of preference for private rooms was seen, the strength of the preference was not assessed, and patients were unable to convey if they had no preference regarding room type. The survey instrument was designed for the purposes of this study and has not been validated. Some patients were assisted with filling out their surveys and may have responded more positively about their roommates if they thought their roommates could overhear their answers to questions. Because patient enrollment was slower than expected, fewer patients were enrolled than planned and the power achieved was insufficient for detecting differences of 20% or less between groups. This is unlikely to have affected our positive findings, but may have prevented us from detecting additional differences among the study groups.

Our data shows that for a population of veteran patients, private rooms are preferred, and that increased privacy is a primary advantage. It also demonstrates that multioccupancy rooms have several positive aspects. Given the need for sound public policy in regard to this issue, further research is needed to better evaluate objective outcomes of room type, such as fall rates, nosocomial infection rates, and lengths of hospitalization. In the meantime, efforts should be taken to address some of the known problems of shared rooms. For instance, establishing more substantial dividers between beds would be an intervention that could increase privacy, reduce noise, and minimize unwelcome smells. Additionally, given that patients have a wide variety of feelings toward multioccupancy rooms, incorporating patient choice into room assignments when logistically feasible is a step that could lead to increased patient satisfaction with hospitalization.

Originally championed in the form of large multibed wards by Florence Nightingale in the nineteenth century, multioccupancy hospital rooms have recently been criticized for concerns about their cost, safety, lack of privacy, and unpopularity among patients.13 Specifically, they have been linked to longer hospital stays and the morbidity that those stays produceinfections, falls, and medical errors.13 In 2006, the Health Guidelines Revision Committee, the body that establishes guidelines for the construction of healthcare facilities in the United States, moved to a position recommending private rooms as the minimum standard for medical/surgical and postpartum hospital beds. The evidence supporting their position, however, has remained limited. Given the substantial cost and effort involved in converting hospital rooms, it is important to understand the impact of such policies on patients.

In a recent literature review by van de Glind et al.,4 the authors were able to identify only 4 randomized controlled trials comparing private and shared rooms. While they found that private rooms had a moderately positive impact on patient satisfaction and privacy, data on infection control was mixed, and data on patient safety was lacking.48 Though suggested, the association between private rooms and shortened hospital stays has not clearly been shown.13, 9 Certain advantages of shared rooms, such as decreased loneliness and increased patient social interaction, have also been identified.10 In addition, specialized multioccupancy wards for dementia and delirium have been demonstrated to be useful in the management of inpatients with those conditions, partly because of increased nursing presence.11 Finally, an additional theoretical benefit of shared rooms is the possibility of assistance between roommates in emergency situations.

For the sake of providing high‐quality patient‐centered care, patient concerns and preferences are also important to consider when establishing health policy regarding rooms. Previously, patient surveys and interviews have been helpful in identifying issues of concern to patients and understanding their preferences. Kirk12 identified privacy, quiet, improved sleep, and ability to have a family member stay in the room as reasons why 18 of 24 hospice patients preferred private hospice rooms.12 Conversely, in a 2002 Welsh survey of palliative patients, 68% wanted to be in an open area (4‐bed bay), identifying companionship as the major reason.13 Most recently, palliative care patients in the United Kingdom expressed a preference for shared rooms while well enough to interact with others but preferred private rooms when very ill or dying.14 These findings, given the lack of solid medical justification for 1 room type, raise multiple concerns about whether universal adherence to 1 room type will lead to the best patient care in the United States, and more specifically the veteran population. First, nearly all of the preference data described has been collected outside of the United States and veteran's populations. Second, studies have targeted highly‐specialized patient populations, and their applicability in a general medical/surgical population is unknown.10, 1217 Finally, the diversity of preferences exhibited in these studies underscores the difficulty in applying preference data to outside populations, such as the veteran population.

With 7.8 million patients enrolled in its healthcare system and operating 153 medical centers, the U.S. Department of Veterans Affairs has a large interest in the ongoing debate about shared vs. private hospital rooms. Because veterans share a common military background, they may place a higher value on companionship relative to privacy, and so they are an important group to study separately from the general U.S. population. To date, we have been unable to identify any studies that have looked at veterans' experiences and preferences in regards to room type. Through an anonymous survey, we sought to learn about these experiences and preferences and assess whether previously found advantages of each room type (eg, companionship in shared rooms, privacy in private rooms) could be confirmed in the veteran population. By doing so, we also aimed to better inform public policy about a general continued role for shared hospital rooms in the United States.

Methods

Our study used a point prevalence survey to examine patient preferences and experiences with shared (2‐person or 3‐person occupancy) vs. private rooms.

Results

A total of 162 surveys were completed with a participation rate of approximately 73%. Eighty three patients (51%) reported a shared room stay, while 70 patients (43%) reported a private room stay (17 of which were isolation). Nine did not report room type (Figure 1). Excluding isolation patients and those respondents who did not indicate room type, 5% of respondents reported age less than 45 years, 56% between 45 and 65 years, and 39% greater than 65 years. Sixty percent of patients identified themselves as African‐American, 28% as white, and 5% as other (7% did not answer). The majority (56%) reported having some formal higher education after high school. Median total household income was between $11,000 and $20,000.

Figure 1

Discussion

In this observational study of U.S. veterans, we found that patients in private rooms were more satisfied with privacy than their shared‐room counterparts, a finding consistent with previous studies.4, 5 We also found that private room patients were more likely to feel that nursing was available, a finding that had not previously been reported. Although increased loneliness and fear of death had been noted in private room patients before, our study did not support such findings.10 We found a strong overall preference for private rooms, adding to previously mixed evidence about patient room preferences.10, 1217 Although we had hypothesized a substantial demand for shared rooms among veterans based upon their common military experience, we found the demand modest at best.

The increased satisfaction with privacy among private room patients was expected.4, 5 That nearly one‐half of the patients in shared rooms found the privacy to be inadequate, however, was a strong finding that points to a fundamental problem with these rooms. That private room patient also felt more that nurses were available was surprising, since the higher number of patients in shared rooms might have been expected to generate a more visible nursing presence. Whether this finding is reproducible and whether it is based upon an objective difference in nursing behavior is something that should be studied further. It may simply have been a marker of increased satisfaction among the private room patients.

Although loneliness and increased fear of death had previously been identified as possible drawbacks of private rooms, our survey found no such suggestion.10 This may have been because our survey instrument simply asked patients to answer yes or no whether they had experienced the particular feelings; the previous study had used interviews. A more sensitive survey instrument would be needed to investigate questions of loneliness and fear of death further.

Except for the problem of privacy, the shared room experience generally appeared to be a positive one for patients, particularly in the exchange of conversation and help. A genuine value of these interactions to the patients was suggested by the finding that patients in shared rooms were more likely to prefer shared rooms than patients in private rooms. It is possible that once in them, patients found shared rooms to be better than expected, or that they warmed to them as a matter of resolving cognitive dissonance between their preconceptions and their placement. Although peace and quiet was a frequently cited reason for preferring private rooms, relatively few patients reported being bothered by their roommates. The high rate of assistance among roommates was perhaps the most surprising positive about the shared room experience, even if patients recalled helping others more than they remembered being helped themselves.

Preference for private rooms was greater than that found in previous studies, and appears to have been based largely upon concerns for privacy.1217 We suspect it may reflect changing societal values that have placed an increased priority on personal privacy; this is suggested by the greater willingness of elderly patients to stay in shared rooms. Whether the preference for private rooms is even higher in the general nonveteran population is something that should be assessed with future studies.

Our study was limited by a number of factors. While room assignments were based on bed availability, it is possible that patients requested particular rooms or were moved due to other factors. By selecting a study population of male veterans of the armed services, we limited the extent to which our results can be generalized. However, this particular population is an important one to study due to the size of the VA healthcare system, the prevalence of shared rooms in VA hospitals, and the unique cultural values of veterans. We were unable to obtain information from approximately 27% of our target population because they were unable or unwilling to participate, thus excluding a population subset from analysis and potentially biasing our results. Although a high prevalence of preference for private rooms was seen, the strength of the preference was not assessed, and patients were unable to convey if they had no preference regarding room type. The survey instrument was designed for the purposes of this study and has not been validated. Some patients were assisted with filling out their surveys and may have responded more positively about their roommates if they thought their roommates could overhear their answers to questions. Because patient enrollment was slower than expected, fewer patients were enrolled than planned and the power achieved was insufficient for detecting differences of 20% or less between groups. This is unlikely to have affected our positive findings, but may have prevented us from detecting additional differences among the study groups.

Our data shows that for a population of veteran patients, private rooms are preferred, and that increased privacy is a primary advantage. It also demonstrates that multioccupancy rooms have several positive aspects. Given the need for sound public policy in regard to this issue, further research is needed to better evaluate objective outcomes of room type, such as fall rates, nosocomial infection rates, and lengths of hospitalization. In the meantime, efforts should be taken to address some of the known problems of shared rooms. For instance, establishing more substantial dividers between beds would be an intervention that could increase privacy, reduce noise, and minimize unwelcome smells. Additionally, given that patients have a wide variety of feelings toward multioccupancy rooms, incorporating patient choice into room assignments when logistically feasible is a step that could lead to increased patient satisfaction with hospitalization.

Originally championed in the form of large multibed wards by Florence Nightingale in the nineteenth century, multioccupancy hospital rooms have recently been criticized for concerns about their cost, safety, lack of privacy, and unpopularity among patients.13 Specifically, they have been linked to longer hospital stays and the morbidity that those stays produceinfections, falls, and medical errors.13 In 2006, the Health Guidelines Revision Committee, the body that establishes guidelines for the construction of healthcare facilities in the United States, moved to a position recommending private rooms as the minimum standard for medical/surgical and postpartum hospital beds. The evidence supporting their position, however, has remained limited. Given the substantial cost and effort involved in converting hospital rooms, it is important to understand the impact of such policies on patients.

In a recent literature review by van de Glind et al.,4 the authors were able to identify only 4 randomized controlled trials comparing private and shared rooms. While they found that private rooms had a moderately positive impact on patient satisfaction and privacy, data on infection control was mixed, and data on patient safety was lacking.48 Though suggested, the association between private rooms and shortened hospital stays has not clearly been shown.13, 9 Certain advantages of shared rooms, such as decreased loneliness and increased patient social interaction, have also been identified.10 In addition, specialized multioccupancy wards for dementia and delirium have been demonstrated to be useful in the management of inpatients with those conditions, partly because of increased nursing presence.11 Finally, an additional theoretical benefit of shared rooms is the possibility of assistance between roommates in emergency situations.

For the sake of providing high‐quality patient‐centered care, patient concerns and preferences are also important to consider when establishing health policy regarding rooms. Previously, patient surveys and interviews have been helpful in identifying issues of concern to patients and understanding their preferences. Kirk12 identified privacy, quiet, improved sleep, and ability to have a family member stay in the room as reasons why 18 of 24 hospice patients preferred private hospice rooms.12 Conversely, in a 2002 Welsh survey of palliative patients, 68% wanted to be in an open area (4‐bed bay), identifying companionship as the major reason.13 Most recently, palliative care patients in the United Kingdom expressed a preference for shared rooms while well enough to interact with others but preferred private rooms when very ill or dying.14 These findings, given the lack of solid medical justification for 1 room type, raise multiple concerns about whether universal adherence to 1 room type will lead to the best patient care in the United States, and more specifically the veteran population. First, nearly all of the preference data described has been collected outside of the United States and veteran's populations. Second, studies have targeted highly‐specialized patient populations, and their applicability in a general medical/surgical population is unknown.10, 1217 Finally, the diversity of preferences exhibited in these studies underscores the difficulty in applying preference data to outside populations, such as the veteran population.

With 7.8 million patients enrolled in its healthcare system and operating 153 medical centers, the U.S. Department of Veterans Affairs has a large interest in the ongoing debate about shared vs. private hospital rooms. Because veterans share a common military background, they may place a higher value on companionship relative to privacy, and so they are an important group to study separately from the general U.S. population. To date, we have been unable to identify any studies that have looked at veterans' experiences and preferences in regards to room type. Through an anonymous survey, we sought to learn about these experiences and preferences and assess whether previously found advantages of each room type (eg, companionship in shared rooms, privacy in private rooms) could be confirmed in the veteran population. By doing so, we also aimed to better inform public policy about a general continued role for shared hospital rooms in the United States.

Methods

Our study used a point prevalence survey to examine patient preferences and experiences with shared (2‐person or 3‐person occupancy) vs. private rooms.

Results

A total of 162 surveys were completed with a participation rate of approximately 73%. Eighty three patients (51%) reported a shared room stay, while 70 patients (43%) reported a private room stay (17 of which were isolation). Nine did not report room type (Figure 1). Excluding isolation patients and those respondents who did not indicate room type, 5% of respondents reported age less than 45 years, 56% between 45 and 65 years, and 39% greater than 65 years. Sixty percent of patients identified themselves as African‐American, 28% as white, and 5% as other (7% did not answer). The majority (56%) reported having some formal higher education after high school. Median total household income was between $11,000 and $20,000.

