Hospital Medicine

There is little doubt that preventing 30-day readmissions to the hospital results in lower costs for payers. However, reducing costs alone does not make this metric a measure of “high value” care.¹ Rather, it is the improvement in the effectiveness of the discharge process that occurs alongside lower costs that makes readmission reduction efforts “high value” – or a “win-win” for patients and payers.

However, the article by Nuckols and colleagues in this month’s issue of the Journal of Hospital Medicine (JHM) suggests that it might not be that simple and adds nuance to the ongoing discussion about the 30-day readmission metric.² The study used data collected by the federal government to examine changes not only in 30-day readmission rates between 2009-2010 and 2013-2014 but also changes in emergency department (ED) and observation unit visits. What they found is important. In general, despite reductions in 30-day readmissions for patients served by Medicare and private insurance, there were increases in observation unit and ED visits across all payer types (including Medicare and private insurance). These increases in observation unit and ED visits resulted in statistically higher overall “revisit” rates for the uninsured and those insured by Medicaid and offset any improvements in the “revisit” rates resulting from reductions in 30-day readmissions for those with private insurance. Those insured by Medicare—representing about 300,000 of the 420,000 visits analyzed—still had a statistically lower “revisit” rate, but it was only marginally lower (25.0% in 2013-2014 versus 25.3% in 2009-2010).²

The generalizability of the Nuckols’ study was limited in that it examined only index admissions for acute myocardial infarction (AMI), heart failure (HF), and pneumonia and used data from only Georgia, Nebraska, South Carolina, and Tennessee—the four states where observation and ED visit data were available in the federal database.² The study also did not examine hospital-level revisit data; hence, it was not able to determine if hospitals with greater reductions in readmission rates had greater increases in observation or ED visits, as one might predict. Despite these limitations, the rigor of the study was noteworthy. The authors used matching techniques to ensure that the populations examined in the two time periods were comparable. Unlike previous research,^3,4 they also used a comprehensive definition of a hospital “revisit” (including both observation and ED visits) and measured “revisit” rates across several payer types, rather than focusing exclusively on those covered by fee for service Medicare, as in past studies.^4,5

What the study by Nuckols and colleagues suggests is that even though patients may be readmitted less, they may be coming back to the ED or getting admitted to the observation unit more, resulting in overall “revisit” rates that are marginally lower for Medicare patients, but often the same or even higher for other payer groups, particularly disadvantaged payer groups who are uninsured or insured by Medicaid.² Although the authors do not assert causality for these trends, it is worth noting that the much-discussed Hospital Readmission Reduction Program (or “readmission penalty”) applies only to Medicare patients aged more than 65 years. It is likely that this program influenced the differences identified between payer groups in this article.

Beyond the policy implications of these findings, the experience of patients cared for in these different settings is of paramount importance. Unfortunately, there are limited data comparing patient perceptions, preferences, or outcomes resulting from readmission to an inpatient service versus an observation unit or ED visit within 30 days of discharge. However, there is reason to believe that costs could be higher for some patients treated in the ED or an observation unit as compared to those in the inpatient setting,⁶ and that care continuity and quality may be different across these settings. In a recent white paper on observation care published by the Society of Hospital Medicine (SHM) Public Policy Committee,⁷ the SHM reported the results of a 2017 survey of its members about observation care. The results were concerning. An overwhelming majority of respondents (87%) believed that the rules for observation are unclear for patients, and 68% of respondents believed that policy changes mandating informing patients of their observation status have created conflict between the provider and the patient.⁷ As shared by one respondent, “the observation issue can severely damage the therapeutic bond with patient/family, who may conclude that the hospitalist has more interest in saving someone money at the expense of patient care.”⁷ Thus, there is significant concern about the nature of observation stays and the experience for patients and providers. We should take care to better understand these experiences given that readmission reduction efforts may funnel more patients into observation care.

As a next step, we recommend further examination of how “revisit” rates have changed over time for patients with any discharge diagnosis, and not just those with pneumonia, AMI, or HF.⁸ Such examinations should be stratified by payer to identify differential impacts on those with lower socioeconomic status. Analyses should also examine changes in “revisit” types at the hospital level to better understand if hospitals with reductions in readmission rates are simply shifting revisits to the observation unit or ED. It is possible that inpatient readmissions for any given hospital are decreasing without concomitant increases in observation visits, as there are forces independent of the readmission penalty, such as the Recovery Audit Contractor program, that are driving hospitals to more frequently code patients as observation visits rather than inpatient admissions.⁹ Thus, readmissions could decrease and observation unit visits could increase independent of one another. We also recommend further research to examine differences in care quality, clinical outcomes, and costs for those readmitted to the hospital within 30 days of discharge versus those cared for in observation units or the ED. The challenge of such studies will be to identify and examine comparable populations of patients across these three settings. Examining patient perceptions and preferences across these settings is also critical. Finally, when assessing interventions to reduce inpatient readmissions, we need to consider “revisits” as a whole, not simply readmissions.¹⁰ Otherwise, we may simply be promoting the use of interventions that shift inpatient readmissions to observation unit or ED revisits, and there is little that is patient-centered or high value about that.⁹

Disclosures

The authors have nothing to disclose.

There is little doubt that preventing 30-day readmissions to the hospital results in lower costs for payers. However, reducing costs alone does not make this metric a measure of “high value” care.¹ Rather, it is the improvement in the effectiveness of the discharge process that occurs alongside lower costs that makes readmission reduction efforts “high value” – or a “win-win” for patients and payers.

However, the article by Nuckols and colleagues in this month’s issue of the Journal of Hospital Medicine (JHM) suggests that it might not be that simple and adds nuance to the ongoing discussion about the 30-day readmission metric.² The study used data collected by the federal government to examine changes not only in 30-day readmission rates between 2009-2010 and 2013-2014 but also changes in emergency department (ED) and observation unit visits. What they found is important. In general, despite reductions in 30-day readmissions for patients served by Medicare and private insurance, there were increases in observation unit and ED visits across all payer types (including Medicare and private insurance). These increases in observation unit and ED visits resulted in statistically higher overall “revisit” rates for the uninsured and those insured by Medicaid and offset any improvements in the “revisit” rates resulting from reductions in 30-day readmissions for those with private insurance. Those insured by Medicare—representing about 300,000 of the 420,000 visits analyzed—still had a statistically lower “revisit” rate, but it was only marginally lower (25.0% in 2013-2014 versus 25.3% in 2009-2010).²

The generalizability of the Nuckols’ study was limited in that it examined only index admissions for acute myocardial infarction (AMI), heart failure (HF), and pneumonia and used data from only Georgia, Nebraska, South Carolina, and Tennessee—the four states where observation and ED visit data were available in the federal database.² The study also did not examine hospital-level revisit data; hence, it was not able to determine if hospitals with greater reductions in readmission rates had greater increases in observation or ED visits, as one might predict. Despite these limitations, the rigor of the study was noteworthy. The authors used matching techniques to ensure that the populations examined in the two time periods were comparable. Unlike previous research,^3,4 they also used a comprehensive definition of a hospital “revisit” (including both observation and ED visits) and measured “revisit” rates across several payer types, rather than focusing exclusively on those covered by fee for service Medicare, as in past studies.^4,5

What the study by Nuckols and colleagues suggests is that even though patients may be readmitted less, they may be coming back to the ED or getting admitted to the observation unit more, resulting in overall “revisit” rates that are marginally lower for Medicare patients, but often the same or even higher for other payer groups, particularly disadvantaged payer groups who are uninsured or insured by Medicaid.² Although the authors do not assert causality for these trends, it is worth noting that the much-discussed Hospital Readmission Reduction Program (or “readmission penalty”) applies only to Medicare patients aged more than 65 years. It is likely that this program influenced the differences identified between payer groups in this article.

Beyond the policy implications of these findings, the experience of patients cared for in these different settings is of paramount importance. Unfortunately, there are limited data comparing patient perceptions, preferences, or outcomes resulting from readmission to an inpatient service versus an observation unit or ED visit within 30 days of discharge. However, there is reason to believe that costs could be higher for some patients treated in the ED or an observation unit as compared to those in the inpatient setting,⁶ and that care continuity and quality may be different across these settings. In a recent white paper on observation care published by the Society of Hospital Medicine (SHM) Public Policy Committee,⁷ the SHM reported the results of a 2017 survey of its members about observation care. The results were concerning. An overwhelming majority of respondents (87%) believed that the rules for observation are unclear for patients, and 68% of respondents believed that policy changes mandating informing patients of their observation status have created conflict between the provider and the patient.⁷ As shared by one respondent, “the observation issue can severely damage the therapeutic bond with patient/family, who may conclude that the hospitalist has more interest in saving someone money at the expense of patient care.”⁷ Thus, there is significant concern about the nature of observation stays and the experience for patients and providers. We should take care to better understand these experiences given that readmission reduction efforts may funnel more patients into observation care.

As a next step, we recommend further examination of how “revisit” rates have changed over time for patients with any discharge diagnosis, and not just those with pneumonia, AMI, or HF.⁸ Such examinations should be stratified by payer to identify differential impacts on those with lower socioeconomic status. Analyses should also examine changes in “revisit” types at the hospital level to better understand if hospitals with reductions in readmission rates are simply shifting revisits to the observation unit or ED. It is possible that inpatient readmissions for any given hospital are decreasing without concomitant increases in observation visits, as there are forces independent of the readmission penalty, such as the Recovery Audit Contractor program, that are driving hospitals to more frequently code patients as observation visits rather than inpatient admissions.⁹ Thus, readmissions could decrease and observation unit visits could increase independent of one another. We also recommend further research to examine differences in care quality, clinical outcomes, and costs for those readmitted to the hospital within 30 days of discharge versus those cared for in observation units or the ED. The challenge of such studies will be to identify and examine comparable populations of patients across these three settings. Examining patient perceptions and preferences across these settings is also critical. Finally, when assessing interventions to reduce inpatient readmissions, we need to consider “revisits” as a whole, not simply readmissions.¹⁰ Otherwise, we may simply be promoting the use of interventions that shift inpatient readmissions to observation unit or ED revisits, and there is little that is patient-centered or high value about that.⁹

Disclosures

The authors have nothing to disclose.

There is little doubt that preventing 30-day readmissions to the hospital results in lower costs for payers. However, reducing costs alone does not make this metric a measure of “high value” care.¹ Rather, it is the improvement in the effectiveness of the discharge process that occurs alongside lower costs that makes readmission reduction efforts “high value” – or a “win-win” for patients and payers.

However, the article by Nuckols and colleagues in this month’s issue of the Journal of Hospital Medicine (JHM) suggests that it might not be that simple and adds nuance to the ongoing discussion about the 30-day readmission metric.² The study used data collected by the federal government to examine changes not only in 30-day readmission rates between 2009-2010 and 2013-2014 but also changes in emergency department (ED) and observation unit visits. What they found is important. In general, despite reductions in 30-day readmissions for patients served by Medicare and private insurance, there were increases in observation unit and ED visits across all payer types (including Medicare and private insurance). These increases in observation unit and ED visits resulted in statistically higher overall “revisit” rates for the uninsured and those insured by Medicaid and offset any improvements in the “revisit” rates resulting from reductions in 30-day readmissions for those with private insurance. Those insured by Medicare—representing about 300,000 of the 420,000 visits analyzed—still had a statistically lower “revisit” rate, but it was only marginally lower (25.0% in 2013-2014 versus 25.3% in 2009-2010).²

The generalizability of the Nuckols’ study was limited in that it examined only index admissions for acute myocardial infarction (AMI), heart failure (HF), and pneumonia and used data from only Georgia, Nebraska, South Carolina, and Tennessee—the four states where observation and ED visit data were available in the federal database.² The study also did not examine hospital-level revisit data; hence, it was not able to determine if hospitals with greater reductions in readmission rates had greater increases in observation or ED visits, as one might predict. Despite these limitations, the rigor of the study was noteworthy. The authors used matching techniques to ensure that the populations examined in the two time periods were comparable. Unlike previous research,^3,4 they also used a comprehensive definition of a hospital “revisit” (including both observation and ED visits) and measured “revisit” rates across several payer types, rather than focusing exclusively on those covered by fee for service Medicare, as in past studies.^4,5

What the study by Nuckols and colleagues suggests is that even though patients may be readmitted less, they may be coming back to the ED or getting admitted to the observation unit more, resulting in overall “revisit” rates that are marginally lower for Medicare patients, but often the same or even higher for other payer groups, particularly disadvantaged payer groups who are uninsured or insured by Medicaid.² Although the authors do not assert causality for these trends, it is worth noting that the much-discussed Hospital Readmission Reduction Program (or “readmission penalty”) applies only to Medicare patients aged more than 65 years. It is likely that this program influenced the differences identified between payer groups in this article.

Beyond the policy implications of these findings, the experience of patients cared for in these different settings is of paramount importance. Unfortunately, there are limited data comparing patient perceptions, preferences, or outcomes resulting from readmission to an inpatient service versus an observation unit or ED visit within 30 days of discharge. However, there is reason to believe that costs could be higher for some patients treated in the ED or an observation unit as compared to those in the inpatient setting,⁶ and that care continuity and quality may be different across these settings. In a recent white paper on observation care published by the Society of Hospital Medicine (SHM) Public Policy Committee,⁷ the SHM reported the results of a 2017 survey of its members about observation care. The results were concerning. An overwhelming majority of respondents (87%) believed that the rules for observation are unclear for patients, and 68% of respondents believed that policy changes mandating informing patients of their observation status have created conflict between the provider and the patient.⁷ As shared by one respondent, “the observation issue can severely damage the therapeutic bond with patient/family, who may conclude that the hospitalist has more interest in saving someone money at the expense of patient care.”⁷ Thus, there is significant concern about the nature of observation stays and the experience for patients and providers. We should take care to better understand these experiences given that readmission reduction efforts may funnel more patients into observation care.

As a next step, we recommend further examination of how “revisit” rates have changed over time for patients with any discharge diagnosis, and not just those with pneumonia, AMI, or HF.⁸ Such examinations should be stratified by payer to identify differential impacts on those with lower socioeconomic status. Analyses should also examine changes in “revisit” types at the hospital level to better understand if hospitals with reductions in readmission rates are simply shifting revisits to the observation unit or ED. It is possible that inpatient readmissions for any given hospital are decreasing without concomitant increases in observation visits, as there are forces independent of the readmission penalty, such as the Recovery Audit Contractor program, that are driving hospitals to more frequently code patients as observation visits rather than inpatient admissions.⁹ Thus, readmissions could decrease and observation unit visits could increase independent of one another. We also recommend further research to examine differences in care quality, clinical outcomes, and costs for those readmitted to the hospital within 30 days of discharge versus those cared for in observation units or the ED. The challenge of such studies will be to identify and examine comparable populations of patients across these three settings. Examining patient perceptions and preferences across these settings is also critical. Finally, when assessing interventions to reduce inpatient readmissions, we need to consider “revisits” as a whole, not simply readmissions.¹⁰ Otherwise, we may simply be promoting the use of interventions that shift inpatient readmissions to observation unit or ED revisits, and there is little that is patient-centered or high value about that.⁹

Disclosures

The authors have nothing to disclose.

Evidence has emerged over the last decade of the importance of the front line patient care team in improving quality and safety of patient care.^1-3 Improving collaboration and workflow is thought to increase reliability of care delivery.¹ One promising method to improve collaboration is the interdisciplinary ward round (IDR), whereby medical, nursing, and allied health staff attend ward rounds together. IDRs have been shown to reduce the average cost and length of hospital stay,^4,5 although a recent systematic review found inconsistent improvements across studies.⁶ Using the term “interdisciplinary,” however, does not necessarily imply the inclusion of all disciplines necessary for patient care. The challenge of conducting interdisciplinary rounds is considerable in today’s busy clinical environment: health professionals who are spread across multiple locations within the hospital, and who have competing hospital responsibilities and priorities, must come together at the same time and for a set period each day. A survey with respondents from Australia, the United States, and Canada found that only 65% of rounds labelled “interdisciplinary” included a physician.⁷

While IDRs are not new, structured IDRs involve the purposeful inclusion of all disciplinary groups relevant to a patient’s care, alongside a checklist tool to aid comprehensive but concise daily assessment of progress and treatment planning. Novel, structured IDR interventions have been tested recently in various settings, resulting in improved teamwork, hospital performance, and patient outcomes in the US, including the Structured Interdisciplinary Bedside Round (SIBR) model.^8-12

The aim of this study was to assess the impact of the new structure and the associated practice changes on interprofessional working and a set of key patient and hospital outcome measures. As part of the intervention, the hospital established an Acute Medical Unit (AMU) based on the Accountable Care Unit model.¹³

Description of the Intervention

The AMU brought together 2 existing medical wards, a general medical ward and a 48-hour turnaround Medical Assessment Unit (MAU), into 1 geographical location with 26 beds. Prior to the merger, the MAU and general medical ward had separate and distinct cultures and workflows. The MAU was staffed with experienced nurses; nurses worked within a patient allocation model, the workload was shared, and relationships were collegial. In contrast, the medical ward was more typical of the remainder of the hospital: nurses had a heavy workload, managed a large group of longer-term complex patients, and they used a team-based nursing model of care in which senior nurses supervised junior staff. It was decided that because of the seniority of the MAU staff, they should be in charge of the combined AMU, and the patient allocation model of care would be used to facilitate SIBR.

Consultants, junior doctors, nurses, and allied health professionals (including a pharmacist, physiotherapist, occupational therapist, and social worker) were geographically aligned to the new ward, allowing them to participate as a team in daily structured ward rounds. Rounds are scheduled at the same time each day to enable family participation. The ward round is coordinated by a registrar or intern, with input from patient, family, nursing staff, pharmacy, allied health, and other doctors (intern, registrar, and consultant) based on the unit. The patient load is distributed between 2 rounds: 1 scheduled for 10 am and the other for 11 am each weekday.

Data Collection Strategy

The study was set in an AMU in a large tertiary care hospital in regional Australia and used a convergent parallel multimethod approach¹⁴ to evaluate the implementation and effect of SIBR in the AMU. The study population consisted of 32 clinicians employed at the study hospital: (1) the leadership team involved in the development and implementation of the intervention and (2) members of clinical staff who were part of the AMU team.

Qualitative Data

Qualitative measures consisted of semistructured interviews. We utilized multiple strategies to recruit interviewees, including a snowball technique, criterion sampling,¹⁵ and emergent sampling, so that we could seek the views of both the leadership team responsible for the implementation and “frontline” clinical staff whose daily work was directly affected by it. Everyone who was initially recruited agreed to be interviewed, and additional frontline staff asked to be interviewed once they realized that we were asking about how staff experienced the changes in practice.

The research team developed a semistructured interview guide based on an understanding of the merger of the 2 units as well as an understanding of changes in practice of the rounds (provided in Appendix 1). The questions were pilot tested on a separate unit and revised. Questions were structured into 5 topic areas: planning and implementation of AMU/SIBR model, changes in work practices because of the new model, team functioning, job satisfaction, and perceived impact of the new model on patients and families. All interviews were audio-recorded and transcribed verbatim for analysis.

Quantitative Data

Quantitative data were collected on patient outcome measures: length of stay (LOS), discharge date and time, mode of separation (including death), primary diagnostic category, total hospital stay cost and “clinical response calls,” and patient demographic data (age, gender, and Patient Clinical Complexity Level [PCCL]). The PCCL is a standard measure used in Australian public inpatient facilities and is calculated for each episode of care.¹⁶ It measures the cumulative effect of a patient’s complications and/or comorbidities and takes an integer value between 0 (no clinical complexity effect) and 4 (catastrophic clinical complexity effect).

Data regarding LOS, diagnosis (Australian Refined Diagnosis Related Groups [AR-DRG], version 7), discharge date, and mode of separation (including death) were obtained from the New South Wales Ministry of Health’s Health Information Exchange for patients discharged during the year prior to the intervention through 1 year after the implementation of the intervention. The total hospital stay cost for these individuals was obtained from the local Health Service Organizational Performance Management unit. Inclusion criteria were inpatients aged over 15 years experiencing acute episodes of care; patients with a primary diagnostic category of mental diseases and disorders were excluded. LOS was calculated based on ward stay. AMU data were compared with the remaining hospital ward data (the control group). Data on “clinical response calls” per month per ward were also obtained for the 12 months prior to intervention and the 12 months of the intervention.

Analysis

Qualitative Analysis

Qualitative data analysis consisted of a hybrid form of textual analysis, combining inductive and deductive logics.^17,18 Initially, 3 researchers (J.P., J.J., and R.C.W.) independently coded the interview data inductively to identify themes. Discrepancies were resolved through discussion until consensus was reached. Then, to further facilitate analysis, the researchers deductively imposed a matrix categorization, consisting of 4 a priori categories: context/conditions, practices/processes, professional interactions, and consequences.^19,20 Additional a priori categories were used to sort the themes further in terms of experiences prior to, during, and following implementation of the intervention. To compare changes in those different time periods, we wanted to know what themes were related to implementation and whether those themes continued to be applicable to sustainability of the changes.

Quantitative analysis. Distribution of continuous data was examined by using the one-sample Kolmogorov-Smirnov test. We compared pre-SIBR (baseline) measures using the Student t test for normally distributed data, the Mann-Whitney U z test for nonparametric data (denoted as M-W U z), and χ² tests for categorical data. Changes in monthly “clinical response calls” between the AMU and the control wards over time were explored by using analysis of variance (ANOVA). Changes in LOS and cost of stay from the year prior to the intervention to the first year of the intervention were analyzed by using generalized linear models, which are a form of linear regression. Factors, or independent variables, included in the models were time period (before or during intervention), ward (AMU or control), an interaction term (time by ward), patient age, gender, primary diagnosis (major diagnostic categories of the AR-DRG version 7.0), and acuity (PCCL). The estimated marginal means for cost of stay for the 12-month period prior to the intervention and for the first 12 months of the intervention were produced. All statistical analyses were performed by using IBM SPSS version 21 (IBM Corp., Armonk, New York) and with alpha set at P  < .05.

Qualitative Evaluation of the Intervention

Participants.

Three researchers (RCW, JP, and JJ) conducted in-person, semistructured interviews with 32 clinicians (9 male, 23 female) during a 3-day period. The duration of the interviews ranged from 19 minutes to 68 minutes. Participants consisted of 8 doctors, 18 nurses, 5 allied health professionals, and an administrator. Ten of the participants were involved in the leadership group that drove the planning and implementation of SIBR and the AMU.

Themes

Context and Conditions of Work

Practices and Processes

Participants perceived postintervention work processes to be more efficient. A primary example was a near-universal approval of the time saved from not “chasing” other professionals now that they were predictably available on the ward. More timely decision-making was thought to result from this predicted availability and associated improvements in communication.

The SIBR enforced a workflow on all staff, who felt there was less flexibility to work autonomously (doctors) or according to patients’ needs (nurses). More junior staff expressed anxiety about delayed completion of discharge-related administrative tasks because of the midday completion of the round. Allied health professionals who had commitments in other areas of the hospital often faced a dilemma about how to prioritize SIBR attendance and activities on other wards. This was managed differently depending on the specific allied health profession and the individuals within that profession.

Professional Interactions

In terms of interprofessional dynamics on the AMU, the implementation of SIBR resulted in a shift in power between the doctors and the nurses. In the old ward, doctors largely controlled the timing of medical rounding processes. In the new AMU, doctors had to relinquish some control over the timing of personal workflow to comply with the requirements of SIBR. Furthermore, there was evidence that this had some impact on traditional hierarchical models of communication and created a more level playing field, as nonmedical professionals felt more empowered to voice their thoughts during and outside of rounds.

The rounds provided much greater visibility of the “big picture” and each profession’s role within it; this allowed each clinician to adjust their work to fit in and take account of others. The process was not instantaneous, and trust developed over a period of weeks. Better communication meant fewer misunderstandings, and workload dropped.

The participation of allied health professionals in the round enhanced clinician interprofessional skills and knowledge. The more inclusive approach facilitated greater trust between clinical disciplines and a development of increased confidence among nursing, allied health, and administrative professionals.

In contrast to the positive impacts of the new model of care on communication and relationships within the AMU, interdepartmental relationships were seen to have suffered. The processes and practices of the new AMU are different to those in the other hospital departments, resulting in some isolation of the unit and difficulties interacting with other areas of the hospital. For example, the trade-offs that allied health professionals made to participate in SIBR often came at the expense of other units or departments.

Consequences

All interviewees lauded the benefits of the SIBR intervention for patients. Patients were perceived to be better informed and more respected, and they benefited from greater perceived timeliness of treatment and discharge, easier access to doctors, better continuity of treatment and outcomes, improved nurse knowledge of their circumstances, and fewer gaps in their care. Clinicians spoke directly to the patient during SIBR, rather than consulting with professional colleagues over the patient’s head. Some staff felt that doctors were now thinking of patients as “people” rather than “a set of symptoms.” Nurses discovered that informed patients are easier to manage.

Staff members were prepared to compromise on their own needs in the interests of the patient. The emphasis on the patient during rounds resulted in improved advocacy behaviors of clinicians. The nurses became more empowered and able to show greater initiative. Families appeared to find it much easier to access the doctors and obtain information about the patient, resulting in less distress and a greater sense of control and trust in the process.

Quantitative Evaluation of the Intervention

Hospital Outcomes

Patient demographics did not change over time within either the AMU or control wards. However, there were significant increases in Patient Clinical Complexity Level (PCCL) ratings for both the AMU (44.7% to 40.3%; P<0.05) and the control wards (65.2% to 61.6%; P < .001). There was not a statistically significant shift over time in median LoS on the ward prior to (2.16 days, IQR 3.07) and during SIBR in the AMU (2.15 days; IQR 3.28), while LoS increased in the control (pre-SIBR: 1.67, 2.34; during SIBR 1.73, 2.40; M-W U z = -2.46, P = .014). Mortality rates were stable across time for both the AMU (pre-SIBR 2.6% [95% confidence interval {CI}, 1.9-3.5]; during SIBR 2.8% [95% CI, 2.1-3.7]) and the control (pre-SIBR 1.3% [95% CI, 1.0-1.5]; during SIBR 1.2% [95% CI, 1.0-1.4]).

The total number of “clinical response calls” or “flags” per month dropped significantly from pre-SIBR to during SIBR for the AMU from a mean of 63.1 (standard deviation 15.1) to 31.5 (10.8), but remained relatively stable in the control (pre-SIBR 72.5 [17.6]; during SIBR 74.0 [28.3]), and this difference was statistically significant (F (1,44) = 9.03; P = .004). There was no change in monthly “red flags” or “rapid response calls” over time (AMU: 10.5 [3.6] to 9.1 [4.7]; control: 40.3 [11.7] to 41.8 [10.8]). The change in total “clinical response calls” over time was attributable to the “yellow flags” or the decline in “calls for clinical review” in the AMU (from 52.6 [13.5] to 22.4 [9.2]). The average monthly “yellow flags” remained stable in the control (pre-SIBR 32.2 [11.6]; during SIBR 32.3 [22.4]). The AMU and the control wards differed significantly in how the number of monthly “calls for clinical review” changed from pre-SIBR to during SIBR (F (1,44) = 12.18; P = .001).

