Communication is the foundation of medical care.¹ Effective communication can improve health outcomes, safety, adherence, satisfaction, trust, and enable genuine informed consent and decision-making.^2-9 Furthermore, high-quality communication increases provider engagement and workplace satisfaction, while reducing stress and malpractice risk.^10-15

Direct measurement of communication in the healthcare setting can be challenging. The “Four Habits Model,” which is derived from a synthesis of empiric studies^8,16-20 and theoretical models^21-24 of communication, offers 1 framework for assessing healthcare communication. The conceptual model underlying the 4 habits has been validated in studies of physician and patient satisfaction.^1,4,25-27 The 4 habits are: investing in the beginning, eliciting the patient’s perspective, demonstrating empathy, and investing in the end. Each habit is divided into several identifiable tasks or skill sets, which can be reliably measured using validated tools and checklists.²⁸ One such instrument, the Four Habits Coding Scheme (4HCS), has been evaluated against other tools and demonstrated overall satisfactory inter-rater reliability and validity.^29,30

The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey, developed under the direction of the Centers for Medicare and Medicaid Services (CMS) and the Agency for Healthcare Research and Quality, is an established national standard for measuring patient perceptions of care. HCAHPS retrospectively measures global perceptions of communication, support and empathy from physicians and staff, processes of care, and the overall patient experience. HCAHPS scores were first collected nationally in 2006 and have been publicly reported since 2008.³¹ With the introduction of value-based purchasing in 2012, health system revenues are now tied to HCAHPS survey performance.³² As a result, hospitals are financially motivated to improve HCAHPS scores but lack evidence-based methods for doing so. Some healthcare organizations have invested in communication training programs based on the available literature and best practices.^2,33-35 However, it is not known how, if at all, HCAHPS scores relate to physicians’ real-time observed communication skills.

To examine the relationship between physician communication, as reported by global HCAHPS scores, and the quality of physician communication skills in specific encounters, we observed hospitalist physicians during inpatient bedside rounds and measured their communication skills using the 4HCS.

Study Design

The study utilized a cross sectional design; physicians who consented were observed on rounds during 3 separate encounters, and we compared hospitalists’ 4HCS scores to their HCAHPS scores to assess the correlation. The study was approved by the Institutional Review Board of the Cleveland Clinic.

Population

The study was conducted at the main campus of the Cleveland Clinic. All physicians specializing in hospital medicine who had received 10 or more completed HCAHPS survey responses while rounding on a medicine service in the past year were invited to participate in the study. Participation was voluntary; night hospitalists were excluded. A research nurse was trained in the Four Habits Model²⁸ and in the use of the 4HCS coding scheme by the principal investigator. The nurse observed each physician and ascertained the presence of communication behaviors using the 4HCS tool. Physicians were observed between August 2013 and August 2014. Multiple observations per physician could occur on the same day, but only 1 observation per patient was used for analysis. Observations consisted of a physician’s first encounter with a hospitalized patient, with the patient’s consent. Observations were conducted during encounters with English-speaking and cognitively intact patients only. Resident physicians were permitted to stay and conduct rounds per their normal routine. Patient information was not collected as part of the study.

Measures

HCAHPS. For each physician, we extracted all HCAHPS scores that were collected from our hospital’s Press Ganey database. The HCAHPS survey contains 22 core questions divided into 7 themes or domains, 1 of which is doctor communication. The survey uses frequency-based questions with possible answers fixed on a 4-point scale (4=always, 3=usually, 2=sometimes, 1=never). Our primary outcome was the doctor communication domain, which comprises 3 questions: 1) During this hospital stay, how often did the doctors treat you with respect? 2) During this hospital stay, how often did the doctors listen to you? and 3) During this hospital stay, how often did the doctors explain things in a language you can understand? Because CMS counts only the percentage of responses that are graded “always,” so-called “top box” scoring, we used the same measure.

The HCAHPS scores are always attributed to the physician at the time of discharge even if he may not have been responsible for the care of the patient during the entire hospital course. To mitigate contamination from patients seen by multiple providers, we cross-matched length of stay (LOS) data with billing data to determine the proportion of days a patient was seen by a single provider during the entire length of stay. We stratified patients seen by the attending providers to less than 50%, 50% to less than 100%, and at 100% of the LOS. However, we were unable to identify which patients were seen by other consultants or by residents due to limitations in data gathering and the nature of the database.

STATISTICAL ANALYSIS

Physician characteristics were summarized with standard descriptive statistics. Pearson correlation coefficients were computed between HCAHPS and 4HCS scores. All analyses were performed with RStudio (Boston, MA). The Pearson correlation between the averaged HCAHPS and 4HCS scores was also computed. A correlation with a P value less than 0.05 was considered statistically significant. With 28 physicians, the study had a power of 88% to detect a moderate correlation (greater than 0.50) with a 2-sided alpha of 0.05. We also computed the correlations based on the subgroups of data with patients seen by providers for less than 50%, 50% to less than 100%, and 100% of LOS. All analyses were conducted in SAS 9.2 (SAS Institute Inc., Cary, NC).³⁶

There were 31 physicians who met our inclusion criteria. Of 29 volunteers, 28 were observed during 3 separate inpatient encounters and made up the final sample. A total of 1003 HCAHPS survey responses were available for these physicians. Participants were predominantly female (60.7%), with an average age of 39 years. They were in practice for an average of 4 years (12 were in practice more than 5 years), and 9 were observed on a teaching rotation.

The means of the overall 4HCS scores per observation were 17.39 ± 2.33 for the first, 17.00 ± 2.37 for the second, and 17.43 ± 2.36 for third bedside observation. The mean 4HCS scores per observation, broken down by habit, appear in Table 1. The ICC among the repeated scores within the same physician was 0.81. The median number of HCAHPS survey returns was 32 (range = [8, 85], with mean = 35.8, interquartile range = [16, 54]). The median overall HCAHPS doctor communication score was 89.6 (range = 80.9-93.7). Participants scored the highest in the respect subdomain and the lowest in the explain subdomain. Median HCAHPS scores and ranges appear in Table 2.

In this observational study of hospitalist physicians at a large tertiary care center, we found that communication skills, as measured by the 4HCS, varied substantially among physicians but were highly correlated within patients of the same physician. However, there was virtually no correlation between the attending physician of record’s 4HCS scores and their HCAHPS communication scores. When we limited our analysis to patients who saw only 1 hospitalist throughout their stay, there were moderate correlations between demonstration of empathy and both the HCAHPS respect score and overall doctor communication score. There were no trends across the strata of hospitalist involvement. It is important to note that the addition of even 1 different hospitalist to the LOS removes any association. Habits 1 and 2 are close to significance in the 100% subgroup, with a weak correlation. Interestingly, Habit 4, which focuses on creating a plan with the patient, showed no correlation at all with patients reporting that doctors explained things in language they could understand.

Development and testing of the HCAHPS survey began in 2002, commissioned by CMS and the Agency for Healthcare Research and Quality for the purpose of measuring patient experience in the hospital. The HCAHPS survey was endorsed by the National Quality Forum in 2005, with final approval of the national implementation granted by the Office of Management and Budget later that year. The CMS began implementation of the HCAHPS survey in 2006, with the first required public reporting of all hospitals taking place in March 2008.^37-41 Based on CMS’ value-based purchasing initiative, hospitals with low HCAHPS scores have faced substantial penalties since 2012. Under these circumstances, it is important that the HCAHPS measures what it purports to measure. Because HCAHPS was designed to compare hospitals, testing was limited to assessment of internal reliability, hospital-level reliability, and construct validity. External validation with known measures of physician communication was not performed.⁴¹ Our study appears to be the first to compare HCAHPS scores to directly observed measures of physician communication skills. The lack of association between the 2 should sound a cautionary note to hospitals who seek to tie individual compensation to HCAHPS scores to improve them. In particular, the survey asks for a rating for all the patient’s doctors, not just the primary hospitalist. We found that, for hospital stays with just 1 hospitalist, the HCAHPS score reflected observed expression of empathy, although the correlation was only moderate, and HCAHPS were not correlated with other communication skills. Of all communication skills, empathy may be most important. Almost the entire body of research on physician communication cites empathy as a central skill. Empathy improves patient outcomes^{1-9,13-14,16-18,42} such as adherence to treatment, loyalty, and perception of care; and provider outcomes^10-12,15 such as reduced burnout and a decreased likelihood of malpractice litigation.

It is less clear why other communication skills did not correlate with HCAHPS, but several differences in the measures themselves and how they were obtained might be responsible. It is possible that HCAHPS measures something broader than physician communication. In addition, the 4HCS was developed and normed on outpatient encounters as is true for virtually all doctor-patient coding schemes.⁴³ Little is known about inpatient communication best practices. The timing of HCAHPS may also degrade the relationship between observed and reported communication. The HCAHPS questionnaires, collected after discharge, are retrospective reconstructions that are subject to recall bias and recency effects.^44,45 In contrast, our observations took place in real time and were specific to the face-to-face interactions that take place when physicians engage patients at the bedside. Third, the response rate for HCAHPS surveys is only 30%, leading to potential sample bias.⁴⁶ Respondents represent discharged patients who are willing and able to answer surveys, and may not be representative of all hospitalized patients. Finally, as with all global questions, the meaning any individual patient assigns to terms like “respect” may vary.

Our study has several limitations. The HCAHPS and 4HCS scores were not obtained from the same sample of patients. It is possible that the patients who were observed were not representative of the patients who completed the HCAHPS surveys. In addition, the only type of encounter observed was the initial visit between the hospitalist and the patient, and did not include communication during follow-up visits or on the day of discharge. However, there was a strong ICC among the 4HCS scores, implying that the 4HCS measures an inherent physician skill, which should be consistent across patients and encounters. Coding bias of the habits by a single observer could not be excluded. High intra-class correlation could be due in part to observer preferences for particular communication styles. Our sample included only 28 physicians. Although our study was powered to rule out a moderate correlation between 4HCS scores and HCAHPS scores (Pearson correlation coefficient greater than 0.5), we cannot exclude weaker correlations. Most correlations that we observed were so small that they would not be clinically meaningful, even in a much larger sample.

Our findings that HCAHPS scores did not correlate with the communication skills of the attending of record have some important implications. In an environment of value-based purchasing, most hospital systems are interested in identifying modifiable provider behaviors that optimize efficiency and payment structures. This study shows that directly measured communication skills do not correlate with HCAHPS scores as generally reported, indicating that HCAHPS may be measuring a broader domain than only physician communication skills. Better attribution based on the proportion of care provided by an individual physician could make the scores more useful for individual comparisons, but most institutions do not report their data in this way. Given this limitation, hospitals should refrain from comparing and incentivizing individual physicians based on their HCAHPS scores, because this measure was not designed for this purpose and does not appear to reflect an individual’s skills. This is important in the current environment in which hospitals face substantial penalties for underperformance but lack specific tools to improve their scores. Furthermore, there is concern that this type of measurement creates perverse incentives that may adversely alter clinical practice with the aim of improving scores.⁴⁶

Training clinicians in communication and teaming skills is one potential means of increasing overall scores.¹⁵ Improving doctor-patient and team relationships is also the right thing to do. It is increasingly being demanded by patients and has always been a deep source of satisfaction for physicians.^15,47 Moreover, there is an increasingly robust literature that relates face-to-face communication to biomedical and psychosocial outcomes of care.⁴⁸ Identifying individual physicians who need help with communication skills is a worthwhile goal. Unfortunately, the HCAHPS survey does not appear to be the appropriate tool for this purpose.

Disclosure

The Cleveland Clinic Foundation, Division of Clinical Research, Research Programs Committees provided funding support. No funding source had any role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. The authors have no conflicts of interest for this study.

Communication is the foundation of medical care.¹ Effective communication can improve health outcomes, safety, adherence, satisfaction, trust, and enable genuine informed consent and decision-making.^2-9 Furthermore, high-quality communication increases provider engagement and workplace satisfaction, while reducing stress and malpractice risk.^10-15

Direct measurement of communication in the healthcare setting can be challenging. The “Four Habits Model,” which is derived from a synthesis of empiric studies^8,16-20 and theoretical models^21-24 of communication, offers 1 framework for assessing healthcare communication. The conceptual model underlying the 4 habits has been validated in studies of physician and patient satisfaction.^1,4,25-27 The 4 habits are: investing in the beginning, eliciting the patient’s perspective, demonstrating empathy, and investing in the end. Each habit is divided into several identifiable tasks or skill sets, which can be reliably measured using validated tools and checklists.²⁸ One such instrument, the Four Habits Coding Scheme (4HCS), has been evaluated against other tools and demonstrated overall satisfactory inter-rater reliability and validity.^29,30

The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey, developed under the direction of the Centers for Medicare and Medicaid Services (CMS) and the Agency for Healthcare Research and Quality, is an established national standard for measuring patient perceptions of care. HCAHPS retrospectively measures global perceptions of communication, support and empathy from physicians and staff, processes of care, and the overall patient experience. HCAHPS scores were first collected nationally in 2006 and have been publicly reported since 2008.³¹ With the introduction of value-based purchasing in 2012, health system revenues are now tied to HCAHPS survey performance.³² As a result, hospitals are financially motivated to improve HCAHPS scores but lack evidence-based methods for doing so. Some healthcare organizations have invested in communication training programs based on the available literature and best practices.^2,33-35 However, it is not known how, if at all, HCAHPS scores relate to physicians’ real-time observed communication skills.

To examine the relationship between physician communication, as reported by global HCAHPS scores, and the quality of physician communication skills in specific encounters, we observed hospitalist physicians during inpatient bedside rounds and measured their communication skills using the 4HCS.

Study Design

The study utilized a cross sectional design; physicians who consented were observed on rounds during 3 separate encounters, and we compared hospitalists’ 4HCS scores to their HCAHPS scores to assess the correlation. The study was approved by the Institutional Review Board of the Cleveland Clinic.

Population

The study was conducted at the main campus of the Cleveland Clinic. All physicians specializing in hospital medicine who had received 10 or more completed HCAHPS survey responses while rounding on a medicine service in the past year were invited to participate in the study. Participation was voluntary; night hospitalists were excluded. A research nurse was trained in the Four Habits Model²⁸ and in the use of the 4HCS coding scheme by the principal investigator. The nurse observed each physician and ascertained the presence of communication behaviors using the 4HCS tool. Physicians were observed between August 2013 and August 2014. Multiple observations per physician could occur on the same day, but only 1 observation per patient was used for analysis. Observations consisted of a physician’s first encounter with a hospitalized patient, with the patient’s consent. Observations were conducted during encounters with English-speaking and cognitively intact patients only. Resident physicians were permitted to stay and conduct rounds per their normal routine. Patient information was not collected as part of the study.

Measures

HCAHPS. For each physician, we extracted all HCAHPS scores that were collected from our hospital’s Press Ganey database. The HCAHPS survey contains 22 core questions divided into 7 themes or domains, 1 of which is doctor communication. The survey uses frequency-based questions with possible answers fixed on a 4-point scale (4=always, 3=usually, 2=sometimes, 1=never). Our primary outcome was the doctor communication domain, which comprises 3 questions: 1) During this hospital stay, how often did the doctors treat you with respect? 2) During this hospital stay, how often did the doctors listen to you? and 3) During this hospital stay, how often did the doctors explain things in a language you can understand? Because CMS counts only the percentage of responses that are graded “always,” so-called “top box” scoring, we used the same measure.

The HCAHPS scores are always attributed to the physician at the time of discharge even if he may not have been responsible for the care of the patient during the entire hospital course. To mitigate contamination from patients seen by multiple providers, we cross-matched length of stay (LOS) data with billing data to determine the proportion of days a patient was seen by a single provider during the entire length of stay. We stratified patients seen by the attending providers to less than 50%, 50% to less than 100%, and at 100% of the LOS. However, we were unable to identify which patients were seen by other consultants or by residents due to limitations in data gathering and the nature of the database.

STATISTICAL ANALYSIS

Physician characteristics were summarized with standard descriptive statistics. Pearson correlation coefficients were computed between HCAHPS and 4HCS scores. All analyses were performed with RStudio (Boston, MA). The Pearson correlation between the averaged HCAHPS and 4HCS scores was also computed. A correlation with a P value less than 0.05 was considered statistically significant. With 28 physicians, the study had a power of 88% to detect a moderate correlation (greater than 0.50) with a 2-sided alpha of 0.05. We also computed the correlations based on the subgroups of data with patients seen by providers for less than 50%, 50% to less than 100%, and 100% of LOS. All analyses were conducted in SAS 9.2 (SAS Institute Inc., Cary, NC).³⁶

There were 31 physicians who met our inclusion criteria. Of 29 volunteers, 28 were observed during 3 separate inpatient encounters and made up the final sample. A total of 1003 HCAHPS survey responses were available for these physicians. Participants were predominantly female (60.7%), with an average age of 39 years. They were in practice for an average of 4 years (12 were in practice more than 5 years), and 9 were observed on a teaching rotation.

The means of the overall 4HCS scores per observation were 17.39 ± 2.33 for the first, 17.00 ± 2.37 for the second, and 17.43 ± 2.36 for third bedside observation. The mean 4HCS scores per observation, broken down by habit, appear in Table 1. The ICC among the repeated scores within the same physician was 0.81. The median number of HCAHPS survey returns was 32 (range = [8, 85], with mean = 35.8, interquartile range = [16, 54]). The median overall HCAHPS doctor communication score was 89.6 (range = 80.9-93.7). Participants scored the highest in the respect subdomain and the lowest in the explain subdomain. Median HCAHPS scores and ranges appear in Table 2.

In this observational study of hospitalist physicians at a large tertiary care center, we found that communication skills, as measured by the 4HCS, varied substantially among physicians but were highly correlated within patients of the same physician. However, there was virtually no correlation between the attending physician of record’s 4HCS scores and their HCAHPS communication scores. When we limited our analysis to patients who saw only 1 hospitalist throughout their stay, there were moderate correlations between demonstration of empathy and both the HCAHPS respect score and overall doctor communication score. There were no trends across the strata of hospitalist involvement. It is important to note that the addition of even 1 different hospitalist to the LOS removes any association. Habits 1 and 2 are close to significance in the 100% subgroup, with a weak correlation. Interestingly, Habit 4, which focuses on creating a plan with the patient, showed no correlation at all with patients reporting that doctors explained things in language they could understand.

Development and testing of the HCAHPS survey began in 2002, commissioned by CMS and the Agency for Healthcare Research and Quality for the purpose of measuring patient experience in the hospital. The HCAHPS survey was endorsed by the National Quality Forum in 2005, with final approval of the national implementation granted by the Office of Management and Budget later that year. The CMS began implementation of the HCAHPS survey in 2006, with the first required public reporting of all hospitals taking place in March 2008.^37-41 Based on CMS’ value-based purchasing initiative, hospitals with low HCAHPS scores have faced substantial penalties since 2012. Under these circumstances, it is important that the HCAHPS measures what it purports to measure. Because HCAHPS was designed to compare hospitals, testing was limited to assessment of internal reliability, hospital-level reliability, and construct validity. External validation with known measures of physician communication was not performed.⁴¹ Our study appears to be the first to compare HCAHPS scores to directly observed measures of physician communication skills. The lack of association between the 2 should sound a cautionary note to hospitals who seek to tie individual compensation to HCAHPS scores to improve them. In particular, the survey asks for a rating for all the patient’s doctors, not just the primary hospitalist. We found that, for hospital stays with just 1 hospitalist, the HCAHPS score reflected observed expression of empathy, although the correlation was only moderate, and HCAHPS were not correlated with other communication skills. Of all communication skills, empathy may be most important. Almost the entire body of research on physician communication cites empathy as a central skill. Empathy improves patient outcomes^{1-9,13-14,16-18,42} such as adherence to treatment, loyalty, and perception of care; and provider outcomes^10-12,15 such as reduced burnout and a decreased likelihood of malpractice litigation.

It is less clear why other communication skills did not correlate with HCAHPS, but several differences in the measures themselves and how they were obtained might be responsible. It is possible that HCAHPS measures something broader than physician communication. In addition, the 4HCS was developed and normed on outpatient encounters as is true for virtually all doctor-patient coding schemes.⁴³ Little is known about inpatient communication best practices. The timing of HCAHPS may also degrade the relationship between observed and reported communication. The HCAHPS questionnaires, collected after discharge, are retrospective reconstructions that are subject to recall bias and recency effects.^44,45 In contrast, our observations took place in real time and were specific to the face-to-face interactions that take place when physicians engage patients at the bedside. Third, the response rate for HCAHPS surveys is only 30%, leading to potential sample bias.⁴⁶ Respondents represent discharged patients who are willing and able to answer surveys, and may not be representative of all hospitalized patients. Finally, as with all global questions, the meaning any individual patient assigns to terms like “respect” may vary.

Our study has several limitations. The HCAHPS and 4HCS scores were not obtained from the same sample of patients. It is possible that the patients who were observed were not representative of the patients who completed the HCAHPS surveys. In addition, the only type of encounter observed was the initial visit between the hospitalist and the patient, and did not include communication during follow-up visits or on the day of discharge. However, there was a strong ICC among the 4HCS scores, implying that the 4HCS measures an inherent physician skill, which should be consistent across patients and encounters. Coding bias of the habits by a single observer could not be excluded. High intra-class correlation could be due in part to observer preferences for particular communication styles. Our sample included only 28 physicians. Although our study was powered to rule out a moderate correlation between 4HCS scores and HCAHPS scores (Pearson correlation coefficient greater than 0.5), we cannot exclude weaker correlations. Most correlations that we observed were so small that they would not be clinically meaningful, even in a much larger sample.

Our findings that HCAHPS scores did not correlate with the communication skills of the attending of record have some important implications. In an environment of value-based purchasing, most hospital systems are interested in identifying modifiable provider behaviors that optimize efficiency and payment structures. This study shows that directly measured communication skills do not correlate with HCAHPS scores as generally reported, indicating that HCAHPS may be measuring a broader domain than only physician communication skills. Better attribution based on the proportion of care provided by an individual physician could make the scores more useful for individual comparisons, but most institutions do not report their data in this way. Given this limitation, hospitals should refrain from comparing and incentivizing individual physicians based on their HCAHPS scores, because this measure was not designed for this purpose and does not appear to reflect an individual’s skills. This is important in the current environment in which hospitals face substantial penalties for underperformance but lack specific tools to improve their scores. Furthermore, there is concern that this type of measurement creates perverse incentives that may adversely alter clinical practice with the aim of improving scores.⁴⁶

Training clinicians in communication and teaming skills is one potential means of increasing overall scores.¹⁵ Improving doctor-patient and team relationships is also the right thing to do. It is increasingly being demanded by patients and has always been a deep source of satisfaction for physicians.^15,47 Moreover, there is an increasingly robust literature that relates face-to-face communication to biomedical and psychosocial outcomes of care.⁴⁸ Identifying individual physicians who need help with communication skills is a worthwhile goal. Unfortunately, the HCAHPS survey does not appear to be the appropriate tool for this purpose.

Disclosure

The Cleveland Clinic Foundation, Division of Clinical Research, Research Programs Committees provided funding support. No funding source had any role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. The authors have no conflicts of interest for this study.

Communication is the foundation of medical care.¹ Effective communication can improve health outcomes, safety, adherence, satisfaction, trust, and enable genuine informed consent and decision-making.^2-9 Furthermore, high-quality communication increases provider engagement and workplace satisfaction, while reducing stress and malpractice risk.^10-15

Direct measurement of communication in the healthcare setting can be challenging. The “Four Habits Model,” which is derived from a synthesis of empiric studies^8,16-20 and theoretical models^21-24 of communication, offers 1 framework for assessing healthcare communication. The conceptual model underlying the 4 habits has been validated in studies of physician and patient satisfaction.^1,4,25-27 The 4 habits are: investing in the beginning, eliciting the patient’s perspective, demonstrating empathy, and investing in the end. Each habit is divided into several identifiable tasks or skill sets, which can be reliably measured using validated tools and checklists.²⁸ One such instrument, the Four Habits Coding Scheme (4HCS), has been evaluated against other tools and demonstrated overall satisfactory inter-rater reliability and validity.^29,30

The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey, developed under the direction of the Centers for Medicare and Medicaid Services (CMS) and the Agency for Healthcare Research and Quality, is an established national standard for measuring patient perceptions of care. HCAHPS retrospectively measures global perceptions of communication, support and empathy from physicians and staff, processes of care, and the overall patient experience. HCAHPS scores were first collected nationally in 2006 and have been publicly reported since 2008.³¹ With the introduction of value-based purchasing in 2012, health system revenues are now tied to HCAHPS survey performance.³² As a result, hospitals are financially motivated to improve HCAHPS scores but lack evidence-based methods for doing so. Some healthcare organizations have invested in communication training programs based on the available literature and best practices.^2,33-35 However, it is not known how, if at all, HCAHPS scores relate to physicians’ real-time observed communication skills.

To examine the relationship between physician communication, as reported by global HCAHPS scores, and the quality of physician communication skills in specific encounters, we observed hospitalist physicians during inpatient bedside rounds and measured their communication skills using the 4HCS.

Study Design

The study utilized a cross sectional design; physicians who consented were observed on rounds during 3 separate encounters, and we compared hospitalists’ 4HCS scores to their HCAHPS scores to assess the correlation. The study was approved by the Institutional Review Board of the Cleveland Clinic.

Population

The study was conducted at the main campus of the Cleveland Clinic. All physicians specializing in hospital medicine who had received 10 or more completed HCAHPS survey responses while rounding on a medicine service in the past year were invited to participate in the study. Participation was voluntary; night hospitalists were excluded. A research nurse was trained in the Four Habits Model²⁸ and in the use of the 4HCS coding scheme by the principal investigator. The nurse observed each physician and ascertained the presence of communication behaviors using the 4HCS tool. Physicians were observed between August 2013 and August 2014. Multiple observations per physician could occur on the same day, but only 1 observation per patient was used for analysis. Observations consisted of a physician’s first encounter with a hospitalized patient, with the patient’s consent. Observations were conducted during encounters with English-speaking and cognitively intact patients only. Resident physicians were permitted to stay and conduct rounds per their normal routine. Patient information was not collected as part of the study.

Measures

HCAHPS. For each physician, we extracted all HCAHPS scores that were collected from our hospital’s Press Ganey database. The HCAHPS survey contains 22 core questions divided into 7 themes or domains, 1 of which is doctor communication. The survey uses frequency-based questions with possible answers fixed on a 4-point scale (4=always, 3=usually, 2=sometimes, 1=never). Our primary outcome was the doctor communication domain, which comprises 3 questions: 1) During this hospital stay, how often did the doctors treat you with respect? 2) During this hospital stay, how often did the doctors listen to you? and 3) During this hospital stay, how often did the doctors explain things in a language you can understand? Because CMS counts only the percentage of responses that are graded “always,” so-called “top box” scoring, we used the same measure.

The HCAHPS scores are always attributed to the physician at the time of discharge even if he may not have been responsible for the care of the patient during the entire hospital course. To mitigate contamination from patients seen by multiple providers, we cross-matched length of stay (LOS) data with billing data to determine the proportion of days a patient was seen by a single provider during the entire length of stay. We stratified patients seen by the attending providers to less than 50%, 50% to less than 100%, and at 100% of the LOS. However, we were unable to identify which patients were seen by other consultants or by residents due to limitations in data gathering and the nature of the database.

STATISTICAL ANALYSIS

Physician characteristics were summarized with standard descriptive statistics. Pearson correlation coefficients were computed between HCAHPS and 4HCS scores. All analyses were performed with RStudio (Boston, MA). The Pearson correlation between the averaged HCAHPS and 4HCS scores was also computed. A correlation with a P value less than 0.05 was considered statistically significant. With 28 physicians, the study had a power of 88% to detect a moderate correlation (greater than 0.50) with a 2-sided alpha of 0.05. We also computed the correlations based on the subgroups of data with patients seen by providers for less than 50%, 50% to less than 100%, and 100% of LOS. All analyses were conducted in SAS 9.2 (SAS Institute Inc., Cary, NC).³⁶

There were 31 physicians who met our inclusion criteria. Of 29 volunteers, 28 were observed during 3 separate inpatient encounters and made up the final sample. A total of 1003 HCAHPS survey responses were available for these physicians. Participants were predominantly female (60.7%), with an average age of 39 years. They were in practice for an average of 4 years (12 were in practice more than 5 years), and 9 were observed on a teaching rotation.

The means of the overall 4HCS scores per observation were 17.39 ± 2.33 for the first, 17.00 ± 2.37 for the second, and 17.43 ± 2.36 for third bedside observation. The mean 4HCS scores per observation, broken down by habit, appear in Table 1. The ICC among the repeated scores within the same physician was 0.81. The median number of HCAHPS survey returns was 32 (range = [8, 85], with mean = 35.8, interquartile range = [16, 54]). The median overall HCAHPS doctor communication score was 89.6 (range = 80.9-93.7). Participants scored the highest in the respect subdomain and the lowest in the explain subdomain. Median HCAHPS scores and ranges appear in Table 2.

In this observational study of hospitalist physicians at a large tertiary care center, we found that communication skills, as measured by the 4HCS, varied substantially among physicians but were highly correlated within patients of the same physician. However, there was virtually no correlation between the attending physician of record’s 4HCS scores and their HCAHPS communication scores. When we limited our analysis to patients who saw only 1 hospitalist throughout their stay, there were moderate correlations between demonstration of empathy and both the HCAHPS respect score and overall doctor communication score. There were no trends across the strata of hospitalist involvement. It is important to note that the addition of even 1 different hospitalist to the LOS removes any association. Habits 1 and 2 are close to significance in the 100% subgroup, with a weak correlation. Interestingly, Habit 4, which focuses on creating a plan with the patient, showed no correlation at all with patients reporting that doctors explained things in language they could understand.

Development and testing of the HCAHPS survey began in 2002, commissioned by CMS and the Agency for Healthcare Research and Quality for the purpose of measuring patient experience in the hospital. The HCAHPS survey was endorsed by the National Quality Forum in 2005, with final approval of the national implementation granted by the Office of Management and Budget later that year. The CMS began implementation of the HCAHPS survey in 2006, with the first required public reporting of all hospitals taking place in March 2008.^37-41 Based on CMS’ value-based purchasing initiative, hospitals with low HCAHPS scores have faced substantial penalties since 2012. Under these circumstances, it is important that the HCAHPS measures what it purports to measure. Because HCAHPS was designed to compare hospitals, testing was limited to assessment of internal reliability, hospital-level reliability, and construct validity. External validation with known measures of physician communication was not performed.⁴¹ Our study appears to be the first to compare HCAHPS scores to directly observed measures of physician communication skills. The lack of association between the 2 should sound a cautionary note to hospitals who seek to tie individual compensation to HCAHPS scores to improve them. In particular, the survey asks for a rating for all the patient’s doctors, not just the primary hospitalist. We found that, for hospital stays with just 1 hospitalist, the HCAHPS score reflected observed expression of empathy, although the correlation was only moderate, and HCAHPS were not correlated with other communication skills. Of all communication skills, empathy may be most important. Almost the entire body of research on physician communication cites empathy as a central skill. Empathy improves patient outcomes^{1-9,13-14,16-18,42} such as adherence to treatment, loyalty, and perception of care; and provider outcomes^10-12,15 such as reduced burnout and a decreased likelihood of malpractice litigation.

It is less clear why other communication skills did not correlate with HCAHPS, but several differences in the measures themselves and how they were obtained might be responsible. It is possible that HCAHPS measures something broader than physician communication. In addition, the 4HCS was developed and normed on outpatient encounters as is true for virtually all doctor-patient coding schemes.⁴³ Little is known about inpatient communication best practices. The timing of HCAHPS may also degrade the relationship between observed and reported communication. The HCAHPS questionnaires, collected after discharge, are retrospective reconstructions that are subject to recall bias and recency effects.^44,45 In contrast, our observations took place in real time and were specific to the face-to-face interactions that take place when physicians engage patients at the bedside. Third, the response rate for HCAHPS surveys is only 30%, leading to potential sample bias.⁴⁶ Respondents represent discharged patients who are willing and able to answer surveys, and may not be representative of all hospitalized patients. Finally, as with all global questions, the meaning any individual patient assigns to terms like “respect” may vary.

Our study has several limitations. The HCAHPS and 4HCS scores were not obtained from the same sample of patients. It is possible that the patients who were observed were not representative of the patients who completed the HCAHPS surveys. In addition, the only type of encounter observed was the initial visit between the hospitalist and the patient, and did not include communication during follow-up visits or on the day of discharge. However, there was a strong ICC among the 4HCS scores, implying that the 4HCS measures an inherent physician skill, which should be consistent across patients and encounters. Coding bias of the habits by a single observer could not be excluded. High intra-class correlation could be due in part to observer preferences for particular communication styles. Our sample included only 28 physicians. Although our study was powered to rule out a moderate correlation between 4HCS scores and HCAHPS scores (Pearson correlation coefficient greater than 0.5), we cannot exclude weaker correlations. Most correlations that we observed were so small that they would not be clinically meaningful, even in a much larger sample.

Our findings that HCAHPS scores did not correlate with the communication skills of the attending of record have some important implications. In an environment of value-based purchasing, most hospital systems are interested in identifying modifiable provider behaviors that optimize efficiency and payment structures. This study shows that directly measured communication skills do not correlate with HCAHPS scores as generally reported, indicating that HCAHPS may be measuring a broader domain than only physician communication skills. Better attribution based on the proportion of care provided by an individual physician could make the scores more useful for individual comparisons, but most institutions do not report their data in this way. Given this limitation, hospitals should refrain from comparing and incentivizing individual physicians based on their HCAHPS scores, because this measure was not designed for this purpose and does not appear to reflect an individual’s skills. This is important in the current environment in which hospitals face substantial penalties for underperformance but lack specific tools to improve their scores. Furthermore, there is concern that this type of measurement creates perverse incentives that may adversely alter clinical practice with the aim of improving scores.⁴⁶

Training clinicians in communication and teaming skills is one potential means of increasing overall scores.¹⁵ Improving doctor-patient and team relationships is also the right thing to do. It is increasingly being demanded by patients and has always been a deep source of satisfaction for physicians.^15,47 Moreover, there is an increasingly robust literature that relates face-to-face communication to biomedical and psychosocial outcomes of care.⁴⁸ Identifying individual physicians who need help with communication skills is a worthwhile goal. Unfortunately, the HCAHPS survey does not appear to be the appropriate tool for this purpose.

Disclosure

The Cleveland Clinic Foundation, Division of Clinical Research, Research Programs Committees provided funding support. No funding source had any role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. The authors have no conflicts of interest for this study.

Chronic opioid and benzodiazepine use is common and increasing.^1-5 Outpatient use of these medications has been associated with hospital readmission and death,^6-12 with concurrent use associated with particularly increased risk.^13,14 Less is known about outcomes for hospitalized patients receiving these medications.

More than half of hospital inpatients in the United States receive opioids,¹⁵ many of which are new prescriptions rather than continuation of chronic therapy.^16,17 Less is known about inpatient benzodiazepine administration, but the prevalence may exceed 10% among elderly populations.¹⁸ Hospitalized patients often have comorbidities or physiological disturbances that might increase their risk related to use of these medications. Opioids can cause central and obstructive sleep apneas,^19-21 and benzodiazepines contribute to respiratory depression and airway relaxation.²² Benzodiazepines also impair psychomotor function and recall,²³ which could mediate the recognized risk for delirium and falls in the hospital.^24,25 These findings suggest pathways by which these medications might contribute to clinical deterioration.

Most studies in hospitalized patients have been limited to specific populations^15,26-28 and have not explicitly controlled for severity of illness over time. It remains unclear whether associations identified within particular groups of patients hold true for the broader population of general ward inpatients. Therefore, we aimed to determine the independent association between opioid and benzodiazepine administration and clinical deterioration in ward patients.

Setting and Study Population

We performed an observational cohort study at a 500-bed urban academic hospital. Data were obtained from all adults hospitalized on the wards between November 1, 2008, and January 21, 2016. The study protocol was approved by the University of Chicago Institutional Review Board (IRB#15-0195).

Data Collection

The study utilized de-identified data from the electronic health record (EHR; Epic Systems Corporation, Verona, Wisconsin) and administrative databases collected by the University of Chicago Clinical Research Data Warehouse. Patient age, sex, race, body mass index (BMI), and ward admission source (ie, emergency department (ED), transferred from the intensive care unit (ICU), or directly admitted to the wards) were collected. International Classification of Diseases, 9^th Revision, Clinical Modification (ICD-9-CM) codes were used to identify Elixhauser Comorbidity Index categories.^29,30 Because patients with similar diagnoses (eg, active cancer) are cohorted within particular areas in our hospital, we obtained the ward unit for all patients. Patients who underwent surgery were identified using the hospital’s admission-transfer-discharge database.

To determine severity of illness, routinely collected vital signs and laboratory values were utilized to calculate the electronic cardiac arrest risk triage (eCART) score, an accurate risk score we previously developed and validated for predicting adverse events among ward patients.³¹ If any vital sign or laboratory value was missing, the next available measurement was carried forward. If any value remained missing after this change, the median value for that location (ie, wards, ICU, or ED) was imputed.^32,33 Additionally, patient-reported pain scores at the time of opioid administration were extracted from nursing flowsheets. If no pain score was present at the time of opioid administration, the patient’s previous score was carried forward.

We excluded patients with sickle-cell disease or seizure history and admissions with diagnoses of alcohol withdrawal from the analysis, because these diagnoses were expected to be associated with different medication administration practices compared to other inpatients. We also excluded patients with a tracheostomy because we expected their respiratory monitoring to differ from the other patients in our cohort. Finally, because ward deaths resulting from a comfort care scenario often involve opioids and/or benzodiazepines, ward segments involving comfort care deaths (defined as death without attempted resuscitation) were excluded from the analysis (Supplemental Figure 1). Patients with sickle-cell disease were identified using ICD-9 codes, and encounters during which a seizure may have occurred were identified using a combination of ICD-9 codes and receipt of anti-epileptic medication (Supplemental Table 1). Patients at risk for alcohol withdrawal were identified by the presence of any Clinical Institute Withdrawal Assessment for Alcohol score within nursing flowsheets, and patients with tracheostomies were identified using documentation of ventilator support within their first 12 hours on the wards. In addition to these exclusion criteria, patients with obstructive sleep apnea (OSA) were identified by the following ICD-9 codes: 278.03, 327.23, 780.51, 780.53, and 780.57.

Medications

Ward administrations of opioids and benzodiazepines—dose, route, and administration time—were collected from the EHR. We excluded all administrations in nonward locations such as the ED, ICU, operating room, or procedure suite. Additionally, because patients emergently intubated may receive sedative and analgesic medications to facilitate intubation, and because patients experiencing cardiac arrest are frequently intubated periresuscitation, we a priori excluded all administrations within 15 minutes of a ward cardiac arrest or an intubation.

For consistent comparisons, opioid doses were converted to oral morphine equivalents³⁴ and adjusted by a factor of 15 to reflect the smallest routinely available oral morphine tablet in our hospital (Supplemental Table 2). Benzodiazepine doses were converted to oral lorazepam equivalents (Supplemental Table 2).³⁴ Thus, the independent variables were oral morphine or lorazepam equivalents administered within each 6-hour window. We a priori presumed opioid doses greater than the 99^th percentile (1200 mg) or benzodiazepine doses greater than 10 mg oral lorazepam equivalents within a 6-hour window to be erroneous entries, and replaced these outlier values with the median value for each medication category.

Outcomes

The primary outcome was the composite of ICU transfer or cardiac arrest (loss of pulse with attempted resuscitation) on the wards, with individual outcomes investigated secondarily. An ICU transfer (patient movement from a ward directly to the ICU) was identified using the hospital’s admission-transfer-discharge database. Cardiac arrests were identified using a prospectively validated quality improvement database.³⁵

Because deaths on the wards resulted either from cardiac arrest or from a comfort care scenario, mortality was not studied as an outcome.

Statistical Analysis

Patient characteristics were compared using Student t tests, Wilcoxon rank sum tests, and chi-squared statistics, as appropriate. Unadjusted and adjusted models were created using discrete-time survival analysis,^36-39 which involved dividing time into discrete 6-hour intervals and employing the predictor variables chronologically closest to the beginning of each time window to forecast whether the outcome occurred within each interval. Predictor variables in the adjusted model included patient characteristics (age, sex, BMI, and Elixhauser Agency for Healthcare Research and Quality-Web comorbidities³⁰ [a priori excluding comorbidities recorded for fewer than 1000 admissions from the model]), ward unit, surgical status, prior ICU admission during the hospitalization, cumulative opioid or benzodiazepine dose during the previous 24 hours, and severity of illness (measured by eCART score). The adjusted model for opioids also included the patient’s pain score. Age, eCART score, and pain score were entered linearly while race, BMI (underweight, less than 18.5 kg/m²; normal, 18.5-24.9 kg/m²; overweight, 25.0-29.9 kg/m²; obese, 30-39.9 kg/m²; and severely obese, 40 mg/m² or greater), and ward unit were modeled as categorical variables.

Since repeat hospitalization could confound the results of our study, we performed a sensitivity analysis including only 1 randomly selected hospital admission per patient. We also performed a sensitivity analysis including receipt of both opioids and benzodiazepines, and an interaction term within each ward segment, as well as an analysis in which zolpidem—the most commonly administered nonbenzodiazepine hypnotic medication in our hospital—was included along with both opioids and benzodiazepines. Finally, we performed a sensitivity analysis replacing missing pain scores with imputed values ranging from 0 to the median ward pain score.

