Hospital Medicine

In an effort to improve the quality of care delivered to heart failure (HF) patients, the Centers for Medicare & Medicaid Services (CMS) publish hospitals’ 30-day risk-standardized mortality rates (RSMRs) for HF.¹ These mortality rates are also used by CMS to determine the financial penalties and bonuses that hospitals receive as part of the national Hospital Value-based Purchasing program.² Whether or not these efforts effectively direct patients towards high-quality providers or motivate hospitals to provide better care, few would disagree with the overarching goal of decreasing the number of patients who die from HF.

However, for some patients with chronic disease at the end of life, goals of care may change. The quality of days lived may become more important than the quantity of days lived. As a consequence, high-quality care for some patients at the end of life is associated with withdrawing life-sustaining or life-extending therapies. Over time, this therapeutic perspective has become more common, with use of hospice care doubling from 23% to 47% between 2000 and 2012 among Medicare beneficiaries who died.³ For a national cohort of older patients admitted with HF—not just those patients who died in that same year—hospitals’ rates of referral to hospice are considerably lower, averaging 2.9% in 2010 in a national study.⁴ Nevertheless, it is possible that hospitals that more faithfully follow their dying patients’ wishes and withdraw life-prolonging interventions and provide comfort-focused care at the end of life might be unfairly penalized if such efforts resulted in higher mortality rates than other hospitals.

Therefore, we used Medicare data linked to a national HF registry with information about end-of-life care, to address 3 questions: (1) How much do hospitals vary in their rates of early comfort care and how has this changed over time; (2) What hospital and patient factors are associated with higher early comfort care rates; and (3) Is there a correlation between 30-day risk-adjusted mortality rates for HF with hospital rates of early comfort care?

Data Sources

We used data from the American Heart Association’s Get With The Guidelines-Heart Failure (GWTG-HF) registry. GWTG-HF is a voluntary, inpatient, quality improvement registry^5-7 that uses web-based tools and standard questionnaires to collect data on patients with HF admitted to participating hospitals nationwide. The data include information from admission (eg, sociodemographic characteristics, symptoms, medical history, and initial laboratory and test results), the inpatient stay (eg, therapies), and discharge (eg, discharge destination, whether and when comfort care was initiated). We linked the GWTG-HF registry data to Medicare claims data in order to obtain information about Medicare eligibility and patient comorbidities. Additionally, we used data from the American Hospital Association (2008) for hospital characteristics. Quintiles Real-World & Late Phase Research (Cambridge, MA) serves as the data coordinating center for GWTG-HF and the Duke Clinical Research Institute (Durham, NC) serves as the statistical analytic center. GWTG-HF participating sites have a waiver of informed consent because the data are de-identified and primarily used for quality improvement. All analyses performed on this data have been approved by the Duke Medical Center Institutional Review Board.

Study Population

Study Outcomes

Our outcome of interest was the correlation between a hospital’s rate of initiating early CMO for admitted HF patients and a hospital’s 30-day RSMR for HF. The GWTG-HF questionnaire⁸ asks “When is the earliest physician/advanced practice nurse/physician assistant documentation of comfort measures only?” and permits 4 responses: day 0 or 1, day 2 or after, timing unclear, or not documented/unable to determine. We defined early CMO as CMO on day 0 or 1, and late/no CMO as any other response. We chose to examine early comfort care because many hospitalized patients transition to comfort care before they die if the death is in any way predictable. Thus, if comfort care is measured at any time during the hospitalization, hospitals that have high mortality rates are likely to have high comfort care rates. Therefore, we chose to use the more precise measure of early comfort care. We created hospital-level, risk-standardized early comfort care rates using the same risk-adjustment model used for RSMRs but with the outcome of early comfort care instead of mortality.^9,10

RSMRs were calculated using a validated GWTG-HF 30-day risk-standardized mortality model⁹ with additional variables identified from other GWTG-HF analyses.¹⁰ The 30 days are measured as the 30 days after the index admission date.

Statistical Analyses

We described trends in early comfort care rates over time, from February 17, 2008, to February 17, 2014, using the Cochran-Armitage test for trend. We then grouped hospitals into quintiles based on their unadjusted early comfort care rates. We described patient and hospital characteristics for each quintile, using χ² tests to test for differences across quintiles for categorical variables and Wilcoxon rank sum tests to assess for differences across quintiles for continuous variables. We then examined the Spearman’s rank correlation between hospitals’ RSMR and risk-adjusted comfort care rates. Finally, we compared hospital-level RSMRs before and after adjusting for early comfort care.

We performed risk-adjustment for these last 2 analyses as follows. For each patient, covariates were obtained from the GWTG-HF registry. Clinical data captured for the index admission were utilized in the risk-adjustment model (for both RSMRs and risk-adjusted comfort care rates). Included covariates were as follows: age (per 10 years); race (black vs non-black); systolic blood pressure at admission ≤170 (per 10 mm Hg); respiratory rate (per 5 respirations/min); heart rate ≤105 (per 10 beats/min); weight ≤100 (per 5 kg); weight >100 (per 5 kg); blood urea nitrogen (per 10 mg/dl); brain natriuretic peptide ≤2000 (per 500 pg/ml); hemoglobin 10-14 (per 1 g/dl); troponin abnormal (vs normal); creatinine ≤1 (per 1 mg/dl); sodium 130-140 (per 5 mEq/l); and chronic obstructive pulmonary disease or asthma.

Hierarchical logistic regression modeling was used to calculate the hospital-specific RSMR. A predicted/expected ratio similar to an observed/expected (O/E) ratio was calculated using the following modifications: (1) instead of the observed (crude) number of deaths, the numerator is the number of deaths predicted by the hierarchical model among a hospital’s patients given the patients’ risk factors and the hospital-specific effect; (2) the denominator is the expected number of deaths among the hospital’s patients given the patients’ risk factors and the average of all hospital-specific effects overall; and (3) the ratio of the numerator and denominator are then multiplied by the observed overall mortality rate (same as O/E). This calculation is the method used by CMS to derive RSMRs.¹¹ Multiple imputation was used to handle missing data in the models; 25 imputed datasets using the fully conditional specification method were created. Patients with missing prior comorbidities were assumed to not have those conditions. Hospital characteristics were not imputed; therefore, for analyses that required construction of risk-adjusted comfort care rates or RSMRs, we excluded 18,867 patients cared for at 82 hospitals missing hospital characteristics. We ran 2 sets of models for risk-adjusted comfort care rates and RSMRs: the first adjusted only for patient characteristics, and the second adjusted for both patient and hospital characteristics. Results from the 2 models were similar, so we present only results from the latter. Variance inflation factors were all <2, indicating the collinearity between covariates was not an issue.

All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Cary, NC). We tested for statistical significance by using 2-tailed tests and considered P values <.05 to be statistically significant.

Of the 272 hospitals included in our final study cohort, the observed median overall rate of early comfort care in this study was 1.9% (25th to 75th percentile: 0.9% to 4.0%); hospitals varied widely in unadjusted early comfort care rates (0.00% to 0.46% in the lowest quintile, and 4.60% to 39.91% in the highest quintile; Table 1).

Among a national sample of US hospitals, we found wide variation in how frequently health care providers deliver comfort care within the first 2 days of admission for HF. A minority of hospitals reported no early comfort care on any patients throughout the 6-year study period, but hospitals in the highest quintile initiated early comfort care rates for at least 1 in 20 HF patients. Hospitals that were more likely to initiate early comfort care had a higher proportion of female and white patients and were less likely to have the capacity to deliver aggressive surgical interventions such as heart transplants. Hospital-level 30-day RSMRs were not correlated with rates of early comfort care.

While the appropriate rate of early comfort care for patients hospitalized with HF is unknown, given that the average hospital RSMR is approximately 12% for fee-for-service Medicare patients hospitalized with HF,¹² it is surprising that some hospitals initiated early comfort care on none or very few of their HF patients. It is quite possible that many of these hospitals initiated comfort care for some of their patients after 48 hours of hospitalization. We were unable to estimate the average period of time patients received comfort care prior to dying, the degree to which this varies across hospitals or why it might vary, and whether the length of time between comfort care initiation and death is related to satisfaction with end-of-life care. Future research on these topics would help inform providers seeking to deliver better end-of-life care. In this study, we also were unable to estimate how often early comfort care was not initiated because patients had a good prognosis. However, prior studies have suggested low rates of comfort care or hospice referral even among patients at very high estimated mortality risk.⁴ It is also possible that providers and families had concerns about the ability to accurately prognosticate, although several models have been shown to perform acceptably for patients hospitalized with HF.¹³

We found that comfort care rates did not increase over time, even though use of hospice care doubled among Medicare beneficiaries between 2000 and 2012. By way of context, cancer—the second leading cause of death in the US—was responsible for 38% of hospice admissions in 2013, whereas heart disease (including but not limited to HF)—the leading cause of death— was responsible for 13% of hospice admissions.¹⁴ The 2013 American College of Cardiology Foundation and the American Heart Association guidelines for HF recommend consideration of hospice or palliative care for inpatient and transitional care.¹⁵ In future work, it would be important to better understand the drivers behind decisions around comfort care for patients hospitalized with HF.

With regards to the policy implications of our study, we found that on average, adjusting 30-day mortality rates for early comfort care was not associated with a change in hospital mortality rankings. For those hospitals with high comfort care rates, adjusting for comfort care did lower mortality rates, but the change was so small as to be clinically insignificant. CMS’ RSMR for HF excludes patients enrolled in hospice during the 12 months prior to index admission, including the first day of the index admission, acknowledging that death may not be an untoward outcome for such patients.¹⁶ Fee-for-service Medicare beneficiaries excluded for hospice enrollment comprised 1.29% of HF admissions from July 2012 to June 2015¹⁶ and are likely a subset of early comfort care patients in our sample, both because of the inclusiveness of chart review (vs claims-based identification) and because we defined early comfort care as comfort care initiated on day 0 or 1 of hospitalization. Nevertheless, with our data we cannot assess to what degree our findings were due solely to hospice patients excluded from CMS’ current estimates.

Prior research has described the underuse of palliative care among patients with HF¹⁷ and the association of palliative care with better patient and family experiences at the end of life.^18-20 We add to this literature by describing the epidemiology—prevalence, changes over time, and associated factors—of early comfort care for HF in a national sample of hospitals. This serves as a baseline for future work on end-of-life care among patients hospitalized for HF. Our findings also contribute to ongoing discussion about how best to risk-adjust mortality metrics used to assess hospital quality in pay-for-performance programs. Recent research on stroke and pneumonia based on California data suggests that not accounting for do-not-resuscitate (DNR) status biases hospital mortality rates.^21,22 Earlier research examined the impact of adjusting hospital mortality rates for DNR for a broader range of conditions.^23,24 We expand this line of inquiry by examining the hospital-level association of early comfort care with mortality rates for HF, utilizing a national, contemporary cohort of inpatient stays. In addition, while studies have found that DNR rates within the first 24 hours of admission are relatively high (median 15.8% for pneumonia; 13.3% for stroke),^21,22 comfort care is distinct from DNR.

Our findings should be interpreted in the context of several potential limitations. First, we did not have any information about patient or family wishes regarding end-of-life care, or the exact timing of early comfort care (eg, day 0 or day 1). The initiation of comfort care usually follows conversations about end-of-life care involving a patient, his or her family, and the medical team. Thus, we do not know if low early comfort care rates represent the lack of such a conversation (and thus poor-quality care) or the desire by most patients not to initiate early comfort care (and thus high-quality care). This would be an important area for future research. Second, we included only patients admitted to hospitals that participate in GWTG-HF, a voluntary quality improvement initiative. This may limit the generalizability of our findings, but it is unclear how our sample might bias our findings. Hospitals engaged in quality improvement may be more likely to initiate early comfort care aligned with patients’ wishes; on the other hand, hospitals with advanced surgical capabilities are over-represented in our sample and these hospitals are less likely to initiate early comfort care. Third, we examined associations and cannot make conclusions about causality. Residual measured and unmeasured confounding may influence these findings.

In summary, we found that early comfort care rates for fee-for-service Medicare beneficiaries admitted for HF varies widely among hospitals, but median rates of early comfort care have not changed over time. On average, there was no correlation between hospital-level, 30-day, RSMRs and rates of early comfort care. This suggests that current efforts to lower mortality rates have not had unintended consequences for hospitals that institute early comfort care more commonly than their peers.

Acknowledgments

Dr. Chen and Ms. Cox take responsibility for the integrity of the data and the accuracy of the data analysis. Drs. Chen, Levine, and Hayward are responsible for the study concept and design. Drs. Chen and Fonarow acquired the data. Dr. Chen drafted the manuscript. Drs. Chen, Levin, Hayward, Cox, Fonarow, DeVore, Hernandez, Heidenreich, and Yancy revised the manuscript for important intellectual content. Drs. Chen, Hayward, Cox, and Schulte performed the statistical analysis. Drs. Chen and Fonarow obtained funding for the study. Drs. Hayward and Fonarow supervised the study. The authors thank Bailey Green, MPH, for the research assistance she provided. She was compensated for her work.

Disclosure

Dr. Fonarow reports research support from the National Institutes of Health, and consulting for Amgen, Janssen, Novartis, Medtronic, and St Jude Medical. Dr. DeVore reports research support from the American Heart Association, Amgen, and Novartis, and consulting for Amgen. The other authors have no relevant conflicts of interest. Dr. Chen was supported by a Career Development Grant Award (K08HS020671) from the Agency for Healthcare Research and Quality when the manuscript was being prepared. She currently receives support from the Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation for her work there. She also receives support from the Blue Cross Blue Shield of Michigan Foundation’s Investigator Initiated Research Program, the Agency for Healthcare Research and Quality (R01 HS024698), and the National Institute on Aging (P01 AG019783). These funding sources had no role in the preparation, review, or approval of the manuscript. The GWTG-HF program is provided by the American Heart Association. GWTG-HF has been funded in the past through support from Amgen, Medtronic, GlaxoSmithKline, Ortho-McNeil, and the American Heart Association Pharmaceutical Roundtable. These sponsors had no role in the study design, data analysis or manuscript preparation and revision.

In an effort to improve the quality of care delivered to heart failure (HF) patients, the Centers for Medicare & Medicaid Services (CMS) publish hospitals’ 30-day risk-standardized mortality rates (RSMRs) for HF.¹ These mortality rates are also used by CMS to determine the financial penalties and bonuses that hospitals receive as part of the national Hospital Value-based Purchasing program.² Whether or not these efforts effectively direct patients towards high-quality providers or motivate hospitals to provide better care, few would disagree with the overarching goal of decreasing the number of patients who die from HF.

However, for some patients with chronic disease at the end of life, goals of care may change. The quality of days lived may become more important than the quantity of days lived. As a consequence, high-quality care for some patients at the end of life is associated with withdrawing life-sustaining or life-extending therapies. Over time, this therapeutic perspective has become more common, with use of hospice care doubling from 23% to 47% between 2000 and 2012 among Medicare beneficiaries who died.³ For a national cohort of older patients admitted with HF—not just those patients who died in that same year—hospitals’ rates of referral to hospice are considerably lower, averaging 2.9% in 2010 in a national study.⁴ Nevertheless, it is possible that hospitals that more faithfully follow their dying patients’ wishes and withdraw life-prolonging interventions and provide comfort-focused care at the end of life might be unfairly penalized if such efforts resulted in higher mortality rates than other hospitals.

Therefore, we used Medicare data linked to a national HF registry with information about end-of-life care, to address 3 questions: (1) How much do hospitals vary in their rates of early comfort care and how has this changed over time; (2) What hospital and patient factors are associated with higher early comfort care rates; and (3) Is there a correlation between 30-day risk-adjusted mortality rates for HF with hospital rates of early comfort care?

Data Sources

We used data from the American Heart Association’s Get With The Guidelines-Heart Failure (GWTG-HF) registry. GWTG-HF is a voluntary, inpatient, quality improvement registry^5-7 that uses web-based tools and standard questionnaires to collect data on patients with HF admitted to participating hospitals nationwide. The data include information from admission (eg, sociodemographic characteristics, symptoms, medical history, and initial laboratory and test results), the inpatient stay (eg, therapies), and discharge (eg, discharge destination, whether and when comfort care was initiated). We linked the GWTG-HF registry data to Medicare claims data in order to obtain information about Medicare eligibility and patient comorbidities. Additionally, we used data from the American Hospital Association (2008) for hospital characteristics. Quintiles Real-World & Late Phase Research (Cambridge, MA) serves as the data coordinating center for GWTG-HF and the Duke Clinical Research Institute (Durham, NC) serves as the statistical analytic center. GWTG-HF participating sites have a waiver of informed consent because the data are de-identified and primarily used for quality improvement. All analyses performed on this data have been approved by the Duke Medical Center Institutional Review Board.

Study Population

Study Outcomes

Our outcome of interest was the correlation between a hospital’s rate of initiating early CMO for admitted HF patients and a hospital’s 30-day RSMR for HF. The GWTG-HF questionnaire⁸ asks “When is the earliest physician/advanced practice nurse/physician assistant documentation of comfort measures only?” and permits 4 responses: day 0 or 1, day 2 or after, timing unclear, or not documented/unable to determine. We defined early CMO as CMO on day 0 or 1, and late/no CMO as any other response. We chose to examine early comfort care because many hospitalized patients transition to comfort care before they die if the death is in any way predictable. Thus, if comfort care is measured at any time during the hospitalization, hospitals that have high mortality rates are likely to have high comfort care rates. Therefore, we chose to use the more precise measure of early comfort care. We created hospital-level, risk-standardized early comfort care rates using the same risk-adjustment model used for RSMRs but with the outcome of early comfort care instead of mortality.^9,10

RSMRs were calculated using a validated GWTG-HF 30-day risk-standardized mortality model⁹ with additional variables identified from other GWTG-HF analyses.¹⁰ The 30 days are measured as the 30 days after the index admission date.

Statistical Analyses

We described trends in early comfort care rates over time, from February 17, 2008, to February 17, 2014, using the Cochran-Armitage test for trend. We then grouped hospitals into quintiles based on their unadjusted early comfort care rates. We described patient and hospital characteristics for each quintile, using χ² tests to test for differences across quintiles for categorical variables and Wilcoxon rank sum tests to assess for differences across quintiles for continuous variables. We then examined the Spearman’s rank correlation between hospitals’ RSMR and risk-adjusted comfort care rates. Finally, we compared hospital-level RSMRs before and after adjusting for early comfort care.

We performed risk-adjustment for these last 2 analyses as follows. For each patient, covariates were obtained from the GWTG-HF registry. Clinical data captured for the index admission were utilized in the risk-adjustment model (for both RSMRs and risk-adjusted comfort care rates). Included covariates were as follows: age (per 10 years); race (black vs non-black); systolic blood pressure at admission ≤170 (per 10 mm Hg); respiratory rate (per 5 respirations/min); heart rate ≤105 (per 10 beats/min); weight ≤100 (per 5 kg); weight >100 (per 5 kg); blood urea nitrogen (per 10 mg/dl); brain natriuretic peptide ≤2000 (per 500 pg/ml); hemoglobin 10-14 (per 1 g/dl); troponin abnormal (vs normal); creatinine ≤1 (per 1 mg/dl); sodium 130-140 (per 5 mEq/l); and chronic obstructive pulmonary disease or asthma.

Hierarchical logistic regression modeling was used to calculate the hospital-specific RSMR. A predicted/expected ratio similar to an observed/expected (O/E) ratio was calculated using the following modifications: (1) instead of the observed (crude) number of deaths, the numerator is the number of deaths predicted by the hierarchical model among a hospital’s patients given the patients’ risk factors and the hospital-specific effect; (2) the denominator is the expected number of deaths among the hospital’s patients given the patients’ risk factors and the average of all hospital-specific effects overall; and (3) the ratio of the numerator and denominator are then multiplied by the observed overall mortality rate (same as O/E). This calculation is the method used by CMS to derive RSMRs.¹¹ Multiple imputation was used to handle missing data in the models; 25 imputed datasets using the fully conditional specification method were created. Patients with missing prior comorbidities were assumed to not have those conditions. Hospital characteristics were not imputed; therefore, for analyses that required construction of risk-adjusted comfort care rates or RSMRs, we excluded 18,867 patients cared for at 82 hospitals missing hospital characteristics. We ran 2 sets of models for risk-adjusted comfort care rates and RSMRs: the first adjusted only for patient characteristics, and the second adjusted for both patient and hospital characteristics. Results from the 2 models were similar, so we present only results from the latter. Variance inflation factors were all <2, indicating the collinearity between covariates was not an issue.

All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Cary, NC). We tested for statistical significance by using 2-tailed tests and considered P values <.05 to be statistically significant.

Of the 272 hospitals included in our final study cohort, the observed median overall rate of early comfort care in this study was 1.9% (25th to 75th percentile: 0.9% to 4.0%); hospitals varied widely in unadjusted early comfort care rates (0.00% to 0.46% in the lowest quintile, and 4.60% to 39.91% in the highest quintile; Table 1).

Among a national sample of US hospitals, we found wide variation in how frequently health care providers deliver comfort care within the first 2 days of admission for HF. A minority of hospitals reported no early comfort care on any patients throughout the 6-year study period, but hospitals in the highest quintile initiated early comfort care rates for at least 1 in 20 HF patients. Hospitals that were more likely to initiate early comfort care had a higher proportion of female and white patients and were less likely to have the capacity to deliver aggressive surgical interventions such as heart transplants. Hospital-level 30-day RSMRs were not correlated with rates of early comfort care.

While the appropriate rate of early comfort care for patients hospitalized with HF is unknown, given that the average hospital RSMR is approximately 12% for fee-for-service Medicare patients hospitalized with HF,¹² it is surprising that some hospitals initiated early comfort care on none or very few of their HF patients. It is quite possible that many of these hospitals initiated comfort care for some of their patients after 48 hours of hospitalization. We were unable to estimate the average period of time patients received comfort care prior to dying, the degree to which this varies across hospitals or why it might vary, and whether the length of time between comfort care initiation and death is related to satisfaction with end-of-life care. Future research on these topics would help inform providers seeking to deliver better end-of-life care. In this study, we also were unable to estimate how often early comfort care was not initiated because patients had a good prognosis. However, prior studies have suggested low rates of comfort care or hospice referral even among patients at very high estimated mortality risk.⁴ It is also possible that providers and families had concerns about the ability to accurately prognosticate, although several models have been shown to perform acceptably for patients hospitalized with HF.¹³

We found that comfort care rates did not increase over time, even though use of hospice care doubled among Medicare beneficiaries between 2000 and 2012. By way of context, cancer—the second leading cause of death in the US—was responsible for 38% of hospice admissions in 2013, whereas heart disease (including but not limited to HF)—the leading cause of death— was responsible for 13% of hospice admissions.¹⁴ The 2013 American College of Cardiology Foundation and the American Heart Association guidelines for HF recommend consideration of hospice or palliative care for inpatient and transitional care.¹⁵ In future work, it would be important to better understand the drivers behind decisions around comfort care for patients hospitalized with HF.

With regards to the policy implications of our study, we found that on average, adjusting 30-day mortality rates for early comfort care was not associated with a change in hospital mortality rankings. For those hospitals with high comfort care rates, adjusting for comfort care did lower mortality rates, but the change was so small as to be clinically insignificant. CMS’ RSMR for HF excludes patients enrolled in hospice during the 12 months prior to index admission, including the first day of the index admission, acknowledging that death may not be an untoward outcome for such patients.¹⁶ Fee-for-service Medicare beneficiaries excluded for hospice enrollment comprised 1.29% of HF admissions from July 2012 to June 2015¹⁶ and are likely a subset of early comfort care patients in our sample, both because of the inclusiveness of chart review (vs claims-based identification) and because we defined early comfort care as comfort care initiated on day 0 or 1 of hospitalization. Nevertheless, with our data we cannot assess to what degree our findings were due solely to hospice patients excluded from CMS’ current estimates.

Prior research has described the underuse of palliative care among patients with HF¹⁷ and the association of palliative care with better patient and family experiences at the end of life.^18-20 We add to this literature by describing the epidemiology—prevalence, changes over time, and associated factors—of early comfort care for HF in a national sample of hospitals. This serves as a baseline for future work on end-of-life care among patients hospitalized for HF. Our findings also contribute to ongoing discussion about how best to risk-adjust mortality metrics used to assess hospital quality in pay-for-performance programs. Recent research on stroke and pneumonia based on California data suggests that not accounting for do-not-resuscitate (DNR) status biases hospital mortality rates.^21,22 Earlier research examined the impact of adjusting hospital mortality rates for DNR for a broader range of conditions.^23,24 We expand this line of inquiry by examining the hospital-level association of early comfort care with mortality rates for HF, utilizing a national, contemporary cohort of inpatient stays. In addition, while studies have found that DNR rates within the first 24 hours of admission are relatively high (median 15.8% for pneumonia; 13.3% for stroke),^21,22 comfort care is distinct from DNR.

Our findings should be interpreted in the context of several potential limitations. First, we did not have any information about patient or family wishes regarding end-of-life care, or the exact timing of early comfort care (eg, day 0 or day 1). The initiation of comfort care usually follows conversations about end-of-life care involving a patient, his or her family, and the medical team. Thus, we do not know if low early comfort care rates represent the lack of such a conversation (and thus poor-quality care) or the desire by most patients not to initiate early comfort care (and thus high-quality care). This would be an important area for future research. Second, we included only patients admitted to hospitals that participate in GWTG-HF, a voluntary quality improvement initiative. This may limit the generalizability of our findings, but it is unclear how our sample might bias our findings. Hospitals engaged in quality improvement may be more likely to initiate early comfort care aligned with patients’ wishes; on the other hand, hospitals with advanced surgical capabilities are over-represented in our sample and these hospitals are less likely to initiate early comfort care. Third, we examined associations and cannot make conclusions about causality. Residual measured and unmeasured confounding may influence these findings.

In summary, we found that early comfort care rates for fee-for-service Medicare beneficiaries admitted for HF varies widely among hospitals, but median rates of early comfort care have not changed over time. On average, there was no correlation between hospital-level, 30-day, RSMRs and rates of early comfort care. This suggests that current efforts to lower mortality rates have not had unintended consequences for hospitals that institute early comfort care more commonly than their peers.

Acknowledgments

Dr. Chen and Ms. Cox take responsibility for the integrity of the data and the accuracy of the data analysis. Drs. Chen, Levine, and Hayward are responsible for the study concept and design. Drs. Chen and Fonarow acquired the data. Dr. Chen drafted the manuscript. Drs. Chen, Levin, Hayward, Cox, Fonarow, DeVore, Hernandez, Heidenreich, and Yancy revised the manuscript for important intellectual content. Drs. Chen, Hayward, Cox, and Schulte performed the statistical analysis. Drs. Chen and Fonarow obtained funding for the study. Drs. Hayward and Fonarow supervised the study. The authors thank Bailey Green, MPH, for the research assistance she provided. She was compensated for her work.

Disclosure

Dr. Fonarow reports research support from the National Institutes of Health, and consulting for Amgen, Janssen, Novartis, Medtronic, and St Jude Medical. Dr. DeVore reports research support from the American Heart Association, Amgen, and Novartis, and consulting for Amgen. The other authors have no relevant conflicts of interest. Dr. Chen was supported by a Career Development Grant Award (K08HS020671) from the Agency for Healthcare Research and Quality when the manuscript was being prepared. She currently receives support from the Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation for her work there. She also receives support from the Blue Cross Blue Shield of Michigan Foundation’s Investigator Initiated Research Program, the Agency for Healthcare Research and Quality (R01 HS024698), and the National Institute on Aging (P01 AG019783). These funding sources had no role in the preparation, review, or approval of the manuscript. The GWTG-HF program is provided by the American Heart Association. GWTG-HF has been funded in the past through support from Amgen, Medtronic, GlaxoSmithKline, Ortho-McNeil, and the American Heart Association Pharmaceutical Roundtable. These sponsors had no role in the study design, data analysis or manuscript preparation and revision.

In an effort to improve the quality of care delivered to heart failure (HF) patients, the Centers for Medicare & Medicaid Services (CMS) publish hospitals’ 30-day risk-standardized mortality rates (RSMRs) for HF.¹ These mortality rates are also used by CMS to determine the financial penalties and bonuses that hospitals receive as part of the national Hospital Value-based Purchasing program.² Whether or not these efforts effectively direct patients towards high-quality providers or motivate hospitals to provide better care, few would disagree with the overarching goal of decreasing the number of patients who die from HF.

However, for some patients with chronic disease at the end of life, goals of care may change. The quality of days lived may become more important than the quantity of days lived. As a consequence, high-quality care for some patients at the end of life is associated with withdrawing life-sustaining or life-extending therapies. Over time, this therapeutic perspective has become more common, with use of hospice care doubling from 23% to 47% between 2000 and 2012 among Medicare beneficiaries who died.³ For a national cohort of older patients admitted with HF—not just those patients who died in that same year—hospitals’ rates of referral to hospice are considerably lower, averaging 2.9% in 2010 in a national study.⁴ Nevertheless, it is possible that hospitals that more faithfully follow their dying patients’ wishes and withdraw life-prolonging interventions and provide comfort-focused care at the end of life might be unfairly penalized if such efforts resulted in higher mortality rates than other hospitals.

Therefore, we used Medicare data linked to a national HF registry with information about end-of-life care, to address 3 questions: (1) How much do hospitals vary in their rates of early comfort care and how has this changed over time; (2) What hospital and patient factors are associated with higher early comfort care rates; and (3) Is there a correlation between 30-day risk-adjusted mortality rates for HF with hospital rates of early comfort care?

Data Sources

We used data from the American Heart Association’s Get With The Guidelines-Heart Failure (GWTG-HF) registry. GWTG-HF is a voluntary, inpatient, quality improvement registry^5-7 that uses web-based tools and standard questionnaires to collect data on patients with HF admitted to participating hospitals nationwide. The data include information from admission (eg, sociodemographic characteristics, symptoms, medical history, and initial laboratory and test results), the inpatient stay (eg, therapies), and discharge (eg, discharge destination, whether and when comfort care was initiated). We linked the GWTG-HF registry data to Medicare claims data in order to obtain information about Medicare eligibility and patient comorbidities. Additionally, we used data from the American Hospital Association (2008) for hospital characteristics. Quintiles Real-World & Late Phase Research (Cambridge, MA) serves as the data coordinating center for GWTG-HF and the Duke Clinical Research Institute (Durham, NC) serves as the statistical analytic center. GWTG-HF participating sites have a waiver of informed consent because the data are de-identified and primarily used for quality improvement. All analyses performed on this data have been approved by the Duke Medical Center Institutional Review Board.

Study Population

Study Outcomes

Our outcome of interest was the correlation between a hospital’s rate of initiating early CMO for admitted HF patients and a hospital’s 30-day RSMR for HF. The GWTG-HF questionnaire⁸ asks “When is the earliest physician/advanced practice nurse/physician assistant documentation of comfort measures only?” and permits 4 responses: day 0 or 1, day 2 or after, timing unclear, or not documented/unable to determine. We defined early CMO as CMO on day 0 or 1, and late/no CMO as any other response. We chose to examine early comfort care because many hospitalized patients transition to comfort care before they die if the death is in any way predictable. Thus, if comfort care is measured at any time during the hospitalization, hospitals that have high mortality rates are likely to have high comfort care rates. Therefore, we chose to use the more precise measure of early comfort care. We created hospital-level, risk-standardized early comfort care rates using the same risk-adjustment model used for RSMRs but with the outcome of early comfort care instead of mortality.^9,10

RSMRs were calculated using a validated GWTG-HF 30-day risk-standardized mortality model⁹ with additional variables identified from other GWTG-HF analyses.¹⁰ The 30 days are measured as the 30 days after the index admission date.