The 2 main outcome measures, LOS and costs, were analyzed to determine whether changes over time differed between the AMU and the control wards after accounting for age, gender, and PCCL. There was no statistically significant difference between the AMU and control wards in terms of change in LOS over time (Wald χ² = 1.05; degrees of freedom [df] = 1; P = .31). There was a statistically significant interaction for cost of stay, indicating that ward types differed in how they changed over time (with a drop in cost over time observed in the AMU and an increase observed in the control) (Wald χ² = 6.34; df = 1; P = .012.

We report on the implementation of an AMU model of care, including the reorganization of a nursing unit, implementation of IDR, and geographical localization. Our study design allowed a more comprehensive assessment of the implementation of system redesign to include provider perceptions and clinical outcomes.

The 2 very different cultures of the old wards that were combined into the AMU, as well as the fact that the teams had not previously worked together, made the merger of the 2 wards difficult. Historically, the 2 teams had worked in very different ways, and this created barriers to implementation. The SIBR also demanded new ways of working closely with other disciplines, which disrupted older clinical cultures and relationships. While organizational culture is often discussed, and even measured, the full impact of cultural factors when making workplace changes is frequently underestimated.²¹ The development of a new culture takes time, and it can lag organizational structural changes by months or even years.²² As our interviewees expressed, often emotionally, there was a sense of loss during the merger of the 2 units. While this is a potential consequence of any large organizational change, it could be addressed during the planning stages, prior to implementation, by acknowledging and perhaps honoring what is being left behind. It is safe to assume that future units implementing the rounding intervention will not fully realize commensurate levels of culture change until well after the structural and process changes are finalized, and only then if explicit effort is made to engender cultural change.

Overall, however, the interviewees perceived that the SIBR intervention led to improved teamwork and team functioning. These improvements were thought to benefit task performance and patient safety. Our study is consistent with other research in the literature that reported that greater staff empowerment and commitment is associated with interdisciplinary patient care interventions in front line caregiving teams.^23,24 The perception of a more equal nurse-physician relationship resulted in improved job satisfaction, better interprofessional relationships, and perceived improvements in patient care. A flatter power gradient across professions and increased interdisciplinary teamwork has been shown to be associated with improved patient outcomes.^25,26

Changes to clinician workflow can significantly impact the introduction of new models of care. A mandated time each day for structured rounds meant less flexibility in workflow for clinicians and made greater demands on their time management and communication skills. Furthermore, the need for human resource negotiations with nurse representatives was an unexpected component of successfully introducing the changes to workflow. Once the benefits of saved time and better communication became evident, changes to workflow were generally accepted. These challenges can be managed if stakeholders are engaged and supportive of the changes.¹³

Finally, our findings emphasize the importance of combining qualitative and quantitative data when evaluating an intervention. In this case, the qualitative outcomes that include “intangible” positive effects, such as cultural change and improved staff understanding of one another’s roles, might encourage us to continue with the SIBR intervention, which would allow more time to see if the trend of reduced LOS identified in the statistical analysis would translate to a significant effect over time.

We are unable to identify which aspects of the intervention led to the greatest impact on our outcomes. A recent study found that interdisciplinary rounds had no impact on patients’ perceptions of shared decision-making or care satisfaction.²⁷ Although our findings indicated many potential benefits for patients, we were not able to interview patients or their carers to confirm these findings. In addition, we do not have any patient-centered outcomes, which would be important to consider in future work. Although our data on clinical response calls might be seen as a proxy for adverse events, we do not have data on adverse events or errors, and these are important to consider in future work. Finally, our findings are based on data from a single institution.

While there were some criticisms, participants expressed overwhelmingly positive reactions to the SIBR. The biggest reported benefit was perceived improved communication and understanding between and within the clinical professions, and between clinicians and patients. Improved communication was perceived to have fostered improved teamwork and team functioning, with most respondents feeling that they were a valued part of the new team. Improved teamwork was thought to contribute to improved task performance and led interviewees to perceive a higher level of patient safety. This research highlights the need for multimethod evaluations that address contextual factors as well as clinical outcomes.

Acknowledgments

The authors would like to acknowledge the clinicians and staff members who participated in this study. We would also like to acknowledge the support from the NSW Clinical Excellence Commission, in particular, Dr. Peter Kennedy, Mr. Wilson Yeung, Ms. Tracy Clarke, and Mr. Allan Zhang, and also from Ms. Karen Storey and Mr. Steve Shea of the Organisational Performance Management team at the Orange Health Service.

Disclosures

None of the authors had conflicts of interest in relation to the conduct or reporting of this study, with the exception that the lead author’s institution, the Australian Institute of Health Innovation, received a small grant from the New South Wales Clinical Excellence Commission to conduct the work. Ethics approval for the research was granted by the Greater Western Area Health Service Human Research Ethics Committee (HREC/13/GWAHS/22). All interviewees consented to participate in the study. For patient data, consent was not obtained, but presented data are anonymized. The full dataset is available from the corresponding author with restrictions. This research was funded by the NSW Clinical Excellence Commission, who also encouraged submission of the article for publication. The funding source did not have any role in conduct or reporting of the study. R.C.W., J.P., and J.J. conceptualized and conducted the qualitative component of the study, including method, data collection, data analysis, and writing of the manuscript. G.L., C.H., and H.D. conceptualized the quantitative component of the study, including method, data collection, data analysis, and writing of the manuscript. G.S. contributed to conceptualization of the study, and significantly contributed to the revision of the manuscript. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. As the lead author, R.C.W. affirms that the manuscript is an honest, accurate, and transparent account of the study being reported, that no important aspects of the study have been omitted, and that any discrepancies from the study as planned have been explained.

Evidence has emerged over the last decade of the importance of the front line patient care team in improving quality and safety of patient care.^1-3 Improving collaboration and workflow is thought to increase reliability of care delivery.¹ One promising method to improve collaboration is the interdisciplinary ward round (IDR), whereby medical, nursing, and allied health staff attend ward rounds together. IDRs have been shown to reduce the average cost and length of hospital stay,^4,5 although a recent systematic review found inconsistent improvements across studies.⁶ Using the term “interdisciplinary,” however, does not necessarily imply the inclusion of all disciplines necessary for patient care. The challenge of conducting interdisciplinary rounds is considerable in today’s busy clinical environment: health professionals who are spread across multiple locations within the hospital, and who have competing hospital responsibilities and priorities, must come together at the same time and for a set period each day. A survey with respondents from Australia, the United States, and Canada found that only 65% of rounds labelled “interdisciplinary” included a physician.⁷

While IDRs are not new, structured IDRs involve the purposeful inclusion of all disciplinary groups relevant to a patient’s care, alongside a checklist tool to aid comprehensive but concise daily assessment of progress and treatment planning. Novel, structured IDR interventions have been tested recently in various settings, resulting in improved teamwork, hospital performance, and patient outcomes in the US, including the Structured Interdisciplinary Bedside Round (SIBR) model.^8-12

The aim of this study was to assess the impact of the new structure and the associated practice changes on interprofessional working and a set of key patient and hospital outcome measures. As part of the intervention, the hospital established an Acute Medical Unit (AMU) based on the Accountable Care Unit model.¹³

Description of the Intervention

The AMU brought together 2 existing medical wards, a general medical ward and a 48-hour turnaround Medical Assessment Unit (MAU), into 1 geographical location with 26 beds. Prior to the merger, the MAU and general medical ward had separate and distinct cultures and workflows. The MAU was staffed with experienced nurses; nurses worked within a patient allocation model, the workload was shared, and relationships were collegial. In contrast, the medical ward was more typical of the remainder of the hospital: nurses had a heavy workload, managed a large group of longer-term complex patients, and they used a team-based nursing model of care in which senior nurses supervised junior staff. It was decided that because of the seniority of the MAU staff, they should be in charge of the combined AMU, and the patient allocation model of care would be used to facilitate SIBR.

Consultants, junior doctors, nurses, and allied health professionals (including a pharmacist, physiotherapist, occupational therapist, and social worker) were geographically aligned to the new ward, allowing them to participate as a team in daily structured ward rounds. Rounds are scheduled at the same time each day to enable family participation. The ward round is coordinated by a registrar or intern, with input from patient, family, nursing staff, pharmacy, allied health, and other doctors (intern, registrar, and consultant) based on the unit. The patient load is distributed between 2 rounds: 1 scheduled for 10 am and the other for 11 am each weekday.

Data Collection Strategy

The study was set in an AMU in a large tertiary care hospital in regional Australia and used a convergent parallel multimethod approach¹⁴ to evaluate the implementation and effect of SIBR in the AMU. The study population consisted of 32 clinicians employed at the study hospital: (1) the leadership team involved in the development and implementation of the intervention and (2) members of clinical staff who were part of the AMU team.

Qualitative Data

Qualitative measures consisted of semistructured interviews. We utilized multiple strategies to recruit interviewees, including a snowball technique, criterion sampling,¹⁵ and emergent sampling, so that we could seek the views of both the leadership team responsible for the implementation and “frontline” clinical staff whose daily work was directly affected by it. Everyone who was initially recruited agreed to be interviewed, and additional frontline staff asked to be interviewed once they realized that we were asking about how staff experienced the changes in practice.

The research team developed a semistructured interview guide based on an understanding of the merger of the 2 units as well as an understanding of changes in practice of the rounds (provided in Appendix 1). The questions were pilot tested on a separate unit and revised. Questions were structured into 5 topic areas: planning and implementation of AMU/SIBR model, changes in work practices because of the new model, team functioning, job satisfaction, and perceived impact of the new model on patients and families. All interviews were audio-recorded and transcribed verbatim for analysis.

Quantitative Data

Quantitative data were collected on patient outcome measures: length of stay (LOS), discharge date and time, mode of separation (including death), primary diagnostic category, total hospital stay cost and “clinical response calls,” and patient demographic data (age, gender, and Patient Clinical Complexity Level [PCCL]). The PCCL is a standard measure used in Australian public inpatient facilities and is calculated for each episode of care.¹⁶ It measures the cumulative effect of a patient’s complications and/or comorbidities and takes an integer value between 0 (no clinical complexity effect) and 4 (catastrophic clinical complexity effect).

Data regarding LOS, diagnosis (Australian Refined Diagnosis Related Groups [AR-DRG], version 7), discharge date, and mode of separation (including death) were obtained from the New South Wales Ministry of Health’s Health Information Exchange for patients discharged during the year prior to the intervention through 1 year after the implementation of the intervention. The total hospital stay cost for these individuals was obtained from the local Health Service Organizational Performance Management unit. Inclusion criteria were inpatients aged over 15 years experiencing acute episodes of care; patients with a primary diagnostic category of mental diseases and disorders were excluded. LOS was calculated based on ward stay. AMU data were compared with the remaining hospital ward data (the control group). Data on “clinical response calls” per month per ward were also obtained for the 12 months prior to intervention and the 12 months of the intervention.

Analysis

Qualitative Analysis

Qualitative data analysis consisted of a hybrid form of textual analysis, combining inductive and deductive logics.^17,18 Initially, 3 researchers (J.P., J.J., and R.C.W.) independently coded the interview data inductively to identify themes. Discrepancies were resolved through discussion until consensus was reached. Then, to further facilitate analysis, the researchers deductively imposed a matrix categorization, consisting of 4 a priori categories: context/conditions, practices/processes, professional interactions, and consequences.^19,20 Additional a priori categories were used to sort the themes further in terms of experiences prior to, during, and following implementation of the intervention. To compare changes in those different time periods, we wanted to know what themes were related to implementation and whether those themes continued to be applicable to sustainability of the changes.

Quantitative analysis. Distribution of continuous data was examined by using the one-sample Kolmogorov-Smirnov test. We compared pre-SIBR (baseline) measures using the Student t test for normally distributed data, the Mann-Whitney U z test for nonparametric data (denoted as M-W U z), and χ² tests for categorical data. Changes in monthly “clinical response calls” between the AMU and the control wards over time were explored by using analysis of variance (ANOVA). Changes in LOS and cost of stay from the year prior to the intervention to the first year of the intervention were analyzed by using generalized linear models, which are a form of linear regression. Factors, or independent variables, included in the models were time period (before or during intervention), ward (AMU or control), an interaction term (time by ward), patient age, gender, primary diagnosis (major diagnostic categories of the AR-DRG version 7.0), and acuity (PCCL). The estimated marginal means for cost of stay for the 12-month period prior to the intervention and for the first 12 months of the intervention were produced. All statistical analyses were performed by using IBM SPSS version 21 (IBM Corp., Armonk, New York) and with alpha set at P  < .05.

Qualitative Evaluation of the Intervention

Participants.

Three researchers (RCW, JP, and JJ) conducted in-person, semistructured interviews with 32 clinicians (9 male, 23 female) during a 3-day period. The duration of the interviews ranged from 19 minutes to 68 minutes. Participants consisted of 8 doctors, 18 nurses, 5 allied health professionals, and an administrator. Ten of the participants were involved in the leadership group that drove the planning and implementation of SIBR and the AMU.

Themes

Context and Conditions of Work

Practices and Processes

Participants perceived postintervention work processes to be more efficient. A primary example was a near-universal approval of the time saved from not “chasing” other professionals now that they were predictably available on the ward. More timely decision-making was thought to result from this predicted availability and associated improvements in communication.

The SIBR enforced a workflow on all staff, who felt there was less flexibility to work autonomously (doctors) or according to patients’ needs (nurses). More junior staff expressed anxiety about delayed completion of discharge-related administrative tasks because of the midday completion of the round. Allied health professionals who had commitments in other areas of the hospital often faced a dilemma about how to prioritize SIBR attendance and activities on other wards. This was managed differently depending on the specific allied health profession and the individuals within that profession.

Professional Interactions

In terms of interprofessional dynamics on the AMU, the implementation of SIBR resulted in a shift in power between the doctors and the nurses. In the old ward, doctors largely controlled the timing of medical rounding processes. In the new AMU, doctors had to relinquish some control over the timing of personal workflow to comply with the requirements of SIBR. Furthermore, there was evidence that this had some impact on traditional hierarchical models of communication and created a more level playing field, as nonmedical professionals felt more empowered to voice their thoughts during and outside of rounds.

The rounds provided much greater visibility of the “big picture” and each profession’s role within it; this allowed each clinician to adjust their work to fit in and take account of others. The process was not instantaneous, and trust developed over a period of weeks. Better communication meant fewer misunderstandings, and workload dropped.

The participation of allied health professionals in the round enhanced clinician interprofessional skills and knowledge. The more inclusive approach facilitated greater trust between clinical disciplines and a development of increased confidence among nursing, allied health, and administrative professionals.

In contrast to the positive impacts of the new model of care on communication and relationships within the AMU, interdepartmental relationships were seen to have suffered. The processes and practices of the new AMU are different to those in the other hospital departments, resulting in some isolation of the unit and difficulties interacting with other areas of the hospital. For example, the trade-offs that allied health professionals made to participate in SIBR often came at the expense of other units or departments.

Consequences

All interviewees lauded the benefits of the SIBR intervention for patients. Patients were perceived to be better informed and more respected, and they benefited from greater perceived timeliness of treatment and discharge, easier access to doctors, better continuity of treatment and outcomes, improved nurse knowledge of their circumstances, and fewer gaps in their care. Clinicians spoke directly to the patient during SIBR, rather than consulting with professional colleagues over the patient’s head. Some staff felt that doctors were now thinking of patients as “people” rather than “a set of symptoms.” Nurses discovered that informed patients are easier to manage.

Staff members were prepared to compromise on their own needs in the interests of the patient. The emphasis on the patient during rounds resulted in improved advocacy behaviors of clinicians. The nurses became more empowered and able to show greater initiative. Families appeared to find it much easier to access the doctors and obtain information about the patient, resulting in less distress and a greater sense of control and trust in the process.

Quantitative Evaluation of the Intervention

Hospital Outcomes

Patient demographics did not change over time within either the AMU or control wards. However, there were significant increases in Patient Clinical Complexity Level (PCCL) ratings for both the AMU (44.7% to 40.3%; P<0.05) and the control wards (65.2% to 61.6%; P < .001). There was not a statistically significant shift over time in median LoS on the ward prior to (2.16 days, IQR 3.07) and during SIBR in the AMU (2.15 days; IQR 3.28), while LoS increased in the control (pre-SIBR: 1.67, 2.34; during SIBR 1.73, 2.40; M-W U z = -2.46, P = .014). Mortality rates were stable across time for both the AMU (pre-SIBR 2.6% [95% confidence interval {CI}, 1.9-3.5]; during SIBR 2.8% [95% CI, 2.1-3.7]) and the control (pre-SIBR 1.3% [95% CI, 1.0-1.5]; during SIBR 1.2% [95% CI, 1.0-1.4]).

The total number of “clinical response calls” or “flags” per month dropped significantly from pre-SIBR to during SIBR for the AMU from a mean of 63.1 (standard deviation 15.1) to 31.5 (10.8), but remained relatively stable in the control (pre-SIBR 72.5 [17.6]; during SIBR 74.0 [28.3]), and this difference was statistically significant (F (1,44) = 9.03; P = .004). There was no change in monthly “red flags” or “rapid response calls” over time (AMU: 10.5 [3.6] to 9.1 [4.7]; control: 40.3 [11.7] to 41.8 [10.8]). The change in total “clinical response calls” over time was attributable to the “yellow flags” or the decline in “calls for clinical review” in the AMU (from 52.6 [13.5] to 22.4 [9.2]). The average monthly “yellow flags” remained stable in the control (pre-SIBR 32.2 [11.6]; during SIBR 32.3 [22.4]). The AMU and the control wards differed significantly in how the number of monthly “calls for clinical review” changed from pre-SIBR to during SIBR (F (1,44) = 12.18; P = .001).

The 2 main outcome measures, LOS and costs, were analyzed to determine whether changes over time differed between the AMU and the control wards after accounting for age, gender, and PCCL. There was no statistically significant difference between the AMU and control wards in terms of change in LOS over time (Wald χ² = 1.05; degrees of freedom [df] = 1; P = .31). There was a statistically significant interaction for cost of stay, indicating that ward types differed in how they changed over time (with a drop in cost over time observed in the AMU and an increase observed in the control) (Wald χ² = 6.34; df = 1; P = .012.

We report on the implementation of an AMU model of care, including the reorganization of a nursing unit, implementation of IDR, and geographical localization. Our study design allowed a more comprehensive assessment of the implementation of system redesign to include provider perceptions and clinical outcomes.

The 2 very different cultures of the old wards that were combined into the AMU, as well as the fact that the teams had not previously worked together, made the merger of the 2 wards difficult. Historically, the 2 teams had worked in very different ways, and this created barriers to implementation. The SIBR also demanded new ways of working closely with other disciplines, which disrupted older clinical cultures and relationships. While organizational culture is often discussed, and even measured, the full impact of cultural factors when making workplace changes is frequently underestimated.²¹ The development of a new culture takes time, and it can lag organizational structural changes by months or even years.²² As our interviewees expressed, often emotionally, there was a sense of loss during the merger of the 2 units. While this is a potential consequence of any large organizational change, it could be addressed during the planning stages, prior to implementation, by acknowledging and perhaps honoring what is being left behind. It is safe to assume that future units implementing the rounding intervention will not fully realize commensurate levels of culture change until well after the structural and process changes are finalized, and only then if explicit effort is made to engender cultural change.

Overall, however, the interviewees perceived that the SIBR intervention led to improved teamwork and team functioning. These improvements were thought to benefit task performance and patient safety. Our study is consistent with other research in the literature that reported that greater staff empowerment and commitment is associated with interdisciplinary patient care interventions in front line caregiving teams.^23,24 The perception of a more equal nurse-physician relationship resulted in improved job satisfaction, better interprofessional relationships, and perceived improvements in patient care. A flatter power gradient across professions and increased interdisciplinary teamwork has been shown to be associated with improved patient outcomes.^25,26

Changes to clinician workflow can significantly impact the introduction of new models of care. A mandated time each day for structured rounds meant less flexibility in workflow for clinicians and made greater demands on their time management and communication skills. Furthermore, the need for human resource negotiations with nurse representatives was an unexpected component of successfully introducing the changes to workflow. Once the benefits of saved time and better communication became evident, changes to workflow were generally accepted. These challenges can be managed if stakeholders are engaged and supportive of the changes.¹³

Finally, our findings emphasize the importance of combining qualitative and quantitative data when evaluating an intervention. In this case, the qualitative outcomes that include “intangible” positive effects, such as cultural change and improved staff understanding of one another’s roles, might encourage us to continue with the SIBR intervention, which would allow more time to see if the trend of reduced LOS identified in the statistical analysis would translate to a significant effect over time.

We are unable to identify which aspects of the intervention led to the greatest impact on our outcomes. A recent study found that interdisciplinary rounds had no impact on patients’ perceptions of shared decision-making or care satisfaction.²⁷ Although our findings indicated many potential benefits for patients, we were not able to interview patients or their carers to confirm these findings. In addition, we do not have any patient-centered outcomes, which would be important to consider in future work. Although our data on clinical response calls might be seen as a proxy for adverse events, we do not have data on adverse events or errors, and these are important to consider in future work. Finally, our findings are based on data from a single institution.

While there were some criticisms, participants expressed overwhelmingly positive reactions to the SIBR. The biggest reported benefit was perceived improved communication and understanding between and within the clinical professions, and between clinicians and patients. Improved communication was perceived to have fostered improved teamwork and team functioning, with most respondents feeling that they were a valued part of the new team. Improved teamwork was thought to contribute to improved task performance and led interviewees to perceive a higher level of patient safety. This research highlights the need for multimethod evaluations that address contextual factors as well as clinical outcomes.

Acknowledgments

The authors would like to acknowledge the clinicians and staff members who participated in this study. We would also like to acknowledge the support from the NSW Clinical Excellence Commission, in particular, Dr. Peter Kennedy, Mr. Wilson Yeung, Ms. Tracy Clarke, and Mr. Allan Zhang, and also from Ms. Karen Storey and Mr. Steve Shea of the Organisational Performance Management team at the Orange Health Service.

Disclosures

None of the authors had conflicts of interest in relation to the conduct or reporting of this study, with the exception that the lead author’s institution, the Australian Institute of Health Innovation, received a small grant from the New South Wales Clinical Excellence Commission to conduct the work. Ethics approval for the research was granted by the Greater Western Area Health Service Human Research Ethics Committee (HREC/13/GWAHS/22). All interviewees consented to participate in the study. For patient data, consent was not obtained, but presented data are anonymized. The full dataset is available from the corresponding author with restrictions. This research was funded by the NSW Clinical Excellence Commission, who also encouraged submission of the article for publication. The funding source did not have any role in conduct or reporting of the study. R.C.W., J.P., and J.J. conceptualized and conducted the qualitative component of the study, including method, data collection, data analysis, and writing of the manuscript. G.L., C.H., and H.D. conceptualized the quantitative component of the study, including method, data collection, data analysis, and writing of the manuscript. G.S. contributed to conceptualization of the study, and significantly contributed to the revision of the manuscript. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. As the lead author, R.C.W. affirms that the manuscript is an honest, accurate, and transparent account of the study being reported, that no important aspects of the study have been omitted, and that any discrepancies from the study as planned have been explained.

Evidence has emerged over the last decade of the importance of the front line patient care team in improving quality and safety of patient care.^1-3 Improving collaboration and workflow is thought to increase reliability of care delivery.¹ One promising method to improve collaboration is the interdisciplinary ward round (IDR), whereby medical, nursing, and allied health staff attend ward rounds together. IDRs have been shown to reduce the average cost and length of hospital stay,^4,5 although a recent systematic review found inconsistent improvements across studies.⁶ Using the term “interdisciplinary,” however, does not necessarily imply the inclusion of all disciplines necessary for patient care. The challenge of conducting interdisciplinary rounds is considerable in today’s busy clinical environment: health professionals who are spread across multiple locations within the hospital, and who have competing hospital responsibilities and priorities, must come together at the same time and for a set period each day. A survey with respondents from Australia, the United States, and Canada found that only 65% of rounds labelled “interdisciplinary” included a physician.⁷

While IDRs are not new, structured IDRs involve the purposeful inclusion of all disciplinary groups relevant to a patient’s care, alongside a checklist tool to aid comprehensive but concise daily assessment of progress and treatment planning. Novel, structured IDR interventions have been tested recently in various settings, resulting in improved teamwork, hospital performance, and patient outcomes in the US, including the Structured Interdisciplinary Bedside Round (SIBR) model.^8-12

The aim of this study was to assess the impact of the new structure and the associated practice changes on interprofessional working and a set of key patient and hospital outcome measures. As part of the intervention, the hospital established an Acute Medical Unit (AMU) based on the Accountable Care Unit model.¹³

Description of the Intervention

The AMU brought together 2 existing medical wards, a general medical ward and a 48-hour turnaround Medical Assessment Unit (MAU), into 1 geographical location with 26 beds. Prior to the merger, the MAU and general medical ward had separate and distinct cultures and workflows. The MAU was staffed with experienced nurses; nurses worked within a patient allocation model, the workload was shared, and relationships were collegial. In contrast, the medical ward was more typical of the remainder of the hospital: nurses had a heavy workload, managed a large group of longer-term complex patients, and they used a team-based nursing model of care in which senior nurses supervised junior staff. It was decided that because of the seniority of the MAU staff, they should be in charge of the combined AMU, and the patient allocation model of care would be used to facilitate SIBR.

Consultants, junior doctors, nurses, and allied health professionals (including a pharmacist, physiotherapist, occupational therapist, and social worker) were geographically aligned to the new ward, allowing them to participate as a team in daily structured ward rounds. Rounds are scheduled at the same time each day to enable family participation. The ward round is coordinated by a registrar or intern, with input from patient, family, nursing staff, pharmacy, allied health, and other doctors (intern, registrar, and consultant) based on the unit. The patient load is distributed between 2 rounds: 1 scheduled for 10 am and the other for 11 am each weekday.

Data Collection Strategy

The study was set in an AMU in a large tertiary care hospital in regional Australia and used a convergent parallel multimethod approach¹⁴ to evaluate the implementation and effect of SIBR in the AMU. The study population consisted of 32 clinicians employed at the study hospital: (1) the leadership team involved in the development and implementation of the intervention and (2) members of clinical staff who were part of the AMU team.

Qualitative Data

Qualitative measures consisted of semistructured interviews. We utilized multiple strategies to recruit interviewees, including a snowball technique, criterion sampling,¹⁵ and emergent sampling, so that we could seek the views of both the leadership team responsible for the implementation and “frontline” clinical staff whose daily work was directly affected by it. Everyone who was initially recruited agreed to be interviewed, and additional frontline staff asked to be interviewed once they realized that we were asking about how staff experienced the changes in practice.

The research team developed a semistructured interview guide based on an understanding of the merger of the 2 units as well as an understanding of changes in practice of the rounds (provided in Appendix 1). The questions were pilot tested on a separate unit and revised. Questions were structured into 5 topic areas: planning and implementation of AMU/SIBR model, changes in work practices because of the new model, team functioning, job satisfaction, and perceived impact of the new model on patients and families. All interviews were audio-recorded and transcribed verbatim for analysis.