Figure 1

Discussion

In this observational study of U.S. veterans, we found that patients in private rooms were more satisfied with privacy than their shared‐room counterparts, a finding consistent with previous studies.4, 5 We also found that private room patients were more likely to feel that nursing was available, a finding that had not previously been reported. Although increased loneliness and fear of death had been noted in private room patients before, our study did not support such findings.10 We found a strong overall preference for private rooms, adding to previously mixed evidence about patient room preferences.10, 1217 Although we had hypothesized a substantial demand for shared rooms among veterans based upon their common military experience, we found the demand modest at best.

The increased satisfaction with privacy among private room patients was expected.4, 5 That nearly one‐half of the patients in shared rooms found the privacy to be inadequate, however, was a strong finding that points to a fundamental problem with these rooms. That private room patient also felt more that nurses were available was surprising, since the higher number of patients in shared rooms might have been expected to generate a more visible nursing presence. Whether this finding is reproducible and whether it is based upon an objective difference in nursing behavior is something that should be studied further. It may simply have been a marker of increased satisfaction among the private room patients.

Although loneliness and increased fear of death had previously been identified as possible drawbacks of private rooms, our survey found no such suggestion.10 This may have been because our survey instrument simply asked patients to answer yes or no whether they had experienced the particular feelings; the previous study had used interviews. A more sensitive survey instrument would be needed to investigate questions of loneliness and fear of death further.

Except for the problem of privacy, the shared room experience generally appeared to be a positive one for patients, particularly in the exchange of conversation and help. A genuine value of these interactions to the patients was suggested by the finding that patients in shared rooms were more likely to prefer shared rooms than patients in private rooms. It is possible that once in them, patients found shared rooms to be better than expected, or that they warmed to them as a matter of resolving cognitive dissonance between their preconceptions and their placement. Although peace and quiet was a frequently cited reason for preferring private rooms, relatively few patients reported being bothered by their roommates. The high rate of assistance among roommates was perhaps the most surprising positive about the shared room experience, even if patients recalled helping others more than they remembered being helped themselves.

Preference for private rooms was greater than that found in previous studies, and appears to have been based largely upon concerns for privacy.1217 We suspect it may reflect changing societal values that have placed an increased priority on personal privacy; this is suggested by the greater willingness of elderly patients to stay in shared rooms. Whether the preference for private rooms is even higher in the general nonveteran population is something that should be assessed with future studies.

Our study was limited by a number of factors. While room assignments were based on bed availability, it is possible that patients requested particular rooms or were moved due to other factors. By selecting a study population of male veterans of the armed services, we limited the extent to which our results can be generalized. However, this particular population is an important one to study due to the size of the VA healthcare system, the prevalence of shared rooms in VA hospitals, and the unique cultural values of veterans. We were unable to obtain information from approximately 27% of our target population because they were unable or unwilling to participate, thus excluding a population subset from analysis and potentially biasing our results. Although a high prevalence of preference for private rooms was seen, the strength of the preference was not assessed, and patients were unable to convey if they had no preference regarding room type. The survey instrument was designed for the purposes of this study and has not been validated. Some patients were assisted with filling out their surveys and may have responded more positively about their roommates if they thought their roommates could overhear their answers to questions. Because patient enrollment was slower than expected, fewer patients were enrolled than planned and the power achieved was insufficient for detecting differences of 20% or less between groups. This is unlikely to have affected our positive findings, but may have prevented us from detecting additional differences among the study groups.

Our data shows that for a population of veteran patients, private rooms are preferred, and that increased privacy is a primary advantage. It also demonstrates that multioccupancy rooms have several positive aspects. Given the need for sound public policy in regard to this issue, further research is needed to better evaluate objective outcomes of room type, such as fall rates, nosocomial infection rates, and lengths of hospitalization. In the meantime, efforts should be taken to address some of the known problems of shared rooms. For instance, establishing more substantial dividers between beds would be an intervention that could increase privacy, reduce noise, and minimize unwelcome smells. Additionally, given that patients have a wide variety of feelings toward multioccupancy rooms, incorporating patient choice into room assignments when logistically feasible is a step that could lead to increased patient satisfaction with hospitalization.

Effective communication between healthcare providers is essential to patient safety and quality of care.1, 2 Numeric pagers are commonly used communication devices in healthcare, but cannot convey important information such as the reason for the page, urgency of the page, or sender name. Physicians must respond to numeric pages, often disrupting patient encounters or educational activities.36 In a study of medical interns, disruptions to patient care occurred with up to 65% of pages received, two‐thirds of which were not felt to be urgent.5 In addition to causing frustration, frequent disruptions can contribute to medical errors.7, 8

Alphanumeric pagers can display both numbers and text, and may address some of the communication problems associated with numeric pagers. They also lay the groundwork for other patient safety initiatives such as automated paging of critical laboratory values9 and real‐time reporting of user‐requested laboratory data.10 Implementation of alphanumeric paging on a general surgery teaching service reduced disruptions to patient care and the number of pages requiring a return call.11

Our primary aim was to implement an alphanumeric paging system. We will describe our implementation strategies and barriers identified. We evaluated the implementation of alphanumeric paging by measuring (1) the proportion of pages sent as text pages, (2) the source of the pages (other physicians or from the general medicine [GM] ward), (3) the content of the text pages, (4) the number of pages that disrupted scheduled education activities, and (5) satisfaction with the alphanumeric paging system.

Materials and Methods

Implementation Process

We provided alphanumeric pagers to the residents on the General Internal Medicine service in July 2006. Alphanumeric pagers were limited, so each resident traded their numeric pager for an alphanumeric pager at the start of each rotation. Once their rotation ended, they returned their alphanumeric pager for their original numeric pager. The communications department coordinated this process. The chief medical resident spent 10 minutes to teach the residents how to use the system at the beginning of the rotation. In August and September 2006, a member of the communications department trained the nurses on the 4 GM wards how to send a text page using the Internet‐based paging application. We scheduled these 15‐minute sessions throughout the day and evening in order to capture as many nurses as possible. We encouraged the nurses to include standardized information in the text message (eg, patient ID, issue, level of urgency, sender name, call‐back number).

We used rapid‐cycle change methods12 to implement the alphanumeric system (Figure 1). The first change cycle in August 2006 consisted of providing pagers to residents and training the nurses and physicians to send text pages. Users reported that the paging interface was difficult to use. For the second change cycle in September 2006, the communications department modified the paging interface to improve usability and created shortcut icons on the GM ward computers. While the system was easier to access, the nurses reported that 1‐time training was insufficient. For the next change cycle in September 2006, we developed Internet‐based tutorials that could be accessed at any time, and made expert users (charge nurses) available for just‐in‐time training. We asked these charge nurses what they believed would encourage adoption of the system. They suggested that contests worked well with other initiatives. For the final change cycle in October and November 2006, we held a contest and rewarded the GM ward that sent the highest percentage of text pages with a team lunch.

Figure 1

Our results from November 2006 were presented to our hospital medical leaders, who approved widespread implementation of alphanumeric pagers for all residents and medical students in all programs. The cost of this upgrade was approximately $35,000 per year to lease 500 alphanumeric pagers. By July 2007, all residents and students in our hospital had alphanumeric pagers.

Results

Paging Patterns Before, During, and After Implementation

The number of pages per resident was similar before and during implementation, but higher afterwards. (46 16 pages/resident/week in November 2005, 47 20 pages/resident/week in November 2006, and 59 27 pages/resident/week in January 2008; P = 0.17; Table 1). The mean number of admissions per night was 8.0 2.7 before implementation, compared to 10.2 3.5 after implementation (P = 0.009).

Table 1.Paging Patterns Before, During, and After Implementation of an Alphanumeric Paging System

	Before Implementation (all residents had numeric pagers)	During Implementation (all residents had alphanumeric pagers)		After Implementation (all residents had alphanumeric pagers)
Paging Characteristics	November 2005	October 2006	November 2006	January 2008	P Value
Abbreviations: Doc, physician; GM, general medicine; SD, standard deviation. * Student t test comparing November 2005 with January 2008. Fisher's Exact Test comparing November 2005 with January 2008.
Total number of pages	1431	1879	1813	1269
Total number of resident weeks worked	29	33	33	21
Pages per resident week, mean (SD)	46.2 (16.3)	57.9 (19.2)	46.5 (20.2)	59.0 (26.5)	0.17*
Number of patients admitted per night, mean (SD)	8.0 (2.7)	10.2 (3.5)	11.0 (2.9)	10.2 (3.5)	0.009*
Type of page, n (%)
Numeric pages	751 (53)	462 (25)	580 (32)	374 (30)	0.001
GM wards‐to‐Doc	584 (41)	393 (21)	538 (30)	352 (28)	0.001
Doc‐to‐Doc	167 (12)	69 (4)	42 (2)	22 (2)	0.001
Non‐GM ward/Doc pages	680 (47)	1107 (59)	809 (45)	487 (38)	0.001
Text pages	0 (0)	310 (16)	424 (23)	408 (32)	0.001
GM wards‐to‐Doc	0 (0)	175 (9)	221 (12)	129 (10)	0.001
Doc‐to‐Doc	0 (0)	135 (7)	203 (11)	279 (22)	0.001

We observed a significant and sustained increase in the use of text paging during the study (Table 1). After implementation, 32% of all pages sent to our residents were text messages (P 0.001). Physicians almost exclusively sent text pages by the end of implementation (increase from 0% to 83% text paging rate during implementation, and 93% after implementation; P 0.001; Figure 2). GM ward text paging rates also increased from 0% to 29% during implementation, and 27% after implementation (P 0.001; Figure 2). The alphanumeric paging system was used to a greater degree by physicians compared to other workers on the GM ward after full implementation (93% vs. 27%; P 0.001). We explored the proportion of GM ward‐to‐Doc pages sent as text from different GM wards during implementation, and found significant variation, ranging from 14% to 57% (P 0.001).

Figure 2

The most common reasons for text paging from GM wards were to request a patient assessment or for notification of a patient's clinical status (25%), to clarify written orders (20%), and to request a medication prescription (13%) (Table 2). Among physicians, the most common reasons for text paging were to set up meetings for work or teaching rounds (33%), to relay patient‐care related messages (27%), and to sign‐over patients at the end of the day (23%) (Table 3). The remainder of the other Doc‐to‐Doc pages (18%) were mostly personal messages or team communication that was not related to clinical work.

Table 2.Reasons for Paging the Physicians (GM ward‐to‐Doc)

Reason for Paging	Number (%)	Examples
NOTE: Represents data from November 2006. Abbreviations: BP, blood pressure; CXR, chest X‐ray; Doc, physician; GM, general medicine; INR, international normalized ratio; NG, nasogastric tube.
Requests for patient assessment or notification of a patient's clinical status	55 (25)	Patient X. Temp 38.5. No other symptoms. [Nurse's name].
		The repeat CXR on Patient X has been completed. Please call [ward] if NG can be used. Thank you.
Clarification of a written order	45 (20)	Would you like Patient X to get a second dose of Lasix? He has already had 100 mg and his output thus far is 500 cc. [Nurse's name]
		Patient X. BP = 100/62, pulse 54. Patient is supposed to have metoprolol 50 mg tonight. Do you want me to hold it? [Nurse's name]
Request for a medication prescription	28 (13)	Patient X needs an analgesic for pain in his arms and legs. Please call [ward].
		Patient X has a daily coumadin order. INR 2.14. Please call with dosage for 1800. Ask for [Nurse].
Cosigning written order	25 (11)	Please co‐sign neurology suggested orders for Patient X. [Nurse's name]
		Not urgent at your leisure, please co‐sign Patient X orders on [ward] for medical student. Thanks.
Notification of a recent laboratory result	23 (11)	Patient X's potassium is 3.0 today. Please call [ward].
		Patient X. Sodium 163. Troponin unchanged at 1.54. [ward].
Arranging meetings with patients and/or family	18 (8)	Meeting with Patient X's family and social worker at 2 pm tomorrow on [ward].
		[Social worker] is here. [Physiotherapist] expected any minute. We are going to meet in family room for Patient X.
Request to complete paperwork	15 (7)	Patient X is ready to go home, and just needs discharge orders and prescriptions
		Referral form for community palliative doctor on the front of the chart fill in where marked by arrows. Can you please put on form prognosis as well? [Social worker]
Other	12 (6)
Total	221 (100)

Table 3.Reasons for Paging Among Physicians (Doc‐to‐Doc)

Reason for Paging	Number (%)	Examples
NOTE: Represents data from November 2006. Abbreviations: cath, catheterization; CHF, congestive heart failure; Cr, creatinine; DKA, diabetic ketoacidosis; IV, intravenous; Lytes, electrolytes; NS, normal saline; US, ultrasound; WE, weakend.
Setting up meetings for work or teaching rounds	67 (33)	Confirmed diabetes teaching at 1400 hr in [room]. Please let medical students know. Thanks.
		Please come to [lecture theatre] if you can in the next 10 minutes for the teleconferenced noon rounds. Thanks.
We are in the Emergency Department with [attending staff]. Come to meet us here if you can.
Relaying patient‐care related messages	54 (27)	Patient X US query cholangitis, can we ask for a surgical consult. He may benefit from surgery or percutaneous drain. Please repeat his blood work.
		Patient X in [room X]. Presented with DKA. pH 7.2. pCO2 23, bicarbonate 9. Lytes pending. Got IV insulin and NS. Need to check to clinical stability. [Resident]
Signing‐over patients at the end of the day	46 (23)	I'm ready to sign out to you. Where are you? [Resident].
		Patient X is back from cardiac cath, and is stable. Please check Cr over WE, and watch for CHF.
Other	36 (18)
Total	203 (100)

Discussion

We successfully implemented an alphanumeric paging system on a resident inpatient internal medicine teaching service, with 42% of pages from our GM wards or physicians sent as text pages during implementation period. Six months after widespread use of alphanumeric pagers at our hospital (and 14 months after the initial implementation on the General Internal Medicine service) the text paging rate was 52%. Physicians have nearly universally adopted the use of alphanumeric paging as a means of communicating with one another, while the adoption by nursing staff was modest. The implementation of the alphanumeric paging system was associated with a significant reduction in disruptive pages.

