The hospitalist model of care has experienced dramatic growth. In 2003 it was estimated that there were 8000 US hospitalists, a number projected to ultimately reach more than 19 000.1, 2 This rapid growth has largely been driven by improvements in clinical efficiency as a result of hospitalist programs. There is a substantial body of evidence showing that hospitalists reduce length of stay and inpatient costs.3 Despite the rapid growth and proven benefit to clinical efficiency, no studies have evaluated the type and frequency of activities that hospitalists perform during routine work. Although the use of hospitalists improves clinical efficiency for the hospital, relatively little is known about how the hospital can improve efficiency for the hospitalist.

Our institution greatly expanded our hospitalist program in June 2003 to create a resident‐uncovered hospitalist service. The impetus for this change was the need to comply with newly revised Accreditation Council for Graduate Medicine Education (ACGME) program requirements regarding resident duty hours. Many teaching hospitals have implemented similar resident‐uncovered hospitalist services.4 Inefficiencies in their work activities quickly became apparent to our hospitalists. Furthermore, our hospitalists believed that they frequently performed simultaneous activities and that they were excessively interrupted by pages.

To evaluate the type and frequency of activities that the hospitalists performed during routine work, we performed a time‐motion study of hospitalist physicians on the resident‐uncovered hospitalist service. Our goal was to identify areas for systems improvements and activities that were better suited for nonphysician providers and to quantify the time spent multitasking and the frequency of paging interruptions.

METHODS

Northwestern Memorial Hospital (NMH) is a 753‐bed hospital in Chicago, Illinois. NMH is the primary teaching hospital affiliated with the Feinberg School of Medicine of Northwestern University. There are 2 general medicine services at NMH: a traditional resident‐covered ward service and the resident‐uncovered hospitalist service. Patients are admitted to one of these 2 services on the basis of, in order of importance, capacity of the services, preference of the outpatient physician, and potential educational value of the admission. Patients admitted to the hospitalist service are preferentially given beds on specific wards intended for hospitalist service patients. Fourth‐year medical students are frequently paired with hospitalists during their medicine subinternship.

The resident‐uncovered hospitalist service comprises 5 daytime hospitalists on duty at a time. The hospitalists are on service for 7 consecutive days, usually followed by 7 consecutive days off. Hospitalists pick up new patients from the night float hospitalist each morning. Daytime admitting duties rotate on a daily basis. One hospitalist accepts new admissions each morning from 7:00 _AM until noon. Two hospitalists accept admissions from noon until 5:00 _PM. One hospitalist accepts admissions from 5:00 _PM until 9:00 _PM. One hospitalist is free from accepting new admissions each day. All daytime hospitalists begin the workday at 7:00 AM and leave when their duties are completed for the day. One night float hospitalist is on duty each night of the week. The night float hospitalist performs admissions and all cross cover activities from 7:00 _PM until 7:00 _AM.

We first conducted a pilot study to help identify specific activities that our hospitalists routinely perform. Broad categories and subcategories of activities were created based on the results of our pilot study, and a published time‐motion study performed on emergency medicine physicians5 (Table 1). Once activities were defined and codes established, our research assistant unobtrusively shadowed hospitalist physicians for periods lasting 3‐5 hours. The observation periods were distributed in order to sample all activities that a daytime hospitalist would perform throughout a typical week. Observation periods included 2 morning admitting periods, 4 morning nonadmitting periods, 4 afternoon admitting periods, 4 afternoon nonadmitting periods, and 2 admitting periods from 5:00 _PM to 9:00 _PM. Activities were recorded on a standardized data collection form in 1‐minute intervals. When multiple activities were performed at the same time, all activities were recorded in the same 1‐minute interval. Incoming pages were recorded as well. To minimize the possibility that observation would affect hospitalist behavior, the research assistant was instructed not to initiate conversation with the hospitalists.

The data collection forms were manually abstracted and minutes tallied for each category and subcategory, for which summary statistics were converted to percentage of total minutes.

RESULTS

Ten hospitalists were shadowed by a single research assistant for a total of 4467 minutes. Seven hospitalists were male and 3 were female. The hospitalists were a mean age of 31 1.6 years of age and had been practicing as a hospitalist for a mean of 2.1 1.0 years. The hospitalists saw an average of 9.4 4.0 patients on the days they were shadowed by the research assistant. Because simultaneous activities were recorded, a total of 5557 minutes of activities were recorded.

The distribution of total minutes recorded in each activity category is shown in Figure 1. Hospitalists spent 18% of their time doing direct patient care, 69% on indirect patient care, 4% on personal activities, and 3% each on professional development, education, and travel.

Figure 1

Of the time hospitalists directly cared for patients, 18% was spent obtaining histories and performing physical examinations on new patients, 53% seeing patients in follow‐up visits, 16% going over discharge instructions, and 13% in family meetings (Figure 2). Of the time hospitalists spent doing indirect patient care, 37% was taken up by documentation, 21% by reviewing results, 7% by orders, and 35% by communication (Figure 2).

Figure 2

As just explained, communication accounted for 35% of indirect patient care activities; it also accounted for 24% of the total activity minutes. The time spent by hospitalists on communication was further broken down as 23% paging other physicians, 31% returning pages, 34% in face‐to‐face communication, 5% taking report on new admissions, 4% on sign‐out to the night float hospitalist, and 3% receiving sign‐out from the night float hospitalist.

Multitasking, performing more than 1 activity at the same time, was done 21% of the time. Hospitalists received an average of 3.4 1.5 pages per hour, and 7% of total activity time was spent returning pages. Other forms of interruption were not evaluated.

DISCUSSION

Our study had several important findings. First, hospitalists spent most of their time on indirect patient care activities and relatively little time on direct patient care. Time‐motion studies of nonhospitalist physicians have reported similar findings.5, 6 A considerable amount of hospitalist time was spent on documentation. This finding also has been reported in studies of nonhospitalist physicians.5, 7

A unique finding in our study was the large amount of time, 24% of total minutes, spent on communication. A study of emergency medicine physicians by Hollingsworth found that 13% of their time was spent on communication activities.5 The large amount of time spent on communication in our study underscores the need for hospitalists to have outstanding communication skills and systems that support efficient communication. Hospitalists spent 6% of their total time paging other physicians and 7% returning pages. Improvements in the efficiency of paging communication could greatly reduce the amount of time communicating by page. Our paging system provides unidirectional alphanumeric paging. In an effort to improve the efficiency of paging, we have asked nurses and consultants to include FYI and callback in the text of the page so it is clear whether the person who has paged the hospitalist needs to be called back. This simple solution to help reduce the number of unnecessary callbacks has previously been proposed by others.8

Another part of solving this problem is adopting the use of 2‐way pagers instead of alphanumeric pagers. Two‐way paging can increase the efficiency of communication even further. For example, a nurse sends a hospitalist a page that asks if the previous diet orders for a patient just returned from a procedure can be resumed. This hospitalist is on another floor in another patient's room. Rather than spending time leaving the other patient's room, finding a phone, calling the floor, waiting for an answer, and then waiting on hold, the hospitalist simply texts a 1‐word answer, Yes, in the 2‐way paging system. In addition to the time occupied by paging activities, hospitalists spent a large amount of time in face‐to‐face communication (8% of total activity time). On the one hand, having hospitalists discuss patient care with consultants and nurses in person on an ongoing basis throughout the day may improve clinical efficiency. On the other hand, the constant potential for interruption may be problematic. Similarly, 2‐way paging could facilitate communication to such a degree that it could actually increase the frequency of interruptions. Research on improvements in communications systems, interventions to improve communication skills, and team‐based care is warranted in order to evaluate the impact on hospitalist workflow.

An important finding in our study was that multitasking and paging interruptions were common. Although this may come as no surprise to practicing hospitalists, the distraction caused by interruptions and multitasking is an important potential cause of medical errors.911 A thorough examination of the types of activities performed simultaneously and whether they contributed to medical error was beyond the scope of our study. Some activities, such as documenting a note on a patient while reviewing the patient's lab results, are concordant (ie, conducted for the same patient) and therefore may be unlikely to contribute to medical error. Other combinations of activities, such as returning a page about one patient while documenting a note on a different patient or having face‐to‐face communication about one patient while entering an order on another patient, are discordant. Discordant activities may contribute to medical error. Further research of the effect of hospitalist multitasking and interruption on medical error is warranted and should be conducted within the framework of concordant versus discordant activities.

We had hoped to find activities that could be performed by non‐physician providers. No high impact activities were discovered that would be better suited for a non‐physician provider in this study. Clerical tasks, such as calling for radiology orders or obtaining medical records, amounted to a small percentage of hospitalist time (less than 1% combined). We did identify several activities in which automation or process improvement would be helpful. Hospitalists spent 5% of time on the combined activity of documenting discharge instructions and writing out prescriptions. Our institution is in the process of implementing an electronic medical record and computerized physician order entry. We are currently working on an automated process to generate printed discharge instructions and prescriptions. This has the potential not only to improve efficiency, but also to eliminate medication errors, as care is transitioned to the outpatient setting.

Our study had several limitations. First, our findings reflect the experience at one institution. Hospitalist practices vary widely in their staffing and scheduling models as well as in their organizational support. The amount of time that hospitalists spend on activities may differ between practices and between individual hospitalists in the same practice. Another limitation to our study pertains to the workflow of our hospitalists and the locations of their patients. As discussed earlier, patients were assigned to a hospitalist according to time of admission, not location of admission. Because of this, the hospitalists were caring for patients on as many as 5 wards. Although travel time amounted to only 3% of total minutes, it is possible that communication time could have been reduced if patients were distributed to hospitalists on the basis of patient location rather than time of admission of patient. For example, physicians and nurses might spend less time communicating in person compared to communicating via unidirectional paging, which frequently requires waiting for a callback. Finally, our study only observed activities performed by the daytime hospitalists at our hospital. The distribution and types of activities performed by nighttime hospitalists may be somewhat different.

Our study may serve as a model for hospitalist time‐motion studies in other settings. Our findings are of particular importance to resident‐uncovered hospitalist programs in academic hospitals, a setting in which operational inefficiencies may be abundant as house staff members have been poorly positioned in the hospital organization to lobby for process change. We hope that our study is a precursor to research evaluating modifications to the environments and systems in which hospitalists work. Such modifications have the potential to improve productivity and work conditions and promote career satisfaction.

The hospitalist model of care has experienced dramatic growth. In 2003 it was estimated that there were 8000 US hospitalists, a number projected to ultimately reach more than 19 000.1, 2 This rapid growth has largely been driven by improvements in clinical efficiency as a result of hospitalist programs. There is a substantial body of evidence showing that hospitalists reduce length of stay and inpatient costs.3 Despite the rapid growth and proven benefit to clinical efficiency, no studies have evaluated the type and frequency of activities that hospitalists perform during routine work. Although the use of hospitalists improves clinical efficiency for the hospital, relatively little is known about how the hospital can improve efficiency for the hospitalist.

Our institution greatly expanded our hospitalist program in June 2003 to create a resident‐uncovered hospitalist service. The impetus for this change was the need to comply with newly revised Accreditation Council for Graduate Medicine Education (ACGME) program requirements regarding resident duty hours. Many teaching hospitals have implemented similar resident‐uncovered hospitalist services.4 Inefficiencies in their work activities quickly became apparent to our hospitalists. Furthermore, our hospitalists believed that they frequently performed simultaneous activities and that they were excessively interrupted by pages.

To evaluate the type and frequency of activities that the hospitalists performed during routine work, we performed a time‐motion study of hospitalist physicians on the resident‐uncovered hospitalist service. Our goal was to identify areas for systems improvements and activities that were better suited for nonphysician providers and to quantify the time spent multitasking and the frequency of paging interruptions.

METHODS

Northwestern Memorial Hospital (NMH) is a 753‐bed hospital in Chicago, Illinois. NMH is the primary teaching hospital affiliated with the Feinberg School of Medicine of Northwestern University. There are 2 general medicine services at NMH: a traditional resident‐covered ward service and the resident‐uncovered hospitalist service. Patients are admitted to one of these 2 services on the basis of, in order of importance, capacity of the services, preference of the outpatient physician, and potential educational value of the admission. Patients admitted to the hospitalist service are preferentially given beds on specific wards intended for hospitalist service patients. Fourth‐year medical students are frequently paired with hospitalists during their medicine subinternship.

The resident‐uncovered hospitalist service comprises 5 daytime hospitalists on duty at a time. The hospitalists are on service for 7 consecutive days, usually followed by 7 consecutive days off. Hospitalists pick up new patients from the night float hospitalist each morning. Daytime admitting duties rotate on a daily basis. One hospitalist accepts new admissions each morning from 7:00 _AM until noon. Two hospitalists accept admissions from noon until 5:00 _PM. One hospitalist accepts admissions from 5:00 _PM until 9:00 _PM. One hospitalist is free from accepting new admissions each day. All daytime hospitalists begin the workday at 7:00 AM and leave when their duties are completed for the day. One night float hospitalist is on duty each night of the week. The night float hospitalist performs admissions and all cross cover activities from 7:00 _PM until 7:00 _AM.

We first conducted a pilot study to help identify specific activities that our hospitalists routinely perform. Broad categories and subcategories of activities were created based on the results of our pilot study, and a published time‐motion study performed on emergency medicine physicians5 (Table 1). Once activities were defined and codes established, our research assistant unobtrusively shadowed hospitalist physicians for periods lasting 3‐5 hours. The observation periods were distributed in order to sample all activities that a daytime hospitalist would perform throughout a typical week. Observation periods included 2 morning admitting periods, 4 morning nonadmitting periods, 4 afternoon admitting periods, 4 afternoon nonadmitting periods, and 2 admitting periods from 5:00 _PM to 9:00 _PM. Activities were recorded on a standardized data collection form in 1‐minute intervals. When multiple activities were performed at the same time, all activities were recorded in the same 1‐minute interval. Incoming pages were recorded as well. To minimize the possibility that observation would affect hospitalist behavior, the research assistant was instructed not to initiate conversation with the hospitalists.

The data collection forms were manually abstracted and minutes tallied for each category and subcategory, for which summary statistics were converted to percentage of total minutes.

RESULTS

Ten hospitalists were shadowed by a single research assistant for a total of 4467 minutes. Seven hospitalists were male and 3 were female. The hospitalists were a mean age of 31 1.6 years of age and had been practicing as a hospitalist for a mean of 2.1 1.0 years. The hospitalists saw an average of 9.4 4.0 patients on the days they were shadowed by the research assistant. Because simultaneous activities were recorded, a total of 5557 minutes of activities were recorded.

The distribution of total minutes recorded in each activity category is shown in Figure 1. Hospitalists spent 18% of their time doing direct patient care, 69% on indirect patient care, 4% on personal activities, and 3% each on professional development, education, and travel.

Figure 1

Of the time hospitalists directly cared for patients, 18% was spent obtaining histories and performing physical examinations on new patients, 53% seeing patients in follow‐up visits, 16% going over discharge instructions, and 13% in family meetings (Figure 2). Of the time hospitalists spent doing indirect patient care, 37% was taken up by documentation, 21% by reviewing results, 7% by orders, and 35% by communication (Figure 2).

Figure 2

As just explained, communication accounted for 35% of indirect patient care activities; it also accounted for 24% of the total activity minutes. The time spent by hospitalists on communication was further broken down as 23% paging other physicians, 31% returning pages, 34% in face‐to‐face communication, 5% taking report on new admissions, 4% on sign‐out to the night float hospitalist, and 3% receiving sign‐out from the night float hospitalist.

Multitasking, performing more than 1 activity at the same time, was done 21% of the time. Hospitalists received an average of 3.4 1.5 pages per hour, and 7% of total activity time was spent returning pages. Other forms of interruption were not evaluated.

DISCUSSION

Our study had several important findings. First, hospitalists spent most of their time on indirect patient care activities and relatively little time on direct patient care. Time‐motion studies of nonhospitalist physicians have reported similar findings.5, 6 A considerable amount of hospitalist time was spent on documentation. This finding also has been reported in studies of nonhospitalist physicians.5, 7

A unique finding in our study was the large amount of time, 24% of total minutes, spent on communication. A study of emergency medicine physicians by Hollingsworth found that 13% of their time was spent on communication activities.5 The large amount of time spent on communication in our study underscores the need for hospitalists to have outstanding communication skills and systems that support efficient communication. Hospitalists spent 6% of their total time paging other physicians and 7% returning pages. Improvements in the efficiency of paging communication could greatly reduce the amount of time communicating by page. Our paging system provides unidirectional alphanumeric paging. In an effort to improve the efficiency of paging, we have asked nurses and consultants to include FYI and callback in the text of the page so it is clear whether the person who has paged the hospitalist needs to be called back. This simple solution to help reduce the number of unnecessary callbacks has previously been proposed by others.8

Another part of solving this problem is adopting the use of 2‐way pagers instead of alphanumeric pagers. Two‐way paging can increase the efficiency of communication even further. For example, a nurse sends a hospitalist a page that asks if the previous diet orders for a patient just returned from a procedure can be resumed. This hospitalist is on another floor in another patient's room. Rather than spending time leaving the other patient's room, finding a phone, calling the floor, waiting for an answer, and then waiting on hold, the hospitalist simply texts a 1‐word answer, Yes, in the 2‐way paging system. In addition to the time occupied by paging activities, hospitalists spent a large amount of time in face‐to‐face communication (8% of total activity time). On the one hand, having hospitalists discuss patient care with consultants and nurses in person on an ongoing basis throughout the day may improve clinical efficiency. On the other hand, the constant potential for interruption may be problematic. Similarly, 2‐way paging could facilitate communication to such a degree that it could actually increase the frequency of interruptions. Research on improvements in communications systems, interventions to improve communication skills, and team‐based care is warranted in order to evaluate the impact on hospitalist workflow.