Quantitative Data

Quantitative data were collected on patient outcome measures: length of stay (LOS), discharge date and time, mode of separation (including death), primary diagnostic category, total hospital stay cost and “clinical response calls,” and patient demographic data (age, gender, and Patient Clinical Complexity Level [PCCL]). The PCCL is a standard measure used in Australian public inpatient facilities and is calculated for each episode of care.¹⁶ It measures the cumulative effect of a patient’s complications and/or comorbidities and takes an integer value between 0 (no clinical complexity effect) and 4 (catastrophic clinical complexity effect).

Data regarding LOS, diagnosis (Australian Refined Diagnosis Related Groups [AR-DRG], version 7), discharge date, and mode of separation (including death) were obtained from the New South Wales Ministry of Health’s Health Information Exchange for patients discharged during the year prior to the intervention through 1 year after the implementation of the intervention. The total hospital stay cost for these individuals was obtained from the local Health Service Organizational Performance Management unit. Inclusion criteria were inpatients aged over 15 years experiencing acute episodes of care; patients with a primary diagnostic category of mental diseases and disorders were excluded. LOS was calculated based on ward stay. AMU data were compared with the remaining hospital ward data (the control group). Data on “clinical response calls” per month per ward were also obtained for the 12 months prior to intervention and the 12 months of the intervention.

Analysis

Qualitative Analysis

Qualitative data analysis consisted of a hybrid form of textual analysis, combining inductive and deductive logics.^17,18 Initially, 3 researchers (J.P., J.J., and R.C.W.) independently coded the interview data inductively to identify themes. Discrepancies were resolved through discussion until consensus was reached. Then, to further facilitate analysis, the researchers deductively imposed a matrix categorization, consisting of 4 a priori categories: context/conditions, practices/processes, professional interactions, and consequences.^19,20 Additional a priori categories were used to sort the themes further in terms of experiences prior to, during, and following implementation of the intervention. To compare changes in those different time periods, we wanted to know what themes were related to implementation and whether those themes continued to be applicable to sustainability of the changes.

Quantitative analysis. Distribution of continuous data was examined by using the one-sample Kolmogorov-Smirnov test. We compared pre-SIBR (baseline) measures using the Student t test for normally distributed data, the Mann-Whitney U z test for nonparametric data (denoted as M-W U z), and χ² tests for categorical data. Changes in monthly “clinical response calls” between the AMU and the control wards over time were explored by using analysis of variance (ANOVA). Changes in LOS and cost of stay from the year prior to the intervention to the first year of the intervention were analyzed by using generalized linear models, which are a form of linear regression. Factors, or independent variables, included in the models were time period (before or during intervention), ward (AMU or control), an interaction term (time by ward), patient age, gender, primary diagnosis (major diagnostic categories of the AR-DRG version 7.0), and acuity (PCCL). The estimated marginal means for cost of stay for the 12-month period prior to the intervention and for the first 12 months of the intervention were produced. All statistical analyses were performed by using IBM SPSS version 21 (IBM Corp., Armonk, New York) and with alpha set at P  < .05.

Qualitative Evaluation of the Intervention

Participants.

Three researchers (RCW, JP, and JJ) conducted in-person, semistructured interviews with 32 clinicians (9 male, 23 female) during a 3-day period. The duration of the interviews ranged from 19 minutes to 68 minutes. Participants consisted of 8 doctors, 18 nurses, 5 allied health professionals, and an administrator. Ten of the participants were involved in the leadership group that drove the planning and implementation of SIBR and the AMU.

Themes

Context and Conditions of Work

Practices and Processes

Participants perceived postintervention work processes to be more efficient. A primary example was a near-universal approval of the time saved from not “chasing” other professionals now that they were predictably available on the ward. More timely decision-making was thought to result from this predicted availability and associated improvements in communication.

The SIBR enforced a workflow on all staff, who felt there was less flexibility to work autonomously (doctors) or according to patients’ needs (nurses). More junior staff expressed anxiety about delayed completion of discharge-related administrative tasks because of the midday completion of the round. Allied health professionals who had commitments in other areas of the hospital often faced a dilemma about how to prioritize SIBR attendance and activities on other wards. This was managed differently depending on the specific allied health profession and the individuals within that profession.

Professional Interactions

In terms of interprofessional dynamics on the AMU, the implementation of SIBR resulted in a shift in power between the doctors and the nurses. In the old ward, doctors largely controlled the timing of medical rounding processes. In the new AMU, doctors had to relinquish some control over the timing of personal workflow to comply with the requirements of SIBR. Furthermore, there was evidence that this had some impact on traditional hierarchical models of communication and created a more level playing field, as nonmedical professionals felt more empowered to voice their thoughts during and outside of rounds.

The rounds provided much greater visibility of the “big picture” and each profession’s role within it; this allowed each clinician to adjust their work to fit in and take account of others. The process was not instantaneous, and trust developed over a period of weeks. Better communication meant fewer misunderstandings, and workload dropped.

The participation of allied health professionals in the round enhanced clinician interprofessional skills and knowledge. The more inclusive approach facilitated greater trust between clinical disciplines and a development of increased confidence among nursing, allied health, and administrative professionals.

In contrast to the positive impacts of the new model of care on communication and relationships within the AMU, interdepartmental relationships were seen to have suffered. The processes and practices of the new AMU are different to those in the other hospital departments, resulting in some isolation of the unit and difficulties interacting with other areas of the hospital. For example, the trade-offs that allied health professionals made to participate in SIBR often came at the expense of other units or departments.

Consequences

All interviewees lauded the benefits of the SIBR intervention for patients. Patients were perceived to be better informed and more respected, and they benefited from greater perceived timeliness of treatment and discharge, easier access to doctors, better continuity of treatment and outcomes, improved nurse knowledge of their circumstances, and fewer gaps in their care. Clinicians spoke directly to the patient during SIBR, rather than consulting with professional colleagues over the patient’s head. Some staff felt that doctors were now thinking of patients as “people” rather than “a set of symptoms.” Nurses discovered that informed patients are easier to manage.

Staff members were prepared to compromise on their own needs in the interests of the patient. The emphasis on the patient during rounds resulted in improved advocacy behaviors of clinicians. The nurses became more empowered and able to show greater initiative. Families appeared to find it much easier to access the doctors and obtain information about the patient, resulting in less distress and a greater sense of control and trust in the process.

Quantitative Evaluation of the Intervention

Hospital Outcomes

Patient demographics did not change over time within either the AMU or control wards. However, there were significant increases in Patient Clinical Complexity Level (PCCL) ratings for both the AMU (44.7% to 40.3%; P<0.05) and the control wards (65.2% to 61.6%; P < .001). There was not a statistically significant shift over time in median LoS on the ward prior to (2.16 days, IQR 3.07) and during SIBR in the AMU (2.15 days; IQR 3.28), while LoS increased in the control (pre-SIBR: 1.67, 2.34; during SIBR 1.73, 2.40; M-W U z = -2.46, P = .014). Mortality rates were stable across time for both the AMU (pre-SIBR 2.6% [95% confidence interval {CI}, 1.9-3.5]; during SIBR 2.8% [95% CI, 2.1-3.7]) and the control (pre-SIBR 1.3% [95% CI, 1.0-1.5]; during SIBR 1.2% [95% CI, 1.0-1.4]).

The total number of “clinical response calls” or “flags” per month dropped significantly from pre-SIBR to during SIBR for the AMU from a mean of 63.1 (standard deviation 15.1) to 31.5 (10.8), but remained relatively stable in the control (pre-SIBR 72.5 [17.6]; during SIBR 74.0 [28.3]), and this difference was statistically significant (F (1,44) = 9.03; P = .004). There was no change in monthly “red flags” or “rapid response calls” over time (AMU: 10.5 [3.6] to 9.1 [4.7]; control: 40.3 [11.7] to 41.8 [10.8]). The change in total “clinical response calls” over time was attributable to the “yellow flags” or the decline in “calls for clinical review” in the AMU (from 52.6 [13.5] to 22.4 [9.2]). The average monthly “yellow flags” remained stable in the control (pre-SIBR 32.2 [11.6]; during SIBR 32.3 [22.4]). The AMU and the control wards differed significantly in how the number of monthly “calls for clinical review” changed from pre-SIBR to during SIBR (F (1,44) = 12.18; P = .001).

The 2 main outcome measures, LOS and costs, were analyzed to determine whether changes over time differed between the AMU and the control wards after accounting for age, gender, and PCCL. There was no statistically significant difference between the AMU and control wards in terms of change in LOS over time (Wald χ² = 1.05; degrees of freedom [df] = 1; P = .31). There was a statistically significant interaction for cost of stay, indicating that ward types differed in how they changed over time (with a drop in cost over time observed in the AMU and an increase observed in the control) (Wald χ² = 6.34; df = 1; P = .012.

We report on the implementation of an AMU model of care, including the reorganization of a nursing unit, implementation of IDR, and geographical localization. Our study design allowed a more comprehensive assessment of the implementation of system redesign to include provider perceptions and clinical outcomes.

The 2 very different cultures of the old wards that were combined into the AMU, as well as the fact that the teams had not previously worked together, made the merger of the 2 wards difficult. Historically, the 2 teams had worked in very different ways, and this created barriers to implementation. The SIBR also demanded new ways of working closely with other disciplines, which disrupted older clinical cultures and relationships. While organizational culture is often discussed, and even measured, the full impact of cultural factors when making workplace changes is frequently underestimated.²¹ The development of a new culture takes time, and it can lag organizational structural changes by months or even years.²² As our interviewees expressed, often emotionally, there was a sense of loss during the merger of the 2 units. While this is a potential consequence of any large organizational change, it could be addressed during the planning stages, prior to implementation, by acknowledging and perhaps honoring what is being left behind. It is safe to assume that future units implementing the rounding intervention will not fully realize commensurate levels of culture change until well after the structural and process changes are finalized, and only then if explicit effort is made to engender cultural change.

Overall, however, the interviewees perceived that the SIBR intervention led to improved teamwork and team functioning. These improvements were thought to benefit task performance and patient safety. Our study is consistent with other research in the literature that reported that greater staff empowerment and commitment is associated with interdisciplinary patient care interventions in front line caregiving teams.^23,24 The perception of a more equal nurse-physician relationship resulted in improved job satisfaction, better interprofessional relationships, and perceived improvements in patient care. A flatter power gradient across professions and increased interdisciplinary teamwork has been shown to be associated with improved patient outcomes.^25,26

Changes to clinician workflow can significantly impact the introduction of new models of care. A mandated time each day for structured rounds meant less flexibility in workflow for clinicians and made greater demands on their time management and communication skills. Furthermore, the need for human resource negotiations with nurse representatives was an unexpected component of successfully introducing the changes to workflow. Once the benefits of saved time and better communication became evident, changes to workflow were generally accepted. These challenges can be managed if stakeholders are engaged and supportive of the changes.¹³

Finally, our findings emphasize the importance of combining qualitative and quantitative data when evaluating an intervention. In this case, the qualitative outcomes that include “intangible” positive effects, such as cultural change and improved staff understanding of one another’s roles, might encourage us to continue with the SIBR intervention, which would allow more time to see if the trend of reduced LOS identified in the statistical analysis would translate to a significant effect over time.

We are unable to identify which aspects of the intervention led to the greatest impact on our outcomes. A recent study found that interdisciplinary rounds had no impact on patients’ perceptions of shared decision-making or care satisfaction.²⁷ Although our findings indicated many potential benefits for patients, we were not able to interview patients or their carers to confirm these findings. In addition, we do not have any patient-centered outcomes, which would be important to consider in future work. Although our data on clinical response calls might be seen as a proxy for adverse events, we do not have data on adverse events or errors, and these are important to consider in future work. Finally, our findings are based on data from a single institution.

While there were some criticisms, participants expressed overwhelmingly positive reactions to the SIBR. The biggest reported benefit was perceived improved communication and understanding between and within the clinical professions, and between clinicians and patients. Improved communication was perceived to have fostered improved teamwork and team functioning, with most respondents feeling that they were a valued part of the new team. Improved teamwork was thought to contribute to improved task performance and led interviewees to perceive a higher level of patient safety. This research highlights the need for multimethod evaluations that address contextual factors as well as clinical outcomes.

Acknowledgments

The authors would like to acknowledge the clinicians and staff members who participated in this study. We would also like to acknowledge the support from the NSW Clinical Excellence Commission, in particular, Dr. Peter Kennedy, Mr. Wilson Yeung, Ms. Tracy Clarke, and Mr. Allan Zhang, and also from Ms. Karen Storey and Mr. Steve Shea of the Organisational Performance Management team at the Orange Health Service.

Disclosures

None of the authors had conflicts of interest in relation to the conduct or reporting of this study, with the exception that the lead author’s institution, the Australian Institute of Health Innovation, received a small grant from the New South Wales Clinical Excellence Commission to conduct the work. Ethics approval for the research was granted by the Greater Western Area Health Service Human Research Ethics Committee (HREC/13/GWAHS/22). All interviewees consented to participate in the study. For patient data, consent was not obtained, but presented data are anonymized. The full dataset is available from the corresponding author with restrictions. This research was funded by the NSW Clinical Excellence Commission, who also encouraged submission of the article for publication. The funding source did not have any role in conduct or reporting of the study. R.C.W., J.P., and J.J. conceptualized and conducted the qualitative component of the study, including method, data collection, data analysis, and writing of the manuscript. G.L., C.H., and H.D. conceptualized the quantitative component of the study, including method, data collection, data analysis, and writing of the manuscript. G.S. contributed to conceptualization of the study, and significantly contributed to the revision of the manuscript. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. As the lead author, R.C.W. affirms that the manuscript is an honest, accurate, and transparent account of the study being reported, that no important aspects of the study have been omitted, and that any discrepancies from the study as planned have been explained.

The “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but 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/

A 3-week-old, full-term term male febrile infant was evaluated in the emergency department (ED). On the day of admission, he was noted to feel warm to the touch and was found to have a rectal temperature of 101.3°F (38.3°C) at home.

In the ED, the patient was well appearing and had normal physical exam findings. His workup in the ED included a normal chest radiograph, complete blood count (CBC) with differential count, cerebrospinal fluid (CSF) analysis (cell count, protein, and glucose), and urinalysis. Blood, CSF, and catheterized urine cultures were collected, and he was admitted to the hospital on parenteral antibiotics. His provider informed the parents that the infant would be observed in the hospital for 48 hours while monitoring the bacterial cultures. Is it necessary for the hospitalization of this child to last a full 48 hours?

Evaluation and management of fever (T ≥ 38°C) is a common cause of emergency department visits and accounts for up to 20% of pediatric emergency visits.²

Why You Might Think Hospitalization for at Least 48 Hours is Necessary

The evaluation and management of fever in infants aged less than 90 days is challenging due to concern for occult serious bacterial infections. In particular, providers may be concerned that the physical exam lacks sensitivity.⁹

There is also a perceived risk of poor outcomes in young infants if a serious bacterial infection is missed. For these reasons, the evaluation and management of febrile infants has been characterized by practice variability in both outpatient¹⁰ and ED³ settings.

Commonly used febrile infant management protocols vary in approach and do not provide clear guidelines on the recommended duration of hospitalization and empiric antimicrobial treatment.^11-14 Length of hospitalization was widely studied in infants between 1979 and 1999, and results showed that the majority of clinically important bacterial pathogens can be detected within 48 hours.^15-17 Many textbooks and online references, based on this literature, continue to support 48 to 72 hours of observation and empiric antimicrobial treatment for febrile infants.^18,19 A 2012 AAP Clinical Report advocated for limiting the antimicrobial treatment in low-risk infants suspected of early-onset sepsis to 48 hours.²⁰

Why Shorten the Period of In-Hospital Observation to a Maximum of 36 Hours of Culture Incubation

Discharge of low-risk infants with negative enhanced urinalysis and negative bacterial cultures at 36 hours or earlier can reduce costs²¹ and potentially preventable harm (eg, intravenous catheter complications, nosocomial infections) without negatively impacting patient outcomes.²² Early discharge is also patient-centered, given the stress and indirect costs associated with hospitalization, including potential separation of a breastfeeding infant and mother, lost wages from time off work, or childcare for well siblings.²³

Initial studies that evaluated the time-to-positivity (TTP) of bacterial cultures in febrile infants predate the use of continuous monitoring systems for blood cultures. Traditional bacterial culturing techniques require direct observation of broth turbidity and subsequent subculturing onto chocolate and sheep blood agar, typically occurring only once daily.²⁴ Current commercially available continuous monitoring bacterial culture systems decrease TTP by immediately alerting laboratory technicians to bacterial growth through the detection of ¹⁴CO2 released by organisms utilizing radiolabeled glucose in growth media.²⁴ In addition, many studies supporting the evaluation of febrile infants in the hospital for a 48-hour period include those in ICU settings,²⁵ with medically complex histories,²⁴ and aged < 28 days admitted in the NICU,¹⁵ where pathogens with longer incubation times are frequently seen.

When a Longer Observation Period May Be Warranted

What You Should Do Instead: Limit Hospitalization to a Maximum of 36 Hours

For well-appearing febrile infants between 0–90 days of age hospitalized for observation and awaiting bacterial culture results, providers should consider discharge at 36 hours or less, rather than 48 hours, if blood, urine, and CSF cultures do not show bacterial growth. In a large health system, researchers implemented an evidence-based care process model for febrile infants to provide specific guidelines for laboratory testing, criteria for admission, and recommendation for discontinuation of empiric antibiotics and discharge after 36 hours in infants with negative bacterial cultures. These changes led to a 27% reduction in the length of hospital stay and 23% reduction in inpatient costs without any cases of missed bacteremia.²¹ The reduction in the in-hospital observation duration to 24 hours of culture incubation for well-appearing febrile infants has been advocated ³² and is a common practice for infants with appropriate follow up and parental assurance. This recommendation is supported by the following:

• Recent data showing the overwhelming majority of pathogens will be identified by blood culture <24 hours in infants aged 0-90 days³² with blood culture TTP in infants aged 0-30 days being either no different²⁶ or potentially shorter³²
• Studies showing that infants meeting low-risk clinical and laboratory profiles further reduce the likelihood of identifying serious bacterial infection after 24 hours to 0.3%.¹

• Determine if febrile infants aged 0-90 days are at low risk for serious bacterial infection and obtain appropriate bacterial cultures.
• If hospitalized for observation, discharge low-risk febrile infants aged 0–90 days after 36 hours or less if bacterial cultures remain negative.
• If hospitalized for observation, consider reducing the length of inpatient observation for low-risk febrile infants aged 0–90 days with reliable follow-up to 24 hours or less when the culture results are negative.

Monitoring patients in the hospital for greater than 36 hours of bacterial culture incubation is unnecessary for patients similar to the 3 week-old full-term infant in the case presentation, who are at low risk for serious bacterial infection based on available scoring systems and have negative cultures. If patients are not deemed low risk, have an incomplete laboratory evaluation, or have had prior antibiotic treatment, longer observation in the hospital may be warranted. Close reassessment of the rare patients whose blood cultures return positive after 36 hours is necessary, but their outcomes are excellent, especially in well-appearing infants.^7,33

What do you do?

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing[email protected].

Disclosures

There are no conflicts of interest relevant to this work reported by any of the authors.

The “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but 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/

A 3-week-old, full-term term male febrile infant was evaluated in the emergency department (ED). On the day of admission, he was noted to feel warm to the touch and was found to have a rectal temperature of 101.3°F (38.3°C) at home.

In the ED, the patient was well appearing and had normal physical exam findings. His workup in the ED included a normal chest radiograph, complete blood count (CBC) with differential count, cerebrospinal fluid (CSF) analysis (cell count, protein, and glucose), and urinalysis. Blood, CSF, and catheterized urine cultures were collected, and he was admitted to the hospital on parenteral antibiotics. His provider informed the parents that the infant would be observed in the hospital for 48 hours while monitoring the bacterial cultures. Is it necessary for the hospitalization of this child to last a full 48 hours?

Evaluation and management of fever (T ≥ 38°C) is a common cause of emergency department visits and accounts for up to 20% of pediatric emergency visits.²

Why You Might Think Hospitalization for at Least 48 Hours is Necessary

The evaluation and management of fever in infants aged less than 90 days is challenging due to concern for occult serious bacterial infections. In particular, providers may be concerned that the physical exam lacks sensitivity.⁹

There is also a perceived risk of poor outcomes in young infants if a serious bacterial infection is missed. For these reasons, the evaluation and management of febrile infants has been characterized by practice variability in both outpatient¹⁰ and ED³ settings.

Commonly used febrile infant management protocols vary in approach and do not provide clear guidelines on the recommended duration of hospitalization and empiric antimicrobial treatment.^11-14 Length of hospitalization was widely studied in infants between 1979 and 1999, and results showed that the majority of clinically important bacterial pathogens can be detected within 48 hours.^15-17 Many textbooks and online references, based on this literature, continue to support 48 to 72 hours of observation and empiric antimicrobial treatment for febrile infants.^18,19 A 2012 AAP Clinical Report advocated for limiting the antimicrobial treatment in low-risk infants suspected of early-onset sepsis to 48 hours.²⁰

Why Shorten the Period of In-Hospital Observation to a Maximum of 36 Hours of Culture Incubation

Discharge of low-risk infants with negative enhanced urinalysis and negative bacterial cultures at 36 hours or earlier can reduce costs²¹ and potentially preventable harm (eg, intravenous catheter complications, nosocomial infections) without negatively impacting patient outcomes.²² Early discharge is also patient-centered, given the stress and indirect costs associated with hospitalization, including potential separation of a breastfeeding infant and mother, lost wages from time off work, or childcare for well siblings.²³

Initial studies that evaluated the time-to-positivity (TTP) of bacterial cultures in febrile infants predate the use of continuous monitoring systems for blood cultures. Traditional bacterial culturing techniques require direct observation of broth turbidity and subsequent subculturing onto chocolate and sheep blood agar, typically occurring only once daily.²⁴ Current commercially available continuous monitoring bacterial culture systems decrease TTP by immediately alerting laboratory technicians to bacterial growth through the detection of ¹⁴CO2 released by organisms utilizing radiolabeled glucose in growth media.²⁴ In addition, many studies supporting the evaluation of febrile infants in the hospital for a 48-hour period include those in ICU settings,²⁵ with medically complex histories,²⁴ and aged < 28 days admitted in the NICU,¹⁵ where pathogens with longer incubation times are frequently seen.

When a Longer Observation Period May Be Warranted

What You Should Do Instead: Limit Hospitalization to a Maximum of 36 Hours

For well-appearing febrile infants between 0–90 days of age hospitalized for observation and awaiting bacterial culture results, providers should consider discharge at 36 hours or less, rather than 48 hours, if blood, urine, and CSF cultures do not show bacterial growth. In a large health system, researchers implemented an evidence-based care process model for febrile infants to provide specific guidelines for laboratory testing, criteria for admission, and recommendation for discontinuation of empiric antibiotics and discharge after 36 hours in infants with negative bacterial cultures. These changes led to a 27% reduction in the length of hospital stay and 23% reduction in inpatient costs without any cases of missed bacteremia.²¹ The reduction in the in-hospital observation duration to 24 hours of culture incubation for well-appearing febrile infants has been advocated ³² and is a common practice for infants with appropriate follow up and parental assurance. This recommendation is supported by the following:

• Recent data showing the overwhelming majority of pathogens will be identified by blood culture <24 hours in infants aged 0-90 days³² with blood culture TTP in infants aged 0-30 days being either no different²⁶ or potentially shorter³²
• Studies showing that infants meeting low-risk clinical and laboratory profiles further reduce the likelihood of identifying serious bacterial infection after 24 hours to 0.3%.¹

• Determine if febrile infants aged 0-90 days are at low risk for serious bacterial infection and obtain appropriate bacterial cultures.
• If hospitalized for observation, discharge low-risk febrile infants aged 0–90 days after 36 hours or less if bacterial cultures remain negative.
• If hospitalized for observation, consider reducing the length of inpatient observation for low-risk febrile infants aged 0–90 days with reliable follow-up to 24 hours or less when the culture results are negative.

Monitoring patients in the hospital for greater than 36 hours of bacterial culture incubation is unnecessary for patients similar to the 3 week-old full-term infant in the case presentation, who are at low risk for serious bacterial infection based on available scoring systems and have negative cultures. If patients are not deemed low risk, have an incomplete laboratory evaluation, or have had prior antibiotic treatment, longer observation in the hospital may be warranted. Close reassessment of the rare patients whose blood cultures return positive after 36 hours is necessary, but their outcomes are excellent, especially in well-appearing infants.^7,33

What do you do?

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing[email protected].

Disclosures

There are no conflicts of interest relevant to this work reported by any of the authors.

The “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but 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/

A 3-week-old, full-term term male febrile infant was evaluated in the emergency department (ED). On the day of admission, he was noted to feel warm to the touch and was found to have a rectal temperature of 101.3°F (38.3°C) at home.

In the ED, the patient was well appearing and had normal physical exam findings. His workup in the ED included a normal chest radiograph, complete blood count (CBC) with differential count, cerebrospinal fluid (CSF) analysis (cell count, protein, and glucose), and urinalysis. Blood, CSF, and catheterized urine cultures were collected, and he was admitted to the hospital on parenteral antibiotics. His provider informed the parents that the infant would be observed in the hospital for 48 hours while monitoring the bacterial cultures. Is it necessary for the hospitalization of this child to last a full 48 hours?

Evaluation and management of fever (T ≥ 38°C) is a common cause of emergency department visits and accounts for up to 20% of pediatric emergency visits.²

Why You Might Think Hospitalization for at Least 48 Hours is Necessary

The evaluation and management of fever in infants aged less than 90 days is challenging due to concern for occult serious bacterial infections. In particular, providers may be concerned that the physical exam lacks sensitivity.⁹

There is also a perceived risk of poor outcomes in young infants if a serious bacterial infection is missed. For these reasons, the evaluation and management of febrile infants has been characterized by practice variability in both outpatient¹⁰ and ED³ settings.

Commonly used febrile infant management protocols vary in approach and do not provide clear guidelines on the recommended duration of hospitalization and empiric antimicrobial treatment.^11-14 Length of hospitalization was widely studied in infants between 1979 and 1999, and results showed that the majority of clinically important bacterial pathogens can be detected within 48 hours.^15-17 Many textbooks and online references, based on this literature, continue to support 48 to 72 hours of observation and empiric antimicrobial treatment for febrile infants.^18,19 A 2012 AAP Clinical Report advocated for limiting the antimicrobial treatment in low-risk infants suspected of early-onset sepsis to 48 hours.²⁰

Why Shorten the Period of In-Hospital Observation to a Maximum of 36 Hours of Culture Incubation

Discharge of low-risk infants with negative enhanced urinalysis and negative bacterial cultures at 36 hours or earlier can reduce costs²¹ and potentially preventable harm (eg, intravenous catheter complications, nosocomial infections) without negatively impacting patient outcomes.²² Early discharge is also patient-centered, given the stress and indirect costs associated with hospitalization, including potential separation of a breastfeeding infant and mother, lost wages from time off work, or childcare for well siblings.²³

Initial studies that evaluated the time-to-positivity (TTP) of bacterial cultures in febrile infants predate the use of continuous monitoring systems for blood cultures. Traditional bacterial culturing techniques require direct observation of broth turbidity and subsequent subculturing onto chocolate and sheep blood agar, typically occurring only once daily.²⁴ Current commercially available continuous monitoring bacterial culture systems decrease TTP by immediately alerting laboratory technicians to bacterial growth through the detection of ¹⁴CO2 released by organisms utilizing radiolabeled glucose in growth media.²⁴ In addition, many studies supporting the evaluation of febrile infants in the hospital for a 48-hour period include those in ICU settings,²⁵ with medically complex histories,²⁴ and aged < 28 days admitted in the NICU,¹⁵ where pathogens with longer incubation times are frequently seen.