Statistical Analyses

We described trends in early comfort care rates over time, from February 17, 2008, to February 17, 2014, using the Cochran-Armitage test for trend. We then grouped hospitals into quintiles based on their unadjusted early comfort care rates. We described patient and hospital characteristics for each quintile, using χ² tests to test for differences across quintiles for categorical variables and Wilcoxon rank sum tests to assess for differences across quintiles for continuous variables. We then examined the Spearman’s rank correlation between hospitals’ RSMR and risk-adjusted comfort care rates. Finally, we compared hospital-level RSMRs before and after adjusting for early comfort care.

We performed risk-adjustment for these last 2 analyses as follows. For each patient, covariates were obtained from the GWTG-HF registry. Clinical data captured for the index admission were utilized in the risk-adjustment model (for both RSMRs and risk-adjusted comfort care rates). Included covariates were as follows: age (per 10 years); race (black vs non-black); systolic blood pressure at admission ≤170 (per 10 mm Hg); respiratory rate (per 5 respirations/min); heart rate ≤105 (per 10 beats/min); weight ≤100 (per 5 kg); weight >100 (per 5 kg); blood urea nitrogen (per 10 mg/dl); brain natriuretic peptide ≤2000 (per 500 pg/ml); hemoglobin 10-14 (per 1 g/dl); troponin abnormal (vs normal); creatinine ≤1 (per 1 mg/dl); sodium 130-140 (per 5 mEq/l); and chronic obstructive pulmonary disease or asthma.

Hierarchical logistic regression modeling was used to calculate the hospital-specific RSMR. A predicted/expected ratio similar to an observed/expected (O/E) ratio was calculated using the following modifications: (1) instead of the observed (crude) number of deaths, the numerator is the number of deaths predicted by the hierarchical model among a hospital’s patients given the patients’ risk factors and the hospital-specific effect; (2) the denominator is the expected number of deaths among the hospital’s patients given the patients’ risk factors and the average of all hospital-specific effects overall; and (3) the ratio of the numerator and denominator are then multiplied by the observed overall mortality rate (same as O/E). This calculation is the method used by CMS to derive RSMRs.¹¹ Multiple imputation was used to handle missing data in the models; 25 imputed datasets using the fully conditional specification method were created. Patients with missing prior comorbidities were assumed to not have those conditions. Hospital characteristics were not imputed; therefore, for analyses that required construction of risk-adjusted comfort care rates or RSMRs, we excluded 18,867 patients cared for at 82 hospitals missing hospital characteristics. We ran 2 sets of models for risk-adjusted comfort care rates and RSMRs: the first adjusted only for patient characteristics, and the second adjusted for both patient and hospital characteristics. Results from the 2 models were similar, so we present only results from the latter. Variance inflation factors were all <2, indicating the collinearity between covariates was not an issue.

All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Cary, NC). We tested for statistical significance by using 2-tailed tests and considered P values <.05 to be statistically significant.

Of the 272 hospitals included in our final study cohort, the observed median overall rate of early comfort care in this study was 1.9% (25th to 75th percentile: 0.9% to 4.0%); hospitals varied widely in unadjusted early comfort care rates (0.00% to 0.46% in the lowest quintile, and 4.60% to 39.91% in the highest quintile; Table 1).

Among a national sample of US hospitals, we found wide variation in how frequently health care providers deliver comfort care within the first 2 days of admission for HF. A minority of hospitals reported no early comfort care on any patients throughout the 6-year study period, but hospitals in the highest quintile initiated early comfort care rates for at least 1 in 20 HF patients. Hospitals that were more likely to initiate early comfort care had a higher proportion of female and white patients and were less likely to have the capacity to deliver aggressive surgical interventions such as heart transplants. Hospital-level 30-day RSMRs were not correlated with rates of early comfort care.

While the appropriate rate of early comfort care for patients hospitalized with HF is unknown, given that the average hospital RSMR is approximately 12% for fee-for-service Medicare patients hospitalized with HF,¹² it is surprising that some hospitals initiated early comfort care on none or very few of their HF patients. It is quite possible that many of these hospitals initiated comfort care for some of their patients after 48 hours of hospitalization. We were unable to estimate the average period of time patients received comfort care prior to dying, the degree to which this varies across hospitals or why it might vary, and whether the length of time between comfort care initiation and death is related to satisfaction with end-of-life care. Future research on these topics would help inform providers seeking to deliver better end-of-life care. In this study, we also were unable to estimate how often early comfort care was not initiated because patients had a good prognosis. However, prior studies have suggested low rates of comfort care or hospice referral even among patients at very high estimated mortality risk.⁴ It is also possible that providers and families had concerns about the ability to accurately prognosticate, although several models have been shown to perform acceptably for patients hospitalized with HF.¹³

We found that comfort care rates did not increase over time, even though use of hospice care doubled among Medicare beneficiaries between 2000 and 2012. By way of context, cancer—the second leading cause of death in the US—was responsible for 38% of hospice admissions in 2013, whereas heart disease (including but not limited to HF)—the leading cause of death— was responsible for 13% of hospice admissions.¹⁴ The 2013 American College of Cardiology Foundation and the American Heart Association guidelines for HF recommend consideration of hospice or palliative care for inpatient and transitional care.¹⁵ In future work, it would be important to better understand the drivers behind decisions around comfort care for patients hospitalized with HF.

With regards to the policy implications of our study, we found that on average, adjusting 30-day mortality rates for early comfort care was not associated with a change in hospital mortality rankings. For those hospitals with high comfort care rates, adjusting for comfort care did lower mortality rates, but the change was so small as to be clinically insignificant. CMS’ RSMR for HF excludes patients enrolled in hospice during the 12 months prior to index admission, including the first day of the index admission, acknowledging that death may not be an untoward outcome for such patients.¹⁶ Fee-for-service Medicare beneficiaries excluded for hospice enrollment comprised 1.29% of HF admissions from July 2012 to June 2015¹⁶ and are likely a subset of early comfort care patients in our sample, both because of the inclusiveness of chart review (vs claims-based identification) and because we defined early comfort care as comfort care initiated on day 0 or 1 of hospitalization. Nevertheless, with our data we cannot assess to what degree our findings were due solely to hospice patients excluded from CMS’ current estimates.

Prior research has described the underuse of palliative care among patients with HF¹⁷ and the association of palliative care with better patient and family experiences at the end of life.^18-20 We add to this literature by describing the epidemiology—prevalence, changes over time, and associated factors—of early comfort care for HF in a national sample of hospitals. This serves as a baseline for future work on end-of-life care among patients hospitalized for HF. Our findings also contribute to ongoing discussion about how best to risk-adjust mortality metrics used to assess hospital quality in pay-for-performance programs. Recent research on stroke and pneumonia based on California data suggests that not accounting for do-not-resuscitate (DNR) status biases hospital mortality rates.^21,22 Earlier research examined the impact of adjusting hospital mortality rates for DNR for a broader range of conditions.^23,24 We expand this line of inquiry by examining the hospital-level association of early comfort care with mortality rates for HF, utilizing a national, contemporary cohort of inpatient stays. In addition, while studies have found that DNR rates within the first 24 hours of admission are relatively high (median 15.8% for pneumonia; 13.3% for stroke),^21,22 comfort care is distinct from DNR.

Our findings should be interpreted in the context of several potential limitations. First, we did not have any information about patient or family wishes regarding end-of-life care, or the exact timing of early comfort care (eg, day 0 or day 1). The initiation of comfort care usually follows conversations about end-of-life care involving a patient, his or her family, and the medical team. Thus, we do not know if low early comfort care rates represent the lack of such a conversation (and thus poor-quality care) or the desire by most patients not to initiate early comfort care (and thus high-quality care). This would be an important area for future research. Second, we included only patients admitted to hospitals that participate in GWTG-HF, a voluntary quality improvement initiative. This may limit the generalizability of our findings, but it is unclear how our sample might bias our findings. Hospitals engaged in quality improvement may be more likely to initiate early comfort care aligned with patients’ wishes; on the other hand, hospitals with advanced surgical capabilities are over-represented in our sample and these hospitals are less likely to initiate early comfort care. Third, we examined associations and cannot make conclusions about causality. Residual measured and unmeasured confounding may influence these findings.

In summary, we found that early comfort care rates for fee-for-service Medicare beneficiaries admitted for HF varies widely among hospitals, but median rates of early comfort care have not changed over time. On average, there was no correlation between hospital-level, 30-day, RSMRs and rates of early comfort care. This suggests that current efforts to lower mortality rates have not had unintended consequences for hospitals that institute early comfort care more commonly than their peers.

Acknowledgments

Dr. Chen and Ms. Cox take responsibility for the integrity of the data and the accuracy of the data analysis. Drs. Chen, Levine, and Hayward are responsible for the study concept and design. Drs. Chen and Fonarow acquired the data. Dr. Chen drafted the manuscript. Drs. Chen, Levin, Hayward, Cox, Fonarow, DeVore, Hernandez, Heidenreich, and Yancy revised the manuscript for important intellectual content. Drs. Chen, Hayward, Cox, and Schulte performed the statistical analysis. Drs. Chen and Fonarow obtained funding for the study. Drs. Hayward and Fonarow supervised the study. The authors thank Bailey Green, MPH, for the research assistance she provided. She was compensated for her work.

Disclosure

Dr. Fonarow reports research support from the National Institutes of Health, and consulting for Amgen, Janssen, Novartis, Medtronic, and St Jude Medical. Dr. DeVore reports research support from the American Heart Association, Amgen, and Novartis, and consulting for Amgen. The other authors have no relevant conflicts of interest. Dr. Chen was supported by a Career Development Grant Award (K08HS020671) from the Agency for Healthcare Research and Quality when the manuscript was being prepared. She currently receives support from the Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation for her work there. She also receives support from the Blue Cross Blue Shield of Michigan Foundation’s Investigator Initiated Research Program, the Agency for Healthcare Research and Quality (R01 HS024698), and the National Institute on Aging (P01 AG019783). These funding sources had no role in the preparation, review, or approval of the manuscript. The GWTG-HF program is provided by the American Heart Association. GWTG-HF has been funded in the past through support from Amgen, Medtronic, GlaxoSmithKline, Ortho-McNeil, and the American Heart Association Pharmaceutical Roundtable. These sponsors had no role in the study design, data analysis or manuscript preparation and revision.

Bacterial bloodstream infections (BSIs) are a major cause of morbidity and mortality in the United States. Approximately 600,000 BSI cases occur annually, resulting in 85,000 deaths,¹ at a cost exceeding $1 billion.² Traditionally, BSIs have been managed with intravenous antimicrobials, which rapidly achieve therapeutic blood concentrations, and are viewed as more potent than oral alternatives. Indeed, for acutely ill patients with bacteremia and sepsis, timely intravenous antimicrobials are lifesaving.³

However, whether intravenous antimicrobials are essential for the entire treatment course in BSIs, particularly for uncomplicated episodes, is controversial. Patients that are clinically stable or have been stabilized after an initial septic presentation may be appropriate candidates for treatment with oral antimicrobials. There are costs and risks associated with extended courses of intravenous agents, such as the necessity for long-term intravenous catheters, which entail risks for procedural complications, secondary infections, and thrombosis. A prospective study of 192 peripherally inserted central catheter (PICC) episodes reported an overall complication rate of 30.2%, including central line-associated BSIs (CLABSI) or venous thrombosis.⁴ Other studies also identified high rates of thrombosis⁵ and PICC-related CLABSI, particularly in patients with malignancy, where sepsis-related complications approach 25%.⁶ Additionally, appropriate care of indwelling catheters requires significant financial and healthcare resources.

Oral antimicrobial therapy for bacterial BSIs offers several potential benefits. Direct economic and healthcare workforce savings are expected to be significant, and procedural and catheter-related complications would be eliminated.⁷ Moreover, oral therapy provides antimicrobial stewardship by reducing the use of broad-spectrum intravenous agents.⁸ Recent infectious disease “Choosing Wisely” initiatives recommend clinicians “prefer oral formulations of highly bioavailable antimicrobials whenever possible”,⁹ and this approach is supported by the Centers for Disease Control and Prevention antibiotic stewardship program.¹⁰ However, the expected savings and benefits of oral therapy would be lost should they be less effective and result in treatment failure or relapse of the primary BSI. Pathogen susceptibility, gastrointestinal absorption, oral bioavailability, patient tolerability, and adherence with therapy need to be carefully considered before choosing oral antimicrobials. Thus, oral antimicrobial therapy for BSI should be utilized in carefully selected circumstances.

In this narrative review, we highlight areas where oral therapy is safe and effective in treating bloodstream infections, as well as offer guidance to clinicians managing patients experiencing BSI. Given the lack of robust clinical trials on this subject, the evidence for performing a systematic review was insufficient. Thus, the articles and recommendations cited in this review were selected based on the authors’ experiences to represent the best available evidence.

Pathogen and Antimicrobial Factors

Evidence Regarding Bloodstream Infections due to Gram-Negative Rods

BSIs due to gram-negative rods (GNRs) are common and cause significant morbidity and mortality. GNRs represent a broad and diverse array of pathogens. We focus on the Enterobacteriaceae family and Pseudomonas aeruginosa, because they are frequently encountered in clinical practice.¹

Gram-Negative Rods, Enterobacteriaceae Family

The Enterobacteriaceae family includes Escherichia coli, Klebsiella, Salmonella, Proteus, Enterobacter, Serratia, and Citrobacter species. The range of illnesses caused by Enterobacteriaceae is as diverse as the family, encompassing most body sites. Although most Enterobacteriaceae are intrinsically susceptible to antibiotics, there is potential for significant multi-drug resistance. Of particular recent concern has been the emergence of Enterobacteriaceae that produce extended-spectrum β-lactamases (ESBL) and even carbapenem-resistant strains.¹⁴

However, Enterobacteriaceae species susceptible to oral antimicrobials are often suitable candidates for oral BSI therapy. Among 106 patients with GNR BSI treated with a highly bioavailable oral antibiotic (eg, levofloxacin), the treatment failure rate was only 2% (versus 14% when an antimicrobial with only moderate or low bioavailability was selected).¹⁵ Oral treatment of Enterobacteriaceae BSIs secondary to urinary tract infection has been best studied. A prospective randomized, controlled trial evaluated oral versus intravenous ciprofloxacin amongst 141 patients with severe pyelonephritis or complicated urinary tract infections, in which the rate of bacteremia was 38%.¹⁶ Notably, patients with obstruction or renal abscess were excluded from the trial. No significant differences in microbiological failure or unsatisfactory clinical responses were found between the IV and oral treatment groups. Additionally, a Cochrane review reported that oral antibiotic therapy is no less effective than intravenous therapy for severe UTI, although data on BSI frequency were not provided.¹⁷

Resistance to fluoroquinolones such as ciprofloxacin has been identified as a risk factor for GNR BSI oral treatment failure, highlighting the importance of confirming susceptibilities before committing to an oral treatment plan.^18,19 Even ESBL Enterobacteriaceae can be considered for treatment with fluoroquinolones if susceptibilities allow.²⁰

The ideal duration of therapy for GNR BSI is an area of active research. A recent retrospective trial showed no difference in all-cause mortality or recurrent BSI in GNR BSI treated for 8 versus 15 days.²¹ A recent meta-analysis suggested that 7 days of therapy was noninferior to a longer duration therapy (10–14 days) for pyelonephritis, in which a subset was bacteremic.²² However, another trial reported that short course therapy for GNR BSI (<7 days) is associated with higher risk of treatment failure.²² Further data are needed.

Gram-Negative Rods, Pseudomonas aeruginosa

Pseudomonas aeruginosa is a common pathogen, intrinsically resistant to many antimicrobials, and readily develops antimicrobial resistance during therapy. Fluoroquinolones (such as ciprofloxacin, levofloxacin, and delafloxacin) are the only currently available oral agents with antipseudomonal activity. However, fluoroquinolones may not achieve blood concentrations appropriate for P. aeruginosa treatment at standard doses, while higher dose regimens may be associated with increased risk for undesirable side effects.^24,25 Currently, given the minimal trial data comparing oral versus intravenous therapy for P. aeruginosa BSIs, and multiple studies indicating increased mortality when P. aeruginosa is treated inappropriately,^26,27 we prefer a conservative approach and consider oral therapy a less-preferred option.

The majority of bloodstream infections in the United States, and the resultant morbidity and mortality, are from gram-positive cocci (GPCs) such as Staphylococcus, Streptococcus, and Enterococcus species.¹

Gram-Positive Cocci, Streptococcus pneumoniae

Of the approximately 900,000 annual cases of S. pneumoniae infection in the United States, approximately 40,000 are complicated by BSI, with 70% of those cases being secondary to pneumococcal pneumonia.²⁸ In studies on patients with pneumococcal pneumonia, bacteremic cases generally fare worse than those without bacteremia.^29,30 However, several trials demonstrated comparable outcomes in the setting of bacteremic pneumococcal pneumonia when switched early (within 3 days) from intravenous to oral antibiotics to complete a 7-day course.^31,32 Before pneumococcal penicillin resistance became widespread, oral penicillin was shown to be effective, and remains an option for susceptible strains.³³ It is worth noting, however, that other trials have shown a mortality benefit from treating bacteremic pneumococcal pneumonia initially with dual-therapy including a β-lactam and macrolide such as azithromycin. This observation highlights the importance of knowing the final susceptibility data prior to consolidating to monotherapy with an oral agent, and that macrolides may have beneficial anti-inflammatory effects, though further research is needed.^34,35

Although the evidence for treating bacteremic pneumococcal pneumonia using a highly active and absorbable oral agent is fairly robust, S. pneumoniae BSI secondary to other sites of infection sites is less well studied and may require a more conservative approach.

Gram-Positive Cocci, β-hemolytic Streptococcus species

β-Hemolytic Streptococci include groups A to H, of which groups A (S. pyogenes) and B (S. agalactiae) are the most commonly implicated in BSIs.³⁶ Group A Streptococcus (GAS) is classically associated with streptococcal pharyngitis and Group B Streptococcus (GBS) is associated with postpartum endometritis and neonatal meningitis, though both are virulent organisms with a potential to cause invasive infection throughout the body and in all age-groups. Up to 14% of GAS and 41% GBS BSIs have no clear source;^37,38 given these are skin pathogens, such scenarios likely represent invasion via microabrasion. As β-hemolytic streptococcal BSI is often observed in the context of necrotizing skin and soft tissue infections, surgical source control is particularly important.³⁹ GAS remains exquisitely susceptible to penicillin, and intravenous penicillin remains the mainstay for invasive disease; GBS has higher penicillin resistance rates than GAS.⁴⁰ Clindamycin should be added when there is concern for severe disease or toxic shock.⁴¹ Unfortunately, oral penicillin is poorly bioavailable (approximately 50%), and there has been recent concern regarding inducible clindamycin resistance in GAS.⁴² Thus, oral penicillin V and/or clindamycin is a potentially risky strategy, with no clinical trials supporting this approach; however, they may be reasonable options in selected patients with source control and stable hemodynamics. Amoxicillin has high bioavailability (85%) and may be effective; however, there is lack of supporting data. Highly bioavailable agents such as levofloxacin and linezolid have GAS and GBS activity⁴³ and might be expected to produce satisfactory outcomes. Because no clinical trials have compared these agents with intravenous therapy for BSI, caution is advised. Although bacteriostatic against Staphylococcus, linezolid is bactericidal against Streptococcus.⁴⁴ Fluoroquinolone resistance amongst β-hemolytic Streptococcus is rare (approximately 0.5%) but does occur.⁴⁵

Gram-Positive Cocci, Staphylococcus Species

Staphylococcus species include S. aureus (including methicillin susceptible and resistant strains: MSSA and MRSA, respectively) and coagulase-negative species, which include organisms such as S. epidermidis. S. aureus is the most common cause of BSI mortality in the United States,¹ with mortality rates estimated at 20%–40% per episode.⁴⁶ Infectious disease consultation has been associated with decreased mortality and is recommended.⁴⁷ The guidelines of the Infectious Diseases Society of America for the treatment of MRSA recommend the use of parenteral agents for BSI.⁴⁸ It is important to consider if a patient with S. aureus BSI has infective endocarditis.

Oral antibiotic therapy for S. aureus BSI is not currently standard practice. Although trimethoprim-sulfamethoxazole (TMP-SMX) has favorable pharmacokinetics and case series of using it successfully for BSI exist,⁴⁹ TMP-SMX showed inferior outcomes in a randomized trial comparing oral TMP-SMX with intravenous vancomycin in a series of 101 S. aureus infections.⁵⁰ This observation has been replicated.⁵¹ Data on doxycycline or clindamycin for S. aureus BSI are limited, and IDSA guidelines advise against their use in this setting because they are predominantly bacteriostatic.⁴⁸ Linezolid has favorable pharmacokinetics, with approximately 100% bioavailability, and S. aureus resistance to linezolid is rare.⁵² Several randomized trials have compared oral linezolid with intravenous vancomycin for S. aureus BSI; for instance, Stevens et al. randomized 460 patients with S. aureus infection (of whom 18% had BSI) to linezolid versus vancomycin and observed similar clinical cure rates.⁵³ A pooled analysis showed oral linezolid was noninferior to vancomycin specifically for S. aureus BSI.⁵⁴ However, long-term use is often limited by hematologic toxicity, peripheral or optic neuropathy (which can be permanent), and induced serotonin syndrome. Additionally, linezolid is bacteriostatic, not bactericidal against S. aureus. Using oral linezolid as a first-line option for S. aureus BSI would not be recommended; however, it may be used as a second-line treatment option in selected cases. Tedizolid has similar pharmacokinetics and spectrum of activity with fewer side effects; however, clinical data on its use for S. aureus BSI are lacking.⁵⁵ Fluoroquinolones such as levofloxacin and the newer agent delafloxacin have activity against S. aureus, including MRSA, but on-treatment emergence of fluoroquinolone resistance is a concern, and data on delafloxacin for BSI are lacking.^56,57 Older literature suggested the combination of ciprofloxacin and rifampin was effective against right-sided S. aureus endocarditis,⁵⁸ and other oral fluoroquinolone-rifamycin combinations have also been found to be effective⁵⁹ However, this approach is currently not a standard therapy, nor is it widely used. Decisions on the duration of therapy for S. aureus BSI should be made in conjunction with an infectious diseases specialist; 14 days is currently regarded as a minimum.^47,48

Published data regarding oral treatment of coagulase-negative Staphylococcus (CoNS) BSI are limited. Most CoNS bacteremia and up to 80% Staphylococcus epidermidis bacteremia represent blood culture contamination, though true infection from CoNS is not uncommon, particularly in patients with indwelling catheters.⁶⁰ An exception is the CoNS species Staphylococcus lugdunensis, which is more virulent, and bacteremia with this organism usually warrants antibiotics. Oral antimicrobial therapy is currently not a standard treatment practice for CoNS BSI that is felt to represent true infection; however, linezolid has been successfully used in case series.⁶¹

Gram-Positive Cocci, Enterococcus

E. faecium and E. faecalis are commonly implicated in BSI.¹ Similar to S. aureus, infective endocarditis must be ruled out when treating enterococcus BSI; a scoring system has been proposed to assist in deciding if such patients require echocardiography.⁶² Intravenous ampicillin is a preferred, highly effective agent for enterococci treatment when the organism is susceptible.⁴⁴ However, oral ampicillin has poor bioavailability (50%), and data for its use in BSI are lacking. For susceptible strains, amoxicillin has comparable efficacy for enterococci and enhanced bioavailability (85%); high dose oral amoxicillin could be considered, but there is minimal clinical trial data to support this approach. Fluoroquinolones exhibit only modest activity against enterococci and would be an inferior choice for BSI.⁶³ Although often sensitive to oral tetracyclines, data on their use in enterococcal BSI are insufficient. Nitrofurantoin can be used for susceptible enterococcal urinary tract infection; however, it does not achieve high blood concentrations and should not be used for BSI.

There is significant data comparing oral linezolid with intravenous daptomycin for vancomycin-resistant enterococci (VRE) BSI. In a systematic review including 10 trials using 30-day all-cause mortality as the primary outcome, patients treated with daptomycin demonstrated higher odds of death (OR 1.61, 95% CI 1.08–2.40) compared with those treated with linezolid.⁶⁴ However, more recent data suggested that higher daptomycin doses than those used in these earlier trials resulted in improved VRE BSI outcomes.⁶⁵ A subsequent study reported that VRE BSI treatment with linezolid is associated with significantly higher treatment failure and mortality compared with daptomycin therapy.⁶⁶ Further research is needed, but should the side-effect profile of linezolid be tolerable, it remains an effective option for oral treatment of enterococcal BSIs.

Evidence Regarding Anaerobic Bacterial Blood Stream Infection

Anaerobic bacteria include Bacteroides, Prevotella, Porphyromonas, Fusobacterium, Peptostreptococcus, Veillonella, and Clostridium. Anaerobes account for approximately 4% of bacterial BSIs, and are often seen in the context of polymicrobial infection.⁶⁷ Given that anaerobes are difficult to recover, and that antimicrobial resistance testing is more labor intensive, antibiotic therapy choices are often made empirically.⁶⁷ Unfortunately, antibiotic resistance amongst anaerobes is increasing.⁶⁸ However, metronidazole remains highly active against a majority of anaerobes, with only a handful of treatment failures reported,⁶⁹ and has a highly favorable pharmacokinetic profile for oral treatment. Oral metronidazole remains an effective choice for many anaerobic BSIs. Considering the polymicrobial nature of many anaerobic infections, source control is important, and concomitant GNR infection must be ruled out before using metronidazole monotherapy.

Clindamycin has significant anaerobic activity, but Bacteroides resistance has increased significantly in recent years, as high as 26%-44%.⁷⁰ Amoxicillin-clavulanate has good anaerobic coverage, but bioavailability of clavulanate is limited (50%), making it inferior for BSI. Bioavailability is also limited for cephalosporins with anaerobic activity, such as cefuroxime. Moxifloxacin is a fluoroquinolone with some anaerobic coverage and a good oral pharmacokinetic profile, but Bacteroides resistance can be as high as 50%, making it a risky empiric choice.⁶⁸

Bacterial BSIs are common and result in significant morbidity and mortality, with high associated healthcare costs. Although BSIs are traditionally treated with intravenous antimicrobials, many BSIs can be safely and effectively cured using oral antibiotics. When appropriately selected, oral antibiotics offer lower costs, fewer side effects, promote antimicrobial stewardship, and are easier for patients. Ultimately, the decision to use oral versus intravenous antibiotics must consider the characteristics of the pathogen, patient, and drug.

Disclosures

 None of the authors report any conflicts of interest.

Bacterial bloodstream infections (BSIs) are a major cause of morbidity and mortality in the United States. Approximately 600,000 BSI cases occur annually, resulting in 85,000 deaths,¹ at a cost exceeding $1 billion.² Traditionally, BSIs have been managed with intravenous antimicrobials, which rapidly achieve therapeutic blood concentrations, and are viewed as more potent than oral alternatives. Indeed, for acutely ill patients with bacteremia and sepsis, timely intravenous antimicrobials are lifesaving.³

However, whether intravenous antimicrobials are essential for the entire treatment course in BSIs, particularly for uncomplicated episodes, is controversial. Patients that are clinically stable or have been stabilized after an initial septic presentation may be appropriate candidates for treatment with oral antimicrobials. There are costs and risks associated with extended courses of intravenous agents, such as the necessity for long-term intravenous catheters, which entail risks for procedural complications, secondary infections, and thrombosis. A prospective study of 192 peripherally inserted central catheter (PICC) episodes reported an overall complication rate of 30.2%, including central line-associated BSIs (CLABSI) or venous thrombosis.⁴ Other studies also identified high rates of thrombosis⁵ and PICC-related CLABSI, particularly in patients with malignancy, where sepsis-related complications approach 25%.⁶ Additionally, appropriate care of indwelling catheters requires significant financial and healthcare resources.

Oral antimicrobial therapy for bacterial BSIs offers several potential benefits. Direct economic and healthcare workforce savings are expected to be significant, and procedural and catheter-related complications would be eliminated.⁷ Moreover, oral therapy provides antimicrobial stewardship by reducing the use of broad-spectrum intravenous agents.⁸ Recent infectious disease “Choosing Wisely” initiatives recommend clinicians “prefer oral formulations of highly bioavailable antimicrobials whenever possible”,⁹ and this approach is supported by the Centers for Disease Control and Prevention antibiotic stewardship program.¹⁰ However, the expected savings and benefits of oral therapy would be lost should they be less effective and result in treatment failure or relapse of the primary BSI. Pathogen susceptibility, gastrointestinal absorption, oral bioavailability, patient tolerability, and adherence with therapy need to be carefully considered before choosing oral antimicrobials. Thus, oral antimicrobial therapy for BSI should be utilized in carefully selected circumstances.

In this narrative review, we highlight areas where oral therapy is safe and effective in treating bloodstream infections, as well as offer guidance to clinicians managing patients experiencing BSI. Given the lack of robust clinical trials on this subject, the evidence for performing a systematic review was insufficient. Thus, the articles and recommendations cited in this review were selected based on the authors’ experiences to represent the best available evidence.

Pathogen and Antimicrobial Factors

Evidence Regarding Bloodstream Infections due to Gram-Negative Rods

BSIs due to gram-negative rods (GNRs) are common and cause significant morbidity and mortality. GNRs represent a broad and diverse array of pathogens. We focus on the Enterobacteriaceae family and Pseudomonas aeruginosa, because they are frequently encountered in clinical practice.¹

Gram-Negative Rods, Enterobacteriaceae Family

The Enterobacteriaceae family includes Escherichia coli, Klebsiella, Salmonella, Proteus, Enterobacter, Serratia, and Citrobacter species. The range of illnesses caused by Enterobacteriaceae is as diverse as the family, encompassing most body sites. Although most Enterobacteriaceae are intrinsically susceptible to antibiotics, there is potential for significant multi-drug resistance. Of particular recent concern has been the emergence of Enterobacteriaceae that produce extended-spectrum β-lactamases (ESBL) and even carbapenem-resistant strains.¹⁴

However, Enterobacteriaceae species susceptible to oral antimicrobials are often suitable candidates for oral BSI therapy. Among 106 patients with GNR BSI treated with a highly bioavailable oral antibiotic (eg, levofloxacin), the treatment failure rate was only 2% (versus 14% when an antimicrobial with only moderate or low bioavailability was selected).¹⁵ Oral treatment of Enterobacteriaceae BSIs secondary to urinary tract infection has been best studied. A prospective randomized, controlled trial evaluated oral versus intravenous ciprofloxacin amongst 141 patients with severe pyelonephritis or complicated urinary tract infections, in which the rate of bacteremia was 38%.¹⁶ Notably, patients with obstruction or renal abscess were excluded from the trial. No significant differences in microbiological failure or unsatisfactory clinical responses were found between the IV and oral treatment groups. Additionally, a Cochrane review reported that oral antibiotic therapy is no less effective than intravenous therapy for severe UTI, although data on BSI frequency were not provided.¹⁷

Resistance to fluoroquinolones such as ciprofloxacin has been identified as a risk factor for GNR BSI oral treatment failure, highlighting the importance of confirming susceptibilities before committing to an oral treatment plan.^18,19 Even ESBL Enterobacteriaceae can be considered for treatment with fluoroquinolones if susceptibilities allow.²⁰

The ideal duration of therapy for GNR BSI is an area of active research. A recent retrospective trial showed no difference in all-cause mortality or recurrent BSI in GNR BSI treated for 8 versus 15 days.²¹ A recent meta-analysis suggested that 7 days of therapy was noninferior to a longer duration therapy (10–14 days) for pyelonephritis, in which a subset was bacteremic.²² However, another trial reported that short course therapy for GNR BSI (<7 days) is associated with higher risk of treatment failure.²² Further data are needed.