An important finding in our study was that multitasking and paging interruptions were common. Although this may come as no surprise to practicing hospitalists, the distraction caused by interruptions and multitasking is an important potential cause of medical errors.911 A thorough examination of the types of activities performed simultaneously and whether they contributed to medical error was beyond the scope of our study. Some activities, such as documenting a note on a patient while reviewing the patient's lab results, are concordant (ie, conducted for the same patient) and therefore may be unlikely to contribute to medical error. Other combinations of activities, such as returning a page about one patient while documenting a note on a different patient or having face‐to‐face communication about one patient while entering an order on another patient, are discordant. Discordant activities may contribute to medical error. Further research of the effect of hospitalist multitasking and interruption on medical error is warranted and should be conducted within the framework of concordant versus discordant activities.

We had hoped to find activities that could be performed by non‐physician providers. No high impact activities were discovered that would be better suited for a non‐physician provider in this study. Clerical tasks, such as calling for radiology orders or obtaining medical records, amounted to a small percentage of hospitalist time (less than 1% combined). We did identify several activities in which automation or process improvement would be helpful. Hospitalists spent 5% of time on the combined activity of documenting discharge instructions and writing out prescriptions. Our institution is in the process of implementing an electronic medical record and computerized physician order entry. We are currently working on an automated process to generate printed discharge instructions and prescriptions. This has the potential not only to improve efficiency, but also to eliminate medication errors, as care is transitioned to the outpatient setting.

Our study had several limitations. First, our findings reflect the experience at one institution. Hospitalist practices vary widely in their staffing and scheduling models as well as in their organizational support. The amount of time that hospitalists spend on activities may differ between practices and between individual hospitalists in the same practice. Another limitation to our study pertains to the workflow of our hospitalists and the locations of their patients. As discussed earlier, patients were assigned to a hospitalist according to time of admission, not location of admission. Because of this, the hospitalists were caring for patients on as many as 5 wards. Although travel time amounted to only 3% of total minutes, it is possible that communication time could have been reduced if patients were distributed to hospitalists on the basis of patient location rather than time of admission of patient. For example, physicians and nurses might spend less time communicating in person compared to communicating via unidirectional paging, which frequently requires waiting for a callback. Finally, our study only observed activities performed by the daytime hospitalists at our hospital. The distribution and types of activities performed by nighttime hospitalists may be somewhat different.

Our study may serve as a model for hospitalist time‐motion studies in other settings. Our findings are of particular importance to resident‐uncovered hospitalist programs in academic hospitals, a setting in which operational inefficiencies may be abundant as house staff members have been poorly positioned in the hospital organization to lobby for process change. We hope that our study is a precursor to research evaluating modifications to the environments and systems in which hospitalists work. Such modifications have the potential to improve productivity and work conditions and promote career satisfaction.

The hospitalist model of care has experienced dramatic growth. In 2003 it was estimated that there were 8000 US hospitalists, a number projected to ultimately reach more than 19 000.1, 2 This rapid growth has largely been driven by improvements in clinical efficiency as a result of hospitalist programs. There is a substantial body of evidence showing that hospitalists reduce length of stay and inpatient costs.3 Despite the rapid growth and proven benefit to clinical efficiency, no studies have evaluated the type and frequency of activities that hospitalists perform during routine work. Although the use of hospitalists improves clinical efficiency for the hospital, relatively little is known about how the hospital can improve efficiency for the hospitalist.

Our institution greatly expanded our hospitalist program in June 2003 to create a resident‐uncovered hospitalist service. The impetus for this change was the need to comply with newly revised Accreditation Council for Graduate Medicine Education (ACGME) program requirements regarding resident duty hours. Many teaching hospitals have implemented similar resident‐uncovered hospitalist services.4 Inefficiencies in their work activities quickly became apparent to our hospitalists. Furthermore, our hospitalists believed that they frequently performed simultaneous activities and that they were excessively interrupted by pages.

To evaluate the type and frequency of activities that the hospitalists performed during routine work, we performed a time‐motion study of hospitalist physicians on the resident‐uncovered hospitalist service. Our goal was to identify areas for systems improvements and activities that were better suited for nonphysician providers and to quantify the time spent multitasking and the frequency of paging interruptions.

METHODS

Northwestern Memorial Hospital (NMH) is a 753‐bed hospital in Chicago, Illinois. NMH is the primary teaching hospital affiliated with the Feinberg School of Medicine of Northwestern University. There are 2 general medicine services at NMH: a traditional resident‐covered ward service and the resident‐uncovered hospitalist service. Patients are admitted to one of these 2 services on the basis of, in order of importance, capacity of the services, preference of the outpatient physician, and potential educational value of the admission. Patients admitted to the hospitalist service are preferentially given beds on specific wards intended for hospitalist service patients. Fourth‐year medical students are frequently paired with hospitalists during their medicine subinternship.

The resident‐uncovered hospitalist service comprises 5 daytime hospitalists on duty at a time. The hospitalists are on service for 7 consecutive days, usually followed by 7 consecutive days off. Hospitalists pick up new patients from the night float hospitalist each morning. Daytime admitting duties rotate on a daily basis. One hospitalist accepts new admissions each morning from 7:00 _AM until noon. Two hospitalists accept admissions from noon until 5:00 _PM. One hospitalist accepts admissions from 5:00 _PM until 9:00 _PM. One hospitalist is free from accepting new admissions each day. All daytime hospitalists begin the workday at 7:00 AM and leave when their duties are completed for the day. One night float hospitalist is on duty each night of the week. The night float hospitalist performs admissions and all cross cover activities from 7:00 _PM until 7:00 _AM.

We first conducted a pilot study to help identify specific activities that our hospitalists routinely perform. Broad categories and subcategories of activities were created based on the results of our pilot study, and a published time‐motion study performed on emergency medicine physicians5 (Table 1). Once activities were defined and codes established, our research assistant unobtrusively shadowed hospitalist physicians for periods lasting 3‐5 hours. The observation periods were distributed in order to sample all activities that a daytime hospitalist would perform throughout a typical week. Observation periods included 2 morning admitting periods, 4 morning nonadmitting periods, 4 afternoon admitting periods, 4 afternoon nonadmitting periods, and 2 admitting periods from 5:00 _PM to 9:00 _PM. Activities were recorded on a standardized data collection form in 1‐minute intervals. When multiple activities were performed at the same time, all activities were recorded in the same 1‐minute interval. Incoming pages were recorded as well. To minimize the possibility that observation would affect hospitalist behavior, the research assistant was instructed not to initiate conversation with the hospitalists.

The data collection forms were manually abstracted and minutes tallied for each category and subcategory, for which summary statistics were converted to percentage of total minutes.

RESULTS

Ten hospitalists were shadowed by a single research assistant for a total of 4467 minutes. Seven hospitalists were male and 3 were female. The hospitalists were a mean age of 31 1.6 years of age and had been practicing as a hospitalist for a mean of 2.1 1.0 years. The hospitalists saw an average of 9.4 4.0 patients on the days they were shadowed by the research assistant. Because simultaneous activities were recorded, a total of 5557 minutes of activities were recorded.

The distribution of total minutes recorded in each activity category is shown in Figure 1. Hospitalists spent 18% of their time doing direct patient care, 69% on indirect patient care, 4% on personal activities, and 3% each on professional development, education, and travel.

Figure 1

Of the time hospitalists directly cared for patients, 18% was spent obtaining histories and performing physical examinations on new patients, 53% seeing patients in follow‐up visits, 16% going over discharge instructions, and 13% in family meetings (Figure 2). Of the time hospitalists spent doing indirect patient care, 37% was taken up by documentation, 21% by reviewing results, 7% by orders, and 35% by communication (Figure 2).

Figure 2

As just explained, communication accounted for 35% of indirect patient care activities; it also accounted for 24% of the total activity minutes. The time spent by hospitalists on communication was further broken down as 23% paging other physicians, 31% returning pages, 34% in face‐to‐face communication, 5% taking report on new admissions, 4% on sign‐out to the night float hospitalist, and 3% receiving sign‐out from the night float hospitalist.

Multitasking, performing more than 1 activity at the same time, was done 21% of the time. Hospitalists received an average of 3.4 1.5 pages per hour, and 7% of total activity time was spent returning pages. Other forms of interruption were not evaluated.

DISCUSSION

Our study had several important findings. First, hospitalists spent most of their time on indirect patient care activities and relatively little time on direct patient care. Time‐motion studies of nonhospitalist physicians have reported similar findings.5, 6 A considerable amount of hospitalist time was spent on documentation. This finding also has been reported in studies of nonhospitalist physicians.5, 7

A unique finding in our study was the large amount of time, 24% of total minutes, spent on communication. A study of emergency medicine physicians by Hollingsworth found that 13% of their time was spent on communication activities.5 The large amount of time spent on communication in our study underscores the need for hospitalists to have outstanding communication skills and systems that support efficient communication. Hospitalists spent 6% of their total time paging other physicians and 7% returning pages. Improvements in the efficiency of paging communication could greatly reduce the amount of time communicating by page. Our paging system provides unidirectional alphanumeric paging. In an effort to improve the efficiency of paging, we have asked nurses and consultants to include FYI and callback in the text of the page so it is clear whether the person who has paged the hospitalist needs to be called back. This simple solution to help reduce the number of unnecessary callbacks has previously been proposed by others.8

Another part of solving this problem is adopting the use of 2‐way pagers instead of alphanumeric pagers. Two‐way paging can increase the efficiency of communication even further. For example, a nurse sends a hospitalist a page that asks if the previous diet orders for a patient just returned from a procedure can be resumed. This hospitalist is on another floor in another patient's room. Rather than spending time leaving the other patient's room, finding a phone, calling the floor, waiting for an answer, and then waiting on hold, the hospitalist simply texts a 1‐word answer, Yes, in the 2‐way paging system. In addition to the time occupied by paging activities, hospitalists spent a large amount of time in face‐to‐face communication (8% of total activity time). On the one hand, having hospitalists discuss patient care with consultants and nurses in person on an ongoing basis throughout the day may improve clinical efficiency. On the other hand, the constant potential for interruption may be problematic. Similarly, 2‐way paging could facilitate communication to such a degree that it could actually increase the frequency of interruptions. Research on improvements in communications systems, interventions to improve communication skills, and team‐based care is warranted in order to evaluate the impact on hospitalist workflow.

An important finding in our study was that multitasking and paging interruptions were common. Although this may come as no surprise to practicing hospitalists, the distraction caused by interruptions and multitasking is an important potential cause of medical errors.911 A thorough examination of the types of activities performed simultaneously and whether they contributed to medical error was beyond the scope of our study. Some activities, such as documenting a note on a patient while reviewing the patient's lab results, are concordant (ie, conducted for the same patient) and therefore may be unlikely to contribute to medical error. Other combinations of activities, such as returning a page about one patient while documenting a note on a different patient or having face‐to‐face communication about one patient while entering an order on another patient, are discordant. Discordant activities may contribute to medical error. Further research of the effect of hospitalist multitasking and interruption on medical error is warranted and should be conducted within the framework of concordant versus discordant activities.

We had hoped to find activities that could be performed by non‐physician providers. No high impact activities were discovered that would be better suited for a non‐physician provider in this study. Clerical tasks, such as calling for radiology orders or obtaining medical records, amounted to a small percentage of hospitalist time (less than 1% combined). We did identify several activities in which automation or process improvement would be helpful. Hospitalists spent 5% of time on the combined activity of documenting discharge instructions and writing out prescriptions. Our institution is in the process of implementing an electronic medical record and computerized physician order entry. We are currently working on an automated process to generate printed discharge instructions and prescriptions. This has the potential not only to improve efficiency, but also to eliminate medication errors, as care is transitioned to the outpatient setting.

Our study had several limitations. First, our findings reflect the experience at one institution. Hospitalist practices vary widely in their staffing and scheduling models as well as in their organizational support. The amount of time that hospitalists spend on activities may differ between practices and between individual hospitalists in the same practice. Another limitation to our study pertains to the workflow of our hospitalists and the locations of their patients. As discussed earlier, patients were assigned to a hospitalist according to time of admission, not location of admission. Because of this, the hospitalists were caring for patients on as many as 5 wards. Although travel time amounted to only 3% of total minutes, it is possible that communication time could have been reduced if patients were distributed to hospitalists on the basis of patient location rather than time of admission of patient. For example, physicians and nurses might spend less time communicating in person compared to communicating via unidirectional paging, which frequently requires waiting for a callback. Finally, our study only observed activities performed by the daytime hospitalists at our hospital. The distribution and types of activities performed by nighttime hospitalists may be somewhat different.

Our study may serve as a model for hospitalist time‐motion studies in other settings. Our findings are of particular importance to resident‐uncovered hospitalist programs in academic hospitals, a setting in which operational inefficiencies may be abundant as house staff members have been poorly positioned in the hospital organization to lobby for process change. We hope that our study is a precursor to research evaluating modifications to the environments and systems in which hospitalists work. Such modifications have the potential to improve productivity and work conditions and promote career satisfaction.

Aortic dissection is an uncommon but highly lethal disease with an incidence of approximately 2,000 cases per year in the United States.1 It is often mistaken for less serious pathology. In one series, aortic dissection was missed in 38% of patients at presentation, with 28% of patients first diagnosed at autopsy.2 Early recognition and management are crucial. If untreated, the mortality rate for acute aortic dissection increases by approximately 1% per hour over the first 48 hours and may reach 70% at 1 week. As many as 90% of untreated patients who suffer aortic dissection die within 3 months of presentation.3, 4 Generally, cardiothoracic surgeons or cardiologists experienced with managing aortic dissection should direct patient evaluation and treatment. Hospitalists, however, are increasingly assuming responsibility for the initial triage and management of patients with acute chest pain syndromes and therefore must be able to rapidly identify aortic dissection, initiate supportive therapy, and refer patients to appropriate specialty care.

PATHOPHYSIOLOGY

Aortic dissection occurs when layers of the aortic wall separate because of infiltration of high‐pressure arterial blood. The proximate causes are elevated shear stress across the aortic lumen in the setting of a concomitant defect in the aortic media. Shear stress is caused by the rapid increase in luminal pressure per unit of time (dP/dt) that results from cardiac systole. As the aorta traverses away from the heart, an increasing proportion of the kinetic energy of left ventricular systole is stored in the aortic wall as potential energy, which facilitates anterograde propagation of cardiac output during diastole. This conversion of kinetic to potential energy also attenuates shear stress. As the proximal aorta is subject to the steepest fluctuations in pressure, it is at the highest risk of dissection. Degeneration of the aortic media is part of the normal aging process but is accelerated in persons with a bicuspid aortic valve, Turner's syndrome, inflammatory arteritis, or inherited diseases of collagen formation.

Once the aortic intima is compromised, blood dissects longitudinally through the aortic media and propagates proximally or distally, creating a false lumen that may communicate with the true lumen of the aorta. Blood may flow through the true lumen, the false lumen, or both. Propagation of the dissection causes much of the morbidity associated with aortic dissection by disrupting blood flow across branch vessels or by directly compromising the pericardium or aortic valve. Over time, the dissection may traverse the entire aortic wall, causing aortic rupture and exsanguination.

CLASSIFICATION

Acute aortic dissection is classified as any aortic dissection diagnosed within 2 weeks of the onset of symptoms, which is the period of highest risk of mortality. Patients who survive more than 2 weeks without treatment are considered to have chronic dissection. Aortic dissections are further classified according to their anatomic location. The fundamental distinction is whether the dissection is proximal (involving the aortic root or ascending aorta) or distal (below the left subclavian artery). The Stanford and DeBakey classification systems are the classification systems most commonly used (Figure 1).