When a Longer Observation Period May Be Warranted

What You Should Do Instead: Limit Hospitalization to a Maximum of 36 Hours

For well-appearing febrile infants between 0–90 days of age hospitalized for observation and awaiting bacterial culture results, providers should consider discharge at 36 hours or less, rather than 48 hours, if blood, urine, and CSF cultures do not show bacterial growth. In a large health system, researchers implemented an evidence-based care process model for febrile infants to provide specific guidelines for laboratory testing, criteria for admission, and recommendation for discontinuation of empiric antibiotics and discharge after 36 hours in infants with negative bacterial cultures. These changes led to a 27% reduction in the length of hospital stay and 23% reduction in inpatient costs without any cases of missed bacteremia.²¹ The reduction in the in-hospital observation duration to 24 hours of culture incubation for well-appearing febrile infants has been advocated ³² and is a common practice for infants with appropriate follow up and parental assurance. This recommendation is supported by the following:

• Recent data showing the overwhelming majority of pathogens will be identified by blood culture <24 hours in infants aged 0-90 days³² with blood culture TTP in infants aged 0-30 days being either no different²⁶ or potentially shorter³²
• Studies showing that infants meeting low-risk clinical and laboratory profiles further reduce the likelihood of identifying serious bacterial infection after 24 hours to 0.3%.¹

• Determine if febrile infants aged 0-90 days are at low risk for serious bacterial infection and obtain appropriate bacterial cultures.
• If hospitalized for observation, discharge low-risk febrile infants aged 0–90 days after 36 hours or less if bacterial cultures remain negative.
• If hospitalized for observation, consider reducing the length of inpatient observation for low-risk febrile infants aged 0–90 days with reliable follow-up to 24 hours or less when the culture results are negative.

Monitoring patients in the hospital for greater than 36 hours of bacterial culture incubation is unnecessary for patients similar to the 3 week-old full-term infant in the case presentation, who are at low risk for serious bacterial infection based on available scoring systems and have negative cultures. If patients are not deemed low risk, have an incomplete laboratory evaluation, or have had prior antibiotic treatment, longer observation in the hospital may be warranted. Close reassessment of the rare patients whose blood cultures return positive after 36 hours is necessary, but their outcomes are excellent, especially in well-appearing infants.^7,33

What do you do?

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason”? Let us know what you do in your practice and propose ideas for other “Things We Do for No Reason” topics. Please join in the conversation online at Twitter (#TWDFNR)/Facebook and don’t forget to “Like It” on Facebook or retweet it on Twitter. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing[email protected].

Disclosures

There are no conflicts of interest relevant to this work reported by any of the authors.

For adults with serious illness, consulting with a palliative care team within 3 days of hospital admission significantly reduced hospital costs, according to findings from a systematic review and meta-analysis.

In a pooled analysis of six cohort studies, average cost savings per admission were $3,237 (95% confidence interval, –$3,581 to −$2,893) overall, $4,251 for patients with cancer, and $2,105 for patients with other serious illnesses (all P values less than .001), reported Peter May, PhD, of Trinity College Dublin, and his associates.

In this latter group, prompt palliative care consultations saved more when patients had at least four comorbidities rather than two or fewer comorbidities, the reviewers wrote. The report was published in JAMA Internal Medicine.

About one in four Medicare beneficiaries dies in acute care hospitals, often after weeks of intensive, costly care that may not reflect personal wishes, according to an earlier study (JAMA. 2013;309:470-7). Economic studies have tried to pinpoint the cost savings of palliative care. These studies have found it important to consider both the clinical characteristics of patients and the amount of time between admission and palliative consultations, the reviewers noted. However, heterogeneity among older studies had precluded pooled analyses.

The six studies in this meta-analysis were identified by a search of Embase, PsycINFO, CENTRAL, PubMed, CINAHL, and EconLit databases for economic studies of hospital-based palliative care consultations. The studies were published between 2008 and 2017 and included 133,118 adults with cancer, chronic obstructive pulmonary disease, major organ failure, AIDS/HIV, or serious neurodegenerative disease. Patients tended to be in their 60s and were usually Medicare beneficiaries, although one study focused only on Medicaid enrollees. Forty-one percent of patients had a primary diagnosis of cancer, and 93% were discharged alive. Most also had at least two comorbidities. Only 3.6% received a palliative care consultation (range, 2.2% to 22.3%).

The link that they found between more comorbidities and greater cost savings “is the reverse of prior research that assumed that long-stay, high-cost hospitalized patients could not have their care trajectories affected by palliative care,” the researchers wrote. “Current palliative care provision in the United States is characterized by widespread understaffing. Our results suggest that acute care hospitals may be able to reduce costs for this population by increasing palliative care capacity to meet national guidelines.”

Dr. May received grant support from The Atlantic Philanthropies. The reviewers reported having no conflicts of interest.

SOURCE: May P et al. JAMA Intern Med. 2018 Apr 30. doi: 10.1001/jamainternmed.2018.0750.

For adults with serious illness, consulting with a palliative care team within 3 days of hospital admission significantly reduced hospital costs, according to findings from a systematic review and meta-analysis.

In a pooled analysis of six cohort studies, average cost savings per admission were $3,237 (95% confidence interval, –$3,581 to −$2,893) overall, $4,251 for patients with cancer, and $2,105 for patients with other serious illnesses (all P values less than .001), reported Peter May, PhD, of Trinity College Dublin, and his associates.

In this latter group, prompt palliative care consultations saved more when patients had at least four comorbidities rather than two or fewer comorbidities, the reviewers wrote. The report was published in JAMA Internal Medicine.

About one in four Medicare beneficiaries dies in acute care hospitals, often after weeks of intensive, costly care that may not reflect personal wishes, according to an earlier study (JAMA. 2013;309:470-7). Economic studies have tried to pinpoint the cost savings of palliative care. These studies have found it important to consider both the clinical characteristics of patients and the amount of time between admission and palliative consultations, the reviewers noted. However, heterogeneity among older studies had precluded pooled analyses.

The six studies in this meta-analysis were identified by a search of Embase, PsycINFO, CENTRAL, PubMed, CINAHL, and EconLit databases for economic studies of hospital-based palliative care consultations. The studies were published between 2008 and 2017 and included 133,118 adults with cancer, chronic obstructive pulmonary disease, major organ failure, AIDS/HIV, or serious neurodegenerative disease. Patients tended to be in their 60s and were usually Medicare beneficiaries, although one study focused only on Medicaid enrollees. Forty-one percent of patients had a primary diagnosis of cancer, and 93% were discharged alive. Most also had at least two comorbidities. Only 3.6% received a palliative care consultation (range, 2.2% to 22.3%).

The link that they found between more comorbidities and greater cost savings “is the reverse of prior research that assumed that long-stay, high-cost hospitalized patients could not have their care trajectories affected by palliative care,” the researchers wrote. “Current palliative care provision in the United States is characterized by widespread understaffing. Our results suggest that acute care hospitals may be able to reduce costs for this population by increasing palliative care capacity to meet national guidelines.”

Dr. May received grant support from The Atlantic Philanthropies. The reviewers reported having no conflicts of interest.

SOURCE: May P et al. JAMA Intern Med. 2018 Apr 30. doi: 10.1001/jamainternmed.2018.0750.

For adults with serious illness, consulting with a palliative care team within 3 days of hospital admission significantly reduced hospital costs, according to findings from a systematic review and meta-analysis.

In a pooled analysis of six cohort studies, average cost savings per admission were $3,237 (95% confidence interval, –$3,581 to −$2,893) overall, $4,251 for patients with cancer, and $2,105 for patients with other serious illnesses (all P values less than .001), reported Peter May, PhD, of Trinity College Dublin, and his associates.

In this latter group, prompt palliative care consultations saved more when patients had at least four comorbidities rather than two or fewer comorbidities, the reviewers wrote. The report was published in JAMA Internal Medicine.

About one in four Medicare beneficiaries dies in acute care hospitals, often after weeks of intensive, costly care that may not reflect personal wishes, according to an earlier study (JAMA. 2013;309:470-7). Economic studies have tried to pinpoint the cost savings of palliative care. These studies have found it important to consider both the clinical characteristics of patients and the amount of time between admission and palliative consultations, the reviewers noted. However, heterogeneity among older studies had precluded pooled analyses.

The six studies in this meta-analysis were identified by a search of Embase, PsycINFO, CENTRAL, PubMed, CINAHL, and EconLit databases for economic studies of hospital-based palliative care consultations. The studies were published between 2008 and 2017 and included 133,118 adults with cancer, chronic obstructive pulmonary disease, major organ failure, AIDS/HIV, or serious neurodegenerative disease. Patients tended to be in their 60s and were usually Medicare beneficiaries, although one study focused only on Medicaid enrollees. Forty-one percent of patients had a primary diagnosis of cancer, and 93% were discharged alive. Most also had at least two comorbidities. Only 3.6% received a palliative care consultation (range, 2.2% to 22.3%).

The link that they found between more comorbidities and greater cost savings “is the reverse of prior research that assumed that long-stay, high-cost hospitalized patients could not have their care trajectories affected by palliative care,” the researchers wrote. “Current palliative care provision in the United States is characterized by widespread understaffing. Our results suggest that acute care hospitals may be able to reduce costs for this population by increasing palliative care capacity to meet national guidelines.”

Dr. May received grant support from The Atlantic Philanthropies. The reviewers reported having no conflicts of interest.

SOURCE: May P et al. JAMA Intern Med. 2018 Apr 30. doi: 10.1001/jamainternmed.2018.0750.

The research paradigm had been: examine a person’s traits or symptomsthen, then search for genes or gene variants that cause or contribute to them. But it was difficult for researchers to recontact people with genotypes of interest for “downstream” follow-up.

Now, the National Institute of Health (NIH) and Inova Health System are launching The Genomic Ascertainment Cohort (TGAC), a 2-year pilot project that will allow researchers to recall genotyped people and examine the influence of genes and gene variants on phenotypes. “This is essentially match-making between genes and gene variants and the researchers who study them,” says Richard Siegel, MD, PhD, TGAC co-organizer, and clinical director and chief of the Autoimmunity Branch at the National Institute of Arthritis and Musculoskeletal and Skin Diseases. For instance, a researcher might locate a genotype of interest in the database and ask participants with the genotype to come to the NIH Clinical Center in Bethesda.

Participating NIH institutes will contribute genome and exome sequences from existing research programs to the database. Another 1,000 patients will be recruited to have genome sequencing performed. Half of the new recruits will be from Hispanic backgrounds.

The researchers are aiming for 10,000 genomes and exomes to allow recruitment of people with both common and rarer gene variants.

“We’re trying to advance science in a new, creative, and slightly radical way,” says Leslie Biesecker, MD, TGAC co-organizer and chief of the Medical Genomics and Metabolic Genetics Branch at the National Human Genome Research Institute. “We’re especially interested in using this as a platform to test our ability to predict phenotype from genotype—one of the key underpinnings of predictive genomic medicine.”

The research paradigm had been: examine a person’s traits or symptomsthen, then search for genes or gene variants that cause or contribute to them. But it was difficult for researchers to recontact people with genotypes of interest for “downstream” follow-up.

Now, the National Institute of Health (NIH) and Inova Health System are launching The Genomic Ascertainment Cohort (TGAC), a 2-year pilot project that will allow researchers to recall genotyped people and examine the influence of genes and gene variants on phenotypes. “This is essentially match-making between genes and gene variants and the researchers who study them,” says Richard Siegel, MD, PhD, TGAC co-organizer, and clinical director and chief of the Autoimmunity Branch at the National Institute of Arthritis and Musculoskeletal and Skin Diseases. For instance, a researcher might locate a genotype of interest in the database and ask participants with the genotype to come to the NIH Clinical Center in Bethesda.

Participating NIH institutes will contribute genome and exome sequences from existing research programs to the database. Another 1,000 patients will be recruited to have genome sequencing performed. Half of the new recruits will be from Hispanic backgrounds.

The researchers are aiming for 10,000 genomes and exomes to allow recruitment of people with both common and rarer gene variants.

“We’re trying to advance science in a new, creative, and slightly radical way,” says Leslie Biesecker, MD, TGAC co-organizer and chief of the Medical Genomics and Metabolic Genetics Branch at the National Human Genome Research Institute. “We’re especially interested in using this as a platform to test our ability to predict phenotype from genotype—one of the key underpinnings of predictive genomic medicine.”

The research paradigm had been: examine a person’s traits or symptomsthen, then search for genes or gene variants that cause or contribute to them. But it was difficult for researchers to recontact people with genotypes of interest for “downstream” follow-up.

Now, the National Institute of Health (NIH) and Inova Health System are launching The Genomic Ascertainment Cohort (TGAC), a 2-year pilot project that will allow researchers to recall genotyped people and examine the influence of genes and gene variants on phenotypes. “This is essentially match-making between genes and gene variants and the researchers who study them,” says Richard Siegel, MD, PhD, TGAC co-organizer, and clinical director and chief of the Autoimmunity Branch at the National Institute of Arthritis and Musculoskeletal and Skin Diseases. For instance, a researcher might locate a genotype of interest in the database and ask participants with the genotype to come to the NIH Clinical Center in Bethesda.

Participating NIH institutes will contribute genome and exome sequences from existing research programs to the database. Another 1,000 patients will be recruited to have genome sequencing performed. Half of the new recruits will be from Hispanic backgrounds.

The researchers are aiming for 10,000 genomes and exomes to allow recruitment of people with both common and rarer gene variants.

“We’re trying to advance science in a new, creative, and slightly radical way,” says Leslie Biesecker, MD, TGAC co-organizer and chief of the Medical Genomics and Metabolic Genetics Branch at the National Human Genome Research Institute. “We’re especially interested in using this as a platform to test our ability to predict phenotype from genotype—one of the key underpinnings of predictive genomic medicine.”

When Surgeon General Jerome Adams issued an advisory calling for more people to carry naloxone — not just people at overdose risk, but also friends and family — experts and advocates were almost giddy.

This is an “unequivocally positive” step forward, said Leo Beletsky, an associate professor of law and health sciences at Northeastern University.

And not necessarily a surprise. Adams, who previously was Indiana’s health commissioner, was recruited to be the nation’s top doctor in part because of his work with then-Gov. Mike Pence, now the vice president. In Indiana, Adams pushed for harm-reduction approaches, which included expanded access to naloxone and the implementation of a needle exchange to combat the state’s much-publicized HIV outbreak, which began in 2015 and was linked to injection drug use.

Others cautioned, though, that his have-naloxone-will-carry recommendation is at best limited in what it can achieve, in part because the drug is relatively expensive.

Kaiser Health News breaks down what the advisory means, experts’ concerns and what policy approaches may be in the pipeline.

Many public health advocates applaud the surgeon general’s position.

Naloxone, which is a drug that can keep drug users alive by reversing opioid overdoses, is viewed by many as the cornerstone of the harm-reduction approach to the epidemic. Experts say people with addiction problems should carry it, and so should their family, friends and acquaintances.

“We want to put it more in reach,” said Traci Green, an associate professor of emergency medicine and community health sciences at Boston University, who has extensively researched the opioid abuse crisis. “It could not have been a better endorsement.”

Others, including Diane Goodman, who penned a recent Medscape commentary reflecting on the advisory, wonder whether this is a “rational” response to the scourge, since opioid addiction is one of many health problems people might encounter in everyday life and for which treatment options are still limited.

“I’m not sure it makes much more sense than any of us carrying a bottle of nitroglycerin to treat patients with end-stage angina,” wrote Goodman, an acute-care nurse practitioner, referring to chest pain.

“What, exactly, are we offering to addicts once their condition has been reversed?” she asked, noting that without treatment and therapy programs that help wean people from addiction “the odds of survival for any length of time remain low, no matter how much reversal medication is kept nearby.”

Results would likely be limited by naloxone’s price tag.

Take Baltimore, which has been hit particularly hard by the opioid epidemic. Its health department already has pushed for more people to carry naloxone.

But the drug’s price is an issue, said Dr. Leana Wen, the city’s health commissioner, and an emergency physician. She suggested that the federal government negotiate directly for a lower price, or give more money to organizations and agencies like hers so they can afford to maintain an adequate supply.

“Every day, people are calling us at the Baltimore City Health Department and requesting naloxone, and I have to tell them I can’t afford for them to have it,” Wen said.

The drug is available in generic form, which can be stored in a vial and injected via a needle, as well as in patented products, such as the nasal spray Narcan, sold by ADAPT Pharmaceuticals, and Kaleo’s Evzio, a talking auto-injector.

Generic naloxone costs $20 to $40 per dose. Narcan, the nasal spray, costs $125 for a two-dose carton, according to ADAPT’s website. A two-pack of Evzio costs close to $4,000, according to GoodRx.

Health departments and first responders qualify for a discounted rate of $75 per carton of Narcan. Kaleo has made Evzio coupons available to consumers, so that some will not have a copay, and it advertises a discount for federal and state agencies.

Skeptics point out that similar methods have been used to build brand loyalty and potentially make a particular product a household name. That’s how Epi-Pen became synonymous with epinephrine for the treatment of anaphylactic shock.

“There’s clearly some overlap” here between the pricing strategies used by naloxone manufacturers and Epi-Pen distributor Mylan, said Richard Evans, co-founder of SSR Health, which tracks the pharmaceutical industry.

But it’s not a perfect comparison. The presence of low-cost generics changes the calculus, he said, as does the different level of demand.

Nonprofit organizations and health care providers keenly feel the pressures of increasing demand and cost.

Experts say price breaks on naloxone are not sufficient to cover the costs on the ground.

“Sixty-four thousand people lost their lives [nationally in 2016] — that’s someone every 12 minutes,” said Justin Phillips, executive director of Overdose Lifeline, an Indianapolis-based nonprofit. “Ten free kits is not going to be enough.”

Phillips said her organization relies on generic naloxone, which is the least expensive formulation. It’s the only feasible option, using dedicated grant money the group received from the state attorney general’s office as part of a program funded by a settlement with pharmaceutical companies.

But that money is almost dried up. “We need to be able to access naloxone — which I’m told is pennies to make — for the pennies it cost to make it,” Phillips said.

Phillips, who worked with Adams when he ran Indiana’s health department, said she has discussed the need for naloxone funding with the surgeon general, but never its price.

Pharmacies assess the hurdles of distribution.

Local pharmacies are key in this chain, but the overdose antidote is new territory for many pharmacists, said Randy Hitchens, the executive vice president of the Indiana Pharmacists Alliance. He said in 2015, when Adams began his push to get naloxone into the hands of drug users and their families, only one or two retail pharmacies carried it.

“This has always been an emergency room drug. Retail pharmacists typically were not used to dealing with [it],” Hitchens said. “A lot were probably saying, ‘What in the devil is naloxone?’”

Today, he estimates 60 to 70 percent of Indiana’s more than 1,100 retail pharmacies carry the drug. Walgreens, the pharmacy chain, has committed to stocking Narcan.

Access, though, is always subject to retail pressures.

“If pharmacies are not seeing a steady stream coming in asking for it, they won’t be incentivized to carry it on their shelves,” said Daniel Raymond, the deputy director of policy and planning for the Harm Reduction Coalition.

A patchwork of other decentralized sources for naloxone exist: syringe-exchange vans, county and state health departments, churches and community centers, all trying to find ways to get overdose medication into the hands of people who need it.

That supply stream “meets people where they are,” Raymond said, but those little programs don’t have the muscle to negotiate discounted prices.

“Individual health programs are trying to navigate the crisis on their own, but when you see … growing demand and limited supply, it’s a role for federal intervention,” Raymond said.

He’d like to see the federal government step in to negotiate prices where smaller programs can’t.

The surgeon general’s message is one part of Washington’s broader response to the epidemic. But even as Congress crafts an opioid epidemic response package, it’s not clear it will tackle these concerns.

In the House of Representatives’ Energy and Commerce Committee, one bill being discussed would require all state Medicaid programs to cover at least one form of naloxone. Currently, not all state Medicaid programs do so.

A Senate bill would authorize $300 million annually to equip first responders with naloxone.

But critics say those approaches still don’t address the underlying problems: cost and funding.

“You can either make naloxone available, at a much discounted price, or we need to have a lot more resources in order to purchase it,” Wen said. “I don’t care which one. My only concern is the health and well-being of our residents.”

Kaiser Health News (KHN) is a national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation which is not affiliated with Kaiser Permanente.

When Surgeon General Jerome Adams issued an advisory calling for more people to carry naloxone — not just people at overdose risk, but also friends and family — experts and advocates were almost giddy.

This is an “unequivocally positive” step forward, said Leo Beletsky, an associate professor of law and health sciences at Northeastern University.

And not necessarily a surprise. Adams, who previously was Indiana’s health commissioner, was recruited to be the nation’s top doctor in part because of his work with then-Gov. Mike Pence, now the vice president. In Indiana, Adams pushed for harm-reduction approaches, which included expanded access to naloxone and the implementation of a needle exchange to combat the state’s much-publicized HIV outbreak, which began in 2015 and was linked to injection drug use.

Others cautioned, though, that his have-naloxone-will-carry recommendation is at best limited in what it can achieve, in part because the drug is relatively expensive.

Kaiser Health News breaks down what the advisory means, experts’ concerns and what policy approaches may be in the pipeline.

Many public health advocates applaud the surgeon general’s position.

Naloxone, which is a drug that can keep drug users alive by reversing opioid overdoses, is viewed by many as the cornerstone of the harm-reduction approach to the epidemic. Experts say people with addiction problems should carry it, and so should their family, friends and acquaintances.

“We want to put it more in reach,” said Traci Green, an associate professor of emergency medicine and community health sciences at Boston University, who has extensively researched the opioid abuse crisis. “It could not have been a better endorsement.”

Others, including Diane Goodman, who penned a recent Medscape commentary reflecting on the advisory, wonder whether this is a “rational” response to the scourge, since opioid addiction is one of many health problems people might encounter in everyday life and for which treatment options are still limited.

“I’m not sure it makes much more sense than any of us carrying a bottle of nitroglycerin to treat patients with end-stage angina,” wrote Goodman, an acute-care nurse practitioner, referring to chest pain.

“What, exactly, are we offering to addicts once their condition has been reversed?” she asked, noting that without treatment and therapy programs that help wean people from addiction “the odds of survival for any length of time remain low, no matter how much reversal medication is kept nearby.”

Results would likely be limited by naloxone’s price tag.

Take Baltimore, which has been hit particularly hard by the opioid epidemic. Its health department already has pushed for more people to carry naloxone.

But the drug’s price is an issue, said Dr. Leana Wen, the city’s health commissioner, and an emergency physician. She suggested that the federal government negotiate directly for a lower price, or give more money to organizations and agencies like hers so they can afford to maintain an adequate supply.

“Every day, people are calling us at the Baltimore City Health Department and requesting naloxone, and I have to tell them I can’t afford for them to have it,” Wen said.

The drug is available in generic form, which can be stored in a vial and injected via a needle, as well as in patented products, such as the nasal spray Narcan, sold by ADAPT Pharmaceuticals, and Kaleo’s Evzio, a talking auto-injector.

Generic naloxone costs $20 to $40 per dose. Narcan, the nasal spray, costs $125 for a two-dose carton, according to ADAPT’s website. A two-pack of Evzio costs close to $4,000, according to GoodRx.

Health departments and first responders qualify for a discounted rate of $75 per carton of Narcan. Kaleo has made Evzio coupons available to consumers, so that some will not have a copay, and it advertises a discount for federal and state agencies.

Skeptics point out that similar methods have been used to build brand loyalty and potentially make a particular product a household name. That’s how Epi-Pen became synonymous with epinephrine for the treatment of anaphylactic shock.

“There’s clearly some overlap” here between the pricing strategies used by naloxone manufacturers and Epi-Pen distributor Mylan, said Richard Evans, co-founder of SSR Health, which tracks the pharmaceutical industry.

But it’s not a perfect comparison. The presence of low-cost generics changes the calculus, he said, as does the different level of demand.

Nonprofit organizations and health care providers keenly feel the pressures of increasing demand and cost.

Experts say price breaks on naloxone are not sufficient to cover the costs on the ground.

“Sixty-four thousand people lost their lives [nationally in 2016] — that’s someone every 12 minutes,” said Justin Phillips, executive director of Overdose Lifeline, an Indianapolis-based nonprofit. “Ten free kits is not going to be enough.”

Phillips said her organization relies on generic naloxone, which is the least expensive formulation. It’s the only feasible option, using dedicated grant money the group received from the state attorney general’s office as part of a program funded by a settlement with pharmaceutical companies.

But that money is almost dried up. “We need to be able to access naloxone — which I’m told is pennies to make — for the pennies it cost to make it,” Phillips said.

Phillips, who worked with Adams when he ran Indiana’s health department, said she has discussed the need for naloxone funding with the surgeon general, but never its price.

Pharmacies assess the hurdles of distribution.

Local pharmacies are key in this chain, but the overdose antidote is new territory for many pharmacists, said Randy Hitchens, the executive vice president of the Indiana Pharmacists Alliance. He said in 2015, when Adams began his push to get naloxone into the hands of drug users and their families, only one or two retail pharmacies carried it.

“This has always been an emergency room drug. Retail pharmacists typically were not used to dealing with [it],” Hitchens said. “A lot were probably saying, ‘What in the devil is naloxone?’”

Today, he estimates 60 to 70 percent of Indiana’s more than 1,100 retail pharmacies carry the drug. Walgreens, the pharmacy chain, has committed to stocking Narcan.

Access, though, is always subject to retail pressures.

“If pharmacies are not seeing a steady stream coming in asking for it, they won’t be incentivized to carry it on their shelves,” said Daniel Raymond, the deputy director of policy and planning for the Harm Reduction Coalition.

A patchwork of other decentralized sources for naloxone exist: syringe-exchange vans, county and state health departments, churches and community centers, all trying to find ways to get overdose medication into the hands of people who need it.

That supply stream “meets people where they are,” Raymond said, but those little programs don’t have the muscle to negotiate discounted prices.

“Individual health programs are trying to navigate the crisis on their own, but when you see … growing demand and limited supply, it’s a role for federal intervention,” Raymond said.

He’d like to see the federal government step in to negotiate prices where smaller programs can’t.

The surgeon general’s message is one part of Washington’s broader response to the epidemic. But even as Congress crafts an opioid epidemic response package, it’s not clear it will tackle these concerns.

In the House of Representatives’ Energy and Commerce Committee, one bill being discussed would require all state Medicaid programs to cover at least one form of naloxone. Currently, not all state Medicaid programs do so.

A Senate bill would authorize $300 million annually to equip first responders with naloxone.

But critics say those approaches still don’t address the underlying problems: cost and funding.