We could identify only one prior study of alphanumeric paging implementation in a hospital setting. A general surgery teaching service in the United States demonstrated that 35% of all pages received by residents were text pages 3 months after implementation,11 similar to our result of 32% text paging rate after full implementation. We found a greater use of text paging among physicians in our study (93% of Doc‐to‐Doc pages were text), compared to 55% in this prior study. This difference may be partially explained by varied methods for identifying Doc‐to‐Doc pages between the studies.

A number of factors influence the adoption of new technology, such as the technology's features, end‐user characteristics, and dissemination strategies.14 During the implementation, even though the web‐based paging system was deemed easy to use, there was a lack of computers available to send text messages at each of our nursing stations. Residents were generally more familiar with technology and the use of computers for communication than nurses, and therefore more likely to use the technology. The presence of innovators influenced the success of adoption, evidenced by the fact that the GM ward that sent the highest percentage of pages as text was also the GM ward where the project leader (B.W.) worked as the attending staff during the implementation period.

Our study has several limitations. First, our method for classifying the source of numeric pages was imperfect. Our method systematically underestimates Doc‐to‐Doc numeric pages, because we assumed that all pages from GM ward phone numbers were not from physicians. We also may have misclassified the origin of some text messages. These limitations would not affect our conclusion that text messaging increased, but may overestimate the increase in Doc‐to‐Doc text messaging, and underestimate the increase in GM ward‐to‐Doc text messaging. The reasons for paging are not known for the numeric pages. The number of pages received per resident increased after implementation, so it is possible that alphanumeric pages increased calls for certain nonurgent issues, such as co‐signing orders. Finally, we assumed that residents were attending scheduled educational rounds, but were unable to confirm attendance, so we cannot be sure that disruptions actually were reduced.

In summary, we implemented an alphanumeric paging system, and observed a sustained use of text messaging after 1 year. The implementation of alphanumeric paging was associated with a reduction in disruptive pages sent during scheduled educational rounds.

Effective communication between healthcare providers is essential to patient safety and quality of care.1, 2 Numeric pagers are commonly used communication devices in healthcare, but cannot convey important information such as the reason for the page, urgency of the page, or sender name. Physicians must respond to numeric pages, often disrupting patient encounters or educational activities.36 In a study of medical interns, disruptions to patient care occurred with up to 65% of pages received, two‐thirds of which were not felt to be urgent.5 In addition to causing frustration, frequent disruptions can contribute to medical errors.7, 8

Alphanumeric pagers can display both numbers and text, and may address some of the communication problems associated with numeric pagers. They also lay the groundwork for other patient safety initiatives such as automated paging of critical laboratory values9 and real‐time reporting of user‐requested laboratory data.10 Implementation of alphanumeric paging on a general surgery teaching service reduced disruptions to patient care and the number of pages requiring a return call.11

Our primary aim was to implement an alphanumeric paging system. We will describe our implementation strategies and barriers identified. We evaluated the implementation of alphanumeric paging by measuring (1) the proportion of pages sent as text pages, (2) the source of the pages (other physicians or from the general medicine [GM] ward), (3) the content of the text pages, (4) the number of pages that disrupted scheduled education activities, and (5) satisfaction with the alphanumeric paging system.

Materials and Methods

Implementation Process

We provided alphanumeric pagers to the residents on the General Internal Medicine service in July 2006. Alphanumeric pagers were limited, so each resident traded their numeric pager for an alphanumeric pager at the start of each rotation. Once their rotation ended, they returned their alphanumeric pager for their original numeric pager. The communications department coordinated this process. The chief medical resident spent 10 minutes to teach the residents how to use the system at the beginning of the rotation. In August and September 2006, a member of the communications department trained the nurses on the 4 GM wards how to send a text page using the Internet‐based paging application. We scheduled these 15‐minute sessions throughout the day and evening in order to capture as many nurses as possible. We encouraged the nurses to include standardized information in the text message (eg, patient ID, issue, level of urgency, sender name, call‐back number).

We used rapid‐cycle change methods12 to implement the alphanumeric system (Figure 1). The first change cycle in August 2006 consisted of providing pagers to residents and training the nurses and physicians to send text pages. Users reported that the paging interface was difficult to use. For the second change cycle in September 2006, the communications department modified the paging interface to improve usability and created shortcut icons on the GM ward computers. While the system was easier to access, the nurses reported that 1‐time training was insufficient. For the next change cycle in September 2006, we developed Internet‐based tutorials that could be accessed at any time, and made expert users (charge nurses) available for just‐in‐time training. We asked these charge nurses what they believed would encourage adoption of the system. They suggested that contests worked well with other initiatives. For the final change cycle in October and November 2006, we held a contest and rewarded the GM ward that sent the highest percentage of text pages with a team lunch.

Figure 1

Our results from November 2006 were presented to our hospital medical leaders, who approved widespread implementation of alphanumeric pagers for all residents and medical students in all programs. The cost of this upgrade was approximately $35,000 per year to lease 500 alphanumeric pagers. By July 2007, all residents and students in our hospital had alphanumeric pagers.

Results

Paging Patterns Before, During, and After Implementation

The number of pages per resident was similar before and during implementation, but higher afterwards. (46 16 pages/resident/week in November 2005, 47 20 pages/resident/week in November 2006, and 59 27 pages/resident/week in January 2008; P = 0.17; Table 1). The mean number of admissions per night was 8.0 2.7 before implementation, compared to 10.2 3.5 after implementation (P = 0.009).

Table 1.Paging Patterns Before, During, and After Implementation of an Alphanumeric Paging System

	Before Implementation (all residents had numeric pagers)	During Implementation (all residents had alphanumeric pagers)		After Implementation (all residents had alphanumeric pagers)
Paging Characteristics	November 2005	October 2006	November 2006	January 2008	P Value
Abbreviations: Doc, physician; GM, general medicine; SD, standard deviation. * Student t test comparing November 2005 with January 2008. Fisher's Exact Test comparing November 2005 with January 2008.
Total number of pages	1431	1879	1813	1269
Total number of resident weeks worked	29	33	33	21
Pages per resident week, mean (SD)	46.2 (16.3)	57.9 (19.2)	46.5 (20.2)	59.0 (26.5)	0.17*
Number of patients admitted per night, mean (SD)	8.0 (2.7)	10.2 (3.5)	11.0 (2.9)	10.2 (3.5)	0.009*
Type of page, n (%)
Numeric pages	751 (53)	462 (25)	580 (32)	374 (30)	0.001
GM wards‐to‐Doc	584 (41)	393 (21)	538 (30)	352 (28)	0.001
Doc‐to‐Doc	167 (12)	69 (4)	42 (2)	22 (2)	0.001
Non‐GM ward/Doc pages	680 (47)	1107 (59)	809 (45)	487 (38)	0.001
Text pages	0 (0)	310 (16)	424 (23)	408 (32)	0.001
GM wards‐to‐Doc	0 (0)	175 (9)	221 (12)	129 (10)	0.001
Doc‐to‐Doc	0 (0)	135 (7)	203 (11)	279 (22)	0.001

We observed a significant and sustained increase in the use of text paging during the study (Table 1). After implementation, 32% of all pages sent to our residents were text messages (P 0.001). Physicians almost exclusively sent text pages by the end of implementation (increase from 0% to 83% text paging rate during implementation, and 93% after implementation; P 0.001; Figure 2). GM ward text paging rates also increased from 0% to 29% during implementation, and 27% after implementation (P 0.001; Figure 2). The alphanumeric paging system was used to a greater degree by physicians compared to other workers on the GM ward after full implementation (93% vs. 27%; P 0.001). We explored the proportion of GM ward‐to‐Doc pages sent as text from different GM wards during implementation, and found significant variation, ranging from 14% to 57% (P 0.001).

Figure 2

The most common reasons for text paging from GM wards were to request a patient assessment or for notification of a patient's clinical status (25%), to clarify written orders (20%), and to request a medication prescription (13%) (Table 2). Among physicians, the most common reasons for text paging were to set up meetings for work or teaching rounds (33%), to relay patient‐care related messages (27%), and to sign‐over patients at the end of the day (23%) (Table 3). The remainder of the other Doc‐to‐Doc pages (18%) were mostly personal messages or team communication that was not related to clinical work.

Table 2.Reasons for Paging the Physicians (GM ward‐to‐Doc)

Reason for Paging	Number (%)	Examples
NOTE: Represents data from November 2006. Abbreviations: BP, blood pressure; CXR, chest X‐ray; Doc, physician; GM, general medicine; INR, international normalized ratio; NG, nasogastric tube.
Requests for patient assessment or notification of a patient's clinical status	55 (25)	Patient X. Temp 38.5. No other symptoms. [Nurse's name].
		The repeat CXR on Patient X has been completed. Please call [ward] if NG can be used. Thank you.
Clarification of a written order	45 (20)	Would you like Patient X to get a second dose of Lasix? He has already had 100 mg and his output thus far is 500 cc. [Nurse's name]
		Patient X. BP = 100/62, pulse 54. Patient is supposed to have metoprolol 50 mg tonight. Do you want me to hold it? [Nurse's name]
Request for a medication prescription	28 (13)	Patient X needs an analgesic for pain in his arms and legs. Please call [ward].
		Patient X has a daily coumadin order. INR 2.14. Please call with dosage for 1800. Ask for [Nurse].
Cosigning written order	25 (11)	Please co‐sign neurology suggested orders for Patient X. [Nurse's name]
		Not urgent at your leisure, please co‐sign Patient X orders on [ward] for medical student. Thanks.
Notification of a recent laboratory result	23 (11)	Patient X's potassium is 3.0 today. Please call [ward].
		Patient X. Sodium 163. Troponin unchanged at 1.54. [ward].
Arranging meetings with patients and/or family	18 (8)	Meeting with Patient X's family and social worker at 2 pm tomorrow on [ward].
		[Social worker] is here. [Physiotherapist] expected any minute. We are going to meet in family room for Patient X.
Request to complete paperwork	15 (7)	Patient X is ready to go home, and just needs discharge orders and prescriptions
		Referral form for community palliative doctor on the front of the chart fill in where marked by arrows. Can you please put on form prognosis as well? [Social worker]
Other	12 (6)
Total	221 (100)

Table 3.Reasons for Paging Among Physicians (Doc‐to‐Doc)

Reason for Paging	Number (%)	Examples
NOTE: Represents data from November 2006. Abbreviations: cath, catheterization; CHF, congestive heart failure; Cr, creatinine; DKA, diabetic ketoacidosis; IV, intravenous; Lytes, electrolytes; NS, normal saline; US, ultrasound; WE, weakend.
Setting up meetings for work or teaching rounds	67 (33)	Confirmed diabetes teaching at 1400 hr in [room]. Please let medical students know. Thanks.
		Please come to [lecture theatre] if you can in the next 10 minutes for the teleconferenced noon rounds. Thanks.
We are in the Emergency Department with [attending staff]. Come to meet us here if you can.
Relaying patient‐care related messages	54 (27)	Patient X US query cholangitis, can we ask for a surgical consult. He may benefit from surgery or percutaneous drain. Please repeat his blood work.
		Patient X in [room X]. Presented with DKA. pH 7.2. pCO2 23, bicarbonate 9. Lytes pending. Got IV insulin and NS. Need to check to clinical stability. [Resident]
Signing‐over patients at the end of the day	46 (23)	I'm ready to sign out to you. Where are you? [Resident].
		Patient X is back from cardiac cath, and is stable. Please check Cr over WE, and watch for CHF.
Other	36 (18)
Total	203 (100)

Discussion

We successfully implemented an alphanumeric paging system on a resident inpatient internal medicine teaching service, with 42% of pages from our GM wards or physicians sent as text pages during implementation period. Six months after widespread use of alphanumeric pagers at our hospital (and 14 months after the initial implementation on the General Internal Medicine service) the text paging rate was 52%. Physicians have nearly universally adopted the use of alphanumeric paging as a means of communicating with one another, while the adoption by nursing staff was modest. The implementation of the alphanumeric paging system was associated with a significant reduction in disruptive pages.