Figure 1

Some variants of aortic dissection are not described in either the Stanford or DeBakey systems. Aortic intramural hematomas (IMH) are caused by intramural hemorrhage of the vasa vasorum without an identifiable intimal tear.57 Penetrating atherosclerotic ulcers (PAUs) are focal defects in the aortic wall with surrounding hematoma but no longitudinal dissection across tissue planes, typically resulting from advanced atherosclerotic disease.8 The pathophysiologic distinctions between IMH, PAU, and classic aortic dissection remain somewhat controversial. Both IMH and PAU may progress to aortic aneurysm formation, frank dissection, or aortic rupture, suggesting that these entities represent a spectrum of diseases with broad overlap (Table 1).9, 10

EPIDEMIOLOGY

Aortic dissection is a rare disease, with an estimated incidence of approximately 5‐30 cases per 1 million people per year.1114 Fewer than 0.5% of patients presenting to an emergency department with chest or back pain suffer from aortic dissection.15 Two thirds of patients are male, with an average age at presentation of approximately 65 years. A history of systemic hypertension, found in up to 72% of patients, is by far the most common risk factor.2, 14, 16 Atherosclerosis, a history of prior cardiac surgery, and known aortic aneurysm are other major risk factors.14 The epidemiology of aortic dissection is substantially different in young patients (40 years of age). Hypertension and atherosclerosis become significantly less common, as other risk factors, such as Marfan syndrome, take precedence17 (Table 2). Other risk factors for aortic dissection include:

• Collagen diseases (eg, Marfan syndrome and Ehlers‐Danlos): In the International Registry of Acute Aortic Dissection (IRAD), the largest prospective analysis of aortic dissection to date, 50% of the young patients presenting with aortic dissection had Marfan syndrome.17

• Bicuspid aortic valve (BAV): Individuals with BAV are 5‐18 times more likely to suffer aortic dissection than those with a trileaflet valve.18, 19 In one survey, 52% of asymptomatic young men with BAV were found to have aortic root dilatation, a frequent precursor of dissection.20 Vascular tissue in individuals with BAV has been found to have increased levels of matrix metalloproteinases, which may degrade elastic matrix components and accelerate medial necrosis.21

• Aortic coarctation: Aortic coarctation is associated with upper extremity hypertension, BAV and aortic dilatation, all of which predispose to aortic dissection.

• Turner syndrome: Aortic root dilatation with or without dissection has been incidentally noted in 6%‐9% of patients with Turner syndrome.22, 23

• Strenuous exercise: Multiple case reports have associated aortic dissection with high‐intensity weightlifting. Many affected individuals were subsequently found to have at least one other risk factor, including hypertension, anabolic steroid abuse, and cocaine abuse.2426

• Large vessel arteritis: Large vessel arteritides, specifically giant cell arteritis, Takayasu's disease, and tertiary syphilis have long been associated with aortic dilatation and dissection.

• Cocaine and methamphetamine ingestion: Sympathomimetic drugs cause rapid increases in heart rate and blood pressure, markedly increasing aortic intraluminal shear stress. Furthermore, cocaine is thought to be directly toxic to vascular endothelium and may accelerate medial necrosis.2730

• Third trimester pregnancy, especially in patients with diseases of collagen31; The significance of pregnancy has recently been called into question by data from the IRAD trial. Of 346 enrolled women with aortic dissection, only 2 were pregnant, suggesting that the previously held association of pregnancy with aortic dissection may be an artifact of selective reporting.1

• Blunt chest trauma or high‐speed deceleration injury.

• Iatrogenic injury, typically from intra‐aortic catheterization.

INITIAL EVALUATION

The differential diagnosis for acute aortic dissection includes acute coronary syndrome, pulmonary embolus, pneumothorax, pneumonia, musculoskeletal pain, acute cholecystitis, esophageal spasm or rupture, acute pancreatitis, and acute pericarditis. Acute aortic dissections are rarely asymptomatic; in fact, the absence of sudden‐onset chest pain decreases the likelihood of dissection (negative LR 0.3).32 In the IRAD trial, approximately 95% of patients with aortic dissection complained of pain in the chest, back, or abdomen, with 90% characterizing their pain as either severe or the worst ever and 64% describing it as sharp.14 Although the presence of tearing or ripping chest or back pain suggests aortic dissection (positive LR 1.2‐10.8), its absence does not reliably exclude this diagnosis.32 The wide variability in the presentation of aortic dissection increases the challenge of establishing a diagnosis. Clinical findings depend largely on the anatomical location of the dissection and may include pulse deficits, neurologic deficits, hypotension, hypertension, and end‐organ ischemia. Women who develop aortic dissection are generally older and present later than men. Their symptoms are less typical and are likely to be confounded by altered mental status.1 A diagnosis of aortic dissection should be strongly considered for patients presenting with acute chest or back pain and otherwise unexplained aortic insufficiency, focal neurologic deficits, pulse deficits, or end‐organ injury (Table 3).

Electrocardiogram: Electrocardiographic abnormalities are commonly seen in aortic dissection and may include ST‐segment or T‐wave abnormalities or left ventricular hypertrophy.14 Proximal aortic dissections may compromise coronary artery perfusion, generating electrocardiogram (ECG) findings compatible with acute myocardial infarction, which may lead the clinician to diagnose and treat myocardial infarction while missing the underlying diagnosis.33 In a recent survey, 9 of 44 patients (21%) presenting with acute aortic dissection were initially diagnosed with acute coronary syndrome and anticoagulated, with 2 deaths.34 ECGs must therefore be interpreted with extreme caution in aortic dissection.

Chest x‐ray: In the emergency department, chest radiography is a mainstay of the evaluation of acute chest pain. Unfortunately, plain‐film radiography has limited utility for diagnosing aortic dissection.35 In the IRAD trial, mediastinal widening (>8 cm) and abnormal aortic contour, the classic radiographic findings in aortic dissection, were present in only 50%‐60% of cases. Twelve percent of patients had a completely normal chest x‐ray.14 A pooled analysis of previous studies demonstrated that the sensitivity of widened mediastinum and abnormal aortic contour was 65% and 71%, respectively.32 Nonspecific radiographic findings, most notably pleural effusion, were common.36 Thus, if the index of suspicion for aortic dissection is elevated, a confirmatory study must be obtained (Figure 2).

Figure 2

CONFIRMATORY IMAGING STUDIES

The ideal confirmatory imaging modality should identify aortic dissection with high sensitivity and specificity. It should also identify the entry and exit points of the dissection and provide information about the extent of compromise of the aortic valve, pericardium, and great vessels. Four imaging modalities sufficiently meet these criteria in order to be considered diagnostically useful.

Aortography: Previously the gold standard for diagnosing aortic dissection, aortography is no longer a first‐line imaging modality. The sensitivity and specificity of aortography are at best equivalent and probably inferior to less invasive imaging modalities.37, 38 False negatives may occur if both the true and false lumens opacify equally with contrast, or if the false lumen is sufficiently thrombosed to preclude any instillation of contrast. Aortography cannot identify aortic intramural hematomas, is invasive and highly operator dependent, requires nephrotoxic contrast, and generally takes longer to obtain than other modalities.39

Aortography uniquely offers excellent visualization of the coronary arteries and great vessels and is preferred when such information is necessary. Percutaneous aortic endovascular stent grafting has been recently employed to repair distal aortic dissections.4043 As a result, aortography is gaining new life as a therapeutic modality.

CT angiography: Spiral CT angiography (CTA) is the most commonly used modality for diagnosing aortic dissection.44 It is emergently available at most hospitals, and images can be obtained in minutes. Sensitivity and specificity may approach 100%, and CTA may be more sensitive than MRA or TEE in evaluating arch vessel involvement.4547 Like conventional angiography, CTA requires administration of nephrotoxic contrast. It frequently cannot visualize the entry and exit sites (intimal flaps) of a dissection and provides limited information about the coronary arteries and no information about the competency of the aortic valve.48, 49 Thus, if aortic dissection is identified by CTA, a second study may be needed to provide further diagnostic information and to guide surgical intervention (Figures 3 and 4).

Figure 3

Figure 4

Magnetic Resonance Angiography: Magnetic resonance angiography (MRA) offers excellent noninvasive evaluation of the thoracic aorta. Sensitivity and specificity are probably superior to spiral CTA, and MRA generally identifies the location of the intimal tear and provides some functional information about the aortic valve.44, 50, 51 MRA is not emergently available at many hospitals. Scanning is time intensive, requiring the patient to remain motionless and relatively inaccessible for up to an hour. Furthermore, patient claustrophobia and the presence of implanted devices such as pacemakers or ferromagnetic foreign bodies may preclude MRA.

Transesophageal echocardiography: The sensitivity and specificity of transesophageal echocardiography (TEE) are also excellenton a par with CTA and MRA. In addition to providing excellent visualization of the thoracic aorta, TEE provides superb images of the pericardium and detailed assessment of aortic valve function.52 It also is extremely effective at visualizing the aortic intimal flap.44, 49, 53 A significant advantage of TEE is its portability, allowing rapid diagnosis at the bedside. For this reason, it is particularly useful for evaluation of patients who are hemodynamically unstable and are suspected to have an aortic dissection. Because of the anatomic relationship of the aorta with the esophagus and the trachea, TEE more effectively identifies proximal than distal dissections.43 TEE is also somewhat invasive, usually requires patient sedation, and is highly operator dependent, requiring the availability of an experienced and technically skilled operator (Figure 5).

Figure 5

Transthoracic echocardiography: Although it is an excellent tool for the evaluation of many aspects of cardiac anatomy and function, surface echocardiography can reliably visualize only limited portions of the ascending and descending aorta.54, 55 As a consequence, it is neither sensitive nor specific enough to diagnose aortic dissection. Transthoracic echocardiography (TTE) does, however, play a role in rapidly assessing patients at the bedside for aortic valve or pericardial compromise when these complications are suspected.

Recommendations

CTA, MRA, and TEE are all highly sensitive and specific modalities for diagnosing aortic dissection. Therefore, the condition of the patient, the information needed, and the resources and expertise immediately available should drive the choice of study. MRA is considered the gold standard diagnostic study and is the preferred modality for hemodynamically stable patients with suspected aortic dissection. Because of slow data acquisition and the inaccessibility of patients in the scanner, it is generally unsuited for unstable patients, including those with ongoing pain. Bedside TEE is an excellent choice for patients who are too unstable for MRA but is less effective at visualizing distal dissections. Arch aortography is generally reserved for the confirmation of questionable diagnoses or to image specific branch arteries (Tables 4 and 5).

Table 4.Comparative Sensitivities of Imaging Modalities in Aortic Dissection

	Overall	Proximal	Distal
Adapted from Moore et al.43
TEE	88%	90%	80%
CTA	93%	93%	93%
MRA	100%	100%	100%
Aortogram	87%	87%	87%

Table 5.Comparative Diagnostic Utility of Imaging Techniques in Aortic Dissection

	TEE	CTA	MRA	Aortography
Adapted from recommendations of the task force on aortic dissection, European Society of Cardiology.61
Sensitivity	++	++	+++	++
Specificity	+++	++	+++	++
Classification	+++	++	++	+
Intimal flap	+++	‐	++	+
Aortic regurgitation	+++		++	++
Pericardial effusion	+++	++	++
Branch vessel involvement	+	++	++	+++
Coronary artery involvement	++	+	+	+++

Most trials comparing CTA, MRA, and TEE were performed in the early 1990s. Computed tomography has evolved significantly over the intervening decade, and some of the diagnostic limitations previously ascribed to CTA, such as the inability to generate 3‐D reconstructed images, no longer exist. Furthermore, CT angiography is widely available and is gaining increasing acceptance as a first‐line imaging modality for patients with noncardiac chest pain.48 Medical centers that maintain round‐the‐clock CT capability may have limited or delayed access to TEE, MRA, or aortography. Given the potential for rapid and dramatic patient deterioration, it is imperative that a diagnosis be established quickly when aortic dissection is suspected. Thus, when the choice is obtaining an immediate CTA or a delayed TEE or MRA, CTA is generally the better choice (Figure 6).

Figure 6

MANAGEMENT

PROGNOSIS

Despite significant medical and surgical advances, aortic dissection remains exceptionally lethal. Patients with proximal dissections are more likely to die than those with distal dissections. Using data from the IRAD trial, Mehta et al determined that age 70 years (OR, 1.70), abrupt onset of chest pain (OR 2.60), hypotension/shock/tamponade (OR, 2.97), renal failure (OR, 4.77), pulse deficit (OR, 2.03), and abnormal ECG (OR, 1.77) were independent determinants of death.59 Medical treatment of proximal dissection is generally reserved for patients too ill, unstable, or frail to undergo surgery. In contrast, most patients with distal dissection are managed medically, with surgery generally reserved for those with acute complications. Hence, patients with proximal dissections who are managed medically and those with distal dissections who are managed surgically have the worst outcomes. Outcomes for women are worse than those for men, which is probably attributable to several factors. Women dissect at an older age, present later after the onset of symptoms, and are more likely to have confounding symptoms that may delay timely diagnosis1 (Table 7).

CONCLUSION

Aortic dissection is a rare and acutely life‐threatening cause of acute chest and back pain. Delays in diagnosis and misdiagnoses are common, frequently with catastrophic consequences. The key to diagnosis is maintaining a high index of suspicion for dissection, especially in patients who present with acute severe chest, back, or abdominal pain in the setting of unexplained acute pulse deficits, neurologic deficits, or acute end‐organ injury. Three clinical findings have been shown to be diagnostically useful: immediate onset of tearing or ripping chest or back pain, mediastinal widening or abnormal aortic contour on chest radiograph, and peripheral pulse deficits or variable pulse pressure (>20 mm Hg). If all 3 findings are absent, acute aortic dissection is unlikely. The presence of any of these findings should prompt further workup. A normal chest radiograph does not rule out aortic dissection. Only TEE, CT, and MR angiography are sufficiently specific to rule out dissection. Aortography is rarely used as a first‐line diagnostic tool but may be useful as a confirmatory test or to provide additional anatomic information. Patients who present with proximal aortic dissection or with any aortic dissection with concomitant hypotension are at exceptionally high risk of death and should be immediately referred for surgical evaluation. Beta‐blockers are the mainstay of acute and chronic therapy of aortic dissection. Survivors of aortic dissection are at a markedly elevated risk for further aortic events and should be followed vigilantly posthospitalization.

Aortic dissection is an uncommon but highly lethal disease with an incidence of approximately 2,000 cases per year in the United States.1 It is often mistaken for less serious pathology. In one series, aortic dissection was missed in 38% of patients at presentation, with 28% of patients first diagnosed at autopsy.2 Early recognition and management are crucial. If untreated, the mortality rate for acute aortic dissection increases by approximately 1% per hour over the first 48 hours and may reach 70% at 1 week. As many as 90% of untreated patients who suffer aortic dissection die within 3 months of presentation.3, 4 Generally, cardiothoracic surgeons or cardiologists experienced with managing aortic dissection should direct patient evaluation and treatment. Hospitalists, however, are increasingly assuming responsibility for the initial triage and management of patients with acute chest pain syndromes and therefore must be able to rapidly identify aortic dissection, initiate supportive therapy, and refer patients to appropriate specialty care.

PATHOPHYSIOLOGY

Aortic dissection occurs when layers of the aortic wall separate because of infiltration of high‐pressure arterial blood. The proximate causes are elevated shear stress across the aortic lumen in the setting of a concomitant defect in the aortic media. Shear stress is caused by the rapid increase in luminal pressure per unit of time (dP/dt) that results from cardiac systole. As the aorta traverses away from the heart, an increasing proportion of the kinetic energy of left ventricular systole is stored in the aortic wall as potential energy, which facilitates anterograde propagation of cardiac output during diastole. This conversion of kinetic to potential energy also attenuates shear stress. As the proximal aorta is subject to the steepest fluctuations in pressure, it is at the highest risk of dissection. Degeneration of the aortic media is part of the normal aging process but is accelerated in persons with a bicuspid aortic valve, Turner's syndrome, inflammatory arteritis, or inherited diseases of collagen formation.

Once the aortic intima is compromised, blood dissects longitudinally through the aortic media and propagates proximally or distally, creating a false lumen that may communicate with the true lumen of the aorta. Blood may flow through the true lumen, the false lumen, or both. Propagation of the dissection causes much of the morbidity associated with aortic dissection by disrupting blood flow across branch vessels or by directly compromising the pericardium or aortic valve. Over time, the dissection may traverse the entire aortic wall, causing aortic rupture and exsanguination.

CLASSIFICATION

Acute aortic dissection is classified as any aortic dissection diagnosed within 2 weeks of the onset of symptoms, which is the period of highest risk of mortality. Patients who survive more than 2 weeks without treatment are considered to have chronic dissection. Aortic dissections are further classified according to their anatomic location. The fundamental distinction is whether the dissection is proximal (involving the aortic root or ascending aorta) or distal (below the left subclavian artery). The Stanford and DeBakey classification systems are the classification systems most commonly used (Figure 1).