We also performed subgroup analyses of adjusted models across age quartiles and for each BMI category, as well as for surgical status, OSA status, gender, time of medication administration, and route of administration (intravenous vs. oral). We also performed an analysis across pain score severity⁴⁰ to determine whether these medications produce differential effects at various levels of pain.

All tests of significance used a 2-sided P value less than 0.05. Statistical analyses were completed using Stata version 14.1 (StataCorp, LLC, College Station, Texas).

Patient Characteristics

A total of 144,895 admissions, from 75,369 patients, had ward vital signs or laboratory values documented during the study period. Ward segments from 634 admissions were excluded due to comfort care status, which resulted in exclusion of 479 complete patient admissions. Additionally, 139 patients with tracheostomies were excluded. Furthermore, 2934 patient admissions with a sickle-cell diagnosis were excluded, of which 95% (n = 2791) received an opioid and 11% (n = 310) received a benzodiazepine. Another 14,029 admissions associated with seizures, 6134 admissions involving alcohol withdrawal, and 1332 with both were excluded, of which 66% (n = 14,174) received an opioid and 35% (n = 7504) received a benzodiazepine. After exclusions, 120,518 admissions were included in the final analysis, with 67% (n = 80,463) associated with at least 1 administration of an opioid and 21% (n = 25,279) associated with at least 1 benzodiazepine administration.

In total, there were 672,851 intervals when an opioid was administered during the study, with a median dose of 12 mg oral morphine equivalents (interquartile range, 8-30). Of these, 21,634 doses were replaced due to outlier status outside the 99^th percentile. Patients receiving opioids were younger (median age 56 vs 61 years), less likely to be African American (48% vs 59%), more likely to have undergone surgery (18% vs 6%), and less likely to have most noncancer medical comorbidities than those who never received an opioid (all P < 0.001) (Table 1).

Patient Outcomes

The incidence of the composite outcome was higher in admissions with at least 1 opioid medication than those without an opioid (7% vs. 4%, P < 0.001), and in admissions with at least 1 dose of benzodiazepines compared to those without a benzodiazepine (11% vs. 4%, P < 0.001) (Table 2).

Within 6-hour segments, increasing doses of opioids were associated with an initial decrease in the frequency of the composite outcome followed by a dose-related increase in the frequency of the composite outcome with morphine equivalents greater than 45 mg. By contrast, the frequency of the composite outcome increased with additional benzodiazepine equivalents (Figure).

In the adjusted model, opioid administration was associated with increased risk for the composite outcome (Table 3) in a dose-dependent fashion, with each 15 mg oral morphine equivalent associated with a 1.9% increase in the odds of ICU transfer or cardiac arrest within the subsequent 6-hour time interval (odds ratio [OR], 1.019; 95% confidence interval [CI], 1.013-1.026; P < 0.001).

Sensitivity Analyses

A sensitivity analysis including 1 randomly selected hospitalization per patient involved 67,097 admissions and found results similar to the primary analysis, with each 15 mg oral morphine equivalent associated with a 1.9% increase in the odds of the composite outcome (OR, 1.019; 95% CI, 1.011-1.028; P < 0.001) and each 1 mg oral lorazepam equivalent associated with a 41% increase in the odds of the composite outcome (OR, 1.41; 95% CI, 1.21-1.65; P < 0.001). Inclusion of both opioids and benzodiazepines in the adjusted model again yielded results similar to the main analysis for both opioids (OR, 1.020; 95% CI, 1.013-1.026; P < 0.001) and benzodiazepines (OR, 1.35; 95% CI, 1.18-1.54; P < 0.001), without a significant interaction detected (P = 0.09). These results were unchanged with the addition of zolpidem to the model as an additional potential confounder, and zolpidem did not increase the risk of the study outcomes (P = 0.2).

A final sensitivity analysis for the opioid model involved replacing missing pain scores with imputed values ranging from 0 to the median ward score, which was 5. The results of these analyses did not differ from the primary model and were consistent regardless of imputation value (OR, 1.018; 95% CI, 1.012-1.023; P < 0.001).

Subgroup Analyses

Analyses of opioid administration by subgroup (sex, age quartiles, BMI categories, OSA diagnosis, surgical status, daytime/nighttime medication administration, IV/PO administration, and pain severity) yielded similar results to the overall analysis (Supplemental Figure 2). Subgroup analysis of patients receiving benzodiazepines revealed similarly increased adjusted odds of the composite outcome across strata of gender, BMI, surgical status, and medication administration time (Supplemental Figure 3). Notably, patients older than 70 years who received a benzodiazepine were at 64% increased odds of the composite outcome (OR, 1.64; 95% CI, 1.30-2.08), compared to 2% to 38% increased risk for patients under 70 years. Finally, IV doses of benzodiazepines were associated with 48% increased odds for deterioration (OR, 1.48; 95% CI, 1.18-1.84; P = 0.001), compared to a nonsignificant 14% increase in the odds for PO doses (OR, 1.14; 95% CI, 0.99-1.31; P = 0.066).

In a large, single-center, observational study of ward inpatients, we found that opioid use was associated with a small but significant increased risk for clinical deterioration on the wards, with every 15 mg oral morphine equivalent increasing the odds of ICU transfer or cardiac arrest in the next 6 hours by 1.9%. Benzodiazepines were associated with a much higher risk: each equivalent of 1 mg of oral lorazepam increased the odds of ICU transfer or cardiac arrest by almost 30%. These results have important implications for care at the bedside of hospitalized ward patients and suggest the need for closer monitoring after receipt of these medications, particularly benzodiazepines.

Previous work has described negative effects of opioid medications among select inpatient populations. In surgical patients, opioids have been associated with hospital readmission, increased length of stay, and hospital mortality.^26,28 More recently, Herzig et al.¹⁵ found more adverse events in nonsurgical ward patients within the hospitals prescribing opioids the most frequently. These studies may have been limited by the populations studied and the inability to control for confounders such as severity of illness and pain score. Our study expands these findings to a more generalizable population and shows that even after adjustment for potential confounders, such as severity of illness, pain score, and medication dose, opioids are associated with increased short-term risk of clinical deterioration.

By contrast, few studies have characterized the risks associated with benzodiazepine use among ward inpatients. Recently, Overdyk et al.²⁷ found that inpatient use of opioids and sedatives was associated with increased risk for cardiac arrest and hospital death. However, this study included ICU patients, which may confound the results, as ICU patients often receive high doses of opioids or benzodiazepines to facilitate mechanical ventilation or other invasive procedures, while also having a particularly high risk of adverse outcomes like cardiac arrest and inhospital death.

Several mechanisms may explain the magnitude of effect seen with regard to benzodiazepines. First, benzodiazepines may directly produce clinical deterioration by decreased respiratory drive, diminished airway tone, or hemodynamic decompensation. It is possible that the broad spectrum of cardiorespiratory side effects of benzodiazepines—and potential unpredictability of these effects—increases the difficulty of observation and management for patients receiving them. This difficulty may be compounded with intravenous administration of benzodiazepines, which was associated with a higher risk for deterioration than oral doses in our cohort. Alternatively, benzodiazepines may contribute to clinical decompensation by masking signs of deterioration such as encephalopathy or vital sign instability like tachycardia or tachypnea that may be mistaken as anxiety. Notably, while our hospital has a nursing-driven protocol for monitoring patients receiving opioids (in which pain is serially assessed, leading to additional bedside observation), we do not have protocols for ward patients receiving benzodiazepines. Finally, although we found that orders for opioids and benzodiazepines were more common in white patients than African American patients, this finding may be due to differences in the types or number of medical comorbidities experienced by these patients.

Our study has several strengths, including the large number of admissions we included. Additionally, we included a broad range of medical and surgical ward admissions, which should increase the generalizability of our results. Further, our rates of ICU transfer are in line with data reported from other groups,^41,42 which again may add to the generalizability of our findings. We also addressed many potential confounders by including patient characteristics, individual ward units, and (for opioids) pain score in our model, and by controlling for severity of illness with the eCART score, an accurate predictor of ICU transfer and ward cardiac arrest within our population.^32,37 Finally, our robust methodology allowed us to include acute and cumulative medication doses, as well as time, in the model. By performing a discrete-time survival analysis, we were able to evaluate receipt of opioids and benzodiazepines—as well as risk for clinical deterioration—longitudinally, lending strength to our results.

Limitations of our study include its single-center cohort, which may reduce generalizability to other populations. Additionally, because we could not validate the accuracy of—or adherence to—outpatient medication lists, we were unable to identify chronic opioid or benzodiazepine users by these lists. However, patients chronically taking opioids or benzodiazepines would likely receive doses each hospital day; by including 24-hour cumulative doses in our model, we attempted to adjust for some portion of their chronic use. Also, because evaluation of delirium was not objectively recorded in our dataset, we were unable to evaluate the relationship between receipt of these medications and development of delirium, which is an important outcome for hospitalized patients. Finally, neither the diagnoses for which these medications were prescribed, nor the reason for ICU transfer, were present in our dataset, which leaves open the possibility of unmeasured confounding.

After adjustment for important confounders including severity of illness, medication dose, and time, opioids were associated with a slight increase in clinical deterioration on the wards, while benzodiazepines were associated with a much larger risk for deterioration. This finding raises concern about the safety of benzodiazepine use among ward patients and suggests that increased monitoring of patients receiving these medications may be warranted.

Acknowledgment

The authors thank Nicole Twu for administrative support.

Disclosure

Drs. Churpek and Edelson have a patent pending (ARCD. P0535US.P2) for risk stratification algorithms for hospitalized patients. Dr. Churpek is supported by a career development award from the National Heart, Lung, and Blood Institute (K08 HL121080). Dr. Churpek has received honoraria from Chest for invited speaking engagements. In addition, Dr. Edelson has received research support from Philips Healthcare (Andover, Massachusetts), research support from the American Heart Association (Dallas, Texas) and Laerdal Medical (Stavanger, Norway), and research support from Early Sense (Tel Aviv, Israel). She has ownership interest in Quant HC (Chicago, Illinois), which is developing products for risk stratification of hospitalized patients. Dr. Mokhlesi is supported by National Institutes of Health grant R01HL119161. Dr. Mokhlesi has served as a consultant to Philips/Respironics and has received research support from Philips/Respironics. Preliminary versions of these data were presented as a poster presentation at the 2016 meeting of the American Thoracic Society, May 17, 2016; San Francisco, California.

Chronic opioid and benzodiazepine use is common and increasing.^1-5 Outpatient use of these medications has been associated with hospital readmission and death,^6-12 with concurrent use associated with particularly increased risk.^13,14 Less is known about outcomes for hospitalized patients receiving these medications.

More than half of hospital inpatients in the United States receive opioids,¹⁵ many of which are new prescriptions rather than continuation of chronic therapy.^16,17 Less is known about inpatient benzodiazepine administration, but the prevalence may exceed 10% among elderly populations.¹⁸ Hospitalized patients often have comorbidities or physiological disturbances that might increase their risk related to use of these medications. Opioids can cause central and obstructive sleep apneas,^19-21 and benzodiazepines contribute to respiratory depression and airway relaxation.²² Benzodiazepines also impair psychomotor function and recall,²³ which could mediate the recognized risk for delirium and falls in the hospital.^24,25 These findings suggest pathways by which these medications might contribute to clinical deterioration.

Most studies in hospitalized patients have been limited to specific populations^15,26-28 and have not explicitly controlled for severity of illness over time. It remains unclear whether associations identified within particular groups of patients hold true for the broader population of general ward inpatients. Therefore, we aimed to determine the independent association between opioid and benzodiazepine administration and clinical deterioration in ward patients.

Setting and Study Population

We performed an observational cohort study at a 500-bed urban academic hospital. Data were obtained from all adults hospitalized on the wards between November 1, 2008, and January 21, 2016. The study protocol was approved by the University of Chicago Institutional Review Board (IRB#15-0195).

Data Collection

The study utilized de-identified data from the electronic health record (EHR; Epic Systems Corporation, Verona, Wisconsin) and administrative databases collected by the University of Chicago Clinical Research Data Warehouse. Patient age, sex, race, body mass index (BMI), and ward admission source (ie, emergency department (ED), transferred from the intensive care unit (ICU), or directly admitted to the wards) were collected. International Classification of Diseases, 9^th Revision, Clinical Modification (ICD-9-CM) codes were used to identify Elixhauser Comorbidity Index categories.^29,30 Because patients with similar diagnoses (eg, active cancer) are cohorted within particular areas in our hospital, we obtained the ward unit for all patients. Patients who underwent surgery were identified using the hospital’s admission-transfer-discharge database.

To determine severity of illness, routinely collected vital signs and laboratory values were utilized to calculate the electronic cardiac arrest risk triage (eCART) score, an accurate risk score we previously developed and validated for predicting adverse events among ward patients.³¹ If any vital sign or laboratory value was missing, the next available measurement was carried forward. If any value remained missing after this change, the median value for that location (ie, wards, ICU, or ED) was imputed.^32,33 Additionally, patient-reported pain scores at the time of opioid administration were extracted from nursing flowsheets. If no pain score was present at the time of opioid administration, the patient’s previous score was carried forward.

We excluded patients with sickle-cell disease or seizure history and admissions with diagnoses of alcohol withdrawal from the analysis, because these diagnoses were expected to be associated with different medication administration practices compared to other inpatients. We also excluded patients with a tracheostomy because we expected their respiratory monitoring to differ from the other patients in our cohort. Finally, because ward deaths resulting from a comfort care scenario often involve opioids and/or benzodiazepines, ward segments involving comfort care deaths (defined as death without attempted resuscitation) were excluded from the analysis (Supplemental Figure 1). Patients with sickle-cell disease were identified using ICD-9 codes, and encounters during which a seizure may have occurred were identified using a combination of ICD-9 codes and receipt of anti-epileptic medication (Supplemental Table 1). Patients at risk for alcohol withdrawal were identified by the presence of any Clinical Institute Withdrawal Assessment for Alcohol score within nursing flowsheets, and patients with tracheostomies were identified using documentation of ventilator support within their first 12 hours on the wards. In addition to these exclusion criteria, patients with obstructive sleep apnea (OSA) were identified by the following ICD-9 codes: 278.03, 327.23, 780.51, 780.53, and 780.57.

Medications

Ward administrations of opioids and benzodiazepines—dose, route, and administration time—were collected from the EHR. We excluded all administrations in nonward locations such as the ED, ICU, operating room, or procedure suite. Additionally, because patients emergently intubated may receive sedative and analgesic medications to facilitate intubation, and because patients experiencing cardiac arrest are frequently intubated periresuscitation, we a priori excluded all administrations within 15 minutes of a ward cardiac arrest or an intubation.

For consistent comparisons, opioid doses were converted to oral morphine equivalents³⁴ and adjusted by a factor of 15 to reflect the smallest routinely available oral morphine tablet in our hospital (Supplemental Table 2). Benzodiazepine doses were converted to oral lorazepam equivalents (Supplemental Table 2).³⁴ Thus, the independent variables were oral morphine or lorazepam equivalents administered within each 6-hour window. We a priori presumed opioid doses greater than the 99^th percentile (1200 mg) or benzodiazepine doses greater than 10 mg oral lorazepam equivalents within a 6-hour window to be erroneous entries, and replaced these outlier values with the median value for each medication category.

Outcomes

The primary outcome was the composite of ICU transfer or cardiac arrest (loss of pulse with attempted resuscitation) on the wards, with individual outcomes investigated secondarily. An ICU transfer (patient movement from a ward directly to the ICU) was identified using the hospital’s admission-transfer-discharge database. Cardiac arrests were identified using a prospectively validated quality improvement database.³⁵

Because deaths on the wards resulted either from cardiac arrest or from a comfort care scenario, mortality was not studied as an outcome.

Statistical Analysis

Patient characteristics were compared using Student t tests, Wilcoxon rank sum tests, and chi-squared statistics, as appropriate. Unadjusted and adjusted models were created using discrete-time survival analysis,^36-39 which involved dividing time into discrete 6-hour intervals and employing the predictor variables chronologically closest to the beginning of each time window to forecast whether the outcome occurred within each interval. Predictor variables in the adjusted model included patient characteristics (age, sex, BMI, and Elixhauser Agency for Healthcare Research and Quality-Web comorbidities³⁰ [a priori excluding comorbidities recorded for fewer than 1000 admissions from the model]), ward unit, surgical status, prior ICU admission during the hospitalization, cumulative opioid or benzodiazepine dose during the previous 24 hours, and severity of illness (measured by eCART score). The adjusted model for opioids also included the patient’s pain score. Age, eCART score, and pain score were entered linearly while race, BMI (underweight, less than 18.5 kg/m²; normal, 18.5-24.9 kg/m²; overweight, 25.0-29.9 kg/m²; obese, 30-39.9 kg/m²; and severely obese, 40 mg/m² or greater), and ward unit were modeled as categorical variables.

Since repeat hospitalization could confound the results of our study, we performed a sensitivity analysis including only 1 randomly selected hospital admission per patient. We also performed a sensitivity analysis including receipt of both opioids and benzodiazepines, and an interaction term within each ward segment, as well as an analysis in which zolpidem—the most commonly administered nonbenzodiazepine hypnotic medication in our hospital—was included along with both opioids and benzodiazepines. Finally, we performed a sensitivity analysis replacing missing pain scores with imputed values ranging from 0 to the median ward pain score.

We also performed subgroup analyses of adjusted models across age quartiles and for each BMI category, as well as for surgical status, OSA status, gender, time of medication administration, and route of administration (intravenous vs. oral). We also performed an analysis across pain score severity⁴⁰ to determine whether these medications produce differential effects at various levels of pain.

All tests of significance used a 2-sided P value less than 0.05. Statistical analyses were completed using Stata version 14.1 (StataCorp, LLC, College Station, Texas).

Patient Characteristics

A total of 144,895 admissions, from 75,369 patients, had ward vital signs or laboratory values documented during the study period. Ward segments from 634 admissions were excluded due to comfort care status, which resulted in exclusion of 479 complete patient admissions. Additionally, 139 patients with tracheostomies were excluded. Furthermore, 2934 patient admissions with a sickle-cell diagnosis were excluded, of which 95% (n = 2791) received an opioid and 11% (n = 310) received a benzodiazepine. Another 14,029 admissions associated with seizures, 6134 admissions involving alcohol withdrawal, and 1332 with both were excluded, of which 66% (n = 14,174) received an opioid and 35% (n = 7504) received a benzodiazepine. After exclusions, 120,518 admissions were included in the final analysis, with 67% (n = 80,463) associated with at least 1 administration of an opioid and 21% (n = 25,279) associated with at least 1 benzodiazepine administration.

In total, there were 672,851 intervals when an opioid was administered during the study, with a median dose of 12 mg oral morphine equivalents (interquartile range, 8-30). Of these, 21,634 doses were replaced due to outlier status outside the 99^th percentile. Patients receiving opioids were younger (median age 56 vs 61 years), less likely to be African American (48% vs 59%), more likely to have undergone surgery (18% vs 6%), and less likely to have most noncancer medical comorbidities than those who never received an opioid (all P < 0.001) (Table 1).

Patient Outcomes

The incidence of the composite outcome was higher in admissions with at least 1 opioid medication than those without an opioid (7% vs. 4%, P < 0.001), and in admissions with at least 1 dose of benzodiazepines compared to those without a benzodiazepine (11% vs. 4%, P < 0.001) (Table 2).

Within 6-hour segments, increasing doses of opioids were associated with an initial decrease in the frequency of the composite outcome followed by a dose-related increase in the frequency of the composite outcome with morphine equivalents greater than 45 mg. By contrast, the frequency of the composite outcome increased with additional benzodiazepine equivalents (Figure).

In the adjusted model, opioid administration was associated with increased risk for the composite outcome (Table 3) in a dose-dependent fashion, with each 15 mg oral morphine equivalent associated with a 1.9% increase in the odds of ICU transfer or cardiac arrest within the subsequent 6-hour time interval (odds ratio [OR], 1.019; 95% confidence interval [CI], 1.013-1.026; P < 0.001).

Sensitivity Analyses

A sensitivity analysis including 1 randomly selected hospitalization per patient involved 67,097 admissions and found results similar to the primary analysis, with each 15 mg oral morphine equivalent associated with a 1.9% increase in the odds of the composite outcome (OR, 1.019; 95% CI, 1.011-1.028; P < 0.001) and each 1 mg oral lorazepam equivalent associated with a 41% increase in the odds of the composite outcome (OR, 1.41; 95% CI, 1.21-1.65; P < 0.001). Inclusion of both opioids and benzodiazepines in the adjusted model again yielded results similar to the main analysis for both opioids (OR, 1.020; 95% CI, 1.013-1.026; P < 0.001) and benzodiazepines (OR, 1.35; 95% CI, 1.18-1.54; P < 0.001), without a significant interaction detected (P = 0.09). These results were unchanged with the addition of zolpidem to the model as an additional potential confounder, and zolpidem did not increase the risk of the study outcomes (P = 0.2).

A final sensitivity analysis for the opioid model involved replacing missing pain scores with imputed values ranging from 0 to the median ward score, which was 5. The results of these analyses did not differ from the primary model and were consistent regardless of imputation value (OR, 1.018; 95% CI, 1.012-1.023; P < 0.001).

Subgroup Analyses

Analyses of opioid administration by subgroup (sex, age quartiles, BMI categories, OSA diagnosis, surgical status, daytime/nighttime medication administration, IV/PO administration, and pain severity) yielded similar results to the overall analysis (Supplemental Figure 2). Subgroup analysis of patients receiving benzodiazepines revealed similarly increased adjusted odds of the composite outcome across strata of gender, BMI, surgical status, and medication administration time (Supplemental Figure 3). Notably, patients older than 70 years who received a benzodiazepine were at 64% increased odds of the composite outcome (OR, 1.64; 95% CI, 1.30-2.08), compared to 2% to 38% increased risk for patients under 70 years. Finally, IV doses of benzodiazepines were associated with 48% increased odds for deterioration (OR, 1.48; 95% CI, 1.18-1.84; P = 0.001), compared to a nonsignificant 14% increase in the odds for PO doses (OR, 1.14; 95% CI, 0.99-1.31; P = 0.066).

In a large, single-center, observational study of ward inpatients, we found that opioid use was associated with a small but significant increased risk for clinical deterioration on the wards, with every 15 mg oral morphine equivalent increasing the odds of ICU transfer or cardiac arrest in the next 6 hours by 1.9%. Benzodiazepines were associated with a much higher risk: each equivalent of 1 mg of oral lorazepam increased the odds of ICU transfer or cardiac arrest by almost 30%. These results have important implications for care at the bedside of hospitalized ward patients and suggest the need for closer monitoring after receipt of these medications, particularly benzodiazepines.

Previous work has described negative effects of opioid medications among select inpatient populations. In surgical patients, opioids have been associated with hospital readmission, increased length of stay, and hospital mortality.^26,28 More recently, Herzig et al.¹⁵ found more adverse events in nonsurgical ward patients within the hospitals prescribing opioids the most frequently. These studies may have been limited by the populations studied and the inability to control for confounders such as severity of illness and pain score. Our study expands these findings to a more generalizable population and shows that even after adjustment for potential confounders, such as severity of illness, pain score, and medication dose, opioids are associated with increased short-term risk of clinical deterioration.

By contrast, few studies have characterized the risks associated with benzodiazepine use among ward inpatients. Recently, Overdyk et al.²⁷ found that inpatient use of opioids and sedatives was associated with increased risk for cardiac arrest and hospital death. However, this study included ICU patients, which may confound the results, as ICU patients often receive high doses of opioids or benzodiazepines to facilitate mechanical ventilation or other invasive procedures, while also having a particularly high risk of adverse outcomes like cardiac arrest and inhospital death.

Several mechanisms may explain the magnitude of effect seen with regard to benzodiazepines. First, benzodiazepines may directly produce clinical deterioration by decreased respiratory drive, diminished airway tone, or hemodynamic decompensation. It is possible that the broad spectrum of cardiorespiratory side effects of benzodiazepines—and potential unpredictability of these effects—increases the difficulty of observation and management for patients receiving them. This difficulty may be compounded with intravenous administration of benzodiazepines, which was associated with a higher risk for deterioration than oral doses in our cohort. Alternatively, benzodiazepines may contribute to clinical decompensation by masking signs of deterioration such as encephalopathy or vital sign instability like tachycardia or tachypnea that may be mistaken as anxiety. Notably, while our hospital has a nursing-driven protocol for monitoring patients receiving opioids (in which pain is serially assessed, leading to additional bedside observation), we do not have protocols for ward patients receiving benzodiazepines. Finally, although we found that orders for opioids and benzodiazepines were more common in white patients than African American patients, this finding may be due to differences in the types or number of medical comorbidities experienced by these patients.

Our study has several strengths, including the large number of admissions we included. Additionally, we included a broad range of medical and surgical ward admissions, which should increase the generalizability of our results. Further, our rates of ICU transfer are in line with data reported from other groups,^41,42 which again may add to the generalizability of our findings. We also addressed many potential confounders by including patient characteristics, individual ward units, and (for opioids) pain score in our model, and by controlling for severity of illness with the eCART score, an accurate predictor of ICU transfer and ward cardiac arrest within our population.^32,37 Finally, our robust methodology allowed us to include acute and cumulative medication doses, as well as time, in the model. By performing a discrete-time survival analysis, we were able to evaluate receipt of opioids and benzodiazepines—as well as risk for clinical deterioration—longitudinally, lending strength to our results.

Limitations of our study include its single-center cohort, which may reduce generalizability to other populations. Additionally, because we could not validate the accuracy of—or adherence to—outpatient medication lists, we were unable to identify chronic opioid or benzodiazepine users by these lists. However, patients chronically taking opioids or benzodiazepines would likely receive doses each hospital day; by including 24-hour cumulative doses in our model, we attempted to adjust for some portion of their chronic use. Also, because evaluation of delirium was not objectively recorded in our dataset, we were unable to evaluate the relationship between receipt of these medications and development of delirium, which is an important outcome for hospitalized patients. Finally, neither the diagnoses for which these medications were prescribed, nor the reason for ICU transfer, were present in our dataset, which leaves open the possibility of unmeasured confounding.

After adjustment for important confounders including severity of illness, medication dose, and time, opioids were associated with a slight increase in clinical deterioration on the wards, while benzodiazepines were associated with a much larger risk for deterioration. This finding raises concern about the safety of benzodiazepine use among ward patients and suggests that increased monitoring of patients receiving these medications may be warranted.

Acknowledgment

The authors thank Nicole Twu for administrative support.

Disclosure

Drs. Churpek and Edelson have a patent pending (ARCD. P0535US.P2) for risk stratification algorithms for hospitalized patients. Dr. Churpek is supported by a career development award from the National Heart, Lung, and Blood Institute (K08 HL121080). Dr. Churpek has received honoraria from Chest for invited speaking engagements. In addition, Dr. Edelson has received research support from Philips Healthcare (Andover, Massachusetts), research support from the American Heart Association (Dallas, Texas) and Laerdal Medical (Stavanger, Norway), and research support from Early Sense (Tel Aviv, Israel). She has ownership interest in Quant HC (Chicago, Illinois), which is developing products for risk stratification of hospitalized patients. Dr. Mokhlesi is supported by National Institutes of Health grant R01HL119161. Dr. Mokhlesi has served as a consultant to Philips/Respironics and has received research support from Philips/Respironics. Preliminary versions of these data were presented as a poster presentation at the 2016 meeting of the American Thoracic Society, May 17, 2016; San Francisco, California.

Chronic opioid and benzodiazepine use is common and increasing.^1-5 Outpatient use of these medications has been associated with hospital readmission and death,^6-12 with concurrent use associated with particularly increased risk.^13,14 Less is known about outcomes for hospitalized patients receiving these medications.

More than half of hospital inpatients in the United States receive opioids,¹⁵ many of which are new prescriptions rather than continuation of chronic therapy.^16,17 Less is known about inpatient benzodiazepine administration, but the prevalence may exceed 10% among elderly populations.¹⁸ Hospitalized patients often have comorbidities or physiological disturbances that might increase their risk related to use of these medications. Opioids can cause central and obstructive sleep apneas,^19-21 and benzodiazepines contribute to respiratory depression and airway relaxation.²² Benzodiazepines also impair psychomotor function and recall,²³ which could mediate the recognized risk for delirium and falls in the hospital.^24,25 These findings suggest pathways by which these medications might contribute to clinical deterioration.

Most studies in hospitalized patients have been limited to specific populations^15,26-28 and have not explicitly controlled for severity of illness over time. It remains unclear whether associations identified within particular groups of patients hold true for the broader population of general ward inpatients. Therefore, we aimed to determine the independent association between opioid and benzodiazepine administration and clinical deterioration in ward patients.

Setting and Study Population

We performed an observational cohort study at a 500-bed urban academic hospital. Data were obtained from all adults hospitalized on the wards between November 1, 2008, and January 21, 2016. The study protocol was approved by the University of Chicago Institutional Review Board (IRB#15-0195).

Data Collection

The study utilized de-identified data from the electronic health record (EHR; Epic Systems Corporation, Verona, Wisconsin) and administrative databases collected by the University of Chicago Clinical Research Data Warehouse. Patient age, sex, race, body mass index (BMI), and ward admission source (ie, emergency department (ED), transferred from the intensive care unit (ICU), or directly admitted to the wards) were collected. International Classification of Diseases, 9^th Revision, Clinical Modification (ICD-9-CM) codes were used to identify Elixhauser Comorbidity Index categories.^29,30 Because patients with similar diagnoses (eg, active cancer) are cohorted within particular areas in our hospital, we obtained the ward unit for all patients. Patients who underwent surgery were identified using the hospital’s admission-transfer-discharge database.

To determine severity of illness, routinely collected vital signs and laboratory values were utilized to calculate the electronic cardiac arrest risk triage (eCART) score, an accurate risk score we previously developed and validated for predicting adverse events among ward patients.³¹ If any vital sign or laboratory value was missing, the next available measurement was carried forward. If any value remained missing after this change, the median value for that location (ie, wards, ICU, or ED) was imputed.^32,33 Additionally, patient-reported pain scores at the time of opioid administration were extracted from nursing flowsheets. If no pain score was present at the time of opioid administration, the patient’s previous score was carried forward.

We excluded patients with sickle-cell disease or seizure history and admissions with diagnoses of alcohol withdrawal from the analysis, because these diagnoses were expected to be associated with different medication administration practices compared to other inpatients. We also excluded patients with a tracheostomy because we expected their respiratory monitoring to differ from the other patients in our cohort. Finally, because ward deaths resulting from a comfort care scenario often involve opioids and/or benzodiazepines, ward segments involving comfort care deaths (defined as death without attempted resuscitation) were excluded from the analysis (Supplemental Figure 1). Patients with sickle-cell disease were identified using ICD-9 codes, and encounters during which a seizure may have occurred were identified using a combination of ICD-9 codes and receipt of anti-epileptic medication (Supplemental Table 1). Patients at risk for alcohol withdrawal were identified by the presence of any Clinical Institute Withdrawal Assessment for Alcohol score within nursing flowsheets, and patients with tracheostomies were identified using documentation of ventilator support within their first 12 hours on the wards. In addition to these exclusion criteria, patients with obstructive sleep apnea (OSA) were identified by the following ICD-9 codes: 278.03, 327.23, 780.51, 780.53, and 780.57.

Medications

Ward administrations of opioids and benzodiazepines—dose, route, and administration time—were collected from the EHR. We excluded all administrations in nonward locations such as the ED, ICU, operating room, or procedure suite. Additionally, because patients emergently intubated may receive sedative and analgesic medications to facilitate intubation, and because patients experiencing cardiac arrest are frequently intubated periresuscitation, we a priori excluded all administrations within 15 minutes of a ward cardiac arrest or an intubation.

For consistent comparisons, opioid doses were converted to oral morphine equivalents³⁴ and adjusted by a factor of 15 to reflect the smallest routinely available oral morphine tablet in our hospital (Supplemental Table 2). Benzodiazepine doses were converted to oral lorazepam equivalents (Supplemental Table 2).³⁴ Thus, the independent variables were oral morphine or lorazepam equivalents administered within each 6-hour window. We a priori presumed opioid doses greater than the 99^th percentile (1200 mg) or benzodiazepine doses greater than 10 mg oral lorazepam equivalents within a 6-hour window to be erroneous entries, and replaced these outlier values with the median value for each medication category.

Outcomes

The primary outcome was the composite of ICU transfer or cardiac arrest (loss of pulse with attempted resuscitation) on the wards, with individual outcomes investigated secondarily. An ICU transfer (patient movement from a ward directly to the ICU) was identified using the hospital’s admission-transfer-discharge database. Cardiac arrests were identified using a prospectively validated quality improvement database.³⁵

Because deaths on the wards resulted either from cardiac arrest or from a comfort care scenario, mortality was not studied as an outcome.

Statistical Analysis

Patient characteristics were compared using Student t tests, Wilcoxon rank sum tests, and chi-squared statistics, as appropriate. Unadjusted and adjusted models were created using discrete-time survival analysis,^36-39 which involved dividing time into discrete 6-hour intervals and employing the predictor variables chronologically closest to the beginning of each time window to forecast whether the outcome occurred within each interval. Predictor variables in the adjusted model included patient characteristics (age, sex, BMI, and Elixhauser Agency for Healthcare Research and Quality-Web comorbidities³⁰ [a priori excluding comorbidities recorded for fewer than 1000 admissions from the model]), ward unit, surgical status, prior ICU admission during the hospitalization, cumulative opioid or benzodiazepine dose during the previous 24 hours, and severity of illness (measured by eCART score). The adjusted model for opioids also included the patient’s pain score. Age, eCART score, and pain score were entered linearly while race, BMI (underweight, less than 18.5 kg/m²; normal, 18.5-24.9 kg/m²; overweight, 25.0-29.9 kg/m²; obese, 30-39.9 kg/m²; and severely obese, 40 mg/m² or greater), and ward unit were modeled as categorical variables.

Since repeat hospitalization could confound the results of our study, we performed a sensitivity analysis including only 1 randomly selected hospital admission per patient. We also performed a sensitivity analysis including receipt of both opioids and benzodiazepines, and an interaction term within each ward segment, as well as an analysis in which zolpidem—the most commonly administered nonbenzodiazepine hypnotic medication in our hospital—was included along with both opioids and benzodiazepines. Finally, we performed a sensitivity analysis replacing missing pain scores with imputed values ranging from 0 to the median ward pain score.

We also performed subgroup analyses of adjusted models across age quartiles and for each BMI category, as well as for surgical status, OSA status, gender, time of medication administration, and route of administration (intravenous vs. oral). We also performed an analysis across pain score severity⁴⁰ to determine whether these medications produce differential effects at various levels of pain.

All tests of significance used a 2-sided P value less than 0.05. Statistical analyses were completed using Stata version 14.1 (StataCorp, LLC, College Station, Texas).

Patient Characteristics

A total of 144,895 admissions, from 75,369 patients, had ward vital signs or laboratory values documented during the study period. Ward segments from 634 admissions were excluded due to comfort care status, which resulted in exclusion of 479 complete patient admissions. Additionally, 139 patients with tracheostomies were excluded. Furthermore, 2934 patient admissions with a sickle-cell diagnosis were excluded, of which 95% (n = 2791) received an opioid and 11% (n = 310) received a benzodiazepine. Another 14,029 admissions associated with seizures, 6134 admissions involving alcohol withdrawal, and 1332 with both were excluded, of which 66% (n = 14,174) received an opioid and 35% (n = 7504) received a benzodiazepine. After exclusions, 120,518 admissions were included in the final analysis, with 67% (n = 80,463) associated with at least 1 administration of an opioid and 21% (n = 25,279) associated with at least 1 benzodiazepine administration.

In total, there were 672,851 intervals when an opioid was administered during the study, with a median dose of 12 mg oral morphine equivalents (interquartile range, 8-30). Of these, 21,634 doses were replaced due to outlier status outside the 99^th percentile. Patients receiving opioids were younger (median age 56 vs 61 years), less likely to be African American (48% vs 59%), more likely to have undergone surgery (18% vs 6%), and less likely to have most noncancer medical comorbidities than those who never received an opioid (all P < 0.001) (Table 1).

Patient Outcomes

The incidence of the composite outcome was higher in admissions with at least 1 opioid medication than those without an opioid (7% vs. 4%, P < 0.001), and in admissions with at least 1 dose of benzodiazepines compared to those without a benzodiazepine (11% vs. 4%, P < 0.001) (Table 2).

Within 6-hour segments, increasing doses of opioids were associated with an initial decrease in the frequency of the composite outcome followed by a dose-related increase in the frequency of the composite outcome with morphine equivalents greater than 45 mg. By contrast, the frequency of the composite outcome increased with additional benzodiazepine equivalents (Figure).

In the adjusted model, opioid administration was associated with increased risk for the composite outcome (Table 3) in a dose-dependent fashion, with each 15 mg oral morphine equivalent associated with a 1.9% increase in the odds of ICU transfer or cardiac arrest within the subsequent 6-hour time interval (odds ratio [OR], 1.019; 95% confidence interval [CI], 1.013-1.026; P < 0.001).

Sensitivity Analyses

A sensitivity analysis including 1 randomly selected hospitalization per patient involved 67,097 admissions and found results similar to the primary analysis, with each 15 mg oral morphine equivalent associated with a 1.9% increase in the odds of the composite outcome (OR, 1.019; 95% CI, 1.011-1.028; P < 0.001) and each 1 mg oral lorazepam equivalent associated with a 41% increase in the odds of the composite outcome (OR, 1.41; 95% CI, 1.21-1.65; P < 0.001). Inclusion of both opioids and benzodiazepines in the adjusted model again yielded results similar to the main analysis for both opioids (OR, 1.020; 95% CI, 1.013-1.026; P < 0.001) and benzodiazepines (OR, 1.35; 95% CI, 1.18-1.54; P < 0.001), without a significant interaction detected (P = 0.09). These results were unchanged with the addition of zolpidem to the model as an additional potential confounder, and zolpidem did not increase the risk of the study outcomes (P = 0.2).

A final sensitivity analysis for the opioid model involved replacing missing pain scores with imputed values ranging from 0 to the median ward score, which was 5. The results of these analyses did not differ from the primary model and were consistent regardless of imputation value (OR, 1.018; 95% CI, 1.012-1.023; P < 0.001).

Subgroup Analyses

Analyses of opioid administration by subgroup (sex, age quartiles, BMI categories, OSA diagnosis, surgical status, daytime/nighttime medication administration, IV/PO administration, and pain severity) yielded similar results to the overall analysis (Supplemental Figure 2). Subgroup analysis of patients receiving benzodiazepines revealed similarly increased adjusted odds of the composite outcome across strata of gender, BMI, surgical status, and medication administration time (Supplemental Figure 3). Notably, patients older than 70 years who received a benzodiazepine were at 64% increased odds of the composite outcome (OR, 1.64; 95% CI, 1.30-2.08), compared to 2% to 38% increased risk for patients under 70 years. Finally, IV doses of benzodiazepines were associated with 48% increased odds for deterioration (OR, 1.48; 95% CI, 1.18-1.84; P = 0.001), compared to a nonsignificant 14% increase in the odds for PO doses (OR, 1.14; 95% CI, 0.99-1.31; P = 0.066).

In a large, single-center, observational study of ward inpatients, we found that opioid use was associated with a small but significant increased risk for clinical deterioration on the wards, with every 15 mg oral morphine equivalent increasing the odds of ICU transfer or cardiac arrest in the next 6 hours by 1.9%. Benzodiazepines were associated with a much higher risk: each equivalent of 1 mg of oral lorazepam increased the odds of ICU transfer or cardiac arrest by almost 30%. These results have important implications for care at the bedside of hospitalized ward patients and suggest the need for closer monitoring after receipt of these medications, particularly benzodiazepines.

Previous work has described negative effects of opioid medications among select inpatient populations. In surgical patients, opioids have been associated with hospital readmission, increased length of stay, and hospital mortality.^26,28 More recently, Herzig et al.¹⁵ found more adverse events in nonsurgical ward patients within the hospitals prescribing opioids the most frequently. These studies may have been limited by the populations studied and the inability to control for confounders such as severity of illness and pain score. Our study expands these findings to a more generalizable population and shows that even after adjustment for potential confounders, such as severity of illness, pain score, and medication dose, opioids are associated with increased short-term risk of clinical deterioration.

By contrast, few studies have characterized the risks associated with benzodiazepine use among ward inpatients. Recently, Overdyk et al.²⁷ found that inpatient use of opioids and sedatives was associated with increased risk for cardiac arrest and hospital death. However, this study included ICU patients, which may confound the results, as ICU patients often receive high doses of opioids or benzodiazepines to facilitate mechanical ventilation or other invasive procedures, while also having a particularly high risk of adverse outcomes like cardiac arrest and inhospital death.

Several mechanisms may explain the magnitude of effect seen with regard to benzodiazepines. First, benzodiazepines may directly produce clinical deterioration by decreased respiratory drive, diminished airway tone, or hemodynamic decompensation. It is possible that the broad spectrum of cardiorespiratory side effects of benzodiazepines—and potential unpredictability of these effects—increases the difficulty of observation and management for patients receiving them. This difficulty may be compounded with intravenous administration of benzodiazepines, which was associated with a higher risk for deterioration than oral doses in our cohort. Alternatively, benzodiazepines may contribute to clinical decompensation by masking signs of deterioration such as encephalopathy or vital sign instability like tachycardia or tachypnea that may be mistaken as anxiety. Notably, while our hospital has a nursing-driven protocol for monitoring patients receiving opioids (in which pain is serially assessed, leading to additional bedside observation), we do not have protocols for ward patients receiving benzodiazepines. Finally, although we found that orders for opioids and benzodiazepines were more common in white patients than African American patients, this finding may be due to differences in the types or number of medical comorbidities experienced by these patients.