We could identify only one prior study of alphanumeric paging implementation in a hospital setting. A general surgery teaching service in the United States demonstrated that 35% of all pages received by residents were text pages 3 months after implementation,11 similar to our result of 32% text paging rate after full implementation. We found a greater use of text paging among physicians in our study (93% of Doc‐to‐Doc pages were text), compared to 55% in this prior study. This difference may be partially explained by varied methods for identifying Doc‐to‐Doc pages between the studies.

A number of factors influence the adoption of new technology, such as the technology's features, end‐user characteristics, and dissemination strategies.14 During the implementation, even though the web‐based paging system was deemed easy to use, there was a lack of computers available to send text messages at each of our nursing stations. Residents were generally more familiar with technology and the use of computers for communication than nurses, and therefore more likely to use the technology. The presence of innovators influenced the success of adoption, evidenced by the fact that the GM ward that sent the highest percentage of pages as text was also the GM ward where the project leader (B.W.) worked as the attending staff during the implementation period.

Our study has several limitations. First, our method for classifying the source of numeric pages was imperfect. Our method systematically underestimates Doc‐to‐Doc numeric pages, because we assumed that all pages from GM ward phone numbers were not from physicians. We also may have misclassified the origin of some text messages. These limitations would not affect our conclusion that text messaging increased, but may overestimate the increase in Doc‐to‐Doc text messaging, and underestimate the increase in GM ward‐to‐Doc text messaging. The reasons for paging are not known for the numeric pages. The number of pages received per resident increased after implementation, so it is possible that alphanumeric pages increased calls for certain nonurgent issues, such as co‐signing orders. Finally, we assumed that residents were attending scheduled educational rounds, but were unable to confirm attendance, so we cannot be sure that disruptions actually were reduced.

In summary, we implemented an alphanumeric paging system, and observed a sustained use of text messaging after 1 year. The implementation of alphanumeric paging was associated with a reduction in disruptive pages sent during scheduled educational rounds.

Effective communication between healthcare providers is essential to patient safety and quality of care.1, 2 Numeric pagers are commonly used communication devices in healthcare, but cannot convey important information such as the reason for the page, urgency of the page, or sender name. Physicians must respond to numeric pages, often disrupting patient encounters or educational activities.36 In a study of medical interns, disruptions to patient care occurred with up to 65% of pages received, two‐thirds of which were not felt to be urgent.5 In addition to causing frustration, frequent disruptions can contribute to medical errors.7, 8

Alphanumeric pagers can display both numbers and text, and may address some of the communication problems associated with numeric pagers. They also lay the groundwork for other patient safety initiatives such as automated paging of critical laboratory values9 and real‐time reporting of user‐requested laboratory data.10 Implementation of alphanumeric paging on a general surgery teaching service reduced disruptions to patient care and the number of pages requiring a return call.11

Our primary aim was to implement an alphanumeric paging system. We will describe our implementation strategies and barriers identified. We evaluated the implementation of alphanumeric paging by measuring (1) the proportion of pages sent as text pages, (2) the source of the pages (other physicians or from the general medicine [GM] ward), (3) the content of the text pages, (4) the number of pages that disrupted scheduled education activities, and (5) satisfaction with the alphanumeric paging system.

Materials and Methods

Implementation Process

We provided alphanumeric pagers to the residents on the General Internal Medicine service in July 2006. Alphanumeric pagers were limited, so each resident traded their numeric pager for an alphanumeric pager at the start of each rotation. Once their rotation ended, they returned their alphanumeric pager for their original numeric pager. The communications department coordinated this process. The chief medical resident spent 10 minutes to teach the residents how to use the system at the beginning of the rotation. In August and September 2006, a member of the communications department trained the nurses on the 4 GM wards how to send a text page using the Internet‐based paging application. We scheduled these 15‐minute sessions throughout the day and evening in order to capture as many nurses as possible. We encouraged the nurses to include standardized information in the text message (eg, patient ID, issue, level of urgency, sender name, call‐back number).

We used rapid‐cycle change methods12 to implement the alphanumeric system (Figure 1). The first change cycle in August 2006 consisted of providing pagers to residents and training the nurses and physicians to send text pages. Users reported that the paging interface was difficult to use. For the second change cycle in September 2006, the communications department modified the paging interface to improve usability and created shortcut icons on the GM ward computers. While the system was easier to access, the nurses reported that 1‐time training was insufficient. For the next change cycle in September 2006, we developed Internet‐based tutorials that could be accessed at any time, and made expert users (charge nurses) available for just‐in‐time training. We asked these charge nurses what they believed would encourage adoption of the system. They suggested that contests worked well with other initiatives. For the final change cycle in October and November 2006, we held a contest and rewarded the GM ward that sent the highest percentage of text pages with a team lunch.

Figure 1

Our results from November 2006 were presented to our hospital medical leaders, who approved widespread implementation of alphanumeric pagers for all residents and medical students in all programs. The cost of this upgrade was approximately $35,000 per year to lease 500 alphanumeric pagers. By July 2007, all residents and students in our hospital had alphanumeric pagers.

Results

Paging Patterns Before, During, and After Implementation

The number of pages per resident was similar before and during implementation, but higher afterwards. (46 16 pages/resident/week in November 2005, 47 20 pages/resident/week in November 2006, and 59 27 pages/resident/week in January 2008; P = 0.17; Table 1). The mean number of admissions per night was 8.0 2.7 before implementation, compared to 10.2 3.5 after implementation (P = 0.009).

Table 1.Paging Patterns Before, During, and After Implementation of an Alphanumeric Paging System

	Before Implementation (all residents had numeric pagers)	During Implementation (all residents had alphanumeric pagers)		After Implementation (all residents had alphanumeric pagers)
Paging Characteristics	November 2005	October 2006	November 2006	January 2008	P Value
Abbreviations: Doc, physician; GM, general medicine; SD, standard deviation. * Student t test comparing November 2005 with January 2008. Fisher's Exact Test comparing November 2005 with January 2008.
Total number of pages	1431	1879	1813	1269
Total number of resident weeks worked	29	33	33	21
Pages per resident week, mean (SD)	46.2 (16.3)	57.9 (19.2)	46.5 (20.2)	59.0 (26.5)	0.17*
Number of patients admitted per night, mean (SD)	8.0 (2.7)	10.2 (3.5)	11.0 (2.9)	10.2 (3.5)	0.009*
Type of page, n (%)
Numeric pages	751 (53)	462 (25)	580 (32)	374 (30)	0.001
GM wards‐to‐Doc	584 (41)	393 (21)	538 (30)	352 (28)	0.001
Doc‐to‐Doc	167 (12)	69 (4)	42 (2)	22 (2)	0.001
Non‐GM ward/Doc pages	680 (47)	1107 (59)	809 (45)	487 (38)	0.001
Text pages	0 (0)	310 (16)	424 (23)	408 (32)	0.001
GM wards‐to‐Doc	0 (0)	175 (9)	221 (12)	129 (10)	0.001
Doc‐to‐Doc	0 (0)	135 (7)	203 (11)	279 (22)	0.001

We observed a significant and sustained increase in the use of text paging during the study (Table 1). After implementation, 32% of all pages sent to our residents were text messages (P 0.001). Physicians almost exclusively sent text pages by the end of implementation (increase from 0% to 83% text paging rate during implementation, and 93% after implementation; P 0.001; Figure 2). GM ward text paging rates also increased from 0% to 29% during implementation, and 27% after implementation (P 0.001; Figure 2). The alphanumeric paging system was used to a greater degree by physicians compared to other workers on the GM ward after full implementation (93% vs. 27%; P 0.001). We explored the proportion of GM ward‐to‐Doc pages sent as text from different GM wards during implementation, and found significant variation, ranging from 14% to 57% (P 0.001).

Figure 2

The most common reasons for text paging from GM wards were to request a patient assessment or for notification of a patient's clinical status (25%), to clarify written orders (20%), and to request a medication prescription (13%) (Table 2). Among physicians, the most common reasons for text paging were to set up meetings for work or teaching rounds (33%), to relay patient‐care related messages (27%), and to sign‐over patients at the end of the day (23%) (Table 3). The remainder of the other Doc‐to‐Doc pages (18%) were mostly personal messages or team communication that was not related to clinical work.

Table 2.Reasons for Paging the Physicians (GM ward‐to‐Doc)

Reason for Paging	Number (%)	Examples
NOTE: Represents data from November 2006. Abbreviations: BP, blood pressure; CXR, chest X‐ray; Doc, physician; GM, general medicine; INR, international normalized ratio; NG, nasogastric tube.
Requests for patient assessment or notification of a patient's clinical status	55 (25)	Patient X. Temp 38.5. No other symptoms. [Nurse's name].
		The repeat CXR on Patient X has been completed. Please call [ward] if NG can be used. Thank you.
Clarification of a written order	45 (20)	Would you like Patient X to get a second dose of Lasix? He has already had 100 mg and his output thus far is 500 cc. [Nurse's name]
		Patient X. BP = 100/62, pulse 54. Patient is supposed to have metoprolol 50 mg tonight. Do you want me to hold it? [Nurse's name]
Request for a medication prescription	28 (13)	Patient X needs an analgesic for pain in his arms and legs. Please call [ward].
		Patient X has a daily coumadin order. INR 2.14. Please call with dosage for 1800. Ask for [Nurse].
Cosigning written order	25 (11)	Please co‐sign neurology suggested orders for Patient X. [Nurse's name]
		Not urgent at your leisure, please co‐sign Patient X orders on [ward] for medical student. Thanks.
Notification of a recent laboratory result	23 (11)	Patient X's potassium is 3.0 today. Please call [ward].
		Patient X. Sodium 163. Troponin unchanged at 1.54. [ward].
Arranging meetings with patients and/or family	18 (8)	Meeting with Patient X's family and social worker at 2 pm tomorrow on [ward].
		[Social worker] is here. [Physiotherapist] expected any minute. We are going to meet in family room for Patient X.
Request to complete paperwork	15 (7)	Patient X is ready to go home, and just needs discharge orders and prescriptions
		Referral form for community palliative doctor on the front of the chart fill in where marked by arrows. Can you please put on form prognosis as well? [Social worker]
Other	12 (6)
Total	221 (100)

Table 3.Reasons for Paging Among Physicians (Doc‐to‐Doc)

Reason for Paging	Number (%)	Examples
NOTE: Represents data from November 2006. Abbreviations: cath, catheterization; CHF, congestive heart failure; Cr, creatinine; DKA, diabetic ketoacidosis; IV, intravenous; Lytes, electrolytes; NS, normal saline; US, ultrasound; WE, weakend.
Setting up meetings for work or teaching rounds	67 (33)	Confirmed diabetes teaching at 1400 hr in [room]. Please let medical students know. Thanks.
		Please come to [lecture theatre] if you can in the next 10 minutes for the teleconferenced noon rounds. Thanks.
We are in the Emergency Department with [attending staff]. Come to meet us here if you can.
Relaying patient‐care related messages	54 (27)	Patient X US query cholangitis, can we ask for a surgical consult. He may benefit from surgery or percutaneous drain. Please repeat his blood work.
		Patient X in [room X]. Presented with DKA. pH 7.2. pCO2 23, bicarbonate 9. Lytes pending. Got IV insulin and NS. Need to check to clinical stability. [Resident]
Signing‐over patients at the end of the day	46 (23)	I'm ready to sign out to you. Where are you? [Resident].
		Patient X is back from cardiac cath, and is stable. Please check Cr over WE, and watch for CHF.
Other	36 (18)
Total	203 (100)

Discussion

We successfully implemented an alphanumeric paging system on a resident inpatient internal medicine teaching service, with 42% of pages from our GM wards or physicians sent as text pages during implementation period. Six months after widespread use of alphanumeric pagers at our hospital (and 14 months after the initial implementation on the General Internal Medicine service) the text paging rate was 52%. Physicians have nearly universally adopted the use of alphanumeric paging as a means of communicating with one another, while the adoption by nursing staff was modest. The implementation of the alphanumeric paging system was associated with a significant reduction in disruptive pages.

We could identify only one prior study of alphanumeric paging implementation in a hospital setting. A general surgery teaching service in the United States demonstrated that 35% of all pages received by residents were text pages 3 months after implementation,11 similar to our result of 32% text paging rate after full implementation. We found a greater use of text paging among physicians in our study (93% of Doc‐to‐Doc pages were text), compared to 55% in this prior study. This difference may be partially explained by varied methods for identifying Doc‐to‐Doc pages between the studies.

A number of factors influence the adoption of new technology, such as the technology's features, end‐user characteristics, and dissemination strategies.14 During the implementation, even though the web‐based paging system was deemed easy to use, there was a lack of computers available to send text messages at each of our nursing stations. Residents were generally more familiar with technology and the use of computers for communication than nurses, and therefore more likely to use the technology. The presence of innovators influenced the success of adoption, evidenced by the fact that the GM ward that sent the highest percentage of pages as text was also the GM ward where the project leader (B.W.) worked as the attending staff during the implementation period.