Figure 1

Some variants of aortic dissection are not described in either the Stanford or DeBakey systems. Aortic intramural hematomas (IMH) are caused by intramural hemorrhage of the vasa vasorum without an identifiable intimal tear.57 Penetrating atherosclerotic ulcers (PAUs) are focal defects in the aortic wall with surrounding hematoma but no longitudinal dissection across tissue planes, typically resulting from advanced atherosclerotic disease.8 The pathophysiologic distinctions between IMH, PAU, and classic aortic dissection remain somewhat controversial. Both IMH and PAU may progress to aortic aneurysm formation, frank dissection, or aortic rupture, suggesting that these entities represent a spectrum of diseases with broad overlap (Table 1).9, 10

EPIDEMIOLOGY

Aortic dissection is a rare disease, with an estimated incidence of approximately 5‐30 cases per 1 million people per year.1114 Fewer than 0.5% of patients presenting to an emergency department with chest or back pain suffer from aortic dissection.15 Two thirds of patients are male, with an average age at presentation of approximately 65 years. A history of systemic hypertension, found in up to 72% of patients, is by far the most common risk factor.2, 14, 16 Atherosclerosis, a history of prior cardiac surgery, and known aortic aneurysm are other major risk factors.14 The epidemiology of aortic dissection is substantially different in young patients (40 years of age). Hypertension and atherosclerosis become significantly less common, as other risk factors, such as Marfan syndrome, take precedence17 (Table 2). Other risk factors for aortic dissection include:

• Collagen diseases (eg, Marfan syndrome and Ehlers‐Danlos): In the International Registry of Acute Aortic Dissection (IRAD), the largest prospective analysis of aortic dissection to date, 50% of the young patients presenting with aortic dissection had Marfan syndrome.17

• Bicuspid aortic valve (BAV): Individuals with BAV are 5‐18 times more likely to suffer aortic dissection than those with a trileaflet valve.18, 19 In one survey, 52% of asymptomatic young men with BAV were found to have aortic root dilatation, a frequent precursor of dissection.20 Vascular tissue in individuals with BAV has been found to have increased levels of matrix metalloproteinases, which may degrade elastic matrix components and accelerate medial necrosis.21

• Aortic coarctation: Aortic coarctation is associated with upper extremity hypertension, BAV and aortic dilatation, all of which predispose to aortic dissection.

• Turner syndrome: Aortic root dilatation with or without dissection has been incidentally noted in 6%‐9% of patients with Turner syndrome.22, 23

• Strenuous exercise: Multiple case reports have associated aortic dissection with high‐intensity weightlifting. Many affected individuals were subsequently found to have at least one other risk factor, including hypertension, anabolic steroid abuse, and cocaine abuse.2426

• Large vessel arteritis: Large vessel arteritides, specifically giant cell arteritis, Takayasu's disease, and tertiary syphilis have long been associated with aortic dilatation and dissection.

• Cocaine and methamphetamine ingestion: Sympathomimetic drugs cause rapid increases in heart rate and blood pressure, markedly increasing aortic intraluminal shear stress. Furthermore, cocaine is thought to be directly toxic to vascular endothelium and may accelerate medial necrosis.2730

• Third trimester pregnancy, especially in patients with diseases of collagen31; The significance of pregnancy has recently been called into question by data from the IRAD trial. Of 346 enrolled women with aortic dissection, only 2 were pregnant, suggesting that the previously held association of pregnancy with aortic dissection may be an artifact of selective reporting.1

• Blunt chest trauma or high‐speed deceleration injury.

• Iatrogenic injury, typically from intra‐aortic catheterization.

INITIAL EVALUATION

The differential diagnosis for acute aortic dissection includes acute coronary syndrome, pulmonary embolus, pneumothorax, pneumonia, musculoskeletal pain, acute cholecystitis, esophageal spasm or rupture, acute pancreatitis, and acute pericarditis. Acute aortic dissections are rarely asymptomatic; in fact, the absence of sudden‐onset chest pain decreases the likelihood of dissection (negative LR 0.3).32 In the IRAD trial, approximately 95% of patients with aortic dissection complained of pain in the chest, back, or abdomen, with 90% characterizing their pain as either severe or the worst ever and 64% describing it as sharp.14 Although the presence of tearing or ripping chest or back pain suggests aortic dissection (positive LR 1.2‐10.8), its absence does not reliably exclude this diagnosis.32 The wide variability in the presentation of aortic dissection increases the challenge of establishing a diagnosis. Clinical findings depend largely on the anatomical location of the dissection and may include pulse deficits, neurologic deficits, hypotension, hypertension, and end‐organ ischemia. Women who develop aortic dissection are generally older and present later than men. Their symptoms are less typical and are likely to be confounded by altered mental status.1 A diagnosis of aortic dissection should be strongly considered for patients presenting with acute chest or back pain and otherwise unexplained aortic insufficiency, focal neurologic deficits, pulse deficits, or end‐organ injury (Table 3).

Electrocardiogram: Electrocardiographic abnormalities are commonly seen in aortic dissection and may include ST‐segment or T‐wave abnormalities or left ventricular hypertrophy.14 Proximal aortic dissections may compromise coronary artery perfusion, generating electrocardiogram (ECG) findings compatible with acute myocardial infarction, which may lead the clinician to diagnose and treat myocardial infarction while missing the underlying diagnosis.33 In a recent survey, 9 of 44 patients (21%) presenting with acute aortic dissection were initially diagnosed with acute coronary syndrome and anticoagulated, with 2 deaths.34 ECGs must therefore be interpreted with extreme caution in aortic dissection.

Chest x‐ray: In the emergency department, chest radiography is a mainstay of the evaluation of acute chest pain. Unfortunately, plain‐film radiography has limited utility for diagnosing aortic dissection.35 In the IRAD trial, mediastinal widening (>8 cm) and abnormal aortic contour, the classic radiographic findings in aortic dissection, were present in only 50%‐60% of cases. Twelve percent of patients had a completely normal chest x‐ray.14 A pooled analysis of previous studies demonstrated that the sensitivity of widened mediastinum and abnormal aortic contour was 65% and 71%, respectively.32 Nonspecific radiographic findings, most notably pleural effusion, were common.36 Thus, if the index of suspicion for aortic dissection is elevated, a confirmatory study must be obtained (Figure 2).

Figure 2

CONFIRMATORY IMAGING STUDIES

The ideal confirmatory imaging modality should identify aortic dissection with high sensitivity and specificity. It should also identify the entry and exit points of the dissection and provide information about the extent of compromise of the aortic valve, pericardium, and great vessels. Four imaging modalities sufficiently meet these criteria in order to be considered diagnostically useful.

Aortography: Previously the gold standard for diagnosing aortic dissection, aortography is no longer a first‐line imaging modality. The sensitivity and specificity of aortography are at best equivalent and probably inferior to less invasive imaging modalities.37, 38 False negatives may occur if both the true and false lumens opacify equally with contrast, or if the false lumen is sufficiently thrombosed to preclude any instillation of contrast. Aortography cannot identify aortic intramural hematomas, is invasive and highly operator dependent, requires nephrotoxic contrast, and generally takes longer to obtain than other modalities.39

Aortography uniquely offers excellent visualization of the coronary arteries and great vessels and is preferred when such information is necessary. Percutaneous aortic endovascular stent grafting has been recently employed to repair distal aortic dissections.4043 As a result, aortography is gaining new life as a therapeutic modality.

CT angiography: Spiral CT angiography (CTA) is the most commonly used modality for diagnosing aortic dissection.44 It is emergently available at most hospitals, and images can be obtained in minutes. Sensitivity and specificity may approach 100%, and CTA may be more sensitive than MRA or TEE in evaluating arch vessel involvement.4547 Like conventional angiography, CTA requires administration of nephrotoxic contrast. It frequently cannot visualize the entry and exit sites (intimal flaps) of a dissection and provides limited information about the coronary arteries and no information about the competency of the aortic valve.48, 49 Thus, if aortic dissection is identified by CTA, a second study may be needed to provide further diagnostic information and to guide surgical intervention (Figures 3 and 4).

Figure 3

Figure 4

Magnetic Resonance Angiography: Magnetic resonance angiography (MRA) offers excellent noninvasive evaluation of the thoracic aorta. Sensitivity and specificity are probably superior to spiral CTA, and MRA generally identifies the location of the intimal tear and provides some functional information about the aortic valve.44, 50, 51 MRA is not emergently available at many hospitals. Scanning is time intensive, requiring the patient to remain motionless and relatively inaccessible for up to an hour. Furthermore, patient claustrophobia and the presence of implanted devices such as pacemakers or ferromagnetic foreign bodies may preclude MRA.

Transesophageal echocardiography: The sensitivity and specificity of transesophageal echocardiography (TEE) are also excellenton a par with CTA and MRA. In addition to providing excellent visualization of the thoracic aorta, TEE provides superb images of the pericardium and detailed assessment of aortic valve function.52 It also is extremely effective at visualizing the aortic intimal flap.44, 49, 53 A significant advantage of TEE is its portability, allowing rapid diagnosis at the bedside. For this reason, it is particularly useful for evaluation of patients who are hemodynamically unstable and are suspected to have an aortic dissection. Because of the anatomic relationship of the aorta with the esophagus and the trachea, TEE more effectively identifies proximal than distal dissections.43 TEE is also somewhat invasive, usually requires patient sedation, and is highly operator dependent, requiring the availability of an experienced and technically skilled operator (Figure 5).

Figure 5

Transthoracic echocardiography: Although it is an excellent tool for the evaluation of many aspects of cardiac anatomy and function, surface echocardiography can reliably visualize only limited portions of the ascending and descending aorta.54, 55 As a consequence, it is neither sensitive nor specific enough to diagnose aortic dissection. Transthoracic echocardiography (TTE) does, however, play a role in rapidly assessing patients at the bedside for aortic valve or pericardial compromise when these complications are suspected.

Recommendations

CTA, MRA, and TEE are all highly sensitive and specific modalities for diagnosing aortic dissection. Therefore, the condition of the patient, the information needed, and the resources and expertise immediately available should drive the choice of study. MRA is considered the gold standard diagnostic study and is the preferred modality for hemodynamically stable patients with suspected aortic dissection. Because of slow data acquisition and the inaccessibility of patients in the scanner, it is generally unsuited for unstable patients, including those with ongoing pain. Bedside TEE is an excellent choice for patients who are too unstable for MRA but is less effective at visualizing distal dissections. Arch aortography is generally reserved for the confirmation of questionable diagnoses or to image specific branch arteries (Tables 4 and 5).

Table 4.Comparative Sensitivities of Imaging Modalities in Aortic Dissection

	Overall	Proximal	Distal
Adapted from Moore et al.43
TEE	88%	90%	80%
CTA	93%	93%	93%
MRA	100%	100%	100%
Aortogram	87%	87%	87%

Table 5.Comparative Diagnostic Utility of Imaging Techniques in Aortic Dissection

	TEE	CTA	MRA	Aortography
Adapted from recommendations of the task force on aortic dissection, European Society of Cardiology.61
Sensitivity	++	++	+++	++
Specificity	+++	++	+++	++
Classification	+++	++	++	+
Intimal flap	+++	‐	++	+
Aortic regurgitation	+++		++	++
Pericardial effusion	+++	++	++
Branch vessel involvement	+	++	++	+++
Coronary artery involvement	++	+	+	+++

Most trials comparing CTA, MRA, and TEE were performed in the early 1990s. Computed tomography has evolved significantly over the intervening decade, and some of the diagnostic limitations previously ascribed to CTA, such as the inability to generate 3‐D reconstructed images, no longer exist. Furthermore, CT angiography is widely available and is gaining increasing acceptance as a first‐line imaging modality for patients with noncardiac chest pain.48 Medical centers that maintain round‐the‐clock CT capability may have limited or delayed access to TEE, MRA, or aortography. Given the potential for rapid and dramatic patient deterioration, it is imperative that a diagnosis be established quickly when aortic dissection is suspected. Thus, when the choice is obtaining an immediate CTA or a delayed TEE or MRA, CTA is generally the better choice (Figure 6).

Figure 6

MANAGEMENT

PROGNOSIS

Despite significant medical and surgical advances, aortic dissection remains exceptionally lethal. Patients with proximal dissections are more likely to die than those with distal dissections. Using data from the IRAD trial, Mehta et al determined that age 70 years (OR, 1.70), abrupt onset of chest pain (OR 2.60), hypotension/shock/tamponade (OR, 2.97), renal failure (OR, 4.77), pulse deficit (OR, 2.03), and abnormal ECG (OR, 1.77) were independent determinants of death.59 Medical treatment of proximal dissection is generally reserved for patients too ill, unstable, or frail to undergo surgery. In contrast, most patients with distal dissection are managed medically, with surgery generally reserved for those with acute complications. Hence, patients with proximal dissections who are managed medically and those with distal dissections who are managed surgically have the worst outcomes. Outcomes for women are worse than those for men, which is probably attributable to several factors. Women dissect at an older age, present later after the onset of symptoms, and are more likely to have confounding symptoms that may delay timely diagnosis1 (Table 7).

CONCLUSION

Aortic dissection is a rare and acutely life‐threatening cause of acute chest and back pain. Delays in diagnosis and misdiagnoses are common, frequently with catastrophic consequences. The key to diagnosis is maintaining a high index of suspicion for dissection, especially in patients who present with acute severe chest, back, or abdominal pain in the setting of unexplained acute pulse deficits, neurologic deficits, or acute end‐organ injury. Three clinical findings have been shown to be diagnostically useful: immediate onset of tearing or ripping chest or back pain, mediastinal widening or abnormal aortic contour on chest radiograph, and peripheral pulse deficits or variable pulse pressure (>20 mm Hg). If all 3 findings are absent, acute aortic dissection is unlikely. The presence of any of these findings should prompt further workup. A normal chest radiograph does not rule out aortic dissection. Only TEE, CT, and MR angiography are sufficiently specific to rule out dissection. Aortography is rarely used as a first‐line diagnostic tool but may be useful as a confirmatory test or to provide additional anatomic information. Patients who present with proximal aortic dissection or with any aortic dissection with concomitant hypotension are at exceptionally high risk of death and should be immediately referred for surgical evaluation. Beta‐blockers are the mainstay of acute and chronic therapy of aortic dissection. Survivors of aortic dissection are at a markedly elevated risk for further aortic events and should be followed vigilantly posthospitalization.

Aortic dissection is an uncommon but highly lethal disease with an incidence of approximately 2,000 cases per year in the United States.1 It is often mistaken for less serious pathology. In one series, aortic dissection was missed in 38% of patients at presentation, with 28% of patients first diagnosed at autopsy.2 Early recognition and management are crucial. If untreated, the mortality rate for acute aortic dissection increases by approximately 1% per hour over the first 48 hours and may reach 70% at 1 week. As many as 90% of untreated patients who suffer aortic dissection die within 3 months of presentation.3, 4 Generally, cardiothoracic surgeons or cardiologists experienced with managing aortic dissection should direct patient evaluation and treatment. Hospitalists, however, are increasingly assuming responsibility for the initial triage and management of patients with acute chest pain syndromes and therefore must be able to rapidly identify aortic dissection, initiate supportive therapy, and refer patients to appropriate specialty care.

PATHOPHYSIOLOGY

Aortic dissection occurs when layers of the aortic wall separate because of infiltration of high‐pressure arterial blood. The proximate causes are elevated shear stress across the aortic lumen in the setting of a concomitant defect in the aortic media. Shear stress is caused by the rapid increase in luminal pressure per unit of time (dP/dt) that results from cardiac systole. As the aorta traverses away from the heart, an increasing proportion of the kinetic energy of left ventricular systole is stored in the aortic wall as potential energy, which facilitates anterograde propagation of cardiac output during diastole. This conversion of kinetic to potential energy also attenuates shear stress. As the proximal aorta is subject to the steepest fluctuations in pressure, it is at the highest risk of dissection. Degeneration of the aortic media is part of the normal aging process but is accelerated in persons with a bicuspid aortic valve, Turner's syndrome, inflammatory arteritis, or inherited diseases of collagen formation.

Once the aortic intima is compromised, blood dissects longitudinally through the aortic media and propagates proximally or distally, creating a false lumen that may communicate with the true lumen of the aorta. Blood may flow through the true lumen, the false lumen, or both. Propagation of the dissection causes much of the morbidity associated with aortic dissection by disrupting blood flow across branch vessels or by directly compromising the pericardium or aortic valve. Over time, the dissection may traverse the entire aortic wall, causing aortic rupture and exsanguination.