Our study has several strengths, including the large number of admissions we included. Additionally, we included a broad range of medical and surgical ward admissions, which should increase the generalizability of our results. Further, our rates of ICU transfer are in line with data reported from other groups,^41,42 which again may add to the generalizability of our findings. We also addressed many potential confounders by including patient characteristics, individual ward units, and (for opioids) pain score in our model, and by controlling for severity of illness with the eCART score, an accurate predictor of ICU transfer and ward cardiac arrest within our population.^32,37 Finally, our robust methodology allowed us to include acute and cumulative medication doses, as well as time, in the model. By performing a discrete-time survival analysis, we were able to evaluate receipt of opioids and benzodiazepines—as well as risk for clinical deterioration—longitudinally, lending strength to our results.

Limitations of our study include its single-center cohort, which may reduce generalizability to other populations. Additionally, because we could not validate the accuracy of—or adherence to—outpatient medication lists, we were unable to identify chronic opioid or benzodiazepine users by these lists. However, patients chronically taking opioids or benzodiazepines would likely receive doses each hospital day; by including 24-hour cumulative doses in our model, we attempted to adjust for some portion of their chronic use. Also, because evaluation of delirium was not objectively recorded in our dataset, we were unable to evaluate the relationship between receipt of these medications and development of delirium, which is an important outcome for hospitalized patients. Finally, neither the diagnoses for which these medications were prescribed, nor the reason for ICU transfer, were present in our dataset, which leaves open the possibility of unmeasured confounding.

After adjustment for important confounders including severity of illness, medication dose, and time, opioids were associated with a slight increase in clinical deterioration on the wards, while benzodiazepines were associated with a much larger risk for deterioration. This finding raises concern about the safety of benzodiazepine use among ward patients and suggests that increased monitoring of patients receiving these medications may be warranted.

Acknowledgment

The authors thank Nicole Twu for administrative support.

Disclosure

Drs. Churpek and Edelson have a patent pending (ARCD. P0535US.P2) for risk stratification algorithms for hospitalized patients. Dr. Churpek is supported by a career development award from the National Heart, Lung, and Blood Institute (K08 HL121080). Dr. Churpek has received honoraria from Chest for invited speaking engagements. In addition, Dr. Edelson has received research support from Philips Healthcare (Andover, Massachusetts), research support from the American Heart Association (Dallas, Texas) and Laerdal Medical (Stavanger, Norway), and research support from Early Sense (Tel Aviv, Israel). She has ownership interest in Quant HC (Chicago, Illinois), which is developing products for risk stratification of hospitalized patients. Dr. Mokhlesi is supported by National Institutes of Health grant R01HL119161. Dr. Mokhlesi has served as a consultant to Philips/Respironics and has received research support from Philips/Respironics. Preliminary versions of these data were presented as a poster presentation at the 2016 meeting of the American Thoracic Society, May 17, 2016; San Francisco, California.

For over a decade, private and public payers have implemented policies aimed at reducing rates of inpatient hospitalization. One approach for doing so is to improve ambulatory care, which can reduce the need for hospital-based acute care. Another approach is to stabilize acutely ill patients and discharge them from the emergency department (ED) or following a period of observation.¹ Private payers are entering into value-based contracting arrangements with hospitals and health systems to improve the quality of ambulatory care and lower healthcare expenditures.² Enrollment in managed care programs has grown among Medicaid recipients for similar reasons.³ Policies of the Centers for Medicare & Medicaid Services (CMS) encourage improvements in ambulatory care as well as observation of Medicare beneficiaries instead of inpatient admission in certain situations.⁴

Recent studies have documented declines in inpatient admissions and increases in treat-and-release observation stays and ED visits among Medicare beneficiaries.^4-7 However, almost half of all hospitalizations unrelated to childbirth occur among patients with private insurance, Medicaid, or no insurance.⁸ Less is known about shifts in the nature of hospital-based acute care among these populations. Such shifts would have implications for quality of care, patient outcomes, and costs. Therefore, further investigation is warranted.

Our objective was to investigate recent trends in payer-specific population-based rates of adults using inpatient, observation, and ED services. We focused on 10 medical conditions that are common reasons for hospital-based acute care: heart failure, bacterial pneumonia, chronic obstructive pulmonary disease, asthma, dehydration, urinary tract infection, uncontrolled diabetes, diabetes with long-term complications, diabetes with short-term complications, and hypertension. These conditions constitute more than 20% of inpatient stays in the general medical service line, can be affected by improvements in ambulatory care, and provided a consistent set of diagnoses to track trends over time.⁹ We used 2009 and 2013 data from four states to examine trends among individuals with private insurance, Medicare, Medicaid, and no insurance.

We obtained encounter-level data for Georgia, Nebraska, South Carolina, and Tennessee from the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP).¹⁰ Using encrypted patient identifiers, we linked inpatient admissions from the 2009 and 2013 State Inpatient Databases, observation stays from the State Ambulatory Surgery and Services Databases, and ED visits from State Emergency Department Databases.

We defined the 10 medical conditions using numerator specifications from the ICD-9-CM v 5.0 AHRQ Prevention Quality Indicators (see Appendix). At most, 1 inpatient admission, 1 observation stay, and 1 ED visit for a study condition was counted for each adult in each year. Limiting the number of visits minimized the skew caused by multiple uses of the same service.

Using the American Community Survey, we calculated utilization rates for each type of service per 100,000 population in four payer and age groups: privately insured adults, Medicaid recipients, and uninsured adults 18 to 64 years, as well as Medicare beneficiaries 65 years and older. For each group, we also examined the origin of inpatient admissions—those who were directly admitted without evaluation in the ED, those admitted from the ED, and ED visits leading to observation stays and then inpatient admission.

Comparing 2009 and 2013, population-based rates of adults with 1 or more inpatient admissions for 10 common medical conditions declined, whereas rates of adults with treat-and-release observation stays rose. Changes in rates of treat-and-release ED visits varied across payers but were small relative to the substantial declines in inpatient admissions (Figure 1). In addition, a growing percentage of inpatient admissions began as observation stays and fewer adults were admitted directly, except among uninsured individuals (Figure 2).

Private Payers, 18 to 64 Years

The rate of adults with treat-and-release observation stays rose (+12.0%, 30 to 33 per 100,000 private payer population, P < 0.001). The rate of adults with treat-and-release ED visits declined (–9.0%, 713 to 648 per 100,000 population, P < 0.001), but by less than for inpatient admissions (–28.2%, 231 to 166 per 100,000 population, P < 0.001; Figure 1A). The percentage of inpatient admissions that began as observation stays rose (from 4.1% to 5.4%, P = 0.041), as did the percentage of admissions originating in the ED (from 66.4% to 71.5%, P ≤ 0.001; Figure 2).

Medicare, 65 Years and Older

The rate of adults with inpatient admissions declined (–17.0%, 2669 to 2216 per 100,000 Medicare population, P < 0.001). Rates rose for adults with treat-and-release ED visits (+3.9%, 1887 to 1961 per 100,000 population, P < 0.001) and treat-and-release observation stays (+32.9%, 234 to 311 per 100,000 population, P < 0.001; Figure 1B). The percentage of inpatient admissions that began as observation stays also rose (5.4% to 9.1%, P < 0.001; Figure 2).

Medicaid, 18 to 64 Years

The rate of adults with inpatient admissions declined (–15.3%, 1100 to 931 per 100,000 Medicaid population, P < 0.001), whereas treat-and-release ED visits remained flat (–1.5%, 4867 to 4792 per 100,000 population, P = 0.413) and treat-and-release observation stays rose (+18.1%, 196 to 232 per 100,000 population, P < 0.001; Figure 1C). The percentage of inpatient admissions that began as observation stays rose (5.9% to 8.1%, P = 0.022; Figure 2).

Uninsured, 18 to 64 Years

The rate of adults with inpatient admissions declined (–5.2%, 296 to 281 per 100,000 uninsured population, P = 0.003), whereas rates rose for treat-and-release ED visits (+8.9%, 1888 to 2057 per 100,000 population, P < 0.001) and treat-and-release observation stays (34.7%, 54 to 73 per 100,000 population, P < 0.001; Figure 1D). The source of inpatient admissions remained stable (Figure 2).

Data on hospital encounters from four states show that both ED visits and observation stays are playing an increasing role in hospital-based acute care for 10 common conditions among populations insured by private payers, Medicare, and Medicaid, as well as those without insurance. Compared with 2009, in 2013 substantially fewer individuals had inpatient admissions, and patients were more likely to be discharged from the ED or discharged following observation without receiving inpatient care. Additionally, an increasing percentage of inpatient admissions followed observation stays, whereas direct admissions declined.

Previous authors also have reported declines in inpatient stays for these same conditions.¹¹ Others have reported increases in the use of observation stays for diverse conditions among patients with private insurance, Medicare beneficiaries, and veterans.^4,12,13 The unique attributes of HCUP databases from these four states (eg, all-payer data including patient linkage numbers across inpatient, observation, and ED care) enabled us to assess concurrent shifts in hospital-based acute care from inpatient to outpatient care among multiple payer populations. A recent analysis reported declines in readmissions and increases in observation visits occurring within 30 days after hospitalization among Medicare beneficiaries with heart failure, acute myocardial infarction, or pneumonia.¹⁴ Future research should examine trends in readmissions and observation visits following hospitalization among multiple payer populations.

These shifts raise two important questions. The first pertains to quality of care, including outcomes. Although dedicated observation units with condition-specific care pathways can be associated with shorter stays and fewer admissions, many patients placed under observation are neither in dedicated units nor subject to care pathways.^15,16 Systems for monitoring quality of care are less developed for observation care. The CMS publicly reports hospital-level data on quality of ED and inpatient care, including for several of the conditions we studied, but no measures apply to observation stays.¹⁷ Little is known about whether shifts from inpatient care to observation status or discharge from the ED are associated with different health outcomes.

The second issue is patients’ out-of-pocket costs. Although shifts from inpatient admissions to observation stays can reduce costs to payers,¹⁵ effects on patient out-of-pocket costs are uncertain and may vary. For privately insured patients, out-of-pocket costs may be up to four times higher for observation than for inpatient care.¹⁸ For Medicare beneficiaries, out-of-pocket costs can be higher for observation than for inpatient stays, particularly when patients receive costly medications or are discharged to skilled nursing facilities;^19,20 however, having secondary insurance dramatically reduces out-of-pocket costs.²¹ We are not aware of data on Medicaid recipients or uninsured individuals.

This study has limitations. Only four states had data needed for these analyses, so generalization to other states is limited. Our analysis was descriptive and did not control for case mix, evaluate specific policies by any payer, or assess the full volume of visits among high utilizers. Movement of healthier or sicker individuals across payers could have contributed to temporal trends, but findings were similar across payers.

In conclusion, among 10 common medical conditions and three major payer populations and uninsured individuals in four states, inpatient admissions declined, and care shifted toward treat-and-release ED visits and observation stays. The number of inpatient admissions that began as observation stays also increased. Given these trends and the possibility that such shifts may be widespread and continue beyond 2013, quality of care, outcomes, and costs to patients warrant further evaluation.

Acknowledgments

The authors gratefully acknowledge Minya Sheng, MS (Truven Health Analytics) for assistance in programming and data management, and Paige Jackson, MS and Linda Lee, PhD, (Truven Health Analytics) for providing editorial review of the manuscript. They also wish to acknowledge the four HCUP Partner organizations that contributed to the 2009 and 2013 HCUP state databases used in this study: Georgia Hospital Association, Nebraska Hospital Association, South Carolina Revenue and Fiscal Affairs Office, and Tennessee Hospital Association.

Disclosure

Funding for this study was provided by the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP) (Contract No. HHSA-290-2013-00002-C). The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services. The authors have no conflicts of interest to declare or financial disclosures.

For over a decade, private and public payers have implemented policies aimed at reducing rates of inpatient hospitalization. One approach for doing so is to improve ambulatory care, which can reduce the need for hospital-based acute care. Another approach is to stabilize acutely ill patients and discharge them from the emergency department (ED) or following a period of observation.¹ Private payers are entering into value-based contracting arrangements with hospitals and health systems to improve the quality of ambulatory care and lower healthcare expenditures.² Enrollment in managed care programs has grown among Medicaid recipients for similar reasons.³ Policies of the Centers for Medicare & Medicaid Services (CMS) encourage improvements in ambulatory care as well as observation of Medicare beneficiaries instead of inpatient admission in certain situations.⁴

Recent studies have documented declines in inpatient admissions and increases in treat-and-release observation stays and ED visits among Medicare beneficiaries.^4-7 However, almost half of all hospitalizations unrelated to childbirth occur among patients with private insurance, Medicaid, or no insurance.⁸ Less is known about shifts in the nature of hospital-based acute care among these populations. Such shifts would have implications for quality of care, patient outcomes, and costs. Therefore, further investigation is warranted.

Our objective was to investigate recent trends in payer-specific population-based rates of adults using inpatient, observation, and ED services. We focused on 10 medical conditions that are common reasons for hospital-based acute care: heart failure, bacterial pneumonia, chronic obstructive pulmonary disease, asthma, dehydration, urinary tract infection, uncontrolled diabetes, diabetes with long-term complications, diabetes with short-term complications, and hypertension. These conditions constitute more than 20% of inpatient stays in the general medical service line, can be affected by improvements in ambulatory care, and provided a consistent set of diagnoses to track trends over time.⁹ We used 2009 and 2013 data from four states to examine trends among individuals with private insurance, Medicare, Medicaid, and no insurance.

We obtained encounter-level data for Georgia, Nebraska, South Carolina, and Tennessee from the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP).¹⁰ Using encrypted patient identifiers, we linked inpatient admissions from the 2009 and 2013 State Inpatient Databases, observation stays from the State Ambulatory Surgery and Services Databases, and ED visits from State Emergency Department Databases.

We defined the 10 medical conditions using numerator specifications from the ICD-9-CM v 5.0 AHRQ Prevention Quality Indicators (see Appendix). At most, 1 inpatient admission, 1 observation stay, and 1 ED visit for a study condition was counted for each adult in each year. Limiting the number of visits minimized the skew caused by multiple uses of the same service.

Using the American Community Survey, we calculated utilization rates for each type of service per 100,000 population in four payer and age groups: privately insured adults, Medicaid recipients, and uninsured adults 18 to 64 years, as well as Medicare beneficiaries 65 years and older. For each group, we also examined the origin of inpatient admissions—those who were directly admitted without evaluation in the ED, those admitted from the ED, and ED visits leading to observation stays and then inpatient admission.

Comparing 2009 and 2013, population-based rates of adults with 1 or more inpatient admissions for 10 common medical conditions declined, whereas rates of adults with treat-and-release observation stays rose. Changes in rates of treat-and-release ED visits varied across payers but were small relative to the substantial declines in inpatient admissions (Figure 1). In addition, a growing percentage of inpatient admissions began as observation stays and fewer adults were admitted directly, except among uninsured individuals (Figure 2).

Private Payers, 18 to 64 Years

The rate of adults with treat-and-release observation stays rose (+12.0%, 30 to 33 per 100,000 private payer population, P < 0.001). The rate of adults with treat-and-release ED visits declined (–9.0%, 713 to 648 per 100,000 population, P < 0.001), but by less than for inpatient admissions (–28.2%, 231 to 166 per 100,000 population, P < 0.001; Figure 1A). The percentage of inpatient admissions that began as observation stays rose (from 4.1% to 5.4%, P = 0.041), as did the percentage of admissions originating in the ED (from 66.4% to 71.5%, P ≤ 0.001; Figure 2).

Medicare, 65 Years and Older

The rate of adults with inpatient admissions declined (–17.0%, 2669 to 2216 per 100,000 Medicare population, P < 0.001). Rates rose for adults with treat-and-release ED visits (+3.9%, 1887 to 1961 per 100,000 population, P < 0.001) and treat-and-release observation stays (+32.9%, 234 to 311 per 100,000 population, P < 0.001; Figure 1B). The percentage of inpatient admissions that began as observation stays also rose (5.4% to 9.1%, P < 0.001; Figure 2).

Medicaid, 18 to 64 Years

The rate of adults with inpatient admissions declined (–15.3%, 1100 to 931 per 100,000 Medicaid population, P < 0.001), whereas treat-and-release ED visits remained flat (–1.5%, 4867 to 4792 per 100,000 population, P = 0.413) and treat-and-release observation stays rose (+18.1%, 196 to 232 per 100,000 population, P < 0.001; Figure 1C). The percentage of inpatient admissions that began as observation stays rose (5.9% to 8.1%, P = 0.022; Figure 2).

Uninsured, 18 to 64 Years

The rate of adults with inpatient admissions declined (–5.2%, 296 to 281 per 100,000 uninsured population, P = 0.003), whereas rates rose for treat-and-release ED visits (+8.9%, 1888 to 2057 per 100,000 population, P < 0.001) and treat-and-release observation stays (34.7%, 54 to 73 per 100,000 population, P < 0.001; Figure 1D). The source of inpatient admissions remained stable (Figure 2).

Data on hospital encounters from four states show that both ED visits and observation stays are playing an increasing role in hospital-based acute care for 10 common conditions among populations insured by private payers, Medicare, and Medicaid, as well as those without insurance. Compared with 2009, in 2013 substantially fewer individuals had inpatient admissions, and patients were more likely to be discharged from the ED or discharged following observation without receiving inpatient care. Additionally, an increasing percentage of inpatient admissions followed observation stays, whereas direct admissions declined.

Previous authors also have reported declines in inpatient stays for these same conditions.¹¹ Others have reported increases in the use of observation stays for diverse conditions among patients with private insurance, Medicare beneficiaries, and veterans.^4,12,13 The unique attributes of HCUP databases from these four states (eg, all-payer data including patient linkage numbers across inpatient, observation, and ED care) enabled us to assess concurrent shifts in hospital-based acute care from inpatient to outpatient care among multiple payer populations. A recent analysis reported declines in readmissions and increases in observation visits occurring within 30 days after hospitalization among Medicare beneficiaries with heart failure, acute myocardial infarction, or pneumonia.¹⁴ Future research should examine trends in readmissions and observation visits following hospitalization among multiple payer populations.

These shifts raise two important questions. The first pertains to quality of care, including outcomes. Although dedicated observation units with condition-specific care pathways can be associated with shorter stays and fewer admissions, many patients placed under observation are neither in dedicated units nor subject to care pathways.^15,16 Systems for monitoring quality of care are less developed for observation care. The CMS publicly reports hospital-level data on quality of ED and inpatient care, including for several of the conditions we studied, but no measures apply to observation stays.¹⁷ Little is known about whether shifts from inpatient care to observation status or discharge from the ED are associated with different health outcomes.

The second issue is patients’ out-of-pocket costs. Although shifts from inpatient admissions to observation stays can reduce costs to payers,¹⁵ effects on patient out-of-pocket costs are uncertain and may vary. For privately insured patients, out-of-pocket costs may be up to four times higher for observation than for inpatient care.¹⁸ For Medicare beneficiaries, out-of-pocket costs can be higher for observation than for inpatient stays, particularly when patients receive costly medications or are discharged to skilled nursing facilities;^19,20 however, having secondary insurance dramatically reduces out-of-pocket costs.²¹ We are not aware of data on Medicaid recipients or uninsured individuals.

This study has limitations. Only four states had data needed for these analyses, so generalization to other states is limited. Our analysis was descriptive and did not control for case mix, evaluate specific policies by any payer, or assess the full volume of visits among high utilizers. Movement of healthier or sicker individuals across payers could have contributed to temporal trends, but findings were similar across payers.

In conclusion, among 10 common medical conditions and three major payer populations and uninsured individuals in four states, inpatient admissions declined, and care shifted toward treat-and-release ED visits and observation stays. The number of inpatient admissions that began as observation stays also increased. Given these trends and the possibility that such shifts may be widespread and continue beyond 2013, quality of care, outcomes, and costs to patients warrant further evaluation.

Acknowledgments

The authors gratefully acknowledge Minya Sheng, MS (Truven Health Analytics) for assistance in programming and data management, and Paige Jackson, MS and Linda Lee, PhD, (Truven Health Analytics) for providing editorial review of the manuscript. They also wish to acknowledge the four HCUP Partner organizations that contributed to the 2009 and 2013 HCUP state databases used in this study: Georgia Hospital Association, Nebraska Hospital Association, South Carolina Revenue and Fiscal Affairs Office, and Tennessee Hospital Association.

Disclosure

Funding for this study was provided by the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP) (Contract No. HHSA-290-2013-00002-C). The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services. The authors have no conflicts of interest to declare or financial disclosures.

For over a decade, private and public payers have implemented policies aimed at reducing rates of inpatient hospitalization. One approach for doing so is to improve ambulatory care, which can reduce the need for hospital-based acute care. Another approach is to stabilize acutely ill patients and discharge them from the emergency department (ED) or following a period of observation.¹ Private payers are entering into value-based contracting arrangements with hospitals and health systems to improve the quality of ambulatory care and lower healthcare expenditures.² Enrollment in managed care programs has grown among Medicaid recipients for similar reasons.³ Policies of the Centers for Medicare & Medicaid Services (CMS) encourage improvements in ambulatory care as well as observation of Medicare beneficiaries instead of inpatient admission in certain situations.⁴

Recent studies have documented declines in inpatient admissions and increases in treat-and-release observation stays and ED visits among Medicare beneficiaries.^4-7 However, almost half of all hospitalizations unrelated to childbirth occur among patients with private insurance, Medicaid, or no insurance.⁸ Less is known about shifts in the nature of hospital-based acute care among these populations. Such shifts would have implications for quality of care, patient outcomes, and costs. Therefore, further investigation is warranted.

Our objective was to investigate recent trends in payer-specific population-based rates of adults using inpatient, observation, and ED services. We focused on 10 medical conditions that are common reasons for hospital-based acute care: heart failure, bacterial pneumonia, chronic obstructive pulmonary disease, asthma, dehydration, urinary tract infection, uncontrolled diabetes, diabetes with long-term complications, diabetes with short-term complications, and hypertension. These conditions constitute more than 20% of inpatient stays in the general medical service line, can be affected by improvements in ambulatory care, and provided a consistent set of diagnoses to track trends over time.⁹ We used 2009 and 2013 data from four states to examine trends among individuals with private insurance, Medicare, Medicaid, and no insurance.

We obtained encounter-level data for Georgia, Nebraska, South Carolina, and Tennessee from the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP).¹⁰ Using encrypted patient identifiers, we linked inpatient admissions from the 2009 and 2013 State Inpatient Databases, observation stays from the State Ambulatory Surgery and Services Databases, and ED visits from State Emergency Department Databases.

We defined the 10 medical conditions using numerator specifications from the ICD-9-CM v 5.0 AHRQ Prevention Quality Indicators (see Appendix). At most, 1 inpatient admission, 1 observation stay, and 1 ED visit for a study condition was counted for each adult in each year. Limiting the number of visits minimized the skew caused by multiple uses of the same service.

Using the American Community Survey, we calculated utilization rates for each type of service per 100,000 population in four payer and age groups: privately insured adults, Medicaid recipients, and uninsured adults 18 to 64 years, as well as Medicare beneficiaries 65 years and older. For each group, we also examined the origin of inpatient admissions—those who were directly admitted without evaluation in the ED, those admitted from the ED, and ED visits leading to observation stays and then inpatient admission.

Comparing 2009 and 2013, population-based rates of adults with 1 or more inpatient admissions for 10 common medical conditions declined, whereas rates of adults with treat-and-release observation stays rose. Changes in rates of treat-and-release ED visits varied across payers but were small relative to the substantial declines in inpatient admissions (Figure 1). In addition, a growing percentage of inpatient admissions began as observation stays and fewer adults were admitted directly, except among uninsured individuals (Figure 2).

Private Payers, 18 to 64 Years

The rate of adults with treat-and-release observation stays rose (+12.0%, 30 to 33 per 100,000 private payer population, P < 0.001). The rate of adults with treat-and-release ED visits declined (–9.0%, 713 to 648 per 100,000 population, P < 0.001), but by less than for inpatient admissions (–28.2%, 231 to 166 per 100,000 population, P < 0.001; Figure 1A). The percentage of inpatient admissions that began as observation stays rose (from 4.1% to 5.4%, P = 0.041), as did the percentage of admissions originating in the ED (from 66.4% to 71.5%, P ≤ 0.001; Figure 2).

Medicare, 65 Years and Older

The rate of adults with inpatient admissions declined (–17.0%, 2669 to 2216 per 100,000 Medicare population, P < 0.001). Rates rose for adults with treat-and-release ED visits (+3.9%, 1887 to 1961 per 100,000 population, P < 0.001) and treat-and-release observation stays (+32.9%, 234 to 311 per 100,000 population, P < 0.001; Figure 1B). The percentage of inpatient admissions that began as observation stays also rose (5.4% to 9.1%, P < 0.001; Figure 2).

Medicaid, 18 to 64 Years

The rate of adults with inpatient admissions declined (–15.3%, 1100 to 931 per 100,000 Medicaid population, P < 0.001), whereas treat-and-release ED visits remained flat (–1.5%, 4867 to 4792 per 100,000 population, P = 0.413) and treat-and-release observation stays rose (+18.1%, 196 to 232 per 100,000 population, P < 0.001; Figure 1C). The percentage of inpatient admissions that began as observation stays rose (5.9% to 8.1%, P = 0.022; Figure 2).

Uninsured, 18 to 64 Years

The rate of adults with inpatient admissions declined (–5.2%, 296 to 281 per 100,000 uninsured population, P = 0.003), whereas rates rose for treat-and-release ED visits (+8.9%, 1888 to 2057 per 100,000 population, P < 0.001) and treat-and-release observation stays (34.7%, 54 to 73 per 100,000 population, P < 0.001; Figure 1D). The source of inpatient admissions remained stable (Figure 2).

Data on hospital encounters from four states show that both ED visits and observation stays are playing an increasing role in hospital-based acute care for 10 common conditions among populations insured by private payers, Medicare, and Medicaid, as well as those without insurance. Compared with 2009, in 2013 substantially fewer individuals had inpatient admissions, and patients were more likely to be discharged from the ED or discharged following observation without receiving inpatient care. Additionally, an increasing percentage of inpatient admissions followed observation stays, whereas direct admissions declined.

Previous authors also have reported declines in inpatient stays for these same conditions.¹¹ Others have reported increases in the use of observation stays for diverse conditions among patients with private insurance, Medicare beneficiaries, and veterans.^4,12,13 The unique attributes of HCUP databases from these four states (eg, all-payer data including patient linkage numbers across inpatient, observation, and ED care) enabled us to assess concurrent shifts in hospital-based acute care from inpatient to outpatient care among multiple payer populations. A recent analysis reported declines in readmissions and increases in observation visits occurring within 30 days after hospitalization among Medicare beneficiaries with heart failure, acute myocardial infarction, or pneumonia.¹⁴ Future research should examine trends in readmissions and observation visits following hospitalization among multiple payer populations.

These shifts raise two important questions. The first pertains to quality of care, including outcomes. Although dedicated observation units with condition-specific care pathways can be associated with shorter stays and fewer admissions, many patients placed under observation are neither in dedicated units nor subject to care pathways.^15,16 Systems for monitoring quality of care are less developed for observation care. The CMS publicly reports hospital-level data on quality of ED and inpatient care, including for several of the conditions we studied, but no measures apply to observation stays.¹⁷ Little is known about whether shifts from inpatient care to observation status or discharge from the ED are associated with different health outcomes.

The second issue is patients’ out-of-pocket costs. Although shifts from inpatient admissions to observation stays can reduce costs to payers,¹⁵ effects on patient out-of-pocket costs are uncertain and may vary. For privately insured patients, out-of-pocket costs may be up to four times higher for observation than for inpatient care.¹⁸ For Medicare beneficiaries, out-of-pocket costs can be higher for observation than for inpatient stays, particularly when patients receive costly medications or are discharged to skilled nursing facilities;^19,20 however, having secondary insurance dramatically reduces out-of-pocket costs.²¹ We are not aware of data on Medicaid recipients or uninsured individuals.

This study has limitations. Only four states had data needed for these analyses, so generalization to other states is limited. Our analysis was descriptive and did not control for case mix, evaluate specific policies by any payer, or assess the full volume of visits among high utilizers. Movement of healthier or sicker individuals across payers could have contributed to temporal trends, but findings were similar across payers.

In conclusion, among 10 common medical conditions and three major payer populations and uninsured individuals in four states, inpatient admissions declined, and care shifted toward treat-and-release ED visits and observation stays. The number of inpatient admissions that began as observation stays also increased. Given these trends and the possibility that such shifts may be widespread and continue beyond 2013, quality of care, outcomes, and costs to patients warrant further evaluation.

Acknowledgments

The authors gratefully acknowledge Minya Sheng, MS (Truven Health Analytics) for assistance in programming and data management, and Paige Jackson, MS and Linda Lee, PhD, (Truven Health Analytics) for providing editorial review of the manuscript. They also wish to acknowledge the four HCUP Partner organizations that contributed to the 2009 and 2013 HCUP state databases used in this study: Georgia Hospital Association, Nebraska Hospital Association, South Carolina Revenue and Fiscal Affairs Office, and Tennessee Hospital Association.

Disclosure

Funding for this study was provided by the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP) (Contract No. HHSA-290-2013-00002-C). The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services. The authors have no conflicts of interest to declare or financial disclosures.

Cardiac telemetry, designed to monitor hospitalized patients with active cardiac conditions, is highly utilized outside the intensive care unit (ICU) and generates a large number of automated alarms. Telemetry is also costly and requires substantial time and attention commitments from nursing and technician staff, who place and maintain the recording devices and address monitoring results. ^1,2 The staff address and dismiss invalid alarms caused by telemetry artifacts, ² such as the misreporting of patient movement as ventricular tachycardia/fibrillation (VT/VF) or the mimicking of asystole by a lead disconnection.

One strategy for addressing telemetry alarms is to have dedicated staff observe telemetry monitors and notify nurses with any events or findings. Studies conducted in the 1990s showed that dedicated monitor watchers, compared with automatically generated alarms alone, did not affect most outcomes ³ but can improve accuracy of arrhythmia detection. ⁴ Since then, given the advances in telemetry detection software, the effect of monitor watchers has not been evaluated. Mindful of the perceived burden of nonactionable telemetry alerts, we wanted to quantify the frequency of automated telemetry alerts in the wards and analyze the proportion of alerts deemed nonactionable by monitor watchers.

We conducted this retrospective study at a 545-bed urban academic hospital in the United States. We reviewed the cases of all non-ICU patients with telemetry monitoring ordered. The telemetry order requires providers specify the indication for monitoring and adjust alert parameters for variables such as heart rate (preset to 60 and 100 beats per minute) and baseline rhythm (preset to normal sinus). Once a telemetry order is received, 5 leads are attached to the patient, and electrocardiographic data begin transmitting to a portable wireless telemetry monitor, or telemeter (Philips Intellispace Telemetry System), which in turn transmits to a central monitoring station in the progressive care unit (PCU; cardiac/pulmonary unit). The majority of patients on telemetry are in the PCU. Telemeters are also located in the general medicine, surgical, and neurologic non-ICU units. Data from a maximum of 96 telemeters in the hospital are simultaneously displayed in the central monitoring station.

At all times, two dedicated monitor watchers oversee the central monitoring station. Watchers are certified medical assistants with extra telemetry-specific training. Each receives a salary of $17 per hour (no benefits), or about $800 per 24-hour day for two watchers. Their role is to respond to audiovisual alerts triggered by the monitoring system—they either contact the bedside nurse or intercept the alert if deemed nonactionable. Consistent with the literature, ⁵ nonactionable alerts and alarms were defined as either “invalid” or “nuisance.” Invalid alerts and alarms misrepresent patient status (eg, patient motion is electronically interpreted as VT/VF), and nuisance alerts and alarms do not require clinical intervention (eg, persistent sinus tachycardia has already been communicated to the nurse or provider). Monitor watchers must intercept the alert within a limited amount of time: 15 seconds for suspected lethal alerts (asystole, VT/VF), 30 seconds for extreme tachycardia/bradycardia, and 60 seconds for lead displacement or low battery.

Within the 2-month study period, there were 1917 admissions to, and 1370 transfers to, non-ICU floors, for a total of 3287 unique patient-admissions and 9704 total patient-days. There were 1199 patient admissions with telemetry orders (36.5% of all admissions), 4044 total patient-days of telemetry, and an average of 66.3 patients monitored per day. In addition, the system generated 20,775 alerts, an average of 341 per day, 5.1 per patient-day, 1 every 4 minutes. Overall, 18,051 alerts (87%) were intercepted by monitor watchers, preventing nurse text-alarms. Of all alerts, 91% were from patients on medicine services, including pulmonary and cardiology; 6% were from patients on the neurology floor; and 3% were from patients on the surgery floor.

Forty percent of all alerts were for heart rates deviating outside the ranges set by the provider; of these, the overwhelming majority were intercepted as nuisance alerts (Table). In addition, 26% of all alerts were for maintenance reasons, including issues with batteries or leads. Finally, 34% (6954) were suspected lethal alerts (asystole, VT/VF); of these, 74% (5170) were intercepted by monitor watchers, suggesting they were deemed invalid. None of the suspected lethal alerts triggered a code team activation, indicating there were no telemetry-documented asystole or VT/VF episodes prompting resuscitative efforts. During the study period, there were 7 code team activations. Of the 7 patients, 2 were on telemetry, and their code team activation was for hypoxia detected by pulse oximetry; the other 5 patients, not on telemetry, were found unresponsive or apneic, and 4 of them had confirmed pulseless electrical activity.

In small studies, other investigators have directly observed nurses for hours at a time and assessed their response to telemetry-related alarms. ^1,2 In the present study, we found a very large number of telemetry-detected alerts over a continuous 2-month period. The large majority (87%) of alerts were manually intercepted by monitor watchers before being communicated to a nurse or provider, indicating these alerts did not affect clinical management and likely were either false positives or nonactionable. It is possible that repeat nonactionable alerts, like continued sinus tachycardia or bradycardia, affect decision making, but this may be outside the role of continuous cardiac telemetry. In addition, it is likely that all the lethal alarms (asystole, VT/VF) forwarded to the nurses were invalid, as none resulted in code team activations.

Addressing these alerts is a major issue, as frequent telemetry alarms can lead to alarm fatigue, a widely acknowledged safety concern. ⁶ Furthermore, nonactionable alarms are a time sink, diverting nursing attention from other patient care needs. Finally, nonactionable alarms, especially invalid alarms, can lead to adverse patient outcomes. Although we did not specifically evaluate for harm, an earlier case series found a potential for unnecessary interventions and device implantation as a result of reporting artifactual arrhythmias. ⁷

Our results also highlight the role of monitor watchers in intercepting nonactionable alarms and reducing the alarm burden on nurses. Other investigators have reported on computerized paging systems that directly alert only nurses, ⁸ or on escalated alarm paging systems that let noncrisis alarms self-resolve. ⁹ In contrast, our study used a hybrid 2-step telemetry-monitoring system—an escalated paging system designed to be sensitive and less likely than human monitoring to overlook events, followed by dedicated monitor watchers who are first-responders for a large number of alarms and who increase the specificity of alarms by screening for nonactionable alarms, thereby reducing the number of alarms transmitted to nurses. We think that, for most hospitals, monitor watchers are cost-effective, as their hourly wage is lower than that of registered nurses. Furthermore, monitor watchers can screen alerts faster because they are always at the monitoring station. Their presence reduces the amount of time that nurses need to divert from other clinical tasks in order to walk to the monitoring station to evaluate alerts.

Nonetheless, there remains a large number of nonactionable alerts forwarded as alarms to nurses, likely because of monitor watchers’ inability to address the multitude of alerts, and perhaps because of alarm fatigue. Although this study showed the utility of monitor watchers in decreasing telemetry alarms to nurses, other steps can be taken to reduce telemetry alarm fatigue. A systematic review of alarm frequency interventions ⁵ noted that detection algorithms can be improved to decrease telemetry alert false positives. Another solution, likely easier to implement, is to encourage appropriate alterations in telemetry alarm parameters, which can decrease the alarm proportion. ¹⁰ An essential step is to decrease inappropriate telemetry use regarding the indication for and duration of monitoring, as emphasized by the Choosing Wisely campaign championing American Heart Association (AHA) guidelines for appropriate telemetry use. ¹¹ At our institution, 20.2% of telemetry orders were for indications outside AHA guidelines, and that percentage likely is an underestimate, as this was required self-reporting on ordering. ¹² Telemetry may not frequently result in changes in management in the non-ICU setting, ¹³ and may lead to other harms such as worsening delirium, ¹⁴ so it needs to be evaluated for harm versus benefit per patient before order.

Cardiac telemetry in the non-ICU setting produces a large number of alerts and alarms. The vast majority are not seen or addressed by nurses or physicians, leading to a negligible impact on patient care decisions. Monitor watchers reduce the nursing burden in dealing with telemetry alerts, but we emphasize the need to take additional measures to reduce telemetry-related alerts and thereby reduce alarm-related harms and alarm fatigue.

Acknowledgments

The authors thank Torberg Tonnessen, who was instrumental in providing the telemetry and clinical data used in this study, as well as the numerous Johns Hopkins Bayview Medical Center nurses, patient care technicians, and monitor watchers who answered questions about telemetry processes and allowed their work to be observed.

Disclosure

Nothing to report.

Cardiac telemetry, designed to monitor hospitalized patients with active cardiac conditions, is highly utilized outside the intensive care unit (ICU) and generates a large number of automated alarms. Telemetry is also costly and requires substantial time and attention commitments from nursing and technician staff, who place and maintain the recording devices and address monitoring results. ^1,2 The staff address and dismiss invalid alarms caused by telemetry artifacts, ² such as the misreporting of patient movement as ventricular tachycardia/fibrillation (VT/VF) or the mimicking of asystole by a lead disconnection.

One strategy for addressing telemetry alarms is to have dedicated staff observe telemetry monitors and notify nurses with any events or findings. Studies conducted in the 1990s showed that dedicated monitor watchers, compared with automatically generated alarms alone, did not affect most outcomes ³ but can improve accuracy of arrhythmia detection. ⁴ Since then, given the advances in telemetry detection software, the effect of monitor watchers has not been evaluated. Mindful of the perceived burden of nonactionable telemetry alerts, we wanted to quantify the frequency of automated telemetry alerts in the wards and analyze the proportion of alerts deemed nonactionable by monitor watchers.

We conducted this retrospective study at a 545-bed urban academic hospital in the United States. We reviewed the cases of all non-ICU patients with telemetry monitoring ordered. The telemetry order requires providers specify the indication for monitoring and adjust alert parameters for variables such as heart rate (preset to 60 and 100 beats per minute) and baseline rhythm (preset to normal sinus). Once a telemetry order is received, 5 leads are attached to the patient, and electrocardiographic data begin transmitting to a portable wireless telemetry monitor, or telemeter (Philips Intellispace Telemetry System), which in turn transmits to a central monitoring station in the progressive care unit (PCU; cardiac/pulmonary unit). The majority of patients on telemetry are in the PCU. Telemeters are also located in the general medicine, surgical, and neurologic non-ICU units. Data from a maximum of 96 telemeters in the hospital are simultaneously displayed in the central monitoring station.

At all times, two dedicated monitor watchers oversee the central monitoring station. Watchers are certified medical assistants with extra telemetry-specific training. Each receives a salary of $17 per hour (no benefits), or about $800 per 24-hour day for two watchers. Their role is to respond to audiovisual alerts triggered by the monitoring system—they either contact the bedside nurse or intercept the alert if deemed nonactionable. Consistent with the literature, ⁵ nonactionable alerts and alarms were defined as either “invalid” or “nuisance.” Invalid alerts and alarms misrepresent patient status (eg, patient motion is electronically interpreted as VT/VF), and nuisance alerts and alarms do not require clinical intervention (eg, persistent sinus tachycardia has already been communicated to the nurse or provider). Monitor watchers must intercept the alert within a limited amount of time: 15 seconds for suspected lethal alerts (asystole, VT/VF), 30 seconds for extreme tachycardia/bradycardia, and 60 seconds for lead displacement or low battery.

Within the 2-month study period, there were 1917 admissions to, and 1370 transfers to, non-ICU floors, for a total of 3287 unique patient-admissions and 9704 total patient-days. There were 1199 patient admissions with telemetry orders (36.5% of all admissions), 4044 total patient-days of telemetry, and an average of 66.3 patients monitored per day. In addition, the system generated 20,775 alerts, an average of 341 per day, 5.1 per patient-day, 1 every 4 minutes. Overall, 18,051 alerts (87%) were intercepted by monitor watchers, preventing nurse text-alarms. Of all alerts, 91% were from patients on medicine services, including pulmonary and cardiology; 6% were from patients on the neurology floor; and 3% were from patients on the surgery floor.

Forty percent of all alerts were for heart rates deviating outside the ranges set by the provider; of these, the overwhelming majority were intercepted as nuisance alerts (Table). In addition, 26% of all alerts were for maintenance reasons, including issues with batteries or leads. Finally, 34% (6954) were suspected lethal alerts (asystole, VT/VF); of these, 74% (5170) were intercepted by monitor watchers, suggesting they were deemed invalid. None of the suspected lethal alerts triggered a code team activation, indicating there were no telemetry-documented asystole or VT/VF episodes prompting resuscitative efforts. During the study period, there were 7 code team activations. Of the 7 patients, 2 were on telemetry, and their code team activation was for hypoxia detected by pulse oximetry; the other 5 patients, not on telemetry, were found unresponsive or apneic, and 4 of them had confirmed pulseless electrical activity.