Our study has several limitations. First, our method for classifying the source of numeric pages was imperfect. Our method systematically underestimates Doc‐to‐Doc numeric pages, because we assumed that all pages from GM ward phone numbers were not from physicians. We also may have misclassified the origin of some text messages. These limitations would not affect our conclusion that text messaging increased, but may overestimate the increase in Doc‐to‐Doc text messaging, and underestimate the increase in GM ward‐to‐Doc text messaging. The reasons for paging are not known for the numeric pages. The number of pages received per resident increased after implementation, so it is possible that alphanumeric pages increased calls for certain nonurgent issues, such as co‐signing orders. Finally, we assumed that residents were attending scheduled educational rounds, but were unable to confirm attendance, so we cannot be sure that disruptions actually were reduced.

In summary, we implemented an alphanumeric paging system, and observed a sustained use of text messaging after 1 year. The implementation of alphanumeric paging was associated with a reduction in disruptive pages sent during scheduled educational rounds.

It is estimated that the Medicare Part A trust fund will be exhausted by 2016 to 2019; also, the quality of care delivered in the United States is highly variable.13 Value is typically defined as the quality achieved for a given cost (ie, value = quality/cost). The focus on the 3 related concepts of value, quality, and cost of health care is likely to continue to increase. Previously, the U.S. Department of Health and Human Services (HHS) made value‐driven health care one of the Department's top priorities.4 Policymakers are in a period of transition but the publicly available plans of the President and Senate leadership indicate that the focus on value‐based initiatives will likely continue to increase as our nation strives to achieve better outcomes for our health care dollar.5, 6 Specifically, the federal government and other payers increasingly align payment incentives with value and quality, encourage public reporting on quality and Medicare payment costs, such as on the Hospital Compare website (http://www.hospitalcompare.hhs.gov), and implement and evaluate demonstrations to test mechanisms such as health information technology (HIT) to improve value‐based performance.

Since hospital care represented $648 billion in 2006, which is 37% of the total patient‐related U.S. health care expenditure, the trend to pay for value will likely have significant impact on hospitals and hospitalists.7 The Society of Hospital Medicine has a public policy committee that provides feedback to government on programs and policies related to value‐driven health care. The policies and programs need consideration and input from the broader community of hospitalists. This work outlines some of the major national initiatives and policies focused on value‐driven health care and their implications for hospitalists. Hospitalists will need to understand the policy landscape and trends, lead improvement in their individual hospitals to receive value‐based incentives, and assess the opportunities and challenges of current and potential payer programs and policies.

Policies and Initiatives: Implications for Hospitals and Hospitalists

Within the portfolio of value‐driven health care, there are at least 6 major government programs, initiatives, and policies with implications for hospitals and hospitalists: value‐based purchasing (VBP), quality and cost public reporting, Medicare demonstrations, hospital‐acquired conditions, incentives for use of effective HIT, and the physician quality reporting initiative (PQRI) (Table 1).

Future Considerations

The political leadership at the federal and state level is beginning a new transition; however, the focus on quality and value for our health care dollar will likely continue to increase.5, 6 The U.S. health care system has untenable cost estimates, significant quality gaps, and a fractured payment system that fails to reward effective care coordination.2, 21, 22 This increased focus on quality and value should be viewed as an opportunity for hospitalists and hospitals. Hospitalist groups that can achieve high‐quality performance will be increasingly valued, and hospitals should further recognize the critical role hospitalists play in achieving high performance and the associated financial rewards. Hospitalists often lead quality improvement and safety programs in hospitals, and these programs are likely to be seen as progressively more important as payment is linked to performance. The Society of Hospital Medicine engages with policymakers and this role is increasingly significant as more policy and payment decisions impact hospitalists. The Society has focused on collaborative work with payers, policymakers, and other providers to find joint shared solutions. Hospitalists can serve as a link between providers and a focal point of care coordination, especially for the hospitalized patient. Finally, as our system and its incentives continue to progress toward alignment with value‐based high quality care, hospitalists should be leading the change and be an essential part of the solution to transform our health care system to provide high‐quality, efficient care to all Americans.

It is estimated that the Medicare Part A trust fund will be exhausted by 2016 to 2019; also, the quality of care delivered in the United States is highly variable.13 Value is typically defined as the quality achieved for a given cost (ie, value = quality/cost). The focus on the 3 related concepts of value, quality, and cost of health care is likely to continue to increase. Previously, the U.S. Department of Health and Human Services (HHS) made value‐driven health care one of the Department's top priorities.4 Policymakers are in a period of transition but the publicly available plans of the President and Senate leadership indicate that the focus on value‐based initiatives will likely continue to increase as our nation strives to achieve better outcomes for our health care dollar.5, 6 Specifically, the federal government and other payers increasingly align payment incentives with value and quality, encourage public reporting on quality and Medicare payment costs, such as on the Hospital Compare website (http://www.hospitalcompare.hhs.gov), and implement and evaluate demonstrations to test mechanisms such as health information technology (HIT) to improve value‐based performance.

Since hospital care represented $648 billion in 2006, which is 37% of the total patient‐related U.S. health care expenditure, the trend to pay for value will likely have significant impact on hospitals and hospitalists.7 The Society of Hospital Medicine has a public policy committee that provides feedback to government on programs and policies related to value‐driven health care. The policies and programs need consideration and input from the broader community of hospitalists. This work outlines some of the major national initiatives and policies focused on value‐driven health care and their implications for hospitalists. Hospitalists will need to understand the policy landscape and trends, lead improvement in their individual hospitals to receive value‐based incentives, and assess the opportunities and challenges of current and potential payer programs and policies.

Policies and Initiatives: Implications for Hospitals and Hospitalists

Within the portfolio of value‐driven health care, there are at least 6 major government programs, initiatives, and policies with implications for hospitals and hospitalists: value‐based purchasing (VBP), quality and cost public reporting, Medicare demonstrations, hospital‐acquired conditions, incentives for use of effective HIT, and the physician quality reporting initiative (PQRI) (Table 1).

Future Considerations

The political leadership at the federal and state level is beginning a new transition; however, the focus on quality and value for our health care dollar will likely continue to increase.5, 6 The U.S. health care system has untenable cost estimates, significant quality gaps, and a fractured payment system that fails to reward effective care coordination.2, 21, 22 This increased focus on quality and value should be viewed as an opportunity for hospitalists and hospitals. Hospitalist groups that can achieve high‐quality performance will be increasingly valued, and hospitals should further recognize the critical role hospitalists play in achieving high performance and the associated financial rewards. Hospitalists often lead quality improvement and safety programs in hospitals, and these programs are likely to be seen as progressively more important as payment is linked to performance. The Society of Hospital Medicine engages with policymakers and this role is increasingly significant as more policy and payment decisions impact hospitalists. The Society has focused on collaborative work with payers, policymakers, and other providers to find joint shared solutions. Hospitalists can serve as a link between providers and a focal point of care coordination, especially for the hospitalized patient. Finally, as our system and its incentives continue to progress toward alignment with value‐based high quality care, hospitalists should be leading the change and be an essential part of the solution to transform our health care system to provide high‐quality, efficient care to all Americans.

It is estimated that the Medicare Part A trust fund will be exhausted by 2016 to 2019; also, the quality of care delivered in the United States is highly variable.13 Value is typically defined as the quality achieved for a given cost (ie, value = quality/cost). The focus on the 3 related concepts of value, quality, and cost of health care is likely to continue to increase. Previously, the U.S. Department of Health and Human Services (HHS) made value‐driven health care one of the Department's top priorities.4 Policymakers are in a period of transition but the publicly available plans of the President and Senate leadership indicate that the focus on value‐based initiatives will likely continue to increase as our nation strives to achieve better outcomes for our health care dollar.5, 6 Specifically, the federal government and other payers increasingly align payment incentives with value and quality, encourage public reporting on quality and Medicare payment costs, such as on the Hospital Compare website (http://www.hospitalcompare.hhs.gov), and implement and evaluate demonstrations to test mechanisms such as health information technology (HIT) to improve value‐based performance.

Since hospital care represented $648 billion in 2006, which is 37% of the total patient‐related U.S. health care expenditure, the trend to pay for value will likely have significant impact on hospitals and hospitalists.7 The Society of Hospital Medicine has a public policy committee that provides feedback to government on programs and policies related to value‐driven health care. The policies and programs need consideration and input from the broader community of hospitalists. This work outlines some of the major national initiatives and policies focused on value‐driven health care and their implications for hospitalists. Hospitalists will need to understand the policy landscape and trends, lead improvement in their individual hospitals to receive value‐based incentives, and assess the opportunities and challenges of current and potential payer programs and policies.

Policies and Initiatives: Implications for Hospitals and Hospitalists

Within the portfolio of value‐driven health care, there are at least 6 major government programs, initiatives, and policies with implications for hospitals and hospitalists: value‐based purchasing (VBP), quality and cost public reporting, Medicare demonstrations, hospital‐acquired conditions, incentives for use of effective HIT, and the physician quality reporting initiative (PQRI) (Table 1).

Future Considerations

The political leadership at the federal and state level is beginning a new transition; however, the focus on quality and value for our health care dollar will likely continue to increase.5, 6 The U.S. health care system has untenable cost estimates, significant quality gaps, and a fractured payment system that fails to reward effective care coordination.2, 21, 22 This increased focus on quality and value should be viewed as an opportunity for hospitalists and hospitals. Hospitalist groups that can achieve high‐quality performance will be increasingly valued, and hospitals should further recognize the critical role hospitalists play in achieving high performance and the associated financial rewards. Hospitalists often lead quality improvement and safety programs in hospitals, and these programs are likely to be seen as progressively more important as payment is linked to performance. The Society of Hospital Medicine engages with policymakers and this role is increasingly significant as more policy and payment decisions impact hospitalists. The Society has focused on collaborative work with payers, policymakers, and other providers to find joint shared solutions. Hospitalists can serve as a link between providers and a focal point of care coordination, especially for the hospitalized patient. Finally, as our system and its incentives continue to progress toward alignment with value‐based high quality care, hospitalists should be leading the change and be an essential part of the solution to transform our health care system to provide high‐quality, efficient care to all Americans.

By the year 2010, more than 20,000 hospitalists will be in practice, compared to 5000 rheumatologists and 8000 pulmonologists.13 The growth of this career option has been driven by an industry need to reduce healthcare costs, increase the emphasis on quality improvement of healthcare services, and improve the efficiency and delivery of care between that provided in the hospital and that provided by the primary care physician.49

While hospitalists' roots can be traced back to community hospitals in the late 1980s and early 1990s, this career option is now flourishing in academic centers, with the rise of hospitalist faculty and hospitalist faculty tracks.10, 11 One potential advantage of having faculty who are hospitalists is the availability and expertise of physicians who specialize in the care of hospitalized patients.8, 12 Additionally, hospitalist faculty have been reported to achieve high resident satisfaction scores, increase understanding of cost‐effective measures, and improve the supervision of hospital procedures.12 To our knowledge, the medical literature does not provide an estimate of the percentage of internal medicine residency programs utilizing hospitalist faculty.

Critics of hospitalist faculty point to the potential loss of teaching opportunities from shorter hospital stays, bemoan the decreased physician‐patient continuity between inpatient and outpatient arenas, and fear that the hospitalists' presence may decrease resident autonomy and decrease subspecialty consultations by fellows.5, 13, 14 Hospitalist faculty need development, education, and training to match the teaching activities they are expected to fulfill. The challenge is for hospitalist societies and national residency organizations to define, plan for, and meet their faculty development needs.

Goals

The goals of this study were to describe the current involvement of hospitalists in internal medicine residencies. More specifically, we wanted to determine: (1) the percentage of programs with hospitalists as faculty, (2) the teaching activities of hospitalists, (3) regional differences in academic hospitalist activity, and (4) the number of programs with hospitalist training tracks.

Materials and Methods

Results

A total of 272 (response rate 70%) programs completed the 2005 survey, and 236 (response rate 62%) completed the 2007 survey. A total of 171 programs completed both surveys. In 2007, 15 (6%) program directors reported that they were hospitalists while 118 (50%) claimed to be traditional general internists.

For the program directors who answered both surveys, 57% indicated that their primary teaching hospital employed hospitalists before the residency work‐hour limits were implemented (before June 2002). At the time of the survey in 2005, 77% said that hospitalists were employed, a 20% increase in 3 years. When we surveyed these same programs again in 2007, the proportion had risen to 81% (Fisher's exact P = 0.02 compared to before work‐hour limits; Figure 1).

Figure 1

Discussion

This is the first study to document the national rise of hospitalist faculty in internal medicine residency programs. Program directors noted a 20% increase in teaching hospitals that employed hospitalists after the work‐hour regulations went into effecta trend that continued to rise. The tendency was seen first on the coasts, where managed care has higher penetration, and particularly in the Northeast, where New York's resident work‐hour reforms occurred by state mandate prior to the residency accreditation action that affected the rest of the country. Not only have hospitalists picked up the burden of service at these hospitals,19, 20 but the vast majority of programs (>80%) have utilized hospitalists as teachers in important areas of their residency. The magnitude of hospitalist involvement in residency training may have important implications.