CLASSIFICATION

Acute aortic dissection is classified as any aortic dissection diagnosed within 2 weeks of the onset of symptoms, which is the period of highest risk of mortality. Patients who survive more than 2 weeks without treatment are considered to have chronic dissection. Aortic dissections are further classified according to their anatomic location. The fundamental distinction is whether the dissection is proximal (involving the aortic root or ascending aorta) or distal (below the left subclavian artery). The Stanford and DeBakey classification systems are the classification systems most commonly used (Figure 1).

Figure 1

Some variants of aortic dissection are not described in either the Stanford or DeBakey systems. Aortic intramural hematomas (IMH) are caused by intramural hemorrhage of the vasa vasorum without an identifiable intimal tear.57 Penetrating atherosclerotic ulcers (PAUs) are focal defects in the aortic wall with surrounding hematoma but no longitudinal dissection across tissue planes, typically resulting from advanced atherosclerotic disease.8 The pathophysiologic distinctions between IMH, PAU, and classic aortic dissection remain somewhat controversial. Both IMH and PAU may progress to aortic aneurysm formation, frank dissection, or aortic rupture, suggesting that these entities represent a spectrum of diseases with broad overlap (Table 1).9, 10

EPIDEMIOLOGY

Aortic dissection is a rare disease, with an estimated incidence of approximately 5‐30 cases per 1 million people per year.1114 Fewer than 0.5% of patients presenting to an emergency department with chest or back pain suffer from aortic dissection.15 Two thirds of patients are male, with an average age at presentation of approximately 65 years. A history of systemic hypertension, found in up to 72% of patients, is by far the most common risk factor.2, 14, 16 Atherosclerosis, a history of prior cardiac surgery, and known aortic aneurysm are other major risk factors.14 The epidemiology of aortic dissection is substantially different in young patients (40 years of age). Hypertension and atherosclerosis become significantly less common, as other risk factors, such as Marfan syndrome, take precedence17 (Table 2). Other risk factors for aortic dissection include:

• Collagen diseases (eg, Marfan syndrome and Ehlers‐Danlos): In the International Registry of Acute Aortic Dissection (IRAD), the largest prospective analysis of aortic dissection to date, 50% of the young patients presenting with aortic dissection had Marfan syndrome.17

• Bicuspid aortic valve (BAV): Individuals with BAV are 5‐18 times more likely to suffer aortic dissection than those with a trileaflet valve.18, 19 In one survey, 52% of asymptomatic young men with BAV were found to have aortic root dilatation, a frequent precursor of dissection.20 Vascular tissue in individuals with BAV has been found to have increased levels of matrix metalloproteinases, which may degrade elastic matrix components and accelerate medial necrosis.21

• Aortic coarctation: Aortic coarctation is associated with upper extremity hypertension, BAV and aortic dilatation, all of which predispose to aortic dissection.

• Turner syndrome: Aortic root dilatation with or without dissection has been incidentally noted in 6%‐9% of patients with Turner syndrome.22, 23

• Strenuous exercise: Multiple case reports have associated aortic dissection with high‐intensity weightlifting. Many affected individuals were subsequently found to have at least one other risk factor, including hypertension, anabolic steroid abuse, and cocaine abuse.2426

• Large vessel arteritis: Large vessel arteritides, specifically giant cell arteritis, Takayasu's disease, and tertiary syphilis have long been associated with aortic dilatation and dissection.

• Cocaine and methamphetamine ingestion: Sympathomimetic drugs cause rapid increases in heart rate and blood pressure, markedly increasing aortic intraluminal shear stress. Furthermore, cocaine is thought to be directly toxic to vascular endothelium and may accelerate medial necrosis.2730

• Third trimester pregnancy, especially in patients with diseases of collagen31; The significance of pregnancy has recently been called into question by data from the IRAD trial. Of 346 enrolled women with aortic dissection, only 2 were pregnant, suggesting that the previously held association of pregnancy with aortic dissection may be an artifact of selective reporting.1

• Blunt chest trauma or high‐speed deceleration injury.

• Iatrogenic injury, typically from intra‐aortic catheterization.

INITIAL EVALUATION

The differential diagnosis for acute aortic dissection includes acute coronary syndrome, pulmonary embolus, pneumothorax, pneumonia, musculoskeletal pain, acute cholecystitis, esophageal spasm or rupture, acute pancreatitis, and acute pericarditis. Acute aortic dissections are rarely asymptomatic; in fact, the absence of sudden‐onset chest pain decreases the likelihood of dissection (negative LR 0.3).32 In the IRAD trial, approximately 95% of patients with aortic dissection complained of pain in the chest, back, or abdomen, with 90% characterizing their pain as either severe or the worst ever and 64% describing it as sharp.14 Although the presence of tearing or ripping chest or back pain suggests aortic dissection (positive LR 1.2‐10.8), its absence does not reliably exclude this diagnosis.32 The wide variability in the presentation of aortic dissection increases the challenge of establishing a diagnosis. Clinical findings depend largely on the anatomical location of the dissection and may include pulse deficits, neurologic deficits, hypotension, hypertension, and end‐organ ischemia. Women who develop aortic dissection are generally older and present later than men. Their symptoms are less typical and are likely to be confounded by altered mental status.1 A diagnosis of aortic dissection should be strongly considered for patients presenting with acute chest or back pain and otherwise unexplained aortic insufficiency, focal neurologic deficits, pulse deficits, or end‐organ injury (Table 3).

Electrocardiogram: Electrocardiographic abnormalities are commonly seen in aortic dissection and may include ST‐segment or T‐wave abnormalities or left ventricular hypertrophy.14 Proximal aortic dissections may compromise coronary artery perfusion, generating electrocardiogram (ECG) findings compatible with acute myocardial infarction, which may lead the clinician to diagnose and treat myocardial infarction while missing the underlying diagnosis.33 In a recent survey, 9 of 44 patients (21%) presenting with acute aortic dissection were initially diagnosed with acute coronary syndrome and anticoagulated, with 2 deaths.34 ECGs must therefore be interpreted with extreme caution in aortic dissection.

Chest x‐ray: In the emergency department, chest radiography is a mainstay of the evaluation of acute chest pain. Unfortunately, plain‐film radiography has limited utility for diagnosing aortic dissection.35 In the IRAD trial, mediastinal widening (>8 cm) and abnormal aortic contour, the classic radiographic findings in aortic dissection, were present in only 50%‐60% of cases. Twelve percent of patients had a completely normal chest x‐ray.14 A pooled analysis of previous studies demonstrated that the sensitivity of widened mediastinum and abnormal aortic contour was 65% and 71%, respectively.32 Nonspecific radiographic findings, most notably pleural effusion, were common.36 Thus, if the index of suspicion for aortic dissection is elevated, a confirmatory study must be obtained (Figure 2).

Figure 2

CONFIRMATORY IMAGING STUDIES

The ideal confirmatory imaging modality should identify aortic dissection with high sensitivity and specificity. It should also identify the entry and exit points of the dissection and provide information about the extent of compromise of the aortic valve, pericardium, and great vessels. Four imaging modalities sufficiently meet these criteria in order to be considered diagnostically useful.

Aortography: Previously the gold standard for diagnosing aortic dissection, aortography is no longer a first‐line imaging modality. The sensitivity and specificity of aortography are at best equivalent and probably inferior to less invasive imaging modalities.37, 38 False negatives may occur if both the true and false lumens opacify equally with contrast, or if the false lumen is sufficiently thrombosed to preclude any instillation of contrast. Aortography cannot identify aortic intramural hematomas, is invasive and highly operator dependent, requires nephrotoxic contrast, and generally takes longer to obtain than other modalities.39

Aortography uniquely offers excellent visualization of the coronary arteries and great vessels and is preferred when such information is necessary. Percutaneous aortic endovascular stent grafting has been recently employed to repair distal aortic dissections.4043 As a result, aortography is gaining new life as a therapeutic modality.

CT angiography: Spiral CT angiography (CTA) is the most commonly used modality for diagnosing aortic dissection.44 It is emergently available at most hospitals, and images can be obtained in minutes. Sensitivity and specificity may approach 100%, and CTA may be more sensitive than MRA or TEE in evaluating arch vessel involvement.4547 Like conventional angiography, CTA requires administration of nephrotoxic contrast. It frequently cannot visualize the entry and exit sites (intimal flaps) of a dissection and provides limited information about the coronary arteries and no information about the competency of the aortic valve.48, 49 Thus, if aortic dissection is identified by CTA, a second study may be needed to provide further diagnostic information and to guide surgical intervention (Figures 3 and 4).

Figure 3

Figure 4

Magnetic Resonance Angiography: Magnetic resonance angiography (MRA) offers excellent noninvasive evaluation of the thoracic aorta. Sensitivity and specificity are probably superior to spiral CTA, and MRA generally identifies the location of the intimal tear and provides some functional information about the aortic valve.44, 50, 51 MRA is not emergently available at many hospitals. Scanning is time intensive, requiring the patient to remain motionless and relatively inaccessible for up to an hour. Furthermore, patient claustrophobia and the presence of implanted devices such as pacemakers or ferromagnetic foreign bodies may preclude MRA.

Transesophageal echocardiography: The sensitivity and specificity of transesophageal echocardiography (TEE) are also excellenton a par with CTA and MRA. In addition to providing excellent visualization of the thoracic aorta, TEE provides superb images of the pericardium and detailed assessment of aortic valve function.52 It also is extremely effective at visualizing the aortic intimal flap.44, 49, 53 A significant advantage of TEE is its portability, allowing rapid diagnosis at the bedside. For this reason, it is particularly useful for evaluation of patients who are hemodynamically unstable and are suspected to have an aortic dissection. Because of the anatomic relationship of the aorta with the esophagus and the trachea, TEE more effectively identifies proximal than distal dissections.43 TEE is also somewhat invasive, usually requires patient sedation, and is highly operator dependent, requiring the availability of an experienced and technically skilled operator (Figure 5).

Figure 5

Transthoracic echocardiography: Although it is an excellent tool for the evaluation of many aspects of cardiac anatomy and function, surface echocardiography can reliably visualize only limited portions of the ascending and descending aorta.54, 55 As a consequence, it is neither sensitive nor specific enough to diagnose aortic dissection. Transthoracic echocardiography (TTE) does, however, play a role in rapidly assessing patients at the bedside for aortic valve or pericardial compromise when these complications are suspected.

Recommendations

CTA, MRA, and TEE are all highly sensitive and specific modalities for diagnosing aortic dissection. Therefore, the condition of the patient, the information needed, and the resources and expertise immediately available should drive the choice of study. MRA is considered the gold standard diagnostic study and is the preferred modality for hemodynamically stable patients with suspected aortic dissection. Because of slow data acquisition and the inaccessibility of patients in the scanner, it is generally unsuited for unstable patients, including those with ongoing pain. Bedside TEE is an excellent choice for patients who are too unstable for MRA but is less effective at visualizing distal dissections. Arch aortography is generally reserved for the confirmation of questionable diagnoses or to image specific branch arteries (Tables 4 and 5).

Table 4.Comparative Sensitivities of Imaging Modalities in Aortic Dissection

	Overall	Proximal	Distal
Adapted from Moore et al.43
TEE	88%	90%	80%
CTA	93%	93%	93%
MRA	100%	100%	100%
Aortogram	87%	87%	87%

Table 5.Comparative Diagnostic Utility of Imaging Techniques in Aortic Dissection

	TEE	CTA	MRA	Aortography
Adapted from recommendations of the task force on aortic dissection, European Society of Cardiology.61
Sensitivity	++	++	+++	++
Specificity	+++	++	+++	++
Classification	+++	++	++	+
Intimal flap	+++	‐	++	+
Aortic regurgitation	+++		++	++
Pericardial effusion	+++	++	++
Branch vessel involvement	+	++	++	+++
Coronary artery involvement	++	+	+	+++

Most trials comparing CTA, MRA, and TEE were performed in the early 1990s. Computed tomography has evolved significantly over the intervening decade, and some of the diagnostic limitations previously ascribed to CTA, such as the inability to generate 3‐D reconstructed images, no longer exist. Furthermore, CT angiography is widely available and is gaining increasing acceptance as a first‐line imaging modality for patients with noncardiac chest pain.48 Medical centers that maintain round‐the‐clock CT capability may have limited or delayed access to TEE, MRA, or aortography. Given the potential for rapid and dramatic patient deterioration, it is imperative that a diagnosis be established quickly when aortic dissection is suspected. Thus, when the choice is obtaining an immediate CTA or a delayed TEE or MRA, CTA is generally the better choice (Figure 6).

Figure 6

MANAGEMENT

PROGNOSIS

Despite significant medical and surgical advances, aortic dissection remains exceptionally lethal. Patients with proximal dissections are more likely to die than those with distal dissections. Using data from the IRAD trial, Mehta et al determined that age 70 years (OR, 1.70), abrupt onset of chest pain (OR 2.60), hypotension/shock/tamponade (OR, 2.97), renal failure (OR, 4.77), pulse deficit (OR, 2.03), and abnormal ECG (OR, 1.77) were independent determinants of death.59 Medical treatment of proximal dissection is generally reserved for patients too ill, unstable, or frail to undergo surgery. In contrast, most patients with distal dissection are managed medically, with surgery generally reserved for those with acute complications. Hence, patients with proximal dissections who are managed medically and those with distal dissections who are managed surgically have the worst outcomes. Outcomes for women are worse than those for men, which is probably attributable to several factors. Women dissect at an older age, present later after the onset of symptoms, and are more likely to have confounding symptoms that may delay timely diagnosis1 (Table 7).

CONCLUSION

Aortic dissection is a rare and acutely life‐threatening cause of acute chest and back pain. Delays in diagnosis and misdiagnoses are common, frequently with catastrophic consequences. The key to diagnosis is maintaining a high index of suspicion for dissection, especially in patients who present with acute severe chest, back, or abdominal pain in the setting of unexplained acute pulse deficits, neurologic deficits, or acute end‐organ injury. Three clinical findings have been shown to be diagnostically useful: immediate onset of tearing or ripping chest or back pain, mediastinal widening or abnormal aortic contour on chest radiograph, and peripheral pulse deficits or variable pulse pressure (>20 mm Hg). If all 3 findings are absent, acute aortic dissection is unlikely. The presence of any of these findings should prompt further workup. A normal chest radiograph does not rule out aortic dissection. Only TEE, CT, and MR angiography are sufficiently specific to rule out dissection. Aortography is rarely used as a first‐line diagnostic tool but may be useful as a confirmatory test or to provide additional anatomic information. Patients who present with proximal aortic dissection or with any aortic dissection with concomitant hypotension are at exceptionally high risk of death and should be immediately referred for surgical evaluation. Beta‐blockers are the mainstay of acute and chronic therapy of aortic dissection. Survivors of aortic dissection are at a markedly elevated risk for further aortic events and should be followed vigilantly posthospitalization.

Six years ago, at the age of 76, I suffered a type B aortic dissection while lifting weights. At the time, I managed a busy architecture practice, jogged a 9‐minute mile, and had never experienced a serious illness. My only medication was a baby aspirin every morning. The dissection was diagnosed promptly and treated medically with a stay of less than a week in the hospital. My son Eric, a hospitalist, immediately flew in from Denver and stayed with me throughout my hospitalization. Within days of my discharge, however, I began experiencing a series of mild discomforts. I had difficulty sleeping, mild indigestion, and a burning sensation in my thighs after walking relatively short distances. My primary care physician and cardiologist didn't seem concerned about these symptoms, and as they were initially mild, I accepted them as the residual effects of sleep deprivation, hospital food, and muscle atrophy. Nobody recognized that my dissection had propagated, effectively cutting off blood flow below my diaphragm.

To my good fortune, Eric had previously scheduled a second visit long before I had become ill. I vividly recall him walking into our house in the early evening, taking one look at me, and saying, I don't like what I see. My most notable memory of the return ambulance trip was how cold my feet were. I had never experienced that intense a sensation of cold before. I recall arriving at a very busy, crowded ER and Eric aggressively trying to get priority attention. The next thing I remember was waking up in the ICU the next afternoon. My kidneys had failed, and my intestines were not getting blood flow. By then, I had become too unstable to undergo aortic surgery. As a last‐ditch effort, an interventional radiologist tried to open my aorta using four large biliary stents, none of which deployed properly. Then a vascular surgeon suggested performing an axillo‐bifemoral bypass, which is much less invasive than aortic surgery, in order to restore blood flow to my kidneys and intestines. It saved my life.

There are many things about the hospital environment that elevate anxiety and vulnerability, and perhaps that is inevitable. The concentration of sick people is depressing. I had several roommates, each much younger than I, with prognoses that appeared far less favorable than mine. I listened to their doctors, some with bedside manners so clinical that they bordered on insensitivity. In one instance, when there was clearly a communication barrier, I saw a patient and family become bewildered and overwhelmed. Watching this unfold only heightened my own sense of vulnerability.