“You can either make naloxone available, at a much discounted price, or we need to have a lot more resources in order to purchase it,” Wen said. “I don’t care which one. My only concern is the health and well-being of our residents.”

Kaiser Health News (KHN) is a national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation which is not affiliated with Kaiser Permanente.

When Surgeon General Jerome Adams issued an advisory calling for more people to carry naloxone — not just people at overdose risk, but also friends and family — experts and advocates were almost giddy.

This is an “unequivocally positive” step forward, said Leo Beletsky, an associate professor of law and health sciences at Northeastern University.

And not necessarily a surprise. Adams, who previously was Indiana’s health commissioner, was recruited to be the nation’s top doctor in part because of his work with then-Gov. Mike Pence, now the vice president. In Indiana, Adams pushed for harm-reduction approaches, which included expanded access to naloxone and the implementation of a needle exchange to combat the state’s much-publicized HIV outbreak, which began in 2015 and was linked to injection drug use.

Others cautioned, though, that his have-naloxone-will-carry recommendation is at best limited in what it can achieve, in part because the drug is relatively expensive.

Kaiser Health News breaks down what the advisory means, experts’ concerns and what policy approaches may be in the pipeline.

Many public health advocates applaud the surgeon general’s position.

Naloxone, which is a drug that can keep drug users alive by reversing opioid overdoses, is viewed by many as the cornerstone of the harm-reduction approach to the epidemic. Experts say people with addiction problems should carry it, and so should their family, friends and acquaintances.

“We want to put it more in reach,” said Traci Green, an associate professor of emergency medicine and community health sciences at Boston University, who has extensively researched the opioid abuse crisis. “It could not have been a better endorsement.”

Others, including Diane Goodman, who penned a recent Medscape commentary reflecting on the advisory, wonder whether this is a “rational” response to the scourge, since opioid addiction is one of many health problems people might encounter in everyday life and for which treatment options are still limited.

“I’m not sure it makes much more sense than any of us carrying a bottle of nitroglycerin to treat patients with end-stage angina,” wrote Goodman, an acute-care nurse practitioner, referring to chest pain.

“What, exactly, are we offering to addicts once their condition has been reversed?” she asked, noting that without treatment and therapy programs that help wean people from addiction “the odds of survival for any length of time remain low, no matter how much reversal medication is kept nearby.”

Results would likely be limited by naloxone’s price tag.

Take Baltimore, which has been hit particularly hard by the opioid epidemic. Its health department already has pushed for more people to carry naloxone.

But the drug’s price is an issue, said Dr. Leana Wen, the city’s health commissioner, and an emergency physician. She suggested that the federal government negotiate directly for a lower price, or give more money to organizations and agencies like hers so they can afford to maintain an adequate supply.

“Every day, people are calling us at the Baltimore City Health Department and requesting naloxone, and I have to tell them I can’t afford for them to have it,” Wen said.

The drug is available in generic form, which can be stored in a vial and injected via a needle, as well as in patented products, such as the nasal spray Narcan, sold by ADAPT Pharmaceuticals, and Kaleo’s Evzio, a talking auto-injector.

Generic naloxone costs $20 to $40 per dose. Narcan, the nasal spray, costs $125 for a two-dose carton, according to ADAPT’s website. A two-pack of Evzio costs close to $4,000, according to GoodRx.

Health departments and first responders qualify for a discounted rate of $75 per carton of Narcan. Kaleo has made Evzio coupons available to consumers, so that some will not have a copay, and it advertises a discount for federal and state agencies.

Skeptics point out that similar methods have been used to build brand loyalty and potentially make a particular product a household name. That’s how Epi-Pen became synonymous with epinephrine for the treatment of anaphylactic shock.

“There’s clearly some overlap” here between the pricing strategies used by naloxone manufacturers and Epi-Pen distributor Mylan, said Richard Evans, co-founder of SSR Health, which tracks the pharmaceutical industry.

But it’s not a perfect comparison. The presence of low-cost generics changes the calculus, he said, as does the different level of demand.

Nonprofit organizations and health care providers keenly feel the pressures of increasing demand and cost.

Experts say price breaks on naloxone are not sufficient to cover the costs on the ground.

“Sixty-four thousand people lost their lives [nationally in 2016] — that’s someone every 12 minutes,” said Justin Phillips, executive director of Overdose Lifeline, an Indianapolis-based nonprofit. “Ten free kits is not going to be enough.”

Phillips said her organization relies on generic naloxone, which is the least expensive formulation. It’s the only feasible option, using dedicated grant money the group received from the state attorney general’s office as part of a program funded by a settlement with pharmaceutical companies.

But that money is almost dried up. “We need to be able to access naloxone — which I’m told is pennies to make — for the pennies it cost to make it,” Phillips said.

Phillips, who worked with Adams when he ran Indiana’s health department, said she has discussed the need for naloxone funding with the surgeon general, but never its price.

Pharmacies assess the hurdles of distribution.

Local pharmacies are key in this chain, but the overdose antidote is new territory for many pharmacists, said Randy Hitchens, the executive vice president of the Indiana Pharmacists Alliance. He said in 2015, when Adams began his push to get naloxone into the hands of drug users and their families, only one or two retail pharmacies carried it.

“This has always been an emergency room drug. Retail pharmacists typically were not used to dealing with [it],” Hitchens said. “A lot were probably saying, ‘What in the devil is naloxone?’”

Today, he estimates 60 to 70 percent of Indiana’s more than 1,100 retail pharmacies carry the drug. Walgreens, the pharmacy chain, has committed to stocking Narcan.

Access, though, is always subject to retail pressures.

“If pharmacies are not seeing a steady stream coming in asking for it, they won’t be incentivized to carry it on their shelves,” said Daniel Raymond, the deputy director of policy and planning for the Harm Reduction Coalition.

A patchwork of other decentralized sources for naloxone exist: syringe-exchange vans, county and state health departments, churches and community centers, all trying to find ways to get overdose medication into the hands of people who need it.

That supply stream “meets people where they are,” Raymond said, but those little programs don’t have the muscle to negotiate discounted prices.

“Individual health programs are trying to navigate the crisis on their own, but when you see … growing demand and limited supply, it’s a role for federal intervention,” Raymond said.

He’d like to see the federal government step in to negotiate prices where smaller programs can’t.

The surgeon general’s message is one part of Washington’s broader response to the epidemic. But even as Congress crafts an opioid epidemic response package, it’s not clear it will tackle these concerns.

In the House of Representatives’ Energy and Commerce Committee, one bill being discussed would require all state Medicaid programs to cover at least one form of naloxone. Currently, not all state Medicaid programs do so.

A Senate bill would authorize $300 million annually to equip first responders with naloxone.

But critics say those approaches still don’t address the underlying problems: cost and funding.

“You can either make naloxone available, at a much discounted price, or we need to have a lot more resources in order to purchase it,” Wen said. “I don’t care which one. My only concern is the health and well-being of our residents.”

Kaiser Health News (KHN) is a national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation which is not affiliated with Kaiser Permanente.

As a hospital-acquired condition responsible for a significant share of preventable deaths in the United States,¹ venous thromboembolism (VTE) prevention should remain a high priority for healthcare organizations. Pursuant to the goal of reducing the frequency of this and other hospital-acquired conditions, several performance measures have been developed by third-party payers in the United States to provide incentives for inpatients to receive prophylaxis measures appropriate to their specific level of risk. Perhaps the best known of these is the Hospital Value-Based Purchasing Program, initiated by the Center for Medicare and Medicaid Studies (CMS) in 2013 as a provision of the Affordable Care Act.² The Joint Commission, as steward of the 6 VTE-related measures,³ dictates the criteria for assessing performance. However, recent adjustments to one of these measures have been performed in such a way that neglects real-world considerations faced by providers and threatens to delegitimize the important role that value-based purchasing should have in reimbursement.

Effective in 2017, the guidelines pertaining to abstraction-based reporting added a new component to the VTE-6 measure, which applies to those inpatients not ordered to receive mechanical or pharmacologic prophylaxis who go on to suffer VTE. Specifically, it is concerned with how accurately hospitals stratify such patients as low risk before the decision is made to not order either method of prophylaxis. With the update, to satisfy the measure, a formal assessment confirming a patient’s low-risk status must have been documented between arrival and the time the VTE diagnostic test was performed. The guidelines explicitly note that only 3 risk assessment models (RAMs) are accepted, including the Caprini DVT Prediction Score, Padua Prediction Score, and IMPROVE VTE Risk Score.⁴ The rationale for this addition to the measure clearly is to protect patients from being incorrectly designated as low risk and subsequently receiving inadequate prophylaxis that could increase their likelihood of developing preventable VTE. Unfortunately, in its current form, it imposes a substantial burden on providers and healthcare organizations, without much promise of significantly reducing rates of this pervasive threat to patient safety.

Although the aim of reducing the incidence of VTE is laudable, this updated requirement for VTE-6 is problematic on several levels. First, there is considerable uncertainty regarding how to implement the RAMs clinically in a user-friendly way that is conducive to their intended use. Due to limitations in most computerized physician order entry systems, it is not feasible to mandate the RAMs for only those patients not ordered for VTE prophylaxis (nor would it be sensible to restrict performing the assessment to low-risk patients, as the point of RAMs is to help risk stratify and not simply validate whatever determinations were already made by other means). As virtually every class of inpatient has some risk of VTE development, these factors effectively require that a score be tabulated on all admitted patients, giving the measure an enormous footprint on clinical operations. This is important because the permissible RAMs can sometimes be quite burdensome to complete faithfully. For instance, the Caprini Score necessitates the fairly prodigious collection and input of up to 26 data points. Some of the questions require exceedingly granular data, such as whether there is any “history of unexplained stillborn infant, recurrent spontaneous abortion (more than 3), premature birth with toxemia or growth restricted infant.”⁵ This clearly is far outside the scope of most focused admission assessments. Already deluged with the number of clicks inherent to the workflow of most electronic health records,⁶ it seems likely that some providers default to selecting “no” for such prompts as a time-saving measure, potentially sabotaging the goal of linking patients with a risk-appropriate method of prophylaxis. Meanwhile, those who are diligent about completing the assessment honestly will find themselves rewarded with less time to dedicate to other critical aspects of patient care.⁷

The small number of RAMs accepted under the measure also fails to account for the breadth of clinical circumstances providers faced. Although the permitted models are validated in certain patient populations, they exclude some that might be better suited for many practice environments. The University of California San Diego “3 bucket” design, for instance, has been shown to result in high levels of risk-appropriate prophylaxis, has high inter-user agreement, and perhaps most importantly, is relatively quick and easy to use.⁸ Also critical, it is easier to integrate into the admission workflow for under-resourced hospitals that might not have the ability to incorporate a point-based risk score calculator into their electronic health records.

Finally, the relative abruptness with which the changes were made complicated the task for institutions to integrate the RAMs into their applicable order sets in a user-friendly fashion. The new guidelines were released only 6 months before taking effect,⁹ and the RAM requirement was not widely advertised. This left a fairly short window that does not seem to reflect an understanding by the Joint Commission of the process required by hospitals to make such a transition responsibly. This should involve obtaining inputs from multiple specialty stakeholders on which RAM to employ, working with information system specialists on how to restructure key order sets, and education of end-users on how to apply them correctly.¹⁰

For these reasons, the rollout of the VTE-6 update falls well short of its ambitions. Satisfying the measure necessitates a substantial investment of time and effort by providers and yet forcing the use of such decidedly imperfect RAMs could paradoxically worsen accurate risk stratification and appropriate use of prophylaxis. Also, while it represents only a small slice of pay-for-performance initiatives, its broader impact should not be underestimated. Unlike many of the more specific items, the VTE measures affect the workflow related to virtually all hospitalized patients. Therefore, it is imperative that regulators “get it right,” as it might only take one poorly conceived mandate of this type to risk permanently souring providers and hospitals on the idea of value-based purchasing. The Joint Commission and CMS ought to seriously consider retracting the new provisions until the role of RAMs for VTE prevention is better understood. This would buy time to reconfigure the measure in a way that is compatible with actual clinical care and for hospitals to thoughtfully design how new requirements can best be implemented.

Disclosurses

The author has nothing to disclose.

As a hospital-acquired condition responsible for a significant share of preventable deaths in the United States,¹ venous thromboembolism (VTE) prevention should remain a high priority for healthcare organizations. Pursuant to the goal of reducing the frequency of this and other hospital-acquired conditions, several performance measures have been developed by third-party payers in the United States to provide incentives for inpatients to receive prophylaxis measures appropriate to their specific level of risk. Perhaps the best known of these is the Hospital Value-Based Purchasing Program, initiated by the Center for Medicare and Medicaid Studies (CMS) in 2013 as a provision of the Affordable Care Act.² The Joint Commission, as steward of the 6 VTE-related measures,³ dictates the criteria for assessing performance. However, recent adjustments to one of these measures have been performed in such a way that neglects real-world considerations faced by providers and threatens to delegitimize the important role that value-based purchasing should have in reimbursement.

Effective in 2017, the guidelines pertaining to abstraction-based reporting added a new component to the VTE-6 measure, which applies to those inpatients not ordered to receive mechanical or pharmacologic prophylaxis who go on to suffer VTE. Specifically, it is concerned with how accurately hospitals stratify such patients as low risk before the decision is made to not order either method of prophylaxis. With the update, to satisfy the measure, a formal assessment confirming a patient’s low-risk status must have been documented between arrival and the time the VTE diagnostic test was performed. The guidelines explicitly note that only 3 risk assessment models (RAMs) are accepted, including the Caprini DVT Prediction Score, Padua Prediction Score, and IMPROVE VTE Risk Score.⁴ The rationale for this addition to the measure clearly is to protect patients from being incorrectly designated as low risk and subsequently receiving inadequate prophylaxis that could increase their likelihood of developing preventable VTE. Unfortunately, in its current form, it imposes a substantial burden on providers and healthcare organizations, without much promise of significantly reducing rates of this pervasive threat to patient safety.

Although the aim of reducing the incidence of VTE is laudable, this updated requirement for VTE-6 is problematic on several levels. First, there is considerable uncertainty regarding how to implement the RAMs clinically in a user-friendly way that is conducive to their intended use. Due to limitations in most computerized physician order entry systems, it is not feasible to mandate the RAMs for only those patients not ordered for VTE prophylaxis (nor would it be sensible to restrict performing the assessment to low-risk patients, as the point of RAMs is to help risk stratify and not simply validate whatever determinations were already made by other means). As virtually every class of inpatient has some risk of VTE development, these factors effectively require that a score be tabulated on all admitted patients, giving the measure an enormous footprint on clinical operations. This is important because the permissible RAMs can sometimes be quite burdensome to complete faithfully. For instance, the Caprini Score necessitates the fairly prodigious collection and input of up to 26 data points. Some of the questions require exceedingly granular data, such as whether there is any “history of unexplained stillborn infant, recurrent spontaneous abortion (more than 3), premature birth with toxemia or growth restricted infant.”⁵ This clearly is far outside the scope of most focused admission assessments. Already deluged with the number of clicks inherent to the workflow of most electronic health records,⁶ it seems likely that some providers default to selecting “no” for such prompts as a time-saving measure, potentially sabotaging the goal of linking patients with a risk-appropriate method of prophylaxis. Meanwhile, those who are diligent about completing the assessment honestly will find themselves rewarded with less time to dedicate to other critical aspects of patient care.⁷

The small number of RAMs accepted under the measure also fails to account for the breadth of clinical circumstances providers faced. Although the permitted models are validated in certain patient populations, they exclude some that might be better suited for many practice environments. The University of California San Diego “3 bucket” design, for instance, has been shown to result in high levels of risk-appropriate prophylaxis, has high inter-user agreement, and perhaps most importantly, is relatively quick and easy to use.⁸ Also critical, it is easier to integrate into the admission workflow for under-resourced hospitals that might not have the ability to incorporate a point-based risk score calculator into their electronic health records.

Finally, the relative abruptness with which the changes were made complicated the task for institutions to integrate the RAMs into their applicable order sets in a user-friendly fashion. The new guidelines were released only 6 months before taking effect,⁹ and the RAM requirement was not widely advertised. This left a fairly short window that does not seem to reflect an understanding by the Joint Commission of the process required by hospitals to make such a transition responsibly. This should involve obtaining inputs from multiple specialty stakeholders on which RAM to employ, working with information system specialists on how to restructure key order sets, and education of end-users on how to apply them correctly.¹⁰

For these reasons, the rollout of the VTE-6 update falls well short of its ambitions. Satisfying the measure necessitates a substantial investment of time and effort by providers and yet forcing the use of such decidedly imperfect RAMs could paradoxically worsen accurate risk stratification and appropriate use of prophylaxis. Also, while it represents only a small slice of pay-for-performance initiatives, its broader impact should not be underestimated. Unlike many of the more specific items, the VTE measures affect the workflow related to virtually all hospitalized patients. Therefore, it is imperative that regulators “get it right,” as it might only take one poorly conceived mandate of this type to risk permanently souring providers and hospitals on the idea of value-based purchasing. The Joint Commission and CMS ought to seriously consider retracting the new provisions until the role of RAMs for VTE prevention is better understood. This would buy time to reconfigure the measure in a way that is compatible with actual clinical care and for hospitals to thoughtfully design how new requirements can best be implemented.

Disclosurses

The author has nothing to disclose.

As a hospital-acquired condition responsible for a significant share of preventable deaths in the United States,¹ venous thromboembolism (VTE) prevention should remain a high priority for healthcare organizations. Pursuant to the goal of reducing the frequency of this and other hospital-acquired conditions, several performance measures have been developed by third-party payers in the United States to provide incentives for inpatients to receive prophylaxis measures appropriate to their specific level of risk. Perhaps the best known of these is the Hospital Value-Based Purchasing Program, initiated by the Center for Medicare and Medicaid Studies (CMS) in 2013 as a provision of the Affordable Care Act.² The Joint Commission, as steward of the 6 VTE-related measures,³ dictates the criteria for assessing performance. However, recent adjustments to one of these measures have been performed in such a way that neglects real-world considerations faced by providers and threatens to delegitimize the important role that value-based purchasing should have in reimbursement.

Effective in 2017, the guidelines pertaining to abstraction-based reporting added a new component to the VTE-6 measure, which applies to those inpatients not ordered to receive mechanical or pharmacologic prophylaxis who go on to suffer VTE. Specifically, it is concerned with how accurately hospitals stratify such patients as low risk before the decision is made to not order either method of prophylaxis. With the update, to satisfy the measure, a formal assessment confirming a patient’s low-risk status must have been documented between arrival and the time the VTE diagnostic test was performed. The guidelines explicitly note that only 3 risk assessment models (RAMs) are accepted, including the Caprini DVT Prediction Score, Padua Prediction Score, and IMPROVE VTE Risk Score.⁴ The rationale for this addition to the measure clearly is to protect patients from being incorrectly designated as low risk and subsequently receiving inadequate prophylaxis that could increase their likelihood of developing preventable VTE. Unfortunately, in its current form, it imposes a substantial burden on providers and healthcare organizations, without much promise of significantly reducing rates of this pervasive threat to patient safety.

Although the aim of reducing the incidence of VTE is laudable, this updated requirement for VTE-6 is problematic on several levels. First, there is considerable uncertainty regarding how to implement the RAMs clinically in a user-friendly way that is conducive to their intended use. Due to limitations in most computerized physician order entry systems, it is not feasible to mandate the RAMs for only those patients not ordered for VTE prophylaxis (nor would it be sensible to restrict performing the assessment to low-risk patients, as the point of RAMs is to help risk stratify and not simply validate whatever determinations were already made by other means). As virtually every class of inpatient has some risk of VTE development, these factors effectively require that a score be tabulated on all admitted patients, giving the measure an enormous footprint on clinical operations. This is important because the permissible RAMs can sometimes be quite burdensome to complete faithfully. For instance, the Caprini Score necessitates the fairly prodigious collection and input of up to 26 data points. Some of the questions require exceedingly granular data, such as whether there is any “history of unexplained stillborn infant, recurrent spontaneous abortion (more than 3), premature birth with toxemia or growth restricted infant.”⁵ This clearly is far outside the scope of most focused admission assessments. Already deluged with the number of clicks inherent to the workflow of most electronic health records,⁶ it seems likely that some providers default to selecting “no” for such prompts as a time-saving measure, potentially sabotaging the goal of linking patients with a risk-appropriate method of prophylaxis. Meanwhile, those who are diligent about completing the assessment honestly will find themselves rewarded with less time to dedicate to other critical aspects of patient care.⁷

The small number of RAMs accepted under the measure also fails to account for the breadth of clinical circumstances providers faced. Although the permitted models are validated in certain patient populations, they exclude some that might be better suited for many practice environments. The University of California San Diego “3 bucket” design, for instance, has been shown to result in high levels of risk-appropriate prophylaxis, has high inter-user agreement, and perhaps most importantly, is relatively quick and easy to use.⁸ Also critical, it is easier to integrate into the admission workflow for under-resourced hospitals that might not have the ability to incorporate a point-based risk score calculator into their electronic health records.

Finally, the relative abruptness with which the changes were made complicated the task for institutions to integrate the RAMs into their applicable order sets in a user-friendly fashion. The new guidelines were released only 6 months before taking effect,⁹ and the RAM requirement was not widely advertised. This left a fairly short window that does not seem to reflect an understanding by the Joint Commission of the process required by hospitals to make such a transition responsibly. This should involve obtaining inputs from multiple specialty stakeholders on which RAM to employ, working with information system specialists on how to restructure key order sets, and education of end-users on how to apply them correctly.¹⁰

For these reasons, the rollout of the VTE-6 update falls well short of its ambitions. Satisfying the measure necessitates a substantial investment of time and effort by providers and yet forcing the use of such decidedly imperfect RAMs could paradoxically worsen accurate risk stratification and appropriate use of prophylaxis. Also, while it represents only a small slice of pay-for-performance initiatives, its broader impact should not be underestimated. Unlike many of the more specific items, the VTE measures affect the workflow related to virtually all hospitalized patients. Therefore, it is imperative that regulators “get it right,” as it might only take one poorly conceived mandate of this type to risk permanently souring providers and hospitals on the idea of value-based purchasing. The Joint Commission and CMS ought to seriously consider retracting the new provisions until the role of RAMs for VTE prevention is better understood. This would buy time to reconfigure the measure in a way that is compatible with actual clinical care and for hospitals to thoughtfully design how new requirements can best be implemented.

Disclosurses

The author has nothing to disclose.

Cirrhosis is a morbid condition characterized by complications such as ascites, gastrointestinal bleeding, and hepatic encephalopathy. These complications frequently require hospitalization, which is a substantial burden to the healthcare system. In 2012, liver disease was responsible for nearly 250,000 admissions across the United States, costing $3 billion.¹ Despite this substantial resource utilization, outcomes remain poor, with an inpatient mortality of 6%. For those that survive, many experience hospital readmission.

More generally, early readmission reflects poor quality of care in the US. In 2004, 30-day readmissions occurred in nearly 20% of Medicare beneficiaries and costed over $17 billion.² In response to this problem, the Affordable Care Act established the Hospital Readmissions Reduction Program (HRRP), which reduces Centers for Medicare & Medicaid Services (CMS) payments to hospitals with excess 30-day readmissions for high-risk conditions, including pneumonia and heart failure.³ Heart failure, in particular, has been the subject of numerous studies detailing risk factors and interventions to predict and prevent readmission.^4-6 Based on this extensive evidence, guidelines recommend disease management programs to reduce readmissions in this population.⁷ In contrast, readmission in the cirrhosis population has received limited attention.

We therefore conducted a systematic review aiming to examine the range of readmission risk noted in the literature, with a focus on the model for end-stage liver disease (MELD) score as a risk factor for readmission.

Search Strategy

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for conducting and reporting systematic reviews.⁸ A literature search was performed by a medical librarian using the following databases: Ovid MEDLINE, PubMed, EMBASE, CINAHL, the full Cochrane Library, Scopus, Google Scholar, and ClinicalTrials.gov. All the databases were searched from 2000 to May 2017. We did not include older reports because the review focused on contemporary care; earlier studies may not reflect current cirrhosis management. To ensure literature saturation, included articles’ reference lists were reviewed.

Search strategies were developed by combining database-specific subject headings and keywords for readmissions with those for cirrhosis or its complications (Supplementary Material). Google Scholar and ClinicalTrials.gov were searched using keywords only. All results were limited to the English language and those published in 2000 or later, but no other limits were applied.

Identified records were reviewed based on strict criteria. We excluded case reports, case series, reviews, editorials, letters, and meeting abstracts without final peer-reviewed publication. We also excluded studies of pediatric populations (age < 18 years), patients without cirrhosis, and patients with liver transplants. We excluded studies in which patients were not hospitalized at study onset and those where the index admission was for an elective procedure. Because our interest was to identify factors associated with early readmission, we excluded studies that did not report readmissions within 90 days or those with a mean or median follow-up of less than 30 days. We also excluded studies that did not examine the association between readmission and at least 1 independent variable or intervention. Duplicate reports of a common sample were excluded unless the duplicate provided additional information, and such reports were examined together in our synthesis.

Two authors identified potentially eligible records by independently screening titles and abstracts. At this stage, records that did not meet the eligibility criteria were excluded, and the reasons for exclusion were not recorded. Records with disagreement were retained for full-text review. After this initial exclusion of records, the remaining full-text records were reviewed independently. For this full-text review, we recorded exclusion reasons and disagreements were resolved through discussion.

Data Collection

Data were abstracted from each study by 2 authors independently and recorded in a REDCap database.⁹ Discrepancies were resolved through discussion. We recorded study characteristics, including study design, setting, population (including the inclusion/exclusion criteria, sample size, and patient and hospitalization characteristics), interventions, and comparisons. To facilitate comparisons across studies, we employed validated methods to approximate means and standard deviations (SD).¹⁰ We recorded detailed information on outcomes including readmissions, preventability, independent variables, and mortality. Studies that focused on a single independent factor or intervention were classified as “focused,” while those that examined multiple factors were classified as “broad.” We used the Newcastle–Ottawa Scale to assess the risk of bias in each study.¹¹ This instrument uses a 9-point scale to gauge methodological quality based on selection, group comparability, and exposure/outcome assessment.

Statistical Analysis

Analyses were performed using Stata 13.1 (StataCorp LP, College Station, Texas). We determined the pooled proportion of patients with 30-day readmission using a random-effects model, with the Freeman–Tukey double-arcsine transformation for meta-analysis of proportions.¹² We investigated the heterogeneity by stratifying analyses according to prespecified study characteristics, including “broad” versus “focused.” However, the readmission risk was not different in the stratified analysis; therefore, we chose to pool the findings. For point estimates, 95% confidence intervals (CIs) were calculated, and a P-value < .05 was considered statistically significant.

Search Results

Study Characteristics

Two studies were performed in Australia, 4 in Europe, and the remainder in North America. Twenty one of the 26 studies were retrospective cohort studies (Table 1). Twenty studies were single-center studies (of which half were performed at transplant centers), and 4 of the 6 multicenter studies were based on administrative data with large samples (173,254 patients). The inclusion/exclusion criteria varied widely (Supplementary Material). Some studies only included patients admitted for specific cirrhosis complications, while others included those admitted for any reason. Two studies excluded patients admitted in the prior 30 days, and 6 excluded patients discharged to hospice. The mean risk of bias score was 7.5 (SD 1.3) out of a possible 9 points, with most lacking an adequate description of follow-up and several lacking adjustment for confounders.