Gram-Negative Rods, Pseudomonas aeruginosa

Pseudomonas aeruginosa is a common pathogen, intrinsically resistant to many antimicrobials, and readily develops antimicrobial resistance during therapy. Fluoroquinolones (such as ciprofloxacin, levofloxacin, and delafloxacin) are the only currently available oral agents with antipseudomonal activity. However, fluoroquinolones may not achieve blood concentrations appropriate for P. aeruginosa treatment at standard doses, while higher dose regimens may be associated with increased risk for undesirable side effects.^24,25 Currently, given the minimal trial data comparing oral versus intravenous therapy for P. aeruginosa BSIs, and multiple studies indicating increased mortality when P. aeruginosa is treated inappropriately,^26,27 we prefer a conservative approach and consider oral therapy a less-preferred option.

The majority of bloodstream infections in the United States, and the resultant morbidity and mortality, are from gram-positive cocci (GPCs) such as Staphylococcus, Streptococcus, and Enterococcus species.¹

Gram-Positive Cocci, Streptococcus pneumoniae

Of the approximately 900,000 annual cases of S. pneumoniae infection in the United States, approximately 40,000 are complicated by BSI, with 70% of those cases being secondary to pneumococcal pneumonia.²⁸ In studies on patients with pneumococcal pneumonia, bacteremic cases generally fare worse than those without bacteremia.^29,30 However, several trials demonstrated comparable outcomes in the setting of bacteremic pneumococcal pneumonia when switched early (within 3 days) from intravenous to oral antibiotics to complete a 7-day course.^31,32 Before pneumococcal penicillin resistance became widespread, oral penicillin was shown to be effective, and remains an option for susceptible strains.³³ It is worth noting, however, that other trials have shown a mortality benefit from treating bacteremic pneumococcal pneumonia initially with dual-therapy including a β-lactam and macrolide such as azithromycin. This observation highlights the importance of knowing the final susceptibility data prior to consolidating to monotherapy with an oral agent, and that macrolides may have beneficial anti-inflammatory effects, though further research is needed.^34,35

Although the evidence for treating bacteremic pneumococcal pneumonia using a highly active and absorbable oral agent is fairly robust, S. pneumoniae BSI secondary to other sites of infection sites is less well studied and may require a more conservative approach.

Gram-Positive Cocci, β-hemolytic Streptococcus species

β-Hemolytic Streptococci include groups A to H, of which groups A (S. pyogenes) and B (S. agalactiae) are the most commonly implicated in BSIs.³⁶ Group A Streptococcus (GAS) is classically associated with streptococcal pharyngitis and Group B Streptococcus (GBS) is associated with postpartum endometritis and neonatal meningitis, though both are virulent organisms with a potential to cause invasive infection throughout the body and in all age-groups. Up to 14% of GAS and 41% GBS BSIs have no clear source;^37,38 given these are skin pathogens, such scenarios likely represent invasion via microabrasion. As β-hemolytic streptococcal BSI is often observed in the context of necrotizing skin and soft tissue infections, surgical source control is particularly important.³⁹ GAS remains exquisitely susceptible to penicillin, and intravenous penicillin remains the mainstay for invasive disease; GBS has higher penicillin resistance rates than GAS.⁴⁰ Clindamycin should be added when there is concern for severe disease or toxic shock.⁴¹ Unfortunately, oral penicillin is poorly bioavailable (approximately 50%), and there has been recent concern regarding inducible clindamycin resistance in GAS.⁴² Thus, oral penicillin V and/or clindamycin is a potentially risky strategy, with no clinical trials supporting this approach; however, they may be reasonable options in selected patients with source control and stable hemodynamics. Amoxicillin has high bioavailability (85%) and may be effective; however, there is lack of supporting data. Highly bioavailable agents such as levofloxacin and linezolid have GAS and GBS activity⁴³ and might be expected to produce satisfactory outcomes. Because no clinical trials have compared these agents with intravenous therapy for BSI, caution is advised. Although bacteriostatic against Staphylococcus, linezolid is bactericidal against Streptococcus.⁴⁴ Fluoroquinolone resistance amongst β-hemolytic Streptococcus is rare (approximately 0.5%) but does occur.⁴⁵

Gram-Positive Cocci, Staphylococcus Species

Staphylococcus species include S. aureus (including methicillin susceptible and resistant strains: MSSA and MRSA, respectively) and coagulase-negative species, which include organisms such as S. epidermidis. S. aureus is the most common cause of BSI mortality in the United States,¹ with mortality rates estimated at 20%–40% per episode.⁴⁶ Infectious disease consultation has been associated with decreased mortality and is recommended.⁴⁷ The guidelines of the Infectious Diseases Society of America for the treatment of MRSA recommend the use of parenteral agents for BSI.⁴⁸ It is important to consider if a patient with S. aureus BSI has infective endocarditis.

Oral antibiotic therapy for S. aureus BSI is not currently standard practice. Although trimethoprim-sulfamethoxazole (TMP-SMX) has favorable pharmacokinetics and case series of using it successfully for BSI exist,⁴⁹ TMP-SMX showed inferior outcomes in a randomized trial comparing oral TMP-SMX with intravenous vancomycin in a series of 101 S. aureus infections.⁵⁰ This observation has been replicated.⁵¹ Data on doxycycline or clindamycin for S. aureus BSI are limited, and IDSA guidelines advise against their use in this setting because they are predominantly bacteriostatic.⁴⁸ Linezolid has favorable pharmacokinetics, with approximately 100% bioavailability, and S. aureus resistance to linezolid is rare.⁵² Several randomized trials have compared oral linezolid with intravenous vancomycin for S. aureus BSI; for instance, Stevens et al. randomized 460 patients with S. aureus infection (of whom 18% had BSI) to linezolid versus vancomycin and observed similar clinical cure rates.⁵³ A pooled analysis showed oral linezolid was noninferior to vancomycin specifically for S. aureus BSI.⁵⁴ However, long-term use is often limited by hematologic toxicity, peripheral or optic neuropathy (which can be permanent), and induced serotonin syndrome. Additionally, linezolid is bacteriostatic, not bactericidal against S. aureus. Using oral linezolid as a first-line option for S. aureus BSI would not be recommended; however, it may be used as a second-line treatment option in selected cases. Tedizolid has similar pharmacokinetics and spectrum of activity with fewer side effects; however, clinical data on its use for S. aureus BSI are lacking.⁵⁵ Fluoroquinolones such as levofloxacin and the newer agent delafloxacin have activity against S. aureus, including MRSA, but on-treatment emergence of fluoroquinolone resistance is a concern, and data on delafloxacin for BSI are lacking.^56,57 Older literature suggested the combination of ciprofloxacin and rifampin was effective against right-sided S. aureus endocarditis,⁵⁸ and other oral fluoroquinolone-rifamycin combinations have also been found to be effective⁵⁹ However, this approach is currently not a standard therapy, nor is it widely used. Decisions on the duration of therapy for S. aureus BSI should be made in conjunction with an infectious diseases specialist; 14 days is currently regarded as a minimum.^47,48

Published data regarding oral treatment of coagulase-negative Staphylococcus (CoNS) BSI are limited. Most CoNS bacteremia and up to 80% Staphylococcus epidermidis bacteremia represent blood culture contamination, though true infection from CoNS is not uncommon, particularly in patients with indwelling catheters.⁶⁰ An exception is the CoNS species Staphylococcus lugdunensis, which is more virulent, and bacteremia with this organism usually warrants antibiotics. Oral antimicrobial therapy is currently not a standard treatment practice for CoNS BSI that is felt to represent true infection; however, linezolid has been successfully used in case series.⁶¹

Gram-Positive Cocci, Enterococcus

E. faecium and E. faecalis are commonly implicated in BSI.¹ Similar to S. aureus, infective endocarditis must be ruled out when treating enterococcus BSI; a scoring system has been proposed to assist in deciding if such patients require echocardiography.⁶² Intravenous ampicillin is a preferred, highly effective agent for enterococci treatment when the organism is susceptible.⁴⁴ However, oral ampicillin has poor bioavailability (50%), and data for its use in BSI are lacking. For susceptible strains, amoxicillin has comparable efficacy for enterococci and enhanced bioavailability (85%); high dose oral amoxicillin could be considered, but there is minimal clinical trial data to support this approach. Fluoroquinolones exhibit only modest activity against enterococci and would be an inferior choice for BSI.⁶³ Although often sensitive to oral tetracyclines, data on their use in enterococcal BSI are insufficient. Nitrofurantoin can be used for susceptible enterococcal urinary tract infection; however, it does not achieve high blood concentrations and should not be used for BSI.

There is significant data comparing oral linezolid with intravenous daptomycin for vancomycin-resistant enterococci (VRE) BSI. In a systematic review including 10 trials using 30-day all-cause mortality as the primary outcome, patients treated with daptomycin demonstrated higher odds of death (OR 1.61, 95% CI 1.08–2.40) compared with those treated with linezolid.⁶⁴ However, more recent data suggested that higher daptomycin doses than those used in these earlier trials resulted in improved VRE BSI outcomes.⁶⁵ A subsequent study reported that VRE BSI treatment with linezolid is associated with significantly higher treatment failure and mortality compared with daptomycin therapy.⁶⁶ Further research is needed, but should the side-effect profile of linezolid be tolerable, it remains an effective option for oral treatment of enterococcal BSIs.

Evidence Regarding Anaerobic Bacterial Blood Stream Infection

Anaerobic bacteria include Bacteroides, Prevotella, Porphyromonas, Fusobacterium, Peptostreptococcus, Veillonella, and Clostridium. Anaerobes account for approximately 4% of bacterial BSIs, and are often seen in the context of polymicrobial infection.⁶⁷ Given that anaerobes are difficult to recover, and that antimicrobial resistance testing is more labor intensive, antibiotic therapy choices are often made empirically.⁶⁷ Unfortunately, antibiotic resistance amongst anaerobes is increasing.⁶⁸ However, metronidazole remains highly active against a majority of anaerobes, with only a handful of treatment failures reported,⁶⁹ and has a highly favorable pharmacokinetic profile for oral treatment. Oral metronidazole remains an effective choice for many anaerobic BSIs. Considering the polymicrobial nature of many anaerobic infections, source control is important, and concomitant GNR infection must be ruled out before using metronidazole monotherapy.

Clindamycin has significant anaerobic activity, but Bacteroides resistance has increased significantly in recent years, as high as 26%-44%.⁷⁰ Amoxicillin-clavulanate has good anaerobic coverage, but bioavailability of clavulanate is limited (50%), making it inferior for BSI. Bioavailability is also limited for cephalosporins with anaerobic activity, such as cefuroxime. Moxifloxacin is a fluoroquinolone with some anaerobic coverage and a good oral pharmacokinetic profile, but Bacteroides resistance can be as high as 50%, making it a risky empiric choice.⁶⁸

Bacterial BSIs are common and result in significant morbidity and mortality, with high associated healthcare costs. Although BSIs are traditionally treated with intravenous antimicrobials, many BSIs can be safely and effectively cured using oral antibiotics. When appropriately selected, oral antibiotics offer lower costs, fewer side effects, promote antimicrobial stewardship, and are easier for patients. Ultimately, the decision to use oral versus intravenous antibiotics must consider the characteristics of the pathogen, patient, and drug.

Disclosures

 None of the authors report any conflicts of interest.

Bacterial bloodstream infections (BSIs) are a major cause of morbidity and mortality in the United States. Approximately 600,000 BSI cases occur annually, resulting in 85,000 deaths,¹ at a cost exceeding $1 billion.² Traditionally, BSIs have been managed with intravenous antimicrobials, which rapidly achieve therapeutic blood concentrations, and are viewed as more potent than oral alternatives. Indeed, for acutely ill patients with bacteremia and sepsis, timely intravenous antimicrobials are lifesaving.³

However, whether intravenous antimicrobials are essential for the entire treatment course in BSIs, particularly for uncomplicated episodes, is controversial. Patients that are clinically stable or have been stabilized after an initial septic presentation may be appropriate candidates for treatment with oral antimicrobials. There are costs and risks associated with extended courses of intravenous agents, such as the necessity for long-term intravenous catheters, which entail risks for procedural complications, secondary infections, and thrombosis. A prospective study of 192 peripherally inserted central catheter (PICC) episodes reported an overall complication rate of 30.2%, including central line-associated BSIs (CLABSI) or venous thrombosis.⁴ Other studies also identified high rates of thrombosis⁵ and PICC-related CLABSI, particularly in patients with malignancy, where sepsis-related complications approach 25%.⁶ Additionally, appropriate care of indwelling catheters requires significant financial and healthcare resources.

Oral antimicrobial therapy for bacterial BSIs offers several potential benefits. Direct economic and healthcare workforce savings are expected to be significant, and procedural and catheter-related complications would be eliminated.⁷ Moreover, oral therapy provides antimicrobial stewardship by reducing the use of broad-spectrum intravenous agents.⁸ Recent infectious disease “Choosing Wisely” initiatives recommend clinicians “prefer oral formulations of highly bioavailable antimicrobials whenever possible”,⁹ and this approach is supported by the Centers for Disease Control and Prevention antibiotic stewardship program.¹⁰ However, the expected savings and benefits of oral therapy would be lost should they be less effective and result in treatment failure or relapse of the primary BSI. Pathogen susceptibility, gastrointestinal absorption, oral bioavailability, patient tolerability, and adherence with therapy need to be carefully considered before choosing oral antimicrobials. Thus, oral antimicrobial therapy for BSI should be utilized in carefully selected circumstances.

In this narrative review, we highlight areas where oral therapy is safe and effective in treating bloodstream infections, as well as offer guidance to clinicians managing patients experiencing BSI. Given the lack of robust clinical trials on this subject, the evidence for performing a systematic review was insufficient. Thus, the articles and recommendations cited in this review were selected based on the authors’ experiences to represent the best available evidence.

Pathogen and Antimicrobial Factors

Evidence Regarding Bloodstream Infections due to Gram-Negative Rods

BSIs due to gram-negative rods (GNRs) are common and cause significant morbidity and mortality. GNRs represent a broad and diverse array of pathogens. We focus on the Enterobacteriaceae family and Pseudomonas aeruginosa, because they are frequently encountered in clinical practice.¹

Gram-Negative Rods, Enterobacteriaceae Family

The Enterobacteriaceae family includes Escherichia coli, Klebsiella, Salmonella, Proteus, Enterobacter, Serratia, and Citrobacter species. The range of illnesses caused by Enterobacteriaceae is as diverse as the family, encompassing most body sites. Although most Enterobacteriaceae are intrinsically susceptible to antibiotics, there is potential for significant multi-drug resistance. Of particular recent concern has been the emergence of Enterobacteriaceae that produce extended-spectrum β-lactamases (ESBL) and even carbapenem-resistant strains.¹⁴

However, Enterobacteriaceae species susceptible to oral antimicrobials are often suitable candidates for oral BSI therapy. Among 106 patients with GNR BSI treated with a highly bioavailable oral antibiotic (eg, levofloxacin), the treatment failure rate was only 2% (versus 14% when an antimicrobial with only moderate or low bioavailability was selected).¹⁵ Oral treatment of Enterobacteriaceae BSIs secondary to urinary tract infection has been best studied. A prospective randomized, controlled trial evaluated oral versus intravenous ciprofloxacin amongst 141 patients with severe pyelonephritis or complicated urinary tract infections, in which the rate of bacteremia was 38%.¹⁶ Notably, patients with obstruction or renal abscess were excluded from the trial. No significant differences in microbiological failure or unsatisfactory clinical responses were found between the IV and oral treatment groups. Additionally, a Cochrane review reported that oral antibiotic therapy is no less effective than intravenous therapy for severe UTI, although data on BSI frequency were not provided.¹⁷

Resistance to fluoroquinolones such as ciprofloxacin has been identified as a risk factor for GNR BSI oral treatment failure, highlighting the importance of confirming susceptibilities before committing to an oral treatment plan.^18,19 Even ESBL Enterobacteriaceae can be considered for treatment with fluoroquinolones if susceptibilities allow.²⁰

The ideal duration of therapy for GNR BSI is an area of active research. A recent retrospective trial showed no difference in all-cause mortality or recurrent BSI in GNR BSI treated for 8 versus 15 days.²¹ A recent meta-analysis suggested that 7 days of therapy was noninferior to a longer duration therapy (10–14 days) for pyelonephritis, in which a subset was bacteremic.²² However, another trial reported that short course therapy for GNR BSI (<7 days) is associated with higher risk of treatment failure.²² Further data are needed.

Gram-Negative Rods, Pseudomonas aeruginosa

Pseudomonas aeruginosa is a common pathogen, intrinsically resistant to many antimicrobials, and readily develops antimicrobial resistance during therapy. Fluoroquinolones (such as ciprofloxacin, levofloxacin, and delafloxacin) are the only currently available oral agents with antipseudomonal activity. However, fluoroquinolones may not achieve blood concentrations appropriate for P. aeruginosa treatment at standard doses, while higher dose regimens may be associated with increased risk for undesirable side effects.^24,25 Currently, given the minimal trial data comparing oral versus intravenous therapy for P. aeruginosa BSIs, and multiple studies indicating increased mortality when P. aeruginosa is treated inappropriately,^26,27 we prefer a conservative approach and consider oral therapy a less-preferred option.

The majority of bloodstream infections in the United States, and the resultant morbidity and mortality, are from gram-positive cocci (GPCs) such as Staphylococcus, Streptococcus, and Enterococcus species.¹

Gram-Positive Cocci, Streptococcus pneumoniae

Of the approximately 900,000 annual cases of S. pneumoniae infection in the United States, approximately 40,000 are complicated by BSI, with 70% of those cases being secondary to pneumococcal pneumonia.²⁸ In studies on patients with pneumococcal pneumonia, bacteremic cases generally fare worse than those without bacteremia.^29,30 However, several trials demonstrated comparable outcomes in the setting of bacteremic pneumococcal pneumonia when switched early (within 3 days) from intravenous to oral antibiotics to complete a 7-day course.^31,32 Before pneumococcal penicillin resistance became widespread, oral penicillin was shown to be effective, and remains an option for susceptible strains.³³ It is worth noting, however, that other trials have shown a mortality benefit from treating bacteremic pneumococcal pneumonia initially with dual-therapy including a β-lactam and macrolide such as azithromycin. This observation highlights the importance of knowing the final susceptibility data prior to consolidating to monotherapy with an oral agent, and that macrolides may have beneficial anti-inflammatory effects, though further research is needed.^34,35

Although the evidence for treating bacteremic pneumococcal pneumonia using a highly active and absorbable oral agent is fairly robust, S. pneumoniae BSI secondary to other sites of infection sites is less well studied and may require a more conservative approach.

Gram-Positive Cocci, β-hemolytic Streptococcus species

β-Hemolytic Streptococci include groups A to H, of which groups A (S. pyogenes) and B (S. agalactiae) are the most commonly implicated in BSIs.³⁶ Group A Streptococcus (GAS) is classically associated with streptococcal pharyngitis and Group B Streptococcus (GBS) is associated with postpartum endometritis and neonatal meningitis, though both are virulent organisms with a potential to cause invasive infection throughout the body and in all age-groups. Up to 14% of GAS and 41% GBS BSIs have no clear source;^37,38 given these are skin pathogens, such scenarios likely represent invasion via microabrasion. As β-hemolytic streptococcal BSI is often observed in the context of necrotizing skin and soft tissue infections, surgical source control is particularly important.³⁹ GAS remains exquisitely susceptible to penicillin, and intravenous penicillin remains the mainstay for invasive disease; GBS has higher penicillin resistance rates than GAS.⁴⁰ Clindamycin should be added when there is concern for severe disease or toxic shock.⁴¹ Unfortunately, oral penicillin is poorly bioavailable (approximately 50%), and there has been recent concern regarding inducible clindamycin resistance in GAS.⁴² Thus, oral penicillin V and/or clindamycin is a potentially risky strategy, with no clinical trials supporting this approach; however, they may be reasonable options in selected patients with source control and stable hemodynamics. Amoxicillin has high bioavailability (85%) and may be effective; however, there is lack of supporting data. Highly bioavailable agents such as levofloxacin and linezolid have GAS and GBS activity⁴³ and might be expected to produce satisfactory outcomes. Because no clinical trials have compared these agents with intravenous therapy for BSI, caution is advised. Although bacteriostatic against Staphylococcus, linezolid is bactericidal against Streptococcus.⁴⁴ Fluoroquinolone resistance amongst β-hemolytic Streptococcus is rare (approximately 0.5%) but does occur.⁴⁵

Gram-Positive Cocci, Staphylococcus Species

Staphylococcus species include S. aureus (including methicillin susceptible and resistant strains: MSSA and MRSA, respectively) and coagulase-negative species, which include organisms such as S. epidermidis. S. aureus is the most common cause of BSI mortality in the United States,¹ with mortality rates estimated at 20%–40% per episode.⁴⁶ Infectious disease consultation has been associated with decreased mortality and is recommended.⁴⁷ The guidelines of the Infectious Diseases Society of America for the treatment of MRSA recommend the use of parenteral agents for BSI.⁴⁸ It is important to consider if a patient with S. aureus BSI has infective endocarditis.

Oral antibiotic therapy for S. aureus BSI is not currently standard practice. Although trimethoprim-sulfamethoxazole (TMP-SMX) has favorable pharmacokinetics and case series of using it successfully for BSI exist,⁴⁹ TMP-SMX showed inferior outcomes in a randomized trial comparing oral TMP-SMX with intravenous vancomycin in a series of 101 S. aureus infections.⁵⁰ This observation has been replicated.⁵¹ Data on doxycycline or clindamycin for S. aureus BSI are limited, and IDSA guidelines advise against their use in this setting because they are predominantly bacteriostatic.⁴⁸ Linezolid has favorable pharmacokinetics, with approximately 100% bioavailability, and S. aureus resistance to linezolid is rare.⁵² Several randomized trials have compared oral linezolid with intravenous vancomycin for S. aureus BSI; for instance, Stevens et al. randomized 460 patients with S. aureus infection (of whom 18% had BSI) to linezolid versus vancomycin and observed similar clinical cure rates.⁵³ A pooled analysis showed oral linezolid was noninferior to vancomycin specifically for S. aureus BSI.⁵⁴ However, long-term use is often limited by hematologic toxicity, peripheral or optic neuropathy (which can be permanent), and induced serotonin syndrome. Additionally, linezolid is bacteriostatic, not bactericidal against S. aureus. Using oral linezolid as a first-line option for S. aureus BSI would not be recommended; however, it may be used as a second-line treatment option in selected cases. Tedizolid has similar pharmacokinetics and spectrum of activity with fewer side effects; however, clinical data on its use for S. aureus BSI are lacking.⁵⁵ Fluoroquinolones such as levofloxacin and the newer agent delafloxacin have activity against S. aureus, including MRSA, but on-treatment emergence of fluoroquinolone resistance is a concern, and data on delafloxacin for BSI are lacking.^56,57 Older literature suggested the combination of ciprofloxacin and rifampin was effective against right-sided S. aureus endocarditis,⁵⁸ and other oral fluoroquinolone-rifamycin combinations have also been found to be effective⁵⁹ However, this approach is currently not a standard therapy, nor is it widely used. Decisions on the duration of therapy for S. aureus BSI should be made in conjunction with an infectious diseases specialist; 14 days is currently regarded as a minimum.^47,48

Published data regarding oral treatment of coagulase-negative Staphylococcus (CoNS) BSI are limited. Most CoNS bacteremia and up to 80% Staphylococcus epidermidis bacteremia represent blood culture contamination, though true infection from CoNS is not uncommon, particularly in patients with indwelling catheters.⁶⁰ An exception is the CoNS species Staphylococcus lugdunensis, which is more virulent, and bacteremia with this organism usually warrants antibiotics. Oral antimicrobial therapy is currently not a standard treatment practice for CoNS BSI that is felt to represent true infection; however, linezolid has been successfully used in case series.⁶¹

Gram-Positive Cocci, Enterococcus

E. faecium and E. faecalis are commonly implicated in BSI.¹ Similar to S. aureus, infective endocarditis must be ruled out when treating enterococcus BSI; a scoring system has been proposed to assist in deciding if such patients require echocardiography.⁶² Intravenous ampicillin is a preferred, highly effective agent for enterococci treatment when the organism is susceptible.⁴⁴ However, oral ampicillin has poor bioavailability (50%), and data for its use in BSI are lacking. For susceptible strains, amoxicillin has comparable efficacy for enterococci and enhanced bioavailability (85%); high dose oral amoxicillin could be considered, but there is minimal clinical trial data to support this approach. Fluoroquinolones exhibit only modest activity against enterococci and would be an inferior choice for BSI.⁶³ Although often sensitive to oral tetracyclines, data on their use in enterococcal BSI are insufficient. Nitrofurantoin can be used for susceptible enterococcal urinary tract infection; however, it does not achieve high blood concentrations and should not be used for BSI.

There is significant data comparing oral linezolid with intravenous daptomycin for vancomycin-resistant enterococci (VRE) BSI. In a systematic review including 10 trials using 30-day all-cause mortality as the primary outcome, patients treated with daptomycin demonstrated higher odds of death (OR 1.61, 95% CI 1.08–2.40) compared with those treated with linezolid.⁶⁴ However, more recent data suggested that higher daptomycin doses than those used in these earlier trials resulted in improved VRE BSI outcomes.⁶⁵ A subsequent study reported that VRE BSI treatment with linezolid is associated with significantly higher treatment failure and mortality compared with daptomycin therapy.⁶⁶ Further research is needed, but should the side-effect profile of linezolid be tolerable, it remains an effective option for oral treatment of enterococcal BSIs.

Evidence Regarding Anaerobic Bacterial Blood Stream Infection

Anaerobic bacteria include Bacteroides, Prevotella, Porphyromonas, Fusobacterium, Peptostreptococcus, Veillonella, and Clostridium. Anaerobes account for approximately 4% of bacterial BSIs, and are often seen in the context of polymicrobial infection.⁶⁷ Given that anaerobes are difficult to recover, and that antimicrobial resistance testing is more labor intensive, antibiotic therapy choices are often made empirically.⁶⁷ Unfortunately, antibiotic resistance amongst anaerobes is increasing.⁶⁸ However, metronidazole remains highly active against a majority of anaerobes, with only a handful of treatment failures reported,⁶⁹ and has a highly favorable pharmacokinetic profile for oral treatment. Oral metronidazole remains an effective choice for many anaerobic BSIs. Considering the polymicrobial nature of many anaerobic infections, source control is important, and concomitant GNR infection must be ruled out before using metronidazole monotherapy.

Clindamycin has significant anaerobic activity, but Bacteroides resistance has increased significantly in recent years, as high as 26%-44%.⁷⁰ Amoxicillin-clavulanate has good anaerobic coverage, but bioavailability of clavulanate is limited (50%), making it inferior for BSI. Bioavailability is also limited for cephalosporins with anaerobic activity, such as cefuroxime. Moxifloxacin is a fluoroquinolone with some anaerobic coverage and a good oral pharmacokinetic profile, but Bacteroides resistance can be as high as 50%, making it a risky empiric choice.⁶⁸

Bacterial BSIs are common and result in significant morbidity and mortality, with high associated healthcare costs. Although BSIs are traditionally treated with intravenous antimicrobials, many BSIs can be safely and effectively cured using oral antibiotics. When appropriately selected, oral antibiotics offer lower costs, fewer side effects, promote antimicrobial stewardship, and are easier for patients. Ultimately, the decision to use oral versus intravenous antibiotics must consider the characteristics of the pathogen, patient, and drug.

Disclosures

 None of the authors report any conflicts of interest.

Healthcare is a “dirty” business with widespread effects on the environment. In the US, healthcare is estimated to generate 9.8% of our greenhouse gases and 9% of our particulate matter emissions.¹ Hazardous wastes must be incinerated, emitting carbon dioxide, nitrogen oxides, and volatile substances into the atmosphere.² Similarly, hospitals are responsible for 7% of commercial water use in the US.³ Conventional water treatment systems are not designed to remove heavy metals, pharmaceuticals, and disinfectants in hospital wastewaters; these compounds have been detected in rivers and streams throughout the US.^4,5 Furthermore, pharmaceutical compounds such as antibiotics, anti-epileptics, and narcotics have even been isolated in our drinking water.⁵

As hospitalists, we are the directors of inpatient care, yet we only witness brief moments in the lives of our patients and the products we use for their care. For example, we are unaware of particulate matter emissions needed to power an extra imaging study or the contribution of unused materials to a growing landfill. However, pollution, including that from our clinical practice, is detrimental to human health in many ways. Exposure to particulate matter and toxic wastes has been linked to increased rates of reproductive and developmental disorders, cancer, and respiratory disease. ⁶ Particles <2.5 µm in diameter can diffuse through alveoli into the bloodstream, contributing to heart disease, stroke, and lung disease.⁷ Climate change has been linked to a wide range of adverse cardiovascular, respiratory, infectious, and mental health outcomes.^8,9 These examples of the health impacts of pollution are illustrative but not exhaustive.

The environmental impact of US healthcare accounts for an estimated 470,000 disability-adjusted life years lost; this figure is on par with the burden of preventable medical errors.¹ Clearly, change is necessary at all levels in the healthcare system to address our impact on human health. Fortunately, healthcare systems and hospital administrators have begun to address this issue. This perspective describes sustainability efforts in hospitals and healthcare systems and seeks to motivate hospitalists to build upon these efforts.

With the ability to affect change from the top down, health systems are playing an important role in healthcare’s environmental sustainability. Ambitiously, Kaiser Permanente outlined eight environmental stewardship goals, which include becoming net carbon positive and recycling, reusing, or composting 100% of their non-hazardous waste by 2025.¹⁰ The Cleveland Clinic has pledged to become carbon neutral within the next 10 years.¹¹ Other healthcare systems may follow suite. Many “green” interventions aimed at reducing waste and pollution also protect population health and reduce hospital operating costs.

From 2011 to 2015, a group of Boston Hospitals decreased energy use by 9.4% compared with a historical growth of 1.5% per year and saved over 15 million dollars.¹² Similarly, Virginia Mason reduced landfill waste by reprocessing single-use medical devices, thereby decreasing purchasing costs by $3 million.¹³ As part of a regional campaign to protect the St. Croix River, Hudson Hospital and Clinic in Wisconsin saved over $20,000 with new recycling and waste reduction programs.¹³ Notably, these programs not only benefit hospitals but also patients and payers by reducing costs of care.

These examples illustrate that a greener healthcare industry is achievable. Despite the potential benefits, sustainability efforts in US hospitals are the exception, not the rule, and the diffusion of such innovations must be encouraged from within.

In addition to the moral case for environmentally sustainable healthcare,^14,15 such efforts can also improve our quality of care. The conversation around healthcare waste has focused on costs. Yet, examining our waste from a new perspective may reveal new ways to increase the value of patient care while protecting population health. Our communities and families are not immune to the health impacts of pollution, including that generated by our industry. However, predicted effects of climate change including altered patterns of vector-borne disease and frequent hurricanes and forest fires are upon us, affecting our communities, hospitals, and health delivery enterprise today. These challenges represent educational, academic, and economic opportunities that hospitalists should embrace.

Education and Awareness

The first step to engagement is to promote awareness of the effects of healthcare waste. Physicians remain one of the most trusted sources of information about the health impacts of climate change.¹⁶ By educating ourselves, we can spread accurate knowledge to our patients and communities. Furthermore, we have the ability to advocate for our hospitals to follow institutions such as Kaiser Permanente and the Cleveland Clinic.

Given that hospitalists play a key role in educating students and residents, they are ideal vehicles for such dissemination. Education should begin in medical and nursing schools, where curricula detailing the importance and impact of healthcare pollution may be introduced. As hospitalists, we should champion such efforts.