When it became evident that I was not going to die, my emotions ranged from elation at having beaten the odds to outright fear. I had come to the hospital healthy, fit, and independent. Now a host of new concerns emerged. What would my limitations be, and to what kind of lifestyle could I look forward? Would I be self‐sufficient, or would I become a burden to my family? Perhaps these were ungrateful responses to having just dodged the bullet, but I take no responsibility for my subconscious. Nights were the worst, especially when sleep was elusive. My days were filled with tests and visitors, but I had ample time to court my anxieties after dark. As an artillery reconnaissance officer during World War II, I had known fear, but this was different. Fear during combat was shared by all and functionally accommodated by most. It became part of a common bond. Maybe because of our youth and inexperience, we only worried about being killeda singular event, and then it was all over. Thoughts of permanent disability and its consequences never crossed our minds.

My physical recovery was far more rapid than my psychological and emotional recovery. My physicians told me that they had never encountered a case like mine and that we were in uncharted waters. Although I appreciated their candor, this was less than reassuring. Furthermore, the mild symptoms I had experienced during my re‐dissection sensitized me to every new little pain, twinge, or discomfort. How was I to differentiate relevant and significant new symptoms from hypochondria? It took a very long time to recover my sense of well‐being. The support of my wife and family was invaluable, but ultimately this is something one must sort out for oneself. I recognized that I could not face the rest of my life with fear and anxiety. I tried professional counseling without noticeable benefit. Eventually, I learned to analyze each concern as it surfaced, recognize it for what it was, and put it in perspective. The passage of time and daily meditation also contributed to my emotional healing. Today, some of the old ghosts still emerge from the shadows, but now they have no substance and rapidly disappear.

I attribute my survival and recovery over the past six years to the marvels of modern medicine (accompanied by some miscues and imperfections), the forceful advocacy of a loving wife and a physician son who were always there at critical moments, and a significant dose of pure good luck. As I write this, a few days before my 82nd birthday, I remain actively engaged in my practice, still work out, albeit more prudently than before, and walk a brisk 16‐minute mile. I will forever be grateful for the professionalism and dedication of the health care personnel I encountered and for the astonishing technology and infrastructure that saved my life.

Six years ago, at the age of 76, I suffered a type B aortic dissection while lifting weights. At the time, I managed a busy architecture practice, jogged a 9‐minute mile, and had never experienced a serious illness. My only medication was a baby aspirin every morning. The dissection was diagnosed promptly and treated medically with a stay of less than a week in the hospital. My son Eric, a hospitalist, immediately flew in from Denver and stayed with me throughout my hospitalization. Within days of my discharge, however, I began experiencing a series of mild discomforts. I had difficulty sleeping, mild indigestion, and a burning sensation in my thighs after walking relatively short distances. My primary care physician and cardiologist didn't seem concerned about these symptoms, and as they were initially mild, I accepted them as the residual effects of sleep deprivation, hospital food, and muscle atrophy. Nobody recognized that my dissection had propagated, effectively cutting off blood flow below my diaphragm.

To my good fortune, Eric had previously scheduled a second visit long before I had become ill. I vividly recall him walking into our house in the early evening, taking one look at me, and saying, I don't like what I see. My most notable memory of the return ambulance trip was how cold my feet were. I had never experienced that intense a sensation of cold before. I recall arriving at a very busy, crowded ER and Eric aggressively trying to get priority attention. The next thing I remember was waking up in the ICU the next afternoon. My kidneys had failed, and my intestines were not getting blood flow. By then, I had become too unstable to undergo aortic surgery. As a last‐ditch effort, an interventional radiologist tried to open my aorta using four large biliary stents, none of which deployed properly. Then a vascular surgeon suggested performing an axillo‐bifemoral bypass, which is much less invasive than aortic surgery, in order to restore blood flow to my kidneys and intestines. It saved my life.

There are many things about the hospital environment that elevate anxiety and vulnerability, and perhaps that is inevitable. The concentration of sick people is depressing. I had several roommates, each much younger than I, with prognoses that appeared far less favorable than mine. I listened to their doctors, some with bedside manners so clinical that they bordered on insensitivity. In one instance, when there was clearly a communication barrier, I saw a patient and family become bewildered and overwhelmed. Watching this unfold only heightened my own sense of vulnerability.

When it became evident that I was not going to die, my emotions ranged from elation at having beaten the odds to outright fear. I had come to the hospital healthy, fit, and independent. Now a host of new concerns emerged. What would my limitations be, and to what kind of lifestyle could I look forward? Would I be self‐sufficient, or would I become a burden to my family? Perhaps these were ungrateful responses to having just dodged the bullet, but I take no responsibility for my subconscious. Nights were the worst, especially when sleep was elusive. My days were filled with tests and visitors, but I had ample time to court my anxieties after dark. As an artillery reconnaissance officer during World War II, I had known fear, but this was different. Fear during combat was shared by all and functionally accommodated by most. It became part of a common bond. Maybe because of our youth and inexperience, we only worried about being killeda singular event, and then it was all over. Thoughts of permanent disability and its consequences never crossed our minds.

My physical recovery was far more rapid than my psychological and emotional recovery. My physicians told me that they had never encountered a case like mine and that we were in uncharted waters. Although I appreciated their candor, this was less than reassuring. Furthermore, the mild symptoms I had experienced during my re‐dissection sensitized me to every new little pain, twinge, or discomfort. How was I to differentiate relevant and significant new symptoms from hypochondria? It took a very long time to recover my sense of well‐being. The support of my wife and family was invaluable, but ultimately this is something one must sort out for oneself. I recognized that I could not face the rest of my life with fear and anxiety. I tried professional counseling without noticeable benefit. Eventually, I learned to analyze each concern as it surfaced, recognize it for what it was, and put it in perspective. The passage of time and daily meditation also contributed to my emotional healing. Today, some of the old ghosts still emerge from the shadows, but now they have no substance and rapidly disappear.

I attribute my survival and recovery over the past six years to the marvels of modern medicine (accompanied by some miscues and imperfections), the forceful advocacy of a loving wife and a physician son who were always there at critical moments, and a significant dose of pure good luck. As I write this, a few days before my 82nd birthday, I remain actively engaged in my practice, still work out, albeit more prudently than before, and walk a brisk 16‐minute mile. I will forever be grateful for the professionalism and dedication of the health care personnel I encountered and for the astonishing technology and infrastructure that saved my life.

Six years ago, at the age of 76, I suffered a type B aortic dissection while lifting weights. At the time, I managed a busy architecture practice, jogged a 9‐minute mile, and had never experienced a serious illness. My only medication was a baby aspirin every morning. The dissection was diagnosed promptly and treated medically with a stay of less than a week in the hospital. My son Eric, a hospitalist, immediately flew in from Denver and stayed with me throughout my hospitalization. Within days of my discharge, however, I began experiencing a series of mild discomforts. I had difficulty sleeping, mild indigestion, and a burning sensation in my thighs after walking relatively short distances. My primary care physician and cardiologist didn't seem concerned about these symptoms, and as they were initially mild, I accepted them as the residual effects of sleep deprivation, hospital food, and muscle atrophy. Nobody recognized that my dissection had propagated, effectively cutting off blood flow below my diaphragm.

To my good fortune, Eric had previously scheduled a second visit long before I had become ill. I vividly recall him walking into our house in the early evening, taking one look at me, and saying, I don't like what I see. My most notable memory of the return ambulance trip was how cold my feet were. I had never experienced that intense a sensation of cold before. I recall arriving at a very busy, crowded ER and Eric aggressively trying to get priority attention. The next thing I remember was waking up in the ICU the next afternoon. My kidneys had failed, and my intestines were not getting blood flow. By then, I had become too unstable to undergo aortic surgery. As a last‐ditch effort, an interventional radiologist tried to open my aorta using four large biliary stents, none of which deployed properly. Then a vascular surgeon suggested performing an axillo‐bifemoral bypass, which is much less invasive than aortic surgery, in order to restore blood flow to my kidneys and intestines. It saved my life.

There are many things about the hospital environment that elevate anxiety and vulnerability, and perhaps that is inevitable. The concentration of sick people is depressing. I had several roommates, each much younger than I, with prognoses that appeared far less favorable than mine. I listened to their doctors, some with bedside manners so clinical that they bordered on insensitivity. In one instance, when there was clearly a communication barrier, I saw a patient and family become bewildered and overwhelmed. Watching this unfold only heightened my own sense of vulnerability.

When it became evident that I was not going to die, my emotions ranged from elation at having beaten the odds to outright fear. I had come to the hospital healthy, fit, and independent. Now a host of new concerns emerged. What would my limitations be, and to what kind of lifestyle could I look forward? Would I be self‐sufficient, or would I become a burden to my family? Perhaps these were ungrateful responses to having just dodged the bullet, but I take no responsibility for my subconscious. Nights were the worst, especially when sleep was elusive. My days were filled with tests and visitors, but I had ample time to court my anxieties after dark. As an artillery reconnaissance officer during World War II, I had known fear, but this was different. Fear during combat was shared by all and functionally accommodated by most. It became part of a common bond. Maybe because of our youth and inexperience, we only worried about being killeda singular event, and then it was all over. Thoughts of permanent disability and its consequences never crossed our minds.

My physical recovery was far more rapid than my psychological and emotional recovery. My physicians told me that they had never encountered a case like mine and that we were in uncharted waters. Although I appreciated their candor, this was less than reassuring. Furthermore, the mild symptoms I had experienced during my re‐dissection sensitized me to every new little pain, twinge, or discomfort. How was I to differentiate relevant and significant new symptoms from hypochondria? It took a very long time to recover my sense of well‐being. The support of my wife and family was invaluable, but ultimately this is something one must sort out for oneself. I recognized that I could not face the rest of my life with fear and anxiety. I tried professional counseling without noticeable benefit. Eventually, I learned to analyze each concern as it surfaced, recognize it for what it was, and put it in perspective. The passage of time and daily meditation also contributed to my emotional healing. Today, some of the old ghosts still emerge from the shadows, but now they have no substance and rapidly disappear.

I attribute my survival and recovery over the past six years to the marvels of modern medicine (accompanied by some miscues and imperfections), the forceful advocacy of a loving wife and a physician son who were always there at critical moments, and a significant dose of pure good luck. As I write this, a few days before my 82nd birthday, I remain actively engaged in my practice, still work out, albeit more prudently than before, and walk a brisk 16‐minute mile. I will forever be grateful for the professionalism and dedication of the health care personnel I encountered and for the astonishing technology and infrastructure that saved my life.

Spiral computed tomographic pulmonary angiography (CTPA) is a common first‐line test for the evaluation of suspected pulmonary embolism (PE). At our institution CTPA became the initial diagnostic study in 83% of patients with suspected PE within 3 years of the introduction of CT,1 and by 2001 CTPA had become the most common diagnostic test performed nationwide in patients diagnosed with PE.2 Most scans are interpreted as either positive or negative for pulmonary embolism, providing clinicians with a greater sense of diagnostic certainty than with the probabilistic results of lung scintigraphy. Initial studies of CTPA supported this appearance of diagnostic certainty, reporting sensitivity and specificity of greater than 90%,3, 4 but several subsequent studies have failed to reproduce these results.57 Newer multidetector CT scans are believed to be more accurate than earlier single‐detector CT,8 but true estimates of CTPA test characteristics will not be known until publication of the forthcoming PIOPED II study.9

Even without these data, CT‐based diagnostic algorithms have already appeared.1014 These algorithms generally focus on minimizing the false‐negative rate through use of serial testing (involving combinations of serum D‐dimer, lower‐extremity ultrasound, and CTPA). A recent meta‐analysis demonstrated that negative CTPA is highly accurate at ruling out PE, with test characteristics similar to conventional pulmonary angiography.15 Another meta‐analysis found that the 3‐month rate of subsequent venous thromboembolism after negative CTPA was 1.4% (95% CI 1.1%‐1.8%),16 supporting the strategy of withholding anticoagulants after negative CTPA in combination with other tests. However, use of serial testing to establish the diagnosis of PE and initiate anticoagulation has not been systematically evaluated or recommended, even for patients with a low pretest probability of PE.17

To assess the potential impact of these algorithms on the diagnosis of PE in clinical practice, we analyzed the clinical presentation and treatment of a cohort of patients at our institution who underwent CTPA for suspected PE.1 We calculated a range of posttest probabilities for pulmonary embolism for these patients, given the pretest probabilities, test results, and estimates of CTPA test characteristics. We then compared the treatment decisions of clinicians to the posttest probabilities of PE in order to establish the potential frequency of false‐positive and false‐negative diagnoses and to determine if patients were treated appropriately based on these estimates.

METHODS

Sites and Subjects

Details of the sites, subjects, and methods used to collect patient‐level data in this analysis have been previously published.1 The study was performed at Moffitt‐Long Hospital and San Francisco General Hospital, teaching hospitals affiliated with the University of California San Francisco School of Medicine. At both sites, single‐detector CT scans were available 24 hours a day throughout the study period and were read by attending radiologists who specialized in thoracic imaging. We excluded patients whose CTPA was not completed as the initial test in the evaluation of suspected PE, those who underwent testing for any indication other than suspected acute PE, and those with incomplete medical records or technically inadequate CTPA.

We randomly selected 345 patients who underwent CTPA between January 1, 1998, and December 31, 2000, from the Radiology Department databases. One investigator (R.L.T.) then abstracted charts of all patients. For each subject, we collected data about history and clinical presentation, diagnostic impressions of the treating clinicians, treatments administered both before and after diagnostic testing, CTPA result, results of other diagnostic tests for PE, and final clinical diagnosis. During the study period, there were no institution‐ or department‐specific guidelines or decision aids available for the diagnosis of PE. Ventilation‐perfusion scan, lower extremity ultrasound, and pulmonary angiography were available, but highly sensitive D‐dimer assays were not in use. The study was approved by the Institutional Review Boards of both sites, and requirement for written informed consent from patients was waived.

Estimates of Pretest Probabilities of Pulmonary Embolism and CTPA Test Characteristics

Several prediction rules1820 generate clinical pretest probabilities for patients with suspected PE. We used the Wells score18 to assign a pretest probability of low, moderate, or high to each patient on the basis of the following clinical variables: leg swelling, hemoptysis, tachycardia, history of recent immobilization, history of prior DVT or PE, active malignancy, and lack of a more likely alternative diagnosis. We chose this rule as (unlike other prediction rules such as the Geneva rule20) the Wells score has been validated for hospitalized patients with suspected PE and does not require arterial blood gas measurements. The prevalence of PE reported in the evaluation of the Wells score was 3.4%, 27.8%, and 78.3% for low, moderate, and high pretest probabilities, respectively.18

As in our previous study,1 we assumed CTPA to be 90% sensitive and 95% specific based on published estimates.3, 17 These values correspond to a positive likelihood ratio of 18 and a negative likelihood ratio of 0.1.21 We chose these values as a best‐case estimate of the test characteristics of CTPA, although other studies have found less impressive results.7 Using these pretest probabilities and likelihood ratios, we then used Bayes' theorem (Figure 1) to calculate the range of expected posttest probabilities of pulmonary embolism.

Figure 1

RESULTS

DISCUSSION

Physicians at our institution regarded CTPA results as definitive, anticoagulating 96.5% of patients with a positive CT and withholding treatment in 95.8% of patients with a negative scan. This practice pattern may result in unnecessary anticoagulation of many patients with a low pretest probability of PE who may have had false‐positive CTPA findings. In contrast, the rate of possible false‐negative diagnosis of PE was low, consistent with the results of several other studies.16

The use of CTPA is likely to increase because of the publication of multiple algorithms advocating that CTPA be the chief imaging study used in the diagnosis of PE.1014 These algorithms recommend serial testing on patients with a negative CTPA in order to minimize the false‐negative rate, but they do not require systematic follow‐up in patients with a positive scan, even if the pretest probability was low. In management trials, this approach resulted in a low false‐negative rate (1.0%‐1.8% at 3‐month follow‐up).1114 However, the rate of major bleeding in patients treated for PE was 3.2%‐6.0% at 3 months,1214 illustrating the potential risk of anticoagulating patients who may have false‐positive diagnoses. Furthermore, premature diagnostic closure after a CTPA positive for PE may result in additional morbidity as a result of missing the true diagnosis.

One potential explanation for the large number of potential false‐positive emboli seen in low‐risk patients is that it is difficult to accurately diagnose distal pulmonary emboli with CTPA. The interrater reliability of CTPA for diagnosis of subsegmental PE is suboptimal,23 and the clinical significance of these emboli remains uncertain.24 Thus, many emboli found in patients with low pretest probability actually may have been subsegmental PE that would not have been diagnosed by another radiologist. As CTPA is more accurate for diagnosing central PE,25 clinicians should consider reviewing positive scans with the interpreting radiologist, especially when the pretest probability was low and the filling defects identified are in distal vessels.