Outcomes

Readmission and MELD

Hospital readmission is a costly and common problem in the US.² In addition to the negative impact that readmissions have on patients’ lives,⁴⁰ readmissions are increasingly being used to measure quality. Unplanned 30-day readmissions are posted publicly, and excess readmissions for high-risk conditions are penalized through HRRP.³ Although HRRP does not currently include cirrhosis, the program has expanded to include several conditions that were not included in the initial iteration. Whether cirrhosis will be included in future iterations remains to be seen; however, increasing scrutiny is likely to continue. Of specific populations at risk, patients with cirrhosis are particularly vulnerable due to several features. Ascites management often requires hospitalization due to diuretic titration and poor access to paracentesis, and hepatic encephalopathy treatment requires complex lactulose titration.¹⁶ Other features of cirrhosis, such as gastrointestinal bleeding, infections, and renal failure, also place patients at risk of poor outcomes. The resulting readmission burden is high, with a pooled 30-day readmission rate of 26%. Other associated outcomes are also poor, with a consistent relationship between readmission and subsequent mortality.

We found striking heterogeneity in various aspects. First, the inclusion/exclusion criteria varied widely, both cirrhosis-specific (eg, spontaneous bacterial peritonitis) and more general (patients admitted within the prior 30 days). Some of these criteria may bias readmission estimates; the risk of readmission may be reduced in those on hospice, as patients forgo curative therapy. Additionally, an established risk factor for readmission is prior hospitalization⁴¹; excluding patients with prior admissions prohibits analysis of this variable. Another aspect is the capture of readmissions: readmissions outside of the index hospital were not included in most studies. In those that did include outside readmissions, the burden was sizeable: 17% in 1 single-center study and 23% in a multistate administrative database.^16,36 These outside readmissions must be included in future studies; they are as important as same-center readmissions both to patients and CMS.³ Despite this heterogeneity, the studies scored relatively high on the Newcastle–Ottawa risk of bias scale, with the only common deficiency being an inadequate description of follow-up.

Building on the findings of this review, an important step will be the design of interventions to reduce readmissions. Such interventions require a full understanding of this population’s characteristics and needs. Critically, we found a lack of data on social determinants of health. Impairments in these factors are well-established contributors to readmission risk in other populations,^4,40 and are highly prevalent in cirrhosis.⁴² Indeed, CMS has focused resources toward social determinants of health in the effort to reduce utilization and improve outcomes. This lack of data on social determinants of health, as well as other understudied factors, represents an important opportunity for future research efforts to better define the modifiable features that could be targeted in the future to prevent readmissions. Such research is urgently needed and will likely require prospective studies to gather these important factors. Notably, most studies in this systematic review were retrospective and therefore unable to examine many of these understudied factors. Another important aspect that has received little attention is readmission preventability: only 2 studies assessed preventability, both through unstructured chart review. Preventability assessments in noncirrhotic populations have used wide-ranging methodologies, yielding inconsistent results.⁴³ This variability prompted recommendations that preventability should be assessed by multiple reviewers guided by explicit parameters.⁴³ Such detailed attention to preventability is urgently needed to better inform interventions.

In contrast to the lack of data on social factors, we found that the MELD score was examined in most studies and was frequently associated with readmission. Despite this consistent association, differences in the MELD scores between studies limit inferences into specific cutoff values that could identify the highest risk patients. Because of its existing widespread clinical use, the MELD score may prove to be important in readmission risk stratification. Efforts to develop a useful model including the MELD score are needed to target interventions to the highest risk patients.

This review has several limitations. Although we used a broad search strategy to capture studies, some may not have been included due to our selection criteria. For instance, 1 retrospective paper described factors associated with high admission density during 1 year but did not specifically report the frequency of early readmissions.⁴⁴ Similarly, a randomized trial of a disease management program did not specifically examine early readmissions.⁴⁵ Another quasi-experimental study of a quality improvement initiative was not included because a large proportion of their subjects was post liver transplant.⁴⁶ However, the inclusion of these papers is unlikely to change our conclusions; the retrospective study identified factors similar to those in the included studies, and the quasi-experimental study overlapped with the included study that assessed frailty.²⁷ Another potential limitation is the exclusion of studies published in abstract form only. Such studies may be important, as the field of cirrhosis readmissions is relatively young. However, including only full-paper publications ensures the inclusion of only higher quality studies scrutinized during the peer-review process. Similarly, newer published studies may have been missed due to the abundant interest in this topic and ongoing research. Lastly, the significant heterogeneity of the studies limits conclusions that can be made regarding the pooled readmission rates.

In summary, we found that patients with cirrhosis experience a high incidence of hospital readmissions. Several processes of care may be associated with readmissions, suggesting room for improvement in caring for this population and reducing readmissions. However, we identified several gaps in the literature, which does not adequately describe social factors and is lacking details on readmission preventability assessment. Future studies should attempt to address these issues so that interventions can be targeted to the highest risk patients and designed to best meet the needs of patients with cirrhosis.

Disclosures

Dr. Orman, Dr. Ghabril, and Dr. Emmett report no potential conflicts of interest. Dr. Chalasani reports personal fees from Lilly, personal fees from Abbvie, personal fees from Tobira/Allergan, personal fees from Ardelyx, personal fees from Amarin, personal fees from Shire, personal fees from Madrigal, personal fees from DS Biopharma (Afimmune), personal fees from Cempra, personal fees from NuSirt, grants from Galectin, grants from Gilead, grants from Intercept, grants from Cumberland, grants from Conatus, personal fees from Immuron, and personal fees from Axovant, outside the submitted work.

Funding Information

This work was supported, in part, by the National Institutes of Health, KL2 TR001106 and K23 DK109202

Cirrhosis is a morbid condition characterized by complications such as ascites, gastrointestinal bleeding, and hepatic encephalopathy. These complications frequently require hospitalization, which is a substantial burden to the healthcare system. In 2012, liver disease was responsible for nearly 250,000 admissions across the United States, costing $3 billion.¹ Despite this substantial resource utilization, outcomes remain poor, with an inpatient mortality of 6%. For those that survive, many experience hospital readmission.

More generally, early readmission reflects poor quality of care in the US. In 2004, 30-day readmissions occurred in nearly 20% of Medicare beneficiaries and costed over $17 billion.² In response to this problem, the Affordable Care Act established the Hospital Readmissions Reduction Program (HRRP), which reduces Centers for Medicare & Medicaid Services (CMS) payments to hospitals with excess 30-day readmissions for high-risk conditions, including pneumonia and heart failure.³ Heart failure, in particular, has been the subject of numerous studies detailing risk factors and interventions to predict and prevent readmission.^4-6 Based on this extensive evidence, guidelines recommend disease management programs to reduce readmissions in this population.⁷ In contrast, readmission in the cirrhosis population has received limited attention.

We therefore conducted a systematic review aiming to examine the range of readmission risk noted in the literature, with a focus on the model for end-stage liver disease (MELD) score as a risk factor for readmission.

Search Strategy

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for conducting and reporting systematic reviews.⁸ A literature search was performed by a medical librarian using the following databases: Ovid MEDLINE, PubMed, EMBASE, CINAHL, the full Cochrane Library, Scopus, Google Scholar, and ClinicalTrials.gov. All the databases were searched from 2000 to May 2017. We did not include older reports because the review focused on contemporary care; earlier studies may not reflect current cirrhosis management. To ensure literature saturation, included articles’ reference lists were reviewed.

Search strategies were developed by combining database-specific subject headings and keywords for readmissions with those for cirrhosis or its complications (Supplementary Material). Google Scholar and ClinicalTrials.gov were searched using keywords only. All results were limited to the English language and those published in 2000 or later, but no other limits were applied.

Identified records were reviewed based on strict criteria. We excluded case reports, case series, reviews, editorials, letters, and meeting abstracts without final peer-reviewed publication. We also excluded studies of pediatric populations (age < 18 years), patients without cirrhosis, and patients with liver transplants. We excluded studies in which patients were not hospitalized at study onset and those where the index admission was for an elective procedure. Because our interest was to identify factors associated with early readmission, we excluded studies that did not report readmissions within 90 days or those with a mean or median follow-up of less than 30 days. We also excluded studies that did not examine the association between readmission and at least 1 independent variable or intervention. Duplicate reports of a common sample were excluded unless the duplicate provided additional information, and such reports were examined together in our synthesis.

Two authors identified potentially eligible records by independently screening titles and abstracts. At this stage, records that did not meet the eligibility criteria were excluded, and the reasons for exclusion were not recorded. Records with disagreement were retained for full-text review. After this initial exclusion of records, the remaining full-text records were reviewed independently. For this full-text review, we recorded exclusion reasons and disagreements were resolved through discussion.

Data Collection

Data were abstracted from each study by 2 authors independently and recorded in a REDCap database.⁹ Discrepancies were resolved through discussion. We recorded study characteristics, including study design, setting, population (including the inclusion/exclusion criteria, sample size, and patient and hospitalization characteristics), interventions, and comparisons. To facilitate comparisons across studies, we employed validated methods to approximate means and standard deviations (SD).¹⁰ We recorded detailed information on outcomes including readmissions, preventability, independent variables, and mortality. Studies that focused on a single independent factor or intervention were classified as “focused,” while those that examined multiple factors were classified as “broad.” We used the Newcastle–Ottawa Scale to assess the risk of bias in each study.¹¹ This instrument uses a 9-point scale to gauge methodological quality based on selection, group comparability, and exposure/outcome assessment.

Statistical Analysis

Analyses were performed using Stata 13.1 (StataCorp LP, College Station, Texas). We determined the pooled proportion of patients with 30-day readmission using a random-effects model, with the Freeman–Tukey double-arcsine transformation for meta-analysis of proportions.¹² We investigated the heterogeneity by stratifying analyses according to prespecified study characteristics, including “broad” versus “focused.” However, the readmission risk was not different in the stratified analysis; therefore, we chose to pool the findings. For point estimates, 95% confidence intervals (CIs) were calculated, and a P-value < .05 was considered statistically significant.

Search Results

Study Characteristics

Two studies were performed in Australia, 4 in Europe, and the remainder in North America. Twenty one of the 26 studies were retrospective cohort studies (Table 1). Twenty studies were single-center studies (of which half were performed at transplant centers), and 4 of the 6 multicenter studies were based on administrative data with large samples (173,254 patients). The inclusion/exclusion criteria varied widely (Supplementary Material). Some studies only included patients admitted for specific cirrhosis complications, while others included those admitted for any reason. Two studies excluded patients admitted in the prior 30 days, and 6 excluded patients discharged to hospice. The mean risk of bias score was 7.5 (SD 1.3) out of a possible 9 points, with most lacking an adequate description of follow-up and several lacking adjustment for confounders.

Outcomes

Readmission and MELD

Hospital readmission is a costly and common problem in the US.² In addition to the negative impact that readmissions have on patients’ lives,⁴⁰ readmissions are increasingly being used to measure quality. Unplanned 30-day readmissions are posted publicly, and excess readmissions for high-risk conditions are penalized through HRRP.³ Although HRRP does not currently include cirrhosis, the program has expanded to include several conditions that were not included in the initial iteration. Whether cirrhosis will be included in future iterations remains to be seen; however, increasing scrutiny is likely to continue. Of specific populations at risk, patients with cirrhosis are particularly vulnerable due to several features. Ascites management often requires hospitalization due to diuretic titration and poor access to paracentesis, and hepatic encephalopathy treatment requires complex lactulose titration.¹⁶ Other features of cirrhosis, such as gastrointestinal bleeding, infections, and renal failure, also place patients at risk of poor outcomes. The resulting readmission burden is high, with a pooled 30-day readmission rate of 26%. Other associated outcomes are also poor, with a consistent relationship between readmission and subsequent mortality.

We found striking heterogeneity in various aspects. First, the inclusion/exclusion criteria varied widely, both cirrhosis-specific (eg, spontaneous bacterial peritonitis) and more general (patients admitted within the prior 30 days). Some of these criteria may bias readmission estimates; the risk of readmission may be reduced in those on hospice, as patients forgo curative therapy. Additionally, an established risk factor for readmission is prior hospitalization⁴¹; excluding patients with prior admissions prohibits analysis of this variable. Another aspect is the capture of readmissions: readmissions outside of the index hospital were not included in most studies. In those that did include outside readmissions, the burden was sizeable: 17% in 1 single-center study and 23% in a multistate administrative database.^16,36 These outside readmissions must be included in future studies; they are as important as same-center readmissions both to patients and CMS.³ Despite this heterogeneity, the studies scored relatively high on the Newcastle–Ottawa risk of bias scale, with the only common deficiency being an inadequate description of follow-up.

Building on the findings of this review, an important step will be the design of interventions to reduce readmissions. Such interventions require a full understanding of this population’s characteristics and needs. Critically, we found a lack of data on social determinants of health. Impairments in these factors are well-established contributors to readmission risk in other populations,^4,40 and are highly prevalent in cirrhosis.⁴² Indeed, CMS has focused resources toward social determinants of health in the effort to reduce utilization and improve outcomes. This lack of data on social determinants of health, as well as other understudied factors, represents an important opportunity for future research efforts to better define the modifiable features that could be targeted in the future to prevent readmissions. Such research is urgently needed and will likely require prospective studies to gather these important factors. Notably, most studies in this systematic review were retrospective and therefore unable to examine many of these understudied factors. Another important aspect that has received little attention is readmission preventability: only 2 studies assessed preventability, both through unstructured chart review. Preventability assessments in noncirrhotic populations have used wide-ranging methodologies, yielding inconsistent results.⁴³ This variability prompted recommendations that preventability should be assessed by multiple reviewers guided by explicit parameters.⁴³ Such detailed attention to preventability is urgently needed to better inform interventions.

In contrast to the lack of data on social factors, we found that the MELD score was examined in most studies and was frequently associated with readmission. Despite this consistent association, differences in the MELD scores between studies limit inferences into specific cutoff values that could identify the highest risk patients. Because of its existing widespread clinical use, the MELD score may prove to be important in readmission risk stratification. Efforts to develop a useful model including the MELD score are needed to target interventions to the highest risk patients.

This review has several limitations. Although we used a broad search strategy to capture studies, some may not have been included due to our selection criteria. For instance, 1 retrospective paper described factors associated with high admission density during 1 year but did not specifically report the frequency of early readmissions.⁴⁴ Similarly, a randomized trial of a disease management program did not specifically examine early readmissions.⁴⁵ Another quasi-experimental study of a quality improvement initiative was not included because a large proportion of their subjects was post liver transplant.⁴⁶ However, the inclusion of these papers is unlikely to change our conclusions; the retrospective study identified factors similar to those in the included studies, and the quasi-experimental study overlapped with the included study that assessed frailty.²⁷ Another potential limitation is the exclusion of studies published in abstract form only. Such studies may be important, as the field of cirrhosis readmissions is relatively young. However, including only full-paper publications ensures the inclusion of only higher quality studies scrutinized during the peer-review process. Similarly, newer published studies may have been missed due to the abundant interest in this topic and ongoing research. Lastly, the significant heterogeneity of the studies limits conclusions that can be made regarding the pooled readmission rates.

In summary, we found that patients with cirrhosis experience a high incidence of hospital readmissions. Several processes of care may be associated with readmissions, suggesting room for improvement in caring for this population and reducing readmissions. However, we identified several gaps in the literature, which does not adequately describe social factors and is lacking details on readmission preventability assessment. Future studies should attempt to address these issues so that interventions can be targeted to the highest risk patients and designed to best meet the needs of patients with cirrhosis.

Disclosures

Dr. Orman, Dr. Ghabril, and Dr. Emmett report no potential conflicts of interest. Dr. Chalasani reports personal fees from Lilly, personal fees from Abbvie, personal fees from Tobira/Allergan, personal fees from Ardelyx, personal fees from Amarin, personal fees from Shire, personal fees from Madrigal, personal fees from DS Biopharma (Afimmune), personal fees from Cempra, personal fees from NuSirt, grants from Galectin, grants from Gilead, grants from Intercept, grants from Cumberland, grants from Conatus, personal fees from Immuron, and personal fees from Axovant, outside the submitted work.

Funding Information

This work was supported, in part, by the National Institutes of Health, KL2 TR001106 and K23 DK109202

Cirrhosis is a morbid condition characterized by complications such as ascites, gastrointestinal bleeding, and hepatic encephalopathy. These complications frequently require hospitalization, which is a substantial burden to the healthcare system. In 2012, liver disease was responsible for nearly 250,000 admissions across the United States, costing $3 billion.¹ Despite this substantial resource utilization, outcomes remain poor, with an inpatient mortality of 6%. For those that survive, many experience hospital readmission.

More generally, early readmission reflects poor quality of care in the US. In 2004, 30-day readmissions occurred in nearly 20% of Medicare beneficiaries and costed over $17 billion.² In response to this problem, the Affordable Care Act established the Hospital Readmissions Reduction Program (HRRP), which reduces Centers for Medicare & Medicaid Services (CMS) payments to hospitals with excess 30-day readmissions for high-risk conditions, including pneumonia and heart failure.³ Heart failure, in particular, has been the subject of numerous studies detailing risk factors and interventions to predict and prevent readmission.^4-6 Based on this extensive evidence, guidelines recommend disease management programs to reduce readmissions in this population.⁷ In contrast, readmission in the cirrhosis population has received limited attention.

We therefore conducted a systematic review aiming to examine the range of readmission risk noted in the literature, with a focus on the model for end-stage liver disease (MELD) score as a risk factor for readmission.

Search Strategy

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for conducting and reporting systematic reviews.⁸ A literature search was performed by a medical librarian using the following databases: Ovid MEDLINE, PubMed, EMBASE, CINAHL, the full Cochrane Library, Scopus, Google Scholar, and ClinicalTrials.gov. All the databases were searched from 2000 to May 2017. We did not include older reports because the review focused on contemporary care; earlier studies may not reflect current cirrhosis management. To ensure literature saturation, included articles’ reference lists were reviewed.

Search strategies were developed by combining database-specific subject headings and keywords for readmissions with those for cirrhosis or its complications (Supplementary Material). Google Scholar and ClinicalTrials.gov were searched using keywords only. All results were limited to the English language and those published in 2000 or later, but no other limits were applied.

Identified records were reviewed based on strict criteria. We excluded case reports, case series, reviews, editorials, letters, and meeting abstracts without final peer-reviewed publication. We also excluded studies of pediatric populations (age < 18 years), patients without cirrhosis, and patients with liver transplants. We excluded studies in which patients were not hospitalized at study onset and those where the index admission was for an elective procedure. Because our interest was to identify factors associated with early readmission, we excluded studies that did not report readmissions within 90 days or those with a mean or median follow-up of less than 30 days. We also excluded studies that did not examine the association between readmission and at least 1 independent variable or intervention. Duplicate reports of a common sample were excluded unless the duplicate provided additional information, and such reports were examined together in our synthesis.

Two authors identified potentially eligible records by independently screening titles and abstracts. At this stage, records that did not meet the eligibility criteria were excluded, and the reasons for exclusion were not recorded. Records with disagreement were retained for full-text review. After this initial exclusion of records, the remaining full-text records were reviewed independently. For this full-text review, we recorded exclusion reasons and disagreements were resolved through discussion.

Data Collection

Data were abstracted from each study by 2 authors independently and recorded in a REDCap database.⁹ Discrepancies were resolved through discussion. We recorded study characteristics, including study design, setting, population (including the inclusion/exclusion criteria, sample size, and patient and hospitalization characteristics), interventions, and comparisons. To facilitate comparisons across studies, we employed validated methods to approximate means and standard deviations (SD).¹⁰ We recorded detailed information on outcomes including readmissions, preventability, independent variables, and mortality. Studies that focused on a single independent factor or intervention were classified as “focused,” while those that examined multiple factors were classified as “broad.” We used the Newcastle–Ottawa Scale to assess the risk of bias in each study.¹¹ This instrument uses a 9-point scale to gauge methodological quality based on selection, group comparability, and exposure/outcome assessment.

Statistical Analysis

Analyses were performed using Stata 13.1 (StataCorp LP, College Station, Texas). We determined the pooled proportion of patients with 30-day readmission using a random-effects model, with the Freeman–Tukey double-arcsine transformation for meta-analysis of proportions.¹² We investigated the heterogeneity by stratifying analyses according to prespecified study characteristics, including “broad” versus “focused.” However, the readmission risk was not different in the stratified analysis; therefore, we chose to pool the findings. For point estimates, 95% confidence intervals (CIs) were calculated, and a P-value < .05 was considered statistically significant.

Search Results

Study Characteristics

Two studies were performed in Australia, 4 in Europe, and the remainder in North America. Twenty one of the 26 studies were retrospective cohort studies (Table 1). Twenty studies were single-center studies (of which half were performed at transplant centers), and 4 of the 6 multicenter studies were based on administrative data with large samples (173,254 patients). The inclusion/exclusion criteria varied widely (Supplementary Material). Some studies only included patients admitted for specific cirrhosis complications, while others included those admitted for any reason. Two studies excluded patients admitted in the prior 30 days, and 6 excluded patients discharged to hospice. The mean risk of bias score was 7.5 (SD 1.3) out of a possible 9 points, with most lacking an adequate description of follow-up and several lacking adjustment for confounders.

Outcomes

Readmission and MELD

Hospital readmission is a costly and common problem in the US.² In addition to the negative impact that readmissions have on patients’ lives,⁴⁰ readmissions are increasingly being used to measure quality. Unplanned 30-day readmissions are posted publicly, and excess readmissions for high-risk conditions are penalized through HRRP.³ Although HRRP does not currently include cirrhosis, the program has expanded to include several conditions that were not included in the initial iteration. Whether cirrhosis will be included in future iterations remains to be seen; however, increasing scrutiny is likely to continue. Of specific populations at risk, patients with cirrhosis are particularly vulnerable due to several features. Ascites management often requires hospitalization due to diuretic titration and poor access to paracentesis, and hepatic encephalopathy treatment requires complex lactulose titration.¹⁶ Other features of cirrhosis, such as gastrointestinal bleeding, infections, and renal failure, also place patients at risk of poor outcomes. The resulting readmission burden is high, with a pooled 30-day readmission rate of 26%. Other associated outcomes are also poor, with a consistent relationship between readmission and subsequent mortality.

We found striking heterogeneity in various aspects. First, the inclusion/exclusion criteria varied widely, both cirrhosis-specific (eg, spontaneous bacterial peritonitis) and more general (patients admitted within the prior 30 days). Some of these criteria may bias readmission estimates; the risk of readmission may be reduced in those on hospice, as patients forgo curative therapy. Additionally, an established risk factor for readmission is prior hospitalization⁴¹; excluding patients with prior admissions prohibits analysis of this variable. Another aspect is the capture of readmissions: readmissions outside of the index hospital were not included in most studies. In those that did include outside readmissions, the burden was sizeable: 17% in 1 single-center study and 23% in a multistate administrative database.^16,36 These outside readmissions must be included in future studies; they are as important as same-center readmissions both to patients and CMS.³ Despite this heterogeneity, the studies scored relatively high on the Newcastle–Ottawa risk of bias scale, with the only common deficiency being an inadequate description of follow-up.

Building on the findings of this review, an important step will be the design of interventions to reduce readmissions. Such interventions require a full understanding of this population’s characteristics and needs. Critically, we found a lack of data on social determinants of health. Impairments in these factors are well-established contributors to readmission risk in other populations,^4,40 and are highly prevalent in cirrhosis.⁴² Indeed, CMS has focused resources toward social determinants of health in the effort to reduce utilization and improve outcomes. This lack of data on social determinants of health, as well as other understudied factors, represents an important opportunity for future research efforts to better define the modifiable features that could be targeted in the future to prevent readmissions. Such research is urgently needed and will likely require prospective studies to gather these important factors. Notably, most studies in this systematic review were retrospective and therefore unable to examine many of these understudied factors. Another important aspect that has received little attention is readmission preventability: only 2 studies assessed preventability, both through unstructured chart review. Preventability assessments in noncirrhotic populations have used wide-ranging methodologies, yielding inconsistent results.⁴³ This variability prompted recommendations that preventability should be assessed by multiple reviewers guided by explicit parameters.⁴³ Such detailed attention to preventability is urgently needed to better inform interventions.

In contrast to the lack of data on social factors, we found that the MELD score was examined in most studies and was frequently associated with readmission. Despite this consistent association, differences in the MELD scores between studies limit inferences into specific cutoff values that could identify the highest risk patients. Because of its existing widespread clinical use, the MELD score may prove to be important in readmission risk stratification. Efforts to develop a useful model including the MELD score are needed to target interventions to the highest risk patients.

This review has several limitations. Although we used a broad search strategy to capture studies, some may not have been included due to our selection criteria. For instance, 1 retrospective paper described factors associated with high admission density during 1 year but did not specifically report the frequency of early readmissions.⁴⁴ Similarly, a randomized trial of a disease management program did not specifically examine early readmissions.⁴⁵ Another quasi-experimental study of a quality improvement initiative was not included because a large proportion of their subjects was post liver transplant.⁴⁶ However, the inclusion of these papers is unlikely to change our conclusions; the retrospective study identified factors similar to those in the included studies, and the quasi-experimental study overlapped with the included study that assessed frailty.²⁷ Another potential limitation is the exclusion of studies published in abstract form only. Such studies may be important, as the field of cirrhosis readmissions is relatively young. However, including only full-paper publications ensures the inclusion of only higher quality studies scrutinized during the peer-review process. Similarly, newer published studies may have been missed due to the abundant interest in this topic and ongoing research. Lastly, the significant heterogeneity of the studies limits conclusions that can be made regarding the pooled readmission rates.

In summary, we found that patients with cirrhosis experience a high incidence of hospital readmissions. Several processes of care may be associated with readmissions, suggesting room for improvement in caring for this population and reducing readmissions. However, we identified several gaps in the literature, which does not adequately describe social factors and is lacking details on readmission preventability assessment. Future studies should attempt to address these issues so that interventions can be targeted to the highest risk patients and designed to best meet the needs of patients with cirrhosis.

Disclosures

Dr. Orman, Dr. Ghabril, and Dr. Emmett report no potential conflicts of interest. Dr. Chalasani reports personal fees from Lilly, personal fees from Abbvie, personal fees from Tobira/Allergan, personal fees from Ardelyx, personal fees from Amarin, personal fees from Shire, personal fees from Madrigal, personal fees from DS Biopharma (Afimmune), personal fees from Cempra, personal fees from NuSirt, grants from Galectin, grants from Gilead, grants from Intercept, grants from Cumberland, grants from Conatus, personal fees from Immuron, and personal fees from Axovant, outside the submitted work.

Funding Information

This work was supported, in part, by the National Institutes of Health, KL2 TR001106 and K23 DK109202

A 43-year-old man with a history of asplenia, hepatitis C, and nephrolithiasis reported right-flank pain. He described severe, sharp pain that came in waves and radiated to the right groin, associated with nausea and nonbloody emesis. He noted “pink urine” but no dysuria. He had 4prior similar episodes during which he had passed kidney stones, although stone analysis had never been performed. He denied having fevers or chills.