Beyond the financial significance of hospitalists at academic teaching hospitals,21 only a few studies have addressed their impact on resident education. On the monthly evaluations at the University of California, San Francisco (San Francisco, CA), residents' satisfaction with their attendings was significantly higher when the physician was a hospitalist rather than a traditional faculty member.22 Residents believed hospitalists were more effective teachers, and provided more effective feedback. At Emory University (Atlanta, GA), a methodologically more rigorous study of postrotation assessment of faculty demonstrated that ratings of hospitalists were not different from traditional general internists; both scored higher than subspecialists.23 The hospitalists as a group had completed training more recently, which also was associated with higher scores.

Even community hospitals that sponsor residency programs have benefited from hospitalist faculty. At Norwalk Hospital (Norwalk, CT), the program had used resident teams led by a group of community physicians and a small group of employed internists. But time pressures and reimbursement concerns created tension between the workload and education balance. After hiring 2 hospitalist clinician‐educators, the length of stay and cost per case were substantially reduced, while resident evaluations indicated improved teaching rounds, conferences, and bedside teaching.24

The results of our study fit with the role of hospitalists as well as what individual programs have reported about hospitalist faculty in the past. Hospitalist faculty serve by and large (92%) as attendings on the hospital ward services. Theoretically, who better to have round with residents in the hospital than the specialists of hospital medicine. For a pulmonary curricular experience, residents work with pulmonologists. But beyond serving as attendings in the hospital, they perform the traditional functions of hospital attendings: providing teaching rounds (>80%), evaluating clinical skills (67%), and even lecturing to residents (>65%). The Southern region trends toward slower adoption of hospitalists as faculty, particularly compared with the Northeast. Overall, what is striking is how much hospitalist faculty already are filling the roles expected of all academic faculty.

We found that only 11% of programs have a hospitalist track through which internal medicine residents may develop the specialized skills and knowledge needed to function optimally in a hospitalist career.25 But given the rapid growth of this specialty, we might expect to see a similar rise in programs providing such specialized training.

Are there risks to having hospitalists teaching residents? One concern is the potential to model fragmented medical care to trainees when hospitals and ambulatory health systems neglect to ensure quality handoffs.26 In an era that heralds the demise of the primary care general internist,27 the impact of hospitalist faculty on general internal medicine nationally, the gravitation of residents toward or away from hospitalist and ambulatory careers, and the role of the traditional general internist in residency training programs in the future remain to be seen. These were not addressed by our data, but are ripe areas for study.

This study has several limitations. It relies on self‐reported data from program directors that, while knowing the intimate details of their educational program, may not have exact knowledge of the number of hospitalists employed by their hospitals. There is also the potential for recall bias by asking the group to remember the number of hospitalists before duty‐hour implementation. Both points in time of our 2 surveys (2005 and 2007) were after the incident growth of hospital medicine as evidenced by the high prevalence of hospitalists in both surveys. Yet, most program directors know that the service needs of the hospitals were acutely increased when the duty‐hours policies went into affect, and probably were fairly involved in hospital decisions to utilize hospitalist physicians to meet these needs. Finally, our study does not address hospitalism within the family medicine and pediatric specialties, both of which have a significant stake in hospital medicine.

In conclusion, our study documents the recent growth and current prevalence of hospitalists' activities in the teaching hospitals of internal medicine residencies in the United States, the duties they perform in resident education, and the magnitude of their penetration in the geographic regions of the country, both in community‐based and university‐based programs. The high degree of involvement of hospitalists in resident education may have important implications for the future of internal medicine as a discipline both with regard to the need for academic faculty development of this important sector of the education community as well as for the education and career development of the residents whom they train.

By the year 2010, more than 20,000 hospitalists will be in practice, compared to 5000 rheumatologists and 8000 pulmonologists.13 The growth of this career option has been driven by an industry need to reduce healthcare costs, increase the emphasis on quality improvement of healthcare services, and improve the efficiency and delivery of care between that provided in the hospital and that provided by the primary care physician.49

While hospitalists' roots can be traced back to community hospitals in the late 1980s and early 1990s, this career option is now flourishing in academic centers, with the rise of hospitalist faculty and hospitalist faculty tracks.10, 11 One potential advantage of having faculty who are hospitalists is the availability and expertise of physicians who specialize in the care of hospitalized patients.8, 12 Additionally, hospitalist faculty have been reported to achieve high resident satisfaction scores, increase understanding of cost‐effective measures, and improve the supervision of hospital procedures.12 To our knowledge, the medical literature does not provide an estimate of the percentage of internal medicine residency programs utilizing hospitalist faculty.

Critics of hospitalist faculty point to the potential loss of teaching opportunities from shorter hospital stays, bemoan the decreased physician‐patient continuity between inpatient and outpatient arenas, and fear that the hospitalists' presence may decrease resident autonomy and decrease subspecialty consultations by fellows.5, 13, 14 Hospitalist faculty need development, education, and training to match the teaching activities they are expected to fulfill. The challenge is for hospitalist societies and national residency organizations to define, plan for, and meet their faculty development needs.

Goals

The goals of this study were to describe the current involvement of hospitalists in internal medicine residencies. More specifically, we wanted to determine: (1) the percentage of programs with hospitalists as faculty, (2) the teaching activities of hospitalists, (3) regional differences in academic hospitalist activity, and (4) the number of programs with hospitalist training tracks.

Materials and Methods

Results

A total of 272 (response rate 70%) programs completed the 2005 survey, and 236 (response rate 62%) completed the 2007 survey. A total of 171 programs completed both surveys. In 2007, 15 (6%) program directors reported that they were hospitalists while 118 (50%) claimed to be traditional general internists.

For the program directors who answered both surveys, 57% indicated that their primary teaching hospital employed hospitalists before the residency work‐hour limits were implemented (before June 2002). At the time of the survey in 2005, 77% said that hospitalists were employed, a 20% increase in 3 years. When we surveyed these same programs again in 2007, the proportion had risen to 81% (Fisher's exact P = 0.02 compared to before work‐hour limits; Figure 1).

Figure 1

Discussion

This is the first study to document the national rise of hospitalist faculty in internal medicine residency programs. Program directors noted a 20% increase in teaching hospitals that employed hospitalists after the work‐hour regulations went into effecta trend that continued to rise. The tendency was seen first on the coasts, where managed care has higher penetration, and particularly in the Northeast, where New York's resident work‐hour reforms occurred by state mandate prior to the residency accreditation action that affected the rest of the country. Not only have hospitalists picked up the burden of service at these hospitals,19, 20 but the vast majority of programs (>80%) have utilized hospitalists as teachers in important areas of their residency. The magnitude of hospitalist involvement in residency training may have important implications.

Beyond the financial significance of hospitalists at academic teaching hospitals,21 only a few studies have addressed their impact on resident education. On the monthly evaluations at the University of California, San Francisco (San Francisco, CA), residents' satisfaction with their attendings was significantly higher when the physician was a hospitalist rather than a traditional faculty member.22 Residents believed hospitalists were more effective teachers, and provided more effective feedback. At Emory University (Atlanta, GA), a methodologically more rigorous study of postrotation assessment of faculty demonstrated that ratings of hospitalists were not different from traditional general internists; both scored higher than subspecialists.23 The hospitalists as a group had completed training more recently, which also was associated with higher scores.

Even community hospitals that sponsor residency programs have benefited from hospitalist faculty. At Norwalk Hospital (Norwalk, CT), the program had used resident teams led by a group of community physicians and a small group of employed internists. But time pressures and reimbursement concerns created tension between the workload and education balance. After hiring 2 hospitalist clinician‐educators, the length of stay and cost per case were substantially reduced, while resident evaluations indicated improved teaching rounds, conferences, and bedside teaching.24

The results of our study fit with the role of hospitalists as well as what individual programs have reported about hospitalist faculty in the past. Hospitalist faculty serve by and large (92%) as attendings on the hospital ward services. Theoretically, who better to have round with residents in the hospital than the specialists of hospital medicine. For a pulmonary curricular experience, residents work with pulmonologists. But beyond serving as attendings in the hospital, they perform the traditional functions of hospital attendings: providing teaching rounds (>80%), evaluating clinical skills (67%), and even lecturing to residents (>65%). The Southern region trends toward slower adoption of hospitalists as faculty, particularly compared with the Northeast. Overall, what is striking is how much hospitalist faculty already are filling the roles expected of all academic faculty.

We found that only 11% of programs have a hospitalist track through which internal medicine residents may develop the specialized skills and knowledge needed to function optimally in a hospitalist career.25 But given the rapid growth of this specialty, we might expect to see a similar rise in programs providing such specialized training.

Are there risks to having hospitalists teaching residents? One concern is the potential to model fragmented medical care to trainees when hospitals and ambulatory health systems neglect to ensure quality handoffs.26 In an era that heralds the demise of the primary care general internist,27 the impact of hospitalist faculty on general internal medicine nationally, the gravitation of residents toward or away from hospitalist and ambulatory careers, and the role of the traditional general internist in residency training programs in the future remain to be seen. These were not addressed by our data, but are ripe areas for study.

This study has several limitations. It relies on self‐reported data from program directors that, while knowing the intimate details of their educational program, may not have exact knowledge of the number of hospitalists employed by their hospitals. There is also the potential for recall bias by asking the group to remember the number of hospitalists before duty‐hour implementation. Both points in time of our 2 surveys (2005 and 2007) were after the incident growth of hospital medicine as evidenced by the high prevalence of hospitalists in both surveys. Yet, most program directors know that the service needs of the hospitals were acutely increased when the duty‐hours policies went into affect, and probably were fairly involved in hospital decisions to utilize hospitalist physicians to meet these needs. Finally, our study does not address hospitalism within the family medicine and pediatric specialties, both of which have a significant stake in hospital medicine.

In conclusion, our study documents the recent growth and current prevalence of hospitalists' activities in the teaching hospitals of internal medicine residencies in the United States, the duties they perform in resident education, and the magnitude of their penetration in the geographic regions of the country, both in community‐based and university‐based programs. The high degree of involvement of hospitalists in resident education may have important implications for the future of internal medicine as a discipline both with regard to the need for academic faculty development of this important sector of the education community as well as for the education and career development of the residents whom they train.

By the year 2010, more than 20,000 hospitalists will be in practice, compared to 5000 rheumatologists and 8000 pulmonologists.13 The growth of this career option has been driven by an industry need to reduce healthcare costs, increase the emphasis on quality improvement of healthcare services, and improve the efficiency and delivery of care between that provided in the hospital and that provided by the primary care physician.49

While hospitalists' roots can be traced back to community hospitals in the late 1980s and early 1990s, this career option is now flourishing in academic centers, with the rise of hospitalist faculty and hospitalist faculty tracks.10, 11 One potential advantage of having faculty who are hospitalists is the availability and expertise of physicians who specialize in the care of hospitalized patients.8, 12 Additionally, hospitalist faculty have been reported to achieve high resident satisfaction scores, increase understanding of cost‐effective measures, and improve the supervision of hospital procedures.12 To our knowledge, the medical literature does not provide an estimate of the percentage of internal medicine residency programs utilizing hospitalist faculty.

Critics of hospitalist faculty point to the potential loss of teaching opportunities from shorter hospital stays, bemoan the decreased physician‐patient continuity between inpatient and outpatient arenas, and fear that the hospitalists' presence may decrease resident autonomy and decrease subspecialty consultations by fellows.5, 13, 14 Hospitalist faculty need development, education, and training to match the teaching activities they are expected to fulfill. The challenge is for hospitalist societies and national residency organizations to define, plan for, and meet their faculty development needs.

Goals

The goals of this study were to describe the current involvement of hospitalists in internal medicine residencies. More specifically, we wanted to determine: (1) the percentage of programs with hospitalists as faculty, (2) the teaching activities of hospitalists, (3) regional differences in academic hospitalist activity, and (4) the number of programs with hospitalist training tracks.

Materials and Methods

Results

A total of 272 (response rate 70%) programs completed the 2005 survey, and 236 (response rate 62%) completed the 2007 survey. A total of 171 programs completed both surveys. In 2007, 15 (6%) program directors reported that they were hospitalists while 118 (50%) claimed to be traditional general internists.

For the program directors who answered both surveys, 57% indicated that their primary teaching hospital employed hospitalists before the residency work‐hour limits were implemented (before June 2002). At the time of the survey in 2005, 77% said that hospitalists were employed, a 20% increase in 3 years. When we surveyed these same programs again in 2007, the proportion had risen to 81% (Fisher's exact P = 0.02 compared to before work‐hour limits; Figure 1).