Measurement and Amelioration

Second, resource use, waste production, and areas for improvement must be systematically quantified. At a national level, the Sustainable Development Unit of the National Health System (NHS) measures and reports water use, waste production, and energy consumption of the UK’s healthcare sector. Consequently, the NHS has surpassed their 2015 goal of reducing their carbon footprint by 10%.¹⁷ By establishing a baseline understanding of our carbon emissions, waste production, and water consumption, areas where physicians and hospitals can target improvement can similarly be identified.

Hospitalists appreciate the practical tradeoffs between clinical work and change efforts; thus, they are critical in establishing pragmatic policies. Physicians, often in collaboration with environmental engineers, have used evidence-based methods such as life-cycle analysis (LCA) to evaluate the environmental impacts of the pharmaceuticals and procedures that they use.^18-20 An LCA is a cost-benefit analysis that examines multiple parameters of a product, namely, emissions, water use, costs, and waste production, from production to disposal. For example, an LCA of disposable custom packs for hysterectomies, vaginal deliveries, and laryngeal masks found costs savings and environmental benefits from choosing reusable over single-use items and removing unnecessary materials such as extra towels in this setting. ^18-20 By considering the full life cycle of a procedure, LCAs reveal important information about the value and safety of care. LCAs, along with other sustainable design strategies, are tools that can provide hospitalists with new insights for quality improvement.

Public Reporting

Numerous physicians are known for educating their communities about the impacts of pollution on health. Recently, a pediatrician brought the presence of lead in Flint’s water supply to the public’s attention, instigating government action and policy change.²¹ A group called Utah Physicians for a Healthy Environment publishes online summaries of peer-reviewed information on air pollution and health. The Huma Lung Foundation led by a pulmonologist in Chennai, India, is working with a local radio station to report daily air quality measurements along with health advisories for the city.

We must now extend this paradigm to encompass transparency about healthcare’s practices and their impact on health. Indeed, the public is comfortable with this idea: a survey of 1011 respondents in the UK found that 92% indicated that the healthcare system should be environmentally sustainable.²² One idea may be a public-facing scorecard for hospitals, akin to publicly reported quality metrics. We can look to the example of the SDU and corporations such as Apple, which publicly report their carbon emissions, waste production, water use, and other metrics of their environmental impact. By galvanizing efforts to quantify and report our impact, hospitalists have the opportunity to be a role model for the industry and increase trust within their communities.

Individual Actions

What can a hospitalist do today? First, simple measures, like turning off idle electronics, recycling appropriately, or avoiding the use of unnecessary supplies or tests, are behavioral steps in the right direction. Second, just as education, goal setting, and feedback have met success in improving hand hygiene,²³ we must begin the hard work of developing programs to monitor our environmental impact. Individual hospitalist carbon scores may help intensify efforts and spur improvement. Finally, we should learn and celebrate each other’s success. Renewed focus on this topic with increased reporting of interventions and outcomes is needed.

As hospitalists, we must look within ourselves to protect our planet and advocate for solutions that assure a sustainable future. By recognizing that a healthy environment is crucial to human health, we can set an example for other industries and create a safer world for our patients. Eliminating the harm we do is the first step in this process.

Disclosures 

The authors have nothing to disclose.

Funding 

Dr. Chopra is supported by a Career Development Award from the Agency for Healthcare Quality and Research (1-K08-HS-022835-01).

Healthcare is a “dirty” business with widespread effects on the environment. In the US, healthcare is estimated to generate 9.8% of our greenhouse gases and 9% of our particulate matter emissions.¹ Hazardous wastes must be incinerated, emitting carbon dioxide, nitrogen oxides, and volatile substances into the atmosphere.² Similarly, hospitals are responsible for 7% of commercial water use in the US.³ Conventional water treatment systems are not designed to remove heavy metals, pharmaceuticals, and disinfectants in hospital wastewaters; these compounds have been detected in rivers and streams throughout the US.^4,5 Furthermore, pharmaceutical compounds such as antibiotics, anti-epileptics, and narcotics have even been isolated in our drinking water.⁵

As hospitalists, we are the directors of inpatient care, yet we only witness brief moments in the lives of our patients and the products we use for their care. For example, we are unaware of particulate matter emissions needed to power an extra imaging study or the contribution of unused materials to a growing landfill. However, pollution, including that from our clinical practice, is detrimental to human health in many ways. Exposure to particulate matter and toxic wastes has been linked to increased rates of reproductive and developmental disorders, cancer, and respiratory disease. ⁶ Particles <2.5 µm in diameter can diffuse through alveoli into the bloodstream, contributing to heart disease, stroke, and lung disease.⁷ Climate change has been linked to a wide range of adverse cardiovascular, respiratory, infectious, and mental health outcomes.^8,9 These examples of the health impacts of pollution are illustrative but not exhaustive.

The environmental impact of US healthcare accounts for an estimated 470,000 disability-adjusted life years lost; this figure is on par with the burden of preventable medical errors.¹ Clearly, change is necessary at all levels in the healthcare system to address our impact on human health. Fortunately, healthcare systems and hospital administrators have begun to address this issue. This perspective describes sustainability efforts in hospitals and healthcare systems and seeks to motivate hospitalists to build upon these efforts.

With the ability to affect change from the top down, health systems are playing an important role in healthcare’s environmental sustainability. Ambitiously, Kaiser Permanente outlined eight environmental stewardship goals, which include becoming net carbon positive and recycling, reusing, or composting 100% of their non-hazardous waste by 2025.¹⁰ The Cleveland Clinic has pledged to become carbon neutral within the next 10 years.¹¹ Other healthcare systems may follow suite. Many “green” interventions aimed at reducing waste and pollution also protect population health and reduce hospital operating costs.

From 2011 to 2015, a group of Boston Hospitals decreased energy use by 9.4% compared with a historical growth of 1.5% per year and saved over 15 million dollars.¹² Similarly, Virginia Mason reduced landfill waste by reprocessing single-use medical devices, thereby decreasing purchasing costs by $3 million.¹³ As part of a regional campaign to protect the St. Croix River, Hudson Hospital and Clinic in Wisconsin saved over $20,000 with new recycling and waste reduction programs.¹³ Notably, these programs not only benefit hospitals but also patients and payers by reducing costs of care.

These examples illustrate that a greener healthcare industry is achievable. Despite the potential benefits, sustainability efforts in US hospitals are the exception, not the rule, and the diffusion of such innovations must be encouraged from within.

In addition to the moral case for environmentally sustainable healthcare,^14,15 such efforts can also improve our quality of care. The conversation around healthcare waste has focused on costs. Yet, examining our waste from a new perspective may reveal new ways to increase the value of patient care while protecting population health. Our communities and families are not immune to the health impacts of pollution, including that generated by our industry. However, predicted effects of climate change including altered patterns of vector-borne disease and frequent hurricanes and forest fires are upon us, affecting our communities, hospitals, and health delivery enterprise today. These challenges represent educational, academic, and economic opportunities that hospitalists should embrace.

Education and Awareness

The first step to engagement is to promote awareness of the effects of healthcare waste. Physicians remain one of the most trusted sources of information about the health impacts of climate change.¹⁶ By educating ourselves, we can spread accurate knowledge to our patients and communities. Furthermore, we have the ability to advocate for our hospitals to follow institutions such as Kaiser Permanente and the Cleveland Clinic.

Given that hospitalists play a key role in educating students and residents, they are ideal vehicles for such dissemination. Education should begin in medical and nursing schools, where curricula detailing the importance and impact of healthcare pollution may be introduced. As hospitalists, we should champion such efforts.

Measurement and Amelioration

Second, resource use, waste production, and areas for improvement must be systematically quantified. At a national level, the Sustainable Development Unit of the National Health System (NHS) measures and reports water use, waste production, and energy consumption of the UK’s healthcare sector. Consequently, the NHS has surpassed their 2015 goal of reducing their carbon footprint by 10%.¹⁷ By establishing a baseline understanding of our carbon emissions, waste production, and water consumption, areas where physicians and hospitals can target improvement can similarly be identified.

Hospitalists appreciate the practical tradeoffs between clinical work and change efforts; thus, they are critical in establishing pragmatic policies. Physicians, often in collaboration with environmental engineers, have used evidence-based methods such as life-cycle analysis (LCA) to evaluate the environmental impacts of the pharmaceuticals and procedures that they use.^18-20 An LCA is a cost-benefit analysis that examines multiple parameters of a product, namely, emissions, water use, costs, and waste production, from production to disposal. For example, an LCA of disposable custom packs for hysterectomies, vaginal deliveries, and laryngeal masks found costs savings and environmental benefits from choosing reusable over single-use items and removing unnecessary materials such as extra towels in this setting. ^18-20 By considering the full life cycle of a procedure, LCAs reveal important information about the value and safety of care. LCAs, along with other sustainable design strategies, are tools that can provide hospitalists with new insights for quality improvement.

Public Reporting

Numerous physicians are known for educating their communities about the impacts of pollution on health. Recently, a pediatrician brought the presence of lead in Flint’s water supply to the public’s attention, instigating government action and policy change.²¹ A group called Utah Physicians for a Healthy Environment publishes online summaries of peer-reviewed information on air pollution and health. The Huma Lung Foundation led by a pulmonologist in Chennai, India, is working with a local radio station to report daily air quality measurements along with health advisories for the city.

We must now extend this paradigm to encompass transparency about healthcare’s practices and their impact on health. Indeed, the public is comfortable with this idea: a survey of 1011 respondents in the UK found that 92% indicated that the healthcare system should be environmentally sustainable.²² One idea may be a public-facing scorecard for hospitals, akin to publicly reported quality metrics. We can look to the example of the SDU and corporations such as Apple, which publicly report their carbon emissions, waste production, water use, and other metrics of their environmental impact. By galvanizing efforts to quantify and report our impact, hospitalists have the opportunity to be a role model for the industry and increase trust within their communities.

Individual Actions

What can a hospitalist do today? First, simple measures, like turning off idle electronics, recycling appropriately, or avoiding the use of unnecessary supplies or tests, are behavioral steps in the right direction. Second, just as education, goal setting, and feedback have met success in improving hand hygiene,²³ we must begin the hard work of developing programs to monitor our environmental impact. Individual hospitalist carbon scores may help intensify efforts and spur improvement. Finally, we should learn and celebrate each other’s success. Renewed focus on this topic with increased reporting of interventions and outcomes is needed.

As hospitalists, we must look within ourselves to protect our planet and advocate for solutions that assure a sustainable future. By recognizing that a healthy environment is crucial to human health, we can set an example for other industries and create a safer world for our patients. Eliminating the harm we do is the first step in this process.

Disclosures 

The authors have nothing to disclose.

Funding 

Dr. Chopra is supported by a Career Development Award from the Agency for Healthcare Quality and Research (1-K08-HS-022835-01).

Healthcare is a “dirty” business with widespread effects on the environment. In the US, healthcare is estimated to generate 9.8% of our greenhouse gases and 9% of our particulate matter emissions.¹ Hazardous wastes must be incinerated, emitting carbon dioxide, nitrogen oxides, and volatile substances into the atmosphere.² Similarly, hospitals are responsible for 7% of commercial water use in the US.³ Conventional water treatment systems are not designed to remove heavy metals, pharmaceuticals, and disinfectants in hospital wastewaters; these compounds have been detected in rivers and streams throughout the US.^4,5 Furthermore, pharmaceutical compounds such as antibiotics, anti-epileptics, and narcotics have even been isolated in our drinking water.⁵

As hospitalists, we are the directors of inpatient care, yet we only witness brief moments in the lives of our patients and the products we use for their care. For example, we are unaware of particulate matter emissions needed to power an extra imaging study or the contribution of unused materials to a growing landfill. However, pollution, including that from our clinical practice, is detrimental to human health in many ways. Exposure to particulate matter and toxic wastes has been linked to increased rates of reproductive and developmental disorders, cancer, and respiratory disease. ⁶ Particles <2.5 µm in diameter can diffuse through alveoli into the bloodstream, contributing to heart disease, stroke, and lung disease.⁷ Climate change has been linked to a wide range of adverse cardiovascular, respiratory, infectious, and mental health outcomes.^8,9 These examples of the health impacts of pollution are illustrative but not exhaustive.

The environmental impact of US healthcare accounts for an estimated 470,000 disability-adjusted life years lost; this figure is on par with the burden of preventable medical errors.¹ Clearly, change is necessary at all levels in the healthcare system to address our impact on human health. Fortunately, healthcare systems and hospital administrators have begun to address this issue. This perspective describes sustainability efforts in hospitals and healthcare systems and seeks to motivate hospitalists to build upon these efforts.

With the ability to affect change from the top down, health systems are playing an important role in healthcare’s environmental sustainability. Ambitiously, Kaiser Permanente outlined eight environmental stewardship goals, which include becoming net carbon positive and recycling, reusing, or composting 100% of their non-hazardous waste by 2025.¹⁰ The Cleveland Clinic has pledged to become carbon neutral within the next 10 years.¹¹ Other healthcare systems may follow suite. Many “green” interventions aimed at reducing waste and pollution also protect population health and reduce hospital operating costs.

From 2011 to 2015, a group of Boston Hospitals decreased energy use by 9.4% compared with a historical growth of 1.5% per year and saved over 15 million dollars.¹² Similarly, Virginia Mason reduced landfill waste by reprocessing single-use medical devices, thereby decreasing purchasing costs by $3 million.¹³ As part of a regional campaign to protect the St. Croix River, Hudson Hospital and Clinic in Wisconsin saved over $20,000 with new recycling and waste reduction programs.¹³ Notably, these programs not only benefit hospitals but also patients and payers by reducing costs of care.

These examples illustrate that a greener healthcare industry is achievable. Despite the potential benefits, sustainability efforts in US hospitals are the exception, not the rule, and the diffusion of such innovations must be encouraged from within.

In addition to the moral case for environmentally sustainable healthcare,^14,15 such efforts can also improve our quality of care. The conversation around healthcare waste has focused on costs. Yet, examining our waste from a new perspective may reveal new ways to increase the value of patient care while protecting population health. Our communities and families are not immune to the health impacts of pollution, including that generated by our industry. However, predicted effects of climate change including altered patterns of vector-borne disease and frequent hurricanes and forest fires are upon us, affecting our communities, hospitals, and health delivery enterprise today. These challenges represent educational, academic, and economic opportunities that hospitalists should embrace.

Education and Awareness

The first step to engagement is to promote awareness of the effects of healthcare waste. Physicians remain one of the most trusted sources of information about the health impacts of climate change.¹⁶ By educating ourselves, we can spread accurate knowledge to our patients and communities. Furthermore, we have the ability to advocate for our hospitals to follow institutions such as Kaiser Permanente and the Cleveland Clinic.

Given that hospitalists play a key role in educating students and residents, they are ideal vehicles for such dissemination. Education should begin in medical and nursing schools, where curricula detailing the importance and impact of healthcare pollution may be introduced. As hospitalists, we should champion such efforts.

Measurement and Amelioration

Second, resource use, waste production, and areas for improvement must be systematically quantified. At a national level, the Sustainable Development Unit of the National Health System (NHS) measures and reports water use, waste production, and energy consumption of the UK’s healthcare sector. Consequently, the NHS has surpassed their 2015 goal of reducing their carbon footprint by 10%.¹⁷ By establishing a baseline understanding of our carbon emissions, waste production, and water consumption, areas where physicians and hospitals can target improvement can similarly be identified.

Hospitalists appreciate the practical tradeoffs between clinical work and change efforts; thus, they are critical in establishing pragmatic policies. Physicians, often in collaboration with environmental engineers, have used evidence-based methods such as life-cycle analysis (LCA) to evaluate the environmental impacts of the pharmaceuticals and procedures that they use.^18-20 An LCA is a cost-benefit analysis that examines multiple parameters of a product, namely, emissions, water use, costs, and waste production, from production to disposal. For example, an LCA of disposable custom packs for hysterectomies, vaginal deliveries, and laryngeal masks found costs savings and environmental benefits from choosing reusable over single-use items and removing unnecessary materials such as extra towels in this setting. ^18-20 By considering the full life cycle of a procedure, LCAs reveal important information about the value and safety of care. LCAs, along with other sustainable design strategies, are tools that can provide hospitalists with new insights for quality improvement.

Public Reporting

Numerous physicians are known for educating their communities about the impacts of pollution on health. Recently, a pediatrician brought the presence of lead in Flint’s water supply to the public’s attention, instigating government action and policy change.²¹ A group called Utah Physicians for a Healthy Environment publishes online summaries of peer-reviewed information on air pollution and health. The Huma Lung Foundation led by a pulmonologist in Chennai, India, is working with a local radio station to report daily air quality measurements along with health advisories for the city.

We must now extend this paradigm to encompass transparency about healthcare’s practices and their impact on health. Indeed, the public is comfortable with this idea: a survey of 1011 respondents in the UK found that 92% indicated that the healthcare system should be environmentally sustainable.²² One idea may be a public-facing scorecard for hospitals, akin to publicly reported quality metrics. We can look to the example of the SDU and corporations such as Apple, which publicly report their carbon emissions, waste production, water use, and other metrics of their environmental impact. By galvanizing efforts to quantify and report our impact, hospitalists have the opportunity to be a role model for the industry and increase trust within their communities.

Individual Actions

What can a hospitalist do today? First, simple measures, like turning off idle electronics, recycling appropriately, or avoiding the use of unnecessary supplies or tests, are behavioral steps in the right direction. Second, just as education, goal setting, and feedback have met success in improving hand hygiene,²³ we must begin the hard work of developing programs to monitor our environmental impact. Individual hospitalist carbon scores may help intensify efforts and spur improvement. Finally, we should learn and celebrate each other’s success. Renewed focus on this topic with increased reporting of interventions and outcomes is needed.

As hospitalists, we must look within ourselves to protect our planet and advocate for solutions that assure a sustainable future. By recognizing that a healthy environment is crucial to human health, we can set an example for other industries and create a safer world for our patients. Eliminating the harm we do is the first step in this process.

Disclosures 

The authors have nothing to disclose.

Funding 

Dr. Chopra is supported by a Career Development Award from the Agency for Healthcare Quality and Research (1-K08-HS-022835-01).

More than 80% of children who were given omalizumab with oral immunotherapy (OIT) for 36 weeks could safely consume portions of at least 2 foods they were causing an allergic reaction, according to findings from a phase 2 study funded by the National Institute of Allergy and Infectious Diseases. Omalizumab, an injectable antibody drug approved for moderate-to-severe allergic asthma, blocks the activity of IgE.

Researchers from Stanford University School of Medicine in California enrolled 48 children aged 4 years to 15 years with confirmed allergy to multiple foods, such as milk, egg, wheat, soy, sesame seeds, peanuts, and tree nuts. The children received omalizumab or placebo injections for the first 16 weeks. At week 8, all participants began eating small, gradually increasing amounts of an allergenic food. They continued OIT until week 36, when they underwent an oral food challenge.

Of the 36 children who received omalizumab, 30 were able to eat at least 2 grams of ≥ 2 allergenic foods, compared with that of only 4 of 12 children (33%) who received placebo. Children who received omalizumab also had fewer adverse events from OIT

More than 80% of children who were given omalizumab with oral immunotherapy (OIT) for 36 weeks could safely consume portions of at least 2 foods they were causing an allergic reaction, according to findings from a phase 2 study funded by the National Institute of Allergy and Infectious Diseases. Omalizumab, an injectable antibody drug approved for moderate-to-severe allergic asthma, blocks the activity of IgE.

Researchers from Stanford University School of Medicine in California enrolled 48 children aged 4 years to 15 years with confirmed allergy to multiple foods, such as milk, egg, wheat, soy, sesame seeds, peanuts, and tree nuts. The children received omalizumab or placebo injections for the first 16 weeks. At week 8, all participants began eating small, gradually increasing amounts of an allergenic food. They continued OIT until week 36, when they underwent an oral food challenge.

Of the 36 children who received omalizumab, 30 were able to eat at least 2 grams of ≥ 2 allergenic foods, compared with that of only 4 of 12 children (33%) who received placebo. Children who received omalizumab also had fewer adverse events from OIT

More than 80% of children who were given omalizumab with oral immunotherapy (OIT) for 36 weeks could safely consume portions of at least 2 foods they were causing an allergic reaction, according to findings from a phase 2 study funded by the National Institute of Allergy and Infectious Diseases. Omalizumab, an injectable antibody drug approved for moderate-to-severe allergic asthma, blocks the activity of IgE.

Researchers from Stanford University School of Medicine in California enrolled 48 children aged 4 years to 15 years with confirmed allergy to multiple foods, such as milk, egg, wheat, soy, sesame seeds, peanuts, and tree nuts. The children received omalizumab or placebo injections for the first 16 weeks. At week 8, all participants began eating small, gradually increasing amounts of an allergenic food. They continued OIT until week 36, when they underwent an oral food challenge.

Of the 36 children who received omalizumab, 30 were able to eat at least 2 grams of ≥ 2 allergenic foods, compared with that of only 4 of 12 children (33%) who received placebo. Children who received omalizumab also had fewer adverse events from OIT

The results are in: 93% of soldiers, retirees, and family members report very high overall satisfaction with their experience at Army medical treatment facilities.

Survey responses were for the DoD’s 2017 Joint Outpatient Experience Survey (JOES), which also asked about ease of access to Army providers (83% positive response) and overall experience with Army pharmacies (78% positive).

The results showed an increase in satisfaction of about 2% for those 3 questions compared with the results of 2016, the first time the Army participated in the survey, according to Melissa Gliner, senior health policy analyst with the Office of the Army Surgeon General, in an article for Defense.gov. The survey goes to about 10% of patients who have visited a military health facility.

Besides sharing the survey results with the facilities, Gliner advises them on how to improve the patient experience. For instance, she looks at civilian treatment facilities to see what works. One insight she culled was that it helps to have staff members circulate in the waiting area to chat with patients so they do not feel they are being ignored. Another was that facilities should retrain scheduling clerks to set up appointments without making the patient call back.

Gliner says the U.S. Army Medical Command also is working on a website that will help military health facilities share their ideas and “further elevate patient experience and survey scores.”

The results are in: 93% of soldiers, retirees, and family members report very high overall satisfaction with their experience at Army medical treatment facilities.

Survey responses were for the DoD’s 2017 Joint Outpatient Experience Survey (JOES), which also asked about ease of access to Army providers (83% positive response) and overall experience with Army pharmacies (78% positive).

The results showed an increase in satisfaction of about 2% for those 3 questions compared with the results of 2016, the first time the Army participated in the survey, according to Melissa Gliner, senior health policy analyst with the Office of the Army Surgeon General, in an article for Defense.gov. The survey goes to about 10% of patients who have visited a military health facility.

Besides sharing the survey results with the facilities, Gliner advises them on how to improve the patient experience. For instance, she looks at civilian treatment facilities to see what works. One insight she culled was that it helps to have staff members circulate in the waiting area to chat with patients so they do not feel they are being ignored. Another was that facilities should retrain scheduling clerks to set up appointments without making the patient call back.

Gliner says the U.S. Army Medical Command also is working on a website that will help military health facilities share their ideas and “further elevate patient experience and survey scores.”

The results are in: 93% of soldiers, retirees, and family members report very high overall satisfaction with their experience at Army medical treatment facilities.

Survey responses were for the DoD’s 2017 Joint Outpatient Experience Survey (JOES), which also asked about ease of access to Army providers (83% positive response) and overall experience with Army pharmacies (78% positive).

The results showed an increase in satisfaction of about 2% for those 3 questions compared with the results of 2016, the first time the Army participated in the survey, according to Melissa Gliner, senior health policy analyst with the Office of the Army Surgeon General, in an article for Defense.gov. The survey goes to about 10% of patients who have visited a military health facility.

Besides sharing the survey results with the facilities, Gliner advises them on how to improve the patient experience. For instance, she looks at civilian treatment facilities to see what works. One insight she culled was that it helps to have staff members circulate in the waiting area to chat with patients so they do not feel they are being ignored. Another was that facilities should retrain scheduling clerks to set up appointments without making the patient call back.

Gliner says the U.S. Army Medical Command also is working on a website that will help military health facilities share their ideas and “further elevate patient experience and survey scores.”

Deep venous thrombosis and pulmonary embolism, collectively known as venous thromboembolism (VTE), affect up to 600,000 Americans a year.¹ Most of these are hospital-associated venous thromboembolisms (HA-VTE).^1,2 VTE poses a substantial risk of mortality and long-term morbidity, and its treatment poses a risk of major bleeding.¹ As appropriate VTE prophylaxis (“prophylaxis”) can reduce the risk of VTE by 40% to 80% depending on the patient population,³ VTE risk assessment and prophylaxis is endorsed by multiple guidelines^4-7 and supported by regulatory agencies.^8-10

However, despite extensive study, consensus about the impact of prophylaxis^4,11 and the optimal method of risk assessment^4,5,7,12 is lacking. Meanwhile, implementation of prophylaxis in real-world settings is poor; only 40% to 60% of at-risk patients receive prophylaxis,¹³ and as few as <20% receive optimal prophylaxis.¹⁴ Both systematic reviews^15,16 and experience with VTE prevention collaboratives^17,18 found that multifaceted interventions and alerts may be most effective in improving prophylaxis rates, but without proof of improved VTE rates.¹⁵ There is limited experience with large-scale VTE prevention. Organizations like The Joint Commission (TJC)⁸ and the Surgical Care Improvement Project have promoted quality measures but without clear evidence of improvement.¹⁹ In addition, an analysis of over 20,000 medical patients at 35 hospitals found no difference in VTE rates between high- and low-performing hospitals,²⁰ suggesting that aggressive prophylaxis efforts may not reduce VTE, at least among medical patients.²¹ However, a 5-hospital University of California collaborative was associated with improved VTE rates, chiefly among surgical patients.²²

In 2011, Dignity Health targeted VTE for improvement after investigations of potentially preventable HA-VTE revealed variable patterns of prophylaxis. In addition, improvement seemed feasible because there is a proven framework for VTE quality improvement (QI) projects^17,18 and a record of success with the following 3 specific strategies: quality mentorship,²³ use of a simple VTE risk assessment method, and active surveillance (real-time monitoring targeting suboptimal prophylaxis with concurrent intervention). This active surveillance technique has been used successfully in prior improvement efforts, often termed measure-vention.^17,18,22,24

Setting and Participants

The QI collaborative was performed at 35 Dignity Health community hospitals in California, Arizona, and Nevada. Facilities ranged from 25 to 571 beds in size with a mixture of teaching and nonteaching hospitals. Prior to the initiative, prophylaxis improvement efforts were incomplete and inconsistent at study facilities. All adult acute care inpatients at all facilities were included except rehabilitation, behavioral health, skilled nursing, hospice, other nonacute care, and inpatient deliveries.

Design Overview

We performed a prospective, unblinded, open-intervention study of a QI collaborative in 35 community hospitals and studied the effect on prophylaxis and VTE rates with historical controls. The 35 hospitals were organized into 2 cohorts. In the “pilot” cohort, 9 hospitals (chosen to be representative of the various settings, size, and teaching status within the Dignity system) received funding from the Gordon and Betty Moore Foundation (GBMF) for intensive, individualized QI mentorship from experts as well as active surveillance (see “Interventions”). The pilot sites led the development of the VTE risk assessment and prophylaxis protocol (“VTE protocol”), measures, order sets, implementation tactics, and lessons learned, assisted by the mentor experts. Dissemination to the 26-hospital “spread” cohort was facilitated by the Dignity Health Hospital Engagement Network (HEN) infrastructure.

Timeline

Two of the pilot sites, acting as leads on the development of protocol and order set tools, formed improvement teams in March 2011, 6 to 12 months earlier than other Dignity sites. Planning and design work occurred from March 2011 to September 2012. Most implementation at the 35 hospitals occurred in a staggered fashion during calendar year (CY) 2012 and 2013 (see Figure 1). As few changes were made until mid-2012, we considered CY 2011 the baseline for comparison, CY 2012 to 2013 the implementation years, and CY 2014 the postimplementation period.

The project was reviewed by the Institutional Review Board (IRB) of Dignity Health and determined to be an IRB-exempt QI project.

Interventions

Collaborative Infrastructure

Multidisciplinary Teams

Improvement teams formed between March 2011 and September 2012. Members included a physician champion, frontline nurses and physicians, an administrative liaison, pharmacists, quality and data specialists, clinical informatics staff, and stakeholders from key clinical services. Teams met at least monthly at each site.

Physician Mentors

The 9 pilot sites received individualized mentorship provided by outside experts (IJ or GM) based on a model pioneered by the Society of Hospital Medicine’s (SHM) Mentored Implementation programs.²³ Each pilot site completed a self-assessment survey¹⁷ (see supplementary Appendix A) about past efforts, team composition, current performance, aims, barriers, and opportunities. The mentors reviewed the completed questionnaire with each hospital and provided advice on the VTE protocol and order set design, measurement, and benchmarking during 3 webinar meetings scheduled at 0, 3, and 9 months, plus as-needed e-mail and phone correspondence. After each webinar, the mentors provided detailed improvement suggestions (see supplementary Appendix B). Several hospitals received mentor site visits, which focused on unit rounding, active surveillance, staff and provider education, and problem-solving sessions with senior leadership, physician leadership, and the improvement team.

VTE Protocol

After a literature review and consultation with the mentors, Dignity Health developed and implemented a VTE protocol, modified from a model used in previous improvement efforts.^18,22-24 Its risk assessment method is often referred to as a “3 bucket” model because it assigns patients to high-, moderate-, or low-risk categories based on clinical factors (eg, major orthopedic surgery, prior VTE, and others), and the VTE protocol recommends interventions based on the risk category (see supplementary Appendix C). Dignity Health was transitioning to a single electronic health record (Cerner Corporation, North Kansas City, MO) during the study, and study hospitals were using multiple platforms, necessitating the development of both paper and electronic versions of the VTE protocol. The electronic version required completion of the VTE protocol for all inpatient admissions and transfers. The VTE protocol was completed in November 2011 and disseminated to other sites in a staggered fashion through November 2012. Completed protocols and improvement tips were shared by the project lead and by webinar sessions. Sites were also encouraged to implement a standardized practice that allowed nurses to apply sequential compression devices to at-risk patients without physician orders when indicated by protocol, when contraindications such as vascular disease or ulceration were absent.

Education

Staff were educated about the VTE protocol by local teams, starting between late 2011 and September 2012. The audience (physicians, nurses, pharmacists, etc.) and methods (conferences, fliers, etc.) were determined by local teams, following guidance by mentors and webinar content. Active surveillance provided opportunities for in-the-moment, patient-specific education and protocol reinforcement. Both mentors delivered educational presentations at pilot sites.

Active Surveillance

Sites were encouraged to perform daily review of prophylaxis adequacy for inpatients and correct lapses in real time (both under- and overprophylaxis). Inappropriate prophylaxis orders were addressed by contacting providers to change the order or document the rationale not to. Lapses in adherence to prophylaxis were addressed by nursing correction and education of involved staff. Active surveillance was funded for 10 hours a week at pilot sites. Spread sites received only minimal support from HEN monies. All sites used daily prophylaxis reports, enhanced to include contraindications like thrombocytopenia and coagulopathy, to facilitate efforts. Active surveillance began in May 2012 in the lead pilot hospitals and was implemented in other sites between October 2012 and February 2013.

Metrics

Prophylaxis Rates

Measurement of prophylaxis did not begin until 2012 to 2013; thus, the true baseline rate for prophylaxis was not captured. TJC metrics (VTE-1 and VTE-2)²⁵ were consolidated into a composite TJC prophylaxis rate from January 2012 to December 2014 for both pilot and spread hospitals. These measures assess the percentage of adult inpatients who received VTE prophylaxis or have documentation of why no prophylaxis was given the day of or day after hospital admission (VTE-1) or the day of or day after ICU admission or transfer (VTE-2). These measures are met if any mechanical or pharmacologic prophylaxis was delivered.