The patient had been involved in a remote motor vehicle accident complicated by splenic laceration, for which he underwent splenectomy. He was appropriately immunized. The patient also suffered from bipolar affective disorder and untreated chronic hepatitis C infection with no evidence of cirrhosis. He smoked one pack of tobacco per day for the last 10 years and reported distant alcohol and methamphetamine use.

Right-flank pain can arise from conditions affecting the lower thorax (effusion, pneumonia, pulmonary embolism), abdomen (hepatobiliary or intestinal disease), retroperitoneum (hemorrhage or infection), musculoskeletal system, peripheral nerves (herpes zoster), or the genitourinary system (pyelonephritis). Pain radiating to the groin, discolored urine (suggesting hematuria), and history of kidney stones increase the likelihood of renal colic from nephrolithiasis.

Less commonly, flank pain and hematuria may present as initial symptoms of renal cell carcinoma, renal infarction, or aortic dissection. The patient’s immunosuppression from asplenia and active injection drug use could predispose him to septic emboli to his kidneys. Prior trauma causing aortic injury could predispose himto subsequent dissection.

The patient appeared well with a heart rate of 100 beats per minute, blood pressure 122/76 mmHg, temperature 36.8°C, respiratory rate 16 breaths per minute, and oxygen saturation 96% on room air. His cardiopulmonary and abdominal examinations were normal, and he had no costovertebral angle tenderness. His skin was warm and dry without rashes. His white blood cell (WBC) count was 26,000/μL; absolute neutrophil count was 22,000/μL. Serum chemistries were normal, including creatinine 0.63 mg/dL, calcium 8.8 mg/dL, and phosphorus 3.1 mg/dL. Lactate was 0.8 mmol/L (reference range: 0-2.0 mmol/L). Urinalysis revealed large ketones, >50 red blood cells (RBC) per high power field (HPF), <5 WBC per HPF, 1+ calcium oxalate crystals and pH 6.0. A bedside ultrasound showed mild right hydronephrosis. Computed tomography (CT) with intravenous contrast of his abdomen and pelvis demonstrated diffuse, mildly prominent subcentimeter mesenteric lymphadenopathy and no kidney stones. He was treated with intravenous fluids and pain control, and was discharged with a presumptive diagnosis of a passed kidney stone.

A passed stone would not explain this degree of leukocytosis. The CT results reduce the likelihood of a renal neoplasm, renal infarction, or pyelonephritis. Mesenteric lymphadenopathy is nonspecific, but it may signal underlying infection or malignancy with spread to lymph nodes, or it may be part of a systemic disorder causing generalized lymphadenopathy. Malignant causes of mesenteric lymphadenopathy (with no apparent primary tumor) include testicular cancer, lymphoma, and primary urogenital neoplasms.

Noninfectious illnesses to consider include neoplasms and connective tissue diseases. Malignancy would be unlikely to manifest this abruptly or produce a paraneoplastic disorder with these features. Among the connective tissue diseases, sarcoidosis warrants consideration in this patient with adenopathy, erythema nodosum, arthritis, and a predilection for skin changes in prior scars. However, it is uncommon for sarcoidosis to present so explosively. Painful oral and genital ulcers, pustular rash, polyarthritis, and erythema nodosum occur in Behçet’s disease, which is associated with pathergy (an exaggerated cutaneous response to minor trauma). Patients with Behçet’s may have eye involvement, including uveitis and a hypopion, and may develop vascular aneurysms in the pulmonary, intracranial, or visceral arteries. Renal artery involvement could cause hematuria and flank pain.

The patient described severe fatigue and drenching night sweats for two months prior to admission. He denied dyspnea or cough. He was born in the southwestern United States and had lived in California for almost a decade. He had been incarcerated for a few years and released three years prior. He had intermittently lived in homeless shelters, but currently lived alone in downtown San Francisco. He had traveled remotely to the Caribbean, and more recently traveled frequently to the Central Valley in California. The patient formerly worked as a pipe-fitter and welder. He denied animal exposure or recent sick contacts. He was sexually active with women, and intermittently used barrier protection.

His years in the southwestern United States may have exposed the patient to blastomycosis or histoplasmosis; both can mimic mycobacterial disease. Blastomycosis demonstrates a slightly stronger predilection for spreading to the bones, genitourinary tract, and central nervous system, whereas histoplasmosis is a more frequent cause of polyarthrtitis and mesenteric adenopathy. The patient’s travel to the Central Valley, California raises the possibility of coccidioidomycosis, which typically starts with pulmonary disease prior to dissemination to bones, skin, and other less common sites. Pipe-fitters are predisposed to asbestos-related illnesses, including lung cancer and mesothelioma, which would not explain this patient’s presentation. Incarceration and high-risk sexual practices increase his risk for tuberculosis, HIV, and syphilis. Widespread skin involvement is more characteristic of syphilis or primary HIV infection than of disseminated fungal or mycobacterial infection.

WBC measured 29,000/uL with a neutrophilic predominance. His peripheral blood smear was unremarkable. A comprehensive metabolic panel was normal. Lactate dehydrogenase (LDH) was 317 U/L (reference range 140-280 U/L). Erythrocyte sedimentation rate (ESR) was 39 mm/hr (reference range < 20 mm/hr) and C-reactive protein (CRP) was 66 mg/L (reference range <6.3 mg/L). Blood, urine, and throat cultures were sent. Chest radiograph showed clear lungs without adenopathy. Ankle and knee radiographs identified small effusions bilaterally without bony abnormalities. CT of his brain showed a small, hypodense lesion in the right lacrimal gland. A lumbar puncture with cerebrospinal fluid (CSF) analysis showed absence of RBCs; WBC, 2/µL; protein, 35 mg/dL; glucose, 62 mg/dL; negative gram stain. CSF bacterial and fungal cultures, venereal disease research laboratory (VDRL), herpes simplex virus polymerase chain reaction (HSV PCR), and cryptococcal antigen were sent for laboratory analysis. The patient was started on vancomycin and aztreonam.

Lesions of the lacrimal gland feature multiple causes, including autoimmune diseases (Sjögren’s, Behçet’s disease), granulomatous diseases (sarcoidosis, granulomatosis with polyangiitis), neoplasms (salivary gland tumors, lymphoma), and infections. Initiating broad-spectrum antibiotics is reasonable while awaiting additional information from blood and urine cultures, serologies for HIV and syphilis, and purified protein derivative or interferon-gamma release assay (IGRA).

If these tests fail to reveal a diagnosis, the search for atypical infections and noninfectious possibilities should expand. Histoplasmosis and blastomycosis would be the most likely fungal diseases to account for his arthritis and adenopathy. Coccidioidomycosis is less likely in light of the normal chest radiograph. Computed tomography of the chest would be reasonable to look for adenopathy, which would strengthen the case for lymphoma or sarcoidosis, and may also identify a potential site to biopsy to establish these diagnoses.

The patient continued to have intermittent fevers, sweats, and malaise over the next 3 days. All bacterial and fungal cultures remained negative, and antibiotics were discontinued. Rheumatoid factor, anticyclic citrullinated peptide, antinuclear antibody, and cryoglobulins were negative. Serum C3, C4, and angiotensin-converting enzyme (ACE) levels were normal. A rapid plasma reagin (RPR), HIV antibody, IGRA, and serum antibodies for Coccidioides, histoplasmosis, and West Nile virus were negative. Urine nucleic acid amplification testing for gonorrhea and chlamydia was negative. CSF VDRL, HSV PCR and cryptococcal antigen were negative. HSV culture from an oral ulcer showed no growth. The patient had a reactive hepatitis C antibody with a viral load of 3 million virus equivalents/mL.

The additional test results lower the likelihood of an acute infection. Uncontrolled hepatitis C increases the risk of several noninfectious manifestations. The normal results for serum complements and cryoglobulins effectively rule out cryoglobulinemia. Patients with hepatitis C have an increased risk of lymphoma, which could account for the subacute fevers, night sweats, adenopathy, elevated LDH, and the right orbital mass, but less likely for the oral ulcers, arthritis, and skin manifestations. Sarcoidosis is less likely given the lack of hilar adenopathy, relatively abrupt onset of multisystem disease, and the presence of oral ulcers. Behçet’s disease could account for his oral ulcers, erythema nodosum, and distribution of papules, pustules, and plaques with the predilection for scars and tattoos. Behçet’s could also explain the arthritis, the hematuria if the patient had renal artery involvement, and the orbital lesion. However, lymphadenopathy is not a prominent feature. At this point, tissue should be obtained for histopathology (to assess for vasculitis or granulomatous infiltration) and flow cytometry.

The most likely diagnosis is Löfgren’s syndrome, a variant of sarcoidosis characterized by erythema nodosum, bilateral hilar lymphadenopathy, and polyarthralgias or polyarthritis. Löfgren’s syndrome may include fevers, uveitis, widespread skin lesions and other systemic manifestations. Sarcoidosis could explain the lacrimal gland lesion, and could manifest with recurrent kidney stones. Oral lesions may occur in sarcoidosis. A normal serum ACE level may be observed in up to half of patients. The lack of visualized granulomas on the submental node FNA may reflect sampling error, lower likelihood of visualizing granulomas on FNA (compared with excisional biopsy), or biopsy location (hilar nodes are more likely to demonstrate sarcoid granulomas).

Although Löfgren’s syndrome is often self-limited, treatment can ameliorate symptoms. Nonsteroidal anti-inflammatory medication can be tried first, with prednisone reserved for refractory cases.

The constellation of bilateral hilar adenopathy, arthritis, and erythema nodosum was consistent with Löfgren’s syndrome, further supported by granulomatous infiltrates on biopsy. The patient’s symptoms resolved with naproxen. He was scheduled for follow-up in dermatology and rheumatology clinics and was referred to hepatology for management of hepatitis C.

Sarcoidosis is a multisystem granulomatous disease of unclear etiology. The disease derives its name from Boeck’s 1899 report describing benign cutaneous lesions that resembled sarcomas.¹ Sarcoidosis most commonly manifests as bilateral hilar adenopathy and pulmonary infiltrates, but may impact any tissue or organ, including the eyes, nonhilar lymph nodes, liver, spleen, joints, mucous membranes, and skin. Nephrolithiasis may result from hypercalcemia and/or hypercalciuria (related to granulomatous production of 1,25 vitamin D) and can be the presenting feature of sarcoidosis.² Less common presentations include neurologic sarcoidosis (which can present with seizures, aseptic meningitis, encephalopathy, neuroendocrine dysfunction, myelopathy and peripheral neuropathies), cardiac sarcoidosis (which may present with arrhythmias, valvular dysfunction, heart failure, ischemia, or pericardial disease), and Heerfordt syndrome (the constellation of parotid gland enlargement, facial palsy, anterior uveitis, and fever). Sarcoidosis may mimic other diseases, including malignancy, idiopathic pulmonary fibrosis, and infiltrative tuberculosis.³ Sarcoidosis-like reactions have occurred in response to malignancy and medications.⁴

The patient’s rash demonstrated a predilection for areas of prior scarring, which has a limited differential diagnosis. Keloids and hypertrophic scars occur at sites of former surgical wounds, lacerations, or areas of inflammation. Pruritic urticarial papules and plaques of pregnancy (PUPPP) is a benign inflammatory condition where papules cluster in areas of prior striae. Cutaneous lesions of Behçet’s syndrome display pathergy, where pustular response is observed at sites of injury. Granulomatous infiltration in sarcoidosis may demonstrate a predilection for scars and tattoos (ie, scar or tattoo sarcoidosis).⁵ Sarcoidosis can have other cutaneous manifestations, including psoriaform, ulcerative, or erythrodermic lesions; subcutaneous nodules; scarring or nonscarring alopecia; and lupus pernio – violaceous, nodular and plaque-like lesions on the nose, earlobes, cheeks, and digits.⁵

Löfgren’s syndrome is a distinct variant of sarcoidosis.In 1952, Dr. Löfgren described a case series of patients with bilateral hilar lymphadenopathy and coexisting erythema nodosum and polyarthralgia.⁶ The epidemiology favors young women.⁷ Patients with Löfgren’s syndrome present acutely (as in this case), which differs from the typical subacute course observed with sarcoidosis. In addition to the classic presentation described above, patients with Löfgren’s syndrome may demonstrate additional manifestations of sarcoidosis, including fevers, peripheral adenopathy, arthritis, and granulomatous skin lesions. Painful symptoms may require short-term anti-inflammatory treatments. Most patients do not require systemic immunosuppression. Symptoms usually decrease over several months, and the majority of patients experience complete remission within years. Rare recurrences have been described up to several years.⁸

In confirming the diagnosis of sarcoidosis, current guidelines recommend exclusion of other diseases that present similarly, a work-up that generally includes compatible laboratory tests and imaging, and histologic demonstration of noncaseating granulomas.⁹ However, Löfgren’s syndrome is a notable exception. The constellation of fever, bilateral hilar adenopathy, polyarthralgia, and erythema nodosum suffices to diagnose Löfgren’s syndrome as long as the disease remits rapidly and spontaneously.⁹ Thus, in this case, although granulomatous infiltrates were confirmed on biopsy, the diagnosis of Löfgren’s syndrome could have been based on clinical and radiologic features alone.

• Sarcoidosis is a multisystem granulomatous disease that most commonly presents with bilateral hilar adenopathy and pulmonary infiltrates but can also present atypically, including with nephrolithiasis from hypercalcemia, neurologic syndromes, and cardiac involvement.
• Löfgren’s syndrome, a variant of sarcoidosis, is characterized by relatively acute onset of fevers, erythema nodosum, bilateral hilar adenopathy, and polyarthralgia or polyarthritis. Most patients recover and manifest complete remission.
• A limited differential exists for rashes with a predilection for areas of tattoos and prior scarring, including keloids, PUPPP, Behçet’s disease, and granulomatous infiltration.

Disclosure

There are no conflicts of interest or financial disclosures to report.

A 43-year-old man with a history of asplenia, hepatitis C, and nephrolithiasis reported right-flank pain. He described severe, sharp pain that came in waves and radiated to the right groin, associated with nausea and nonbloody emesis. He noted “pink urine” but no dysuria. He had 4prior similar episodes during which he had passed kidney stones, although stone analysis had never been performed. He denied having fevers or chills.

The patient had been involved in a remote motor vehicle accident complicated by splenic laceration, for which he underwent splenectomy. He was appropriately immunized. The patient also suffered from bipolar affective disorder and untreated chronic hepatitis C infection with no evidence of cirrhosis. He smoked one pack of tobacco per day for the last 10 years and reported distant alcohol and methamphetamine use.

Right-flank pain can arise from conditions affecting the lower thorax (effusion, pneumonia, pulmonary embolism), abdomen (hepatobiliary or intestinal disease), retroperitoneum (hemorrhage or infection), musculoskeletal system, peripheral nerves (herpes zoster), or the genitourinary system (pyelonephritis). Pain radiating to the groin, discolored urine (suggesting hematuria), and history of kidney stones increase the likelihood of renal colic from nephrolithiasis.

Less commonly, flank pain and hematuria may present as initial symptoms of renal cell carcinoma, renal infarction, or aortic dissection. The patient’s immunosuppression from asplenia and active injection drug use could predispose him to septic emboli to his kidneys. Prior trauma causing aortic injury could predispose himto subsequent dissection.

The patient appeared well with a heart rate of 100 beats per minute, blood pressure 122/76 mmHg, temperature 36.8°C, respiratory rate 16 breaths per minute, and oxygen saturation 96% on room air. His cardiopulmonary and abdominal examinations were normal, and he had no costovertebral angle tenderness. His skin was warm and dry without rashes. His white blood cell (WBC) count was 26,000/μL; absolute neutrophil count was 22,000/μL. Serum chemistries were normal, including creatinine 0.63 mg/dL, calcium 8.8 mg/dL, and phosphorus 3.1 mg/dL. Lactate was 0.8 mmol/L (reference range: 0-2.0 mmol/L). Urinalysis revealed large ketones, >50 red blood cells (RBC) per high power field (HPF), <5 WBC per HPF, 1+ calcium oxalate crystals and pH 6.0. A bedside ultrasound showed mild right hydronephrosis. Computed tomography (CT) with intravenous contrast of his abdomen and pelvis demonstrated diffuse, mildly prominent subcentimeter mesenteric lymphadenopathy and no kidney stones. He was treated with intravenous fluids and pain control, and was discharged with a presumptive diagnosis of a passed kidney stone.

A passed stone would not explain this degree of leukocytosis. The CT results reduce the likelihood of a renal neoplasm, renal infarction, or pyelonephritis. Mesenteric lymphadenopathy is nonspecific, but it may signal underlying infection or malignancy with spread to lymph nodes, or it may be part of a systemic disorder causing generalized lymphadenopathy. Malignant causes of mesenteric lymphadenopathy (with no apparent primary tumor) include testicular cancer, lymphoma, and primary urogenital neoplasms.

Noninfectious illnesses to consider include neoplasms and connective tissue diseases. Malignancy would be unlikely to manifest this abruptly or produce a paraneoplastic disorder with these features. Among the connective tissue diseases, sarcoidosis warrants consideration in this patient with adenopathy, erythema nodosum, arthritis, and a predilection for skin changes in prior scars. However, it is uncommon for sarcoidosis to present so explosively. Painful oral and genital ulcers, pustular rash, polyarthritis, and erythema nodosum occur in Behçet’s disease, which is associated with pathergy (an exaggerated cutaneous response to minor trauma). Patients with Behçet’s may have eye involvement, including uveitis and a hypopion, and may develop vascular aneurysms in the pulmonary, intracranial, or visceral arteries. Renal artery involvement could cause hematuria and flank pain.

The patient described severe fatigue and drenching night sweats for two months prior to admission. He denied dyspnea or cough. He was born in the southwestern United States and had lived in California for almost a decade. He had been incarcerated for a few years and released three years prior. He had intermittently lived in homeless shelters, but currently lived alone in downtown San Francisco. He had traveled remotely to the Caribbean, and more recently traveled frequently to the Central Valley in California. The patient formerly worked as a pipe-fitter and welder. He denied animal exposure or recent sick contacts. He was sexually active with women, and intermittently used barrier protection.

His years in the southwestern United States may have exposed the patient to blastomycosis or histoplasmosis; both can mimic mycobacterial disease. Blastomycosis demonstrates a slightly stronger predilection for spreading to the bones, genitourinary tract, and central nervous system, whereas histoplasmosis is a more frequent cause of polyarthrtitis and mesenteric adenopathy. The patient’s travel to the Central Valley, California raises the possibility of coccidioidomycosis, which typically starts with pulmonary disease prior to dissemination to bones, skin, and other less common sites. Pipe-fitters are predisposed to asbestos-related illnesses, including lung cancer and mesothelioma, which would not explain this patient’s presentation. Incarceration and high-risk sexual practices increase his risk for tuberculosis, HIV, and syphilis. Widespread skin involvement is more characteristic of syphilis or primary HIV infection than of disseminated fungal or mycobacterial infection.

WBC measured 29,000/uL with a neutrophilic predominance. His peripheral blood smear was unremarkable. A comprehensive metabolic panel was normal. Lactate dehydrogenase (LDH) was 317 U/L (reference range 140-280 U/L). Erythrocyte sedimentation rate (ESR) was 39 mm/hr (reference range < 20 mm/hr) and C-reactive protein (CRP) was 66 mg/L (reference range <6.3 mg/L). Blood, urine, and throat cultures were sent. Chest radiograph showed clear lungs without adenopathy. Ankle and knee radiographs identified small effusions bilaterally without bony abnormalities. CT of his brain showed a small, hypodense lesion in the right lacrimal gland. A lumbar puncture with cerebrospinal fluid (CSF) analysis showed absence of RBCs; WBC, 2/µL; protein, 35 mg/dL; glucose, 62 mg/dL; negative gram stain. CSF bacterial and fungal cultures, venereal disease research laboratory (VDRL), herpes simplex virus polymerase chain reaction (HSV PCR), and cryptococcal antigen were sent for laboratory analysis. The patient was started on vancomycin and aztreonam.

Lesions of the lacrimal gland feature multiple causes, including autoimmune diseases (Sjögren’s, Behçet’s disease), granulomatous diseases (sarcoidosis, granulomatosis with polyangiitis), neoplasms (salivary gland tumors, lymphoma), and infections. Initiating broad-spectrum antibiotics is reasonable while awaiting additional information from blood and urine cultures, serologies for HIV and syphilis, and purified protein derivative or interferon-gamma release assay (IGRA).

If these tests fail to reveal a diagnosis, the search for atypical infections and noninfectious possibilities should expand. Histoplasmosis and blastomycosis would be the most likely fungal diseases to account for his arthritis and adenopathy. Coccidioidomycosis is less likely in light of the normal chest radiograph. Computed tomography of the chest would be reasonable to look for adenopathy, which would strengthen the case for lymphoma or sarcoidosis, and may also identify a potential site to biopsy to establish these diagnoses.

The patient continued to have intermittent fevers, sweats, and malaise over the next 3 days. All bacterial and fungal cultures remained negative, and antibiotics were discontinued. Rheumatoid factor, anticyclic citrullinated peptide, antinuclear antibody, and cryoglobulins were negative. Serum C3, C4, and angiotensin-converting enzyme (ACE) levels were normal. A rapid plasma reagin (RPR), HIV antibody, IGRA, and serum antibodies for Coccidioides, histoplasmosis, and West Nile virus were negative. Urine nucleic acid amplification testing for gonorrhea and chlamydia was negative. CSF VDRL, HSV PCR and cryptococcal antigen were negative. HSV culture from an oral ulcer showed no growth. The patient had a reactive hepatitis C antibody with a viral load of 3 million virus equivalents/mL.

The additional test results lower the likelihood of an acute infection. Uncontrolled hepatitis C increases the risk of several noninfectious manifestations. The normal results for serum complements and cryoglobulins effectively rule out cryoglobulinemia. Patients with hepatitis C have an increased risk of lymphoma, which could account for the subacute fevers, night sweats, adenopathy, elevated LDH, and the right orbital mass, but less likely for the oral ulcers, arthritis, and skin manifestations. Sarcoidosis is less likely given the lack of hilar adenopathy, relatively abrupt onset of multisystem disease, and the presence of oral ulcers. Behçet’s disease could account for his oral ulcers, erythema nodosum, and distribution of papules, pustules, and plaques with the predilection for scars and tattoos. Behçet’s could also explain the arthritis, the hematuria if the patient had renal artery involvement, and the orbital lesion. However, lymphadenopathy is not a prominent feature. At this point, tissue should be obtained for histopathology (to assess for vasculitis or granulomatous infiltration) and flow cytometry.

The most likely diagnosis is Löfgren’s syndrome, a variant of sarcoidosis characterized by erythema nodosum, bilateral hilar lymphadenopathy, and polyarthralgias or polyarthritis. Löfgren’s syndrome may include fevers, uveitis, widespread skin lesions and other systemic manifestations. Sarcoidosis could explain the lacrimal gland lesion, and could manifest with recurrent kidney stones. Oral lesions may occur in sarcoidosis. A normal serum ACE level may be observed in up to half of patients. The lack of visualized granulomas on the submental node FNA may reflect sampling error, lower likelihood of visualizing granulomas on FNA (compared with excisional biopsy), or biopsy location (hilar nodes are more likely to demonstrate sarcoid granulomas).

Although Löfgren’s syndrome is often self-limited, treatment can ameliorate symptoms. Nonsteroidal anti-inflammatory medication can be tried first, with prednisone reserved for refractory cases.

The constellation of bilateral hilar adenopathy, arthritis, and erythema nodosum was consistent with Löfgren’s syndrome, further supported by granulomatous infiltrates on biopsy. The patient’s symptoms resolved with naproxen. He was scheduled for follow-up in dermatology and rheumatology clinics and was referred to hepatology for management of hepatitis C.

Sarcoidosis is a multisystem granulomatous disease of unclear etiology. The disease derives its name from Boeck’s 1899 report describing benign cutaneous lesions that resembled sarcomas.¹ Sarcoidosis most commonly manifests as bilateral hilar adenopathy and pulmonary infiltrates, but may impact any tissue or organ, including the eyes, nonhilar lymph nodes, liver, spleen, joints, mucous membranes, and skin. Nephrolithiasis may result from hypercalcemia and/or hypercalciuria (related to granulomatous production of 1,25 vitamin D) and can be the presenting feature of sarcoidosis.² Less common presentations include neurologic sarcoidosis (which can present with seizures, aseptic meningitis, encephalopathy, neuroendocrine dysfunction, myelopathy and peripheral neuropathies), cardiac sarcoidosis (which may present with arrhythmias, valvular dysfunction, heart failure, ischemia, or pericardial disease), and Heerfordt syndrome (the constellation of parotid gland enlargement, facial palsy, anterior uveitis, and fever). Sarcoidosis may mimic other diseases, including malignancy, idiopathic pulmonary fibrosis, and infiltrative tuberculosis.³ Sarcoidosis-like reactions have occurred in response to malignancy and medications.⁴

The patient’s rash demonstrated a predilection for areas of prior scarring, which has a limited differential diagnosis. Keloids and hypertrophic scars occur at sites of former surgical wounds, lacerations, or areas of inflammation. Pruritic urticarial papules and plaques of pregnancy (PUPPP) is a benign inflammatory condition where papules cluster in areas of prior striae. Cutaneous lesions of Behçet’s syndrome display pathergy, where pustular response is observed at sites of injury. Granulomatous infiltration in sarcoidosis may demonstrate a predilection for scars and tattoos (ie, scar or tattoo sarcoidosis).⁵ Sarcoidosis can have other cutaneous manifestations, including psoriaform, ulcerative, or erythrodermic lesions; subcutaneous nodules; scarring or nonscarring alopecia; and lupus pernio – violaceous, nodular and plaque-like lesions on the nose, earlobes, cheeks, and digits.⁵

Löfgren’s syndrome is a distinct variant of sarcoidosis.In 1952, Dr. Löfgren described a case series of patients with bilateral hilar lymphadenopathy and coexisting erythema nodosum and polyarthralgia.⁶ The epidemiology favors young women.⁷ Patients with Löfgren’s syndrome present acutely (as in this case), which differs from the typical subacute course observed with sarcoidosis. In addition to the classic presentation described above, patients with Löfgren’s syndrome may demonstrate additional manifestations of sarcoidosis, including fevers, peripheral adenopathy, arthritis, and granulomatous skin lesions. Painful symptoms may require short-term anti-inflammatory treatments. Most patients do not require systemic immunosuppression. Symptoms usually decrease over several months, and the majority of patients experience complete remission within years. Rare recurrences have been described up to several years.⁸

In confirming the diagnosis of sarcoidosis, current guidelines recommend exclusion of other diseases that present similarly, a work-up that generally includes compatible laboratory tests and imaging, and histologic demonstration of noncaseating granulomas.⁹ However, Löfgren’s syndrome is a notable exception. The constellation of fever, bilateral hilar adenopathy, polyarthralgia, and erythema nodosum suffices to diagnose Löfgren’s syndrome as long as the disease remits rapidly and spontaneously.⁹ Thus, in this case, although granulomatous infiltrates were confirmed on biopsy, the diagnosis of Löfgren’s syndrome could have been based on clinical and radiologic features alone.