Figure 1

Discussion

This is the first study to document the national rise of hospitalist faculty in internal medicine residency programs. Program directors noted a 20% increase in teaching hospitals that employed hospitalists after the work‐hour regulations went into effecta trend that continued to rise. The tendency was seen first on the coasts, where managed care has higher penetration, and particularly in the Northeast, where New York's resident work‐hour reforms occurred by state mandate prior to the residency accreditation action that affected the rest of the country. Not only have hospitalists picked up the burden of service at these hospitals,19, 20 but the vast majority of programs (>80%) have utilized hospitalists as teachers in important areas of their residency. The magnitude of hospitalist involvement in residency training may have important implications.

Beyond the financial significance of hospitalists at academic teaching hospitals,21 only a few studies have addressed their impact on resident education. On the monthly evaluations at the University of California, San Francisco (San Francisco, CA), residents' satisfaction with their attendings was significantly higher when the physician was a hospitalist rather than a traditional faculty member.22 Residents believed hospitalists were more effective teachers, and provided more effective feedback. At Emory University (Atlanta, GA), a methodologically more rigorous study of postrotation assessment of faculty demonstrated that ratings of hospitalists were not different from traditional general internists; both scored higher than subspecialists.23 The hospitalists as a group had completed training more recently, which also was associated with higher scores.

Even community hospitals that sponsor residency programs have benefited from hospitalist faculty. At Norwalk Hospital (Norwalk, CT), the program had used resident teams led by a group of community physicians and a small group of employed internists. But time pressures and reimbursement concerns created tension between the workload and education balance. After hiring 2 hospitalist clinician‐educators, the length of stay and cost per case were substantially reduced, while resident evaluations indicated improved teaching rounds, conferences, and bedside teaching.24

The results of our study fit with the role of hospitalists as well as what individual programs have reported about hospitalist faculty in the past. Hospitalist faculty serve by and large (92%) as attendings on the hospital ward services. Theoretically, who better to have round with residents in the hospital than the specialists of hospital medicine. For a pulmonary curricular experience, residents work with pulmonologists. But beyond serving as attendings in the hospital, they perform the traditional functions of hospital attendings: providing teaching rounds (>80%), evaluating clinical skills (67%), and even lecturing to residents (>65%). The Southern region trends toward slower adoption of hospitalists as faculty, particularly compared with the Northeast. Overall, what is striking is how much hospitalist faculty already are filling the roles expected of all academic faculty.

We found that only 11% of programs have a hospitalist track through which internal medicine residents may develop the specialized skills and knowledge needed to function optimally in a hospitalist career.25 But given the rapid growth of this specialty, we might expect to see a similar rise in programs providing such specialized training.

Are there risks to having hospitalists teaching residents? One concern is the potential to model fragmented medical care to trainees when hospitals and ambulatory health systems neglect to ensure quality handoffs.26 In an era that heralds the demise of the primary care general internist,27 the impact of hospitalist faculty on general internal medicine nationally, the gravitation of residents toward or away from hospitalist and ambulatory careers, and the role of the traditional general internist in residency training programs in the future remain to be seen. These were not addressed by our data, but are ripe areas for study.

This study has several limitations. It relies on self‐reported data from program directors that, while knowing the intimate details of their educational program, may not have exact knowledge of the number of hospitalists employed by their hospitals. There is also the potential for recall bias by asking the group to remember the number of hospitalists before duty‐hour implementation. Both points in time of our 2 surveys (2005 and 2007) were after the incident growth of hospital medicine as evidenced by the high prevalence of hospitalists in both surveys. Yet, most program directors know that the service needs of the hospitals were acutely increased when the duty‐hours policies went into affect, and probably were fairly involved in hospital decisions to utilize hospitalist physicians to meet these needs. Finally, our study does not address hospitalism within the family medicine and pediatric specialties, both of which have a significant stake in hospital medicine.

In conclusion, our study documents the recent growth and current prevalence of hospitalists' activities in the teaching hospitals of internal medicine residencies in the United States, the duties they perform in resident education, and the magnitude of their penetration in the geographic regions of the country, both in community‐based and university‐based programs. The high degree of involvement of hospitalists in resident education may have important implications for the future of internal medicine as a discipline both with regard to the need for academic faculty development of this important sector of the education community as well as for the education and career development of the residents whom they train.

Hospitalists are increasingly assuming a primary role in medical education in the hospital setting, as they also steadily care for a larger portion of hospitalized patients.1 This issue of the Journal of Hospital Medicine highlights the role of hospitalists as teachers in academic medical centers, confirming their expanding and positive role in resident and medical student education. A survey of academic medical centers, a systematic review, an evaluation of the implementation of an educational curriculum, and a survey of residents hint at the challenges hospitalists face in teaching, but also expose us to a more advanced yet facile approach to evaluating the effectiveness of a teaching intervention.25 These publications provoke interesting questions about clinical teaching that hospital medicine educators and researchers should pursue answering. I believe they will also encourage us to innovate in medical education and assessment of that teaching.

Traditionally, teaching attendings for resident teams on medicine or pediatric services rotated through these duties for 1 to 3 months each year, while spending the majority of their time in clinic or research activities. The increasing complexity of hospitalized patients and the pressure to reduce length of stay prompted closer oversight of trainees. With the advent of resident work‐hour restrictions, the need for greater clinical involvement by attending physicians made it increasingly difficult to maintain the traditional model of limited engagement by faculty attendings. Simply put, the dwindling pool of willing and able teaching attendings encouraged teaching hospitals to employ hospitalists to fill the gap in teaching and supervision, as well as clinical coverage.6

Beasley et al.2 report that resident work‐hour restrictions were associated with an increase in the number of teaching hospitals employing hospitalists to 79% of 193 surveyed hospitals in 2007. Of those hospitals with hospitalists, 92% reported that hospitalists serve as attendings on the teaching service. Hospitalists also teach in a number of other venues within these programs, including formal teaching rounds without direct care responsibility, along with delivering didactic lectures and clinical skills education.

How well are teaching hospitalists performing compared to traditional teaching attendings? Natarajan et al.4 provide an important summary of the evidence in a systematic review of studies comparing teaching efforts of hospitalist attendings to those of nonhospitalist attendings. Eight studies from a variety of institutions measured trainee (resident or medical student) attitudes. It is gratifying to learn that hospitalists were generally rated higher at overall teaching effectiveness, provision of feedback, knowledge base, and involvement of the learner in patient care. It seems likely that publication bias would overestimate the positive effect of hospitalists on learner attitudes. However, there are plausible reasons that the positive effect is accurate. Because their professional responsibilities are focused in the hospital, hospitalists should naturally be more available to learners for teaching and feedback. Hospitalists tend to be younger in their academic careers, placing them closer to the cutting edge of knowledge gained during residency and possibly fellowship. They may be more in tune with the needs and pressures faced by their learners, having dealt with these same challenges either during recent training or during nonteaching rotations.

As a relatively young specialty with young and developing academic hospitalists, will the advantage suggested by the Natarajan et al.4 systematic review be sustained over the long term as careers in hospital medicine mature? A 2005 systematic review studying this question among practicing clinicians found, somewhat paradoxically, that older, more experienced clinicians appeared to be at risk for providing lower‐quality care.7 To avoid this decline in clinical effectiveness, hospitalists should proactively seek innovative ways to refresh and update their knowledge and skills throughout their careers. This is particularly critical for teaching physicians. We should seize the opportunity to study the relationship between advancing clinical/teaching experience and educational quality within our teaching programs.

The review by Natarajan et al.4 should also challenge the hospitalist community to achieve even higher levels of proficiency as teachers of medicine. The review alludes to bedside teaching and attention to psychosocial aspects of care as opportunities for improvement by hospitalist teachers. A recent study suggested that physical examination instruction receives declining attention from inpatient teachers and that there are opportunities to increase the amount of bedside teaching.8 A provocative study of inpatients admitted to a teaching service found that physical examination could substantially impact patient care, but that trainees often failed to appreciate significant findings on initial examination.9 How do teaching hospitalists become proficient at physical examination and bedside teaching? Are there models around the country that are successfully developing outstanding clinician educators, incorporating teach‐the‐teacher models to improve physical examination and bedside teaching?

A practical limitation of attitude surveys and learner evaluation is the well‐known phenomenon of grade inflation that resulted in high ratings for all attending groups in the studies summarized by Natarajan et al.4 This limits the ability of surveys or evaluations to distinguish truly outstanding teachers and consequently makes it difficult to analyze the attributes of these teachers. We need better tools to detect and learn teaching techniques from great teachers in the clinical environment. We need studies evaluating the effect of teaching hospitalists on learner knowledge or, even more importantly, learner outcomes. Ultimately, we need studies of educational interventions that evaluate the impact of these interventions on patient outcome.

Wright et al.5 provide guidance as they describe the evaluation of a teaching intervention that moves beyond measurement of knowledge or attitudes. The Johns Hopkins Bayview hospitalist group sought to improve the quality of medical consultations performed by hospitalists and by residents rotating on the consultation service using a case‐based teaching module with audits of recent notes. The participants then audited their most recent consultation notes with feedback from the module teacher. The study employed pretests and posttests of knowledgea standard evaluation for educational interventions. This tells us little about the true impact of the teaching module. However, the study then assessed the quality of written consultations done by hospitalists before and after the educational interventions. Scores of consult notes improved significantly after the intervention, although the number of assessments for each physician was limited. Importantly, we need to know if interventions such as this are sustained over time. Wright's well‐established medical education research group's study design assessed the impact of an intervention on physician performance and moves us closer to assessment of the impact on actual patient outcomes. As clinical teachers, we would like to believe that our teaching and our educational innovations are having a positive impact on patient care. Can we demonstrate this?

As academic medical centers contend with further resident work‐hour restrictions proposed by the Institute of Medicine (IOM),10 how will this affect hospitalist teachers? The study by Mazotti et al.3 from the University of California at San Francisco residency program found that about one‐quarter of residents reported spending less time teaching after implementation of the Accreditation Council for Graduate Medical Education (ACGME) duty‐hour restrictions in 2003. Interestingly, those residents reporting less time spent teaching also reported less emotional exhaustion and perceived that they were delivering higher‐quality patient care. This raises a fascinating question for academic hospitalists. Would these findings be similar among teaching hospitalists and nonteaching hospitalists? What about hospitalists who rotate through months of teaching and nonteaching services? Is teaching emotionally exhausting for experienced teachers? A Mayo Clinic study suggested that the extent that faculty physicians are able to engage in work that is most meaningful to them as individuals is a strong determinant of faculty burnout.11 Is the hospitalist who finds teaching most rewarding at risk of burnout if they are assigned only 2 weeks a year as a teaching attending? The answers to these questions will be critical to hospitalist program leaders trying to assure sustainable careers for hospitalists in their programs.

Although the study by Mazotti et al.3 did not assess the impact of the reduction in resident teaching time on the teaching responsibilities for academic hospitalists, previous studies suggest that faculty are also teaching less since the introduction of work‐hour restrictions.12, 13 If the new IOM recommendations are enacted, who will teach? Although the reported experience following the 2003 work‐hour restrictions begs pessimism, the anticipated changes represent an opportunity for creative hospitalist teachers to demonstrate effective adaptations to the changing and compressed inpatient teaching environment.

In summary, this issue of the Journal presents studies that praise the role hospitalists play in teaching the next generation of physicians, but also gives a glimpse of future challenges and opportunities. We should take advantage of hospitalists' central position in clinical education in the hospital to innovate, study the effect on both learner outcomes and patient outcomes, and share our experiences with the hospitalist and medical education communities.

Hospitalists are increasingly assuming a primary role in medical education in the hospital setting, as they also steadily care for a larger portion of hospitalized patients.1 This issue of the Journal of Hospital Medicine highlights the role of hospitalists as teachers in academic medical centers, confirming their expanding and positive role in resident and medical student education. A survey of academic medical centers, a systematic review, an evaluation of the implementation of an educational curriculum, and a survey of residents hint at the challenges hospitalists face in teaching, but also expose us to a more advanced yet facile approach to evaluating the effectiveness of a teaching intervention.25 These publications provoke interesting questions about clinical teaching that hospital medicine educators and researchers should pursue answering. I believe they will also encourage us to innovate in medical education and assessment of that teaching.

Traditionally, teaching attendings for resident teams on medicine or pediatric services rotated through these duties for 1 to 3 months each year, while spending the majority of their time in clinic or research activities. The increasing complexity of hospitalized patients and the pressure to reduce length of stay prompted closer oversight of trainees. With the advent of resident work‐hour restrictions, the need for greater clinical involvement by attending physicians made it increasingly difficult to maintain the traditional model of limited engagement by faculty attendings. Simply put, the dwindling pool of willing and able teaching attendings encouraged teaching hospitals to employ hospitalists to fill the gap in teaching and supervision, as well as clinical coverage.6

Beasley et al.2 report that resident work‐hour restrictions were associated with an increase in the number of teaching hospitals employing hospitalists to 79% of 193 surveyed hospitals in 2007. Of those hospitals with hospitalists, 92% reported that hospitalists serve as attendings on the teaching service. Hospitalists also teach in a number of other venues within these programs, including formal teaching rounds without direct care responsibility, along with delivering didactic lectures and clinical skills education.