In addition to the TJC metric, the 9 pilot hospitals monitored rates of protocol-compliant prophylaxis for 12 to 20 months. Each patient’s prophylaxis was considered protocol compliant if it was consistent with the prophylaxis protocol at the time of the audit or if contraindications were documented (eg, patients eligible for, but with contraindications to, pharmacologic prophylaxis had to have an order for mechanical prophylaxis or documented contraindication to both modalities). As this measure was initiated in a staggered fashion, the rate of protocol-compliant prophylaxis is summarized for consecutive months of measurement rather than consecutive calendar months.

HA-VTE Rates

VTE events were captured by review of electronic coding data for the International Classification of Diseases, 9th Revision (ICD-9) codes 415.11-415.19, 453.2, 453.40-453.42, and 453.8-453.89. HA-VTE was defined as either new VTE not present on admission (NPOA HA-VTE) or new VTE presenting in a readmitted patient within 30 days of discharge (Readmit HA-VTE). Cases were stratified based on whether the patient had undergone a major operation (surgery patients) or not (medical patients) as identified by Medicare Services diagnosis-related group codes.

Control Measures

Potential adverse events were captured by review of electronic coding data for ICD-9 codes 289.84 (heparin-induced thrombocytopenia [HIT]) and E934.2 (adverse effects because of anticoagulants).

Statistical Analysis

Statistical process control charts were used to depict changes in prophylaxis rates over the 3 years for which data was collected. For VTE and safety outcomes, Pearson χ² value with relative risk (RR) calculations and 95% confidence intervals (CIs) were used to compare proportions between groups at baseline (CY 2011) versus postimplementation (CY 2014). Differences between the means of normally distributed data were calculated, and a 95% CI for the difference between the means was performed to assess statistical difference. Nonparametric characteristics were described by quartiles and interquartile range, and the 2-sided Mann-Whitney U test was performed to assess statistical difference between the CY 2011 and CY 2014 period.

Role of the Funding Source

The GBMF funded the collaborative and supported authorship of the manuscript but had no role in the design or conduct of the intervention, the collection or analysis of data, or the drafting of the manuscript.

Population Demographics

There were 1,155,069 adult inpatient admissions during the 4-year study period (264,280 in the 9 pilot sites, 890,789 in the 26 spread sites). There were no clinically relevant changes in gender distribution, mortality rate, median age, case mix index, or hospital length of stay in 2011 versus 2014. Men comprised 47.1% of the patient population in 2011 and 47.7% in 2014. The mortality rate was 2.7% in both years. Median age was 62 in 2011 and 63 in 2014. The mean case mix index (1.58 vs 1.65) and mean length of stay (4.29 vs 4.33 days) were similar in the 2 time periods.

Prophylaxis Rates

TJC Prophylaxis rates

Rates of Protocol-Compliant Prophylaxis

There were 34,071 active surveillance audits across the 20 months of reporting in the pilot cohort (mean, 1817 audits per month). The rate of protocol-compliant prophylaxis improved from 89% at month 1 of observation to 93% during month 2 and 97% by the last 3 months (Pearson χ² P < .001 for both comparisons).

HA-VTE

HA-VTE characteristics

Improved HA-VTE over Time

Four hundred twenty-eight fewer HA-VTEs occurred in 2014 than in 2011 (RR 0.78; 95% CI, 0.73-0.85) (Table and Figure 3). Readmission HA-VTEs were reduced by 315 (RR 0.72; 95% CI, 0.65-0.80), while the reduction in NPOA HA-VTEs was less robust (RR 0.88; 95% CI, 0.79-0.99). Pilot sites enjoyed a more robust reduction in HA-VTEs than spread sites (26% vs 20%), largely because the pilot cohort enjoyed a 34% reduction in NPOA HA-VTEs and a 20% reduction in Readmit HA-VTEs, while the spread cohort only achieved reductions in Readmit HA-VTEs.

Safety

Rates of HIT and adverse effects because of anticoagulants were low (Table). The rate of HIT declined from 178 events in 2011 to 109 in 2014 (RR 0.66; 95% CI, 0.52-0.84), and the RR of anticoagulant adverse events remained stable (RR 1.01; 95% CI, 0.87-1.15).

Our QI project, based on a proven collaborative approach and mentorship,^18,22,24 order set redesign, and active surveillance, was associated with 26% less VTEs in the pilot cohort and 20% less VTEs in the spread cohort. These gains, down to a final rate of approximately 4 HA-VTEs per 1000 admissions, occurred despite a low baseline HA-VTE rate. Dignity Health achieved these improvements in 35 hospitals with varied sizes, settings, ordering systems, and teaching statuses, achieving what is to our knowledge the largest VTE QI initiative yet reported.

Implementation experiences were not systematically recorded, and techniques were not compared with a control group. However, we believe that Dignity Health’s organizational commitment to improvement and centralized support were crucial for success. In addition, the pilot sites received grant support from the GBMF for intensive quality mentoring, a strategy with demonstrated value.²³ Mentors and team members noted that system-wide revision to the computerized physician order entry system was easiest to implement, while active surveillance represented the most labor-intensive intervention. Other experiences echoed lessons from previous VTE mentorship efforts.^17,18

The selection of a VTE protocol conducive to implementation and provider use was a key strategy. The ideal approach to VTE risk assessment is not known,^12,26 but guidelines either offer no specific guidance⁷ or would require implementation of 3 different systems per hospital.^4,5 Several of these are point scoring systems, which may have lower clinician acceptance or require programming to improve real-world use^18,26,27; the Padua score was derived from a patient population that differs significantly from those in the United States.¹² Our study provides more practical experience with a “3-bucket” model, which has previously shown high interobserver reliability, good clinician acceptance, and meaningful reductions of VTE, including in American patient populations.^18,22,24

The value of VTE prophylaxis is still disputed in many inpatient groups. The overall rate of HA-VTE is low, so the per-patient benefit of prophylaxis is low, and many patients may be overprophylaxed.^4,11,12 Recently, Flanders et al.²⁰ reported that HA-VTE rates among 20,800 medical inpatients in Michigan were low (about 1%) and similar at hospitals in the top (mean prophylaxis rate 86%) or bottom (mean prophylaxis rate 56%) tertiles of performance. Possible explanations for the differences between their multicenter experience and ours include our sample size (55 times larger) and the possibility that targeting prophylaxis to patients at highest need (captured in our protocol-compliant prophylaxis rates) matters more than prophylaxing a percent of the population.

Further research is needed to develop simple, easy-to-implement methods to identify inpatients who do not, or no longer, require prophylaxis.¹² Hospital systems also need methods to determine if prophylaxis improvement efforts can lower their HA-VTE rates and in which subpopulations. For example, a collaborative effort at the University of California lowered HA-VTE rates toward a common improved rate of 0.65% to 0.73%,²² while Dignity Health achieved improvement despite starting with an even lower baseline. In the University of California collaborative, benefits were limited chiefly to surgical patients, while Dignity Health achieved most improvement in medical patients, particularly in Readmit HA-VTE. If future research uncovers the reasons for these differences, it could help hospitals decide where to target improvement efforts.

Our study has several limitations. First, we used a nonrandomized time series design, so we cannot exclude other potential explanations for the change in VTE rates. However, there were no major changes in patient populations or concurrent projects likely to have influenced event rates. While we did not collect detailed demographic information on subjects, the broad inclusion criteria and multicenter design suggests a high degree of generalizability. Second, we followed inpatient VTE events and VTE-related readmissions, but not VTE treated in the outpatient setting. This did not change over the study, but the availability of all-oral therapy for VTE could have caused underdetection if clinic or emergency room doctors sent home more patients on oral therapy instead of readmitting them to the hospital. Third, implementation was enhanced by GBMF funds (at 9 sites, with the remainder benefitting from their experience), a shared electronic medical record at many sites, and a strong organizational safety culture, which may limit generalizability. However, spread sites showed similar improvement, paper-based sites were included, and the mentorship and quality collaborative models are scalable at low cost. Fourth, some QI efforts began at some pilot sites in CY 2011, so we could not compare completely clean pre- and postproject timeframes. However, early improvement would have resulted in an underestimation of the project’s impact. Lastly, the reason for a decline in HIT rates is not known. Standardized order sets promoted preferential use of low molecular weight heparin, which is less likely to induce HIT, and active surveillance targeted overprophylaxis as well as underprophylaxis, but we do not have data on heparin utilization patterns to confirm or refute these possibilities.

Strengths of our study include reductions in HA-VTE, both with and without access to GBMF funds, by using broadly available QI strategies.¹⁷ This real-world success and ease of dissemination are particularly important because the clinical trials of prophylaxis have been criticized for using highly selected patient populations,¹¹ and prophylaxis QI studies show an inconsistent impact on VTE outcomes.¹⁵ In previous studies, two of the authors monitored orders for prophylaxis^22,24; during this project, delivery for both pharmacologic and mechanical VTE prophylaxis was monitored, confirming that patient care actually changed.

Our multicenter VTE prophylaxis initiative, featuring a “3-bucket” VTE protocol, QI mentorship, and active surveillance as key interventions, was associated with improved prophylaxis rates and a reduction in HA-VTE by 22% with no increase in adverse events. This project provides a model for hospital systems seeking to optimize their prophylaxis efforts, and it supports the use of collaborative QI initiatives and SHM’s quality mentorship program as methods to drive improvement across health systems.

Disclosure

None of the authors have any conflicts of interest related to any topics or products discussed in the article. Dignity Health provided a stipend for writing the manuscript to GM and IJ, as noted in the article, but had no role in data analysis, writing, or decision to submit.

Deep venous thrombosis and pulmonary embolism, collectively known as venous thromboembolism (VTE), affect up to 600,000 Americans a year.¹ Most of these are hospital-associated venous thromboembolisms (HA-VTE).^1,2 VTE poses a substantial risk of mortality and long-term morbidity, and its treatment poses a risk of major bleeding.¹ As appropriate VTE prophylaxis (“prophylaxis”) can reduce the risk of VTE by 40% to 80% depending on the patient population,³ VTE risk assessment and prophylaxis is endorsed by multiple guidelines^4-7 and supported by regulatory agencies.^8-10

However, despite extensive study, consensus about the impact of prophylaxis^4,11 and the optimal method of risk assessment^4,5,7,12 is lacking. Meanwhile, implementation of prophylaxis in real-world settings is poor; only 40% to 60% of at-risk patients receive prophylaxis,¹³ and as few as <20% receive optimal prophylaxis.¹⁴ Both systematic reviews^15,16 and experience with VTE prevention collaboratives^17,18 found that multifaceted interventions and alerts may be most effective in improving prophylaxis rates, but without proof of improved VTE rates.¹⁵ There is limited experience with large-scale VTE prevention. Organizations like The Joint Commission (TJC)⁸ and the Surgical Care Improvement Project have promoted quality measures but without clear evidence of improvement.¹⁹ In addition, an analysis of over 20,000 medical patients at 35 hospitals found no difference in VTE rates between high- and low-performing hospitals,²⁰ suggesting that aggressive prophylaxis efforts may not reduce VTE, at least among medical patients.²¹ However, a 5-hospital University of California collaborative was associated with improved VTE rates, chiefly among surgical patients.²²

In 2011, Dignity Health targeted VTE for improvement after investigations of potentially preventable HA-VTE revealed variable patterns of prophylaxis. In addition, improvement seemed feasible because there is a proven framework for VTE quality improvement (QI) projects^17,18 and a record of success with the following 3 specific strategies: quality mentorship,²³ use of a simple VTE risk assessment method, and active surveillance (real-time monitoring targeting suboptimal prophylaxis with concurrent intervention). This active surveillance technique has been used successfully in prior improvement efforts, often termed measure-vention.^17,18,22,24

Setting and Participants

The QI collaborative was performed at 35 Dignity Health community hospitals in California, Arizona, and Nevada. Facilities ranged from 25 to 571 beds in size with a mixture of teaching and nonteaching hospitals. Prior to the initiative, prophylaxis improvement efforts were incomplete and inconsistent at study facilities. All adult acute care inpatients at all facilities were included except rehabilitation, behavioral health, skilled nursing, hospice, other nonacute care, and inpatient deliveries.

Design Overview

We performed a prospective, unblinded, open-intervention study of a QI collaborative in 35 community hospitals and studied the effect on prophylaxis and VTE rates with historical controls. The 35 hospitals were organized into 2 cohorts. In the “pilot” cohort, 9 hospitals (chosen to be representative of the various settings, size, and teaching status within the Dignity system) received funding from the Gordon and Betty Moore Foundation (GBMF) for intensive, individualized QI mentorship from experts as well as active surveillance (see “Interventions”). The pilot sites led the development of the VTE risk assessment and prophylaxis protocol (“VTE protocol”), measures, order sets, implementation tactics, and lessons learned, assisted by the mentor experts. Dissemination to the 26-hospital “spread” cohort was facilitated by the Dignity Health Hospital Engagement Network (HEN) infrastructure.

Timeline

Two of the pilot sites, acting as leads on the development of protocol and order set tools, formed improvement teams in March 2011, 6 to 12 months earlier than other Dignity sites. Planning and design work occurred from March 2011 to September 2012. Most implementation at the 35 hospitals occurred in a staggered fashion during calendar year (CY) 2012 and 2013 (see Figure 1). As few changes were made until mid-2012, we considered CY 2011 the baseline for comparison, CY 2012 to 2013 the implementation years, and CY 2014 the postimplementation period.

The project was reviewed by the Institutional Review Board (IRB) of Dignity Health and determined to be an IRB-exempt QI project.

Interventions

Collaborative Infrastructure

Multidisciplinary Teams

Improvement teams formed between March 2011 and September 2012. Members included a physician champion, frontline nurses and physicians, an administrative liaison, pharmacists, quality and data specialists, clinical informatics staff, and stakeholders from key clinical services. Teams met at least monthly at each site.

Physician Mentors

The 9 pilot sites received individualized mentorship provided by outside experts (IJ or GM) based on a model pioneered by the Society of Hospital Medicine’s (SHM) Mentored Implementation programs.²³ Each pilot site completed a self-assessment survey¹⁷ (see supplementary Appendix A) about past efforts, team composition, current performance, aims, barriers, and opportunities. The mentors reviewed the completed questionnaire with each hospital and provided advice on the VTE protocol and order set design, measurement, and benchmarking during 3 webinar meetings scheduled at 0, 3, and 9 months, plus as-needed e-mail and phone correspondence. After each webinar, the mentors provided detailed improvement suggestions (see supplementary Appendix B). Several hospitals received mentor site visits, which focused on unit rounding, active surveillance, staff and provider education, and problem-solving sessions with senior leadership, physician leadership, and the improvement team.

VTE Protocol

After a literature review and consultation with the mentors, Dignity Health developed and implemented a VTE protocol, modified from a model used in previous improvement efforts.^18,22-24 Its risk assessment method is often referred to as a “3 bucket” model because it assigns patients to high-, moderate-, or low-risk categories based on clinical factors (eg, major orthopedic surgery, prior VTE, and others), and the VTE protocol recommends interventions based on the risk category (see supplementary Appendix C). Dignity Health was transitioning to a single electronic health record (Cerner Corporation, North Kansas City, MO) during the study, and study hospitals were using multiple platforms, necessitating the development of both paper and electronic versions of the VTE protocol. The electronic version required completion of the VTE protocol for all inpatient admissions and transfers. The VTE protocol was completed in November 2011 and disseminated to other sites in a staggered fashion through November 2012. Completed protocols and improvement tips were shared by the project lead and by webinar sessions. Sites were also encouraged to implement a standardized practice that allowed nurses to apply sequential compression devices to at-risk patients without physician orders when indicated by protocol, when contraindications such as vascular disease or ulceration were absent.

Education

Staff were educated about the VTE protocol by local teams, starting between late 2011 and September 2012. The audience (physicians, nurses, pharmacists, etc.) and methods (conferences, fliers, etc.) were determined by local teams, following guidance by mentors and webinar content. Active surveillance provided opportunities for in-the-moment, patient-specific education and protocol reinforcement. Both mentors delivered educational presentations at pilot sites.

Active Surveillance

Sites were encouraged to perform daily review of prophylaxis adequacy for inpatients and correct lapses in real time (both under- and overprophylaxis). Inappropriate prophylaxis orders were addressed by contacting providers to change the order or document the rationale not to. Lapses in adherence to prophylaxis were addressed by nursing correction and education of involved staff. Active surveillance was funded for 10 hours a week at pilot sites. Spread sites received only minimal support from HEN monies. All sites used daily prophylaxis reports, enhanced to include contraindications like thrombocytopenia and coagulopathy, to facilitate efforts. Active surveillance began in May 2012 in the lead pilot hospitals and was implemented in other sites between October 2012 and February 2013.

Metrics

Prophylaxis Rates

Measurement of prophylaxis did not begin until 2012 to 2013; thus, the true baseline rate for prophylaxis was not captured. TJC metrics (VTE-1 and VTE-2)²⁵ were consolidated into a composite TJC prophylaxis rate from January 2012 to December 2014 for both pilot and spread hospitals. These measures assess the percentage of adult inpatients who received VTE prophylaxis or have documentation of why no prophylaxis was given the day of or day after hospital admission (VTE-1) or the day of or day after ICU admission or transfer (VTE-2). These measures are met if any mechanical or pharmacologic prophylaxis was delivered.

In addition to the TJC metric, the 9 pilot hospitals monitored rates of protocol-compliant prophylaxis for 12 to 20 months. Each patient’s prophylaxis was considered protocol compliant if it was consistent with the prophylaxis protocol at the time of the audit or if contraindications were documented (eg, patients eligible for, but with contraindications to, pharmacologic prophylaxis had to have an order for mechanical prophylaxis or documented contraindication to both modalities). As this measure was initiated in a staggered fashion, the rate of protocol-compliant prophylaxis is summarized for consecutive months of measurement rather than consecutive calendar months.

HA-VTE Rates

VTE events were captured by review of electronic coding data for the International Classification of Diseases, 9th Revision (ICD-9) codes 415.11-415.19, 453.2, 453.40-453.42, and 453.8-453.89. HA-VTE was defined as either new VTE not present on admission (NPOA HA-VTE) or new VTE presenting in a readmitted patient within 30 days of discharge (Readmit HA-VTE). Cases were stratified based on whether the patient had undergone a major operation (surgery patients) or not (medical patients) as identified by Medicare Services diagnosis-related group codes.

Control Measures

Potential adverse events were captured by review of electronic coding data for ICD-9 codes 289.84 (heparin-induced thrombocytopenia [HIT]) and E934.2 (adverse effects because of anticoagulants).

Statistical Analysis

Statistical process control charts were used to depict changes in prophylaxis rates over the 3 years for which data was collected. For VTE and safety outcomes, Pearson χ² value with relative risk (RR) calculations and 95% confidence intervals (CIs) were used to compare proportions between groups at baseline (CY 2011) versus postimplementation (CY 2014). Differences between the means of normally distributed data were calculated, and a 95% CI for the difference between the means was performed to assess statistical difference. Nonparametric characteristics were described by quartiles and interquartile range, and the 2-sided Mann-Whitney U test was performed to assess statistical difference between the CY 2011 and CY 2014 period.

Role of the Funding Source

The GBMF funded the collaborative and supported authorship of the manuscript but had no role in the design or conduct of the intervention, the collection or analysis of data, or the drafting of the manuscript.

Population Demographics

There were 1,155,069 adult inpatient admissions during the 4-year study period (264,280 in the 9 pilot sites, 890,789 in the 26 spread sites). There were no clinically relevant changes in gender distribution, mortality rate, median age, case mix index, or hospital length of stay in 2011 versus 2014. Men comprised 47.1% of the patient population in 2011 and 47.7% in 2014. The mortality rate was 2.7% in both years. Median age was 62 in 2011 and 63 in 2014. The mean case mix index (1.58 vs 1.65) and mean length of stay (4.29 vs 4.33 days) were similar in the 2 time periods.

Prophylaxis Rates

TJC Prophylaxis rates

Rates of Protocol-Compliant Prophylaxis

There were 34,071 active surveillance audits across the 20 months of reporting in the pilot cohort (mean, 1817 audits per month). The rate of protocol-compliant prophylaxis improved from 89% at month 1 of observation to 93% during month 2 and 97% by the last 3 months (Pearson χ² P < .001 for both comparisons).

HA-VTE

HA-VTE characteristics

Improved HA-VTE over Time

Four hundred twenty-eight fewer HA-VTEs occurred in 2014 than in 2011 (RR 0.78; 95% CI, 0.73-0.85) (Table and Figure 3). Readmission HA-VTEs were reduced by 315 (RR 0.72; 95% CI, 0.65-0.80), while the reduction in NPOA HA-VTEs was less robust (RR 0.88; 95% CI, 0.79-0.99). Pilot sites enjoyed a more robust reduction in HA-VTEs than spread sites (26% vs 20%), largely because the pilot cohort enjoyed a 34% reduction in NPOA HA-VTEs and a 20% reduction in Readmit HA-VTEs, while the spread cohort only achieved reductions in Readmit HA-VTEs.

Safety

Rates of HIT and adverse effects because of anticoagulants were low (Table). The rate of HIT declined from 178 events in 2011 to 109 in 2014 (RR 0.66; 95% CI, 0.52-0.84), and the RR of anticoagulant adverse events remained stable (RR 1.01; 95% CI, 0.87-1.15).

Our QI project, based on a proven collaborative approach and mentorship,^18,22,24 order set redesign, and active surveillance, was associated with 26% less VTEs in the pilot cohort and 20% less VTEs in the spread cohort. These gains, down to a final rate of approximately 4 HA-VTEs per 1000 admissions, occurred despite a low baseline HA-VTE rate. Dignity Health achieved these improvements in 35 hospitals with varied sizes, settings, ordering systems, and teaching statuses, achieving what is to our knowledge the largest VTE QI initiative yet reported.

Implementation experiences were not systematically recorded, and techniques were not compared with a control group. However, we believe that Dignity Health’s organizational commitment to improvement and centralized support were crucial for success. In addition, the pilot sites received grant support from the GBMF for intensive quality mentoring, a strategy with demonstrated value.²³ Mentors and team members noted that system-wide revision to the computerized physician order entry system was easiest to implement, while active surveillance represented the most labor-intensive intervention. Other experiences echoed lessons from previous VTE mentorship efforts.^17,18

The selection of a VTE protocol conducive to implementation and provider use was a key strategy. The ideal approach to VTE risk assessment is not known,^12,26 but guidelines either offer no specific guidance⁷ or would require implementation of 3 different systems per hospital.^4,5 Several of these are point scoring systems, which may have lower clinician acceptance or require programming to improve real-world use^18,26,27; the Padua score was derived from a patient population that differs significantly from those in the United States.¹² Our study provides more practical experience with a “3-bucket” model, which has previously shown high interobserver reliability, good clinician acceptance, and meaningful reductions of VTE, including in American patient populations.^18,22,24

The value of VTE prophylaxis is still disputed in many inpatient groups. The overall rate of HA-VTE is low, so the per-patient benefit of prophylaxis is low, and many patients may be overprophylaxed.^4,11,12 Recently, Flanders et al.²⁰ reported that HA-VTE rates among 20,800 medical inpatients in Michigan were low (about 1%) and similar at hospitals in the top (mean prophylaxis rate 86%) or bottom (mean prophylaxis rate 56%) tertiles of performance. Possible explanations for the differences between their multicenter experience and ours include our sample size (55 times larger) and the possibility that targeting prophylaxis to patients at highest need (captured in our protocol-compliant prophylaxis rates) matters more than prophylaxing a percent of the population.

Further research is needed to develop simple, easy-to-implement methods to identify inpatients who do not, or no longer, require prophylaxis.¹² Hospital systems also need methods to determine if prophylaxis improvement efforts can lower their HA-VTE rates and in which subpopulations. For example, a collaborative effort at the University of California lowered HA-VTE rates toward a common improved rate of 0.65% to 0.73%,²² while Dignity Health achieved improvement despite starting with an even lower baseline. In the University of California collaborative, benefits were limited chiefly to surgical patients, while Dignity Health achieved most improvement in medical patients, particularly in Readmit HA-VTE. If future research uncovers the reasons for these differences, it could help hospitals decide where to target improvement efforts.

Our study has several limitations. First, we used a nonrandomized time series design, so we cannot exclude other potential explanations for the change in VTE rates. However, there were no major changes in patient populations or concurrent projects likely to have influenced event rates. While we did not collect detailed demographic information on subjects, the broad inclusion criteria and multicenter design suggests a high degree of generalizability. Second, we followed inpatient VTE events and VTE-related readmissions, but not VTE treated in the outpatient setting. This did not change over the study, but the availability of all-oral therapy for VTE could have caused underdetection if clinic or emergency room doctors sent home more patients on oral therapy instead of readmitting them to the hospital. Third, implementation was enhanced by GBMF funds (at 9 sites, with the remainder benefitting from their experience), a shared electronic medical record at many sites, and a strong organizational safety culture, which may limit generalizability. However, spread sites showed similar improvement, paper-based sites were included, and the mentorship and quality collaborative models are scalable at low cost. Fourth, some QI efforts began at some pilot sites in CY 2011, so we could not compare completely clean pre- and postproject timeframes. However, early improvement would have resulted in an underestimation of the project’s impact. Lastly, the reason for a decline in HIT rates is not known. Standardized order sets promoted preferential use of low molecular weight heparin, which is less likely to induce HIT, and active surveillance targeted overprophylaxis as well as underprophylaxis, but we do not have data on heparin utilization patterns to confirm or refute these possibilities.

Strengths of our study include reductions in HA-VTE, both with and without access to GBMF funds, by using broadly available QI strategies.¹⁷ This real-world success and ease of dissemination are particularly important because the clinical trials of prophylaxis have been criticized for using highly selected patient populations,¹¹ and prophylaxis QI studies show an inconsistent impact on VTE outcomes.¹⁵ In previous studies, two of the authors monitored orders for prophylaxis^22,24; during this project, delivery for both pharmacologic and mechanical VTE prophylaxis was monitored, confirming that patient care actually changed.

Our multicenter VTE prophylaxis initiative, featuring a “3-bucket” VTE protocol, QI mentorship, and active surveillance as key interventions, was associated with improved prophylaxis rates and a reduction in HA-VTE by 22% with no increase in adverse events. This project provides a model for hospital systems seeking to optimize their prophylaxis efforts, and it supports the use of collaborative QI initiatives and SHM’s quality mentorship program as methods to drive improvement across health systems.

Disclosure

None of the authors have any conflicts of interest related to any topics or products discussed in the article. Dignity Health provided a stipend for writing the manuscript to GM and IJ, as noted in the article, but had no role in data analysis, writing, or decision to submit.

Deep venous thrombosis and pulmonary embolism, collectively known as venous thromboembolism (VTE), affect up to 600,000 Americans a year.¹ Most of these are hospital-associated venous thromboembolisms (HA-VTE).^1,2 VTE poses a substantial risk of mortality and long-term morbidity, and its treatment poses a risk of major bleeding.¹ As appropriate VTE prophylaxis (“prophylaxis”) can reduce the risk of VTE by 40% to 80% depending on the patient population,³ VTE risk assessment and prophylaxis is endorsed by multiple guidelines^4-7 and supported by regulatory agencies.^8-10

However, despite extensive study, consensus about the impact of prophylaxis^4,11 and the optimal method of risk assessment^4,5,7,12 is lacking. Meanwhile, implementation of prophylaxis in real-world settings is poor; only 40% to 60% of at-risk patients receive prophylaxis,¹³ and as few as <20% receive optimal prophylaxis.¹⁴ Both systematic reviews^15,16 and experience with VTE prevention collaboratives^17,18 found that multifaceted interventions and alerts may be most effective in improving prophylaxis rates, but without proof of improved VTE rates.¹⁵ There is limited experience with large-scale VTE prevention. Organizations like The Joint Commission (TJC)⁸ and the Surgical Care Improvement Project have promoted quality measures but without clear evidence of improvement.¹⁹ In addition, an analysis of over 20,000 medical patients at 35 hospitals found no difference in VTE rates between high- and low-performing hospitals,²⁰ suggesting that aggressive prophylaxis efforts may not reduce VTE, at least among medical patients.²¹ However, a 5-hospital University of California collaborative was associated with improved VTE rates, chiefly among surgical patients.²²

In 2011, Dignity Health targeted VTE for improvement after investigations of potentially preventable HA-VTE revealed variable patterns of prophylaxis. In addition, improvement seemed feasible because there is a proven framework for VTE quality improvement (QI) projects^17,18 and a record of success with the following 3 specific strategies: quality mentorship,²³ use of a simple VTE risk assessment method, and active surveillance (real-time monitoring targeting suboptimal prophylaxis with concurrent intervention). This active surveillance technique has been used successfully in prior improvement efforts, often termed measure-vention.^17,18,22,24

Setting and Participants

The QI collaborative was performed at 35 Dignity Health community hospitals in California, Arizona, and Nevada. Facilities ranged from 25 to 571 beds in size with a mixture of teaching and nonteaching hospitals. Prior to the initiative, prophylaxis improvement efforts were incomplete and inconsistent at study facilities. All adult acute care inpatients at all facilities were included except rehabilitation, behavioral health, skilled nursing, hospice, other nonacute care, and inpatient deliveries.

Design Overview

We performed a prospective, unblinded, open-intervention study of a QI collaborative in 35 community hospitals and studied the effect on prophylaxis and VTE rates with historical controls. The 35 hospitals were organized into 2 cohorts. In the “pilot” cohort, 9 hospitals (chosen to be representative of the various settings, size, and teaching status within the Dignity system) received funding from the Gordon and Betty Moore Foundation (GBMF) for intensive, individualized QI mentorship from experts as well as active surveillance (see “Interventions”). The pilot sites led the development of the VTE risk assessment and prophylaxis protocol (“VTE protocol”), measures, order sets, implementation tactics, and lessons learned, assisted by the mentor experts. Dissemination to the 26-hospital “spread” cohort was facilitated by the Dignity Health Hospital Engagement Network (HEN) infrastructure.

Timeline

Two of the pilot sites, acting as leads on the development of protocol and order set tools, formed improvement teams in March 2011, 6 to 12 months earlier than other Dignity sites. Planning and design work occurred from March 2011 to September 2012. Most implementation at the 35 hospitals occurred in a staggered fashion during calendar year (CY) 2012 and 2013 (see Figure 1). As few changes were made until mid-2012, we considered CY 2011 the baseline for comparison, CY 2012 to 2013 the implementation years, and CY 2014 the postimplementation period.

The project was reviewed by the Institutional Review Board (IRB) of Dignity Health and determined to be an IRB-exempt QI project.

Interventions

Collaborative Infrastructure

Multidisciplinary Teams

Improvement teams formed between March 2011 and September 2012. Members included a physician champion, frontline nurses and physicians, an administrative liaison, pharmacists, quality and data specialists, clinical informatics staff, and stakeholders from key clinical services. Teams met at least monthly at each site.

Physician Mentors

The 9 pilot sites received individualized mentorship provided by outside experts (IJ or GM) based on a model pioneered by the Society of Hospital Medicine’s (SHM) Mentored Implementation programs.²³ Each pilot site completed a self-assessment survey¹⁷ (see supplementary Appendix A) about past efforts, team composition, current performance, aims, barriers, and opportunities. The mentors reviewed the completed questionnaire with each hospital and provided advice on the VTE protocol and order set design, measurement, and benchmarking during 3 webinar meetings scheduled at 0, 3, and 9 months, plus as-needed e-mail and phone correspondence. After each webinar, the mentors provided detailed improvement suggestions (see supplementary Appendix B). Several hospitals received mentor site visits, which focused on unit rounding, active surveillance, staff and provider education, and problem-solving sessions with senior leadership, physician leadership, and the improvement team.