• Sarcoidosis is a multisystem granulomatous disease that most commonly presents with bilateral hilar adenopathy and pulmonary infiltrates but can also present atypically, including with nephrolithiasis from hypercalcemia, neurologic syndromes, and cardiac involvement.
• Löfgren’s syndrome, a variant of sarcoidosis, is characterized by relatively acute onset of fevers, erythema nodosum, bilateral hilar adenopathy, and polyarthralgia or polyarthritis. Most patients recover and manifest complete remission.
• A limited differential exists for rashes with a predilection for areas of tattoos and prior scarring, including keloids, PUPPP, Behçet’s disease, and granulomatous infiltration.

Disclosure

There are no conflicts of interest or financial disclosures to report.

A 43-year-old man with a history of asplenia, hepatitis C, and nephrolithiasis reported right-flank pain. He described severe, sharp pain that came in waves and radiated to the right groin, associated with nausea and nonbloody emesis. He noted “pink urine” but no dysuria. He had 4prior similar episodes during which he had passed kidney stones, although stone analysis had never been performed. He denied having fevers or chills.

The patient had been involved in a remote motor vehicle accident complicated by splenic laceration, for which he underwent splenectomy. He was appropriately immunized. The patient also suffered from bipolar affective disorder and untreated chronic hepatitis C infection with no evidence of cirrhosis. He smoked one pack of tobacco per day for the last 10 years and reported distant alcohol and methamphetamine use.

Right-flank pain can arise from conditions affecting the lower thorax (effusion, pneumonia, pulmonary embolism), abdomen (hepatobiliary or intestinal disease), retroperitoneum (hemorrhage or infection), musculoskeletal system, peripheral nerves (herpes zoster), or the genitourinary system (pyelonephritis). Pain radiating to the groin, discolored urine (suggesting hematuria), and history of kidney stones increase the likelihood of renal colic from nephrolithiasis.

Less commonly, flank pain and hematuria may present as initial symptoms of renal cell carcinoma, renal infarction, or aortic dissection. The patient’s immunosuppression from asplenia and active injection drug use could predispose him to septic emboli to his kidneys. Prior trauma causing aortic injury could predispose himto subsequent dissection.

The patient appeared well with a heart rate of 100 beats per minute, blood pressure 122/76 mmHg, temperature 36.8°C, respiratory rate 16 breaths per minute, and oxygen saturation 96% on room air. His cardiopulmonary and abdominal examinations were normal, and he had no costovertebral angle tenderness. His skin was warm and dry without rashes. His white blood cell (WBC) count was 26,000/μL; absolute neutrophil count was 22,000/μL. Serum chemistries were normal, including creatinine 0.63 mg/dL, calcium 8.8 mg/dL, and phosphorus 3.1 mg/dL. Lactate was 0.8 mmol/L (reference range: 0-2.0 mmol/L). Urinalysis revealed large ketones, >50 red blood cells (RBC) per high power field (HPF), <5 WBC per HPF, 1+ calcium oxalate crystals and pH 6.0. A bedside ultrasound showed mild right hydronephrosis. Computed tomography (CT) with intravenous contrast of his abdomen and pelvis demonstrated diffuse, mildly prominent subcentimeter mesenteric lymphadenopathy and no kidney stones. He was treated with intravenous fluids and pain control, and was discharged with a presumptive diagnosis of a passed kidney stone.

A passed stone would not explain this degree of leukocytosis. The CT results reduce the likelihood of a renal neoplasm, renal infarction, or pyelonephritis. Mesenteric lymphadenopathy is nonspecific, but it may signal underlying infection or malignancy with spread to lymph nodes, or it may be part of a systemic disorder causing generalized lymphadenopathy. Malignant causes of mesenteric lymphadenopathy (with no apparent primary tumor) include testicular cancer, lymphoma, and primary urogenital neoplasms.

Noninfectious illnesses to consider include neoplasms and connective tissue diseases. Malignancy would be unlikely to manifest this abruptly or produce a paraneoplastic disorder with these features. Among the connective tissue diseases, sarcoidosis warrants consideration in this patient with adenopathy, erythema nodosum, arthritis, and a predilection for skin changes in prior scars. However, it is uncommon for sarcoidosis to present so explosively. Painful oral and genital ulcers, pustular rash, polyarthritis, and erythema nodosum occur in Behçet’s disease, which is associated with pathergy (an exaggerated cutaneous response to minor trauma). Patients with Behçet’s may have eye involvement, including uveitis and a hypopion, and may develop vascular aneurysms in the pulmonary, intracranial, or visceral arteries. Renal artery involvement could cause hematuria and flank pain.

The patient described severe fatigue and drenching night sweats for two months prior to admission. He denied dyspnea or cough. He was born in the southwestern United States and had lived in California for almost a decade. He had been incarcerated for a few years and released three years prior. He had intermittently lived in homeless shelters, but currently lived alone in downtown San Francisco. He had traveled remotely to the Caribbean, and more recently traveled frequently to the Central Valley in California. The patient formerly worked as a pipe-fitter and welder. He denied animal exposure or recent sick contacts. He was sexually active with women, and intermittently used barrier protection.

His years in the southwestern United States may have exposed the patient to blastomycosis or histoplasmosis; both can mimic mycobacterial disease. Blastomycosis demonstrates a slightly stronger predilection for spreading to the bones, genitourinary tract, and central nervous system, whereas histoplasmosis is a more frequent cause of polyarthrtitis and mesenteric adenopathy. The patient’s travel to the Central Valley, California raises the possibility of coccidioidomycosis, which typically starts with pulmonary disease prior to dissemination to bones, skin, and other less common sites. Pipe-fitters are predisposed to asbestos-related illnesses, including lung cancer and mesothelioma, which would not explain this patient’s presentation. Incarceration and high-risk sexual practices increase his risk for tuberculosis, HIV, and syphilis. Widespread skin involvement is more characteristic of syphilis or primary HIV infection than of disseminated fungal or mycobacterial infection.

WBC measured 29,000/uL with a neutrophilic predominance. His peripheral blood smear was unremarkable. A comprehensive metabolic panel was normal. Lactate dehydrogenase (LDH) was 317 U/L (reference range 140-280 U/L). Erythrocyte sedimentation rate (ESR) was 39 mm/hr (reference range < 20 mm/hr) and C-reactive protein (CRP) was 66 mg/L (reference range <6.3 mg/L). Blood, urine, and throat cultures were sent. Chest radiograph showed clear lungs without adenopathy. Ankle and knee radiographs identified small effusions bilaterally without bony abnormalities. CT of his brain showed a small, hypodense lesion in the right lacrimal gland. A lumbar puncture with cerebrospinal fluid (CSF) analysis showed absence of RBCs; WBC, 2/µL; protein, 35 mg/dL; glucose, 62 mg/dL; negative gram stain. CSF bacterial and fungal cultures, venereal disease research laboratory (VDRL), herpes simplex virus polymerase chain reaction (HSV PCR), and cryptococcal antigen were sent for laboratory analysis. The patient was started on vancomycin and aztreonam.

Lesions of the lacrimal gland feature multiple causes, including autoimmune diseases (Sjögren’s, Behçet’s disease), granulomatous diseases (sarcoidosis, granulomatosis with polyangiitis), neoplasms (salivary gland tumors, lymphoma), and infections. Initiating broad-spectrum antibiotics is reasonable while awaiting additional information from blood and urine cultures, serologies for HIV and syphilis, and purified protein derivative or interferon-gamma release assay (IGRA).

If these tests fail to reveal a diagnosis, the search for atypical infections and noninfectious possibilities should expand. Histoplasmosis and blastomycosis would be the most likely fungal diseases to account for his arthritis and adenopathy. Coccidioidomycosis is less likely in light of the normal chest radiograph. Computed tomography of the chest would be reasonable to look for adenopathy, which would strengthen the case for lymphoma or sarcoidosis, and may also identify a potential site to biopsy to establish these diagnoses.

The patient continued to have intermittent fevers, sweats, and malaise over the next 3 days. All bacterial and fungal cultures remained negative, and antibiotics were discontinued. Rheumatoid factor, anticyclic citrullinated peptide, antinuclear antibody, and cryoglobulins were negative. Serum C3, C4, and angiotensin-converting enzyme (ACE) levels were normal. A rapid plasma reagin (RPR), HIV antibody, IGRA, and serum antibodies for Coccidioides, histoplasmosis, and West Nile virus were negative. Urine nucleic acid amplification testing for gonorrhea and chlamydia was negative. CSF VDRL, HSV PCR and cryptococcal antigen were negative. HSV culture from an oral ulcer showed no growth. The patient had a reactive hepatitis C antibody with a viral load of 3 million virus equivalents/mL.

The additional test results lower the likelihood of an acute infection. Uncontrolled hepatitis C increases the risk of several noninfectious manifestations. The normal results for serum complements and cryoglobulins effectively rule out cryoglobulinemia. Patients with hepatitis C have an increased risk of lymphoma, which could account for the subacute fevers, night sweats, adenopathy, elevated LDH, and the right orbital mass, but less likely for the oral ulcers, arthritis, and skin manifestations. Sarcoidosis is less likely given the lack of hilar adenopathy, relatively abrupt onset of multisystem disease, and the presence of oral ulcers. Behçet’s disease could account for his oral ulcers, erythema nodosum, and distribution of papules, pustules, and plaques with the predilection for scars and tattoos. Behçet’s could also explain the arthritis, the hematuria if the patient had renal artery involvement, and the orbital lesion. However, lymphadenopathy is not a prominent feature. At this point, tissue should be obtained for histopathology (to assess for vasculitis or granulomatous infiltration) and flow cytometry.

The most likely diagnosis is Löfgren’s syndrome, a variant of sarcoidosis characterized by erythema nodosum, bilateral hilar lymphadenopathy, and polyarthralgias or polyarthritis. Löfgren’s syndrome may include fevers, uveitis, widespread skin lesions and other systemic manifestations. Sarcoidosis could explain the lacrimal gland lesion, and could manifest with recurrent kidney stones. Oral lesions may occur in sarcoidosis. A normal serum ACE level may be observed in up to half of patients. The lack of visualized granulomas on the submental node FNA may reflect sampling error, lower likelihood of visualizing granulomas on FNA (compared with excisional biopsy), or biopsy location (hilar nodes are more likely to demonstrate sarcoid granulomas).

Although Löfgren’s syndrome is often self-limited, treatment can ameliorate symptoms. Nonsteroidal anti-inflammatory medication can be tried first, with prednisone reserved for refractory cases.

The constellation of bilateral hilar adenopathy, arthritis, and erythema nodosum was consistent with Löfgren’s syndrome, further supported by granulomatous infiltrates on biopsy. The patient’s symptoms resolved with naproxen. He was scheduled for follow-up in dermatology and rheumatology clinics and was referred to hepatology for management of hepatitis C.

Sarcoidosis is a multisystem granulomatous disease of unclear etiology. The disease derives its name from Boeck’s 1899 report describing benign cutaneous lesions that resembled sarcomas.¹ Sarcoidosis most commonly manifests as bilateral hilar adenopathy and pulmonary infiltrates, but may impact any tissue or organ, including the eyes, nonhilar lymph nodes, liver, spleen, joints, mucous membranes, and skin. Nephrolithiasis may result from hypercalcemia and/or hypercalciuria (related to granulomatous production of 1,25 vitamin D) and can be the presenting feature of sarcoidosis.² Less common presentations include neurologic sarcoidosis (which can present with seizures, aseptic meningitis, encephalopathy, neuroendocrine dysfunction, myelopathy and peripheral neuropathies), cardiac sarcoidosis (which may present with arrhythmias, valvular dysfunction, heart failure, ischemia, or pericardial disease), and Heerfordt syndrome (the constellation of parotid gland enlargement, facial palsy, anterior uveitis, and fever). Sarcoidosis may mimic other diseases, including malignancy, idiopathic pulmonary fibrosis, and infiltrative tuberculosis.³ Sarcoidosis-like reactions have occurred in response to malignancy and medications.⁴

The patient’s rash demonstrated a predilection for areas of prior scarring, which has a limited differential diagnosis. Keloids and hypertrophic scars occur at sites of former surgical wounds, lacerations, or areas of inflammation. Pruritic urticarial papules and plaques of pregnancy (PUPPP) is a benign inflammatory condition where papules cluster in areas of prior striae. Cutaneous lesions of Behçet’s syndrome display pathergy, where pustular response is observed at sites of injury. Granulomatous infiltration in sarcoidosis may demonstrate a predilection for scars and tattoos (ie, scar or tattoo sarcoidosis).⁵ Sarcoidosis can have other cutaneous manifestations, including psoriaform, ulcerative, or erythrodermic lesions; subcutaneous nodules; scarring or nonscarring alopecia; and lupus pernio – violaceous, nodular and plaque-like lesions on the nose, earlobes, cheeks, and digits.⁵

Löfgren’s syndrome is a distinct variant of sarcoidosis.In 1952, Dr. Löfgren described a case series of patients with bilateral hilar lymphadenopathy and coexisting erythema nodosum and polyarthralgia.⁶ The epidemiology favors young women.⁷ Patients with Löfgren’s syndrome present acutely (as in this case), which differs from the typical subacute course observed with sarcoidosis. In addition to the classic presentation described above, patients with Löfgren’s syndrome may demonstrate additional manifestations of sarcoidosis, including fevers, peripheral adenopathy, arthritis, and granulomatous skin lesions. Painful symptoms may require short-term anti-inflammatory treatments. Most patients do not require systemic immunosuppression. Symptoms usually decrease over several months, and the majority of patients experience complete remission within years. Rare recurrences have been described up to several years.⁸

In confirming the diagnosis of sarcoidosis, current guidelines recommend exclusion of other diseases that present similarly, a work-up that generally includes compatible laboratory tests and imaging, and histologic demonstration of noncaseating granulomas.⁹ However, Löfgren’s syndrome is a notable exception. The constellation of fever, bilateral hilar adenopathy, polyarthralgia, and erythema nodosum suffices to diagnose Löfgren’s syndrome as long as the disease remits rapidly and spontaneously.⁹ Thus, in this case, although granulomatous infiltrates were confirmed on biopsy, the diagnosis of Löfgren’s syndrome could have been based on clinical and radiologic features alone.

• Sarcoidosis is a multisystem granulomatous disease that most commonly presents with bilateral hilar adenopathy and pulmonary infiltrates but can also present atypically, including with nephrolithiasis from hypercalcemia, neurologic syndromes, and cardiac involvement.
• Löfgren’s syndrome, a variant of sarcoidosis, is characterized by relatively acute onset of fevers, erythema nodosum, bilateral hilar adenopathy, and polyarthralgia or polyarthritis. Most patients recover and manifest complete remission.
• A limited differential exists for rashes with a predilection for areas of tattoos and prior scarring, including keloids, PUPPP, Behçet’s disease, and granulomatous infiltration.

Disclosure

There are no conflicts of interest or financial disclosures to report.

What is the most common intervention to which hospitalized children are exposed? Acetaminophen? IV access? Phlebotomy? Or is it being connected to a monitor?

In a study conducted in five children’s hospitals, Schondelmeyer et al found that exposure to continuous electronic physiologic monitoring was extremely common. During a selected 24-hour window of observation, nearly 100% of PICU and NICU patients and 26%-48% of medical–surgical patients were exposed to continuous monitoring.¹ The latter is undoubtedly an underestimate given that monitoring periods less than 24 hours were not captured, patients may have been exposed before or after the 24-hour study window, and monitoring in the emergency department was not included.

The omnipresence of electronic physiologic monitoring in children’s hospitals is striking, particularly because we know very little about its benefits. Outside of the perioperative period, there is a dearth of evidence demonstrating improved outcomes for hospitalized children as a result of continuous physiologic monitoring. Guidelines for the most common inpatient pediatric conditions do not advocate for continuous physiologic monitoring. Presumably, this practice has become so pervasive in the absence of a strong evidence base and guideline recommendations because it is a passive, seemingly innocuous intervention that continuously collects important components of the physical examination (after all, they are known as “vital” signs). It is tempting to assume that providing clinicians with this information will make patients safer.

The danger of routinely exposing children to an intervention for which the benefits are unproven is that the net effect of the intervention may be harm. What could be harmful? The simple act of monitoring is distressing to children; sticky electrode pads stuck to their skin and a tangle of wires that restrict their movement–all impeding physical activity and contact with loved ones.

Then, there are the alarms. Schondelmeyer et al report a staggering number of them: between 42 and 152 alarms per monitored day on the floor; between 54 and 351 alarms in the intensive care units. The vast majority are false alarms, triggered by inappropriate preselected thresholds or displaced leads. This cacophony of noise only amplifies an already stressful environment for our patients–and their parents. Nurses and physicians are similarly stressed by alarms, not only by the noise but also by the frequent need to respond to them. The combination of frequent and largely unnecessary interruptions leads to alarm fatigue, whereby providers are desensitized to the alarms and may be slower to recognize a truly decompensating patient.

Continuous monitoring also risks overdiagnosis, the accurate detection of abnormalities that are not destined to cause problems, but nonetheless trigger interventions that can cause harm.² Studies in adult populations have demonstrated that continuous monitoring can produce overdiagnosis. Repeated Cochrane reviews conclude that continuous electronic fetal monitoring during labor is associated with overdiagnosis of fetal distress—with attendant increase in cesarean sections without decreasing the risk for important neonatal outcomes such as cerebral palsy and mortality.³ A recent randomized trial of continuous pulmonary impedance monitoring intended to reduce readmission rates in patients with CHF instead found that continuous monitoring resulted in overdiagnosis of CHF exacerbations—paradoxically increasing hospital admission with no significant change in mortality (in fact, mortality was nominally higher in the monitoring group).⁴

Pediatric providers are probably no less susceptible to the impulse to act in the face of abnormalities detected by continuous monitoring. EKGs and electrolyte panels may be ordered in response to transient arrhythmias. Similarly, it is challenging for providers to watch a monitor flashing elevated respiratory rates in an otherwise healthy infant with bronchiolitis and not seek an escalation in care, including increased oxygen flow or transfer to a higher acuity unit. Although arrhythmia and respiratory rate alarms were common in Schondelmeyer et al’s study, low oxygen level was far and away the most common alarm. Indeed, the poster child of pediatric overdiagnosis in the setting of electronic physiologic monitoring is hypoxemia. The present body of literature suggests that overreliance on pulse oximetry among patients with bronchiolitis increases admission rates to the hospital and prolongs length of stay, without a measurable improvement in morbidity or mortality.⁵

Few patients cared for at American children’s hospitals will be discharged without exposure to prolonged periods of continuous physiologic monitoring. Undoubtedly, there are inpatients who benefit from this technology, such as children on mechanical ventilators. Unfortunately, there are also patients who are undoubtedly harmed by it. Greater understanding of which types of patients are more likely to benefit and which are more likely to be harmed is needed to determine whether continuous physiologic monitoring should remain our most common hospital intervention.

Disclosures

The authors have no financial relationships relevant to this article to disclose.

Funding

No external funding was secured for this study.

What is the most common intervention to which hospitalized children are exposed? Acetaminophen? IV access? Phlebotomy? Or is it being connected to a monitor?

In a study conducted in five children’s hospitals, Schondelmeyer et al found that exposure to continuous electronic physiologic monitoring was extremely common. During a selected 24-hour window of observation, nearly 100% of PICU and NICU patients and 26%-48% of medical–surgical patients were exposed to continuous monitoring.¹ The latter is undoubtedly an underestimate given that monitoring periods less than 24 hours were not captured, patients may have been exposed before or after the 24-hour study window, and monitoring in the emergency department was not included.

The omnipresence of electronic physiologic monitoring in children’s hospitals is striking, particularly because we know very little about its benefits. Outside of the perioperative period, there is a dearth of evidence demonstrating improved outcomes for hospitalized children as a result of continuous physiologic monitoring. Guidelines for the most common inpatient pediatric conditions do not advocate for continuous physiologic monitoring. Presumably, this practice has become so pervasive in the absence of a strong evidence base and guideline recommendations because it is a passive, seemingly innocuous intervention that continuously collects important components of the physical examination (after all, they are known as “vital” signs). It is tempting to assume that providing clinicians with this information will make patients safer.

The danger of routinely exposing children to an intervention for which the benefits are unproven is that the net effect of the intervention may be harm. What could be harmful? The simple act of monitoring is distressing to children; sticky electrode pads stuck to their skin and a tangle of wires that restrict their movement–all impeding physical activity and contact with loved ones.

Then, there are the alarms. Schondelmeyer et al report a staggering number of them: between 42 and 152 alarms per monitored day on the floor; between 54 and 351 alarms in the intensive care units. The vast majority are false alarms, triggered by inappropriate preselected thresholds or displaced leads. This cacophony of noise only amplifies an already stressful environment for our patients–and their parents. Nurses and physicians are similarly stressed by alarms, not only by the noise but also by the frequent need to respond to them. The combination of frequent and largely unnecessary interruptions leads to alarm fatigue, whereby providers are desensitized to the alarms and may be slower to recognize a truly decompensating patient.

Continuous monitoring also risks overdiagnosis, the accurate detection of abnormalities that are not destined to cause problems, but nonetheless trigger interventions that can cause harm.² Studies in adult populations have demonstrated that continuous monitoring can produce overdiagnosis. Repeated Cochrane reviews conclude that continuous electronic fetal monitoring during labor is associated with overdiagnosis of fetal distress—with attendant increase in cesarean sections without decreasing the risk for important neonatal outcomes such as cerebral palsy and mortality.³ A recent randomized trial of continuous pulmonary impedance monitoring intended to reduce readmission rates in patients with CHF instead found that continuous monitoring resulted in overdiagnosis of CHF exacerbations—paradoxically increasing hospital admission with no significant change in mortality (in fact, mortality was nominally higher in the monitoring group).⁴

Pediatric providers are probably no less susceptible to the impulse to act in the face of abnormalities detected by continuous monitoring. EKGs and electrolyte panels may be ordered in response to transient arrhythmias. Similarly, it is challenging for providers to watch a monitor flashing elevated respiratory rates in an otherwise healthy infant with bronchiolitis and not seek an escalation in care, including increased oxygen flow or transfer to a higher acuity unit. Although arrhythmia and respiratory rate alarms were common in Schondelmeyer et al’s study, low oxygen level was far and away the most common alarm. Indeed, the poster child of pediatric overdiagnosis in the setting of electronic physiologic monitoring is hypoxemia. The present body of literature suggests that overreliance on pulse oximetry among patients with bronchiolitis increases admission rates to the hospital and prolongs length of stay, without a measurable improvement in morbidity or mortality.⁵

Few patients cared for at American children’s hospitals will be discharged without exposure to prolonged periods of continuous physiologic monitoring. Undoubtedly, there are inpatients who benefit from this technology, such as children on mechanical ventilators. Unfortunately, there are also patients who are undoubtedly harmed by it. Greater understanding of which types of patients are more likely to benefit and which are more likely to be harmed is needed to determine whether continuous physiologic monitoring should remain our most common hospital intervention.

Disclosures

The authors have no financial relationships relevant to this article to disclose.

Funding

No external funding was secured for this study.

What is the most common intervention to which hospitalized children are exposed? Acetaminophen? IV access? Phlebotomy? Or is it being connected to a monitor?

In a study conducted in five children’s hospitals, Schondelmeyer et al found that exposure to continuous electronic physiologic monitoring was extremely common. During a selected 24-hour window of observation, nearly 100% of PICU and NICU patients and 26%-48% of medical–surgical patients were exposed to continuous monitoring.¹ The latter is undoubtedly an underestimate given that monitoring periods less than 24 hours were not captured, patients may have been exposed before or after the 24-hour study window, and monitoring in the emergency department was not included.

The omnipresence of electronic physiologic monitoring in children’s hospitals is striking, particularly because we know very little about its benefits. Outside of the perioperative period, there is a dearth of evidence demonstrating improved outcomes for hospitalized children as a result of continuous physiologic monitoring. Guidelines for the most common inpatient pediatric conditions do not advocate for continuous physiologic monitoring. Presumably, this practice has become so pervasive in the absence of a strong evidence base and guideline recommendations because it is a passive, seemingly innocuous intervention that continuously collects important components of the physical examination (after all, they are known as “vital” signs). It is tempting to assume that providing clinicians with this information will make patients safer.

The danger of routinely exposing children to an intervention for which the benefits are unproven is that the net effect of the intervention may be harm. What could be harmful? The simple act of monitoring is distressing to children; sticky electrode pads stuck to their skin and a tangle of wires that restrict their movement–all impeding physical activity and contact with loved ones.

Then, there are the alarms. Schondelmeyer et al report a staggering number of them: between 42 and 152 alarms per monitored day on the floor; between 54 and 351 alarms in the intensive care units. The vast majority are false alarms, triggered by inappropriate preselected thresholds or displaced leads. This cacophony of noise only amplifies an already stressful environment for our patients–and their parents. Nurses and physicians are similarly stressed by alarms, not only by the noise but also by the frequent need to respond to them. The combination of frequent and largely unnecessary interruptions leads to alarm fatigue, whereby providers are desensitized to the alarms and may be slower to recognize a truly decompensating patient.

Continuous monitoring also risks overdiagnosis, the accurate detection of abnormalities that are not destined to cause problems, but nonetheless trigger interventions that can cause harm.² Studies in adult populations have demonstrated that continuous monitoring can produce overdiagnosis. Repeated Cochrane reviews conclude that continuous electronic fetal monitoring during labor is associated with overdiagnosis of fetal distress—with attendant increase in cesarean sections without decreasing the risk for important neonatal outcomes such as cerebral palsy and mortality.³ A recent randomized trial of continuous pulmonary impedance monitoring intended to reduce readmission rates in patients with CHF instead found that continuous monitoring resulted in overdiagnosis of CHF exacerbations—paradoxically increasing hospital admission with no significant change in mortality (in fact, mortality was nominally higher in the monitoring group).⁴

Pediatric providers are probably no less susceptible to the impulse to act in the face of abnormalities detected by continuous monitoring. EKGs and electrolyte panels may be ordered in response to transient arrhythmias. Similarly, it is challenging for providers to watch a monitor flashing elevated respiratory rates in an otherwise healthy infant with bronchiolitis and not seek an escalation in care, including increased oxygen flow or transfer to a higher acuity unit. Although arrhythmia and respiratory rate alarms were common in Schondelmeyer et al’s study, low oxygen level was far and away the most common alarm. Indeed, the poster child of pediatric overdiagnosis in the setting of electronic physiologic monitoring is hypoxemia. The present body of literature suggests that overreliance on pulse oximetry among patients with bronchiolitis increases admission rates to the hospital and prolongs length of stay, without a measurable improvement in morbidity or mortality.⁵

Few patients cared for at American children’s hospitals will be discharged without exposure to prolonged periods of continuous physiologic monitoring. Undoubtedly, there are inpatients who benefit from this technology, such as children on mechanical ventilators. Unfortunately, there are also patients who are undoubtedly harmed by it. Greater understanding of which types of patients are more likely to benefit and which are more likely to be harmed is needed to determine whether continuous physiologic monitoring should remain our most common hospital intervention.

Disclosures

The authors have no financial relationships relevant to this article to disclose.

Funding

No external funding was secured for this study.