In small studies, other investigators have directly observed nurses for hours at a time and assessed their response to telemetry-related alarms. ^1,2 In the present study, we found a very large number of telemetry-detected alerts over a continuous 2-month period. The large majority (87%) of alerts were manually intercepted by monitor watchers before being communicated to a nurse or provider, indicating these alerts did not affect clinical management and likely were either false positives or nonactionable. It is possible that repeat nonactionable alerts, like continued sinus tachycardia or bradycardia, affect decision making, but this may be outside the role of continuous cardiac telemetry. In addition, it is likely that all the lethal alarms (asystole, VT/VF) forwarded to the nurses were invalid, as none resulted in code team activations.

Addressing these alerts is a major issue, as frequent telemetry alarms can lead to alarm fatigue, a widely acknowledged safety concern. ⁶ Furthermore, nonactionable alarms are a time sink, diverting nursing attention from other patient care needs. Finally, nonactionable alarms, especially invalid alarms, can lead to adverse patient outcomes. Although we did not specifically evaluate for harm, an earlier case series found a potential for unnecessary interventions and device implantation as a result of reporting artifactual arrhythmias. ⁷

Our results also highlight the role of monitor watchers in intercepting nonactionable alarms and reducing the alarm burden on nurses. Other investigators have reported on computerized paging systems that directly alert only nurses, ⁸ or on escalated alarm paging systems that let noncrisis alarms self-resolve. ⁹ In contrast, our study used a hybrid 2-step telemetry-monitoring system—an escalated paging system designed to be sensitive and less likely than human monitoring to overlook events, followed by dedicated monitor watchers who are first-responders for a large number of alarms and who increase the specificity of alarms by screening for nonactionable alarms, thereby reducing the number of alarms transmitted to nurses. We think that, for most hospitals, monitor watchers are cost-effective, as their hourly wage is lower than that of registered nurses. Furthermore, monitor watchers can screen alerts faster because they are always at the monitoring station. Their presence reduces the amount of time that nurses need to divert from other clinical tasks in order to walk to the monitoring station to evaluate alerts.

Nonetheless, there remains a large number of nonactionable alerts forwarded as alarms to nurses, likely because of monitor watchers’ inability to address the multitude of alerts, and perhaps because of alarm fatigue. Although this study showed the utility of monitor watchers in decreasing telemetry alarms to nurses, other steps can be taken to reduce telemetry alarm fatigue. A systematic review of alarm frequency interventions ⁵ noted that detection algorithms can be improved to decrease telemetry alert false positives. Another solution, likely easier to implement, is to encourage appropriate alterations in telemetry alarm parameters, which can decrease the alarm proportion. ¹⁰ An essential step is to decrease inappropriate telemetry use regarding the indication for and duration of monitoring, as emphasized by the Choosing Wisely campaign championing American Heart Association (AHA) guidelines for appropriate telemetry use. ¹¹ At our institution, 20.2% of telemetry orders were for indications outside AHA guidelines, and that percentage likely is an underestimate, as this was required self-reporting on ordering. ¹² Telemetry may not frequently result in changes in management in the non-ICU setting, ¹³ and may lead to other harms such as worsening delirium, ¹⁴ so it needs to be evaluated for harm versus benefit per patient before order.

Cardiac telemetry in the non-ICU setting produces a large number of alerts and alarms. The vast majority are not seen or addressed by nurses or physicians, leading to a negligible impact on patient care decisions. Monitor watchers reduce the nursing burden in dealing with telemetry alerts, but we emphasize the need to take additional measures to reduce telemetry-related alerts and thereby reduce alarm-related harms and alarm fatigue.

Acknowledgments

The authors thank Torberg Tonnessen, who was instrumental in providing the telemetry and clinical data used in this study, as well as the numerous Johns Hopkins Bayview Medical Center nurses, patient care technicians, and monitor watchers who answered questions about telemetry processes and allowed their work to be observed.

Disclosure

Nothing to report.

Cardiac telemetry, designed to monitor hospitalized patients with active cardiac conditions, is highly utilized outside the intensive care unit (ICU) and generates a large number of automated alarms. Telemetry is also costly and requires substantial time and attention commitments from nursing and technician staff, who place and maintain the recording devices and address monitoring results. ^1,2 The staff address and dismiss invalid alarms caused by telemetry artifacts, ² such as the misreporting of patient movement as ventricular tachycardia/fibrillation (VT/VF) or the mimicking of asystole by a lead disconnection.

One strategy for addressing telemetry alarms is to have dedicated staff observe telemetry monitors and notify nurses with any events or findings. Studies conducted in the 1990s showed that dedicated monitor watchers, compared with automatically generated alarms alone, did not affect most outcomes ³ but can improve accuracy of arrhythmia detection. ⁴ Since then, given the advances in telemetry detection software, the effect of monitor watchers has not been evaluated. Mindful of the perceived burden of nonactionable telemetry alerts, we wanted to quantify the frequency of automated telemetry alerts in the wards and analyze the proportion of alerts deemed nonactionable by monitor watchers.

We conducted this retrospective study at a 545-bed urban academic hospital in the United States. We reviewed the cases of all non-ICU patients with telemetry monitoring ordered. The telemetry order requires providers specify the indication for monitoring and adjust alert parameters for variables such as heart rate (preset to 60 and 100 beats per minute) and baseline rhythm (preset to normal sinus). Once a telemetry order is received, 5 leads are attached to the patient, and electrocardiographic data begin transmitting to a portable wireless telemetry monitor, or telemeter (Philips Intellispace Telemetry System), which in turn transmits to a central monitoring station in the progressive care unit (PCU; cardiac/pulmonary unit). The majority of patients on telemetry are in the PCU. Telemeters are also located in the general medicine, surgical, and neurologic non-ICU units. Data from a maximum of 96 telemeters in the hospital are simultaneously displayed in the central monitoring station.

At all times, two dedicated monitor watchers oversee the central monitoring station. Watchers are certified medical assistants with extra telemetry-specific training. Each receives a salary of $17 per hour (no benefits), or about $800 per 24-hour day for two watchers. Their role is to respond to audiovisual alerts triggered by the monitoring system—they either contact the bedside nurse or intercept the alert if deemed nonactionable. Consistent with the literature, ⁵ nonactionable alerts and alarms were defined as either “invalid” or “nuisance.” Invalid alerts and alarms misrepresent patient status (eg, patient motion is electronically interpreted as VT/VF), and nuisance alerts and alarms do not require clinical intervention (eg, persistent sinus tachycardia has already been communicated to the nurse or provider). Monitor watchers must intercept the alert within a limited amount of time: 15 seconds for suspected lethal alerts (asystole, VT/VF), 30 seconds for extreme tachycardia/bradycardia, and 60 seconds for lead displacement or low battery.

Within the 2-month study period, there were 1917 admissions to, and 1370 transfers to, non-ICU floors, for a total of 3287 unique patient-admissions and 9704 total patient-days. There were 1199 patient admissions with telemetry orders (36.5% of all admissions), 4044 total patient-days of telemetry, and an average of 66.3 patients monitored per day. In addition, the system generated 20,775 alerts, an average of 341 per day, 5.1 per patient-day, 1 every 4 minutes. Overall, 18,051 alerts (87%) were intercepted by monitor watchers, preventing nurse text-alarms. Of all alerts, 91% were from patients on medicine services, including pulmonary and cardiology; 6% were from patients on the neurology floor; and 3% were from patients on the surgery floor.

Forty percent of all alerts were for heart rates deviating outside the ranges set by the provider; of these, the overwhelming majority were intercepted as nuisance alerts (Table). In addition, 26% of all alerts were for maintenance reasons, including issues with batteries or leads. Finally, 34% (6954) were suspected lethal alerts (asystole, VT/VF); of these, 74% (5170) were intercepted by monitor watchers, suggesting they were deemed invalid. None of the suspected lethal alerts triggered a code team activation, indicating there were no telemetry-documented asystole or VT/VF episodes prompting resuscitative efforts. During the study period, there were 7 code team activations. Of the 7 patients, 2 were on telemetry, and their code team activation was for hypoxia detected by pulse oximetry; the other 5 patients, not on telemetry, were found unresponsive or apneic, and 4 of them had confirmed pulseless electrical activity.

In small studies, other investigators have directly observed nurses for hours at a time and assessed their response to telemetry-related alarms. ^1,2 In the present study, we found a very large number of telemetry-detected alerts over a continuous 2-month period. The large majority (87%) of alerts were manually intercepted by monitor watchers before being communicated to a nurse or provider, indicating these alerts did not affect clinical management and likely were either false positives or nonactionable. It is possible that repeat nonactionable alerts, like continued sinus tachycardia or bradycardia, affect decision making, but this may be outside the role of continuous cardiac telemetry. In addition, it is likely that all the lethal alarms (asystole, VT/VF) forwarded to the nurses were invalid, as none resulted in code team activations.

Addressing these alerts is a major issue, as frequent telemetry alarms can lead to alarm fatigue, a widely acknowledged safety concern. ⁶ Furthermore, nonactionable alarms are a time sink, diverting nursing attention from other patient care needs. Finally, nonactionable alarms, especially invalid alarms, can lead to adverse patient outcomes. Although we did not specifically evaluate for harm, an earlier case series found a potential for unnecessary interventions and device implantation as a result of reporting artifactual arrhythmias. ⁷

Our results also highlight the role of monitor watchers in intercepting nonactionable alarms and reducing the alarm burden on nurses. Other investigators have reported on computerized paging systems that directly alert only nurses, ⁸ or on escalated alarm paging systems that let noncrisis alarms self-resolve. ⁹ In contrast, our study used a hybrid 2-step telemetry-monitoring system—an escalated paging system designed to be sensitive and less likely than human monitoring to overlook events, followed by dedicated monitor watchers who are first-responders for a large number of alarms and who increase the specificity of alarms by screening for nonactionable alarms, thereby reducing the number of alarms transmitted to nurses. We think that, for most hospitals, monitor watchers are cost-effective, as their hourly wage is lower than that of registered nurses. Furthermore, monitor watchers can screen alerts faster because they are always at the monitoring station. Their presence reduces the amount of time that nurses need to divert from other clinical tasks in order to walk to the monitoring station to evaluate alerts.

Nonetheless, there remains a large number of nonactionable alerts forwarded as alarms to nurses, likely because of monitor watchers’ inability to address the multitude of alerts, and perhaps because of alarm fatigue. Although this study showed the utility of monitor watchers in decreasing telemetry alarms to nurses, other steps can be taken to reduce telemetry alarm fatigue. A systematic review of alarm frequency interventions ⁵ noted that detection algorithms can be improved to decrease telemetry alert false positives. Another solution, likely easier to implement, is to encourage appropriate alterations in telemetry alarm parameters, which can decrease the alarm proportion. ¹⁰ An essential step is to decrease inappropriate telemetry use regarding the indication for and duration of monitoring, as emphasized by the Choosing Wisely campaign championing American Heart Association (AHA) guidelines for appropriate telemetry use. ¹¹ At our institution, 20.2% of telemetry orders were for indications outside AHA guidelines, and that percentage likely is an underestimate, as this was required self-reporting on ordering. ¹² Telemetry may not frequently result in changes in management in the non-ICU setting, ¹³ and may lead to other harms such as worsening delirium, ¹⁴ so it needs to be evaluated for harm versus benefit per patient before order.

Cardiac telemetry in the non-ICU setting produces a large number of alerts and alarms. The vast majority are not seen or addressed by nurses or physicians, leading to a negligible impact on patient care decisions. Monitor watchers reduce the nursing burden in dealing with telemetry alerts, but we emphasize the need to take additional measures to reduce telemetry-related alerts and thereby reduce alarm-related harms and alarm fatigue.

Acknowledgments

The authors thank Torberg Tonnessen, who was instrumental in providing the telemetry and clinical data used in this study, as well as the numerous Johns Hopkins Bayview Medical Center nurses, patient care technicians, and monitor watchers who answered questions about telemetry processes and allowed their work to be observed.

Disclosure

Nothing to report.

In the United States, patients 65 years old or older accounted for more than one third of inpatient stays and 42% of inpatient care spending in 2012.¹ Despite the identification of risk factors, the implementation of an array of interventions, and the institution of penalties on hospitals, a subset of older adults continues to spend significant time in the hospital.^2,3

Hospital dependency is a concept that was only recently described. It identifies patients who improve while in the hospital but quickly deteriorate after leaving the hospital, resulting in recurring hospitalizations.⁴ Although little is known about hospital-dependent patients, studies have explored patients’ perspectives on readmissions.^5,6 Nevertheless, it remains unclear whether there are individuals for whom frequent and prolonged hospitalizations are appropriate, and whether there are undisclosed factors that, if addressed, could decrease their hospital dependency. We conducted an exploratory study to ascertain hospital-dependent patients’ perspectives on their needs for hospitalizations.

Study Design

This study was approved by the Yale University Institutional Review Board. From March 2015 to September 2015, Dr. Liu conducted semistructured explorative interviews with patients on the medical units of an academic medical center. Dr. Liu was not directly involved in the care of these patients. An interview guide that includes open-ended questions was created to elicit patients’ perspectives on their need for hospitalizations, health status, and outside-hospital support. This guide was pilot-tested with 6 patients, whose transcripts were not included in the final analysis, to assess for ease of understanding. After the pilot interviews, the questions were revised, and the final guide consists of 12 questions (Supplemental Table).

Recruitment

We used predetermined criteria and a purposeful sampling strategy to select potential study participants. We identified participants by periodically (~once a week) reviewing the electronic medical records of all patients admitted to the medicine service during the study period. Eligible patients were 65 years old or older and had at least 3 hospitalizations over the preceding 6 months. Patients were excluded if they met our chronic critical illness criteria: mechanical ventilation for more than 21 days, history of tracheotomy for failed weaning from mechanical ventilation,⁷ presence of a conservator, or admission only for comfort measures. Participants were recruited until no new themes emerged.

Data Collection

Twenty-nine patients were eligible. We obtained permission from their inpatient providers to approach them about the study. Of the 29 patients, 26 agreed to be interviewed, and 3 declined. Of the 26 participants, 6 underwent pilot interviews, and 20 underwent formal interviews with use of the finalized interview guide. The interviews, conducted in the hospital while the participants were hospitalized, lasted 17 minutes on average. The interviews were transcribed and iteratively analyzed. The themes that emerged from the initial interviews were further explored and validated in subsequent interviews. Interviews were conducted until theoretical saturation was reached and no new themes were derived from them. Demographic information, including age, sex, ethnicity, and marital status, was also collected.

Analysis

Interviews were digitally recorded and transcribed. Independently, two investigators used Atlas Ti software to analyze and code the interview transcriptions. An inductive approach was used to identify new codes from the data.⁸ The coders then met to discuss repeating ideas based on the codes. When a code was identified by one coder but not the other, or when there was disagreement about interpretation of a code, the coders returned to the relevant text to reach consensus and to determine whether to include or discard the code.⁹ We then organized and reorganized repeating ideas based on their conceptual similarities to determine the themes and subthemes.⁹

Twenty patients participated in the formal interviews. Participants’ baseline characteristics are listed in Table 1, and four dominant themes, and their subthemes and exemplary quotations, are listed in Table 2.

Perspectives on Hospital Care

Participants perceived their hospitalizations as inevitable and necessary for survival: “I think if I haven’t come to the hospital, I probably would have died.” Furthermore, participants thought only the hospital had the resources to help them (“The medications they were giving me … you can get that in the hospital but not outside the hospital”) and sustain them (“You are like an old car, and it breaks down little by little, so you have to go in periodically and get the problem fixed, so you will drive it around for a while”).

Feeling Safe in Hospital. Asked how being in the hospital makes them feel, participants attributed their feelings of safety to the constant observation, the availability of providers and nurses, and the idea that hospital care is helping. As one participant stated, “Makes me feel safer in case you go into something like cardiac arrest. You are right here where they can help you.”

Outside-Hospital Support. Despite multiple hospitalizations, most participants reported having social support (“I have the aide. I got the nurses come in. I have my daughter …”), physical support, and medical support (“I have all the doctors”) outside the hospital. A minority of participants questioned the usefulness of the services. One participant described declining the help of visiting nurses because she wanted to be independent and thought that, despite recurrent hospitalizations for physical symptoms, she still had the ability to manage her own medications.

Goals-of-Care Discussion. Some participants reported inadequate discussions about goals of care, health priorities, and health trajectories. In their reports, this inadequacy included not thinking about their goals, despite continued health decline. One participant stated, “Oh, God, I don’t know if I had any conversation like that. … I think until it is really brought to the front, you don’t make a decision really if you don’t have to.” Citing the value of a more established relationship and deeper trust, participants preferred having these serious and personal discussions with their ambulatory care clinicians: “Because I know my doctor much closer. I have been with him for a number of years. The doctors in the hospital seem to be nice and competent, but I don’t know them.”

Participants considered their hospitalizations a necessity and reported feeling safe in the hospital. Given that most already had support outside the hospital, increasing community services may be inadequate to alter participants’ perceived hospital care needs. On the other hand, a few participants reported declining services that might have prevented hospitalizations. Although there has been a study of treatment refusal among older adults with advanced illnesses,¹⁰ not much is known about refusal of services among this population. Investigators should examine the reasons for refusing services and the effect that refusal has on hospitalizations. Furthermore, although it would have been informative to ascertain clinician perspectives as well, we focused on patient perspectives because less is known on this topic.

Some participants noted their lack of discussion with their clinicians about healthcare goals and probable health trajectories. Barriers to goals-of-care discussion among this highly vulnerable population have been researched from the perspectives of clinicians and other health professionals but not patients themselves.^11,12 Of particular concern in our study is the participant-noted lack of discussion about health trajectories and health priorities, given the decline that occurs in this population and even in those with good care. This inadequacy in discussion suggests continued hospital care may not always be consistent with a patient’s goals. Patients’ desire to have this discussion with their clinicians, with whom they have a relationship, supports the need to involve ambulatory care clinicians, or ensure these patients are cared for by the same clinicians, across healthcare settings.^13,14 Whoever provides the care, the clinician must align treatment with the patient’s goal, whether it is to continue hospital-level care or to transition to palliative care. Such an approach also reflects the core elements of person-centered care.¹⁵

Study Limitations

Participants were recruited from the medicine service at a single large academic center, limiting the study’s generalizability to patients admitted to surgical services or community hospitals. The patients in this small sample were English-speaking and predominantly Caucasian, so our findings may not represent the perspectives of non-English-speaking or minority patients. We did not perform statistical analysis to quantify intercoder reliability. Last, as this was a qualitative study, we cannot comment on the relative importance or prevalence of the reasons cited for frequent hospitalizations, and we cannot estimate the proportion of patients who had recurrent hospitalizations and were hospital-dependent.

Implication

Although quantitative research is needed to confirm our findings, the hospital-dependent patients in this study thought their survival required hospital-level care and resources. From their perspective, increasing posthospital and community support may be insufficient to prevent some hospitalizations. The lack of goals-of-care discussion supports attempts to increase efforts to facilitate discussion about health trajectories and health priorities between patients and their preferred clinicians.

Acknowledgments

The authors thank Dr. Grace Jenq for providing feedback on the study design.

Disclosure

Nothing to report.

In the United States, patients 65 years old or older accounted for more than one third of inpatient stays and 42% of inpatient care spending in 2012.¹ Despite the identification of risk factors, the implementation of an array of interventions, and the institution of penalties on hospitals, a subset of older adults continues to spend significant time in the hospital.^2,3

Hospital dependency is a concept that was only recently described. It identifies patients who improve while in the hospital but quickly deteriorate after leaving the hospital, resulting in recurring hospitalizations.⁴ Although little is known about hospital-dependent patients, studies have explored patients’ perspectives on readmissions.^5,6 Nevertheless, it remains unclear whether there are individuals for whom frequent and prolonged hospitalizations are appropriate, and whether there are undisclosed factors that, if addressed, could decrease their hospital dependency. We conducted an exploratory study to ascertain hospital-dependent patients’ perspectives on their needs for hospitalizations.

Study Design

This study was approved by the Yale University Institutional Review Board. From March 2015 to September 2015, Dr. Liu conducted semistructured explorative interviews with patients on the medical units of an academic medical center. Dr. Liu was not directly involved in the care of these patients. An interview guide that includes open-ended questions was created to elicit patients’ perspectives on their need for hospitalizations, health status, and outside-hospital support. This guide was pilot-tested with 6 patients, whose transcripts were not included in the final analysis, to assess for ease of understanding. After the pilot interviews, the questions were revised, and the final guide consists of 12 questions (Supplemental Table).

Recruitment

We used predetermined criteria and a purposeful sampling strategy to select potential study participants. We identified participants by periodically (~once a week) reviewing the electronic medical records of all patients admitted to the medicine service during the study period. Eligible patients were 65 years old or older and had at least 3 hospitalizations over the preceding 6 months. Patients were excluded if they met our chronic critical illness criteria: mechanical ventilation for more than 21 days, history of tracheotomy for failed weaning from mechanical ventilation,⁷ presence of a conservator, or admission only for comfort measures. Participants were recruited until no new themes emerged.

Data Collection

Twenty-nine patients were eligible. We obtained permission from their inpatient providers to approach them about the study. Of the 29 patients, 26 agreed to be interviewed, and 3 declined. Of the 26 participants, 6 underwent pilot interviews, and 20 underwent formal interviews with use of the finalized interview guide. The interviews, conducted in the hospital while the participants were hospitalized, lasted 17 minutes on average. The interviews were transcribed and iteratively analyzed. The themes that emerged from the initial interviews were further explored and validated in subsequent interviews. Interviews were conducted until theoretical saturation was reached and no new themes were derived from them. Demographic information, including age, sex, ethnicity, and marital status, was also collected.

Analysis

Interviews were digitally recorded and transcribed. Independently, two investigators used Atlas Ti software to analyze and code the interview transcriptions. An inductive approach was used to identify new codes from the data.⁸ The coders then met to discuss repeating ideas based on the codes. When a code was identified by one coder but not the other, or when there was disagreement about interpretation of a code, the coders returned to the relevant text to reach consensus and to determine whether to include or discard the code.⁹ We then organized and reorganized repeating ideas based on their conceptual similarities to determine the themes and subthemes.⁹

Twenty patients participated in the formal interviews. Participants’ baseline characteristics are listed in Table 1, and four dominant themes, and their subthemes and exemplary quotations, are listed in Table 2.

Perspectives on Hospital Care

Participants perceived their hospitalizations as inevitable and necessary for survival: “I think if I haven’t come to the hospital, I probably would have died.” Furthermore, participants thought only the hospital had the resources to help them (“The medications they were giving me … you can get that in the hospital but not outside the hospital”) and sustain them (“You are like an old car, and it breaks down little by little, so you have to go in periodically and get the problem fixed, so you will drive it around for a while”).

Feeling Safe in Hospital. Asked how being in the hospital makes them feel, participants attributed their feelings of safety to the constant observation, the availability of providers and nurses, and the idea that hospital care is helping. As one participant stated, “Makes me feel safer in case you go into something like cardiac arrest. You are right here where they can help you.”

Outside-Hospital Support. Despite multiple hospitalizations, most participants reported having social support (“I have the aide. I got the nurses come in. I have my daughter …”), physical support, and medical support (“I have all the doctors”) outside the hospital. A minority of participants questioned the usefulness of the services. One participant described declining the help of visiting nurses because she wanted to be independent and thought that, despite recurrent hospitalizations for physical symptoms, she still had the ability to manage her own medications.

Goals-of-Care Discussion. Some participants reported inadequate discussions about goals of care, health priorities, and health trajectories. In their reports, this inadequacy included not thinking about their goals, despite continued health decline. One participant stated, “Oh, God, I don’t know if I had any conversation like that. … I think until it is really brought to the front, you don’t make a decision really if you don’t have to.” Citing the value of a more established relationship and deeper trust, participants preferred having these serious and personal discussions with their ambulatory care clinicians: “Because I know my doctor much closer. I have been with him for a number of years. The doctors in the hospital seem to be nice and competent, but I don’t know them.”

Participants considered their hospitalizations a necessity and reported feeling safe in the hospital. Given that most already had support outside the hospital, increasing community services may be inadequate to alter participants’ perceived hospital care needs. On the other hand, a few participants reported declining services that might have prevented hospitalizations. Although there has been a study of treatment refusal among older adults with advanced illnesses,¹⁰ not much is known about refusal of services among this population. Investigators should examine the reasons for refusing services and the effect that refusal has on hospitalizations. Furthermore, although it would have been informative to ascertain clinician perspectives as well, we focused on patient perspectives because less is known on this topic.

Some participants noted their lack of discussion with their clinicians about healthcare goals and probable health trajectories. Barriers to goals-of-care discussion among this highly vulnerable population have been researched from the perspectives of clinicians and other health professionals but not patients themselves.^11,12 Of particular concern in our study is the participant-noted lack of discussion about health trajectories and health priorities, given the decline that occurs in this population and even in those with good care. This inadequacy in discussion suggests continued hospital care may not always be consistent with a patient’s goals. Patients’ desire to have this discussion with their clinicians, with whom they have a relationship, supports the need to involve ambulatory care clinicians, or ensure these patients are cared for by the same clinicians, across healthcare settings.^13,14 Whoever provides the care, the clinician must align treatment with the patient’s goal, whether it is to continue hospital-level care or to transition to palliative care. Such an approach also reflects the core elements of person-centered care.¹⁵

Study Limitations

Participants were recruited from the medicine service at a single large academic center, limiting the study’s generalizability to patients admitted to surgical services or community hospitals. The patients in this small sample were English-speaking and predominantly Caucasian, so our findings may not represent the perspectives of non-English-speaking or minority patients. We did not perform statistical analysis to quantify intercoder reliability. Last, as this was a qualitative study, we cannot comment on the relative importance or prevalence of the reasons cited for frequent hospitalizations, and we cannot estimate the proportion of patients who had recurrent hospitalizations and were hospital-dependent.

Implication

Although quantitative research is needed to confirm our findings, the hospital-dependent patients in this study thought their survival required hospital-level care and resources. From their perspective, increasing posthospital and community support may be insufficient to prevent some hospitalizations. The lack of goals-of-care discussion supports attempts to increase efforts to facilitate discussion about health trajectories and health priorities between patients and their preferred clinicians.

Acknowledgments

The authors thank Dr. Grace Jenq for providing feedback on the study design.

Disclosure

Nothing to report.

In the United States, patients 65 years old or older accounted for more than one third of inpatient stays and 42% of inpatient care spending in 2012.¹ Despite the identification of risk factors, the implementation of an array of interventions, and the institution of penalties on hospitals, a subset of older adults continues to spend significant time in the hospital.^2,3

Hospital dependency is a concept that was only recently described. It identifies patients who improve while in the hospital but quickly deteriorate after leaving the hospital, resulting in recurring hospitalizations.⁴ Although little is known about hospital-dependent patients, studies have explored patients’ perspectives on readmissions.^5,6 Nevertheless, it remains unclear whether there are individuals for whom frequent and prolonged hospitalizations are appropriate, and whether there are undisclosed factors that, if addressed, could decrease their hospital dependency. We conducted an exploratory study to ascertain hospital-dependent patients’ perspectives on their needs for hospitalizations.

Study Design

This study was approved by the Yale University Institutional Review Board. From March 2015 to September 2015, Dr. Liu conducted semistructured explorative interviews with patients on the medical units of an academic medical center. Dr. Liu was not directly involved in the care of these patients. An interview guide that includes open-ended questions was created to elicit patients’ perspectives on their need for hospitalizations, health status, and outside-hospital support. This guide was pilot-tested with 6 patients, whose transcripts were not included in the final analysis, to assess for ease of understanding. After the pilot interviews, the questions were revised, and the final guide consists of 12 questions (Supplemental Table).

Recruitment

We used predetermined criteria and a purposeful sampling strategy to select potential study participants. We identified participants by periodically (~once a week) reviewing the electronic medical records of all patients admitted to the medicine service during the study period. Eligible patients were 65 years old or older and had at least 3 hospitalizations over the preceding 6 months. Patients were excluded if they met our chronic critical illness criteria: mechanical ventilation for more than 21 days, history of tracheotomy for failed weaning from mechanical ventilation,⁷ presence of a conservator, or admission only for comfort measures. Participants were recruited until no new themes emerged.

Data Collection

Twenty-nine patients were eligible. We obtained permission from their inpatient providers to approach them about the study. Of the 29 patients, 26 agreed to be interviewed, and 3 declined. Of the 26 participants, 6 underwent pilot interviews, and 20 underwent formal interviews with use of the finalized interview guide. The interviews, conducted in the hospital while the participants were hospitalized, lasted 17 minutes on average. The interviews were transcribed and iteratively analyzed. The themes that emerged from the initial interviews were further explored and validated in subsequent interviews. Interviews were conducted until theoretical saturation was reached and no new themes were derived from them. Demographic information, including age, sex, ethnicity, and marital status, was also collected.

Analysis

Interviews were digitally recorded and transcribed. Independently, two investigators used Atlas Ti software to analyze and code the interview transcriptions. An inductive approach was used to identify new codes from the data.⁸ The coders then met to discuss repeating ideas based on the codes. When a code was identified by one coder but not the other, or when there was disagreement about interpretation of a code, the coders returned to the relevant text to reach consensus and to determine whether to include or discard the code.⁹ We then organized and reorganized repeating ideas based on their conceptual similarities to determine the themes and subthemes.⁹

Twenty patients participated in the formal interviews. Participants’ baseline characteristics are listed in Table 1, and four dominant themes, and their subthemes and exemplary quotations, are listed in Table 2.

Perspectives on Hospital Care

Participants perceived their hospitalizations as inevitable and necessary for survival: “I think if I haven’t come to the hospital, I probably would have died.” Furthermore, participants thought only the hospital had the resources to help them (“The medications they were giving me … you can get that in the hospital but not outside the hospital”) and sustain them (“You are like an old car, and it breaks down little by little, so you have to go in periodically and get the problem fixed, so you will drive it around for a while”).

Feeling Safe in Hospital. Asked how being in the hospital makes them feel, participants attributed their feelings of safety to the constant observation, the availability of providers and nurses, and the idea that hospital care is helping. As one participant stated, “Makes me feel safer in case you go into something like cardiac arrest. You are right here where they can help you.”

Outside-Hospital Support. Despite multiple hospitalizations, most participants reported having social support (“I have the aide. I got the nurses come in. I have my daughter …”), physical support, and medical support (“I have all the doctors”) outside the hospital. A minority of participants questioned the usefulness of the services. One participant described declining the help of visiting nurses because she wanted to be independent and thought that, despite recurrent hospitalizations for physical symptoms, she still had the ability to manage her own medications.

Goals-of-Care Discussion. Some participants reported inadequate discussions about goals of care, health priorities, and health trajectories. In their reports, this inadequacy included not thinking about their goals, despite continued health decline. One participant stated, “Oh, God, I don’t know if I had any conversation like that. … I think until it is really brought to the front, you don’t make a decision really if you don’t have to.” Citing the value of a more established relationship and deeper trust, participants preferred having these serious and personal discussions with their ambulatory care clinicians: “Because I know my doctor much closer. I have been with him for a number of years. The doctors in the hospital seem to be nice and competent, but I don’t know them.”

Participants considered their hospitalizations a necessity and reported feeling safe in the hospital. Given that most already had support outside the hospital, increasing community services may be inadequate to alter participants’ perceived hospital care needs. On the other hand, a few participants reported declining services that might have prevented hospitalizations. Although there has been a study of treatment refusal among older adults with advanced illnesses,¹⁰ not much is known about refusal of services among this population. Investigators should examine the reasons for refusing services and the effect that refusal has on hospitalizations. Furthermore, although it would have been informative to ascertain clinician perspectives as well, we focused on patient perspectives because less is known on this topic.

Some participants noted their lack of discussion with their clinicians about healthcare goals and probable health trajectories. Barriers to goals-of-care discussion among this highly vulnerable population have been researched from the perspectives of clinicians and other health professionals but not patients themselves.^11,12 Of particular concern in our study is the participant-noted lack of discussion about health trajectories and health priorities, given the decline that occurs in this population and even in those with good care. This inadequacy in discussion suggests continued hospital care may not always be consistent with a patient’s goals. Patients’ desire to have this discussion with their clinicians, with whom they have a relationship, supports the need to involve ambulatory care clinicians, or ensure these patients are cared for by the same clinicians, across healthcare settings.^13,14 Whoever provides the care, the clinician must align treatment with the patient’s goal, whether it is to continue hospital-level care or to transition to palliative care. Such an approach also reflects the core elements of person-centered care.¹⁵

Study Limitations

Participants were recruited from the medicine service at a single large academic center, limiting the study’s generalizability to patients admitted to surgical services or community hospitals. The patients in this small sample were English-speaking and predominantly Caucasian, so our findings may not represent the perspectives of non-English-speaking or minority patients. We did not perform statistical analysis to quantify intercoder reliability. Last, as this was a qualitative study, we cannot comment on the relative importance or prevalence of the reasons cited for frequent hospitalizations, and we cannot estimate the proportion of patients who had recurrent hospitalizations and were hospital-dependent.

Implication

Although quantitative research is needed to confirm our findings, the hospital-dependent patients in this study thought their survival required hospital-level care and resources. From their perspective, increasing posthospital and community support may be insufficient to prevent some hospitalizations. The lack of goals-of-care discussion supports attempts to increase efforts to facilitate discussion about health trajectories and health priorities between patients and their preferred clinicians.

Acknowledgments

The authors thank Dr. Grace Jenq for providing feedback on the study design.

Disclosure

Nothing to report.

Incidental findings create both medical and logistical challenges regarding communication.^1,2 Pulmonary nodules are among the most frequent and medically relevant incidental findings, being noted in up to 8.4% of abdominal computed tomography (CT) scans.³ There are guidelines regarding proper follow-up and management of such incidental pulmonary nodules, but appropriate evidence-based surveillance imaging is often not performed, and many patients remain uninformed. Collins et al.⁴ reported that, before initiation of a standardized protocol, only 17.7% of incidental findings were communicated to patients admitted to the trauma service; after protocol initiation, the rate increased to 32.4%. The hospital discharge summary provides an opportunity to communicate incidental findings to patients and their medical care providers, but Kripalani et al.⁵ raised questions regarding the current completeness and accuracy of discharge summaries, reporting that 65% of discharge summaries omitted relevant diagnostic testing, and 30% omitted a follow-up plan.

We conducted a study to determine how often incidental pulmonary nodules found on abdominal CT are documented in the discharge summary, and to identify factors associated with pulmonary nodule inclusion.

This was a retrospective cohort study of hospitalized patients ≥35 years of age who underwent in-patient abdominal CT between January 1, 2012 and December 31, 2014. Patients were identified by cross-referencing hospital admissions with Current Procedural Terminology (CPT) codes indicating abdominal CT (74176, 74177, 74178, 74160, 74150, 74170). Patients with chest CT (CPT codes 71260, 71250, 71270) during that hospitalization or within 30 days before admission were excluded to ensure that pulmonary nodules were incidental and asymptomatic. The index hospitalization was defined as the first hospitalization during which the patient was diagnosed with an incidental pulmonary nodule on abdominal CT, or the first hospitalization during the study period for patients without pulmonary nodules. All patient charts were manually reviewed, and baseline age, sex, and smoking status data collected.

Radiology reports were electronically screened for the words nodule and nodules and then confirmed through manual review of the full text reports. Nodules described as tiny (without other size description) were assumed to be <4 mm in size, per manual review of a small sample. Nodules were deemed as falling outside the Fleischner Society criteria guidelines (designed for indeterminate pulmonary nodules), and were therefore excluded, if any of seven criteria were met: The nodule was (1) cavitary, (2) associated with a known metastatic disease, (3) associated with a known granulomatous disease, (4) associated with a known inflammatory process, (5) reported likely to represent atelectasis, (6) reported likely to be a lymph node, or (7) previously biopsied.⁴

For each patient with pulmonary nodules, a personal history of cancer was obtained. Nodule size, characteristics, and stability compared with available prior imaging were recorded. Radiology reports were reviewed to determine if pulmonary nodules were mentioned in the summary headings of the reports or in the body of the reports and whether specific follow-up recommendations were provided. Hospital discharge summaries were reviewed for documentation of pulmonary nodule(s) and follow-up recommendations. Discharging service (medical/medical subspecialty, surgical/surgical subspecialty) was noted, along with the patients’ condition at discharge (alive, alive on hospice, deceased).

The frequency of incidental pulmonary nodules on abdominal CT during hospitalization and the frequency of nodules requiring follow-up were reported using a point estimate and corresponding 95% confidence interval (CI). The χ² test was used to compare the frequency of pulmonary nodules across patient groups. In addition, for patients found to have incidental nodules requiring follow-up, the χ² test was used to compare across groups the percentage of patients with discharge documentation of the incidental nodule. In all cases, 2-tailed Ps are reported, with P ≤ 0.05 considered statistically significant.

Between January 1, 2012 and December 31, 2014, 7173 patients ≥35 years old underwent in-patient abdominal CT without concurrent chest CT. Of these patients, 62.2% were ≥60 years old, 50.6% were men, and 45.5% were current or former smokers. Incidental pulmonary nodules were noted in 402 patients (5.6%; 95% CI, 5.1%-6.2%), of whom 68.7% were ≥60 years old, 56.5% were men, and 46.3% were current or former smokers. Increasing age (P = 0.004) and male sex (P = 0.015) were associated with increased frequency of incidental pulmonary nodules, but smoking status (P = 0.586) was not. Of patients with incidental nodules, 71.6% had solitary nodules, and 58.5% had a maximum nodule size of ≤4 mm (Table 1). Based on smoking status, nodule size, and reported size stability, 208 patients (2.9%; 95% CI, 2.5%-3.3%) required follow-up surveillance as per 2005 Fleischner Society guidelines. Among solitary pulmonary nodules requiring further surveillance (n = 147), the mean risk of malignancy based on the Mayo Clinic solitary pulmonary nodule risk calculator was 7.9% (interquartile range, 3.0%-10.5%), with 28% having a malignancy risk of ≥10%.⁶

Of the 208 patients with nodules requiring further surveillance, only 48 (23%) received discharge summaries documenting the nodule; 34 of these summaries included a recommendation for nodule follow-up, with 19 of the recommendations including a time frame for repeat CT. Three factors were positively associated with documentation of the pulmonary nodule in the discharge summary: mention of the pulmonary nodule in the summary headings of the radiology report (P < 0.001), radiologist recommendation for further surveillance (P < 0.001), and medical discharging service (P = 0.016) (Table 2). The highest rate of pulmonary nodule inclusion in the discharge summary (42%) was noted among patients for whom the radiology report included specific recommendations.

The frequency of incidental pulmonary nodules reported on abdominal CT in our study (5.6%) is consistent with frequencies reported in similar studies. Wu et al.⁷ (reviewing 141,406 abdominal CT scans) and Alpert et al.⁸ (reviewing 12,287 abdominal CT scans) reported frequencies of 2.5% and 3%, respectively, while Rinaldi et al.³ (reviewing 243 abdominal CT scans) reported a higher frequency, 8.4%. Variation likely results from patient factors and the individual radiologist’s attention to incidental pulmonary findings. Rinaldi et al. suggested that up to 39% of abdominal CT scans include pulmonary nodules on independent review, raising the possibility of significant underreporting. In our study, we focused on pulmonary nodules included in the radiology report to tailor the relevance of our study to the hospital medicine community. We also included only those incidental nodules falling within the purview of the Fleischner Society criteria in order to analyze only findings with established follow-up guidelines.

The rate of pulmonary nodule documentation in our study was low overall (23%) but consistent with the literature. Collins et al.,⁴ for example, reported that only 17.7% of patients with trauma were notified of incidental CT findings by either the discharge summary or an appropriate specialist consultation. Various contributing factors can be hypothesized. First, incidental pulmonary nodules are discovered largely in the context of evaluation for other symptomatic conditions, which can overshadow their importance. Second, the lack of clear patient-friendly education materials regarding incidental pulmonary nodules can complicate discussions with patients. Third, many electronic health record (EHR) systems cannot automatically pull incidental findings into the discharge summary and instead rely on provider vigilance.

As our study does, the literature highlights the importance of the radiology report in communicating incidental findings. In a review of >1000 pulmonary angiographic CT studies, Blagev et al.⁹ reported an overall follow-up rate of 29% (28/96) among patients with incidental pulmonary nodules, but none of the 12 patients with pulmonary nodules mentioned in the body of the report (rather than in the summary headings) received adequate follow-up. Similarly, in Shuaib et al.,¹⁰ radiology reports that included follow-up recommendations were more likely to change patient treatment than reports without follow-up recommendations (70% vs 2%). However, our data also show that radiologist recommendations alone are insufficient to ensure adequate communication of incidental findings.

The literature regarding the most cost-effective means of addressing this quality gap is limited. Some institutions have integrated their EHR systems to allow radiologists to flag incidental findings for auto-population in a dedicated section of the discharge summary. Although these efforts can be helpful, documentation alone does not save lives without appropriate follow-up and intervention. Some institutions have hired dedicated nursing staff as incidental finding coordinators. For high-risk incidental findings, Sperry et al.¹¹ reported that hiring an incidental findings coordinator helped their level I trauma center achieve nearly complete documentation, patient notification, and confirmation of posthospital follow-up appointments. Such solutions, however, are labor-intensive and still rely on appropriate primary care follow-up.