How well are teaching hospitalists performing compared to traditional teaching attendings? Natarajan et al.4 provide an important summary of the evidence in a systematic review of studies comparing teaching efforts of hospitalist attendings to those of nonhospitalist attendings. Eight studies from a variety of institutions measured trainee (resident or medical student) attitudes. It is gratifying to learn that hospitalists were generally rated higher at overall teaching effectiveness, provision of feedback, knowledge base, and involvement of the learner in patient care. It seems likely that publication bias would overestimate the positive effect of hospitalists on learner attitudes. However, there are plausible reasons that the positive effect is accurate. Because their professional responsibilities are focused in the hospital, hospitalists should naturally be more available to learners for teaching and feedback. Hospitalists tend to be younger in their academic careers, placing them closer to the cutting edge of knowledge gained during residency and possibly fellowship. They may be more in tune with the needs and pressures faced by their learners, having dealt with these same challenges either during recent training or during nonteaching rotations.

As a relatively young specialty with young and developing academic hospitalists, will the advantage suggested by the Natarajan et al.4 systematic review be sustained over the long term as careers in hospital medicine mature? A 2005 systematic review studying this question among practicing clinicians found, somewhat paradoxically, that older, more experienced clinicians appeared to be at risk for providing lower‐quality care.7 To avoid this decline in clinical effectiveness, hospitalists should proactively seek innovative ways to refresh and update their knowledge and skills throughout their careers. This is particularly critical for teaching physicians. We should seize the opportunity to study the relationship between advancing clinical/teaching experience and educational quality within our teaching programs.

The review by Natarajan et al.4 should also challenge the hospitalist community to achieve even higher levels of proficiency as teachers of medicine. The review alludes to bedside teaching and attention to psychosocial aspects of care as opportunities for improvement by hospitalist teachers. A recent study suggested that physical examination instruction receives declining attention from inpatient teachers and that there are opportunities to increase the amount of bedside teaching.8 A provocative study of inpatients admitted to a teaching service found that physical examination could substantially impact patient care, but that trainees often failed to appreciate significant findings on initial examination.9 How do teaching hospitalists become proficient at physical examination and bedside teaching? Are there models around the country that are successfully developing outstanding clinician educators, incorporating teach‐the‐teacher models to improve physical examination and bedside teaching?

A practical limitation of attitude surveys and learner evaluation is the well‐known phenomenon of grade inflation that resulted in high ratings for all attending groups in the studies summarized by Natarajan et al.4 This limits the ability of surveys or evaluations to distinguish truly outstanding teachers and consequently makes it difficult to analyze the attributes of these teachers. We need better tools to detect and learn teaching techniques from great teachers in the clinical environment. We need studies evaluating the effect of teaching hospitalists on learner knowledge or, even more importantly, learner outcomes. Ultimately, we need studies of educational interventions that evaluate the impact of these interventions on patient outcome.

Wright et al.5 provide guidance as they describe the evaluation of a teaching intervention that moves beyond measurement of knowledge or attitudes. The Johns Hopkins Bayview hospitalist group sought to improve the quality of medical consultations performed by hospitalists and by residents rotating on the consultation service using a case‐based teaching module with audits of recent notes. The participants then audited their most recent consultation notes with feedback from the module teacher. The study employed pretests and posttests of knowledgea standard evaluation for educational interventions. This tells us little about the true impact of the teaching module. However, the study then assessed the quality of written consultations done by hospitalists before and after the educational interventions. Scores of consult notes improved significantly after the intervention, although the number of assessments for each physician was limited. Importantly, we need to know if interventions such as this are sustained over time. Wright's well‐established medical education research group's study design assessed the impact of an intervention on physician performance and moves us closer to assessment of the impact on actual patient outcomes. As clinical teachers, we would like to believe that our teaching and our educational innovations are having a positive impact on patient care. Can we demonstrate this?

As academic medical centers contend with further resident work‐hour restrictions proposed by the Institute of Medicine (IOM),10 how will this affect hospitalist teachers? The study by Mazotti et al.3 from the University of California at San Francisco residency program found that about one‐quarter of residents reported spending less time teaching after implementation of the Accreditation Council for Graduate Medical Education (ACGME) duty‐hour restrictions in 2003. Interestingly, those residents reporting less time spent teaching also reported less emotional exhaustion and perceived that they were delivering higher‐quality patient care. This raises a fascinating question for academic hospitalists. Would these findings be similar among teaching hospitalists and nonteaching hospitalists? What about hospitalists who rotate through months of teaching and nonteaching services? Is teaching emotionally exhausting for experienced teachers? A Mayo Clinic study suggested that the extent that faculty physicians are able to engage in work that is most meaningful to them as individuals is a strong determinant of faculty burnout.11 Is the hospitalist who finds teaching most rewarding at risk of burnout if they are assigned only 2 weeks a year as a teaching attending? The answers to these questions will be critical to hospitalist program leaders trying to assure sustainable careers for hospitalists in their programs.

Although the study by Mazotti et al.3 did not assess the impact of the reduction in resident teaching time on the teaching responsibilities for academic hospitalists, previous studies suggest that faculty are also teaching less since the introduction of work‐hour restrictions.12, 13 If the new IOM recommendations are enacted, who will teach? Although the reported experience following the 2003 work‐hour restrictions begs pessimism, the anticipated changes represent an opportunity for creative hospitalist teachers to demonstrate effective adaptations to the changing and compressed inpatient teaching environment.

In summary, this issue of the Journal presents studies that praise the role hospitalists play in teaching the next generation of physicians, but also gives a glimpse of future challenges and opportunities. We should take advantage of hospitalists' central position in clinical education in the hospital to innovate, study the effect on both learner outcomes and patient outcomes, and share our experiences with the hospitalist and medical education communities.

Hospitalists are increasingly assuming a primary role in medical education in the hospital setting, as they also steadily care for a larger portion of hospitalized patients.1 This issue of the Journal of Hospital Medicine highlights the role of hospitalists as teachers in academic medical centers, confirming their expanding and positive role in resident and medical student education. A survey of academic medical centers, a systematic review, an evaluation of the implementation of an educational curriculum, and a survey of residents hint at the challenges hospitalists face in teaching, but also expose us to a more advanced yet facile approach to evaluating the effectiveness of a teaching intervention.25 These publications provoke interesting questions about clinical teaching that hospital medicine educators and researchers should pursue answering. I believe they will also encourage us to innovate in medical education and assessment of that teaching.

Traditionally, teaching attendings for resident teams on medicine or pediatric services rotated through these duties for 1 to 3 months each year, while spending the majority of their time in clinic or research activities. The increasing complexity of hospitalized patients and the pressure to reduce length of stay prompted closer oversight of trainees. With the advent of resident work‐hour restrictions, the need for greater clinical involvement by attending physicians made it increasingly difficult to maintain the traditional model of limited engagement by faculty attendings. Simply put, the dwindling pool of willing and able teaching attendings encouraged teaching hospitals to employ hospitalists to fill the gap in teaching and supervision, as well as clinical coverage.6

Beasley et al.2 report that resident work‐hour restrictions were associated with an increase in the number of teaching hospitals employing hospitalists to 79% of 193 surveyed hospitals in 2007. Of those hospitals with hospitalists, 92% reported that hospitalists serve as attendings on the teaching service. Hospitalists also teach in a number of other venues within these programs, including formal teaching rounds without direct care responsibility, along with delivering didactic lectures and clinical skills education.

How well are teaching hospitalists performing compared to traditional teaching attendings? Natarajan et al.4 provide an important summary of the evidence in a systematic review of studies comparing teaching efforts of hospitalist attendings to those of nonhospitalist attendings. Eight studies from a variety of institutions measured trainee (resident or medical student) attitudes. It is gratifying to learn that hospitalists were generally rated higher at overall teaching effectiveness, provision of feedback, knowledge base, and involvement of the learner in patient care. It seems likely that publication bias would overestimate the positive effect of hospitalists on learner attitudes. However, there are plausible reasons that the positive effect is accurate. Because their professional responsibilities are focused in the hospital, hospitalists should naturally be more available to learners for teaching and feedback. Hospitalists tend to be younger in their academic careers, placing them closer to the cutting edge of knowledge gained during residency and possibly fellowship. They may be more in tune with the needs and pressures faced by their learners, having dealt with these same challenges either during recent training or during nonteaching rotations.

As a relatively young specialty with young and developing academic hospitalists, will the advantage suggested by the Natarajan et al.4 systematic review be sustained over the long term as careers in hospital medicine mature? A 2005 systematic review studying this question among practicing clinicians found, somewhat paradoxically, that older, more experienced clinicians appeared to be at risk for providing lower‐quality care.7 To avoid this decline in clinical effectiveness, hospitalists should proactively seek innovative ways to refresh and update their knowledge and skills throughout their careers. This is particularly critical for teaching physicians. We should seize the opportunity to study the relationship between advancing clinical/teaching experience and educational quality within our teaching programs.

The review by Natarajan et al.4 should also challenge the hospitalist community to achieve even higher levels of proficiency as teachers of medicine. The review alludes to bedside teaching and attention to psychosocial aspects of care as opportunities for improvement by hospitalist teachers. A recent study suggested that physical examination instruction receives declining attention from inpatient teachers and that there are opportunities to increase the amount of bedside teaching.8 A provocative study of inpatients admitted to a teaching service found that physical examination could substantially impact patient care, but that trainees often failed to appreciate significant findings on initial examination.9 How do teaching hospitalists become proficient at physical examination and bedside teaching? Are there models around the country that are successfully developing outstanding clinician educators, incorporating teach‐the‐teacher models to improve physical examination and bedside teaching?

A practical limitation of attitude surveys and learner evaluation is the well‐known phenomenon of grade inflation that resulted in high ratings for all attending groups in the studies summarized by Natarajan et al.4 This limits the ability of surveys or evaluations to distinguish truly outstanding teachers and consequently makes it difficult to analyze the attributes of these teachers. We need better tools to detect and learn teaching techniques from great teachers in the clinical environment. We need studies evaluating the effect of teaching hospitalists on learner knowledge or, even more importantly, learner outcomes. Ultimately, we need studies of educational interventions that evaluate the impact of these interventions on patient outcome.

Wright et al.5 provide guidance as they describe the evaluation of a teaching intervention that moves beyond measurement of knowledge or attitudes. The Johns Hopkins Bayview hospitalist group sought to improve the quality of medical consultations performed by hospitalists and by residents rotating on the consultation service using a case‐based teaching module with audits of recent notes. The participants then audited their most recent consultation notes with feedback from the module teacher. The study employed pretests and posttests of knowledgea standard evaluation for educational interventions. This tells us little about the true impact of the teaching module. However, the study then assessed the quality of written consultations done by hospitalists before and after the educational interventions. Scores of consult notes improved significantly after the intervention, although the number of assessments for each physician was limited. Importantly, we need to know if interventions such as this are sustained over time. Wright's well‐established medical education research group's study design assessed the impact of an intervention on physician performance and moves us closer to assessment of the impact on actual patient outcomes. As clinical teachers, we would like to believe that our teaching and our educational innovations are having a positive impact on patient care. Can we demonstrate this?

As academic medical centers contend with further resident work‐hour restrictions proposed by the Institute of Medicine (IOM),10 how will this affect hospitalist teachers? The study by Mazotti et al.3 from the University of California at San Francisco residency program found that about one‐quarter of residents reported spending less time teaching after implementation of the Accreditation Council for Graduate Medical Education (ACGME) duty‐hour restrictions in 2003. Interestingly, those residents reporting less time spent teaching also reported less emotional exhaustion and perceived that they were delivering higher‐quality patient care. This raises a fascinating question for academic hospitalists. Would these findings be similar among teaching hospitalists and nonteaching hospitalists? What about hospitalists who rotate through months of teaching and nonteaching services? Is teaching emotionally exhausting for experienced teachers? A Mayo Clinic study suggested that the extent that faculty physicians are able to engage in work that is most meaningful to them as individuals is a strong determinant of faculty burnout.11 Is the hospitalist who finds teaching most rewarding at risk of burnout if they are assigned only 2 weeks a year as a teaching attending? The answers to these questions will be critical to hospitalist program leaders trying to assure sustainable careers for hospitalists in their programs.

Although the study by Mazotti et al.3 did not assess the impact of the reduction in resident teaching time on the teaching responsibilities for academic hospitalists, previous studies suggest that faculty are also teaching less since the introduction of work‐hour restrictions.12, 13 If the new IOM recommendations are enacted, who will teach? Although the reported experience following the 2003 work‐hour restrictions begs pessimism, the anticipated changes represent an opportunity for creative hospitalist teachers to demonstrate effective adaptations to the changing and compressed inpatient teaching environment.

In summary, this issue of the Journal presents studies that praise the role hospitalists play in teaching the next generation of physicians, but also gives a glimpse of future challenges and opportunities. We should take advantage of hospitalists' central position in clinical education in the hospital to innovate, study the effect on both learner outcomes and patient outcomes, and share our experiences with the hospitalist and medical education communities.