VTE Protocol

After a literature review and consultation with the mentors, Dignity Health developed and implemented a VTE protocol, modified from a model used in previous improvement efforts.^18,22-24 Its risk assessment method is often referred to as a “3 bucket” model because it assigns patients to high-, moderate-, or low-risk categories based on clinical factors (eg, major orthopedic surgery, prior VTE, and others), and the VTE protocol recommends interventions based on the risk category (see supplementary Appendix C). Dignity Health was transitioning to a single electronic health record (Cerner Corporation, North Kansas City, MO) during the study, and study hospitals were using multiple platforms, necessitating the development of both paper and electronic versions of the VTE protocol. The electronic version required completion of the VTE protocol for all inpatient admissions and transfers. The VTE protocol was completed in November 2011 and disseminated to other sites in a staggered fashion through November 2012. Completed protocols and improvement tips were shared by the project lead and by webinar sessions. Sites were also encouraged to implement a standardized practice that allowed nurses to apply sequential compression devices to at-risk patients without physician orders when indicated by protocol, when contraindications such as vascular disease or ulceration were absent.

Education

Staff were educated about the VTE protocol by local teams, starting between late 2011 and September 2012. The audience (physicians, nurses, pharmacists, etc.) and methods (conferences, fliers, etc.) were determined by local teams, following guidance by mentors and webinar content. Active surveillance provided opportunities for in-the-moment, patient-specific education and protocol reinforcement. Both mentors delivered educational presentations at pilot sites.

Active Surveillance

Sites were encouraged to perform daily review of prophylaxis adequacy for inpatients and correct lapses in real time (both under- and overprophylaxis). Inappropriate prophylaxis orders were addressed by contacting providers to change the order or document the rationale not to. Lapses in adherence to prophylaxis were addressed by nursing correction and education of involved staff. Active surveillance was funded for 10 hours a week at pilot sites. Spread sites received only minimal support from HEN monies. All sites used daily prophylaxis reports, enhanced to include contraindications like thrombocytopenia and coagulopathy, to facilitate efforts. Active surveillance began in May 2012 in the lead pilot hospitals and was implemented in other sites between October 2012 and February 2013.

Metrics

Prophylaxis Rates

Measurement of prophylaxis did not begin until 2012 to 2013; thus, the true baseline rate for prophylaxis was not captured. TJC metrics (VTE-1 and VTE-2)²⁵ were consolidated into a composite TJC prophylaxis rate from January 2012 to December 2014 for both pilot and spread hospitals. These measures assess the percentage of adult inpatients who received VTE prophylaxis or have documentation of why no prophylaxis was given the day of or day after hospital admission (VTE-1) or the day of or day after ICU admission or transfer (VTE-2). These measures are met if any mechanical or pharmacologic prophylaxis was delivered.

In addition to the TJC metric, the 9 pilot hospitals monitored rates of protocol-compliant prophylaxis for 12 to 20 months. Each patient’s prophylaxis was considered protocol compliant if it was consistent with the prophylaxis protocol at the time of the audit or if contraindications were documented (eg, patients eligible for, but with contraindications to, pharmacologic prophylaxis had to have an order for mechanical prophylaxis or documented contraindication to both modalities). As this measure was initiated in a staggered fashion, the rate of protocol-compliant prophylaxis is summarized for consecutive months of measurement rather than consecutive calendar months.

HA-VTE Rates

VTE events were captured by review of electronic coding data for the International Classification of Diseases, 9th Revision (ICD-9) codes 415.11-415.19, 453.2, 453.40-453.42, and 453.8-453.89. HA-VTE was defined as either new VTE not present on admission (NPOA HA-VTE) or new VTE presenting in a readmitted patient within 30 days of discharge (Readmit HA-VTE). Cases were stratified based on whether the patient had undergone a major operation (surgery patients) or not (medical patients) as identified by Medicare Services diagnosis-related group codes.

Control Measures

Potential adverse events were captured by review of electronic coding data for ICD-9 codes 289.84 (heparin-induced thrombocytopenia [HIT]) and E934.2 (adverse effects because of anticoagulants).

Statistical Analysis

Statistical process control charts were used to depict changes in prophylaxis rates over the 3 years for which data was collected. For VTE and safety outcomes, Pearson χ² value with relative risk (RR) calculations and 95% confidence intervals (CIs) were used to compare proportions between groups at baseline (CY 2011) versus postimplementation (CY 2014). Differences between the means of normally distributed data were calculated, and a 95% CI for the difference between the means was performed to assess statistical difference. Nonparametric characteristics were described by quartiles and interquartile range, and the 2-sided Mann-Whitney U test was performed to assess statistical difference between the CY 2011 and CY 2014 period.

Role of the Funding Source

The GBMF funded the collaborative and supported authorship of the manuscript but had no role in the design or conduct of the intervention, the collection or analysis of data, or the drafting of the manuscript.

Population Demographics

There were 1,155,069 adult inpatient admissions during the 4-year study period (264,280 in the 9 pilot sites, 890,789 in the 26 spread sites). There were no clinically relevant changes in gender distribution, mortality rate, median age, case mix index, or hospital length of stay in 2011 versus 2014. Men comprised 47.1% of the patient population in 2011 and 47.7% in 2014. The mortality rate was 2.7% in both years. Median age was 62 in 2011 and 63 in 2014. The mean case mix index (1.58 vs 1.65) and mean length of stay (4.29 vs 4.33 days) were similar in the 2 time periods.

Prophylaxis Rates

TJC Prophylaxis rates

Rates of Protocol-Compliant Prophylaxis

There were 34,071 active surveillance audits across the 20 months of reporting in the pilot cohort (mean, 1817 audits per month). The rate of protocol-compliant prophylaxis improved from 89% at month 1 of observation to 93% during month 2 and 97% by the last 3 months (Pearson χ² P < .001 for both comparisons).

HA-VTE

HA-VTE characteristics

Improved HA-VTE over Time

Four hundred twenty-eight fewer HA-VTEs occurred in 2014 than in 2011 (RR 0.78; 95% CI, 0.73-0.85) (Table and Figure 3). Readmission HA-VTEs were reduced by 315 (RR 0.72; 95% CI, 0.65-0.80), while the reduction in NPOA HA-VTEs was less robust (RR 0.88; 95% CI, 0.79-0.99). Pilot sites enjoyed a more robust reduction in HA-VTEs than spread sites (26% vs 20%), largely because the pilot cohort enjoyed a 34% reduction in NPOA HA-VTEs and a 20% reduction in Readmit HA-VTEs, while the spread cohort only achieved reductions in Readmit HA-VTEs.

Safety

Rates of HIT and adverse effects because of anticoagulants were low (Table). The rate of HIT declined from 178 events in 2011 to 109 in 2014 (RR 0.66; 95% CI, 0.52-0.84), and the RR of anticoagulant adverse events remained stable (RR 1.01; 95% CI, 0.87-1.15).

Our QI project, based on a proven collaborative approach and mentorship,^18,22,24 order set redesign, and active surveillance, was associated with 26% less VTEs in the pilot cohort and 20% less VTEs in the spread cohort. These gains, down to a final rate of approximately 4 HA-VTEs per 1000 admissions, occurred despite a low baseline HA-VTE rate. Dignity Health achieved these improvements in 35 hospitals with varied sizes, settings, ordering systems, and teaching statuses, achieving what is to our knowledge the largest VTE QI initiative yet reported.

Implementation experiences were not systematically recorded, and techniques were not compared with a control group. However, we believe that Dignity Health’s organizational commitment to improvement and centralized support were crucial for success. In addition, the pilot sites received grant support from the GBMF for intensive quality mentoring, a strategy with demonstrated value.²³ Mentors and team members noted that system-wide revision to the computerized physician order entry system was easiest to implement, while active surveillance represented the most labor-intensive intervention. Other experiences echoed lessons from previous VTE mentorship efforts.^17,18

The selection of a VTE protocol conducive to implementation and provider use was a key strategy. The ideal approach to VTE risk assessment is not known,^12,26 but guidelines either offer no specific guidance⁷ or would require implementation of 3 different systems per hospital.^4,5 Several of these are point scoring systems, which may have lower clinician acceptance or require programming to improve real-world use^18,26,27; the Padua score was derived from a patient population that differs significantly from those in the United States.¹² Our study provides more practical experience with a “3-bucket” model, which has previously shown high interobserver reliability, good clinician acceptance, and meaningful reductions of VTE, including in American patient populations.^18,22,24

The value of VTE prophylaxis is still disputed in many inpatient groups. The overall rate of HA-VTE is low, so the per-patient benefit of prophylaxis is low, and many patients may be overprophylaxed.^4,11,12 Recently, Flanders et al.²⁰ reported that HA-VTE rates among 20,800 medical inpatients in Michigan were low (about 1%) and similar at hospitals in the top (mean prophylaxis rate 86%) or bottom (mean prophylaxis rate 56%) tertiles of performance. Possible explanations for the differences between their multicenter experience and ours include our sample size (55 times larger) and the possibility that targeting prophylaxis to patients at highest need (captured in our protocol-compliant prophylaxis rates) matters more than prophylaxing a percent of the population.

Further research is needed to develop simple, easy-to-implement methods to identify inpatients who do not, or no longer, require prophylaxis.¹² Hospital systems also need methods to determine if prophylaxis improvement efforts can lower their HA-VTE rates and in which subpopulations. For example, a collaborative effort at the University of California lowered HA-VTE rates toward a common improved rate of 0.65% to 0.73%,²² while Dignity Health achieved improvement despite starting with an even lower baseline. In the University of California collaborative, benefits were limited chiefly to surgical patients, while Dignity Health achieved most improvement in medical patients, particularly in Readmit HA-VTE. If future research uncovers the reasons for these differences, it could help hospitals decide where to target improvement efforts.

Our study has several limitations. First, we used a nonrandomized time series design, so we cannot exclude other potential explanations for the change in VTE rates. However, there were no major changes in patient populations or concurrent projects likely to have influenced event rates. While we did not collect detailed demographic information on subjects, the broad inclusion criteria and multicenter design suggests a high degree of generalizability. Second, we followed inpatient VTE events and VTE-related readmissions, but not VTE treated in the outpatient setting. This did not change over the study, but the availability of all-oral therapy for VTE could have caused underdetection if clinic or emergency room doctors sent home more patients on oral therapy instead of readmitting them to the hospital. Third, implementation was enhanced by GBMF funds (at 9 sites, with the remainder benefitting from their experience), a shared electronic medical record at many sites, and a strong organizational safety culture, which may limit generalizability. However, spread sites showed similar improvement, paper-based sites were included, and the mentorship and quality collaborative models are scalable at low cost. Fourth, some QI efforts began at some pilot sites in CY 2011, so we could not compare completely clean pre- and postproject timeframes. However, early improvement would have resulted in an underestimation of the project’s impact. Lastly, the reason for a decline in HIT rates is not known. Standardized order sets promoted preferential use of low molecular weight heparin, which is less likely to induce HIT, and active surveillance targeted overprophylaxis as well as underprophylaxis, but we do not have data on heparin utilization patterns to confirm or refute these possibilities.

Strengths of our study include reductions in HA-VTE, both with and without access to GBMF funds, by using broadly available QI strategies.¹⁷ This real-world success and ease of dissemination are particularly important because the clinical trials of prophylaxis have been criticized for using highly selected patient populations,¹¹ and prophylaxis QI studies show an inconsistent impact on VTE outcomes.¹⁵ In previous studies, two of the authors monitored orders for prophylaxis^22,24; during this project, delivery for both pharmacologic and mechanical VTE prophylaxis was monitored, confirming that patient care actually changed.

Our multicenter VTE prophylaxis initiative, featuring a “3-bucket” VTE protocol, QI mentorship, and active surveillance as key interventions, was associated with improved prophylaxis rates and a reduction in HA-VTE by 22% with no increase in adverse events. This project provides a model for hospital systems seeking to optimize their prophylaxis efforts, and it supports the use of collaborative QI initiatives and SHM’s quality mentorship program as methods to drive improvement across health systems.

Disclosure

None of the authors have any conflicts of interest related to any topics or products discussed in the article. Dignity Health provided a stipend for writing the manuscript to GM and IJ, as noted in the article, but had no role in data analysis, writing, or decision to submit.

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/

Chest pain is one of the most common complaints among patients presenting to the emergency department. Moreover, at least 30% of patients who present with chest pain are admitted for observation, and >70% of those admitted with chest pain undergo cardiac stress testing (CST) during hospitalization. Several clinical risk prediction models have validated evaluation processes for managing patients with chest pain, helping to identify those at a low risk of major adverse cardiac events. Among these, the Thrombolysis in Myocardial Infarction or HEART score can identify patients safe to be discharged with outpatient CST within 72 h. It is unnecessary to hospitalize all low-risk patients for cardiac testing because it may expose them to needless risk and avoidable care costs, with little additional benefit.

A 60-year-old man with a history of osteoarthritis and depression presented to our emergency department (ED) with a 1-month history of left-sided chest pain that was present both at rest and exertion. There were no aggravating or relieving factors for the pain and no associated shortness of breath, diaphoresis, nausea, or lightheadedness. He smoked a half pack of cigarettes daily for 5 years in his twenties. The patient was taking aspirin 81 mg daily and paroxetine 40 mg daily, which he had been taking for 10 years. There was a family history of coronary artery disease in his mother, father, and sister. On examination, he was afebrile, with a blood pressure of 138/78 mm Hg and a heart rate of 62 beats/min; he appeared well, with no abnormal cardiopulmonary findings. Investigation revealed a normal initial troponin I level (<0.034 mg/mL) and normal electrocardiogram (ECG) with normal sinus rhythm (75 beats/min), normal axis, no ST changes, and no Q waves. He was therefore admitted to the hospital for further evaluation.

Each year, >7 million patients visit ED for chest pain in the United States,¹ with approximately 13% diagnosed with acute coronary syndromes (ACSs).² Over 30% of patients who present to ED with chest pain are hospitalized for observation, symptom evaluation, and risk stratification.³ In 2012, the mean Medicare reimbursement cost was $1,741 for in-hospital observation,⁴ with up to 70% of admitted patients undergoing cardiac stress testing (CST) before discharge.⁵

A scientific statement by the American Heart Association in 2010 recommended that patients considered to be at low risk for ACS after initial evaluation (based on presenting symptoms, past history, ECG findings, and initial cardiac biomarkers) should undergo CST within 72 h (preferably within 24 h) of presentation to provoke ischemia or detect anatomic coronary artery disease.⁶ Early exercise treadmill testing as part of an accelerated diagnostic pathway can also reduce the length of stays (LOS) in hospital and lower the medical costs.⁷ Moreover, when there is noncompliance or poor accessibility, failure to pursue early exercise testing in a hospital could result in a loss of patients to follow-up. Hospitalization for testing through accelerated diagnostic pathways may improve access to care and reduce clinical and legal risks associated with a major adverse cardiac event (MACE).

Clinical Risk Prediction Models

When a patient initially presents with chest pain, it should be determined if the symptoms are related to ACS or some other diagnosis. Hospitalization is required for patients with ACS but may not be for those without ACS and those with a low risk of inducible ischemia. Clinical risk scores and risk prediction models, such as the Thrombolysis in Myocardial Infarction (TIMI) and HEART scores, have been used in accelerated diagnostic protocols to determine a patient’s likelihood of having ACS. Several large trials of these clinical risk prediction models have validated the processes for evaluating patients with chest pain.

The TIMI risk score, the most well-known model, assesses risk based on the presence or absence of 7 characteristics (Appendix 1). It should be noted that the patient population studied for initial validation of this model comprised high-risk patients with unstable angina or non-ST elevation myocardial infarction who would benefit from early or urgent invasive therapy.⁸ In this population, TIMI scores of 0-1 are associated with low risk, with a 4.7% risk of ACS at 14 days.⁸ In another study of patients presenting to ED with undifferentiated chest pain and a TIMI score of zero, the risk of MACE at 30 days was approximately 2%.⁹

The HEART score is also used for patients presenting to ED with undifferentiated chest pain and assesses 5 separate variables scored 0–2 (Appendix 2). The original research gave a score of 2 to a troponin I level greater than twice the upper limit of the normal level,¹⁰ whereas a subsequent validation study gave a score of 2 to a troponin I or T level greater than or equal to 3 times the upper limit of the normal level.¹¹ Patients are considered at low, intermediate, and high risk based on scores of 0–3, 4–6, and 7–10, respectively.^10,11 Backus et al. performed a prospective randomized trial of 2388 patients who presented to ED with chest pain to validate the HEART score and compare it to the TIMI risk score. The HEART score performed better than the TIMI risk score in low-risk patients, with TIMI scores of 0-1 and HEART scores of 0–3 having a 6-week MACE risk of 2.8% and 1.7%, respectively.¹¹

A HEART pathway was developed that combines the HEART score with serial troponin I assays assessed at the time of initial presentation and approximately 3 h later.¹² Mahler et al. randomized 282 patients presenting to ED with chest pain to either the HEART pathway or conventional care. Patients with low-risk HEART scores and an abnormal troponin I level were admitted for cardiology consultation, whereas discharge was recommended for those with low scores and a normal troponin I level. Despite nearly 20% of the study cohort having a history of myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting, approximately 40% of patients in the HEART pathway were identified as low risk, increasing early discharge rates by 21.3% and decreasing the average LOS by 12 h. No low-risk patient suffered a MACE within 30 days, and the HEART pathway had a sensitivity and a negative predictive value of approximately 99%.

Costs and Harms of Hospitalization for Cardiac Testing

Hospitalization carries measurable risks.^13,14 Between 2008 and 2013, Weinstock et al. evaluated the outcomes of patients presenting with chest pain who were placed in an observation unit for suspected ACS.¹⁵ Low-risk patients were defined as those with normal ECGs (no ischemic changes), 2 negative troponin tests performed 60–420 min apart (no particular troponin assay specified), and stable vital signs. They identified 7266 patients who were considered to have low risk, among whom 4 (0.06%) had an adverse outcome in the hospital (eg, life-threatening arrhythmia, ST-segment elevation myocardial infarction, cardiac or respiratory arrest, or death); 3 among the 4 patients had a cardiac-related adverse outcome. The overall risk of adverse outcomes due to cardiac causes was 1 in 2422 admissions (0.04%). The authors compared their results with the reported risk of 1 in 164 admissions for preventable adverse events contributing to patient death during routine hospitalization (eg, medication or procedure errors).¹⁴

Outpatient CST can be reliably and safely performed for patients with chest pain.^16-18 There is no clear evidence that earlier CST leads to improved patient outcomes, and CST in the absence of acute ischemia (or ACS) increases the rates of angiography and revascularization without improvements in the rate of myocardial infarction.^19-21 Given the costs of in-hospital observation⁴ and the dubious benefits of providing CST for patients with low-risk chest pain, admitting all patients with low-risk chest pain exposes them to costs and harms with little potential benefit.

Patients presenting with chest pain with either dynamic ECG changes or an elevated troponin level require hospitalization for further ACS diagnosis and treatment. When ACS cannot be clearly diagnosed at the initial evaluation, healthcare providers should use clinical risk prediction models to stratify patients. Those deemed to be at an intermediate or high risk by these models should be hospitalized for further evaluation, as should those at low risk but for whom access to outpatient follow-up is difficult (eg, those without health insurance).

A complete history and physical examination, along with ECG and cardiac biomarker testing, are required for all patients presenting with chest pain. Validated clinical risk prediction models should then be used to determine the likelihood of a cardiac event. Fanaroff et al. reported that low-risk HEART scores of 0–3 and TIMI scores of 0-1 gave positive likelihood ratios of 0.2 and 0.31, respectively.²² Using a pre-test probability of 13%, as reported by Bhuiya et al.,² the likelihood of ACS or MACE within 6 weeks is 2.9% for patients with low-risk HEART scores and 4.4% for those with low-risk TIMI scores.²² These risk prediction models allow clinicians to provide a shared decision-making plan with the patient and discuss the risks and benefits of in-hospital versus outpatient cardiac testing, especially among patients with access to appropriate outpatient follow-up.²³ Low-risk patients can be referred for outpatient testing within 72 h, reducing hospitalization-associated costs and harms.

• Patients presenting with chest pain should undergo a complete history taking and physical examination, as well as ECG and cardiac biomarker testing (eg, troponin I level at presentation and approximately 3 h later).
• Clinical risk prediction models, such as TIMI or HEART scores, should then be used to determine the risk of MACE.
• Patients at a low risk may be safely discharged with outpatient CST performed within 72 h.
• Patients at an intermediate or high risk of MACE should be hospitalized for further evaluation, as should those with low-risk chest pain who are unable to attend follow-up for outpatient CST within 72 h.
• Clinicians should provide a shared decision-making plan with each patient, taking care to discuss the risks and benefits of in-hospital versus outpatient CST.

The risk of MACE should be assessed in all patients presenting to ED with low-risk chest pain to avoid unnecessary hospitalization that exposes them to potential costs and harms with few additional benefits. If the risk scoring system was applied to the patient described in our original clinical scenario, he would have had a HEART score of 3 (ie, 1 point for a moderately suspicious history, 1 point for the age of 60 years, and 1 point for a positive family history) and a TIMI score of 1 (ie, 1 point for aspirin use within past 7 days). Therefore, he could be stratified as having a low-risk presentation. With a second negative troponin I test at 3 h, discharge from ED with timely outpatient CST within 72 h would be an appropriate management strategy.

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

Conflicts of Interest

 The authors have no conflicts of interest relevant to this article to disclose.

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/

Chest pain is one of the most common complaints among patients presenting to the emergency department. Moreover, at least 30% of patients who present with chest pain are admitted for observation, and >70% of those admitted with chest pain undergo cardiac stress testing (CST) during hospitalization. Several clinical risk prediction models have validated evaluation processes for managing patients with chest pain, helping to identify those at a low risk of major adverse cardiac events. Among these, the Thrombolysis in Myocardial Infarction or HEART score can identify patients safe to be discharged with outpatient CST within 72 h. It is unnecessary to hospitalize all low-risk patients for cardiac testing because it may expose them to needless risk and avoidable care costs, with little additional benefit.

A 60-year-old man with a history of osteoarthritis and depression presented to our emergency department (ED) with a 1-month history of left-sided chest pain that was present both at rest and exertion. There were no aggravating or relieving factors for the pain and no associated shortness of breath, diaphoresis, nausea, or lightheadedness. He smoked a half pack of cigarettes daily for 5 years in his twenties. The patient was taking aspirin 81 mg daily and paroxetine 40 mg daily, which he had been taking for 10 years. There was a family history of coronary artery disease in his mother, father, and sister. On examination, he was afebrile, with a blood pressure of 138/78 mm Hg and a heart rate of 62 beats/min; he appeared well, with no abnormal cardiopulmonary findings. Investigation revealed a normal initial troponin I level (<0.034 mg/mL) and normal electrocardiogram (ECG) with normal sinus rhythm (75 beats/min), normal axis, no ST changes, and no Q waves. He was therefore admitted to the hospital for further evaluation.

Each year, >7 million patients visit ED for chest pain in the United States,¹ with approximately 13% diagnosed with acute coronary syndromes (ACSs).² Over 30% of patients who present to ED with chest pain are hospitalized for observation, symptom evaluation, and risk stratification.³ In 2012, the mean Medicare reimbursement cost was $1,741 for in-hospital observation,⁴ with up to 70% of admitted patients undergoing cardiac stress testing (CST) before discharge.⁵

A scientific statement by the American Heart Association in 2010 recommended that patients considered to be at low risk for ACS after initial evaluation (based on presenting symptoms, past history, ECG findings, and initial cardiac biomarkers) should undergo CST within 72 h (preferably within 24 h) of presentation to provoke ischemia or detect anatomic coronary artery disease.⁶ Early exercise treadmill testing as part of an accelerated diagnostic pathway can also reduce the length of stays (LOS) in hospital and lower the medical costs.⁷ Moreover, when there is noncompliance or poor accessibility, failure to pursue early exercise testing in a hospital could result in a loss of patients to follow-up. Hospitalization for testing through accelerated diagnostic pathways may improve access to care and reduce clinical and legal risks associated with a major adverse cardiac event (MACE).

Clinical Risk Prediction Models

When a patient initially presents with chest pain, it should be determined if the symptoms are related to ACS or some other diagnosis. Hospitalization is required for patients with ACS but may not be for those without ACS and those with a low risk of inducible ischemia. Clinical risk scores and risk prediction models, such as the Thrombolysis in Myocardial Infarction (TIMI) and HEART scores, have been used in accelerated diagnostic protocols to determine a patient’s likelihood of having ACS. Several large trials of these clinical risk prediction models have validated the processes for evaluating patients with chest pain.

The TIMI risk score, the most well-known model, assesses risk based on the presence or absence of 7 characteristics (Appendix 1). It should be noted that the patient population studied for initial validation of this model comprised high-risk patients with unstable angina or non-ST elevation myocardial infarction who would benefit from early or urgent invasive therapy.⁸ In this population, TIMI scores of 0-1 are associated with low risk, with a 4.7% risk of ACS at 14 days.⁸ In another study of patients presenting to ED with undifferentiated chest pain and a TIMI score of zero, the risk of MACE at 30 days was approximately 2%.⁹

The HEART score is also used for patients presenting to ED with undifferentiated chest pain and assesses 5 separate variables scored 0–2 (Appendix 2). The original research gave a score of 2 to a troponin I level greater than twice the upper limit of the normal level,¹⁰ whereas a subsequent validation study gave a score of 2 to a troponin I or T level greater than or equal to 3 times the upper limit of the normal level.¹¹ Patients are considered at low, intermediate, and high risk based on scores of 0–3, 4–6, and 7–10, respectively.^10,11 Backus et al. performed a prospective randomized trial of 2388 patients who presented to ED with chest pain to validate the HEART score and compare it to the TIMI risk score. The HEART score performed better than the TIMI risk score in low-risk patients, with TIMI scores of 0-1 and HEART scores of 0–3 having a 6-week MACE risk of 2.8% and 1.7%, respectively.¹¹

A HEART pathway was developed that combines the HEART score with serial troponin I assays assessed at the time of initial presentation and approximately 3 h later.¹² Mahler et al. randomized 282 patients presenting to ED with chest pain to either the HEART pathway or conventional care. Patients with low-risk HEART scores and an abnormal troponin I level were admitted for cardiology consultation, whereas discharge was recommended for those with low scores and a normal troponin I level. Despite nearly 20% of the study cohort having a history of myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting, approximately 40% of patients in the HEART pathway were identified as low risk, increasing early discharge rates by 21.3% and decreasing the average LOS by 12 h. No low-risk patient suffered a MACE within 30 days, and the HEART pathway had a sensitivity and a negative predictive value of approximately 99%.

Costs and Harms of Hospitalization for Cardiac Testing

Hospitalization carries measurable risks.^13,14 Between 2008 and 2013, Weinstock et al. evaluated the outcomes of patients presenting with chest pain who were placed in an observation unit for suspected ACS.¹⁵ Low-risk patients were defined as those with normal ECGs (no ischemic changes), 2 negative troponin tests performed 60–420 min apart (no particular troponin assay specified), and stable vital signs. They identified 7266 patients who were considered to have low risk, among whom 4 (0.06%) had an adverse outcome in the hospital (eg, life-threatening arrhythmia, ST-segment elevation myocardial infarction, cardiac or respiratory arrest, or death); 3 among the 4 patients had a cardiac-related adverse outcome. The overall risk of adverse outcomes due to cardiac causes was 1 in 2422 admissions (0.04%). The authors compared their results with the reported risk of 1 in 164 admissions for preventable adverse events contributing to patient death during routine hospitalization (eg, medication or procedure errors).¹⁴

Outpatient CST can be reliably and safely performed for patients with chest pain.^16-18 There is no clear evidence that earlier CST leads to improved patient outcomes, and CST in the absence of acute ischemia (or ACS) increases the rates of angiography and revascularization without improvements in the rate of myocardial infarction.^19-21 Given the costs of in-hospital observation⁴ and the dubious benefits of providing CST for patients with low-risk chest pain, admitting all patients with low-risk chest pain exposes them to costs and harms with little potential benefit.

Patients presenting with chest pain with either dynamic ECG changes or an elevated troponin level require hospitalization for further ACS diagnosis and treatment. When ACS cannot be clearly diagnosed at the initial evaluation, healthcare providers should use clinical risk prediction models to stratify patients. Those deemed to be at an intermediate or high risk by these models should be hospitalized for further evaluation, as should those at low risk but for whom access to outpatient follow-up is difficult (eg, those without health insurance).

A complete history and physical examination, along with ECG and cardiac biomarker testing, are required for all patients presenting with chest pain. Validated clinical risk prediction models should then be used to determine the likelihood of a cardiac event. Fanaroff et al. reported that low-risk HEART scores of 0–3 and TIMI scores of 0-1 gave positive likelihood ratios of 0.2 and 0.31, respectively.²² Using a pre-test probability of 13%, as reported by Bhuiya et al.,² the likelihood of ACS or MACE within 6 weeks is 2.9% for patients with low-risk HEART scores and 4.4% for those with low-risk TIMI scores.²² These risk prediction models allow clinicians to provide a shared decision-making plan with the patient and discuss the risks and benefits of in-hospital versus outpatient cardiac testing, especially among patients with access to appropriate outpatient follow-up.²³ Low-risk patients can be referred for outpatient testing within 72 h, reducing hospitalization-associated costs and harms.

• Patients presenting with chest pain should undergo a complete history taking and physical examination, as well as ECG and cardiac biomarker testing (eg, troponin I level at presentation and approximately 3 h later).
• Clinical risk prediction models, such as TIMI or HEART scores, should then be used to determine the risk of MACE.
• Patients at a low risk may be safely discharged with outpatient CST performed within 72 h.
• Patients at an intermediate or high risk of MACE should be hospitalized for further evaluation, as should those with low-risk chest pain who are unable to attend follow-up for outpatient CST within 72 h.
• Clinicians should provide a shared decision-making plan with each patient, taking care to discuss the risks and benefits of in-hospital versus outpatient CST.

The risk of MACE should be assessed in all patients presenting to ED with low-risk chest pain to avoid unnecessary hospitalization that exposes them to potential costs and harms with few additional benefits. If the risk scoring system was applied to the patient described in our original clinical scenario, he would have had a HEART score of 3 (ie, 1 point for a moderately suspicious history, 1 point for the age of 60 years, and 1 point for a positive family history) and a TIMI score of 1 (ie, 1 point for aspirin use within past 7 days). Therefore, he could be stratified as having a low-risk presentation. With a second negative troponin I test at 3 h, discharge from ED with timely outpatient CST within 72 h would be an appropriate management strategy.

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

Conflicts of Interest

 The authors have no conflicts of interest relevant to this article to disclose.

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/

Chest pain is one of the most common complaints among patients presenting to the emergency department. Moreover, at least 30% of patients who present with chest pain are admitted for observation, and >70% of those admitted with chest pain undergo cardiac stress testing (CST) during hospitalization. Several clinical risk prediction models have validated evaluation processes for managing patients with chest pain, helping to identify those at a low risk of major adverse cardiac events. Among these, the Thrombolysis in Myocardial Infarction or HEART score can identify patients safe to be discharged with outpatient CST within 72 h. It is unnecessary to hospitalize all low-risk patients for cardiac testing because it may expose them to needless risk and avoidable care costs, with little additional benefit.