Strengths of our study include its relatively large size and particular focus on the issues and decisions facing hospital medicine providers. By focusing on incidental pulmonary nodules reported on abdominal CT, and excluding patients with concurrent chest CT, we avoided including patients with symptomatic or previously identified pulmonary findings. Study limitations include the cross-sectional, retrospective design, which did not include follow-up data regarding such outcomes as rates of appropriate follow-up surveillance and subsequent lung cancer diagnoses. Our single-center study findings may not apply to all hospital practice settings, though they are consistent with the literature with comparison data.

Our study results highlight the need for a multidisciplinary systems-based approach to incidental pulmonary nodule documentation, communication, and follow-up surveillance.

Disclosure

Nothing to report.

Incidental findings create both medical and logistical challenges regarding communication.^1,2 Pulmonary nodules are among the most frequent and medically relevant incidental findings, being noted in up to 8.4% of abdominal computed tomography (CT) scans.³ There are guidelines regarding proper follow-up and management of such incidental pulmonary nodules, but appropriate evidence-based surveillance imaging is often not performed, and many patients remain uninformed. Collins et al.⁴ reported that, before initiation of a standardized protocol, only 17.7% of incidental findings were communicated to patients admitted to the trauma service; after protocol initiation, the rate increased to 32.4%. The hospital discharge summary provides an opportunity to communicate incidental findings to patients and their medical care providers, but Kripalani et al.⁵ raised questions regarding the current completeness and accuracy of discharge summaries, reporting that 65% of discharge summaries omitted relevant diagnostic testing, and 30% omitted a follow-up plan.

We conducted a study to determine how often incidental pulmonary nodules found on abdominal CT are documented in the discharge summary, and to identify factors associated with pulmonary nodule inclusion.

This was a retrospective cohort study of hospitalized patients ≥35 years of age who underwent in-patient abdominal CT between January 1, 2012 and December 31, 2014. Patients were identified by cross-referencing hospital admissions with Current Procedural Terminology (CPT) codes indicating abdominal CT (74176, 74177, 74178, 74160, 74150, 74170). Patients with chest CT (CPT codes 71260, 71250, 71270) during that hospitalization or within 30 days before admission were excluded to ensure that pulmonary nodules were incidental and asymptomatic. The index hospitalization was defined as the first hospitalization during which the patient was diagnosed with an incidental pulmonary nodule on abdominal CT, or the first hospitalization during the study period for patients without pulmonary nodules. All patient charts were manually reviewed, and baseline age, sex, and smoking status data collected.

Radiology reports were electronically screened for the words nodule and nodules and then confirmed through manual review of the full text reports. Nodules described as tiny (without other size description) were assumed to be <4 mm in size, per manual review of a small sample. Nodules were deemed as falling outside the Fleischner Society criteria guidelines (designed for indeterminate pulmonary nodules), and were therefore excluded, if any of seven criteria were met: The nodule was (1) cavitary, (2) associated with a known metastatic disease, (3) associated with a known granulomatous disease, (4) associated with a known inflammatory process, (5) reported likely to represent atelectasis, (6) reported likely to be a lymph node, or (7) previously biopsied.⁴

For each patient with pulmonary nodules, a personal history of cancer was obtained. Nodule size, characteristics, and stability compared with available prior imaging were recorded. Radiology reports were reviewed to determine if pulmonary nodules were mentioned in the summary headings of the reports or in the body of the reports and whether specific follow-up recommendations were provided. Hospital discharge summaries were reviewed for documentation of pulmonary nodule(s) and follow-up recommendations. Discharging service (medical/medical subspecialty, surgical/surgical subspecialty) was noted, along with the patients’ condition at discharge (alive, alive on hospice, deceased).

The frequency of incidental pulmonary nodules on abdominal CT during hospitalization and the frequency of nodules requiring follow-up were reported using a point estimate and corresponding 95% confidence interval (CI). The χ² test was used to compare the frequency of pulmonary nodules across patient groups. In addition, for patients found to have incidental nodules requiring follow-up, the χ² test was used to compare across groups the percentage of patients with discharge documentation of the incidental nodule. In all cases, 2-tailed Ps are reported, with P ≤ 0.05 considered statistically significant.

Between January 1, 2012 and December 31, 2014, 7173 patients ≥35 years old underwent in-patient abdominal CT without concurrent chest CT. Of these patients, 62.2% were ≥60 years old, 50.6% were men, and 45.5% were current or former smokers. Incidental pulmonary nodules were noted in 402 patients (5.6%; 95% CI, 5.1%-6.2%), of whom 68.7% were ≥60 years old, 56.5% were men, and 46.3% were current or former smokers. Increasing age (P = 0.004) and male sex (P = 0.015) were associated with increased frequency of incidental pulmonary nodules, but smoking status (P = 0.586) was not. Of patients with incidental nodules, 71.6% had solitary nodules, and 58.5% had a maximum nodule size of ≤4 mm (Table 1). Based on smoking status, nodule size, and reported size stability, 208 patients (2.9%; 95% CI, 2.5%-3.3%) required follow-up surveillance as per 2005 Fleischner Society guidelines. Among solitary pulmonary nodules requiring further surveillance (n = 147), the mean risk of malignancy based on the Mayo Clinic solitary pulmonary nodule risk calculator was 7.9% (interquartile range, 3.0%-10.5%), with 28% having a malignancy risk of ≥10%.⁶

Of the 208 patients with nodules requiring further surveillance, only 48 (23%) received discharge summaries documenting the nodule; 34 of these summaries included a recommendation for nodule follow-up, with 19 of the recommendations including a time frame for repeat CT. Three factors were positively associated with documentation of the pulmonary nodule in the discharge summary: mention of the pulmonary nodule in the summary headings of the radiology report (P < 0.001), radiologist recommendation for further surveillance (P < 0.001), and medical discharging service (P = 0.016) (Table 2). The highest rate of pulmonary nodule inclusion in the discharge summary (42%) was noted among patients for whom the radiology report included specific recommendations.

The frequency of incidental pulmonary nodules reported on abdominal CT in our study (5.6%) is consistent with frequencies reported in similar studies. Wu et al.⁷ (reviewing 141,406 abdominal CT scans) and Alpert et al.⁸ (reviewing 12,287 abdominal CT scans) reported frequencies of 2.5% and 3%, respectively, while Rinaldi et al.³ (reviewing 243 abdominal CT scans) reported a higher frequency, 8.4%. Variation likely results from patient factors and the individual radiologist’s attention to incidental pulmonary findings. Rinaldi et al. suggested that up to 39% of abdominal CT scans include pulmonary nodules on independent review, raising the possibility of significant underreporting. In our study, we focused on pulmonary nodules included in the radiology report to tailor the relevance of our study to the hospital medicine community. We also included only those incidental nodules falling within the purview of the Fleischner Society criteria in order to analyze only findings with established follow-up guidelines.

The rate of pulmonary nodule documentation in our study was low overall (23%) but consistent with the literature. Collins et al.,⁴ for example, reported that only 17.7% of patients with trauma were notified of incidental CT findings by either the discharge summary or an appropriate specialist consultation. Various contributing factors can be hypothesized. First, incidental pulmonary nodules are discovered largely in the context of evaluation for other symptomatic conditions, which can overshadow their importance. Second, the lack of clear patient-friendly education materials regarding incidental pulmonary nodules can complicate discussions with patients. Third, many electronic health record (EHR) systems cannot automatically pull incidental findings into the discharge summary and instead rely on provider vigilance.

As our study does, the literature highlights the importance of the radiology report in communicating incidental findings. In a review of >1000 pulmonary angiographic CT studies, Blagev et al.⁹ reported an overall follow-up rate of 29% (28/96) among patients with incidental pulmonary nodules, but none of the 12 patients with pulmonary nodules mentioned in the body of the report (rather than in the summary headings) received adequate follow-up. Similarly, in Shuaib et al.,¹⁰ radiology reports that included follow-up recommendations were more likely to change patient treatment than reports without follow-up recommendations (70% vs 2%). However, our data also show that radiologist recommendations alone are insufficient to ensure adequate communication of incidental findings.

The literature regarding the most cost-effective means of addressing this quality gap is limited. Some institutions have integrated their EHR systems to allow radiologists to flag incidental findings for auto-population in a dedicated section of the discharge summary. Although these efforts can be helpful, documentation alone does not save lives without appropriate follow-up and intervention. Some institutions have hired dedicated nursing staff as incidental finding coordinators. For high-risk incidental findings, Sperry et al.¹¹ reported that hiring an incidental findings coordinator helped their level I trauma center achieve nearly complete documentation, patient notification, and confirmation of posthospital follow-up appointments. Such solutions, however, are labor-intensive and still rely on appropriate primary care follow-up.

Strengths of our study include its relatively large size and particular focus on the issues and decisions facing hospital medicine providers. By focusing on incidental pulmonary nodules reported on abdominal CT, and excluding patients with concurrent chest CT, we avoided including patients with symptomatic or previously identified pulmonary findings. Study limitations include the cross-sectional, retrospective design, which did not include follow-up data regarding such outcomes as rates of appropriate follow-up surveillance and subsequent lung cancer diagnoses. Our single-center study findings may not apply to all hospital practice settings, though they are consistent with the literature with comparison data.

Our study results highlight the need for a multidisciplinary systems-based approach to incidental pulmonary nodule documentation, communication, and follow-up surveillance.

Disclosure

Nothing to report.

Incidental findings create both medical and logistical challenges regarding communication.^1,2 Pulmonary nodules are among the most frequent and medically relevant incidental findings, being noted in up to 8.4% of abdominal computed tomography (CT) scans.³ There are guidelines regarding proper follow-up and management of such incidental pulmonary nodules, but appropriate evidence-based surveillance imaging is often not performed, and many patients remain uninformed. Collins et al.⁴ reported that, before initiation of a standardized protocol, only 17.7% of incidental findings were communicated to patients admitted to the trauma service; after protocol initiation, the rate increased to 32.4%. The hospital discharge summary provides an opportunity to communicate incidental findings to patients and their medical care providers, but Kripalani et al.⁵ raised questions regarding the current completeness and accuracy of discharge summaries, reporting that 65% of discharge summaries omitted relevant diagnostic testing, and 30% omitted a follow-up plan.

We conducted a study to determine how often incidental pulmonary nodules found on abdominal CT are documented in the discharge summary, and to identify factors associated with pulmonary nodule inclusion.

This was a retrospective cohort study of hospitalized patients ≥35 years of age who underwent in-patient abdominal CT between January 1, 2012 and December 31, 2014. Patients were identified by cross-referencing hospital admissions with Current Procedural Terminology (CPT) codes indicating abdominal CT (74176, 74177, 74178, 74160, 74150, 74170). Patients with chest CT (CPT codes 71260, 71250, 71270) during that hospitalization or within 30 days before admission were excluded to ensure that pulmonary nodules were incidental and asymptomatic. The index hospitalization was defined as the first hospitalization during which the patient was diagnosed with an incidental pulmonary nodule on abdominal CT, or the first hospitalization during the study period for patients without pulmonary nodules. All patient charts were manually reviewed, and baseline age, sex, and smoking status data collected.

Radiology reports were electronically screened for the words nodule and nodules and then confirmed through manual review of the full text reports. Nodules described as tiny (without other size description) were assumed to be <4 mm in size, per manual review of a small sample. Nodules were deemed as falling outside the Fleischner Society criteria guidelines (designed for indeterminate pulmonary nodules), and were therefore excluded, if any of seven criteria were met: The nodule was (1) cavitary, (2) associated with a known metastatic disease, (3) associated with a known granulomatous disease, (4) associated with a known inflammatory process, (5) reported likely to represent atelectasis, (6) reported likely to be a lymph node, or (7) previously biopsied.⁴

For each patient with pulmonary nodules, a personal history of cancer was obtained. Nodule size, characteristics, and stability compared with available prior imaging were recorded. Radiology reports were reviewed to determine if pulmonary nodules were mentioned in the summary headings of the reports or in the body of the reports and whether specific follow-up recommendations were provided. Hospital discharge summaries were reviewed for documentation of pulmonary nodule(s) and follow-up recommendations. Discharging service (medical/medical subspecialty, surgical/surgical subspecialty) was noted, along with the patients’ condition at discharge (alive, alive on hospice, deceased).

The frequency of incidental pulmonary nodules on abdominal CT during hospitalization and the frequency of nodules requiring follow-up were reported using a point estimate and corresponding 95% confidence interval (CI). The χ² test was used to compare the frequency of pulmonary nodules across patient groups. In addition, for patients found to have incidental nodules requiring follow-up, the χ² test was used to compare across groups the percentage of patients with discharge documentation of the incidental nodule. In all cases, 2-tailed Ps are reported, with P ≤ 0.05 considered statistically significant.

Between January 1, 2012 and December 31, 2014, 7173 patients ≥35 years old underwent in-patient abdominal CT without concurrent chest CT. Of these patients, 62.2% were ≥60 years old, 50.6% were men, and 45.5% were current or former smokers. Incidental pulmonary nodules were noted in 402 patients (5.6%; 95% CI, 5.1%-6.2%), of whom 68.7% were ≥60 years old, 56.5% were men, and 46.3% were current or former smokers. Increasing age (P = 0.004) and male sex (P = 0.015) were associated with increased frequency of incidental pulmonary nodules, but smoking status (P = 0.586) was not. Of patients with incidental nodules, 71.6% had solitary nodules, and 58.5% had a maximum nodule size of ≤4 mm (Table 1). Based on smoking status, nodule size, and reported size stability, 208 patients (2.9%; 95% CI, 2.5%-3.3%) required follow-up surveillance as per 2005 Fleischner Society guidelines. Among solitary pulmonary nodules requiring further surveillance (n = 147), the mean risk of malignancy based on the Mayo Clinic solitary pulmonary nodule risk calculator was 7.9% (interquartile range, 3.0%-10.5%), with 28% having a malignancy risk of ≥10%.⁶

Of the 208 patients with nodules requiring further surveillance, only 48 (23%) received discharge summaries documenting the nodule; 34 of these summaries included a recommendation for nodule follow-up, with 19 of the recommendations including a time frame for repeat CT. Three factors were positively associated with documentation of the pulmonary nodule in the discharge summary: mention of the pulmonary nodule in the summary headings of the radiology report (P < 0.001), radiologist recommendation for further surveillance (P < 0.001), and medical discharging service (P = 0.016) (Table 2). The highest rate of pulmonary nodule inclusion in the discharge summary (42%) was noted among patients for whom the radiology report included specific recommendations.

The frequency of incidental pulmonary nodules reported on abdominal CT in our study (5.6%) is consistent with frequencies reported in similar studies. Wu et al.⁷ (reviewing 141,406 abdominal CT scans) and Alpert et al.⁸ (reviewing 12,287 abdominal CT scans) reported frequencies of 2.5% and 3%, respectively, while Rinaldi et al.³ (reviewing 243 abdominal CT scans) reported a higher frequency, 8.4%. Variation likely results from patient factors and the individual radiologist’s attention to incidental pulmonary findings. Rinaldi et al. suggested that up to 39% of abdominal CT scans include pulmonary nodules on independent review, raising the possibility of significant underreporting. In our study, we focused on pulmonary nodules included in the radiology report to tailor the relevance of our study to the hospital medicine community. We also included only those incidental nodules falling within the purview of the Fleischner Society criteria in order to analyze only findings with established follow-up guidelines.

The rate of pulmonary nodule documentation in our study was low overall (23%) but consistent with the literature. Collins et al.,⁴ for example, reported that only 17.7% of patients with trauma were notified of incidental CT findings by either the discharge summary or an appropriate specialist consultation. Various contributing factors can be hypothesized. First, incidental pulmonary nodules are discovered largely in the context of evaluation for other symptomatic conditions, which can overshadow their importance. Second, the lack of clear patient-friendly education materials regarding incidental pulmonary nodules can complicate discussions with patients. Third, many electronic health record (EHR) systems cannot automatically pull incidental findings into the discharge summary and instead rely on provider vigilance.

As our study does, the literature highlights the importance of the radiology report in communicating incidental findings. In a review of >1000 pulmonary angiographic CT studies, Blagev et al.⁹ reported an overall follow-up rate of 29% (28/96) among patients with incidental pulmonary nodules, but none of the 12 patients with pulmonary nodules mentioned in the body of the report (rather than in the summary headings) received adequate follow-up. Similarly, in Shuaib et al.,¹⁰ radiology reports that included follow-up recommendations were more likely to change patient treatment than reports without follow-up recommendations (70% vs 2%). However, our data also show that radiologist recommendations alone are insufficient to ensure adequate communication of incidental findings.

The literature regarding the most cost-effective means of addressing this quality gap is limited. Some institutions have integrated their EHR systems to allow radiologists to flag incidental findings for auto-population in a dedicated section of the discharge summary. Although these efforts can be helpful, documentation alone does not save lives without appropriate follow-up and intervention. Some institutions have hired dedicated nursing staff as incidental finding coordinators. For high-risk incidental findings, Sperry et al.¹¹ reported that hiring an incidental findings coordinator helped their level I trauma center achieve nearly complete documentation, patient notification, and confirmation of posthospital follow-up appointments. Such solutions, however, are labor-intensive and still rely on appropriate primary care follow-up.

Strengths of our study include its relatively large size and particular focus on the issues and decisions facing hospital medicine providers. By focusing on incidental pulmonary nodules reported on abdominal CT, and excluding patients with concurrent chest CT, we avoided including patients with symptomatic or previously identified pulmonary findings. Study limitations include the cross-sectional, retrospective design, which did not include follow-up data regarding such outcomes as rates of appropriate follow-up surveillance and subsequent lung cancer diagnoses. Our single-center study findings may not apply to all hospital practice settings, though they are consistent with the literature with comparison data.

Our study results highlight the need for a multidisciplinary systems-based approach to incidental pulmonary nodule documentation, communication, and follow-up surveillance.

Disclosure

Nothing to report.

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

Evaluation and treatment of the febrile infant 0 to 90 days of age are common clinical issues in pediatrics, family medicine, emergency medicine, and pediatric hospital medicine. Traditional teaching has been that Listeria monocytogenes is 1 of the 3 most common pathogens causing neonatal sepsis. Many practitioners routinely use antibiotic regimens, including ampicillin, to specifically target Listeria. However, a large body of evidence, including a meta-analysis and several multicenter studies, has shown that listeriosis is extremely rare in the United States. The practice of empiric ampicillin thus exposes the patient to harms and costs with little if any potential benefit, while increasing pressure on the bacterial flora in the community to generate antibiotic resistance. Empiric ampicillin for all infants admitted for sepsis evaluation is a tradition-based practice no longer founded on the best available evidence.

A 32-day-old, full-term, previously healthy girl presented with fever of 1 day’s duration. Her parents reported she had appeared well until the evening before admission, when she became a bit less active and spent less time breastfeeding. The morning of admission, she was fussier than usual. Rectal temperature, taken by her parents, was 101°F. There were no other symptoms and no sick contacts.

On examination, the patient’s rectal temperature was 101.5°F. Her other vitals and the physical examination findings were unremarkable. Laboratory test results included a normal urinalysis and a peripheral white blood cell (WBC) count of 21,300 cells/µL. Cerebrospinal fluid (CSF) analysis revealed normal protein and glucose levels with 3 WBCs/µL and a negative gram stain. Due to stratifying at higher risk for serious bacterial infection (SBI), the child was admitted and started on ampicillin and cefotaxime while awaiting culture results.

Evaluation and treatment of febrile infants are common clinical issues in pediatrics, emergency medicine, and general practice. Practice guidelines for evaluation of febrile infants recommend hospitalization and parenteral antibiotics for children younger than 28 days and children 29 to 90 days old if stratified at high risk for SBI.^1,2 Recommendations for empiric antibiotic regimens include ampicillin in addition to either gentamicin or cefotaxime.^1,2

WHY YOU MIGHT THINK AMPICILLIN IS HELPFUL

Generations of pediatrics students have been taught that the 3 pathogens most likely to cause bacterial sepsis in infants are group B Streptococcus (GBS), Escherichia coli, and Listeria monocytogenes. This teaching is still espoused in the latest editions of pediatrics textbooks.³ Ampicillin is specifically recommended for covering Listeria, and studies have found that 62% to 78% of practitioners choose empiric ampicillin-containing antibiotic regimens for the treatment of febrile infants.^4-6

WHY EMPIRIC AMPICILLIN IS UNNECESSARY

In the past, Listeria was a potential though still uncommon infant pathogen. Over the past few decades, however, the epidemiology of infant sepsis has changed significantly. Estimates of the rate of infection with Listeria now range from extremely rare to nonexistent across multiple studies^4,7-15 (Table). In a 4-year retrospective case series at a single urban academic center in Washington, DC, Sadow et al.⁴ reported no instances of Listeria among 121 positive bacterial cultures in infants younger than 60 days seen in the emergency department (ED). Byington et al.⁷ examined all positive cultures for infants 0 to 90 days old at a large academic referral center in Utah over a 3-year period and reported no cases of Listeria (1298 patients, 105 SBI cases). A study at a North Carolina academic center found 1 case of Listeria meningitis among 72 SBIs (668 febrile infants) without a localizing source.⁸ At a large group-practice in northern California, Greenhow et al.⁹ examined all blood cultures (N = 4255) performed over 4 years for otherwise healthy infants 1 week to 3 months old and found no cases of Listeria. In a follow-up study, the same authors examined all blood (n = 5396), urine (n = 4599), and CSF (n = 1796) cultures in the same population and found no Listeria cases.¹⁰ Hassoun et al.¹¹ studied SBI rates among infants younger than 28 days with any blood, urine, or CSF culture performed over 4 years at two Michigan EDs. One (0.08%) of the 1192 infants evaluated had bacteremia caused by Listeria.

Multicenter studies have reported similar results. In a study of 6 hospital systems in geographically diverse areas of the United States, Biondi et al.¹² examined all positive blood cultures (N = 181) for febrile infants younger than 90 days admitted to a general pediatric ward, and found no listeriosis. Mischler et al.¹³ examined all positive blood cultures (N = 392) for otherwise healthy febrile infants 0 to 90 days old admitted to a hospital in 1 of 17 geographically diverse healthcare systems and found no cases of Listeria. A recent meta-analysis of studies that reported SBI rates for febrile infants 0 to 90 days old found the weighted prevalence of Listeria bacteremia to be 0.03% (2/20,703) and that of meningitis to be 0.02% (3/13,375).¹⁴ Veesenmeyer and Edmonson¹⁵ used a national inpatient database to identify all Listeria cases among infants over a 6-year period and estimated listeriosis rates for the US population. Over the 6 years, there were 212 total cases, which were extrapolated to 344 in the United States during that period, yielding a pooled annual incidence rate of 1.41 in 100,000 births.

Ampicillin offers no significant improvement in coverage for GBS or E coli beyond other β-lactam antibiotics, such as cefotaxime. Therefore, though the cost and potential harms of 24 to 48 hours of intravenous ampicillin are low for the individual patient, there is almost no potential benefit. Using the weighted prevalence of 0.03% for Listeria bacteremia reported in the recent meta-analysis,¹⁴ the number needed to treat to cover 1 case of Listeria bacteremia would be 3333. In addition, the increasing incidence of ampicillin resistance, particularly among gram-negative organisms,^4,7,9 argues strongly for better antibiotic stewardship on a national level. A number of expert authors have advocated dropping empiric Listeria coverage as part of the treatment of febrile infants, particularly infants 29 to 90 days old.^16,17 Some authors continue to advocate empiric Listeria coverage.⁶ It is interesting to note, however, that the incidence of Staph aureus bacteremia in recent case series is much higher than that reported for Listeria, accounting for 6-9% of bacteremia cases.^9,11,13 Yet few if any authors advocate for empiric S. aureus coverage.

WHEN EMPIRIC AMPICILLIN COVERAGE MAY BE REASONABLE

The rate of listeriosis remains low across age groups in recent studies, but the rate is slightly higher in very young infants. In the recent national database study of listeriosis cases over a 6-year period, almost half involved infants younger than 7 days, and most of these infants showed no evidence of meningitis.¹⁵ Therefore, it may be reasonable to include empiric Listeria coverage in febrile infants younger than 7 days, though the study authors estimated 22.5 annual cases of Listeria in this age range in the United States. Eighty percent of the Listeria cases were in infants younger than 28 days, but more than 85% of infants 7 to 28 days old had meningitis. Therefore, broad antimicrobial coverage for infants with CSF pleocytosis and/or a high bacterial meningitis score is reasonable, especially for infants younger than 28 days.

Other potential indications for ampicillin are enterococcal infections. Though enteroccocal SBI rates in febrile infants are also quite low,^7-9,11,12 if Enterococcus were highly suspected, such as in an infant with pyuria and gram positive organisms on gram stain, ampicillin offers good additional coverage. In the case of a local outbreak of listeriosis, or a specific exposure to Listeria-contaminated products on a patient history, antibiotics with efficacy against Listeria should be used. Last, in cases in which gentamicin is used as empiric coverage for gram-negative organisms, ampicillin offers important additional coverage for GBS.

Some practitioners advocate ampicillin and gentamicin over cefotaxime regimens on the basis of an often cited study that found a survival benefit for febrile neonates in the intensive care setting.¹⁸ There are a number of reasons that this study should not influence care for typical infants admitted with possible sepsis. First, the study was retrospective and limited by its use of administrative data. The authors acknowledged that a potential explanation for their results is unmeasured confounding. Second, the patients included in the study were dramatically different from the group of well infants admitted with possible sepsis; the study included neonatal critical care unit patients treated with antibiotics within the first 3 days of life. Third, the study’s results have not been replicated in otherwise healthy febrile infants.

WHAT YOU SHOULD USE INSTEAD OF AMPICILLIN FOR EMPIRIC LISTERIA COVERAGE

For febrile children 0 to 90 days old, empiric antibiotic coverage should be aimed at covering the current predominant pathogens, which include E coli and GBS. Therefore, for most children and US regions, a third-generation cephalosporin (eg, cefotaxime) is sufficient.

• Empiric antibiotics for treatment of febrile children 0-90 days should target E. coli and GBS; a third generation cephalosporin, (e.g. cefotaxime) alone is a reasonable choice for most patients.
• Prescribing ampicillin to specifically cover Listeria is unnecessary for the vast majority of febrile infants
• Prescribing ampicillin is reasonable in certain subgroups of febrile infants: those less than seven days of age, those with evidence of bacterial meningitis (especially if also <28 days of age), those in whom enterococcal infection is strongly suspected, and those with specific Listeria exposures related to local outbreaks.

The 32-day-old infant described in the clinical scenario was at extremely low risk for listeriosis. Antibiotic coverage with a third-generation cephalosporin is sufficient for the most likely pathogens. The common practice of empirically covering Listeria in otherwise healthy febrile infants considered to be at higher risk for SBI is no longer based on best available evidence and represents overtreatment with at least theoretical harms. Avoidance of the risks associated with the side effects of antibiotics, costs saved by forgoing multiple antibiotics, a decrease in medication dosing frequency, and improved antibiotic stewardship for the general population all argue forcefully for making empiric Listeria coverage a thing of the past.

Disclosure

Nothing to report.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and Liking It on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

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

Evaluation and treatment of the febrile infant 0 to 90 days of age are common clinical issues in pediatrics, family medicine, emergency medicine, and pediatric hospital medicine. Traditional teaching has been that Listeria monocytogenes is 1 of the 3 most common pathogens causing neonatal sepsis. Many practitioners routinely use antibiotic regimens, including ampicillin, to specifically target Listeria. However, a large body of evidence, including a meta-analysis and several multicenter studies, has shown that listeriosis is extremely rare in the United States. The practice of empiric ampicillin thus exposes the patient to harms and costs with little if any potential benefit, while increasing pressure on the bacterial flora in the community to generate antibiotic resistance. Empiric ampicillin for all infants admitted for sepsis evaluation is a tradition-based practice no longer founded on the best available evidence.

A 32-day-old, full-term, previously healthy girl presented with fever of 1 day’s duration. Her parents reported she had appeared well until the evening before admission, when she became a bit less active and spent less time breastfeeding. The morning of admission, she was fussier than usual. Rectal temperature, taken by her parents, was 101°F. There were no other symptoms and no sick contacts.

On examination, the patient’s rectal temperature was 101.5°F. Her other vitals and the physical examination findings were unremarkable. Laboratory test results included a normal urinalysis and a peripheral white blood cell (WBC) count of 21,300 cells/µL. Cerebrospinal fluid (CSF) analysis revealed normal protein and glucose levels with 3 WBCs/µL and a negative gram stain. Due to stratifying at higher risk for serious bacterial infection (SBI), the child was admitted and started on ampicillin and cefotaxime while awaiting culture results.

Evaluation and treatment of febrile infants are common clinical issues in pediatrics, emergency medicine, and general practice. Practice guidelines for evaluation of febrile infants recommend hospitalization and parenteral antibiotics for children younger than 28 days and children 29 to 90 days old if stratified at high risk for SBI.^1,2 Recommendations for empiric antibiotic regimens include ampicillin in addition to either gentamicin or cefotaxime.^1,2

WHY YOU MIGHT THINK AMPICILLIN IS HELPFUL

Generations of pediatrics students have been taught that the 3 pathogens most likely to cause bacterial sepsis in infants are group B Streptococcus (GBS), Escherichia coli, and Listeria monocytogenes. This teaching is still espoused in the latest editions of pediatrics textbooks.³ Ampicillin is specifically recommended for covering Listeria, and studies have found that 62% to 78% of practitioners choose empiric ampicillin-containing antibiotic regimens for the treatment of febrile infants.^4-6

WHY EMPIRIC AMPICILLIN IS UNNECESSARY

In the past, Listeria was a potential though still uncommon infant pathogen. Over the past few decades, however, the epidemiology of infant sepsis has changed significantly. Estimates of the rate of infection with Listeria now range from extremely rare to nonexistent across multiple studies^4,7-15 (Table). In a 4-year retrospective case series at a single urban academic center in Washington, DC, Sadow et al.⁴ reported no instances of Listeria among 121 positive bacterial cultures in infants younger than 60 days seen in the emergency department (ED). Byington et al.⁷ examined all positive cultures for infants 0 to 90 days old at a large academic referral center in Utah over a 3-year period and reported no cases of Listeria (1298 patients, 105 SBI cases). A study at a North Carolina academic center found 1 case of Listeria meningitis among 72 SBIs (668 febrile infants) without a localizing source.⁸ At a large group-practice in northern California, Greenhow et al.⁹ examined all blood cultures (N = 4255) performed over 4 years for otherwise healthy infants 1 week to 3 months old and found no cases of Listeria. In a follow-up study, the same authors examined all blood (n = 5396), urine (n = 4599), and CSF (n = 1796) cultures in the same population and found no Listeria cases.¹⁰ Hassoun et al.¹¹ studied SBI rates among infants younger than 28 days with any blood, urine, or CSF culture performed over 4 years at two Michigan EDs. One (0.08%) of the 1192 infants evaluated had bacteremia caused by Listeria.

Multicenter studies have reported similar results. In a study of 6 hospital systems in geographically diverse areas of the United States, Biondi et al.¹² examined all positive blood cultures (N = 181) for febrile infants younger than 90 days admitted to a general pediatric ward, and found no listeriosis. Mischler et al.¹³ examined all positive blood cultures (N = 392) for otherwise healthy febrile infants 0 to 90 days old admitted to a hospital in 1 of 17 geographically diverse healthcare systems and found no cases of Listeria. A recent meta-analysis of studies that reported SBI rates for febrile infants 0 to 90 days old found the weighted prevalence of Listeria bacteremia to be 0.03% (2/20,703) and that of meningitis to be 0.02% (3/13,375).¹⁴ Veesenmeyer and Edmonson¹⁵ used a national inpatient database to identify all Listeria cases among infants over a 6-year period and estimated listeriosis rates for the US population. Over the 6 years, there were 212 total cases, which were extrapolated to 344 in the United States during that period, yielding a pooled annual incidence rate of 1.41 in 100,000 births.

Ampicillin offers no significant improvement in coverage for GBS or E coli beyond other β-lactam antibiotics, such as cefotaxime. Therefore, though the cost and potential harms of 24 to 48 hours of intravenous ampicillin are low for the individual patient, there is almost no potential benefit. Using the weighted prevalence of 0.03% for Listeria bacteremia reported in the recent meta-analysis,¹⁴ the number needed to treat to cover 1 case of Listeria bacteremia would be 3333. In addition, the increasing incidence of ampicillin resistance, particularly among gram-negative organisms,^4,7,9 argues strongly for better antibiotic stewardship on a national level. A number of expert authors have advocated dropping empiric Listeria coverage as part of the treatment of febrile infants, particularly infants 29 to 90 days old.^16,17 Some authors continue to advocate empiric Listeria coverage.⁶ It is interesting to note, however, that the incidence of Staph aureus bacteremia in recent case series is much higher than that reported for Listeria, accounting for 6-9% of bacteremia cases.^9,11,13 Yet few if any authors advocate for empiric S. aureus coverage.

WHEN EMPIRIC AMPICILLIN COVERAGE MAY BE REASONABLE

The rate of listeriosis remains low across age groups in recent studies, but the rate is slightly higher in very young infants. In the recent national database study of listeriosis cases over a 6-year period, almost half involved infants younger than 7 days, and most of these infants showed no evidence of meningitis.¹⁵ Therefore, it may be reasonable to include empiric Listeria coverage in febrile infants younger than 7 days, though the study authors estimated 22.5 annual cases of Listeria in this age range in the United States. Eighty percent of the Listeria cases were in infants younger than 28 days, but more than 85% of infants 7 to 28 days old had meningitis. Therefore, broad antimicrobial coverage for infants with CSF pleocytosis and/or a high bacterial meningitis score is reasonable, especially for infants younger than 28 days.

Other potential indications for ampicillin are enterococcal infections. Though enteroccocal SBI rates in febrile infants are also quite low,^7-9,11,12 if Enterococcus were highly suspected, such as in an infant with pyuria and gram positive organisms on gram stain, ampicillin offers good additional coverage. In the case of a local outbreak of listeriosis, or a specific exposure to Listeria-contaminated products on a patient history, antibiotics with efficacy against Listeria should be used. Last, in cases in which gentamicin is used as empiric coverage for gram-negative organisms, ampicillin offers important additional coverage for GBS.

Some practitioners advocate ampicillin and gentamicin over cefotaxime regimens on the basis of an often cited study that found a survival benefit for febrile neonates in the intensive care setting.¹⁸ There are a number of reasons that this study should not influence care for typical infants admitted with possible sepsis. First, the study was retrospective and limited by its use of administrative data. The authors acknowledged that a potential explanation for their results is unmeasured confounding. Second, the patients included in the study were dramatically different from the group of well infants admitted with possible sepsis; the study included neonatal critical care unit patients treated with antibiotics within the first 3 days of life. Third, the study’s results have not been replicated in otherwise healthy febrile infants.

WHAT YOU SHOULD USE INSTEAD OF AMPICILLIN FOR EMPIRIC LISTERIA COVERAGE

For febrile children 0 to 90 days old, empiric antibiotic coverage should be aimed at covering the current predominant pathogens, which include E coli and GBS. Therefore, for most children and US regions, a third-generation cephalosporin (eg, cefotaxime) is sufficient.

• Empiric antibiotics for treatment of febrile children 0-90 days should target E. coli and GBS; a third generation cephalosporin, (e.g. cefotaxime) alone is a reasonable choice for most patients.
• Prescribing ampicillin to specifically cover Listeria is unnecessary for the vast majority of febrile infants
• Prescribing ampicillin is reasonable in certain subgroups of febrile infants: those less than seven days of age, those with evidence of bacterial meningitis (especially if also <28 days of age), those in whom enterococcal infection is strongly suspected, and those with specific Listeria exposures related to local outbreaks.

The 32-day-old infant described in the clinical scenario was at extremely low risk for listeriosis. Antibiotic coverage with a third-generation cephalosporin is sufficient for the most likely pathogens. The common practice of empirically covering Listeria in otherwise healthy febrile infants considered to be at higher risk for SBI is no longer based on best available evidence and represents overtreatment with at least theoretical harms. Avoidance of the risks associated with the side effects of antibiotics, costs saved by forgoing multiple antibiotics, a decrease in medication dosing frequency, and improved antibiotic stewardship for the general population all argue forcefully for making empiric Listeria coverage a thing of the past.

Disclosure

Nothing to report.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and Liking It on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

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

Evaluation and treatment of the febrile infant 0 to 90 days of age are common clinical issues in pediatrics, family medicine, emergency medicine, and pediatric hospital medicine. Traditional teaching has been that Listeria monocytogenes is 1 of the 3 most common pathogens causing neonatal sepsis. Many practitioners routinely use antibiotic regimens, including ampicillin, to specifically target Listeria. However, a large body of evidence, including a meta-analysis and several multicenter studies, has shown that listeriosis is extremely rare in the United States. The practice of empiric ampicillin thus exposes the patient to harms and costs with little if any potential benefit, while increasing pressure on the bacterial flora in the community to generate antibiotic resistance. Empiric ampicillin for all infants admitted for sepsis evaluation is a tradition-based practice no longer founded on the best available evidence.

A 32-day-old, full-term, previously healthy girl presented with fever of 1 day’s duration. Her parents reported she had appeared well until the evening before admission, when she became a bit less active and spent less time breastfeeding. The morning of admission, she was fussier than usual. Rectal temperature, taken by her parents, was 101°F. There were no other symptoms and no sick contacts.

On examination, the patient’s rectal temperature was 101.5°F. Her other vitals and the physical examination findings were unremarkable. Laboratory test results included a normal urinalysis and a peripheral white blood cell (WBC) count of 21,300 cells/µL. Cerebrospinal fluid (CSF) analysis revealed normal protein and glucose levels with 3 WBCs/µL and a negative gram stain. Due to stratifying at higher risk for serious bacterial infection (SBI), the child was admitted and started on ampicillin and cefotaxime while awaiting culture results.

Evaluation and treatment of febrile infants are common clinical issues in pediatrics, emergency medicine, and general practice. Practice guidelines for evaluation of febrile infants recommend hospitalization and parenteral antibiotics for children younger than 28 days and children 29 to 90 days old if stratified at high risk for SBI.^1,2 Recommendations for empiric antibiotic regimens include ampicillin in addition to either gentamicin or cefotaxime.^1,2

WHY YOU MIGHT THINK AMPICILLIN IS HELPFUL

Generations of pediatrics students have been taught that the 3 pathogens most likely to cause bacterial sepsis in infants are group B Streptococcus (GBS), Escherichia coli, and Listeria monocytogenes. This teaching is still espoused in the latest editions of pediatrics textbooks.³ Ampicillin is specifically recommended for covering Listeria, and studies have found that 62% to 78% of practitioners choose empiric ampicillin-containing antibiotic regimens for the treatment of febrile infants.^4-6

WHY EMPIRIC AMPICILLIN IS UNNECESSARY

In the past, Listeria was a potential though still uncommon infant pathogen. Over the past few decades, however, the epidemiology of infant sepsis has changed significantly. Estimates of the rate of infection with Listeria now range from extremely rare to nonexistent across multiple studies^4,7-15 (Table). In a 4-year retrospective case series at a single urban academic center in Washington, DC, Sadow et al.⁴ reported no instances of Listeria among 121 positive bacterial cultures in infants younger than 60 days seen in the emergency department (ED). Byington et al.⁷ examined all positive cultures for infants 0 to 90 days old at a large academic referral center in Utah over a 3-year period and reported no cases of Listeria (1298 patients, 105 SBI cases). A study at a North Carolina academic center found 1 case of Listeria meningitis among 72 SBIs (668 febrile infants) without a localizing source.⁸ At a large group-practice in northern California, Greenhow et al.⁹ examined all blood cultures (N = 4255) performed over 4 years for otherwise healthy infants 1 week to 3 months old and found no cases of Listeria. In a follow-up study, the same authors examined all blood (n = 5396), urine (n = 4599), and CSF (n = 1796) cultures in the same population and found no Listeria cases.¹⁰ Hassoun et al.¹¹ studied SBI rates among infants younger than 28 days with any blood, urine, or CSF culture performed over 4 years at two Michigan EDs. One (0.08%) of the 1192 infants evaluated had bacteremia caused by Listeria.

Multicenter studies have reported similar results. In a study of 6 hospital systems in geographically diverse areas of the United States, Biondi et al.¹² examined all positive blood cultures (N = 181) for febrile infants younger than 90 days admitted to a general pediatric ward, and found no listeriosis. Mischler et al.¹³ examined all positive blood cultures (N = 392) for otherwise healthy febrile infants 0 to 90 days old admitted to a hospital in 1 of 17 geographically diverse healthcare systems and found no cases of Listeria. A recent meta-analysis of studies that reported SBI rates for febrile infants 0 to 90 days old found the weighted prevalence of Listeria bacteremia to be 0.03% (2/20,703) and that of meningitis to be 0.02% (3/13,375).¹⁴ Veesenmeyer and Edmonson¹⁵ used a national inpatient database to identify all Listeria cases among infants over a 6-year period and estimated listeriosis rates for the US population. Over the 6 years, there were 212 total cases, which were extrapolated to 344 in the United States during that period, yielding a pooled annual incidence rate of 1.41 in 100,000 births.