Our results may also illustrate that clinicians have a lower treatment threshold when presented with apparently definitive evidence of pulmonary embolism. Previous proposals on the appropriate treatment threshold for PE, which used Bayesian decision‐making methods similar to ours,22 incorporated PIOPED26 data on the pretest probability of pulmonary embolism, the test characteristics of ventilation‐perfusion scans, and the clinical outcomes of patients in each test result/pretest probability category. However, there is no corresponding data for CTPA, as its test characteristics are still uncertain, and long‐term clinical outcomes have not been documented for patients treated (or not treated) on the basis of CT results.

Our study had several limitations. First, charting bias potentially was introduced by our using a retrospective method of collecting data for calculating pretest probabilities. To address this potential bias, we collected data from the entire medical record, including information available at and preceding the time of the CT scan. We believe this method was effective, as the range of pretest probabilities and the prevalence of PE in our study were very similar to those seen in a number of prospective studies.1820, 26, 27 Although other risk indices exist, the Wells score has been shown to have predictive powers equal to other algorithms and to clinicians; implicit assessments.28, 29 In our cohort, 35.1% of patients with a high pretest probability were diagnosed with PE; although this was lower than that in the initial Wells cohort,18 it was very similar to a subsequent validation study using the Wells algorithm, in which the prevalence of PE in patients with high pretest probability was 37.5%.27 Plasma D‐dimer testing is not routinely used at our hospitals, but it is a component of some CTPA‐based diagnostic algorithms.1114 Although use of D‐dimer testing may have led to fewer scans in patients with negative D‐dimer test results and low pretest probability,30 the high false‐positive rate for D‐dimer assays31 makes it difficult to predict the effect of widespread D‐dimer use on the overall pretest probability distribution. Using our assumptions about CT test characteristics, a pretest probability of more than 30% is required to generate a posttest probability of PE of at least 90% (the traditional treatment threshold for anticoagulant therapy22) with a positive scan. Extensive D‐dimer use would be unlikely to cause such a shift in the distribution of pretest probabilities.

Finally, CT technology has continued to advance, and many institutions now use 64‐slice scanners32 in contrast to the single‐slice scanners in use at the time our data were collected. Our assumptions were that CTPA has a positive likelihood ratio of 18.0 and a negative likelihood ratio of 0.1 (corresponding to a sensitivity of 90% and a specificity of 95%), although many studies of single‐detector CTPA found less impressive values.5, 7 Multidetector CT is thought to be more accurate than was earlier technology, but the true diagnostic performance of multidetector CT is not yet known. However, our findings pertain primarily to clinicians' responses to test results, so even if newer scanners are more accurate, Bayesian analysis will still be required in order to appropriately treat patients. A recent meta‐analysis of diagnostic strategies for PE found CTPA to have a positive likelihood ratio of 24.1, but even using this higher value, patients with a low pretest probability and positive CTPA still have a posttest probability of PE below the traditional treatment threshold.33 As most patients undergoing evaluation for suspected PE have a low pretest probability,17 a substantial number of false‐positive diagnoses of PE may still occur, even with a more accurate diagnostic test.

CT pulmonary angiography has become the first‐line test for pulmonary embolism at our institution, a situation likely mirrored elsewhere. CTPA is safe and rapid and offers the advantage of revealing ancillary lung findings that may be clinically significant.12 Although the test is an important addition to a clinician's diagnostic armamentarium, Bayesian analysis must be used to interpret its results, especially when CTPA is used as the first‐line diagnostic test. Our data raise the troubling concern that reliance on CTPA as the sole diagnostic test for suspected pulmonary embolism may result in a large number of patients with false‐positive CT scans receiving anticoagulation treatment.

Spiral computed tomographic pulmonary angiography (CTPA) is a common first‐line test for the evaluation of suspected pulmonary embolism (PE). At our institution CTPA became the initial diagnostic study in 83% of patients with suspected PE within 3 years of the introduction of CT,1 and by 2001 CTPA had become the most common diagnostic test performed nationwide in patients diagnosed with PE.2 Most scans are interpreted as either positive or negative for pulmonary embolism, providing clinicians with a greater sense of diagnostic certainty than with the probabilistic results of lung scintigraphy. Initial studies of CTPA supported this appearance of diagnostic certainty, reporting sensitivity and specificity of greater than 90%,3, 4 but several subsequent studies have failed to reproduce these results.57 Newer multidetector CT scans are believed to be more accurate than earlier single‐detector CT,8 but true estimates of CTPA test characteristics will not be known until publication of the forthcoming PIOPED II study.9

Even without these data, CT‐based diagnostic algorithms have already appeared.1014 These algorithms generally focus on minimizing the false‐negative rate through use of serial testing (involving combinations of serum D‐dimer, lower‐extremity ultrasound, and CTPA). A recent meta‐analysis demonstrated that negative CTPA is highly accurate at ruling out PE, with test characteristics similar to conventional pulmonary angiography.15 Another meta‐analysis found that the 3‐month rate of subsequent venous thromboembolism after negative CTPA was 1.4% (95% CI 1.1%‐1.8%),16 supporting the strategy of withholding anticoagulants after negative CTPA in combination with other tests. However, use of serial testing to establish the diagnosis of PE and initiate anticoagulation has not been systematically evaluated or recommended, even for patients with a low pretest probability of PE.17

To assess the potential impact of these algorithms on the diagnosis of PE in clinical practice, we analyzed the clinical presentation and treatment of a cohort of patients at our institution who underwent CTPA for suspected PE.1 We calculated a range of posttest probabilities for pulmonary embolism for these patients, given the pretest probabilities, test results, and estimates of CTPA test characteristics. We then compared the treatment decisions of clinicians to the posttest probabilities of PE in order to establish the potential frequency of false‐positive and false‐negative diagnoses and to determine if patients were treated appropriately based on these estimates.

METHODS

Sites and Subjects

Details of the sites, subjects, and methods used to collect patient‐level data in this analysis have been previously published.1 The study was performed at Moffitt‐Long Hospital and San Francisco General Hospital, teaching hospitals affiliated with the University of California San Francisco School of Medicine. At both sites, single‐detector CT scans were available 24 hours a day throughout the study period and were read by attending radiologists who specialized in thoracic imaging. We excluded patients whose CTPA was not completed as the initial test in the evaluation of suspected PE, those who underwent testing for any indication other than suspected acute PE, and those with incomplete medical records or technically inadequate CTPA.

We randomly selected 345 patients who underwent CTPA between January 1, 1998, and December 31, 2000, from the Radiology Department databases. One investigator (R.L.T.) then abstracted charts of all patients. For each subject, we collected data about history and clinical presentation, diagnostic impressions of the treating clinicians, treatments administered both before and after diagnostic testing, CTPA result, results of other diagnostic tests for PE, and final clinical diagnosis. During the study period, there were no institution‐ or department‐specific guidelines or decision aids available for the diagnosis of PE. Ventilation‐perfusion scan, lower extremity ultrasound, and pulmonary angiography were available, but highly sensitive D‐dimer assays were not in use. The study was approved by the Institutional Review Boards of both sites, and requirement for written informed consent from patients was waived.

Estimates of Pretest Probabilities of Pulmonary Embolism and CTPA Test Characteristics

Several prediction rules1820 generate clinical pretest probabilities for patients with suspected PE. We used the Wells score18 to assign a pretest probability of low, moderate, or high to each patient on the basis of the following clinical variables: leg swelling, hemoptysis, tachycardia, history of recent immobilization, history of prior DVT or PE, active malignancy, and lack of a more likely alternative diagnosis. We chose this rule as (unlike other prediction rules such as the Geneva rule20) the Wells score has been validated for hospitalized patients with suspected PE and does not require arterial blood gas measurements. The prevalence of PE reported in the evaluation of the Wells score was 3.4%, 27.8%, and 78.3% for low, moderate, and high pretest probabilities, respectively.18

As in our previous study,1 we assumed CTPA to be 90% sensitive and 95% specific based on published estimates.3, 17 These values correspond to a positive likelihood ratio of 18 and a negative likelihood ratio of 0.1.21 We chose these values as a best‐case estimate of the test characteristics of CTPA, although other studies have found less impressive results.7 Using these pretest probabilities and likelihood ratios, we then used Bayes' theorem (Figure 1) to calculate the range of expected posttest probabilities of pulmonary embolism.

Figure 1

RESULTS

DISCUSSION

Physicians at our institution regarded CTPA results as definitive, anticoagulating 96.5% of patients with a positive CT and withholding treatment in 95.8% of patients with a negative scan. This practice pattern may result in unnecessary anticoagulation of many patients with a low pretest probability of PE who may have had false‐positive CTPA findings. In contrast, the rate of possible false‐negative diagnosis of PE was low, consistent with the results of several other studies.16

The use of CTPA is likely to increase because of the publication of multiple algorithms advocating that CTPA be the chief imaging study used in the diagnosis of PE.1014 These algorithms recommend serial testing on patients with a negative CTPA in order to minimize the false‐negative rate, but they do not require systematic follow‐up in patients with a positive scan, even if the pretest probability was low. In management trials, this approach resulted in a low false‐negative rate (1.0%‐1.8% at 3‐month follow‐up).1114 However, the rate of major bleeding in patients treated for PE was 3.2%‐6.0% at 3 months,1214 illustrating the potential risk of anticoagulating patients who may have false‐positive diagnoses. Furthermore, premature diagnostic closure after a CTPA positive for PE may result in additional morbidity as a result of missing the true diagnosis.

One potential explanation for the large number of potential false‐positive emboli seen in low‐risk patients is that it is difficult to accurately diagnose distal pulmonary emboli with CTPA. The interrater reliability of CTPA for diagnosis of subsegmental PE is suboptimal,23 and the clinical significance of these emboli remains uncertain.24 Thus, many emboli found in patients with low pretest probability actually may have been subsegmental PE that would not have been diagnosed by another radiologist. As CTPA is more accurate for diagnosing central PE,25 clinicians should consider reviewing positive scans with the interpreting radiologist, especially when the pretest probability was low and the filling defects identified are in distal vessels.

Our results may also illustrate that clinicians have a lower treatment threshold when presented with apparently definitive evidence of pulmonary embolism. Previous proposals on the appropriate treatment threshold for PE, which used Bayesian decision‐making methods similar to ours,22 incorporated PIOPED26 data on the pretest probability of pulmonary embolism, the test characteristics of ventilation‐perfusion scans, and the clinical outcomes of patients in each test result/pretest probability category. However, there is no corresponding data for CTPA, as its test characteristics are still uncertain, and long‐term clinical outcomes have not been documented for patients treated (or not treated) on the basis of CT results.

Our study had several limitations. First, charting bias potentially was introduced by our using a retrospective method of collecting data for calculating pretest probabilities. To address this potential bias, we collected data from the entire medical record, including information available at and preceding the time of the CT scan. We believe this method was effective, as the range of pretest probabilities and the prevalence of PE in our study were very similar to those seen in a number of prospective studies.1820, 26, 27 Although other risk indices exist, the Wells score has been shown to have predictive powers equal to other algorithms and to clinicians; implicit assessments.28, 29 In our cohort, 35.1% of patients with a high pretest probability were diagnosed with PE; although this was lower than that in the initial Wells cohort,18 it was very similar to a subsequent validation study using the Wells algorithm, in which the prevalence of PE in patients with high pretest probability was 37.5%.27 Plasma D‐dimer testing is not routinely used at our hospitals, but it is a component of some CTPA‐based diagnostic algorithms.1114 Although use of D‐dimer testing may have led to fewer scans in patients with negative D‐dimer test results and low pretest probability,30 the high false‐positive rate for D‐dimer assays31 makes it difficult to predict the effect of widespread D‐dimer use on the overall pretest probability distribution. Using our assumptions about CT test characteristics, a pretest probability of more than 30% is required to generate a posttest probability of PE of at least 90% (the traditional treatment threshold for anticoagulant therapy22) with a positive scan. Extensive D‐dimer use would be unlikely to cause such a shift in the distribution of pretest probabilities.

Finally, CT technology has continued to advance, and many institutions now use 64‐slice scanners32 in contrast to the single‐slice scanners in use at the time our data were collected. Our assumptions were that CTPA has a positive likelihood ratio of 18.0 and a negative likelihood ratio of 0.1 (corresponding to a sensitivity of 90% and a specificity of 95%), although many studies of single‐detector CTPA found less impressive values.5, 7 Multidetector CT is thought to be more accurate than was earlier technology, but the true diagnostic performance of multidetector CT is not yet known. However, our findings pertain primarily to clinicians' responses to test results, so even if newer scanners are more accurate, Bayesian analysis will still be required in order to appropriately treat patients. A recent meta‐analysis of diagnostic strategies for PE found CTPA to have a positive likelihood ratio of 24.1, but even using this higher value, patients with a low pretest probability and positive CTPA still have a posttest probability of PE below the traditional treatment threshold.33 As most patients undergoing evaluation for suspected PE have a low pretest probability,17 a substantial number of false‐positive diagnoses of PE may still occur, even with a more accurate diagnostic test.

CT pulmonary angiography has become the first‐line test for pulmonary embolism at our institution, a situation likely mirrored elsewhere. CTPA is safe and rapid and offers the advantage of revealing ancillary lung findings that may be clinically significant.12 Although the test is an important addition to a clinician's diagnostic armamentarium, Bayesian analysis must be used to interpret its results, especially when CTPA is used as the first‐line diagnostic test. Our data raise the troubling concern that reliance on CTPA as the sole diagnostic test for suspected pulmonary embolism may result in a large number of patients with false‐positive CT scans receiving anticoagulation treatment.

Spiral computed tomographic pulmonary angiography (CTPA) is a common first‐line test for the evaluation of suspected pulmonary embolism (PE). At our institution CTPA became the initial diagnostic study in 83% of patients with suspected PE within 3 years of the introduction of CT,1 and by 2001 CTPA had become the most common diagnostic test performed nationwide in patients diagnosed with PE.2 Most scans are interpreted as either positive or negative for pulmonary embolism, providing clinicians with a greater sense of diagnostic certainty than with the probabilistic results of lung scintigraphy. Initial studies of CTPA supported this appearance of diagnostic certainty, reporting sensitivity and specificity of greater than 90%,3, 4 but several subsequent studies have failed to reproduce these results.57 Newer multidetector CT scans are believed to be more accurate than earlier single‐detector CT,8 but true estimates of CTPA test characteristics will not be known until publication of the forthcoming PIOPED II study.9

Even without these data, CT‐based diagnostic algorithms have already appeared.1014 These algorithms generally focus on minimizing the false‐negative rate through use of serial testing (involving combinations of serum D‐dimer, lower‐extremity ultrasound, and CTPA). A recent meta‐analysis demonstrated that negative CTPA is highly accurate at ruling out PE, with test characteristics similar to conventional pulmonary angiography.15 Another meta‐analysis found that the 3‐month rate of subsequent venous thromboembolism after negative CTPA was 1.4% (95% CI 1.1%‐1.8%),16 supporting the strategy of withholding anticoagulants after negative CTPA in combination with other tests. However, use of serial testing to establish the diagnosis of PE and initiate anticoagulation has not been systematically evaluated or recommended, even for patients with a low pretest probability of PE.17

To assess the potential impact of these algorithms on the diagnosis of PE in clinical practice, we analyzed the clinical presentation and treatment of a cohort of patients at our institution who underwent CTPA for suspected PE.1 We calculated a range of posttest probabilities for pulmonary embolism for these patients, given the pretest probabilities, test results, and estimates of CTPA test characteristics. We then compared the treatment decisions of clinicians to the posttest probabilities of PE in order to establish the potential frequency of false‐positive and false‐negative diagnoses and to determine if patients were treated appropriately based on these estimates.

METHODS

Sites and Subjects

Details of the sites, subjects, and methods used to collect patient‐level data in this analysis have been previously published.1 The study was performed at Moffitt‐Long Hospital and San Francisco General Hospital, teaching hospitals affiliated with the University of California San Francisco School of Medicine. At both sites, single‐detector CT scans were available 24 hours a day throughout the study period and were read by attending radiologists who specialized in thoracic imaging. We excluded patients whose CTPA was not completed as the initial test in the evaluation of suspected PE, those who underwent testing for any indication other than suspected acute PE, and those with incomplete medical records or technically inadequate CTPA.

We randomly selected 345 patients who underwent CTPA between January 1, 1998, and December 31, 2000, from the Radiology Department databases. One investigator (R.L.T.) then abstracted charts of all patients. For each subject, we collected data about history and clinical presentation, diagnostic impressions of the treating clinicians, treatments administered both before and after diagnostic testing, CTPA result, results of other diagnostic tests for PE, and final clinical diagnosis. During the study period, there were no institution‐ or department‐specific guidelines or decision aids available for the diagnosis of PE. Ventilation‐perfusion scan, lower extremity ultrasound, and pulmonary angiography were available, but highly sensitive D‐dimer assays were not in use. The study was approved by the Institutional Review Boards of both sites, and requirement for written informed consent from patients was waived.