A 60-year-old man with a history of osteoarthritis and depression presented to our emergency department (ED) with a 1-month history of left-sided chest pain that was present both at rest and exertion. There were no aggravating or relieving factors for the pain and no associated shortness of breath, diaphoresis, nausea, or lightheadedness. He smoked a half pack of cigarettes daily for 5 years in his twenties. The patient was taking aspirin 81 mg daily and paroxetine 40 mg daily, which he had been taking for 10 years. There was a family history of coronary artery disease in his mother, father, and sister. On examination, he was afebrile, with a blood pressure of 138/78 mm Hg and a heart rate of 62 beats/min; he appeared well, with no abnormal cardiopulmonary findings. Investigation revealed a normal initial troponin I level (<0.034 mg/mL) and normal electrocardiogram (ECG) with normal sinus rhythm (75 beats/min), normal axis, no ST changes, and no Q waves. He was therefore admitted to the hospital for further evaluation.

Each year, >7 million patients visit ED for chest pain in the United States,¹ with approximately 13% diagnosed with acute coronary syndromes (ACSs).² Over 30% of patients who present to ED with chest pain are hospitalized for observation, symptom evaluation, and risk stratification.³ In 2012, the mean Medicare reimbursement cost was $1,741 for in-hospital observation,⁴ with up to 70% of admitted patients undergoing cardiac stress testing (CST) before discharge.⁵

A scientific statement by the American Heart Association in 2010 recommended that patients considered to be at low risk for ACS after initial evaluation (based on presenting symptoms, past history, ECG findings, and initial cardiac biomarkers) should undergo CST within 72 h (preferably within 24 h) of presentation to provoke ischemia or detect anatomic coronary artery disease.⁶ Early exercise treadmill testing as part of an accelerated diagnostic pathway can also reduce the length of stays (LOS) in hospital and lower the medical costs.⁷ Moreover, when there is noncompliance or poor accessibility, failure to pursue early exercise testing in a hospital could result in a loss of patients to follow-up. Hospitalization for testing through accelerated diagnostic pathways may improve access to care and reduce clinical and legal risks associated with a major adverse cardiac event (MACE).

Clinical Risk Prediction Models

When a patient initially presents with chest pain, it should be determined if the symptoms are related to ACS or some other diagnosis. Hospitalization is required for patients with ACS but may not be for those without ACS and those with a low risk of inducible ischemia. Clinical risk scores and risk prediction models, such as the Thrombolysis in Myocardial Infarction (TIMI) and HEART scores, have been used in accelerated diagnostic protocols to determine a patient’s likelihood of having ACS. Several large trials of these clinical risk prediction models have validated the processes for evaluating patients with chest pain.

The TIMI risk score, the most well-known model, assesses risk based on the presence or absence of 7 characteristics (Appendix 1). It should be noted that the patient population studied for initial validation of this model comprised high-risk patients with unstable angina or non-ST elevation myocardial infarction who would benefit from early or urgent invasive therapy.⁸ In this population, TIMI scores of 0-1 are associated with low risk, with a 4.7% risk of ACS at 14 days.⁸ In another study of patients presenting to ED with undifferentiated chest pain and a TIMI score of zero, the risk of MACE at 30 days was approximately 2%.⁹

The HEART score is also used for patients presenting to ED with undifferentiated chest pain and assesses 5 separate variables scored 0–2 (Appendix 2). The original research gave a score of 2 to a troponin I level greater than twice the upper limit of the normal level,¹⁰ whereas a subsequent validation study gave a score of 2 to a troponin I or T level greater than or equal to 3 times the upper limit of the normal level.¹¹ Patients are considered at low, intermediate, and high risk based on scores of 0–3, 4–6, and 7–10, respectively.^10,11 Backus et al. performed a prospective randomized trial of 2388 patients who presented to ED with chest pain to validate the HEART score and compare it to the TIMI risk score. The HEART score performed better than the TIMI risk score in low-risk patients, with TIMI scores of 0-1 and HEART scores of 0–3 having a 6-week MACE risk of 2.8% and 1.7%, respectively.¹¹

A HEART pathway was developed that combines the HEART score with serial troponin I assays assessed at the time of initial presentation and approximately 3 h later.¹² Mahler et al. randomized 282 patients presenting to ED with chest pain to either the HEART pathway or conventional care. Patients with low-risk HEART scores and an abnormal troponin I level were admitted for cardiology consultation, whereas discharge was recommended for those with low scores and a normal troponin I level. Despite nearly 20% of the study cohort having a history of myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting, approximately 40% of patients in the HEART pathway were identified as low risk, increasing early discharge rates by 21.3% and decreasing the average LOS by 12 h. No low-risk patient suffered a MACE within 30 days, and the HEART pathway had a sensitivity and a negative predictive value of approximately 99%.

Costs and Harms of Hospitalization for Cardiac Testing

Hospitalization carries measurable risks.^13,14 Between 2008 and 2013, Weinstock et al. evaluated the outcomes of patients presenting with chest pain who were placed in an observation unit for suspected ACS.¹⁵ Low-risk patients were defined as those with normal ECGs (no ischemic changes), 2 negative troponin tests performed 60–420 min apart (no particular troponin assay specified), and stable vital signs. They identified 7266 patients who were considered to have low risk, among whom 4 (0.06%) had an adverse outcome in the hospital (eg, life-threatening arrhythmia, ST-segment elevation myocardial infarction, cardiac or respiratory arrest, or death); 3 among the 4 patients had a cardiac-related adverse outcome. The overall risk of adverse outcomes due to cardiac causes was 1 in 2422 admissions (0.04%). The authors compared their results with the reported risk of 1 in 164 admissions for preventable adverse events contributing to patient death during routine hospitalization (eg, medication or procedure errors).¹⁴

Outpatient CST can be reliably and safely performed for patients with chest pain.^16-18 There is no clear evidence that earlier CST leads to improved patient outcomes, and CST in the absence of acute ischemia (or ACS) increases the rates of angiography and revascularization without improvements in the rate of myocardial infarction.^19-21 Given the costs of in-hospital observation⁴ and the dubious benefits of providing CST for patients with low-risk chest pain, admitting all patients with low-risk chest pain exposes them to costs and harms with little potential benefit.

Patients presenting with chest pain with either dynamic ECG changes or an elevated troponin level require hospitalization for further ACS diagnosis and treatment. When ACS cannot be clearly diagnosed at the initial evaluation, healthcare providers should use clinical risk prediction models to stratify patients. Those deemed to be at an intermediate or high risk by these models should be hospitalized for further evaluation, as should those at low risk but for whom access to outpatient follow-up is difficult (eg, those without health insurance).

A complete history and physical examination, along with ECG and cardiac biomarker testing, are required for all patients presenting with chest pain. Validated clinical risk prediction models should then be used to determine the likelihood of a cardiac event. Fanaroff et al. reported that low-risk HEART scores of 0–3 and TIMI scores of 0-1 gave positive likelihood ratios of 0.2 and 0.31, respectively.²² Using a pre-test probability of 13%, as reported by Bhuiya et al.,² the likelihood of ACS or MACE within 6 weeks is 2.9% for patients with low-risk HEART scores and 4.4% for those with low-risk TIMI scores.²² These risk prediction models allow clinicians to provide a shared decision-making plan with the patient and discuss the risks and benefits of in-hospital versus outpatient cardiac testing, especially among patients with access to appropriate outpatient follow-up.²³ Low-risk patients can be referred for outpatient testing within 72 h, reducing hospitalization-associated costs and harms.

• Patients presenting with chest pain should undergo a complete history taking and physical examination, as well as ECG and cardiac biomarker testing (eg, troponin I level at presentation and approximately 3 h later).
• Clinical risk prediction models, such as TIMI or HEART scores, should then be used to determine the risk of MACE.
• Patients at a low risk may be safely discharged with outpatient CST performed within 72 h.
• Patients at an intermediate or high risk of MACE should be hospitalized for further evaluation, as should those with low-risk chest pain who are unable to attend follow-up for outpatient CST within 72 h.
• Clinicians should provide a shared decision-making plan with each patient, taking care to discuss the risks and benefits of in-hospital versus outpatient CST.

The risk of MACE should be assessed in all patients presenting to ED with low-risk chest pain to avoid unnecessary hospitalization that exposes them to potential costs and harms with few additional benefits. If the risk scoring system was applied to the patient described in our original clinical scenario, he would have had a HEART score of 3 (ie, 1 point for a moderately suspicious history, 1 point for the age of 60 years, and 1 point for a positive family history) and a TIMI score of 1 (ie, 1 point for aspirin use within past 7 days). Therefore, he could be stratified as having a low-risk presentation. With a second negative troponin I test at 3 h, discharge from ED with timely outpatient CST within 72 h would be an appropriate management strategy.

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

Conflicts of Interest

 The authors have no conflicts of interest relevant to this article to disclose.

Most studies to identify risk factors for readmission among patients with heart failure (HF) have focused on demographic and clinical characteristics.^1,2 Although easy to extract from administrative databases, this approach fails to capture the complex psychosocial and cognitive factors that influence the ability of HF patients to manage their disease in the postdischarge period, as depicted in the framework by Meyers et al.³ (2014). To date, studies have found low health literacy, decreased social support, and cognitive impairment to be associated with health behaviors and outcomes among HF patients, including decreased self-care,⁴ low HF-specific knowledge,⁵ medication nonadherence,⁶ hospitalizations,⁷ and mortality.^8-10 Less, however, is known about the effect of numeracy on HF outcomes, such as 30-day readmission.

Numeracy, or quantitative literacy, refers to the ability to access, understand, and apply numerical data to health-related decisions.¹¹ It is estimated that 110 million people in the United States have limited numeracy skills.¹² Low numeracy is a risk factor for poor glycemic control among patients with diabetes,¹³ medication adherence in HIV/AIDS,¹⁴ and worse blood pressure control in hypertensives.¹⁵ Much like these conditions, HF requires that patients understand, use, and act on numerical information. Maintaining a low-salt diet, monitoring weight, adjusting diuretic doses, and measuring blood pressure are tasks that HF patients are asked to perform on a daily or near-daily basis. These tasks are particularly important in the posthospitalization period and could be complicated by medication changes, which might create additional challenges for patients with inadequate numeracy. Additionally, cognitive impairment, which is a highly prevalent comorbid condition among adults with HF,^16,17 might impose additional barriers for those with inadequate numeracy who do not have adequate social support. However, to date, numeracy in the context of HF has not been well described.

Herein, we examined the effects of numeracy, alongside health literacy and cognition, on 30-day readmission risk among patients hospitalized for acute decompensated HF (ADHF).

Study Design

The Vanderbilt Inpatient Cohort Study (VICS) is a prospective observational study of patients admitted with cardiovascular disease to Vanderbilt University Medical Center (VUMC), an academic tertiary care hospital. VICS was designed to investigate the impact of social determinants of health on postdischarge health outcomes. A detailed description of the study rationale, design, and methods is described elsewhere.³

Briefly, participants completed a baseline interview while hospitalized, and follow-up phone calls were conducted within 1 week of discharge, at 30 days, and at 90 days. At 30 and 90 days postdischarge, healthcare utilization was ascertained by review of medical records and patient report. Clinical data about the index hospitalization were also abstracted. The Vanderbilt University Institutional Review Board approved the study.

Study Population

Patients hospitalized from 2011 to 2015 with a likely diagnosis of acute coronary syndrome and/or ADHF, as determined by a physician’s review of the medical record, were identified as potentially eligible. Research assistants assessed these patients for the presence of the following exclusion criteria: less than 18 years of age, non-English speaking, unstable psychiatric illness, a low likelihood of follow-up (eg, no reliable telephone number), on hospice, or otherwise too ill to complete an interview. Additionally, those with severe cognitive impairment, as assessed from the medical record (such as seeing a note describing dementia), and those with delirium, as assessed by the brief confusion assessment method, were excluded from enrollment in the study.^18,19 Those who died before discharge or during the 30-day follow-up period were excluded. For this analysis, we restricted our sample to only include participants who were hospitalized for ADHF.

Outcome Measure: 30-Day Readmission

The main outcome was all-cause readmission to any hospital within 30 days of discharge, as determined by patient interview, review of electronic medical records from VUMC, and review of outside hospital records.

Main Exposures: Numeracy, Health Literacy, and Cognitive Impairment

Numeracy was assessed with a 3-item version of the Subjective Numeracy Scale (SNS-3), which quantifies the patients perceived quantitative abilities.²⁰ Other authors have shown that the SNS-3 has a correlation coefficient of 0.88 with the full-length SNS-8 and a Cronbach’s alpha of 0.78.^20-22 The SNS-3 is reported as the mean on a scale from 1 to 6, with higher scores reflecting higher numeracy.

Subjective health literacy was assessed by using the 3-item Brief Health Literacy Screen (BHLS).²³ Scores range from 3 to 15, with higher scores reflecting higher literacy. Objective health literacy was assessed with the short form of the Test of Functional Health Literacy in Adults (sTOFHLA).^24,25 Scores may be categorized as inadequate (0-16), marginal (17-22), or adequate (23-36).

We assessed cognition by using the 10-item Short Portable Mental Status Questionnaire (SPMSQ).²⁶ The SPMSQ, which describes a person’s capacity for memory, structured thought, and orientation, has been validated and has demonstrated good reliability and validity.²⁷ Scores of 0 were considered to reflect intact cognition, and scores of 1 or more were considered to reflect any cognitive impairment, a scoring approach employed by other authors.²⁸ We used this approach, rather than the traditional scoring system developed by Pfeiffer et al.²⁶ (1975), because it would be the most sensitive to detect any cognitive impairment in the VICS cohort, which excluded those with severe cognition impairment, dementia, and delirium.

Covariates

During the hospitalization, participants completed an in-person interviewer-administered baseline assessment composed of demographic information, including age, self-reported race (white and nonwhite), educational attainment, home status (married, not married and living with someone, not married and living alone), and household income.

Clinical and diagnostic characteristics abstracted from the medical record included a medical history of HF, HF subtype (classified by left ventricular ejection fraction [LVEF]), coronary artery disease, chronic obstructive pulmonary disease (COPD), diabetes mellitus (DM), and comorbidity burden as summarized by the van Walraven-Elixhauser score.^29,30 Depressive symptoms were assessed during the 2 weeks prior to the hospitalization by using the first 8 items of the Patient Health Questionnaire.³¹ Scores ranged from 0 to 24, with higher scores reflecting more severe depressive symptoms. Laboratory values included estimated glomerular filtration rate (eGFR), hemoglobin (g/dl), sodium (mg/L), and brain natriuretic peptide (BNP) (pg/ml) from the last laboratory draw before discharge. Smoking status was also assessed (current and former/nonsmokers).

Hospitalization characteristics included length of stay in days, number of prior admissions in the last year, and transfer to the intensive care unit during the index admission.

Statistical Analysis

Descriptive statistics were used to summarize patient characteristics. The Kruskal-Wallis test and the Pearson χ² test were used to determine the association between patient characteristics and levels of numeracy, literacy, and cognition separately. The unadjusted relationship between patient characteristics and 30-day readmission was assessed by using Wilcoxon rank sums tests for continuous variables and Pearson χ² tests for categorical variables. In addition, a correlation matrix was performed to assess the correlations between numeracy, health literacy, and cognition (supplementary Figure 1).

To examine the association between numeracy, health literacy, and cognition and 30-day readmissions, a series of multivariable Poisson (log-linear) regression models were fit.³² Like other studies, numeracy, health literacy, and cognition were examined as categorical and continuous measures in models.³³ Each model was modified with a sandwich estimator for robust standard errors. Log-linear models were chosen over logistic regression models for ease of interpretation because (exponentiated) parameters correspond to risk ratios (RRs) as opposed to odds ratios. Furthermore, the fitting challenges associated with log-linear models when predicted probabilities are near 0 or 1 were not present in these analyses. Redundancy analyses were conducted to ensure that independent variables were not highly correlated with a linear combination of the other independent variables. To avoid case-wise deletion of records with missing covariates, we employed multiple imputation with 10 imputation samples by using predictive mean matching.^34,35 All analyses were conducted in R version 3.1.2 (The R Foundation, Vienna, Austria).³⁶

Overall, 883 patients were included in this analysis (supplementary Figure 2). Of the 883 participants, 46% were female and 76% were white (Table 1). Their median age was 60 years (interdecile range [IDR] 39-78) and the median educational attainment was 13.5 years (IDR 11-18).

This is the first study to examine the effect of numeracy alongside literacy and cognition on 30-day readmission risk among patients hospitalized with ADHF. Overall, we found that 33.9% of participants had inadequate numeracy skills, and 24.6% had inadequate or marginal health literacy. In unadjusted and adjusted models, numeracy was not associated with 30-day readmission. Although (objective) low health literacy was associated with 30-day readmission in unadjusted models, it was not in adjusted models. Additionally, though 53% of participants had any cognitive impairment, readmission did not differ significantly by this factor. Taken together, these findings suggest that other factors may be greater determinants of 30-day readmissions among patients hospitalized for ADHF.

Only 1 other study has examined the effect of numeracy on readmission risk among patients hospitalized for HF. In this multicenter prospective study, McNaughton et al.³⁷ found low numeracy to be associated with higher odds of recidivism to the emergency department (ED) or hospital within 30 days. Our findings may differ from theirs for a few reasons. First, their study had a significantly higher percentage of individuals with low numeracy (55%) compared with ours (33.9%). This may be because they did not exclude individuals with severe cognitive impairment, and their patient population was of lower socioeconomic status (SES) than ours. Low SES is associated with higher 30-day readmissions among HF patients^1,10 throughout the literature, and low numeracy is associated with low SES in other diseases.^13,38,39 Finally, they studied recidivism, which was defined as any unplanned return to the ED or hospital within 30 days of the index ED visit for acute HF. We only focused on 30-day readmissions, which also may explain why our results differed.

We found that health literacy was not associated with 30-day readmissions, which is consistent with the literature. Although an association between health literacy and mortality exists among adults with HF, several studies have not found an association between health literacy and 30- and 90-day readmission among adults hospitalized for HF.^8,9,40 Although we found an association between objective health literacy and 30-day readmission in unadjusted analyses, we did not find one in the multivariable model. This, along with our numeracy finding, suggests that numeracy and literacy may not be driving the 30-day readmission risk among patients hospitalized with ADHF.

We examined cognition alongside numeracy and literacy because it is a prevalent condition among HF patients and because it is associated with adverse outcomes among patients with HF, including readmission.^41,42 Studies have shown that HF preferentially affects certain cognitive domains,⁴³ some of which are vital to HF self-care activities. We found that 53% of patients had any cognitive impairment, which is consistent with the literature of adults hospitalized for ADHF.^44,45 Cognitive impairment was not, however, associated with 30-day readmissions. There may be a couple reasons for this. First, we measured cognitive impairment with the SPMSQ, which, although widely used and well-validated, does not assess executive function, the domain most commonly affected in HF patients with cognitive impairment.⁴⁶ Second, patients with severe cognitive impairment and those with delirium were excluded from this study, which may have limited our ability to detect differences in readmission by this factor.

As in prior studies, we found that a history of DM and more hospitalizations in the prior year were independently associated with 30-day readmissions in fully adjusted models. Like other studies, in adjusted models, we found that LVEF and a history of HF were not independently associated with 30-day readmission.^47-49 This, however, is not surprising because recent studies have shown that, although HF patients are at risk for multiple hospitalizations, early readmission after a hospitalization for ADHF specifically is often because of reasons unrelated to HF or a non-cardiovascular cause in general.^50,51

Although a negative study, several important themes emerged. First, while we were able to assess numeracy, health literacy, and cognition, none of these measures were HF-specific. It is possible that we did not see an effect on readmission because our instruments failed to assess domains specific to HF, such as monitoring weight changes, following a low-salt diet, and interpreting blood pressure. Currently, however, no HF-specific objective numeracy measure exists. With respect to health literacy, only 1 HF-specific measure exists,⁵² although it was only recently developed and validated. Second, while numeracy may not be a driving influence of all-cause 30-day readmissions, it may be associated with other health behaviors and quality metrics that we did not examine here, such as self-care, medication adherence, and HF-specific readmissions. Third, it is likely that the progression of HF itself, as well as the clinical management of patients following discharge, contribute significantly to 30-day readmissions. Increased attention to predischarge processes for HF patients occurred at VUMC during the study period; close follow-up and evidence-directed therapies may have mitigated some of the expected associations. Finally, we were not able to assess numeracy of participants’ primary caregivers who may help patients at home, especially postdischarge. Though a number of studies have examined the role of family caregivers in the management of HF,^53,54 none have examined numeracy levels of caregivers in the context of HF, and this may be worth doing in future studies.

Overall, our study has several strengths. The size of the cohort is large and there were high response rates during the follow-up period. Unlike other HF readmission studies, VICS accounts for readmissions to outside hospitals. Approximately 35% of all hospitalizations in VICS are to outside facilities. Thus, the ascertainment of readmissions to hospitals other than Vanderbilt is more comprehensive than if readmissions to VUMC were only considered. We were able to include a number of clinical comorbidities, laboratory and diagnostic tests from the index admission, and hospitalization characteristics in our analyses. Finally, we performed additional analyses to investigate the correlation between numeracy, literacy, and cognition; ultimately, we found that the majority of these correlations were weak, which supports our ability to study them simultaneously among VICS participants.

Nonetheless, we note some limitations. Although we captured readmissions to outside hospitals, the study took place at a single referral center in Tennessee. Though patients were diverse in age and comorbidities, they were mostly white and of higher SES. Finally, we used home status as a proxy for social support, which may underestimate the support that home care workers provide.

In conclusion, in this prospective longitudinal study of adults hospitalized with ADHF, inadequate numeracy was present in more than a third of patients, and low health literacy was present in roughly a quarter of patients. Neither numeracy nor health literacy, however, were associated with 30-day readmissions in adjusted analyses. Any cognitive impairment, although present in roughly one-half of patients, was not associated with 30-day readmission either. Our findings suggest that other influences may play a more dominant role in determining 30-day readmission rates in patients hospitalized for ADHF than inadequate numeracy, low health literacy, or cognitive impairment as assessed here.

Acknowledgments

This research was supported by the National Heart, Lung, and Blood Institute (R01 HL109388) and in part by the National Center for Advancing Translational Sciences (UL1 TR000445-06). The content is solely the responsibility of the authors and does not necessarily represent official views of the National Institutes of Health. The authors’ funding sources did not participate in the planning, collection, analysis, or interpretation of data or in the decision to submit for publication. Dr. Sterling is supported by T32HS000066 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality. Dr. Mixon has a VA Health Services Research and Development Service Career Development Award at the Tennessee Valley Healthcare System, Department of Veterans Affairs (CDA 12-168). This material was presented at the Society of General Internal Medicine Annual Meeting on April 20, 2017, in Washington, DC.

Disclosure

Dr. Kripalani reports personal fees from Verustat, personal fees from SAI Interactive, and equity from Bioscape Digital, all outside of the submitted work. Dr. Rothman and Dr. Wallston report personal fees from EdLogics outside of the submitted work. All of the other authors have nothing to disclose

Most studies to identify risk factors for readmission among patients with heart failure (HF) have focused on demographic and clinical characteristics.^1,2 Although easy to extract from administrative databases, this approach fails to capture the complex psychosocial and cognitive factors that influence the ability of HF patients to manage their disease in the postdischarge period, as depicted in the framework by Meyers et al.³ (2014). To date, studies have found low health literacy, decreased social support, and cognitive impairment to be associated with health behaviors and outcomes among HF patients, including decreased self-care,⁴ low HF-specific knowledge,⁵ medication nonadherence,⁶ hospitalizations,⁷ and mortality.^8-10 Less, however, is known about the effect of numeracy on HF outcomes, such as 30-day readmission.

Numeracy, or quantitative literacy, refers to the ability to access, understand, and apply numerical data to health-related decisions.¹¹ It is estimated that 110 million people in the United States have limited numeracy skills.¹² Low numeracy is a risk factor for poor glycemic control among patients with diabetes,¹³ medication adherence in HIV/AIDS,¹⁴ and worse blood pressure control in hypertensives.¹⁵ Much like these conditions, HF requires that patients understand, use, and act on numerical information. Maintaining a low-salt diet, monitoring weight, adjusting diuretic doses, and measuring blood pressure are tasks that HF patients are asked to perform on a daily or near-daily basis. These tasks are particularly important in the posthospitalization period and could be complicated by medication changes, which might create additional challenges for patients with inadequate numeracy. Additionally, cognitive impairment, which is a highly prevalent comorbid condition among adults with HF,^16,17 might impose additional barriers for those with inadequate numeracy who do not have adequate social support. However, to date, numeracy in the context of HF has not been well described.

Herein, we examined the effects of numeracy, alongside health literacy and cognition, on 30-day readmission risk among patients hospitalized for acute decompensated HF (ADHF).

Study Design

The Vanderbilt Inpatient Cohort Study (VICS) is a prospective observational study of patients admitted with cardiovascular disease to Vanderbilt University Medical Center (VUMC), an academic tertiary care hospital. VICS was designed to investigate the impact of social determinants of health on postdischarge health outcomes. A detailed description of the study rationale, design, and methods is described elsewhere.³

Briefly, participants completed a baseline interview while hospitalized, and follow-up phone calls were conducted within 1 week of discharge, at 30 days, and at 90 days. At 30 and 90 days postdischarge, healthcare utilization was ascertained by review of medical records and patient report. Clinical data about the index hospitalization were also abstracted. The Vanderbilt University Institutional Review Board approved the study.

Study Population

Patients hospitalized from 2011 to 2015 with a likely diagnosis of acute coronary syndrome and/or ADHF, as determined by a physician’s review of the medical record, were identified as potentially eligible. Research assistants assessed these patients for the presence of the following exclusion criteria: less than 18 years of age, non-English speaking, unstable psychiatric illness, a low likelihood of follow-up (eg, no reliable telephone number), on hospice, or otherwise too ill to complete an interview. Additionally, those with severe cognitive impairment, as assessed from the medical record (such as seeing a note describing dementia), and those with delirium, as assessed by the brief confusion assessment method, were excluded from enrollment in the study.^18,19 Those who died before discharge or during the 30-day follow-up period were excluded. For this analysis, we restricted our sample to only include participants who were hospitalized for ADHF.

Outcome Measure: 30-Day Readmission

The main outcome was all-cause readmission to any hospital within 30 days of discharge, as determined by patient interview, review of electronic medical records from VUMC, and review of outside hospital records.

Main Exposures: Numeracy, Health Literacy, and Cognitive Impairment

Numeracy was assessed with a 3-item version of the Subjective Numeracy Scale (SNS-3), which quantifies the patients perceived quantitative abilities.²⁰ Other authors have shown that the SNS-3 has a correlation coefficient of 0.88 with the full-length SNS-8 and a Cronbach’s alpha of 0.78.^20-22 The SNS-3 is reported as the mean on a scale from 1 to 6, with higher scores reflecting higher numeracy.

Subjective health literacy was assessed by using the 3-item Brief Health Literacy Screen (BHLS).²³ Scores range from 3 to 15, with higher scores reflecting higher literacy. Objective health literacy was assessed with the short form of the Test of Functional Health Literacy in Adults (sTOFHLA).^24,25 Scores may be categorized as inadequate (0-16), marginal (17-22), or adequate (23-36).

We assessed cognition by using the 10-item Short Portable Mental Status Questionnaire (SPMSQ).²⁶ The SPMSQ, which describes a person’s capacity for memory, structured thought, and orientation, has been validated and has demonstrated good reliability and validity.²⁷ Scores of 0 were considered to reflect intact cognition, and scores of 1 or more were considered to reflect any cognitive impairment, a scoring approach employed by other authors.²⁸ We used this approach, rather than the traditional scoring system developed by Pfeiffer et al.²⁶ (1975), because it would be the most sensitive to detect any cognitive impairment in the VICS cohort, which excluded those with severe cognition impairment, dementia, and delirium.

Covariates

During the hospitalization, participants completed an in-person interviewer-administered baseline assessment composed of demographic information, including age, self-reported race (white and nonwhite), educational attainment, home status (married, not married and living with someone, not married and living alone), and household income.

Clinical and diagnostic characteristics abstracted from the medical record included a medical history of HF, HF subtype (classified by left ventricular ejection fraction [LVEF]), coronary artery disease, chronic obstructive pulmonary disease (COPD), diabetes mellitus (DM), and comorbidity burden as summarized by the van Walraven-Elixhauser score.^29,30 Depressive symptoms were assessed during the 2 weeks prior to the hospitalization by using the first 8 items of the Patient Health Questionnaire.³¹ Scores ranged from 0 to 24, with higher scores reflecting more severe depressive symptoms. Laboratory values included estimated glomerular filtration rate (eGFR), hemoglobin (g/dl), sodium (mg/L), and brain natriuretic peptide (BNP) (pg/ml) from the last laboratory draw before discharge. Smoking status was also assessed (current and former/nonsmokers).

Hospitalization characteristics included length of stay in days, number of prior admissions in the last year, and transfer to the intensive care unit during the index admission.

Statistical Analysis

Descriptive statistics were used to summarize patient characteristics. The Kruskal-Wallis test and the Pearson χ² test were used to determine the association between patient characteristics and levels of numeracy, literacy, and cognition separately. The unadjusted relationship between patient characteristics and 30-day readmission was assessed by using Wilcoxon rank sums tests for continuous variables and Pearson χ² tests for categorical variables. In addition, a correlation matrix was performed to assess the correlations between numeracy, health literacy, and cognition (supplementary Figure 1).

To examine the association between numeracy, health literacy, and cognition and 30-day readmissions, a series of multivariable Poisson (log-linear) regression models were fit.³² Like other studies, numeracy, health literacy, and cognition were examined as categorical and continuous measures in models.³³ Each model was modified with a sandwich estimator for robust standard errors. Log-linear models were chosen over logistic regression models for ease of interpretation because (exponentiated) parameters correspond to risk ratios (RRs) as opposed to odds ratios. Furthermore, the fitting challenges associated with log-linear models when predicted probabilities are near 0 or 1 were not present in these analyses. Redundancy analyses were conducted to ensure that independent variables were not highly correlated with a linear combination of the other independent variables. To avoid case-wise deletion of records with missing covariates, we employed multiple imputation with 10 imputation samples by using predictive mean matching.^34,35 All analyses were conducted in R version 3.1.2 (The R Foundation, Vienna, Austria).³⁶

Overall, 883 patients were included in this analysis (supplementary Figure 2). Of the 883 participants, 46% were female and 76% were white (Table 1). Their median age was 60 years (interdecile range [IDR] 39-78) and the median educational attainment was 13.5 years (IDR 11-18).

This is the first study to examine the effect of numeracy alongside literacy and cognition on 30-day readmission risk among patients hospitalized with ADHF. Overall, we found that 33.9% of participants had inadequate numeracy skills, and 24.6% had inadequate or marginal health literacy. In unadjusted and adjusted models, numeracy was not associated with 30-day readmission. Although (objective) low health literacy was associated with 30-day readmission in unadjusted models, it was not in adjusted models. Additionally, though 53% of participants had any cognitive impairment, readmission did not differ significantly by this factor. Taken together, these findings suggest that other factors may be greater determinants of 30-day readmissions among patients hospitalized for ADHF.

Only 1 other study has examined the effect of numeracy on readmission risk among patients hospitalized for HF. In this multicenter prospective study, McNaughton et al.³⁷ found low numeracy to be associated with higher odds of recidivism to the emergency department (ED) or hospital within 30 days. Our findings may differ from theirs for a few reasons. First, their study had a significantly higher percentage of individuals with low numeracy (55%) compared with ours (33.9%). This may be because they did not exclude individuals with severe cognitive impairment, and their patient population was of lower socioeconomic status (SES) than ours. Low SES is associated with higher 30-day readmissions among HF patients^1,10 throughout the literature, and low numeracy is associated with low SES in other diseases.^13,38,39 Finally, they studied recidivism, which was defined as any unplanned return to the ED or hospital within 30 days of the index ED visit for acute HF. We only focused on 30-day readmissions, which also may explain why our results differed.