Ampicillin offers no significant improvement in coverage for GBS or E coli beyond other β-lactam antibiotics, such as cefotaxime. Therefore, though the cost and potential harms of 24 to 48 hours of intravenous ampicillin are low for the individual patient, there is almost no potential benefit. Using the weighted prevalence of 0.03% for Listeria bacteremia reported in the recent meta-analysis,¹⁴ the number needed to treat to cover 1 case of Listeria bacteremia would be 3333. In addition, the increasing incidence of ampicillin resistance, particularly among gram-negative organisms,^4,7,9 argues strongly for better antibiotic stewardship on a national level. A number of expert authors have advocated dropping empiric Listeria coverage as part of the treatment of febrile infants, particularly infants 29 to 90 days old.^16,17 Some authors continue to advocate empiric Listeria coverage.⁶ It is interesting to note, however, that the incidence of Staph aureus bacteremia in recent case series is much higher than that reported for Listeria, accounting for 6-9% of bacteremia cases.^9,11,13 Yet few if any authors advocate for empiric S. aureus coverage.

WHEN EMPIRIC AMPICILLIN COVERAGE MAY BE REASONABLE

The rate of listeriosis remains low across age groups in recent studies, but the rate is slightly higher in very young infants. In the recent national database study of listeriosis cases over a 6-year period, almost half involved infants younger than 7 days, and most of these infants showed no evidence of meningitis.¹⁵ Therefore, it may be reasonable to include empiric Listeria coverage in febrile infants younger than 7 days, though the study authors estimated 22.5 annual cases of Listeria in this age range in the United States. Eighty percent of the Listeria cases were in infants younger than 28 days, but more than 85% of infants 7 to 28 days old had meningitis. Therefore, broad antimicrobial coverage for infants with CSF pleocytosis and/or a high bacterial meningitis score is reasonable, especially for infants younger than 28 days.

Other potential indications for ampicillin are enterococcal infections. Though enteroccocal SBI rates in febrile infants are also quite low,^7-9,11,12 if Enterococcus were highly suspected, such as in an infant with pyuria and gram positive organisms on gram stain, ampicillin offers good additional coverage. In the case of a local outbreak of listeriosis, or a specific exposure to Listeria-contaminated products on a patient history, antibiotics with efficacy against Listeria should be used. Last, in cases in which gentamicin is used as empiric coverage for gram-negative organisms, ampicillin offers important additional coverage for GBS.

Some practitioners advocate ampicillin and gentamicin over cefotaxime regimens on the basis of an often cited study that found a survival benefit for febrile neonates in the intensive care setting.¹⁸ There are a number of reasons that this study should not influence care for typical infants admitted with possible sepsis. First, the study was retrospective and limited by its use of administrative data. The authors acknowledged that a potential explanation for their results is unmeasured confounding. Second, the patients included in the study were dramatically different from the group of well infants admitted with possible sepsis; the study included neonatal critical care unit patients treated with antibiotics within the first 3 days of life. Third, the study’s results have not been replicated in otherwise healthy febrile infants.

WHAT YOU SHOULD USE INSTEAD OF AMPICILLIN FOR EMPIRIC LISTERIA COVERAGE

For febrile children 0 to 90 days old, empiric antibiotic coverage should be aimed at covering the current predominant pathogens, which include E coli and GBS. Therefore, for most children and US regions, a third-generation cephalosporin (eg, cefotaxime) is sufficient.

• Empiric antibiotics for treatment of febrile children 0-90 days should target E. coli and GBS; a third generation cephalosporin, (e.g. cefotaxime) alone is a reasonable choice for most patients.
• Prescribing ampicillin to specifically cover Listeria is unnecessary for the vast majority of febrile infants
• Prescribing ampicillin is reasonable in certain subgroups of febrile infants: those less than seven days of age, those with evidence of bacterial meningitis (especially if also <28 days of age), those in whom enterococcal infection is strongly suspected, and those with specific Listeria exposures related to local outbreaks.

The 32-day-old infant described in the clinical scenario was at extremely low risk for listeriosis. Antibiotic coverage with a third-generation cephalosporin is sufficient for the most likely pathogens. The common practice of empirically covering Listeria in otherwise healthy febrile infants considered to be at higher risk for SBI is no longer based on best available evidence and represents overtreatment with at least theoretical harms. Avoidance of the risks associated with the side effects of antibiotics, costs saved by forgoing multiple antibiotics, a decrease in medication dosing frequency, and improved antibiotic stewardship for the general population all argue forcefully for making empiric Listeria coverage a thing of the past.

Disclosure

Nothing to report.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and Liking It on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

The approach to clinical conundrums by an expert clinician is revealed through the presentation of an actual patient’s case in an approach typical of a morning report. Similarly to patient care, sequential pieces of information are provided to the clinician, who is unfamiliar with the case. The focus is on the thought processes of both the clinical team caring for the patient and the discussant. The bolded text represents the patient’s case. Each paragraph that follows represents the discussant’s thoughts.
 

A 42-year-old Malaysian construction worker with subjective fevers of 4 days’ duration presented to an emergency department in Singapore. He reported nonproductive cough, chills without rigors, sore throat, and body aches. He denied sick contacts. Past medical history included chronic hepatitis B virus (HBV) infection. The patient was not taking any medications.

For this patient presenting acutely with subjective fevers, nonproductive cough, chills, aches, and lethargy, initial considerations include infection with a common virus (influenza virus, adenovirus, Epstein-Barr virus [EBV]), acute human immunodeficiency virus (HIV) infection, emerging infection (severe acute respiratory syndrome [SARS], Middle Eastern respiratory syndrome coronavirus [MERS-CoV] infection, avian influenza), and tropical infection (dengue, chikungunya). Also possible are bacterial infections (eg, with Salmonella typhi or Rickettsia or Mycoplasma species), parasitic infections (eg, malaria), and noninfectious illnesses (eg, autoimmune diseases, thyroiditis, acute leukemia, environmental exposures).

The patient’s temperature was 38.5°C; blood pressure, 133/73 mm Hg; heart rate, 95 beats per minute; respiratory rate, 18 breaths per minute; and oxygen saturation, 100% on ambient air. On physical examination, he appeared comfortable, and heart, lung, abdomen, skin, and extremities were normal. Laboratory test results included white blood cell (WBC) count, 4400/μL (with normal differential); hemoglobin, 16.1 g/dL; and platelet count, 207,000/μL. Serum chemistries were normal. C-reactive protein (CRP) level was 44.6 mg/L (reference range, 0.2-9.1 mg/L), and procalcitonin level was 0.13 ng/mL (reference range, <0.50 ng/mL). Chest radiograph was normal. Dengue antibodies (immunoglobulin M, immunoglobulin G [IgG]) and dengue NS1 antigen were negative. The patient was discharged with a presumptive diagnosis of viral upper respiratory tract infection.

There is no left shift characteristic of bacterial infection or lymphopenia characteristic of rickettsial disease or acute HIV infection. The serologic testing and the patient’s overall appearance make dengue unlikely. The low procalcitonin supports a nonbacterial cause of illness. CRP elevation may indicate an inflammatory process and is relatively nonspecific.

Myalgias, pharyngitis, and cough improved over several days, but fevers persisted, and a rash developed over the lower abdomen. The patient returned to the emergency department and was admitted. He denied weight loss and night sweats. He had multiple female sexual partners, including commercial sex workers, within the previous 6 months. Temperature was 38.5°C. The posterior oropharynx was slightly erythematous. There was no lymphadenopathy. Firm, mildly erythematous macules were present on the anterior abdominal wall (Figure 1). The rest of the physical examination was normal.

Laboratory testing revealed WBC count, 5800/μL (75% neutrophils, 19% lymphocytes, 3% monocytes, 2% atypical mononuclear cells); hemoglobin, 16.3 g/dL; platelet count, 185,000/μL; sodium, 131 mmol/L; potassium, 3.4 mmol/L; creatinine, 0.9 mg/dL; albumin, 3.2 g/dL; alanine aminotransferase (ALT), 99 U/L; aspartate aminotransferase (AST), 137 U/L; alkaline phosphatase (ALP), 63 U/L; and total bilirubin, 1.9 mg/dL. Prothrombin time was 11.1 seconds; partial thromboplastin time, 36.1 seconds; erythrocyte sedimentation rate, 14 mm/h; and CRP, 62.2 mg/L.

EBV, acute HIV, and cytomegalovirus infections often present with adenopathy, which is absent here. Disseminated gonococcal infection can manifest with fever, body aches, and rash, but his rash and the absence of penile discharge, migratory arthritis, and enthesitis are not characteristic. Mycoplasma infection can present with macules, urticaria, or erythema multiforme. Rickettsia illnesses typically cause vasculitis with progression to petechiae or purpura resulting from endothelial damage. Patients with secondary syphilis may have widespread macular lesions, and the accompanying syphilitic hepatitis often manifests with elevations in ALP instead of ALT and AST. The mild elevation in ALT and AST can occur with many systemic viral infections. Sweet syndrome may manifest with febrile illness and rash, but the acuity of this patient’s illness and the rapid evolution favor infection.

The patient’s fevers (35°-40°C) continued without pattern over the next 3 days. Blood and urine cultures were negative. Polymerase chain reaction (PCR) test of the nasal mucosa was negative for respiratory viruses. PCR blood tests for EBV, HIV-1, and cytomegalovirus were also negative. Antistreptolysin O (ASO) titer was 400 IU/mm (reference range, <200 IU/mm). Antinuclear antibodies were negative, and rheumatoid factor was 12.4 U/mL (reference range, <10.3 U/mL). Computed tomography (CT) of the thorax, abdomen, and pelvis was normal. Results of a biopsy of an anterior abdominal wall skin lesion showed perivascular and periadnexal lymphocytic inflammation. Amoxicillin was started for the treatment of possible group A streptococcal infection.

PCR for HIV would be positive at a high level in acute HIV. The skin biopsy is not characteristic of Sweet syndrome, which typically shows neutrophilic infiltrate without leukocytoclastic vasculitis, or of syphilis, which typically shows a plasma cell infiltrate.

The patient’s erythematous oropharynx may indicate recent streptococcal pharyngitis. The fevers, elevated ASO titer, and CRP level are consistent with acute rheumatic fever, but arthritis, carditis, and neurologic manifestations are lacking. Erythema marginatum manifests on the trunk and limbs as macules or papules with central clearing as the lesions spread outward—and differs from the patient’s rash, which is firm and restricted to the abdominal wall.

Fevers persisted through hospital day 7. The WBC count was 1100/μL (75.7% neutrophils, 22.5% lymphocytes), hemoglobin was 10.3 g/dL, and platelet count was 52,000/μL. Additional laboratory test results included ALP, 234 U/L; ALT, 250 U/L; AST, 459 U/L; lactate dehydrogenase, 2303 U/L (reference range, 222-454 U/L); and ferritin, 14,964 ng/mL (reference range, 47-452 ng/mL).

The duration of illness and negative diagnostic tests for infections increases suspicion for a noninfectious illness. Conditions commonly associated with marked hyperferritinemia include adult-onset Still disease (AOSD) and hemophagocytic lymphohistiocytosis (HLH). Of the 9 AOSD diagnostic (Yamaguchi) criteria, 5 are met in this case: fever, rash, sore throat, abnormal liver function tests, and negative rheumatologic tests. However, the patient lacks arthritis, leukocytosis, lymphadenopathy, and hepatosplenomegaly. Except for the elevated ferritin, the AOSD criteria overlap substantially with the criteria for acute rheumatic fever, and still require that infections be adequately excluded. HLH, a state of abnormal immune activation with resultant organ dysfunction, can be a primary disorder, but in adults more often is secondary to underlying infectious, autoimmune, or malignant (often lymphoma) conditions. Elevated ferritin, cytopenias, elevated ALT and AST, elevated CRP and erythrocyte sedimentation rate, and elevated lactate dehydrogenase are consistent with HLH. The HLH diagnosis can be more firmly established with the more specific findings of hypertriglyceridemia, hypofibrinogenemia, and elevated soluble CD25 level. The histopathologic finding of hemophagocytosis in the bone marrow, lymph nodes, or liver may further support the diagnosis of HLH.

Rash and fevers persisted. Hepatitis A, hepatitis C, Rickettsia IgG, Burkholderia pseudomallei (the causative organism of melioidosis), and Leptospira serologies, as well as PCR for herpes simplex virus and parvovirus, were all negative. Hepatitis B viral load was 962 IU/mL (2.98 log), hepatitis B envelope antigen was negative, and hepatitis B envelope antibody was positive. Orientia tsutsugamushi (organism responsible for scrub typhus) IgG titer was elevated at 1:128. Antiliver kidney microsomal antibodies and antineutrophil cytoplasmic antibodies were negative. Fibrinogen level was 0.69 g/L (reference range, 1.8-4.8 g/L), and beta-2 microglobulin level was 5078 ng/mL (reference range, 878-2000 ng/mL). Bone marrow biopsy results showed hypocellular marrow with suppressed myelopoiesis, few atypical lymphoid cells, and few hemophagocytes. Flow cytometry was negative for clonal B lymphocytes and aberrant expression of T lymphocytes. Bone marrow myobacterial PCR and fungal cultures were negative.

The patient’s chronic HBV infection is unlikely to be related to his presentation given his low viral load and absence of signs of hepatic dysfunction. Excluding rickettsial disease requires paired acute and convalescent serologies. O tsutsugamushi, the causative agent of the rickettsial disease scrub typhus, is endemic in Malaysia; thus, his positive O tsutsugamushi IgG may indicate past exposure. His fevers, myalgias, truncal rash, and hepatitis are consistent with scrub typhus, but he lacks the characteristic severe headache and generalized lymphadenopathy. Although eschar formation with evolution of a papular rash is common in scrub typhus, it is often absent in the variant found in Southeast Asia. Although elevated β₂ microglobulin level is used as a prognostic marker in multiple myeloma and Waldenström macroglobulinemia, it can be elevated in many immune-active states. The patient likely has HLH, which is supported by the hemophagocytosis seen on bone marrow biopsy, and the hypofibrinogenemia. Potential HLH triggers include O tsutsugamushi infection or recent streptococcal pharyngitis.

A deep-punch skin biopsy of the anterior abdominal wall skin lesion was performed because of the absence of subcutaneous fat in the first biopsy specimen. The latest biopsy results showed irregular interstitial expansion of medium-size lymphocytes in a lobular panniculated pattern. The lymphocytes contained enlarged, irregularly contoured nucleoli and were positive for T-cell markers CD2 and CD3 with reduction in CD5 expression. The lymphomatous cells were of CD8+ with uniform expression of activated cytotoxic granule protein granzyme B and were positive for T-cell hemireceptor β.

Positron emission tomography (PET) CT, obtained for staging purposes, showed multiple hypermetabolic subcutaneous and cutaneous lesions over the torso and upper and lower limbs—compatible with lymphomatous infiltrates (Figure 2). Examination, pathology, and imaging findings suggested a rare neoplasm: subcutaneous panniculitis-like T-cell lymphoma (SPTCL). SPTCL was confirmed by T-cell receptor gene rearrangements studies.

When residents or visitors to tropical or sub-tropical regions, those located near or between the Tropics of Cancer and Capricorn, present with fever, physicians usually first think of infectious diseases. This patient’s case is a reminder that these important first considerations should not be the last.

Generating a differential diagnosis for tropical illnesses begins with the patient’s history. Factors to be considered include location (regional disease prevalence), exposures (food/water ingestion, outdoor work/recreation, sexual contact, animal contact), and timing (temporal relationship of symptom development to possible exposure). Common tropical infections are malaria, dengue, typhoid, and emerging infections such as chikungunya, avian influenza, and Zika virus infection.¹This case underscores the need to analyze diagnostic tests critically. Interpreting tests as simply positive or negative, irrespective of disease features, epidemiology, and test characteristics, can contribute to diagnostic error. For example, the patient’s positive ASO titer requires an understanding of disease features and a nuanced interpretation based on the clinical presentation. The erythematous posterior oropharynx prompted concern for postinfectious sequelae of streptococcal pharyngitis, but his illness was more severe and more prolonged than is typical of that condition. The isolated elevated O tsutsugamushi IgG titer provides an example of the role of epidemiology in test interpretation. Although a single positive value might indicate a new exposure for a visitor to an endemic region, IgG seropositivity in Singapore, where scrub typhus is endemic, likely reflects prior exposure to the organism. Diagnosing an acute scrub typhus infection in a patient in an endemic region requires PCR testing. The skin biopsy results highlight the importance of understanding test characteristics. A skin biopsy specimen must be adequate in order to draw valid and accurate conclusions. In this case, the initial skin biopsy was superficial, and the specimen inadequate, but the test was not “negative.” In the diagnostic skin biopsy, deeper tissue was sampled, and panniculitis (inflammation of subcutaneous fat), which arises in inflammatory, infectious, traumatic, enzymatic, and malignant conditions, was identified. An adequate biopsy specimen that contains subcutaneous fat is essential in making this diagnosis.²This patient eventually manifested several elements of hemophagocytic lymphohistiocytosis (HLH), a syndrome of excessive inflammation and resultant organ injury relating to abnormal immune activation and excessive inflammation. HLH results from deficient down-regulation of activated macrophages and lymphocytes.³ It was initially described in pediatric patients but is now recognized in adults, and associated with mortality as high as 50%.³ A high ferritin level (>2000 ng/mL) has 70% sensitivity and 68% specificity for pediatric HLH and should trigger consideration of HLH in any age group.⁴ The diagnostic criteria for HLH initially proposed in 2004 by the Histiocyte Society to identify patients for recruitment into a clinical trial included molecular testing consistent with HLH and/or 5 of 8 clinical, laboratory, or histopathologic features (Table 1).⁵ HScore is a more recent validated scoring system that predicts the probability of HLH (Table 2). A score above 169 signifies diagnostic sensitivity of 93% and specificity of 86%.⁶

• A febrile illness acquired in the tropics is not always attributable to a tropical infection.
• To avoid diagnostic error, weigh positive or negative test results against disease features, patient epidemiology, and test characteristics.
• HLH is characterized by fevers, cytopenias, hepatosplenomegaly, hyperferritinemia, hypertriglyceridemia, and hypofibrinogenemia. In tissue specimens, hemophagocytosis may help differentiate HLH from competing conditions.
• After HLH is diagnosed, try to determine its underlying cause, which may be an infection, autoimmunity, or a malignancy (commonly, a lymphoma).

Disclosure

Nothing to report.

The approach to clinical conundrums by an expert clinician is revealed through the presentation of an actual patient’s case in an approach typical of a morning report. Similarly to patient care, sequential pieces of information are provided to the clinician, who is unfamiliar with the case. The focus is on the thought processes of both the clinical team caring for the patient and the discussant. The bolded text represents the patient’s case. Each paragraph that follows represents the discussant’s thoughts.
 

A 42-year-old Malaysian construction worker with subjective fevers of 4 days’ duration presented to an emergency department in Singapore. He reported nonproductive cough, chills without rigors, sore throat, and body aches. He denied sick contacts. Past medical history included chronic hepatitis B virus (HBV) infection. The patient was not taking any medications.

For this patient presenting acutely with subjective fevers, nonproductive cough, chills, aches, and lethargy, initial considerations include infection with a common virus (influenza virus, adenovirus, Epstein-Barr virus [EBV]), acute human immunodeficiency virus (HIV) infection, emerging infection (severe acute respiratory syndrome [SARS], Middle Eastern respiratory syndrome coronavirus [MERS-CoV] infection, avian influenza), and tropical infection (dengue, chikungunya). Also possible are bacterial infections (eg, with Salmonella typhi or Rickettsia or Mycoplasma species), parasitic infections (eg, malaria), and noninfectious illnesses (eg, autoimmune diseases, thyroiditis, acute leukemia, environmental exposures).

The patient’s temperature was 38.5°C; blood pressure, 133/73 mm Hg; heart rate, 95 beats per minute; respiratory rate, 18 breaths per minute; and oxygen saturation, 100% on ambient air. On physical examination, he appeared comfortable, and heart, lung, abdomen, skin, and extremities were normal. Laboratory test results included white blood cell (WBC) count, 4400/μL (with normal differential); hemoglobin, 16.1 g/dL; and platelet count, 207,000/μL. Serum chemistries were normal. C-reactive protein (CRP) level was 44.6 mg/L (reference range, 0.2-9.1 mg/L), and procalcitonin level was 0.13 ng/mL (reference range, <0.50 ng/mL). Chest radiograph was normal. Dengue antibodies (immunoglobulin M, immunoglobulin G [IgG]) and dengue NS1 antigen were negative. The patient was discharged with a presumptive diagnosis of viral upper respiratory tract infection.

There is no left shift characteristic of bacterial infection or lymphopenia characteristic of rickettsial disease or acute HIV infection. The serologic testing and the patient’s overall appearance make dengue unlikely. The low procalcitonin supports a nonbacterial cause of illness. CRP elevation may indicate an inflammatory process and is relatively nonspecific.

Myalgias, pharyngitis, and cough improved over several days, but fevers persisted, and a rash developed over the lower abdomen. The patient returned to the emergency department and was admitted. He denied weight loss and night sweats. He had multiple female sexual partners, including commercial sex workers, within the previous 6 months. Temperature was 38.5°C. The posterior oropharynx was slightly erythematous. There was no lymphadenopathy. Firm, mildly erythematous macules were present on the anterior abdominal wall (Figure 1). The rest of the physical examination was normal.

Laboratory testing revealed WBC count, 5800/μL (75% neutrophils, 19% lymphocytes, 3% monocytes, 2% atypical mononuclear cells); hemoglobin, 16.3 g/dL; platelet count, 185,000/μL; sodium, 131 mmol/L; potassium, 3.4 mmol/L; creatinine, 0.9 mg/dL; albumin, 3.2 g/dL; alanine aminotransferase (ALT), 99 U/L; aspartate aminotransferase (AST), 137 U/L; alkaline phosphatase (ALP), 63 U/L; and total bilirubin, 1.9 mg/dL. Prothrombin time was 11.1 seconds; partial thromboplastin time, 36.1 seconds; erythrocyte sedimentation rate, 14 mm/h; and CRP, 62.2 mg/L.

EBV, acute HIV, and cytomegalovirus infections often present with adenopathy, which is absent here. Disseminated gonococcal infection can manifest with fever, body aches, and rash, but his rash and the absence of penile discharge, migratory arthritis, and enthesitis are not characteristic. Mycoplasma infection can present with macules, urticaria, or erythema multiforme. Rickettsia illnesses typically cause vasculitis with progression to petechiae or purpura resulting from endothelial damage. Patients with secondary syphilis may have widespread macular lesions, and the accompanying syphilitic hepatitis often manifests with elevations in ALP instead of ALT and AST. The mild elevation in ALT and AST can occur with many systemic viral infections. Sweet syndrome may manifest with febrile illness and rash, but the acuity of this patient’s illness and the rapid evolution favor infection.

The patient’s fevers (35°-40°C) continued without pattern over the next 3 days. Blood and urine cultures were negative. Polymerase chain reaction (PCR) test of the nasal mucosa was negative for respiratory viruses. PCR blood tests for EBV, HIV-1, and cytomegalovirus were also negative. Antistreptolysin O (ASO) titer was 400 IU/mm (reference range, <200 IU/mm). Antinuclear antibodies were negative, and rheumatoid factor was 12.4 U/mL (reference range, <10.3 U/mL). Computed tomography (CT) of the thorax, abdomen, and pelvis was normal. Results of a biopsy of an anterior abdominal wall skin lesion showed perivascular and periadnexal lymphocytic inflammation. Amoxicillin was started for the treatment of possible group A streptococcal infection.

PCR for HIV would be positive at a high level in acute HIV. The skin biopsy is not characteristic of Sweet syndrome, which typically shows neutrophilic infiltrate without leukocytoclastic vasculitis, or of syphilis, which typically shows a plasma cell infiltrate.

The patient’s erythematous oropharynx may indicate recent streptococcal pharyngitis. The fevers, elevated ASO titer, and CRP level are consistent with acute rheumatic fever, but arthritis, carditis, and neurologic manifestations are lacking. Erythema marginatum manifests on the trunk and limbs as macules or papules with central clearing as the lesions spread outward—and differs from the patient’s rash, which is firm and restricted to the abdominal wall.

Fevers persisted through hospital day 7. The WBC count was 1100/μL (75.7% neutrophils, 22.5% lymphocytes), hemoglobin was 10.3 g/dL, and platelet count was 52,000/μL. Additional laboratory test results included ALP, 234 U/L; ALT, 250 U/L; AST, 459 U/L; lactate dehydrogenase, 2303 U/L (reference range, 222-454 U/L); and ferritin, 14,964 ng/mL (reference range, 47-452 ng/mL).

The duration of illness and negative diagnostic tests for infections increases suspicion for a noninfectious illness. Conditions commonly associated with marked hyperferritinemia include adult-onset Still disease (AOSD) and hemophagocytic lymphohistiocytosis (HLH). Of the 9 AOSD diagnostic (Yamaguchi) criteria, 5 are met in this case: fever, rash, sore throat, abnormal liver function tests, and negative rheumatologic tests. However, the patient lacks arthritis, leukocytosis, lymphadenopathy, and hepatosplenomegaly. Except for the elevated ferritin, the AOSD criteria overlap substantially with the criteria for acute rheumatic fever, and still require that infections be adequately excluded. HLH, a state of abnormal immune activation with resultant organ dysfunction, can be a primary disorder, but in adults more often is secondary to underlying infectious, autoimmune, or malignant (often lymphoma) conditions. Elevated ferritin, cytopenias, elevated ALT and AST, elevated CRP and erythrocyte sedimentation rate, and elevated lactate dehydrogenase are consistent with HLH. The HLH diagnosis can be more firmly established with the more specific findings of hypertriglyceridemia, hypofibrinogenemia, and elevated soluble CD25 level. The histopathologic finding of hemophagocytosis in the bone marrow, lymph nodes, or liver may further support the diagnosis of HLH.

Rash and fevers persisted. Hepatitis A, hepatitis C, Rickettsia IgG, Burkholderia pseudomallei (the causative organism of melioidosis), and Leptospira serologies, as well as PCR for herpes simplex virus and parvovirus, were all negative. Hepatitis B viral load was 962 IU/mL (2.98 log), hepatitis B envelope antigen was negative, and hepatitis B envelope antibody was positive. Orientia tsutsugamushi (organism responsible for scrub typhus) IgG titer was elevated at 1:128. Antiliver kidney microsomal antibodies and antineutrophil cytoplasmic antibodies were negative. Fibrinogen level was 0.69 g/L (reference range, 1.8-4.8 g/L), and beta-2 microglobulin level was 5078 ng/mL (reference range, 878-2000 ng/mL). Bone marrow biopsy results showed hypocellular marrow with suppressed myelopoiesis, few atypical lymphoid cells, and few hemophagocytes. Flow cytometry was negative for clonal B lymphocytes and aberrant expression of T lymphocytes. Bone marrow myobacterial PCR and fungal cultures were negative.

The patient’s chronic HBV infection is unlikely to be related to his presentation given his low viral load and absence of signs of hepatic dysfunction. Excluding rickettsial disease requires paired acute and convalescent serologies. O tsutsugamushi, the causative agent of the rickettsial disease scrub typhus, is endemic in Malaysia; thus, his positive O tsutsugamushi IgG may indicate past exposure. His fevers, myalgias, truncal rash, and hepatitis are consistent with scrub typhus, but he lacks the characteristic severe headache and generalized lymphadenopathy. Although eschar formation with evolution of a papular rash is common in scrub typhus, it is often absent in the variant found in Southeast Asia. Although elevated β₂ microglobulin level is used as a prognostic marker in multiple myeloma and Waldenström macroglobulinemia, it can be elevated in many immune-active states. The patient likely has HLH, which is supported by the hemophagocytosis seen on bone marrow biopsy, and the hypofibrinogenemia. Potential HLH triggers include O tsutsugamushi infection or recent streptococcal pharyngitis.

A deep-punch skin biopsy of the anterior abdominal wall skin lesion was performed because of the absence of subcutaneous fat in the first biopsy specimen. The latest biopsy results showed irregular interstitial expansion of medium-size lymphocytes in a lobular panniculated pattern. The lymphocytes contained enlarged, irregularly contoured nucleoli and were positive for T-cell markers CD2 and CD3 with reduction in CD5 expression. The lymphomatous cells were of CD8+ with uniform expression of activated cytotoxic granule protein granzyme B and were positive for T-cell hemireceptor β.

Positron emission tomography (PET) CT, obtained for staging purposes, showed multiple hypermetabolic subcutaneous and cutaneous lesions over the torso and upper and lower limbs—compatible with lymphomatous infiltrates (Figure 2). Examination, pathology, and imaging findings suggested a rare neoplasm: subcutaneous panniculitis-like T-cell lymphoma (SPTCL). SPTCL was confirmed by T-cell receptor gene rearrangements studies.

When residents or visitors to tropical or sub-tropical regions, those located near or between the Tropics of Cancer and Capricorn, present with fever, physicians usually first think of infectious diseases. This patient’s case is a reminder that these important first considerations should not be the last.

Generating a differential diagnosis for tropical illnesses begins with the patient’s history. Factors to be considered include location (regional disease prevalence), exposures (food/water ingestion, outdoor work/recreation, sexual contact, animal contact), and timing (temporal relationship of symptom development to possible exposure). Common tropical infections are malaria, dengue, typhoid, and emerging infections such as chikungunya, avian influenza, and Zika virus infection.¹This case underscores the need to analyze diagnostic tests critically. Interpreting tests as simply positive or negative, irrespective of disease features, epidemiology, and test characteristics, can contribute to diagnostic error. For example, the patient’s positive ASO titer requires an understanding of disease features and a nuanced interpretation based on the clinical presentation. The erythematous posterior oropharynx prompted concern for postinfectious sequelae of streptococcal pharyngitis, but his illness was more severe and more prolonged than is typical of that condition. The isolated elevated O tsutsugamushi IgG titer provides an example of the role of epidemiology in test interpretation. Although a single positive value might indicate a new exposure for a visitor to an endemic region, IgG seropositivity in Singapore, where scrub typhus is endemic, likely reflects prior exposure to the organism. Diagnosing an acute scrub typhus infection in a patient in an endemic region requires PCR testing. The skin biopsy results highlight the importance of understanding test characteristics. A skin biopsy specimen must be adequate in order to draw valid and accurate conclusions. In this case, the initial skin biopsy was superficial, and the specimen inadequate, but the test was not “negative.” In the diagnostic skin biopsy, deeper tissue was sampled, and panniculitis (inflammation of subcutaneous fat), which arises in inflammatory, infectious, traumatic, enzymatic, and malignant conditions, was identified. An adequate biopsy specimen that contains subcutaneous fat is essential in making this diagnosis.²This patient eventually manifested several elements of hemophagocytic lymphohistiocytosis (HLH), a syndrome of excessive inflammation and resultant organ injury relating to abnormal immune activation and excessive inflammation. HLH results from deficient down-regulation of activated macrophages and lymphocytes.³ It was initially described in pediatric patients but is now recognized in adults, and associated with mortality as high as 50%.³ A high ferritin level (>2000 ng/mL) has 70% sensitivity and 68% specificity for pediatric HLH and should trigger consideration of HLH in any age group.⁴ The diagnostic criteria for HLH initially proposed in 2004 by the Histiocyte Society to identify patients for recruitment into a clinical trial included molecular testing consistent with HLH and/or 5 of 8 clinical, laboratory, or histopathologic features (Table 1).⁵ HScore is a more recent validated scoring system that predicts the probability of HLH (Table 2). A score above 169 signifies diagnostic sensitivity of 93% and specificity of 86%.⁶

• A febrile illness acquired in the tropics is not always attributable to a tropical infection.
• To avoid diagnostic error, weigh positive or negative test results against disease features, patient epidemiology, and test characteristics.
• HLH is characterized by fevers, cytopenias, hepatosplenomegaly, hyperferritinemia, hypertriglyceridemia, and hypofibrinogenemia. In tissue specimens, hemophagocytosis may help differentiate HLH from competing conditions.
• After HLH is diagnosed, try to determine its underlying cause, which may be an infection, autoimmunity, or a malignancy (commonly, a lymphoma).

Disclosure

Nothing to report.

The approach to clinical conundrums by an expert clinician is revealed through the presentation of an actual patient’s case in an approach typical of a morning report. Similarly to patient care, sequential pieces of information are provided to the clinician, who is unfamiliar with the case. The focus is on the thought processes of both the clinical team caring for the patient and the discussant. The bolded text represents the patient’s case. Each paragraph that follows represents the discussant’s thoughts.
 

A 42-year-old Malaysian construction worker with subjective fevers of 4 days’ duration presented to an emergency department in Singapore. He reported nonproductive cough, chills without rigors, sore throat, and body aches. He denied sick contacts. Past medical history included chronic hepatitis B virus (HBV) infection. The patient was not taking any medications.

For this patient presenting acutely with subjective fevers, nonproductive cough, chills, aches, and lethargy, initial considerations include infection with a common virus (influenza virus, adenovirus, Epstein-Barr virus [EBV]), acute human immunodeficiency virus (HIV) infection, emerging infection (severe acute respiratory syndrome [SARS], Middle Eastern respiratory syndrome coronavirus [MERS-CoV] infection, avian influenza), and tropical infection (dengue, chikungunya). Also possible are bacterial infections (eg, with Salmonella typhi or Rickettsia or Mycoplasma species), parasitic infections (eg, malaria), and noninfectious illnesses (eg, autoimmune diseases, thyroiditis, acute leukemia, environmental exposures).

The patient’s temperature was 38.5°C; blood pressure, 133/73 mm Hg; heart rate, 95 beats per minute; respiratory rate, 18 breaths per minute; and oxygen saturation, 100% on ambient air. On physical examination, he appeared comfortable, and heart, lung, abdomen, skin, and extremities were normal. Laboratory test results included white blood cell (WBC) count, 4400/μL (with normal differential); hemoglobin, 16.1 g/dL; and platelet count, 207,000/μL. Serum chemistries were normal. C-reactive protein (CRP) level was 44.6 mg/L (reference range, 0.2-9.1 mg/L), and procalcitonin level was 0.13 ng/mL (reference range, <0.50 ng/mL). Chest radiograph was normal. Dengue antibodies (immunoglobulin M, immunoglobulin G [IgG]) and dengue NS1 antigen were negative. The patient was discharged with a presumptive diagnosis of viral upper respiratory tract infection.

There is no left shift characteristic of bacterial infection or lymphopenia characteristic of rickettsial disease or acute HIV infection. The serologic testing and the patient’s overall appearance make dengue unlikely. The low procalcitonin supports a nonbacterial cause of illness. CRP elevation may indicate an inflammatory process and is relatively nonspecific.

Myalgias, pharyngitis, and cough improved over several days, but fevers persisted, and a rash developed over the lower abdomen. The patient returned to the emergency department and was admitted. He denied weight loss and night sweats. He had multiple female sexual partners, including commercial sex workers, within the previous 6 months. Temperature was 38.5°C. The posterior oropharynx was slightly erythematous. There was no lymphadenopathy. Firm, mildly erythematous macules were present on the anterior abdominal wall (Figure 1). The rest of the physical examination was normal.

Laboratory testing revealed WBC count, 5800/μL (75% neutrophils, 19% lymphocytes, 3% monocytes, 2% atypical mononuclear cells); hemoglobin, 16.3 g/dL; platelet count, 185,000/μL; sodium, 131 mmol/L; potassium, 3.4 mmol/L; creatinine, 0.9 mg/dL; albumin, 3.2 g/dL; alanine aminotransferase (ALT), 99 U/L; aspartate aminotransferase (AST), 137 U/L; alkaline phosphatase (ALP), 63 U/L; and total bilirubin, 1.9 mg/dL. Prothrombin time was 11.1 seconds; partial thromboplastin time, 36.1 seconds; erythrocyte sedimentation rate, 14 mm/h; and CRP, 62.2 mg/L.

EBV, acute HIV, and cytomegalovirus infections often present with adenopathy, which is absent here. Disseminated gonococcal infection can manifest with fever, body aches, and rash, but his rash and the absence of penile discharge, migratory arthritis, and enthesitis are not characteristic. Mycoplasma infection can present with macules, urticaria, or erythema multiforme. Rickettsia illnesses typically cause vasculitis with progression to petechiae or purpura resulting from endothelial damage. Patients with secondary syphilis may have widespread macular lesions, and the accompanying syphilitic hepatitis often manifests with elevations in ALP instead of ALT and AST. The mild elevation in ALT and AST can occur with many systemic viral infections. Sweet syndrome may manifest with febrile illness and rash, but the acuity of this patient’s illness and the rapid evolution favor infection.

The patient’s fevers (35°-40°C) continued without pattern over the next 3 days. Blood and urine cultures were negative. Polymerase chain reaction (PCR) test of the nasal mucosa was negative for respiratory viruses. PCR blood tests for EBV, HIV-1, and cytomegalovirus were also negative. Antistreptolysin O (ASO) titer was 400 IU/mm (reference range, <200 IU/mm). Antinuclear antibodies were negative, and rheumatoid factor was 12.4 U/mL (reference range, <10.3 U/mL). Computed tomography (CT) of the thorax, abdomen, and pelvis was normal. Results of a biopsy of an anterior abdominal wall skin lesion showed perivascular and periadnexal lymphocytic inflammation. Amoxicillin was started for the treatment of possible group A streptococcal infection.

PCR for HIV would be positive at a high level in acute HIV. The skin biopsy is not characteristic of Sweet syndrome, which typically shows neutrophilic infiltrate without leukocytoclastic vasculitis, or of syphilis, which typically shows a plasma cell infiltrate.

The patient’s erythematous oropharynx may indicate recent streptococcal pharyngitis. The fevers, elevated ASO titer, and CRP level are consistent with acute rheumatic fever, but arthritis, carditis, and neurologic manifestations are lacking. Erythema marginatum manifests on the trunk and limbs as macules or papules with central clearing as the lesions spread outward—and differs from the patient’s rash, which is firm and restricted to the abdominal wall.

Fevers persisted through hospital day 7. The WBC count was 1100/μL (75.7% neutrophils, 22.5% lymphocytes), hemoglobin was 10.3 g/dL, and platelet count was 52,000/μL. Additional laboratory test results included ALP, 234 U/L; ALT, 250 U/L; AST, 459 U/L; lactate dehydrogenase, 2303 U/L (reference range, 222-454 U/L); and ferritin, 14,964 ng/mL (reference range, 47-452 ng/mL).

The duration of illness and negative diagnostic tests for infections increases suspicion for a noninfectious illness. Conditions commonly associated with marked hyperferritinemia include adult-onset Still disease (AOSD) and hemophagocytic lymphohistiocytosis (HLH). Of the 9 AOSD diagnostic (Yamaguchi) criteria, 5 are met in this case: fever, rash, sore throat, abnormal liver function tests, and negative rheumatologic tests. However, the patient lacks arthritis, leukocytosis, lymphadenopathy, and hepatosplenomegaly. Except for the elevated ferritin, the AOSD criteria overlap substantially with the criteria for acute rheumatic fever, and still require that infections be adequately excluded. HLH, a state of abnormal immune activation with resultant organ dysfunction, can be a primary disorder, but in adults more often is secondary to underlying infectious, autoimmune, or malignant (often lymphoma) conditions. Elevated ferritin, cytopenias, elevated ALT and AST, elevated CRP and erythrocyte sedimentation rate, and elevated lactate dehydrogenase are consistent with HLH. The HLH diagnosis can be more firmly established with the more specific findings of hypertriglyceridemia, hypofibrinogenemia, and elevated soluble CD25 level. The histopathologic finding of hemophagocytosis in the bone marrow, lymph nodes, or liver may further support the diagnosis of HLH.

Rash and fevers persisted. Hepatitis A, hepatitis C, Rickettsia IgG, Burkholderia pseudomallei (the causative organism of melioidosis), and Leptospira serologies, as well as PCR for herpes simplex virus and parvovirus, were all negative. Hepatitis B viral load was 962 IU/mL (2.98 log), hepatitis B envelope antigen was negative, and hepatitis B envelope antibody was positive. Orientia tsutsugamushi (organism responsible for scrub typhus) IgG titer was elevated at 1:128. Antiliver kidney microsomal antibodies and antineutrophil cytoplasmic antibodies were negative. Fibrinogen level was 0.69 g/L (reference range, 1.8-4.8 g/L), and beta-2 microglobulin level was 5078 ng/mL (reference range, 878-2000 ng/mL). Bone marrow biopsy results showed hypocellular marrow with suppressed myelopoiesis, few atypical lymphoid cells, and few hemophagocytes. Flow cytometry was negative for clonal B lymphocytes and aberrant expression of T lymphocytes. Bone marrow myobacterial PCR and fungal cultures were negative.

The patient’s chronic HBV infection is unlikely to be related to his presentation given his low viral load and absence of signs of hepatic dysfunction. Excluding rickettsial disease requires paired acute and convalescent serologies. O tsutsugamushi, the causative agent of the rickettsial disease scrub typhus, is endemic in Malaysia; thus, his positive O tsutsugamushi IgG may indicate past exposure. His fevers, myalgias, truncal rash, and hepatitis are consistent with scrub typhus, but he lacks the characteristic severe headache and generalized lymphadenopathy. Although eschar formation with evolution of a papular rash is common in scrub typhus, it is often absent in the variant found in Southeast Asia. Although elevated β₂ microglobulin level is used as a prognostic marker in multiple myeloma and Waldenström macroglobulinemia, it can be elevated in many immune-active states. The patient likely has HLH, which is supported by the hemophagocytosis seen on bone marrow biopsy, and the hypofibrinogenemia. Potential HLH triggers include O tsutsugamushi infection or recent streptococcal pharyngitis.

A deep-punch skin biopsy of the anterior abdominal wall skin lesion was performed because of the absence of subcutaneous fat in the first biopsy specimen. The latest biopsy results showed irregular interstitial expansion of medium-size lymphocytes in a lobular panniculated pattern. The lymphocytes contained enlarged, irregularly contoured nucleoli and were positive for T-cell markers CD2 and CD3 with reduction in CD5 expression. The lymphomatous cells were of CD8+ with uniform expression of activated cytotoxic granule protein granzyme B and were positive for T-cell hemireceptor β.