Estimates of Pretest Probabilities of Pulmonary Embolism and CTPA Test Characteristics

Several prediction rules1820 generate clinical pretest probabilities for patients with suspected PE. We used the Wells score18 to assign a pretest probability of low, moderate, or high to each patient on the basis of the following clinical variables: leg swelling, hemoptysis, tachycardia, history of recent immobilization, history of prior DVT or PE, active malignancy, and lack of a more likely alternative diagnosis. We chose this rule as (unlike other prediction rules such as the Geneva rule20) the Wells score has been validated for hospitalized patients with suspected PE and does not require arterial blood gas measurements. The prevalence of PE reported in the evaluation of the Wells score was 3.4%, 27.8%, and 78.3% for low, moderate, and high pretest probabilities, respectively.18

As in our previous study,1 we assumed CTPA to be 90% sensitive and 95% specific based on published estimates.3, 17 These values correspond to a positive likelihood ratio of 18 and a negative likelihood ratio of 0.1.21 We chose these values as a best‐case estimate of the test characteristics of CTPA, although other studies have found less impressive results.7 Using these pretest probabilities and likelihood ratios, we then used Bayes' theorem (Figure 1) to calculate the range of expected posttest probabilities of pulmonary embolism.

Figure 1

RESULTS

DISCUSSION

Physicians at our institution regarded CTPA results as definitive, anticoagulating 96.5% of patients with a positive CT and withholding treatment in 95.8% of patients with a negative scan. This practice pattern may result in unnecessary anticoagulation of many patients with a low pretest probability of PE who may have had false‐positive CTPA findings. In contrast, the rate of possible false‐negative diagnosis of PE was low, consistent with the results of several other studies.16

The use of CTPA is likely to increase because of the publication of multiple algorithms advocating that CTPA be the chief imaging study used in the diagnosis of PE.1014 These algorithms recommend serial testing on patients with a negative CTPA in order to minimize the false‐negative rate, but they do not require systematic follow‐up in patients with a positive scan, even if the pretest probability was low. In management trials, this approach resulted in a low false‐negative rate (1.0%‐1.8% at 3‐month follow‐up).1114 However, the rate of major bleeding in patients treated for PE was 3.2%‐6.0% at 3 months,1214 illustrating the potential risk of anticoagulating patients who may have false‐positive diagnoses. Furthermore, premature diagnostic closure after a CTPA positive for PE may result in additional morbidity as a result of missing the true diagnosis.

One potential explanation for the large number of potential false‐positive emboli seen in low‐risk patients is that it is difficult to accurately diagnose distal pulmonary emboli with CTPA. The interrater reliability of CTPA for diagnosis of subsegmental PE is suboptimal,23 and the clinical significance of these emboli remains uncertain.24 Thus, many emboli found in patients with low pretest probability actually may have been subsegmental PE that would not have been diagnosed by another radiologist. As CTPA is more accurate for diagnosing central PE,25 clinicians should consider reviewing positive scans with the interpreting radiologist, especially when the pretest probability was low and the filling defects identified are in distal vessels.

Our results may also illustrate that clinicians have a lower treatment threshold when presented with apparently definitive evidence of pulmonary embolism. Previous proposals on the appropriate treatment threshold for PE, which used Bayesian decision‐making methods similar to ours,22 incorporated PIOPED26 data on the pretest probability of pulmonary embolism, the test characteristics of ventilation‐perfusion scans, and the clinical outcomes of patients in each test result/pretest probability category. However, there is no corresponding data for CTPA, as its test characteristics are still uncertain, and long‐term clinical outcomes have not been documented for patients treated (or not treated) on the basis of CT results.

Our study had several limitations. First, charting bias potentially was introduced by our using a retrospective method of collecting data for calculating pretest probabilities. To address this potential bias, we collected data from the entire medical record, including information available at and preceding the time of the CT scan. We believe this method was effective, as the range of pretest probabilities and the prevalence of PE in our study were very similar to those seen in a number of prospective studies.1820, 26, 27 Although other risk indices exist, the Wells score has been shown to have predictive powers equal to other algorithms and to clinicians; implicit assessments.28, 29 In our cohort, 35.1% of patients with a high pretest probability were diagnosed with PE; although this was lower than that in the initial Wells cohort,18 it was very similar to a subsequent validation study using the Wells algorithm, in which the prevalence of PE in patients with high pretest probability was 37.5%.27 Plasma D‐dimer testing is not routinely used at our hospitals, but it is a component of some CTPA‐based diagnostic algorithms.1114 Although use of D‐dimer testing may have led to fewer scans in patients with negative D‐dimer test results and low pretest probability,30 the high false‐positive rate for D‐dimer assays31 makes it difficult to predict the effect of widespread D‐dimer use on the overall pretest probability distribution. Using our assumptions about CT test characteristics, a pretest probability of more than 30% is required to generate a posttest probability of PE of at least 90% (the traditional treatment threshold for anticoagulant therapy22) with a positive scan. Extensive D‐dimer use would be unlikely to cause such a shift in the distribution of pretest probabilities.

Finally, CT technology has continued to advance, and many institutions now use 64‐slice scanners32 in contrast to the single‐slice scanners in use at the time our data were collected. Our assumptions were that CTPA has a positive likelihood ratio of 18.0 and a negative likelihood ratio of 0.1 (corresponding to a sensitivity of 90% and a specificity of 95%), although many studies of single‐detector CTPA found less impressive values.5, 7 Multidetector CT is thought to be more accurate than was earlier technology, but the true diagnostic performance of multidetector CT is not yet known. However, our findings pertain primarily to clinicians' responses to test results, so even if newer scanners are more accurate, Bayesian analysis will still be required in order to appropriately treat patients. A recent meta‐analysis of diagnostic strategies for PE found CTPA to have a positive likelihood ratio of 24.1, but even using this higher value, patients with a low pretest probability and positive CTPA still have a posttest probability of PE below the traditional treatment threshold.33 As most patients undergoing evaluation for suspected PE have a low pretest probability,17 a substantial number of false‐positive diagnoses of PE may still occur, even with a more accurate diagnostic test.

CT pulmonary angiography has become the first‐line test for pulmonary embolism at our institution, a situation likely mirrored elsewhere. CTPA is safe and rapid and offers the advantage of revealing ancillary lung findings that may be clinically significant.12 Although the test is an important addition to a clinician's diagnostic armamentarium, Bayesian analysis must be used to interpret its results, especially when CTPA is used as the first‐line diagnostic test. Our data raise the troubling concern that reliance on CTPA as the sole diagnostic test for suspected pulmonary embolism may result in a large number of patients with false‐positive CT scans receiving anticoagulation treatment.

I recently performed a PubMed search for hospitalists, which returned 561 citations, yet a second search for pediatric hospitalists produced only 37 citations. Growing up in Boston as a sports fan, my memories are filled with images that parallel these findings. One particularly vivid memory is of a Patriots game years ago. During that game, a very dynamic member of the opposing team was caught on camera picking up a phone on the sideline and telling the caller to call out the National Guard because we are killing the Patriots.

Now, pediatric hospital medicine is hardly being killed, and admittedly there were several methodological flaws in how I collected my data. However, this gap in number of publications must shrink if pediatric hospital medicine is to thrive. Like both hospital medicine and emergency medicine before it, pediatric hospital medicine must demonstrate what makes the field distinct and unique if is to be truly recognized as a medical subspecialty. The surest way to succeed in this endeavor is through the dissemination of information via peer‐reviewed journals such as the Journal of Hospital Medicine, potentially an ideal home for us.

It is important to note that dissemination of information is not limited to publication of original research. Pediatric hospital medicine is primarily a clinical field, and as such, practitioners may be spending 80%‐90% of their time (or more) caring for patients. This obviously does not leave much time for other academic pursuits. That being said, sharing many kinds of information can promote excellence in the care of hospitalized pediatric patients. Here are some types of articles that may prove useful.

• Writing that integrates, rather than discovers, new knowledge

• Review articles addressing the diagnosis and treatment of clinical conditions

• Illustrative case reports or series drawn from clinical practice

• Descriptions of best practice

• QI/QA programs

• Patient safety initiatives

• Use of decision support or other information technology tools

• Strategies to maintain physician wellness and career longevity

• Creation of educational curricula or competency assessment methods

• Leadership and professional development

This suggestion to share information of many types is not meant to downplay the importance of original research. As pediatric hospital medicine grows, its research component must grow as well in order to continually define and redefine the field itself, especially with regard to collaborative studies. In the future, it will no longer be acceptable for pediatric hospital programs to be practicing in isolation, without regard for nationally recognized and published benchmarks or other measures of quality. However, I believe that it is equally important for individuals to have outlets for these other forms of scholarship. Both the Society of Hospital Medicine and the Journal of Hospital Medicine are committed to the growth and development of pediatric hospital medicine. We encourage pediatric hospitalists to submit manuscripts and to become reviewers. You can do both at http://mc.manuscriptcentral.com/jhm.

I recently performed a PubMed search for hospitalists, which returned 561 citations, yet a second search for pediatric hospitalists produced only 37 citations. Growing up in Boston as a sports fan, my memories are filled with images that parallel these findings. One particularly vivid memory is of a Patriots game years ago. During that game, a very dynamic member of the opposing team was caught on camera picking up a phone on the sideline and telling the caller to call out the National Guard because we are killing the Patriots.

Now, pediatric hospital medicine is hardly being killed, and admittedly there were several methodological flaws in how I collected my data. However, this gap in number of publications must shrink if pediatric hospital medicine is to thrive. Like both hospital medicine and emergency medicine before it, pediatric hospital medicine must demonstrate what makes the field distinct and unique if is to be truly recognized as a medical subspecialty. The surest way to succeed in this endeavor is through the dissemination of information via peer‐reviewed journals such as the Journal of Hospital Medicine, potentially an ideal home for us.

It is important to note that dissemination of information is not limited to publication of original research. Pediatric hospital medicine is primarily a clinical field, and as such, practitioners may be spending 80%‐90% of their time (or more) caring for patients. This obviously does not leave much time for other academic pursuits. That being said, sharing many kinds of information can promote excellence in the care of hospitalized pediatric patients. Here are some types of articles that may prove useful.

• Writing that integrates, rather than discovers, new knowledge

• Review articles addressing the diagnosis and treatment of clinical conditions

• Illustrative case reports or series drawn from clinical practice

• Descriptions of best practice

• QI/QA programs

• Patient safety initiatives

• Use of decision support or other information technology tools

• Strategies to maintain physician wellness and career longevity

• Creation of educational curricula or competency assessment methods

• Leadership and professional development

This suggestion to share information of many types is not meant to downplay the importance of original research. As pediatric hospital medicine grows, its research component must grow as well in order to continually define and redefine the field itself, especially with regard to collaborative studies. In the future, it will no longer be acceptable for pediatric hospital programs to be practicing in isolation, without regard for nationally recognized and published benchmarks or other measures of quality. However, I believe that it is equally important for individuals to have outlets for these other forms of scholarship. Both the Society of Hospital Medicine and the Journal of Hospital Medicine are committed to the growth and development of pediatric hospital medicine. We encourage pediatric hospitalists to submit manuscripts and to become reviewers. You can do both at http://mc.manuscriptcentral.com/jhm.

I recently performed a PubMed search for hospitalists, which returned 561 citations, yet a second search for pediatric hospitalists produced only 37 citations. Growing up in Boston as a sports fan, my memories are filled with images that parallel these findings. One particularly vivid memory is of a Patriots game years ago. During that game, a very dynamic member of the opposing team was caught on camera picking up a phone on the sideline and telling the caller to call out the National Guard because we are killing the Patriots.

Now, pediatric hospital medicine is hardly being killed, and admittedly there were several methodological flaws in how I collected my data. However, this gap in number of publications must shrink if pediatric hospital medicine is to thrive. Like both hospital medicine and emergency medicine before it, pediatric hospital medicine must demonstrate what makes the field distinct and unique if is to be truly recognized as a medical subspecialty. The surest way to succeed in this endeavor is through the dissemination of information via peer‐reviewed journals such as the Journal of Hospital Medicine, potentially an ideal home for us.

It is important to note that dissemination of information is not limited to publication of original research. Pediatric hospital medicine is primarily a clinical field, and as such, practitioners may be spending 80%‐90% of their time (or more) caring for patients. This obviously does not leave much time for other academic pursuits. That being said, sharing many kinds of information can promote excellence in the care of hospitalized pediatric patients. Here are some types of articles that may prove useful.

• Writing that integrates, rather than discovers, new knowledge

• Review articles addressing the diagnosis and treatment of clinical conditions

• Illustrative case reports or series drawn from clinical practice

• Descriptions of best practice

• QI/QA programs

• Patient safety initiatives

• Use of decision support or other information technology tools

• Strategies to maintain physician wellness and career longevity

• Creation of educational curricula or competency assessment methods

• Leadership and professional development

This suggestion to share information of many types is not meant to downplay the importance of original research. As pediatric hospital medicine grows, its research component must grow as well in order to continually define and redefine the field itself, especially with regard to collaborative studies. In the future, it will no longer be acceptable for pediatric hospital programs to be practicing in isolation, without regard for nationally recognized and published benchmarks or other measures of quality. However, I believe that it is equally important for individuals to have outlets for these other forms of scholarship. Both the Society of Hospital Medicine and the Journal of Hospital Medicine are committed to the growth and development of pediatric hospital medicine. We encourage pediatric hospitalists to submit manuscripts and to become reviewers. You can do both at http://mc.manuscriptcentral.com/jhm.

The term hospitalist was coined in 1996 in an article1 that appeared in the New England Journal of Medicine. Robert M. Wachter, MD, and Lee Goldman, MD, of the University of California, San Francisco, defined hospitalists as hospital‐based physicians who take responsibility for managing medical inpatients. Hospitalists were described as having responsibility for seeing unassigned hospital patients and being available for in‐hospital consultations. Several years later, the Society of Hospital Medicine posted the definition of a hospitalist as someone whose primary professional focus is the medical care of hospitalized patientsin patient care, education, research, and administrative activities.

In January 2002, Wachter and Goldman published a follow‐up article,2 The Hospitalist Movement 5 Years Later, in the Journal of the American Medical Association. This formal review of 19 published studies analyzed the impact of hospital medicine groups on financial and clinical outcomes. Wachter and Goldman concluded, Empirical research supports the premise that hospitalists improve inpatient efficiency without harmful effects on quality or patient satisfaction. These studies indicate an average reduction of cost per stay of 13.4% and an average reduction in length of stay of 16.6%.

The evolution of the hospitalist movement has been fast paced and extensive. Given the recent pace of growth, a scholarly analysis estimated that the mature hospitalist workforce in the United States will eventually total 20,000, making it the equivalent of the cardiology specialty.3 Beyond sheer growth, medical literature has demonstrated positive effects of the hospitalist model on patient quality outcomes, including readmission rates, postoperative complications, and mortality.47

In addition to peer‐reviewed medical literature, there is anecdotal evidence about the growth and effects of the hospitalist movement:

• The Society of Hospital Medicine (SHM), the hospitalist professional society, estimated that in 2003 there were 8000 physicians practicing as hospitalists in the United States.8

• Twelve of the country's top 15 hospitals have hospital medicine groups.8

As hospital medicine groups have proliferated, 4 major employment models have evolved. Hospitalists can be employees of: 1) a hospital or a hospital subsidiary; 2) a multispecialty or primary care physician group; 3) a medical group (local or national) of independent hospitalists; or 4) a university or medical school. However, there is little published data on the prevalence of each of these hospitalist employment models, nationally or by type of hospital.

To better understand the extent and nature of the hospitalist movement, the American Hospital Association (AHA) utilized its 2003 Annual Survey to gather data on hospital medicine groups in the United States

DATA AND METHODS

The data for our analysis came from the 2003 AHA Annual Survey. Conducted since 1946, this survey is the principal data collection mechanism of the American Hospital Association and is a basic source of data on hospitals in the United States about the availability of services, utilization, personnel, finances, and governance. Its main purpose is to provide a cross‐sectional view of hospitals and hospital performance over time. In the 2003 survey, a series of items were added about hospitalists including whether hospitals had hospital medicine groups, the number of hospitalists operating in such groups, and the employment model used.

The study population for this analysis was limited to US community hospitals (n = 4895). Community hospitals are defined as all adult and pediatric nonfederal, short‐term general, and specialty hospitals whose facilities and services are available to the public. Excluded from the analysis were all federal hospitals, long‐term care hospitals, and psychiatric hospitals.

RESULTS