Positron emission tomography (PET) CT, obtained for staging purposes, showed multiple hypermetabolic subcutaneous and cutaneous lesions over the torso and upper and lower limbs—compatible with lymphomatous infiltrates (Figure 2). Examination, pathology, and imaging findings suggested a rare neoplasm: subcutaneous panniculitis-like T-cell lymphoma (SPTCL). SPTCL was confirmed by T-cell receptor gene rearrangements studies.

When residents or visitors to tropical or sub-tropical regions, those located near or between the Tropics of Cancer and Capricorn, present with fever, physicians usually first think of infectious diseases. This patient’s case is a reminder that these important first considerations should not be the last.

Generating a differential diagnosis for tropical illnesses begins with the patient’s history. Factors to be considered include location (regional disease prevalence), exposures (food/water ingestion, outdoor work/recreation, sexual contact, animal contact), and timing (temporal relationship of symptom development to possible exposure). Common tropical infections are malaria, dengue, typhoid, and emerging infections such as chikungunya, avian influenza, and Zika virus infection.¹This case underscores the need to analyze diagnostic tests critically. Interpreting tests as simply positive or negative, irrespective of disease features, epidemiology, and test characteristics, can contribute to diagnostic error. For example, the patient’s positive ASO titer requires an understanding of disease features and a nuanced interpretation based on the clinical presentation. The erythematous posterior oropharynx prompted concern for postinfectious sequelae of streptococcal pharyngitis, but his illness was more severe and more prolonged than is typical of that condition. The isolated elevated O tsutsugamushi IgG titer provides an example of the role of epidemiology in test interpretation. Although a single positive value might indicate a new exposure for a visitor to an endemic region, IgG seropositivity in Singapore, where scrub typhus is endemic, likely reflects prior exposure to the organism. Diagnosing an acute scrub typhus infection in a patient in an endemic region requires PCR testing. The skin biopsy results highlight the importance of understanding test characteristics. A skin biopsy specimen must be adequate in order to draw valid and accurate conclusions. In this case, the initial skin biopsy was superficial, and the specimen inadequate, but the test was not “negative.” In the diagnostic skin biopsy, deeper tissue was sampled, and panniculitis (inflammation of subcutaneous fat), which arises in inflammatory, infectious, traumatic, enzymatic, and malignant conditions, was identified. An adequate biopsy specimen that contains subcutaneous fat is essential in making this diagnosis.²This patient eventually manifested several elements of hemophagocytic lymphohistiocytosis (HLH), a syndrome of excessive inflammation and resultant organ injury relating to abnormal immune activation and excessive inflammation. HLH results from deficient down-regulation of activated macrophages and lymphocytes.³ It was initially described in pediatric patients but is now recognized in adults, and associated with mortality as high as 50%.³ A high ferritin level (>2000 ng/mL) has 70% sensitivity and 68% specificity for pediatric HLH and should trigger consideration of HLH in any age group.⁴ The diagnostic criteria for HLH initially proposed in 2004 by the Histiocyte Society to identify patients for recruitment into a clinical trial included molecular testing consistent with HLH and/or 5 of 8 clinical, laboratory, or histopathologic features (Table 1).⁵ HScore is a more recent validated scoring system that predicts the probability of HLH (Table 2). A score above 169 signifies diagnostic sensitivity of 93% and specificity of 86%.⁶

• A febrile illness acquired in the tropics is not always attributable to a tropical infection.
• To avoid diagnostic error, weigh positive or negative test results against disease features, patient epidemiology, and test characteristics.
• HLH is characterized by fevers, cytopenias, hepatosplenomegaly, hyperferritinemia, hypertriglyceridemia, and hypofibrinogenemia. In tissue specimens, hemophagocytosis may help differentiate HLH from competing conditions.
• After HLH is diagnosed, try to determine its underlying cause, which may be an infection, autoimmunity, or a malignancy (commonly, a lymphoma).

Disclosure

Nothing to report.

Readmission prevention is paramount for hospitals and, by extension, hospitalist programs. Hospitalists see early and reliable outpatient follow-up as a safe landing for their most complicated patient cases. The option of a postdischarge clinic arises from the challenge to arrange adequate postdischarge care for patients who lack easy access because of insurance or provider availability. Guaranteeing postdischarge access by opening a dedicated, hospitalist-led postdischarge clinic appears to be an easy solution, but it is a solution that requires significant investment (including investment in physician and staff training and administrative support) and careful navigation of existing primary care relationships. In addition, a clinic staffed only with physicians may not be well equipped to address the complex social factors in healthcare utilization and readmission. Better understanding of the evidence supporting post discharge physician visits, several models of clinics, and the key operational questions are essential to address before crossing the inpatient-outpatient divide.

A postdischarge outpatient provider visit is often seen as a key factor in reducing readmissions. In 2013, Medicare added strength to this association by establishing transitional care management codes, which provide enhanced reimbursement to providers for a visit within 7 or 14 days of discharge, with focused attention on transitional issues.¹ However, whether a postdischarge visit reduces readmissions remains unclear. Given evidence that higher primary care density is associated with lower healthcare utilization,² CMS’s financial investment in incentivizing post discharge physician visits may be a good bet. On the other hand, simply having a primary care physician (PCP) may be a risk factor for readmission. This association suggests that postdischarge vigilance leads to identification of medical problems that lead to rehospitalization.³ This uncertainty is not resolved in systematic reviews of readmission reduction initiatives, which were not focused solely on the impact of a physician visit.^4,5

The earliest study of postdischarge visits in a general medical population found an association between intensive outpatient follow-up by new providers in a Veterans Affairs population and an increase in hospital readmissions.⁶ This model is similar to some hospitalist models for postdischarge clinics, as the visit was with a noncontinuity provider. The largest recent study, of patients hospitalized with acute myocardial infarction, community-acquired pneumonia, or congestive heart failure (CHF) between 2009 and 2012, found increased frequency of postdischarge follow-up but no concomitant reduction in readmissions.⁷ Although small observational studies⁸ have found a postdischarge primary care visit may reduce the risk for readmission in general medical patients, the bulk of the recent data is negative.

In high-risk patients, however, there may be a clear benefit to postdischarge follow-up. In a North Carolina Medicaid population, a physician visit after discharge was associated with fewer readmissions among high-risk patients, but not among lower risk patients, whose readmission rates were low to start.⁹ The results of that study support the idea that risk stratification may identify patients who can benefit from more intensive outpatient follow-up. In general medical populations, existing studies may suffer from an absence of adequate risk assessment.

The evidence in specific disease states may show a clearer association between a postdischarge physician visit and reduced risk for readmission. One quarter of patients with CHF are rehospitalized within 30 days of discharge.¹⁰ In this disease with frequent exacerbations, a clinic visit to monitor volume status, weight, and medication adherence might reduce the frequency of readmissions or prolong the interval between rehospitalizations. A large observational study observed that earlier post discharge follow up by a cardiologist or a PCP was associated with lower risk of readmission, but only in the quintile with the closest follow-up. In addition, fewer than 40% of patients in this group had a visit within 7 days.¹¹ In another heart failure population, follow-up with either a PCP or cardiologist within 7 days of discharge was again associated with lower risk for readmission.¹² Thus, data suggest a protective effect of postdischarge visits in CHF patients, in contrast to a general medical population. Patients with end-stage renal disease may also fit in this group protected by a postdischarge physician visit, as 1 additional visit within the month after discharge was estimated to reduce rehospitalizations and produce significant cost savings.¹³

With other specific discharge diagnoses, results are varied. Two small observational studies in chronic obstructive pulmonary disease had conflicting results—one found a modest reduction in readmission and emergency department (ED) visits for patients seen by a PCP or pulmonologist within 30 days of discharge,¹⁴ and the other found no effect on readmissions but an associated reduction in mortality.¹⁵ More data are needed to clarify further the interaction of postdischarge visits with mortality, but the association between postdischarge physician visits and readmission reduction is controversial for patients with chronic obstructive pulmonary disease.

Finally, the evidence for dedicated postdischarge clinics is even more limited. A study of a hospitalist-led postdischarge clinic in a Veterans Affairs hospital found reduced length of stay and earlier postdischarge follow-up in a postdischarge clinic, but no effect on readmissions.¹⁶ Other studies have found earlier postdischarge follow-up with dedicated discharge clinics but have not evaluated readmission rates specifically.¹⁷In summary, the effect of postdischarge visits on risk for readmission is an area of active research, but remains unclear. The data reviewed suggest a benefit for the highest risk patients, specifically those with severe chronic illness, or those deemed high-risk with a readmission tool.⁹ At present, because physicians cannot accurately predict which patients will be readmitted,¹⁸ discharging physicians often take a broad approach and schedule outpatient visits for all patients. As readmission tools are further refined, the group of patients who will benefit from postdischarge care will be easier to identify, and a benefit to postdischarge visits may be seen

It is also important to note that this review emphasizes the physician visit and its potential impact on readmissions. Socioeconomic causes are increasingly being recognized as driving readmissions and other utilization.¹⁹ Whether an isolated physician visit is sufficient to prevent readmissions for patients with nonmedical drivers of healthcare utilization is unclear. For those patients, a discharge visit likely is a necessary component of a readmission reduction strategy for high-risk patients, but may be insufficient for patients who require not just an isolated visit but rather a more integrated and comprehensive care program.^8,20,21

Despite the unclear relationship between postdischarge physician care and readmissions, dedicated postdischarge clinics, some staffed by hospitalists, have been adopted over the past 10 years. The three primary types of clinics arise in safety net environments, in academic medical centers, and as comprehensive high-risk patient solutions. Reviewing several types of clinics further clarifies the nature of this structural innovation.

Safety Net Hospital Models

Safety net hospitals and their hospitalists struggle with securing adequate postdischarge access for their population, which has inadequate insurance and poor access to primary care. Patient characteristics also play a role in the complex postdischarge care for this population, given its high rate of ED use (owing to perceived convenience and capabilities) for ambulatory-sensitive conditions.²² In addition, immigrants, particularly those with low English-language proficiency, underuse and have poor access to primary care.^23,24 Postdischarge clinics in this environment focus first on providing a reliable postdischarge plan and then on linking to primary care. Examples of two clinics are at Harborview Medical Center in Seattle, Washington²⁵ and Texas Health in Fort Worth.

Harborview is a 400-bed hospital affiliated with the University of Washington. More than 50% of its patients are considered indigent. The clinic was established in 2007 to provide a postdischarge option for uninsured patients, and a link to primary care in federally qualified health centers. The clinic was staffed 5 days a week with one or two hospitalists or advanced practice nurses. Visit duration was 20 minutes, 270 visits occurred per month, and the no-show rate was 30%. A small subgroup of the hospitalist group staffed the clinic. Particular clinical foci included CHF patients, patients with wound-care needs, and homeless, immigrant, and recently incarcerated patients. A key goal was connecting to longitudinal primary care, and the clinic successfully connected more than 70% of patients to primary care in community health centers. This clinic ultimately transitioned from a hospitalist practice to a primary care practice with a primary focus on post-ED follow-up for unaffiliated patients.²⁶

In 2010, Texas Health faced a similar challenge with unaffiliated patients, and established a nurse practitioner–based clinic with hospitalist oversight to provide care primarily for patients without insurance or without an existing primary care relationship.

Academic Medical Center Models

Another clinical model is designed for patients who receive primary care at practices affiliated with academic medical centers. Although many of these patients have insurance and a PCP, there is often no availability with their continuity provider, because of the resident’s inpatient schedule or the faculty member’s conflicting priorities.^27,28 Academic medical centers, including the University of California at San Francisco, the University of New Mexico, and the Beth Israel Deaconess Medical Center, have established discharge clinics within their faculty primary care practices. A model of this type of clinic was set up at Beth Israel Deaconess in 2010. Staffed by four hospitalists and using 40-minute appointments, this clinic was physically based in the primary care practice. As such, it took advantage of the existing clinic’s administrative and clinical functions, including triage, billing, and scheduling. A visit was scheduled in that clinic by the discharging physician team if a primary care appointment was not available with the patient’s continuity provider. Visits were standardized and focused on outstanding issues at discharge, medication reconciliation, and symptom trajectory. The hospitalists used the clinic’s clinical resources, including nurses, social workers, and pharmacists, but had no other dedicated staff. That there were only four hospitalists meant they were able to gain sufficient exposure to the outpatient setting, provide consistent high-quality care, and gain credibility with the PCPs. As the patients who were seen had PCPs of their own, during the visit significant attention was focused first on the postdischarge concerns, and then on promptly returning the patients to routine primary care. Significant patient outreach was used to address the clinic’s no-show rate, which was almost 50% in the early months. Within a year, the rate was down, closer to 20%. This clinic closed in 2015 after the primary care practice, in which it was based, transitioned to a patient-centered medical home. Since that time, this type of initiative has spread further, with neurohospitalist discharge clinics established, and postdischarge neurology follow-up becoming faster and more reliable.²⁹

Academic medical centers and safety net hospitals substitute for routine primary care to address the basic challenge of primary care access, often without significant enhancements or additional resources, such as dedicated care management and pharmacy, social work, and nursing support. Commonalities of these clinics include dedicated physician staff, appointments generally longer than average outpatient appointments, and visit content concentrated on the key issues at transition (medication reconciliation, outstanding tests, symptom trajectory). As possible, clinics adopted a multidisciplinary approach, with social workers, community health workers, and nurses, to respond to the breadth of patients’ postdischarge needs, which often extend beyond pure medical need. The most frequent barriers encountered included the knowledge gap for hospitalist providers in the outpatient setting (a gap mitigated by using dedicated providers) and the patients’ high no-show rate (not surprising given that the providers are generally new to them). Few clinics have attempted to create continuity across inpatient and outpatient providers, though continuity might reduce no-shows as well as eliminate at least 1 transition.

Comprehensive High-Risk Patient Solutions

At the other end of the clinic spectrum are more integrated postdischarge approaches, which also evolved from the hospitalist model with hospitalist staffing. However, these approaches were introduced in response to the clinical needs of the highest risk patients (who are most vulnerable to frequent provider transitions), not to a systemic inability to provide routine postdischarge care.³⁰

The most long-standing model for this type of clinic is represented by CareMore Health System, a subsidiary of Anthem.^30-32 The extensivist, an expanded-scope hospitalist, acts as primary care coordinator, coordinating a multidisciplinary team for a panel of about 100 patients, representing the sickest 5% of the Medicare Advantage–insured population. Unlike the traditional hospitalist, the extensivist follows patients across all care sites, including hospital, rehabilitation sites, and outpatient clinic. For the most part, this relationship is not designed to evolve into a longitudinal relationship, but rather is an intervention only for the several-months period of acute need. Internal data have shown effects on hospital readmissions as well as length of stay.³⁰

Another integrated clinic was established in 2013, at the University of Chicago. This was an effort to redesign care for patients at highest risk for hospitalization.³³ Similar to the CareMore process, a high-risk population is identified by prior hospitalization and expected high Medicare costs. A comprehensive care physician cares for these patients across care settings. The clinic takes a team-based approach to patient care, with team members selected on the basis of patient need. Physicians have panels limited to only 200 patients, and generally spend part of the day in clinic, and part in seeing their hospitalized patients. Although reminiscent of a traditional primary care setting, this clinic is designed specifically for a high-risk, frequently hospitalized population, and therefore requires physicians with both a skill set akin to that of hospitalists, and an approach of palliative care and holistic patient care. Outcomes from this trial clinic are expected in 2017 or 2018.

Considering some key operational questions (Table) can help guide hospitals, hospitalists, and healthcare systems as they venture into the postdischarge clinic space. Return on investment and sustainability are two key questions for postdischarge clinics.

Return on investment varies by payment structure. In capitated environments with a strong emphasis on readmissions and total medical expenditure, a successful postdischarge clinic would recoup the investment through readmission reduction. However, maintaining adequate patient volume against high no-show rates may strain the group financially. In addition, although a hospitalist group may reap few measurable benefits from this clinical exposure, the unique view of the outpatient world afforded to hospitalists working in this environment could enrich the group as a whole by providing a more well-rounded vantage point.

Another key question surrounds sustainability. The clinic at the Beth Israel Deaconess Medical Center in Boston temporarily closed due to high inpatient volume and corresponding need for those hospitalists in the inpatient setting, early in its inception. It subsequently closed due to evolution in the clinic where it was based, rendering it unnecessary. Clinics that are contingent on other clinics will be vulnerable to external forces. Finally, staffing these clinics may be a stretch for a hospitalist group, as a partly different skill set is required for patient care in the outpatient setting. Hospitalists interested in care transitions are well suited for this role. In addition, hospitalists interested in more clinical variety, or in more schedule variety than that provided in a traditional hospitalist schedule, often enjoy the work. A vast majority of hospitalists think PCPs are responsible for postdischarge problems, and would not be interested in working in the postdischarge world.³⁴ A poor fit for providers may lead to clinic failure.

As evident from this review, gaps in understanding the benefits of postdischarge care have persisted for 10 years. Discharge clinics have been scantly described in the literature. The primary unanswered question remains the effect on readmissions, but this has been the sole research focus to date. Other key research areas are the impact on other patient-centered clinical and system outcomes (eg, patient satisfaction, particularly for patients seeing new providers), postdischarge mortality, the effect on other adverse events, and total medical expenditure.

The healthcare system is evolving in the context of a focus on readmissions, primary care access challenges, and high-risk patients’ specific needs. These forces are spurring innovation in the realm of postdischarge physician clinics, as even the basic need for an appointment may not be met by the existing outpatient primary care system. In this context, multiple new outpatient care structures have arisen, many staffed by hospitalists. Some, such as clinics based in safety net hospitals and academic medical centers, address the simple requirement that patients who lack easy access, because of insurance status or provider availability, can see a doctor after discharge. This type of clinic may be an essential step in alleviating a strained system but may not represent a sustainable long-term solution. More comprehensive solutions for improving patient care and clinical outcomes may be offered by integrated systems, such as CareMore, which also emerged from the hospitalist model. A lasting question is whether these clinics, both the narrowly focused and the comprehensive, will have longevity in the evolving healthcare market. Inevitably, though, hospitalist directors will continue to raise such questions, and should stand to benefit from the experiences of others described in this review.

Disclosure

Nothing to report.

Readmission prevention is paramount for hospitals and, by extension, hospitalist programs. Hospitalists see early and reliable outpatient follow-up as a safe landing for their most complicated patient cases. The option of a postdischarge clinic arises from the challenge to arrange adequate postdischarge care for patients who lack easy access because of insurance or provider availability. Guaranteeing postdischarge access by opening a dedicated, hospitalist-led postdischarge clinic appears to be an easy solution, but it is a solution that requires significant investment (including investment in physician and staff training and administrative support) and careful navigation of existing primary care relationships. In addition, a clinic staffed only with physicians may not be well equipped to address the complex social factors in healthcare utilization and readmission. Better understanding of the evidence supporting post discharge physician visits, several models of clinics, and the key operational questions are essential to address before crossing the inpatient-outpatient divide.

A postdischarge outpatient provider visit is often seen as a key factor in reducing readmissions. In 2013, Medicare added strength to this association by establishing transitional care management codes, which provide enhanced reimbursement to providers for a visit within 7 or 14 days of discharge, with focused attention on transitional issues.¹ However, whether a postdischarge visit reduces readmissions remains unclear. Given evidence that higher primary care density is associated with lower healthcare utilization,² CMS’s financial investment in incentivizing post discharge physician visits may be a good bet. On the other hand, simply having a primary care physician (PCP) may be a risk factor for readmission. This association suggests that postdischarge vigilance leads to identification of medical problems that lead to rehospitalization.³ This uncertainty is not resolved in systematic reviews of readmission reduction initiatives, which were not focused solely on the impact of a physician visit.^4,5

The earliest study of postdischarge visits in a general medical population found an association between intensive outpatient follow-up by new providers in a Veterans Affairs population and an increase in hospital readmissions.⁶ This model is similar to some hospitalist models for postdischarge clinics, as the visit was with a noncontinuity provider. The largest recent study, of patients hospitalized with acute myocardial infarction, community-acquired pneumonia, or congestive heart failure (CHF) between 2009 and 2012, found increased frequency of postdischarge follow-up but no concomitant reduction in readmissions.⁷ Although small observational studies⁸ have found a postdischarge primary care visit may reduce the risk for readmission in general medical patients, the bulk of the recent data is negative.

In high-risk patients, however, there may be a clear benefit to postdischarge follow-up. In a North Carolina Medicaid population, a physician visit after discharge was associated with fewer readmissions among high-risk patients, but not among lower risk patients, whose readmission rates were low to start.⁹ The results of that study support the idea that risk stratification may identify patients who can benefit from more intensive outpatient follow-up. In general medical populations, existing studies may suffer from an absence of adequate risk assessment.

The evidence in specific disease states may show a clearer association between a postdischarge physician visit and reduced risk for readmission. One quarter of patients with CHF are rehospitalized within 30 days of discharge.¹⁰ In this disease with frequent exacerbations, a clinic visit to monitor volume status, weight, and medication adherence might reduce the frequency of readmissions or prolong the interval between rehospitalizations. A large observational study observed that earlier post discharge follow up by a cardiologist or a PCP was associated with lower risk of readmission, but only in the quintile with the closest follow-up. In addition, fewer than 40% of patients in this group had a visit within 7 days.¹¹ In another heart failure population, follow-up with either a PCP or cardiologist within 7 days of discharge was again associated with lower risk for readmission.¹² Thus, data suggest a protective effect of postdischarge visits in CHF patients, in contrast to a general medical population. Patients with end-stage renal disease may also fit in this group protected by a postdischarge physician visit, as 1 additional visit within the month after discharge was estimated to reduce rehospitalizations and produce significant cost savings.¹³

With other specific discharge diagnoses, results are varied. Two small observational studies in chronic obstructive pulmonary disease had conflicting results—one found a modest reduction in readmission and emergency department (ED) visits for patients seen by a PCP or pulmonologist within 30 days of discharge,¹⁴ and the other found no effect on readmissions but an associated reduction in mortality.¹⁵ More data are needed to clarify further the interaction of postdischarge visits with mortality, but the association between postdischarge physician visits and readmission reduction is controversial for patients with chronic obstructive pulmonary disease.

Finally, the evidence for dedicated postdischarge clinics is even more limited. A study of a hospitalist-led postdischarge clinic in a Veterans Affairs hospital found reduced length of stay and earlier postdischarge follow-up in a postdischarge clinic, but no effect on readmissions.¹⁶ Other studies have found earlier postdischarge follow-up with dedicated discharge clinics but have not evaluated readmission rates specifically.¹⁷In summary, the effect of postdischarge visits on risk for readmission is an area of active research, but remains unclear. The data reviewed suggest a benefit for the highest risk patients, specifically those with severe chronic illness, or those deemed high-risk with a readmission tool.⁹ At present, because physicians cannot accurately predict which patients will be readmitted,¹⁸ discharging physicians often take a broad approach and schedule outpatient visits for all patients. As readmission tools are further refined, the group of patients who will benefit from postdischarge care will be easier to identify, and a benefit to postdischarge visits may be seen

It is also important to note that this review emphasizes the physician visit and its potential impact on readmissions. Socioeconomic causes are increasingly being recognized as driving readmissions and other utilization.¹⁹ Whether an isolated physician visit is sufficient to prevent readmissions for patients with nonmedical drivers of healthcare utilization is unclear. For those patients, a discharge visit likely is a necessary component of a readmission reduction strategy for high-risk patients, but may be insufficient for patients who require not just an isolated visit but rather a more integrated and comprehensive care program.^8,20,21

Despite the unclear relationship between postdischarge physician care and readmissions, dedicated postdischarge clinics, some staffed by hospitalists, have been adopted over the past 10 years. The three primary types of clinics arise in safety net environments, in academic medical centers, and as comprehensive high-risk patient solutions. Reviewing several types of clinics further clarifies the nature of this structural innovation.

Safety Net Hospital Models

Safety net hospitals and their hospitalists struggle with securing adequate postdischarge access for their population, which has inadequate insurance and poor access to primary care. Patient characteristics also play a role in the complex postdischarge care for this population, given its high rate of ED use (owing to perceived convenience and capabilities) for ambulatory-sensitive conditions.²² In addition, immigrants, particularly those with low English-language proficiency, underuse and have poor access to primary care.^23,24 Postdischarge clinics in this environment focus first on providing a reliable postdischarge plan and then on linking to primary care. Examples of two clinics are at Harborview Medical Center in Seattle, Washington²⁵ and Texas Health in Fort Worth.

Harborview is a 400-bed hospital affiliated with the University of Washington. More than 50% of its patients are considered indigent. The clinic was established in 2007 to provide a postdischarge option for uninsured patients, and a link to primary care in federally qualified health centers. The clinic was staffed 5 days a week with one or two hospitalists or advanced practice nurses. Visit duration was 20 minutes, 270 visits occurred per month, and the no-show rate was 30%. A small subgroup of the hospitalist group staffed the clinic. Particular clinical foci included CHF patients, patients with wound-care needs, and homeless, immigrant, and recently incarcerated patients. A key goal was connecting to longitudinal primary care, and the clinic successfully connected more than 70% of patients to primary care in community health centers. This clinic ultimately transitioned from a hospitalist practice to a primary care practice with a primary focus on post-ED follow-up for unaffiliated patients.²⁶

In 2010, Texas Health faced a similar challenge with unaffiliated patients, and established a nurse practitioner–based clinic with hospitalist oversight to provide care primarily for patients without insurance or without an existing primary care relationship.

Academic Medical Center Models

Another clinical model is designed for patients who receive primary care at practices affiliated with academic medical centers. Although many of these patients have insurance and a PCP, there is often no availability with their continuity provider, because of the resident’s inpatient schedule or the faculty member’s conflicting priorities.^27,28 Academic medical centers, including the University of California at San Francisco, the University of New Mexico, and the Beth Israel Deaconess Medical Center, have established discharge clinics within their faculty primary care practices. A model of this type of clinic was set up at Beth Israel Deaconess in 2010. Staffed by four hospitalists and using 40-minute appointments, this clinic was physically based in the primary care practice. As such, it took advantage of the existing clinic’s administrative and clinical functions, including triage, billing, and scheduling. A visit was scheduled in that clinic by the discharging physician team if a primary care appointment was not available with the patient’s continuity provider. Visits were standardized and focused on outstanding issues at discharge, medication reconciliation, and symptom trajectory. The hospitalists used the clinic’s clinical resources, including nurses, social workers, and pharmacists, but had no other dedicated staff. That there were only four hospitalists meant they were able to gain sufficient exposure to the outpatient setting, provide consistent high-quality care, and gain credibility with the PCPs. As the patients who were seen had PCPs of their own, during the visit significant attention was focused first on the postdischarge concerns, and then on promptly returning the patients to routine primary care. Significant patient outreach was used to address the clinic’s no-show rate, which was almost 50% in the early months. Within a year, the rate was down, closer to 20%. This clinic closed in 2015 after the primary care practice, in which it was based, transitioned to a patient-centered medical home. Since that time, this type of initiative has spread further, with neurohospitalist discharge clinics established, and postdischarge neurology follow-up becoming faster and more reliable.²⁹

Academic medical centers and safety net hospitals substitute for routine primary care to address the basic challenge of primary care access, often without significant enhancements or additional resources, such as dedicated care management and pharmacy, social work, and nursing support. Commonalities of these clinics include dedicated physician staff, appointments generally longer than average outpatient appointments, and visit content concentrated on the key issues at transition (medication reconciliation, outstanding tests, symptom trajectory). As possible, clinics adopted a multidisciplinary approach, with social workers, community health workers, and nurses, to respond to the breadth of patients’ postdischarge needs, which often extend beyond pure medical need. The most frequent barriers encountered included the knowledge gap for hospitalist providers in the outpatient setting (a gap mitigated by using dedicated providers) and the patients’ high no-show rate (not surprising given that the providers are generally new to them). Few clinics have attempted to create continuity across inpatient and outpatient providers, though continuity might reduce no-shows as well as eliminate at least 1 transition.

Comprehensive High-Risk Patient Solutions

At the other end of the clinic spectrum are more integrated postdischarge approaches, which also evolved from the hospitalist model with hospitalist staffing. However, these approaches were introduced in response to the clinical needs of the highest risk patients (who are most vulnerable to frequent provider transitions), not to a systemic inability to provide routine postdischarge care.³⁰

The most long-standing model for this type of clinic is represented by CareMore Health System, a subsidiary of Anthem.^30-32 The extensivist, an expanded-scope hospitalist, acts as primary care coordinator, coordinating a multidisciplinary team for a panel of about 100 patients, representing the sickest 5% of the Medicare Advantage–insured population. Unlike the traditional hospitalist, the extensivist follows patients across all care sites, including hospital, rehabilitation sites, and outpatient clinic. For the most part, this relationship is not designed to evolve into a longitudinal relationship, but rather is an intervention only for the several-months period of acute need. Internal data have shown effects on hospital readmissions as well as length of stay.³⁰

Another integrated clinic was established in 2013, at the University of Chicago. This was an effort to redesign care for patients at highest risk for hospitalization.³³ Similar to the CareMore process, a high-risk population is identified by prior hospitalization and expected high Medicare costs. A comprehensive care physician cares for these patients across care settings. The clinic takes a team-based approach to patient care, with team members selected on the basis of patient need. Physicians have panels limited to only 200 patients, and generally spend part of the day in clinic, and part in seeing their hospitalized patients. Although reminiscent of a traditional primary care setting, this clinic is designed specifically for a high-risk, frequently hospitalized population, and therefore requires physicians with both a skill set akin to that of hospitalists, and an approach of palliative care and holistic patient care. Outcomes from this trial clinic are expected in 2017 or 2018.

Considering some key operational questions (Table) can help guide hospitals, hospitalists, and healthcare systems as they venture into the postdischarge clinic space. Return on investment and sustainability are two key questions for postdischarge clinics.

Return on investment varies by payment structure. In capitated environments with a strong emphasis on readmissions and total medical expenditure, a successful postdischarge clinic would recoup the investment through readmission reduction. However, maintaining adequate patient volume against high no-show rates may strain the group financially. In addition, although a hospitalist group may reap few measurable benefits from this clinical exposure, the unique view of the outpatient world afforded to hospitalists working in this environment could enrich the group as a whole by providing a more well-rounded vantage point.

Another key question surrounds sustainability. The clinic at the Beth Israel Deaconess Medical Center in Boston temporarily closed due to high inpatient volume and corresponding need for those hospitalists in the inpatient setting, early in its inception. It subsequently closed due to evolution in the clinic where it was based, rendering it unnecessary. Clinics that are contingent on other clinics will be vulnerable to external forces. Finally, staffing these clinics may be a stretch for a hospitalist group, as a partly different skill set is required for patient care in the outpatient setting. Hospitalists interested in care transitions are well suited for this role. In addition, hospitalists interested in more clinical variety, or in more schedule variety than that provided in a traditional hospitalist schedule, often enjoy the work. A vast majority of hospitalists think PCPs are responsible for postdischarge problems, and would not be interested in working in the postdischarge world.³⁴ A poor fit for providers may lead to clinic failure.

As evident from this review, gaps in understanding the benefits of postdischarge care have persisted for 10 years. Discharge clinics have been scantly described in the literature. The primary unanswered question remains the effect on readmissions, but this has been the sole research focus to date. Other key research areas are the impact on other patient-centered clinical and system outcomes (eg, patient satisfaction, particularly for patients seeing new providers), postdischarge mortality, the effect on other adverse events, and total medical expenditure.

The healthcare system is evolving in the context of a focus on readmissions, primary care access challenges, and high-risk patients’ specific needs. These forces are spurring innovation in the realm of postdischarge physician clinics, as even the basic need for an appointment may not be met by the existing outpatient primary care system. In this context, multiple new outpatient care structures have arisen, many staffed by hospitalists. Some, such as clinics based in safety net hospitals and academic medical centers, address the simple requirement that patients who lack easy access, because of insurance status or provider availability, can see a doctor after discharge. This type of clinic may be an essential step in alleviating a strained system but may not represent a sustainable long-term solution. More comprehensive solutions for improving patient care and clinical outcomes may be offered by integrated systems, such as CareMore, which also emerged from the hospitalist model. A lasting question is whether these clinics, both the narrowly focused and the comprehensive, will have longevity in the evolving healthcare market. Inevitably, though, hospitalist directors will continue to raise such questions, and should stand to benefit from the experiences of others described in this review.

Disclosure

Nothing to report.

Readmission prevention is paramount for hospitals and, by extension, hospitalist programs. Hospitalists see early and reliable outpatient follow-up as a safe landing for their most complicated patient cases. The option of a postdischarge clinic arises from the challenge to arrange adequate postdischarge care for patients who lack easy access because of insurance or provider availability. Guaranteeing postdischarge access by opening a dedicated, hospitalist-led postdischarge clinic appears to be an easy solution, but it is a solution that requires significant investment (including investment in physician and staff training and administrative support) and careful navigation of existing primary care relationships. In addition, a clinic staffed only with physicians may not be well equipped to address the complex social factors in healthcare utilization and readmission. Better understanding of the evidence supporting post discharge physician visits, several models of clinics, and the key operational questions are essential to address before crossing the inpatient-outpatient divide.

A postdischarge outpatient provider visit is often seen as a key factor in reducing readmissions. In 2013, Medicare added strength to this association by establishing transitional care management codes, which provide enhanced reimbursement to providers for a visit within 7 or 14 days of discharge, with focused attention on transitional issues.¹ However, whether a postdischarge visit reduces readmissions remains unclear. Given evidence that higher primary care density is associated with lower healthcare utilization,² CMS’s financial investment in incentivizing post discharge physician visits may be a good bet. On the other hand, simply having a primary care physician (PCP) may be a risk factor for readmission. This association suggests that postdischarge vigilance leads to identification of medical problems that lead to rehospitalization.³ This uncertainty is not resolved in systematic reviews of readmission reduction initiatives, which were not focused solely on the impact of a physician visit.^4,5

The earliest study of postdischarge visits in a general medical population found an association between intensive outpatient follow-up by new providers in a Veterans Affairs population and an increase in hospital readmissions.⁶ This model is similar to some hospitalist models for postdischarge clinics, as the visit was with a noncontinuity provider. The largest recent study, of patients hospitalized with acute myocardial infarction, community-acquired pneumonia, or congestive heart failure (CHF) between 2009 and 2012, found increased frequency of postdischarge follow-up but no concomitant reduction in readmissions.⁷ Although small observational studies⁸ have found a postdischarge primary care visit may reduce the risk for readmission in general medical patients, the bulk of the recent data is negative.

In high-risk patients, however, there may be a clear benefit to postdischarge follow-up. In a North Carolina Medicaid population, a physician visit after discharge was associated with fewer readmissions among high-risk patients, but not among lower risk patients, whose readmission rates were low to start.⁹ The results of that study support the idea that risk stratification may identify patients who can benefit from more intensive outpatient follow-up. In general medical populations, existing studies may suffer from an absence of adequate risk assessment.

The evidence in specific disease states may show a clearer association between a postdischarge physician visit and reduced risk for readmission. One quarter of patients with CHF are rehospitalized within 30 days of discharge.¹⁰ In this disease with frequent exacerbations, a clinic visit to monitor volume status, weight, and medication adherence might reduce the frequency of readmissions or prolong the interval between rehospitalizations. A large observational study observed that earlier post discharge follow up by a cardiologist or a PCP was associated with lower risk of readmission, but only in the quintile with the closest follow-up. In addition, fewer than 40% of patients in this group had a visit within 7 days.¹¹ In another heart failure population, follow-up with either a PCP or cardiologist within 7 days of discharge was again associated with lower risk for readmission.¹² Thus, data suggest a protective effect of postdischarge visits in CHF patients, in contrast to a general medical population. Patients with end-stage renal disease may also fit in this group protected by a postdischarge physician visit, as 1 additional visit within the month after discharge was estimated to reduce rehospitalizations and produce significant cost savings.¹³

With other specific discharge diagnoses, results are varied. Two small observational studies in chronic obstructive pulmonary disease had conflicting results—one found a modest reduction in readmission and emergency department (ED) visits for patients seen by a PCP or pulmonologist within 30 days of discharge,¹⁴ and the other found no effect on readmissions but an associated reduction in mortality.¹⁵ More data are needed to clarify further the interaction of postdischarge visits with mortality, but the association between postdischarge physician visits and readmission reduction is controversial for patients with chronic obstructive pulmonary disease.

Finally, the evidence for dedicated postdischarge clinics is even more limited. A study of a hospitalist-led postdischarge clinic in a Veterans Affairs hospital found reduced length of stay and earlier postdischarge follow-up in a postdischarge clinic, but no effect on readmissions.¹⁶ Other studies have found earlier postdischarge follow-up with dedicated discharge clinics but have not evaluated readmission rates specifically.¹⁷In summary, the effect of postdischarge visits on risk for readmission is an area of active research, but remains unclear. The data reviewed suggest a benefit for the highest risk patients, specifically those with severe chronic illness, or those deemed high-risk with a readmission tool.⁹ At present, because physicians cannot accurately predict which patients will be readmitted,¹⁸ discharging physicians often take a broad approach and schedule outpatient visits for all patients. As readmission tools are further refined, the group of patients who will benefit from postdischarge care will be easier to identify, and a benefit to postdischarge visits may be seen

It is also important to note that this review emphasizes the physician visit and its potential impact on readmissions. Socioeconomic causes are increasingly being recognized as driving readmissions and other utilization.¹⁹ Whether an isolated physician visit is sufficient to prevent readmissions for patients with nonmedical drivers of healthcare utilization is unclear. For those patients, a discharge visit likely is a necessary component of a readmission reduction strategy for high-risk patients, but may be insufficient for patients who require not just an isolated visit but rather a more integrated and comprehensive care program.^8,20,21

Despite the unclear relationship between postdischarge physician care and readmissions, dedicated postdischarge clinics, some staffed by hospitalists, have been adopted over the past 10 years. The three primary types of clinics arise in safety net environments, in academic medical centers, and as comprehensive high-risk patient solutions. Reviewing several types of clinics further clarifies the nature of this structural innovation.

Safety Net Hospital Models

Safety net hospitals and their hospitalists struggle with securing adequate postdischarge access for their population, which has inadequate insurance and poor access to primary care. Patient characteristics also play a role in the complex postdischarge care for this population, given its high rate of ED use (owing to perceived convenience and capabilities) for ambulatory-sensitive conditions.²² In addition, immigrants, particularly those with low English-language proficiency, underuse and have poor access to primary care.^23,24 Postdischarge clinics in this environment focus first on providing a reliable postdischarge plan and then on linking to primary care. Examples of two clinics are at Harborview Medical Center in Seattle, Washington²⁵ and Texas Health in Fort Worth.

Harborview is a 400-bed hospital affiliated with the University of Washington. More than 50% of its patients are considered indigent. The clinic was established in 2007 to provide a postdischarge option for uninsured patients, and a link to primary care in federally qualified health centers. The clinic was staffed 5 days a week with one or two hospitalists or advanced practice nurses. Visit duration was 20 minutes, 270 visits occurred per month, and the no-show rate was 30%. A small subgroup of the hospitalist group staffed the clinic. Particular clinical foci included CHF patients, patients with wound-care needs, and homeless, immigrant, and recently incarcerated patients. A key goal was connecting to longitudinal primary care, and the clinic successfully connected more than 70% of patients to primary care in community health centers. This clinic ultimately transitioned from a hospitalist practice to a primary care practice with a primary focus on post-ED follow-up for unaffiliated patients.²⁶

In 2010, Texas Health faced a similar challenge with unaffiliated patients, and established a nurse practitioner–based clinic with hospitalist oversight to provide care primarily for patients without insurance or without an existing primary care relationship.

Academic Medical Center Models

Another clinical model is designed for patients who receive primary care at practices affiliated with academic medical centers. Although many of these patients have insurance and a PCP, there is often no availability with their continuity provider, because of the resident’s inpatient schedule or the faculty member’s conflicting priorities.^27,28 Academic medical centers, including the University of California at San Francisco, the University of New Mexico, and the Beth Israel Deaconess Medical Center, have established discharge clinics within their faculty primary care practices. A model of this type of clinic was set up at Beth Israel Deaconess in 2010. Staffed by four hospitalists and using 40-minute appointments, this clinic was physically based in the primary care practice. As such, it took advantage of the existing clinic’s administrative and clinical functions, including triage, billing, and scheduling. A visit was scheduled in that clinic by the discharging physician team if a primary care appointment was not available with the patient’s continuity provider. Visits were standardized and focused on outstanding issues at discharge, medication reconciliation, and symptom trajectory. The hospitalists used the clinic’s clinical resources, including nurses, social workers, and pharmacists, but had no other dedicated staff. That there were only four hospitalists meant they were able to gain sufficient exposure to the outpatient setting, provide consistent high-quality care, and gain credibility with the PCPs. As the patients who were seen had PCPs of their own, during the visit significant attention was focused first on the postdischarge concerns, and then on promptly returning the patients to routine primary care. Significant patient outreach was used to address the clinic’s no-show rate, which was almost 50% in the early months. Within a year, the rate was down, closer to 20%. This clinic closed in 2015 after the primary care practice, in which it was based, transitioned to a patient-centered medical home. Since that time, this type of initiative has spread further, with neurohospitalist discharge clinics established, and postdischarge neurology follow-up becoming faster and more reliable.²⁹

Academic medical centers and safety net hospitals substitute for routine primary care to address the basic challenge of primary care access, often without significant enhancements or additional resources, such as dedicated care management and pharmacy, social work, and nursing support. Commonalities of these clinics include dedicated physician staff, appointments generally longer than average outpatient appointments, and visit content concentrated on the key issues at transition (medication reconciliation, outstanding tests, symptom trajectory). As possible, clinics adopted a multidisciplinary approach, with social workers, community health workers, and nurses, to respond to the breadth of patients’ postdischarge needs, which often extend beyond pure medical need. The most frequent barriers encountered included the knowledge gap for hospitalist providers in the outpatient setting (a gap mitigated by using dedicated providers) and the patients’ high no-show rate (not surprising given that the providers are generally new to them). Few clinics have attempted to create continuity across inpatient and outpatient providers, though continuity might reduce no-shows as well as eliminate at least 1 transition.

Comprehensive High-Risk Patient Solutions

At the other end of the clinic spectrum are more integrated postdischarge approaches, which also evolved from the hospitalist model with hospitalist staffing. However, these approaches were introduced in response to the clinical needs of the highest risk patients (who are most vulnerable to frequent provider transitions), not to a systemic inability to provide routine postdischarge care.³⁰

The most long-standing model for this type of clinic is represented by CareMore Health System, a subsidiary of Anthem.^30-32 The extensivist, an expanded-scope hospitalist, acts as primary care coordinator, coordinating a multidisciplinary team for a panel of about 100 patients, representing the sickest 5% of the Medicare Advantage–insured population. Unlike the traditional hospitalist, the extensivist follows patients across all care sites, including hospital, rehabilitation sites, and outpatient clinic. For the most part, this relationship is not designed to evolve into a longitudinal relationship, but rather is an intervention only for the several-months period of acute need. Internal data have shown effects on hospital readmissions as well as length of stay.³⁰

Another integrated clinic was established in 2013, at the University of Chicago. This was an effort to redesign care for patients at highest risk for hospitalization.³³ Similar to the CareMore process, a high-risk population is identified by prior hospitalization and expected high Medicare costs. A comprehensive care physician cares for these patients across care settings. The clinic takes a team-based approach to patient care, with team members selected on the basis of patient need. Physicians have panels limited to only 200 patients, and generally spend part of the day in clinic, and part in seeing their hospitalized patients. Although reminiscent of a traditional primary care setting, this clinic is designed specifically for a high-risk, frequently hospitalized population, and therefore requires physicians with both a skill set akin to that of hospitalists, and an approach of palliative care and holistic patient care. Outcomes from this trial clinic are expected in 2017 or 2018.

Considering some key operational questions (Table) can help guide hospitals, hospitalists, and healthcare systems as they venture into the postdischarge clinic space. Return on investment and sustainability are two key questions for postdischarge clinics.

Return on investment varies by payment structure. In capitated environments with a strong emphasis on readmissions and total medical expenditure, a successful postdischarge clinic would recoup the investment through readmission reduction. However, maintaining adequate patient volume against high no-show rates may strain the group financially. In addition, although a hospitalist group may reap few measurable benefits from this clinical exposure, the unique view of the outpatient world afforded to hospitalists working in this environment could enrich the group as a whole by providing a more well-rounded vantage point.

Another key question surrounds sustainability. The clinic at the Beth Israel Deaconess Medical Center in Boston temporarily closed due to high inpatient volume and corresponding need for those hospitalists in the inpatient setting, early in its inception. It subsequently closed due to evolution in the clinic where it was based, rendering it unnecessary. Clinics that are contingent on other clinics will be vulnerable to external forces. Finally, staffing these clinics may be a stretch for a hospitalist group, as a partly different skill set is required for patient care in the outpatient setting. Hospitalists interested in care transitions are well suited for this role. In addition, hospitalists interested in more clinical variety, or in more schedule variety than that provided in a traditional hospitalist schedule, often enjoy the work. A vast majority of hospitalists think PCPs are responsible for postdischarge problems, and would not be interested in working in the postdischarge world.³⁴ A poor fit for providers may lead to clinic failure.

As evident from this review, gaps in understanding the benefits of postdischarge care have persisted for 10 years. Discharge clinics have been scantly described in the literature. The primary unanswered question remains the effect on readmissions, but this has been the sole research focus to date. Other key research areas are the impact on other patient-centered clinical and system outcomes (eg, patient satisfaction, particularly for patients seeing new providers), postdischarge mortality, the effect on other adverse events, and total medical expenditure.

The healthcare system is evolving in the context of a focus on readmissions, primary care access challenges, and high-risk patients’ specific needs. These forces are spurring innovation in the realm of postdischarge physician clinics, as even the basic need for an appointment may not be met by the existing outpatient primary care system. In this context, multiple new outpatient care structures have arisen, many staffed by hospitalists. Some, such as clinics based in safety net hospitals and academic medical centers, address the simple requirement that patients who lack easy access, because of insurance status or provider availability, can see a doctor after discharge. This type of clinic may be an essential step in alleviating a strained system but may not represent a sustainable long-term solution. More comprehensive solutions for improving patient care and clinical outcomes may be offered by integrated systems, such as CareMore, which also emerged from the hospitalist model. A lasting question is whether these clinics, both the narrowly focused and the comprehensive, will have longevity in the evolving healthcare market. Inevitably, though, hospitalist directors will continue to raise such questions, and should stand to benefit from the experiences of others described in this review.

Disclosure

Nothing to report.

Measles is a highly contagious acute respiratory illness that includes a characteristic rash. After exposure, up to 90% of susceptible persons develop measles.¹ Even though it is considered a childhood illness, measles can affect people of all age groups. Measles continues to be major health problem around the world, despite the availability of a safe and effective vaccine, and it remains one of the leading causes of childhood mortality, with nearly 115,000 deaths reported by the World Health Organization² in 2014. In 2000, measles was declared eliminated from the United States, but outbreaks still occasionally occur.^3-6

The disease is self-limited, but some patients develop complications that may require hospitalization for treatment. People at highest risk for complications are children younger than 5 years, adults older than 20 years, pregnant women, and immunocompromised individuals.⁷

During the licensure of live measles vaccine in 1963, an average of 549,000 measles cases and 495 measles deaths, as well as 48,000 hospitalizations and 4000 encephalitis cases, were reported annually in the United States. Almost all Americans were affected by measles by adolescence.

Implementation of the 1-dose vaccine program substantially reduced reported incidence in the United States by 1988, and led to a dramatic decline in measles-related hospitalizations and deaths.^3-6 The 2-dose MMR (measles, mumps, rubella) vaccination was introduced in 1989, and measles was declared eliminated in the United States in 2000.^3-6

National–level one-dose MMR coverage among children 19-35 months has remained above 90% during the last two decades.⁸ NIS-Teen vaccination coverage data for 13- to 17-year-olds since 2008 has been near or above 90%,⁹ and 94% of children enrolled in kindergarten had evidence of 2 MMR doses in the 2014-2015 school year.¹⁰

A large multistate measles outbreak was reported in the United States in 2014-2015.^4,11 One hundred fifty-nine cases were reported in the United States between January 4 and April 5, 2015. The majority of patients either were unvaccinated (45%) or had an unknown vaccination status (38%). Age ranged from 6 weeks to 70 years, and 22 patients (14%) were hospitalized.⁴

Measles is caused by an RNA-containing paramyxovirus that is spread by the respiratory route. Average incubation period from exposure to rash onset is 14 days (range, 7-21 days).^12,13 Peak infectivity occurs during the prodromal phase, before rash onset (Figure 1), but patients are infectious from 4 days before rash onset through 4 days after rash onset.^7,12,13

The disease prodrome consists of a high fever (39°C-40.5°C), coryza, cough, and conjunctivitis followed by Koplik spots (Figure 2A). Koplik spots are pathognomonic for measles but rarely discovered. They appear before the skin rash alongside second molars on the buccal surface of the cheeks. The spots usually disappear when the characteristic maculopapular, nonpruritic rash erupts initially at the hairline and behind the ears, and within four days progresses toward the trunk and limbs, including the palms and soles (Figures 2B, 2C).

The presenting symptoms of primary measles infection are nonspecific, particularly if Koplik spots are not identified. The differential diagnosis for a patient who presents with high fever and rash include Kawasaki disease, dengue, parvovirus B19, serum sickness, syphilis, systemic lupus erythematous, toxic shock syndrome, enterovirus infection, human herpes virus 6 (roseola), viral hemorrhagic fever, drug eruption, infectious mononucleosis, Rocky Mountain spotted fever, rubella, scarlet fever, chikungunya, and Zika virus infection.

Measles complications can affect nearly every organ system (Table). Rates of complications from measles infection depend on age and underlying condition. Coexisting vitamin A deficiency increases complication rates.²⁰

Bacterial infections in the setting of measles infection are more common in adults than in children, and are more severe among people who are malnourished or have an immunodeficiency disorder. The most common infectious complications, which involve the respiratory tract, include pneumonia, laryngotracheitis (“measles croup”), bronchitis, otitis media (most common complication among children in the United States), and sinusitis.^7,13,21

Indications for hospitalizing children include respiratory distress, laryngeal obstruction, dehydration that requires intravenous fluids, diarrhea with more than 10 stools a day or bloody stool, severe anemia, altered mental status, convulsion, severe rash with developing hemorrhagic areas, extensive mouth ulcers, corneal clouding or ulcers, visual disturbance, and mastoiditis.²²

Pneumonia is a common indication for hospitalizing adults.^23,24 Measles-associated interstitial giant cell (Hecht) pneumonia is most often recognized among immunocompromised and malnourished patients.¹³ Primary pneumonia is caused by the measles virus, but bacterial superinfection can occur. The most common bacterial pathogens include Streptococcus, Pneumococcus, and Staphylococcus,^13,24 and less commonly isolated organisms include gram-negative bacteria, such as Haemophilus influenzae, Pseudomonas aeruginosa, Neisseria meningitides, and Enterobacter cloacae.²³

Uncommon complications of measles are myocarditis, glomerulonephritis, acute renal failure, and thrombocytopenic purpura.^25,26

Neurologic complications in measles are an important concern. Measles-associated central nervous system complications are considered a result of an immune-mediated reaction to myelin protein and not from direct viral insult.^26-28 Immunocompromised patients are at risk for developing fatal encephalitis, and those who survive often experience cognitive decline or seizures.

Measles is associated with four different encephalitic diseases: primary measles encephalitis, acute post-measles encephalomyelitis, measles inclusion body encephalitis, and subacute sclerosing panencephalitis.

Primary measles encephalitis is characterized by fever, headache, stiff neck, and meningeal signs. Onset occurs between 1 and 15 days after rash onset, and the disease affects 1/1000 patients. Seizure, altered mental status, and coma can also develop. Viral RNA detection in the cerebrospinal fluid (CSF) confirms the diagnosis.²⁹Acute post-measles encephalomyelitis is more common in adults than in children.¹² It typically develops after the rash fades and the other symptoms subside. Patients suddenly experience a recurrence of fevers or seizures. Deafness, intellectual decline, epilepsy, postencephalitic hyperkinesia, hemiplegia, and/or paraplegia also can develop.^27-29

Measles inclusion body encephalitis is described only in immunocompromised patients, and onset occurs within 1 year of infection. Seizures are an initial and common symptom, and some patients also experience hemiplegia, stupor, hypertonia, and dysarthria.²⁹ Diagnostic findings include seroconversion during the disease course, improvement after withholding of the immunosuppressive regimen, and normal CSF. Brain biopsy confirms the diagnosis.

Subacute sclerosing panencephalitis (SSPE) is a slowly progressing and untreatable degenerative neurologic disorder characterized by demyelination of multiple brain areas. SSPE develops 7 to 10 years after natural measles infection, and usually affects children or adolescents. Clinical presentation includes intellectual decline, frequent rhythmic myoclonic jerks, seizure, and dementia. As the disease progresses, coma, quadriplegia, vegetative state, and autonomic instability develop. Death usually occurs within 2 years of onset.^30,31 In children, the risk for SSPE after measles infection is estimated to be 4 to 11 per 100,000 infections. After the 1989-1991 resurgence of measles in the United States, however, the risk for SSPE was estimated to be 22 per 100,000 infections.^30-32 The pathogenesis of SSPE is not fully understood but is thought to result from persistent aberrant measles virus infection.³²

The SSPE diagnosis is based on clinical presentation, presence of anti-measles antibodies in CSF, typical electroencephalography pattern (periodic paroxysmal bursts) with accompanying myoclonus, tissue analysis, and magnetic resonance imaging.³⁰

Suspicion for measles should prompt immediate consultation with local or state public health officials. Laboratory testing can be carefully considered after consultation, and care is needed in interpreting serologic studies.

The mainstays of measles infection diagnosis are detection of viral RNA by reverse transcriptase–polymerase chain reaction, or isolation of the virus in the clinical specimen, and detection of measles-specific IgM (immunoglobulin M) antibodies. A detailed protocol for collecting specimens for viral isolation appears on the Centers for Disease Control and Prevention website (http://www.cdc.gov/measles/lab-tools/rt-pcr.html).

IgM antibodies are detectable over the 15 weeks after rash onset, but the recommendation is to collect serum between 72 hours and 4 weeks after rash onset.³³ Clinicians should be aware that false-positive IgM results may occur with rheumatologic diseases, parvovirus B19 infection, rubella, and infectious mononucleosis.

IgG (immunoglobulin G) antibodies are usually detectable a week after rash onset. The laboratory can confirm measles by detecting more than a 4-fold increase in IgG titers between the acute phase and the convalescent phase. After measles infection, most adults develop lifelong immunity with positive IgG serology.³⁴

Additional tests, such as IgG avidity and plaque reduction neutralization assay, can be used to confirm suspected cases in previously vaccinated individuals.³⁴

General Principles

Uncomplicated measles treatment is supportive and includes oral fluids and antipyretics.^7,22 Severe bacterial infections, encephalitis, or dehydration may require hospitalization, and in these cases infectious disease consultation is recommended. Patients with pneumonia, purulent otitis media, or tonsillitis should be treated with antibiotics.³⁵ Observational data suggest antibiotics may reduce the occurrence of bacterial infection in children, but there are no usage guidelines.³⁵ Vitamin A supplementation has been associated with a 50% decrease in morbidity and mortality and with blindness prevention.²² This supplementation should be considered in severe measles cases (all hospitalized patients), especially for children, regardless of country of residence, and for adult patients who exhibit clinical signs of vitamin A deficiency.^22,24

Antiviral Treatment

No specific treatment is available.³⁶ Ribavirin demonstrates in vitro activity against the virus, but the Food and Drug Administration has not approved the drug for treatment of measles. Ribavirin has been used for cases of severe measles, and for patients with SSPE along with intrathecal interferon alpha. This antiviral treatment is considered experimental.³⁷

All patients hospitalized with measles infection should be cautioned about the potential downstream complications of the disease and should follow up with their primary care physician for surveillance after discharge.³⁸

If measles symptoms develop, patients should self-quarantine and contact their primary care physician or public health department as soon as possible. Regardless of immune status, family members and other exposed persons should be educated about the measles symptoms that may occur during the 21 days after exposure.³⁸

Both suspected and confirmed cases of measles should be reported immediately to local public health authorities.

Infection Control and Prophylaxis

Current guidelines recommend 2 doses of measles-containing vaccine to all adults at higher risk for contracting measles: international travelers, healthcare personnel, and high school and college students. Infants 6 or 11 months old should receive 1 MMR dose before international travel.^1,38

Strict airborne isolation—use of N95 respirator or respirator with similar effectiveness in preventing airborne transmission—is mandatory from 3 to 5 days before rash onset to 4 days after rash onset (immunocompetent patients) or for the duration of the disease (immunocompromised patients).³⁸

Healthcare workers should have documented presumptive evidence of immunity to measles.³⁹ Healthcare providers without evidence of immunity should be excused from work from day 5 to day 21 of exposure, even if they have received postexposure vaccine or intramuscular immunoglobulin. They should be offered the first MMR dose within 72 hours of measles exposure to prevent or modify the disease. Susceptible family members or visitors should not be allowed in the patient’s room.¹

Postexposure Prophylaxis

Standard MMR vaccination within 72 hours after exposure may protect against disease in people without a contraindication to measles vaccine. The public health department usually identifies these individuals and provides postexposure prophylaxis recommendations.^38,39

People with HIV, patients receiving immunosuppressive therapy, and pregnant women and infants who have been exposed to measles and who are at risk for developing morbid disease can be treated with immunoglobulin (IG). If administered within 6 days of exposure, IG can prevent or modify disease in people who are unvaccinated or severely immunocompromised (ie, not immune). The recommended dose of IG administered intramuscularly is 0.5 mL/kg of body weight (maximum, 15 mL), and the recommended dose of IG given intravenously is 400 mg/kg. Anyone heavier than 30 kg would require intravenous IG to achieve adequate antibody levels.

Physicians should not vaccinate pregnant women, patients with severe immunosuppression from disease or therapy, patients with moderate or severe illness, and people with a history of severe allergic reaction to the vaccine.^1,40 The measles vaccine should be deferred for 6 months after IG administration.³⁶ More details are available in the recommendations made by the Advisory Committee on Immunization Practices.¹

Although rare in the United States, measles remains a common and potentially devastating infection among patients who have not been vaccinated. Diagnosis requires clinical suspicion, engagement of public health authorities, and judicious use of laboratory testing. Hospitalists may encounter infectious and neurologic complications of measles long after the initial infection and should be aware of these associations.

Disclosure

Nothing to report.

Measles is a highly contagious acute respiratory illness that includes a characteristic rash. After exposure, up to 90% of susceptible persons develop measles.¹ Even though it is considered a childhood illness, measles can affect people of all age groups. Measles continues to be major health problem around the world, despite the availability of a safe and effective vaccine, and it remains one of the leading causes of childhood mortality, with nearly 115,000 deaths reported by the World Health Organization² in 2014. In 2000, measles was declared eliminated from the United States, but outbreaks still occasionally occur.^3-6

The disease is self-limited, but some patients develop complications that may require hospitalization for treatment. People at highest risk for complications are children younger than 5 years, adults older than 20 years, pregnant women, and immunocompromised individuals.⁷

During the licensure of live measles vaccine in 1963, an average of 549,000 measles cases and 495 measles deaths, as well as 48,000 hospitalizations and 4000 encephalitis cases, were reported annually in the United States. Almost all Americans were affected by measles by adolescence.

Implementation of the 1-dose vaccine program substantially reduced reported incidence in the United States by 1988, and led to a dramatic decline in measles-related hospitalizations and deaths.^3-6 The 2-dose MMR (measles, mumps, rubella) vaccination was introduced in 1989, and measles was declared eliminated in the United States in 2000.^3-6

National–level one-dose MMR coverage among children 19-35 months has remained above 90% during the last two decades.⁸ NIS-Teen vaccination coverage data for 13- to 17-year-olds since 2008 has been near or above 90%,⁹ and 94% of children enrolled in kindergarten had evidence of 2 MMR doses in the 2014-2015 school year.¹⁰

A large multistate measles outbreak was reported in the United States in 2014-2015.^4,11 One hundred fifty-nine cases were reported in the United States between January 4 and April 5, 2015. The majority of patients either were unvaccinated (45%) or had an unknown vaccination status (38%). Age ranged from 6 weeks to 70 years, and 22 patients (14%) were hospitalized.⁴

Measles is caused by an RNA-containing paramyxovirus that is spread by the respiratory route. Average incubation period from exposure to rash onset is 14 days (range, 7-21 days).^12,13 Peak infectivity occurs during the prodromal phase, before rash onset (Figure 1), but patients are infectious from 4 days before rash onset through 4 days after rash onset.^7,12,13

The disease prodrome consists of a high fever (39°C-40.5°C), coryza, cough, and conjunctivitis followed by Koplik spots (Figure 2A). Koplik spots are pathognomonic for measles but rarely discovered. They appear before the skin rash alongside second molars on the buccal surface of the cheeks. The spots usually disappear when the characteristic maculopapular, nonpruritic rash erupts initially at the hairline and behind the ears, and within four days progresses toward the trunk and limbs, including the palms and soles (Figures 2B, 2C).

The presenting symptoms of primary measles infection are nonspecific, particularly if Koplik spots are not identified. The differential diagnosis for a patient who presents with high fever and rash include Kawasaki disease, dengue, parvovirus B19, serum sickness, syphilis, systemic lupus erythematous, toxic shock syndrome, enterovirus infection, human herpes virus 6 (roseola), viral hemorrhagic fever, drug eruption, infectious mononucleosis, Rocky Mountain spotted fever, rubella, scarlet fever, chikungunya, and Zika virus infection.

Measles complications can affect nearly every organ system (Table). Rates of complications from measles infection depend on age and underlying condition. Coexisting vitamin A deficiency increases complication rates.²⁰

Bacterial infections in the setting of measles infection are more common in adults than in children, and are more severe among people who are malnourished or have an immunodeficiency disorder. The most common infectious complications, which involve the respiratory tract, include pneumonia, laryngotracheitis (“measles croup”), bronchitis, otitis media (most common complication among children in the United States), and sinusitis.^7,13,21

Indications for hospitalizing children include respiratory distress, laryngeal obstruction, dehydration that requires intravenous fluids, diarrhea with more than 10 stools a day or bloody stool, severe anemia, altered mental status, convulsion, severe rash with developing hemorrhagic areas, extensive mouth ulcers, corneal clouding or ulcers, visual disturbance, and mastoiditis.²²

Pneumonia is a common indication for hospitalizing adults.^23,24 Measles-associated interstitial giant cell (Hecht) pneumonia is most often recognized among immunocompromised and malnourished patients.¹³ Primary pneumonia is caused by the measles virus, but bacterial superinfection can occur. The most common bacterial pathogens include Streptococcus, Pneumococcus, and Staphylococcus,^13,24 and less commonly isolated organisms include gram-negative bacteria, such as Haemophilus influenzae, Pseudomonas aeruginosa, Neisseria meningitides, and Enterobacter cloacae.²³

Uncommon complications of measles are myocarditis, glomerulonephritis, acute renal failure, and thrombocytopenic purpura.^25,26

Neurologic complications in measles are an important concern. Measles-associated central nervous system complications are considered a result of an immune-mediated reaction to myelin protein and not from direct viral insult.^26-28 Immunocompromised patients are at risk for developing fatal encephalitis, and those who survive often experience cognitive decline or seizures.

Measles is associated with four different encephalitic diseases: primary measles encephalitis, acute post-measles encephalomyelitis, measles inclusion body encephalitis, and subacute sclerosing panencephalitis.

Primary measles encephalitis is characterized by fever, headache, stiff neck, and meningeal signs. Onset occurs between 1 and 15 days after rash onset, and the disease affects 1/1000 patients. Seizure, altered mental status, and coma can also develop. Viral RNA detection in the cerebrospinal fluid (CSF) confirms the diagnosis.²⁹Acute post-measles encephalomyelitis is more common in adults than in children.¹² It typically develops after the rash fades and the other symptoms subside. Patients suddenly experience a recurrence of fevers or seizures. Deafness, intellectual decline, epilepsy, postencephalitic hyperkinesia, hemiplegia, and/or paraplegia also can develop.^27-29

Measles inclusion body encephalitis is described only in immunocompromised patients, and onset occurs within 1 year of infection. Seizures are an initial and common symptom, and some patients also experience hemiplegia, stupor, hypertonia, and dysarthria.²⁹ Diagnostic findings include seroconversion during the disease course, improvement after withholding of the immunosuppressive regimen, and normal CSF. Brain biopsy confirms the diagnosis.

Subacute sclerosing panencephalitis (SSPE) is a slowly progressing and untreatable degenerative neurologic disorder characterized by demyelination of multiple brain areas. SSPE develops 7 to 10 years after natural measles infection, and usually affects children or adolescents. Clinical presentation includes intellectual decline, frequent rhythmic myoclonic jerks, seizure, and dementia. As the disease progresses, coma, quadriplegia, vegetative state, and autonomic instability develop. Death usually occurs within 2 years of onset.^30,31 In children, the risk for SSPE after measles infection is estimated to be 4 to 11 per 100,000 infections. After the 1989-1991 resurgence of measles in the United States, however, the risk for SSPE was estimated to be 22 per 100,000 infections.^30-32 The pathogenesis of SSPE is not fully understood but is thought to result from persistent aberrant measles virus infection.³²

The SSPE diagnosis is based on clinical presentation, presence of anti-measles antibodies in CSF, typical electroencephalography pattern (periodic paroxysmal bursts) with accompanying myoclonus, tissue analysis, and magnetic resonance imaging.³⁰

Suspicion for measles should prompt immediate consultation with local or state public health officials. Laboratory testing can be carefully considered after consultation, and care is needed in interpreting serologic studies.

The mainstays of measles infection diagnosis are detection of viral RNA by reverse transcriptase–polymerase chain reaction, or isolation of the virus in the clinical specimen, and detection of measles-specific IgM (immunoglobulin M) antibodies. A detailed protocol for collecting specimens for viral isolation appears on the Centers for Disease Control and Prevention website (http://www.cdc.gov/measles/lab-tools/rt-pcr.html).

IgM antibodies are detectable over the 15 weeks after rash onset, but the recommendation is to collect serum between 72 hours and 4 weeks after rash onset.³³ Clinicians should be aware that false-positive IgM results may occur with rheumatologic diseases, parvovirus B19 infection, rubella, and infectious mononucleosis.

IgG (immunoglobulin G) antibodies are usually detectable a week after rash onset. The laboratory can confirm measles by detecting more than a 4-fold increase in IgG titers between the acute phase and the convalescent phase. After measles infection, most adults develop lifelong immunity with positive IgG serology.³⁴

Additional tests, such as IgG avidity and plaque reduction neutralization assay, can be used to confirm suspected cases in previously vaccinated individuals.³⁴

General Principles

Uncomplicated measles treatment is supportive and includes oral fluids and antipyretics.^7,22 Severe bacterial infections, encephalitis, or dehydration may require hospitalization, and in these cases infectious disease consultation is recommended. Patients with pneumonia, purulent otitis media, or tonsillitis should be treated with antibiotics.³⁵ Observational data suggest antibiotics may reduce the occurrence of bacterial infection in children, but there are no usage guidelines.³⁵ Vitamin A supplementation has been associated with a 50% decrease in morbidity and mortality and with blindness prevention.²² This supplementation should be considered in severe measles cases (all hospitalized patients), especially for children, regardless of country of residence, and for adult patients who exhibit clinical signs of vitamin A deficiency.^22,24

Antiviral Treatment

No specific treatment is available.³⁶ Ribavirin demonstrates in vitro activity against the virus, but the Food and Drug Administration has not approved the drug for treatment of measles. Ribavirin has been used for cases of severe measles, and for patients with SSPE along with intrathecal interferon alpha. This antiviral treatment is considered experimental.³⁷

All patients hospitalized with measles infection should be cautioned about the potential downstream complications of the disease and should follow up with their primary care physician for surveillance after discharge.³⁸

If measles symptoms develop, patients should self-quarantine and contact their primary care physician or public health department as soon as possible. Regardless of immune status, family members and other exposed persons should be educated about the measles symptoms that may occur during the 21 days after exposure.³⁸

Both suspected and confirmed cases of measles should be reported immediately to local public health authorities.

Infection Control and Prophylaxis

Current guidelines recommend 2 doses of measles-containing vaccine to all adults at higher risk for contracting measles: international travelers, healthcare personnel, and high school and college students. Infants 6 or 11 months old should receive 1 MMR dose before international travel.^1,38

Strict airborne isolation—use of N95 respirator or respirator with similar effectiveness in preventing airborne transmission—is mandatory from 3 to 5 days before rash onset to 4 days after rash onset (immunocompetent patients) or for the duration of the disease (immunocompromised patients).³⁸

Healthcare workers should have documented presumptive evidence of immunity to measles.³⁹ Healthcare providers without evidence of immunity should be excused from work from day 5 to day 21 of exposure, even if they have received postexposure vaccine or intramuscular immunoglobulin. They should be offered the first MMR dose within 72 hours of measles exposure to prevent or modify the disease. Susceptible family members or visitors should not be allowed in the patient’s room.¹

Postexposure Prophylaxis

Standard MMR vaccination within 72 hours after exposure may protect against disease in people without a contraindication to measles vaccine. The public health department usually identifies these individuals and provides postexposure prophylaxis recommendations.^38,39

People with HIV, patients receiving immunosuppressive therapy, and pregnant women and infants who have been exposed to measles and who are at risk for developing morbid disease can be treated with immunoglobulin (IG). If administered within 6 days of exposure, IG can prevent or modify disease in people who are unvaccinated or severely immunocompromised (ie, not immune). The recommended dose of IG administered intramuscularly is 0.5 mL/kg of body weight (maximum, 15 mL), and the recommended dose of IG given intravenously is 400 mg/kg. Anyone heavier than 30 kg would require intravenous IG to achieve adequate antibody levels.

Physicians should not vaccinate pregnant women, patients with severe immunosuppression from disease or therapy, patients with moderate or severe illness, and people with a history of severe allergic reaction to the vaccine.^1,40 The measles vaccine should be deferred for 6 months after IG administration.³⁶ More details are available in the recommendations made by the Advisory Committee on Immunization Practices.¹

Although rare in the United States, measles remains a common and potentially devastating infection among patients who have not been vaccinated. Diagnosis requires clinical suspicion, engagement of public health authorities, and judicious use of laboratory testing. Hospitalists may encounter infectious and neurologic complications of measles long after the initial infection and should be aware of these associations.

Disclosure

Nothing to report.

Measles is a highly contagious acute respiratory illness that includes a characteristic rash. After exposure, up to 90% of susceptible persons develop measles.¹ Even though it is considered a childhood illness, measles can affect people of all age groups. Measles continues to be major health problem around the world, despite the availability of a safe and effective vaccine, and it remains one of the leading causes of childhood mortality, with nearly 115,000 deaths reported by the World Health Organization² in 2014. In 2000, measles was declared eliminated from the United States, but outbreaks still occasionally occur.^3-6

The disease is self-limited, but some patients develop complications that may require hospitalization for treatment. People at highest risk for complications are children younger than 5 years, adults older than 20 years, pregnant women, and immunocompromised individuals.⁷

During the licensure of live measles vaccine in 1963, an average of 549,000 measles cases and 495 measles deaths, as well as 48,000 hospitalizations and 4000 encephalitis cases, were reported annually in the United States. Almost all Americans were affected by measles by adolescence.

Implementation of the 1-dose vaccine program substantially reduced reported incidence in the United States by 1988, and led to a dramatic decline in measles-related hospitalizations and deaths.^3-6 The 2-dose MMR (measles, mumps, rubella) vaccination was introduced in 1989, and measles was declared eliminated in the United States in 2000.^3-6

National–level one-dose MMR coverage among children 19-35 months has remained above 90% during the last two decades.⁸ NIS-Teen vaccination coverage data for 13- to 17-year-olds since 2008 has been near or above 90%,⁹ and 94% of children enrolled in kindergarten had evidence of 2 MMR doses in the 2014-2015 school year.¹⁰

A large multistate measles outbreak was reported in the United States in 2014-2015.^4,11 One hundred fifty-nine cases were reported in the United States between January 4 and April 5, 2015. The majority of patients either were unvaccinated (45%) or had an unknown vaccination status (38%). Age ranged from 6 weeks to 70 years, and 22 patients (14%) were hospitalized.⁴

Measles is caused by an RNA-containing paramyxovirus that is spread by the respiratory route. Average incubation period from exposure to rash onset is 14 days (range, 7-21 days).^12,13 Peak infectivity occurs during the prodromal phase, before rash onset (Figure 1), but patients are infectious from 4 days before rash onset through 4 days after rash onset.^7,12,13

The disease prodrome consists of a high fever (39°C-40.5°C), coryza, cough, and conjunctivitis followed by Koplik spots (Figure 2A). Koplik spots are pathognomonic for measles but rarely discovered. They appear before the skin rash alongside second molars on the buccal surface of the cheeks. The spots usually disappear when the characteristic maculopapular, nonpruritic rash erupts initially at the hairline and behind the ears, and within four days progresses toward the trunk and limbs, including the palms and soles (Figures 2B, 2C).

The presenting symptoms of primary measles infection are nonspecific, particularly if Koplik spots are not identified. The differential diagnosis for a patient who presents with high fever and rash include Kawasaki disease, dengue, parvovirus B19, serum sickness, syphilis, systemic lupus erythematous, toxic shock syndrome, enterovirus infection, human herpes virus 6 (roseola), viral hemorrhagic fever, drug eruption, infectious mononucleosis, Rocky Mountain spotted fever, rubella, scarlet fever, chikungunya, and Zika virus infection.

Measles complications can affect nearly every organ system (Table). Rates of complications from measles infection depend on age and underlying condition. Coexisting vitamin A deficiency increases complication rates.²⁰

Bacterial infections in the setting of measles infection are more common in adults than in children, and are more severe among people who are malnourished or have an immunodeficiency disorder. The most common infectious complications, which involve the respiratory tract, include pneumonia, laryngotracheitis (“measles croup”), bronchitis, otitis media (most common complication among children in the United States), and sinusitis.^7,13,21

Indications for hospitalizing children include respiratory distress, laryngeal obstruction, dehydration that requires intravenous fluids, diarrhea with more than 10 stools a day or bloody stool, severe anemia, altered mental status, convulsion, severe rash with developing hemorrhagic areas, extensive mouth ulcers, corneal clouding or ulcers, visual disturbance, and mastoiditis.²²

Pneumonia is a common indication for hospitalizing adults.^23,24 Measles-associated interstitial giant cell (Hecht) pneumonia is most often recognized among immunocompromised and malnourished patients.¹³ Primary pneumonia is caused by the measles virus, but bacterial superinfection can occur. The most common bacterial pathogens include Streptococcus, Pneumococcus, and Staphylococcus,^13,24 and less commonly isolated organisms include gram-negative bacteria, such as Haemophilus influenzae, Pseudomonas aeruginosa, Neisseria meningitides, and Enterobacter cloacae.²³

Uncommon complications of measles are myocarditis, glomerulonephritis, acute renal failure, and thrombocytopenic purpura.^25,26

Neurologic complications in measles are an important concern. Measles-associated central nervous system complications are considered a result of an immune-mediated reaction to myelin protein and not from direct viral insult.^26-28 Immunocompromised patients are at risk for developing fatal encephalitis, and those who survive often experience cognitive decline or seizures.

Measles is associated with four different encephalitic diseases: primary measles encephalitis, acute post-measles encephalomyelitis, measles inclusion body encephalitis, and subacute sclerosing panencephalitis.

Primary measles encephalitis is characterized by fever, headache, stiff neck, and meningeal signs. Onset occurs between 1 and 15 days after rash onset, and the disease affects 1/1000 patients. Seizure, altered mental status, and coma can also develop. Viral RNA detection in the cerebrospinal fluid (CSF) confirms the diagnosis.²⁹Acute post-measles encephalomyelitis is more common in adults than in children.¹² It typically develops after the rash fades and the other symptoms subside. Patients suddenly experience a recurrence of fevers or seizures. Deafness, intellectual decline, epilepsy, postencephalitic hyperkinesia, hemiplegia, and/or paraplegia also can develop.^27-29

Measles inclusion body encephalitis is described only in immunocompromised patients, and onset occurs within 1 year of infection. Seizures are an initial and common symptom, and some patients also experience hemiplegia, stupor, hypertonia, and dysarthria.²⁹ Diagnostic findings include seroconversion during the disease course, improvement after withholding of the immunosuppressive regimen, and normal CSF. Brain biopsy confirms the diagnosis.

Subacute sclerosing panencephalitis (SSPE) is a slowly progressing and untreatable degenerative neurologic disorder characterized by demyelination of multiple brain areas. SSPE develops 7 to 10 years after natural measles infection, and usually affects children or adolescents. Clinical presentation includes intellectual decline, frequent rhythmic myoclonic jerks, seizure, and dementia. As the disease progresses, coma, quadriplegia, vegetative state, and autonomic instability develop. Death usually occurs within 2 years of onset.^30,31 In children, the risk for SSPE after measles infection is estimated to be 4 to 11 per 100,000 infections. After the 1989-1991 resurgence of measles in the United States, however, the risk for SSPE was estimated to be 22 per 100,000 infections.^30-32 The pathogenesis of SSPE is not fully understood but is thought to result from persistent aberrant measles virus infection.³²

The SSPE diagnosis is based on clinical presentation, presence of anti-measles antibodies in CSF, typical electroencephalography pattern (periodic paroxysmal bursts) with accompanying myoclonus, tissue analysis, and magnetic resonance imaging.³⁰

Suspicion for measles should prompt immediate consultation with local or state public health officials. Laboratory testing can be carefully considered after consultation, and care is needed in interpreting serologic studies.

The mainstays of measles infection diagnosis are detection of viral RNA by reverse transcriptase–polymerase chain reaction, or isolation of the virus in the clinical specimen, and detection of measles-specific IgM (immunoglobulin M) antibodies. A detailed protocol for collecting specimens for viral isolation appears on the Centers for Disease Control and Prevention website (http://www.cdc.gov/measles/lab-tools/rt-pcr.html).

IgM antibodies are detectable over the 15 weeks after rash onset, but the recommendation is to collect serum between 72 hours and 4 weeks after rash onset.³³ Clinicians should be aware that false-positive IgM results may occur with rheumatologic diseases, parvovirus B19 infection, rubella, and infectious mononucleosis.

IgG (immunoglobulin G) antibodies are usually detectable a week after rash onset. The laboratory can confirm measles by detecting more than a 4-fold increase in IgG titers between the acute phase and the convalescent phase. After measles infection, most adults develop lifelong immunity with positive IgG serology.³⁴

Additional tests, such as IgG avidity and plaque reduction neutralization assay, can be used to confirm suspected cases in previously vaccinated individuals.³⁴

General Principles

Uncomplicated measles treatment is supportive and includes oral fluids and antipyretics.^7,22 Severe bacterial infections, encephalitis, or dehydration may require hospitalization, and in these cases infectious disease consultation is recommended. Patients with pneumonia, purulent otitis media, or tonsillitis should be treated with antibiotics.³⁵ Observational data suggest antibiotics may reduce the occurrence of bacterial infection in children, but there are no usage guidelines.³⁵ Vitamin A supplementation has been associated with a 50% decrease in morbidity and mortality and with blindness prevention.²² This supplementation should be considered in severe measles cases (all hospitalized patients), especially for children, regardless of country of residence, and for adult patients who exhibit clinical signs of vitamin A deficiency.^22,24

Antiviral Treatment

No specific treatment is available.³⁶ Ribavirin demonstrates in vitro activity against the virus, but the Food and Drug Administration has not approved the drug for treatment of measles. Ribavirin has been used for cases of severe measles, and for patients with SSPE along with intrathecal interferon alpha. This antiviral treatment is considered experimental.³⁷

All patients hospitalized with measles infection should be cautioned about the potential downstream complications of the disease and should follow up with their primary care physician for surveillance after discharge.³⁸

If measles symptoms develop, patients should self-quarantine and contact their primary care physician or public health department as soon as possible. Regardless of immune status, family members and other exposed persons should be educated about the measles symptoms that may occur during the 21 days after exposure.³⁸

Both suspected and confirmed cases of measles should be reported immediately to local public health authorities.

Infection Control and Prophylaxis

Current guidelines recommend 2 doses of measles-containing vaccine to all adults at higher risk for contracting measles: international travelers, healthcare personnel, and high school and college students. Infants 6 or 11 months old should receive 1 MMR dose before international travel.^1,38

Strict airborne isolation—use of N95 respirator or respirator with similar effectiveness in preventing airborne transmission—is mandatory from 3 to 5 days before rash onset to 4 days after rash onset (immunocompetent patients) or for the duration of the disease (immunocompromised patients).³⁸

Healthcare workers should have documented presumptive evidence of immunity to measles.³⁹ Healthcare providers without evidence of immunity should be excused from work from day 5 to day 21 of exposure, even if they have received postexposure vaccine or intramuscular immunoglobulin. They should be offered the first MMR dose within 72 hours of measles exposure to prevent or modify the disease. Susceptible family members or visitors should not be allowed in the patient’s room.¹

Postexposure Prophylaxis

Standard MMR vaccination within 72 hours after exposure may protect against disease in people without a contraindication to measles vaccine. The public health department usually identifies these individuals and provides postexposure prophylaxis recommendations.^38,39

People with HIV, patients receiving immunosuppressive therapy, and pregnant women and infants who have been exposed to measles and who are at risk for developing morbid disease can be treated with immunoglobulin (IG). If administered within 6 days of exposure, IG can prevent or modify disease in people who are unvaccinated or severely immunocompromised (ie, not immune). The recommended dose of IG administered intramuscularly is 0.5 mL/kg of body weight (maximum, 15 mL), and the recommended dose of IG given intravenously is 400 mg/kg. Anyone heavier than 30 kg would require intravenous IG to achieve adequate antibody levels.

Physicians should not vaccinate pregnant women, patients with severe immunosuppression from disease or therapy, patients with moderate or severe illness, and people with a history of severe allergic reaction to the vaccine.^1,40 The measles vaccine should be deferred for 6 months after IG administration.³⁶ More details are available in the recommendations made by the Advisory Committee on Immunization Practices.¹

Although rare in the United States, measles remains a common and potentially devastating infection among patients who have not been vaccinated. Diagnosis requires clinical suspicion, engagement of public health authorities, and judicious use of laboratory testing. Hospitalists may encounter infectious and neurologic complications of measles long after the initial infection and should be aware of these associations.

Disclosure

Nothing to report.