We found that health literacy was not associated with 30-day readmissions, which is consistent with the literature. Although an association between health literacy and mortality exists among adults with HF, several studies have not found an association between health literacy and 30- and 90-day readmission among adults hospitalized for HF.^8,9,40 Although we found an association between objective health literacy and 30-day readmission in unadjusted analyses, we did not find one in the multivariable model. This, along with our numeracy finding, suggests that numeracy and literacy may not be driving the 30-day readmission risk among patients hospitalized with ADHF.

We examined cognition alongside numeracy and literacy because it is a prevalent condition among HF patients and because it is associated with adverse outcomes among patients with HF, including readmission.^41,42 Studies have shown that HF preferentially affects certain cognitive domains,⁴³ some of which are vital to HF self-care activities. We found that 53% of patients had any cognitive impairment, which is consistent with the literature of adults hospitalized for ADHF.^44,45 Cognitive impairment was not, however, associated with 30-day readmissions. There may be a couple reasons for this. First, we measured cognitive impairment with the SPMSQ, which, although widely used and well-validated, does not assess executive function, the domain most commonly affected in HF patients with cognitive impairment.⁴⁶ Second, patients with severe cognitive impairment and those with delirium were excluded from this study, which may have limited our ability to detect differences in readmission by this factor.

As in prior studies, we found that a history of DM and more hospitalizations in the prior year were independently associated with 30-day readmissions in fully adjusted models. Like other studies, in adjusted models, we found that LVEF and a history of HF were not independently associated with 30-day readmission.^47-49 This, however, is not surprising because recent studies have shown that, although HF patients are at risk for multiple hospitalizations, early readmission after a hospitalization for ADHF specifically is often because of reasons unrelated to HF or a non-cardiovascular cause in general.^50,51

Although a negative study, several important themes emerged. First, while we were able to assess numeracy, health literacy, and cognition, none of these measures were HF-specific. It is possible that we did not see an effect on readmission because our instruments failed to assess domains specific to HF, such as monitoring weight changes, following a low-salt diet, and interpreting blood pressure. Currently, however, no HF-specific objective numeracy measure exists. With respect to health literacy, only 1 HF-specific measure exists,⁵² although it was only recently developed and validated. Second, while numeracy may not be a driving influence of all-cause 30-day readmissions, it may be associated with other health behaviors and quality metrics that we did not examine here, such as self-care, medication adherence, and HF-specific readmissions. Third, it is likely that the progression of HF itself, as well as the clinical management of patients following discharge, contribute significantly to 30-day readmissions. Increased attention to predischarge processes for HF patients occurred at VUMC during the study period; close follow-up and evidence-directed therapies may have mitigated some of the expected associations. Finally, we were not able to assess numeracy of participants’ primary caregivers who may help patients at home, especially postdischarge. Though a number of studies have examined the role of family caregivers in the management of HF,^53,54 none have examined numeracy levels of caregivers in the context of HF, and this may be worth doing in future studies.

Overall, our study has several strengths. The size of the cohort is large and there were high response rates during the follow-up period. Unlike other HF readmission studies, VICS accounts for readmissions to outside hospitals. Approximately 35% of all hospitalizations in VICS are to outside facilities. Thus, the ascertainment of readmissions to hospitals other than Vanderbilt is more comprehensive than if readmissions to VUMC were only considered. We were able to include a number of clinical comorbidities, laboratory and diagnostic tests from the index admission, and hospitalization characteristics in our analyses. Finally, we performed additional analyses to investigate the correlation between numeracy, literacy, and cognition; ultimately, we found that the majority of these correlations were weak, which supports our ability to study them simultaneously among VICS participants.

Nonetheless, we note some limitations. Although we captured readmissions to outside hospitals, the study took place at a single referral center in Tennessee. Though patients were diverse in age and comorbidities, they were mostly white and of higher SES. Finally, we used home status as a proxy for social support, which may underestimate the support that home care workers provide.

In conclusion, in this prospective longitudinal study of adults hospitalized with ADHF, inadequate numeracy was present in more than a third of patients, and low health literacy was present in roughly a quarter of patients. Neither numeracy nor health literacy, however, were associated with 30-day readmissions in adjusted analyses. Any cognitive impairment, although present in roughly one-half of patients, was not associated with 30-day readmission either. Our findings suggest that other influences may play a more dominant role in determining 30-day readmission rates in patients hospitalized for ADHF than inadequate numeracy, low health literacy, or cognitive impairment as assessed here.

Acknowledgments

This research was supported by the National Heart, Lung, and Blood Institute (R01 HL109388) and in part by the National Center for Advancing Translational Sciences (UL1 TR000445-06). The content is solely the responsibility of the authors and does not necessarily represent official views of the National Institutes of Health. The authors’ funding sources did not participate in the planning, collection, analysis, or interpretation of data or in the decision to submit for publication. Dr. Sterling is supported by T32HS000066 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality. Dr. Mixon has a VA Health Services Research and Development Service Career Development Award at the Tennessee Valley Healthcare System, Department of Veterans Affairs (CDA 12-168). This material was presented at the Society of General Internal Medicine Annual Meeting on April 20, 2017, in Washington, DC.

Disclosure

Dr. Kripalani reports personal fees from Verustat, personal fees from SAI Interactive, and equity from Bioscape Digital, all outside of the submitted work. Dr. Rothman and Dr. Wallston report personal fees from EdLogics outside of the submitted work. All of the other authors have nothing to disclose

Most studies to identify risk factors for readmission among patients with heart failure (HF) have focused on demographic and clinical characteristics.^1,2 Although easy to extract from administrative databases, this approach fails to capture the complex psychosocial and cognitive factors that influence the ability of HF patients to manage their disease in the postdischarge period, as depicted in the framework by Meyers et al.³ (2014). To date, studies have found low health literacy, decreased social support, and cognitive impairment to be associated with health behaviors and outcomes among HF patients, including decreased self-care,⁴ low HF-specific knowledge,⁵ medication nonadherence,⁶ hospitalizations,⁷ and mortality.^8-10 Less, however, is known about the effect of numeracy on HF outcomes, such as 30-day readmission.

Numeracy, or quantitative literacy, refers to the ability to access, understand, and apply numerical data to health-related decisions.¹¹ It is estimated that 110 million people in the United States have limited numeracy skills.¹² Low numeracy is a risk factor for poor glycemic control among patients with diabetes,¹³ medication adherence in HIV/AIDS,¹⁴ and worse blood pressure control in hypertensives.¹⁵ Much like these conditions, HF requires that patients understand, use, and act on numerical information. Maintaining a low-salt diet, monitoring weight, adjusting diuretic doses, and measuring blood pressure are tasks that HF patients are asked to perform on a daily or near-daily basis. These tasks are particularly important in the posthospitalization period and could be complicated by medication changes, which might create additional challenges for patients with inadequate numeracy. Additionally, cognitive impairment, which is a highly prevalent comorbid condition among adults with HF,^16,17 might impose additional barriers for those with inadequate numeracy who do not have adequate social support. However, to date, numeracy in the context of HF has not been well described.

Herein, we examined the effects of numeracy, alongside health literacy and cognition, on 30-day readmission risk among patients hospitalized for acute decompensated HF (ADHF).

Study Design

The Vanderbilt Inpatient Cohort Study (VICS) is a prospective observational study of patients admitted with cardiovascular disease to Vanderbilt University Medical Center (VUMC), an academic tertiary care hospital. VICS was designed to investigate the impact of social determinants of health on postdischarge health outcomes. A detailed description of the study rationale, design, and methods is described elsewhere.³

Briefly, participants completed a baseline interview while hospitalized, and follow-up phone calls were conducted within 1 week of discharge, at 30 days, and at 90 days. At 30 and 90 days postdischarge, healthcare utilization was ascertained by review of medical records and patient report. Clinical data about the index hospitalization were also abstracted. The Vanderbilt University Institutional Review Board approved the study.

Study Population

Patients hospitalized from 2011 to 2015 with a likely diagnosis of acute coronary syndrome and/or ADHF, as determined by a physician’s review of the medical record, were identified as potentially eligible. Research assistants assessed these patients for the presence of the following exclusion criteria: less than 18 years of age, non-English speaking, unstable psychiatric illness, a low likelihood of follow-up (eg, no reliable telephone number), on hospice, or otherwise too ill to complete an interview. Additionally, those with severe cognitive impairment, as assessed from the medical record (such as seeing a note describing dementia), and those with delirium, as assessed by the brief confusion assessment method, were excluded from enrollment in the study.^18,19 Those who died before discharge or during the 30-day follow-up period were excluded. For this analysis, we restricted our sample to only include participants who were hospitalized for ADHF.

Outcome Measure: 30-Day Readmission

The main outcome was all-cause readmission to any hospital within 30 days of discharge, as determined by patient interview, review of electronic medical records from VUMC, and review of outside hospital records.

Main Exposures: Numeracy, Health Literacy, and Cognitive Impairment

Numeracy was assessed with a 3-item version of the Subjective Numeracy Scale (SNS-3), which quantifies the patients perceived quantitative abilities.²⁰ Other authors have shown that the SNS-3 has a correlation coefficient of 0.88 with the full-length SNS-8 and a Cronbach’s alpha of 0.78.^20-22 The SNS-3 is reported as the mean on a scale from 1 to 6, with higher scores reflecting higher numeracy.

Subjective health literacy was assessed by using the 3-item Brief Health Literacy Screen (BHLS).²³ Scores range from 3 to 15, with higher scores reflecting higher literacy. Objective health literacy was assessed with the short form of the Test of Functional Health Literacy in Adults (sTOFHLA).^24,25 Scores may be categorized as inadequate (0-16), marginal (17-22), or adequate (23-36).

We assessed cognition by using the 10-item Short Portable Mental Status Questionnaire (SPMSQ).²⁶ The SPMSQ, which describes a person’s capacity for memory, structured thought, and orientation, has been validated and has demonstrated good reliability and validity.²⁷ Scores of 0 were considered to reflect intact cognition, and scores of 1 or more were considered to reflect any cognitive impairment, a scoring approach employed by other authors.²⁸ We used this approach, rather than the traditional scoring system developed by Pfeiffer et al.²⁶ (1975), because it would be the most sensitive to detect any cognitive impairment in the VICS cohort, which excluded those with severe cognition impairment, dementia, and delirium.

Covariates

During the hospitalization, participants completed an in-person interviewer-administered baseline assessment composed of demographic information, including age, self-reported race (white and nonwhite), educational attainment, home status (married, not married and living with someone, not married and living alone), and household income.

Clinical and diagnostic characteristics abstracted from the medical record included a medical history of HF, HF subtype (classified by left ventricular ejection fraction [LVEF]), coronary artery disease, chronic obstructive pulmonary disease (COPD), diabetes mellitus (DM), and comorbidity burden as summarized by the van Walraven-Elixhauser score.^29,30 Depressive symptoms were assessed during the 2 weeks prior to the hospitalization by using the first 8 items of the Patient Health Questionnaire.³¹ Scores ranged from 0 to 24, with higher scores reflecting more severe depressive symptoms. Laboratory values included estimated glomerular filtration rate (eGFR), hemoglobin (g/dl), sodium (mg/L), and brain natriuretic peptide (BNP) (pg/ml) from the last laboratory draw before discharge. Smoking status was also assessed (current and former/nonsmokers).

Hospitalization characteristics included length of stay in days, number of prior admissions in the last year, and transfer to the intensive care unit during the index admission.

Statistical Analysis

Descriptive statistics were used to summarize patient characteristics. The Kruskal-Wallis test and the Pearson χ² test were used to determine the association between patient characteristics and levels of numeracy, literacy, and cognition separately. The unadjusted relationship between patient characteristics and 30-day readmission was assessed by using Wilcoxon rank sums tests for continuous variables and Pearson χ² tests for categorical variables. In addition, a correlation matrix was performed to assess the correlations between numeracy, health literacy, and cognition (supplementary Figure 1).

To examine the association between numeracy, health literacy, and cognition and 30-day readmissions, a series of multivariable Poisson (log-linear) regression models were fit.³² Like other studies, numeracy, health literacy, and cognition were examined as categorical and continuous measures in models.³³ Each model was modified with a sandwich estimator for robust standard errors. Log-linear models were chosen over logistic regression models for ease of interpretation because (exponentiated) parameters correspond to risk ratios (RRs) as opposed to odds ratios. Furthermore, the fitting challenges associated with log-linear models when predicted probabilities are near 0 or 1 were not present in these analyses. Redundancy analyses were conducted to ensure that independent variables were not highly correlated with a linear combination of the other independent variables. To avoid case-wise deletion of records with missing covariates, we employed multiple imputation with 10 imputation samples by using predictive mean matching.^34,35 All analyses were conducted in R version 3.1.2 (The R Foundation, Vienna, Austria).³⁶

Overall, 883 patients were included in this analysis (supplementary Figure 2). Of the 883 participants, 46% were female and 76% were white (Table 1). Their median age was 60 years (interdecile range [IDR] 39-78) and the median educational attainment was 13.5 years (IDR 11-18).

This is the first study to examine the effect of numeracy alongside literacy and cognition on 30-day readmission risk among patients hospitalized with ADHF. Overall, we found that 33.9% of participants had inadequate numeracy skills, and 24.6% had inadequate or marginal health literacy. In unadjusted and adjusted models, numeracy was not associated with 30-day readmission. Although (objective) low health literacy was associated with 30-day readmission in unadjusted models, it was not in adjusted models. Additionally, though 53% of participants had any cognitive impairment, readmission did not differ significantly by this factor. Taken together, these findings suggest that other factors may be greater determinants of 30-day readmissions among patients hospitalized for ADHF.

Only 1 other study has examined the effect of numeracy on readmission risk among patients hospitalized for HF. In this multicenter prospective study, McNaughton et al.³⁷ found low numeracy to be associated with higher odds of recidivism to the emergency department (ED) or hospital within 30 days. Our findings may differ from theirs for a few reasons. First, their study had a significantly higher percentage of individuals with low numeracy (55%) compared with ours (33.9%). This may be because they did not exclude individuals with severe cognitive impairment, and their patient population was of lower socioeconomic status (SES) than ours. Low SES is associated with higher 30-day readmissions among HF patients^1,10 throughout the literature, and low numeracy is associated with low SES in other diseases.^13,38,39 Finally, they studied recidivism, which was defined as any unplanned return to the ED or hospital within 30 days of the index ED visit for acute HF. We only focused on 30-day readmissions, which also may explain why our results differed.

We found that health literacy was not associated with 30-day readmissions, which is consistent with the literature. Although an association between health literacy and mortality exists among adults with HF, several studies have not found an association between health literacy and 30- and 90-day readmission among adults hospitalized for HF.^8,9,40 Although we found an association between objective health literacy and 30-day readmission in unadjusted analyses, we did not find one in the multivariable model. This, along with our numeracy finding, suggests that numeracy and literacy may not be driving the 30-day readmission risk among patients hospitalized with ADHF.

We examined cognition alongside numeracy and literacy because it is a prevalent condition among HF patients and because it is associated with adverse outcomes among patients with HF, including readmission.^41,42 Studies have shown that HF preferentially affects certain cognitive domains,⁴³ some of which are vital to HF self-care activities. We found that 53% of patients had any cognitive impairment, which is consistent with the literature of adults hospitalized for ADHF.^44,45 Cognitive impairment was not, however, associated with 30-day readmissions. There may be a couple reasons for this. First, we measured cognitive impairment with the SPMSQ, which, although widely used and well-validated, does not assess executive function, the domain most commonly affected in HF patients with cognitive impairment.⁴⁶ Second, patients with severe cognitive impairment and those with delirium were excluded from this study, which may have limited our ability to detect differences in readmission by this factor.

As in prior studies, we found that a history of DM and more hospitalizations in the prior year were independently associated with 30-day readmissions in fully adjusted models. Like other studies, in adjusted models, we found that LVEF and a history of HF were not independently associated with 30-day readmission.^47-49 This, however, is not surprising because recent studies have shown that, although HF patients are at risk for multiple hospitalizations, early readmission after a hospitalization for ADHF specifically is often because of reasons unrelated to HF or a non-cardiovascular cause in general.^50,51

Although a negative study, several important themes emerged. First, while we were able to assess numeracy, health literacy, and cognition, none of these measures were HF-specific. It is possible that we did not see an effect on readmission because our instruments failed to assess domains specific to HF, such as monitoring weight changes, following a low-salt diet, and interpreting blood pressure. Currently, however, no HF-specific objective numeracy measure exists. With respect to health literacy, only 1 HF-specific measure exists,⁵² although it was only recently developed and validated. Second, while numeracy may not be a driving influence of all-cause 30-day readmissions, it may be associated with other health behaviors and quality metrics that we did not examine here, such as self-care, medication adherence, and HF-specific readmissions. Third, it is likely that the progression of HF itself, as well as the clinical management of patients following discharge, contribute significantly to 30-day readmissions. Increased attention to predischarge processes for HF patients occurred at VUMC during the study period; close follow-up and evidence-directed therapies may have mitigated some of the expected associations. Finally, we were not able to assess numeracy of participants’ primary caregivers who may help patients at home, especially postdischarge. Though a number of studies have examined the role of family caregivers in the management of HF,^53,54 none have examined numeracy levels of caregivers in the context of HF, and this may be worth doing in future studies.

Overall, our study has several strengths. The size of the cohort is large and there were high response rates during the follow-up period. Unlike other HF readmission studies, VICS accounts for readmissions to outside hospitals. Approximately 35% of all hospitalizations in VICS are to outside facilities. Thus, the ascertainment of readmissions to hospitals other than Vanderbilt is more comprehensive than if readmissions to VUMC were only considered. We were able to include a number of clinical comorbidities, laboratory and diagnostic tests from the index admission, and hospitalization characteristics in our analyses. Finally, we performed additional analyses to investigate the correlation between numeracy, literacy, and cognition; ultimately, we found that the majority of these correlations were weak, which supports our ability to study them simultaneously among VICS participants.

Nonetheless, we note some limitations. Although we captured readmissions to outside hospitals, the study took place at a single referral center in Tennessee. Though patients were diverse in age and comorbidities, they were mostly white and of higher SES. Finally, we used home status as a proxy for social support, which may underestimate the support that home care workers provide.

In conclusion, in this prospective longitudinal study of adults hospitalized with ADHF, inadequate numeracy was present in more than a third of patients, and low health literacy was present in roughly a quarter of patients. Neither numeracy nor health literacy, however, were associated with 30-day readmissions in adjusted analyses. Any cognitive impairment, although present in roughly one-half of patients, was not associated with 30-day readmission either. Our findings suggest that other influences may play a more dominant role in determining 30-day readmission rates in patients hospitalized for ADHF than inadequate numeracy, low health literacy, or cognitive impairment as assessed here.

Acknowledgments

This research was supported by the National Heart, Lung, and Blood Institute (R01 HL109388) and in part by the National Center for Advancing Translational Sciences (UL1 TR000445-06). The content is solely the responsibility of the authors and does not necessarily represent official views of the National Institutes of Health. The authors’ funding sources did not participate in the planning, collection, analysis, or interpretation of data or in the decision to submit for publication. Dr. Sterling is supported by T32HS000066 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality. Dr. Mixon has a VA Health Services Research and Development Service Career Development Award at the Tennessee Valley Healthcare System, Department of Veterans Affairs (CDA 12-168). This material was presented at the Society of General Internal Medicine Annual Meeting on April 20, 2017, in Washington, DC.

Disclosure

Dr. Kripalani reports personal fees from Verustat, personal fees from SAI Interactive, and equity from Bioscape Digital, all outside of the submitted work. Dr. Rothman and Dr. Wallston report personal fees from EdLogics outside of the submitted work. All of the other authors have nothing to disclose

A 62-year-old man with severe chronic obstructive pulmonary disease (COPD; forced expiratory volume during the first second [FEV1] 40% predicted) and type 2 diabetes mellitus presented to a Veterans Affairs emergency department (ED) with a steadily worsening cough of 4-months’ duration. He also reported subjective fevers, sputum production, shortness of breath, and unintentional 20-pound weight loss. He denied chills, chest pain, nausea, or vomiting.

Cough is classified as acute, subacute, or chronic based on duration of less than 3 weeks, between 3-8 weeks, and greater than 8 weeks, respectively. Common causes of chronic cough include bronchitis, acid reflux, cough-variant asthma, and a side effect of angiotensin converting enzyme inhibitors. Unintentional weight loss suggests a serious disorder, including indolent infection, end-stage COPD, malignancy, and autoimmune causes. Among patients with chronic bronchitis, the microbiology of sputum is often mixed with commensal respiratory flora, including Streptococcus pneumoniae and Haemophilus species. When these organisms are not recovered in sputa, or when patients fail to respond to empiric treatment, the differential diagnosis should be broadened to include pulmonary tuberculosis, nontuberculous mycobacterial infection, lung abscess, pulmonary nocardiosis, or pertussis.

An exposure and social history can focus the differential. For example, coccidioidomycosis or histoplasmosis may present indolently, but have distinct geographic distributions. Bird fanciers may acquire hypersensitivity pneumonitis, psittacosis, or cryptococcosis. Risk factors including smoking history, corticosteroid use, uncontrolled diabetes, and ill contacts should be assessed.

He was discharged from the ED twice in the last 2 weeks after presenting with similar symptoms. On each occasion, he was treated for presumed COPD exacerbations with nebulized albuterol and ipratropium, methylprednisolone followed by oral prednisone, and azithromycin, which did not lead to improvement. Over the last 3 days, he developed lower extremity edema, orthopnea, and dyspnea at rest. He reported worsening fatigue, night sweats, and anorexia. He denied any sick contacts.

Two diagnostic issues have emerged. His edema, orthopnea, and dyspnea at rest suggest a new cause of hypervolemia, perhaps caused by sodium retention from corticosteroids, pulmonary edema from valvular or myocardial disease, or renal failure. More concerning is that he has been treated with azithromycin twice recently but still has night sweats, fatigue, and anorexia. The presence of weight loss despite extracellular volume accumulation suggests an indolent systemic illness. Infection with macrolide-resistant organisms, such as nocardia, mycobacteria, or endemic mycoses, remains high on the differential diagnosis.

His past medical history included hypertension, untreated chronic hepatitis C, tobacco dependence, alcohol use disorder, and extraction of 8 decayed teeth 2 months earlier. He served in a noncombat role during the Vietnam War. He consumed 12 beers weekly with a remote history of alcoholism which required rehabilitation, reported a 50 pack-year smoking history, and denied intravenous (IV) drug use. He lived with an appropriately vaccinated dog and denied recent insect or animal exposures. He had a cat that passed away from an unknown illness 3 years prior. He was in a monogamous relationship with his girlfriend of 35 years. His father had coronary disease. His medications included glyburide, hydrochlorothiazide, lisinopril, theophylline, and meloxicam. Chronic cough, weight loss, diabetes, alcoholism, and history of dental disease raise concern for lung abscess. Oral microbiota such as Streptococcus viridans and Actinomycetes are usually harmless, but when aspirated repeatedly, such as during alcohol intoxication, may evolve into a lung abscess via bronchogenic spread. The combination of unintentional weight loss and smoking history raises concern for lung malignancy. Small cell lung cancer can present with paraneoplastic Cushing’s syndrome and could explain the patient’s volume overload. Finally, human immunodeficiency virus (HIV) serostatus should be determined in all adult patients.

His temperature was 37 °C, blood pressure 161/69 mm Hg, pulse 104 beats per minute, respiratory rate 20 breaths per minute, and oxygen saturation was 95% on room air. On examination, he was an unkempt, ill-appearing man. He had poor dentition, but no oral ulcers or petechiae. Pulmonary exam revealed diffuse rhonchi and scattered wheezes. He developed dyspnea after speaking 2 sentences. Cardiovascular exam showed regular tachycardia, normal S1 and S2 heart sounds, and both an S3 and S4 gallop. A grade III/VI holosystolic murmur at the left lower sternal border with apical radiation, and an early, grade III/IV diastolic murmur at the right upper sternal border were present. Neck exam showed jugular venous distention (JVD) 8 cm above the right clavicle. Lower extremities showed symmetric 3+ pitting edema to the knees. His abdomen was soft, nondistended, and without hepatosplenomegaly. There was no lymphadenopathy. Skin exam showed small, healed excoriations on his anterior shins, forearms, and knuckles. There were no petechiae, Janeway lesions, or Osler’s nodes.

These exam findings change the differential substantially. New regurgitant murmurs strongly suggest infective endocarditis (IE). A diastolic murmur is never normal and suggests aortic regurgitation. The holosystolic murmur with apical radiation suggests mitral regurgitation. Cutaneous stigmata should always be sought, but are found in fewer than half of cases of subacute IE, and their absence does not rule out this diagnosis. Disheveled hygiene and excoriations suggest a skin source of infection, and poor dentition is concerning for an oral source. For the moment, the source does not matter. His clinical condition is serious: tachycardia, JVD, edema, and two-sentence dyspnea indicate congestive heart failure. Even before labs and imaging return, inpatient admission is warranted.

Serum sodium concentration was 140 mEq/L, potassium 3.7 mEq/L, chloride 103 mEq/L, bicarbonate 30 mEq/L, blood urea nitrogen (BUN) 26 mg/dL, creatinine 0.8 mg/dL, glucose 120 mg/dL, and calcium 9.0 mg/dL. The white blood cell count was 7100/µL, hemoglobin 11.8 g/dL, and platelet count 101 K/µL. Brain natriuretic peptide (BNP) was 785 pg/mL (reference range 0-100 pg/mL), aspartate aminotransferase 77 U/L, alanine aminotransferase 57 U/L, alkaline phosphatase 125 U/L, total bilirubin 0.8 mg/dL, total protein 7.7 g/dL, and albumin 3.7 g/dL. Erythrocyte sedimentation (ESR) rate was 38 mm/hour (reference range 0-25 mm/hour) and C-reactive protein (CRP) 0.62 mg/dL (reference range <1.0 mg/dL). Cardiac troponins were 0.03 ng/mL (reference range <0.04 ng/mL). Screening for HIV was negative. Urinalysis showed trace blood by dipstick, but no glucose, protein, dysmorphic red blood cells, or casts. Two sets of peripheral blood cultures were drawn. Two sets of blood cultures from his previous ED visits were negative (drawn 6 and 14 days prior).

These laboratory values are nonspecific, and the differential remains unchanged, with top concern for IE, then lung abscess. Ideally, 3 sets of cultures drawn greater than 12 hours apart should be obtained because the likelihood of pathogen detection rises with the volume of blood tested. Thrombocytopenia and microscopic hematuria suggest microangiopathic hemolytic anemia, and a peripheral blood smear should be examined for schistocytes. Glomerulonephritis from immune complex deposition can occur in IE, but is unlikely with a normal serum creatinine and lack of proteinuria, dysmorphic red blood cells, or casts. The elevated BNP suggests cardiac strain due to a regurgitant valve. ESR and CRP are rarely helpful in this situation, and perhaps previous treatment with azithromycin and steroids prevented significant elevation.

Due to concern for IE, blood cultures were repeated, and IV vancomycin, IV ceftriaxone, and IV gentamicin were initiated. Azithromycin and prednisone were discontinued. His respiratory status continued to improve with IV furosemide, albuterol, ipratropium, and supportive care.

TTE inadequately visualizes the mitral valve, but is useful for tricuspid valve assessment because the right ventricle is closer to the chest wall. Transesophageal echocardiography (TEE) is indicated for a more detailed assessment of the left heart valves for vegetations and perivalvar abscesses. The new regurgitant murmurs satisfy a major criterion of the modified Duke criteria, and valvar vegetations suggests IE. He does not yet fulfill the other major modified Duke criterion for IE, nor does he satisfy enough minor criteria because there are no diagnostic vascular, microbiologic, or immunologic phenomena. However, no diagnostic rubric is perfect, and these results should not supersede clinical judgment. Despite the absence of positive cultures, the concern for bacterial IE remains high. The absence of embolic phenomena fits best with subacute rather than acute IE. Three negative blood cultures to date suggest a fastidious organism is responsible, although oral flora remain on the differential.

There is rarely a need to “hold” blood cultures for prolonged periods because modern instruments typically yield positive results within 7 days for most bacteria, including the HACEK group. Blood culture-negative endocarditis (BCNE) is considered when 3 sets of cultures are negative for at least 5 days. In this situation, one should consider other microorganisms based on the patient’s exposure history. Only certain species with complex growth requirements, such as Brucella and Bartonella, require prolonged holds. Revisiting his exposure history would be helpful in deciding whether serologic testing warranted. If he recalls exposure to parturient animals, then Coxiella is worth pursuing; if he has been bitten by lice, then B. quintana rises as a possibility; if the scratches on his limbs are from recent cat scratches, then B. henselae becomes more likely. Both C. burnetti and Bartonella endocarditis might be partially treated by his courses of azithromycin, confounding the picture.

If the infectious work-up is ultimately negative, one could then consider other etiologies of endocarditis, such as nonbacterial thrombotic endocarditis, which is seen in the context of malignancy and systemic lupus erythematosus (Libman-Sacks endocarditis). Other mimickers of IE include myxomatous valve degeneration, ruptured mitral chordae, and eosinophilic heart disease (Löffler’s endocarditis).

A transesophageal echocardiogram confirmed the presence of small echodensities on the aortic valve’s right and left coronary cusps, consistent with vegetations. The vegetation on the anterior leaflet of the mitral valve from the TTE also showed an aneurysm with a small perforation (Figure 2).

The combination of aortic regurgitation and the mitral valve aneurysm supports IE, because the aortic regurgitant jet directly strikes the anterior mitral valve leaflet, seeding the valve with infection from the aortic cusps. A positive serum PCR is diagnostic, but if it had been negative or unavailable, the serology would remain very helpful. In this context, the elevated IgG titer implicates B. henselae, the agent responsible for cat scratch disease (CSD). Out of context, these titers would not be diagnostic, because anti-Bartonella IgG may be increased due to a prior subclinical episode of CSD. Anti-Bartonella IgM is an unreliable indicator of recent infection because it may wane within weeks, and this IgG titer is higher than what is observed with most remote infections.

Revisiting previous cat exposure is warranted. He lost his cat to an illness 3 years prior, however it would be appropriate to inquire about other animals, such as a stray kitten with fleas, which his skin scratches suggest. Up to 50% of all cats in flea endemic regions harbor Bartonella and are asymptomatic. Rarely, dogs can serve as reservoirs of this organism, with a presumed transmission route via flea, louse, or tick. Regardless of the route of infection, treatment should be focused on B. henselae IE.

Azithromycin can treat CSD, and its use for his presumed COPD exacerbation may have temporized his infection. However, azithromycin monotherapy is not recommended for B. henselae IE. Treatment is usually with 2 antibiotics, including an aminoglycoside (gentamicin) for the first 2 weeks, combined with either a tetracycline, a macrolide, or a beta-lactam for a minimum of 4-6 weeks. Oral rifampin can be considered if gentamicin is not tolerated. After completing IV treatment, an additional 6 months of oral doxycycline or azithromycin should be considered, especially for those who have not undergone valve surgery.

His symptoms are compatible with subacute IE, which is typically more difficult to diagnose than acute IE due to its insidious onset. He meets criteria for blood culture negative IE based on 3 sets of negative blood cultures for greater than 5 days and major criteria for IE. The pathologic changes are consistent with B. henselae infection.

The incidence of IE in the United States is 40,000 cases per year¹ with an in-hospital mortality of 15%-20% and a 1-year mortality of up to 40%.^2,3 Five to 20% of patients with IE never develop positive blood cultures⁴ due to receipt of antibiotics prior to culture, inadequate microbiologic testing, or infection caused by noncultivable bacteria (eg, Tropheryma whipplei), fastidious extracellular bacteria (eg, HACEK group and nutritionally variant streptococci), or by intracellular pathogens with complex nutrient requirements (eg, Bartonella, Chlamydia, Brucella, or Coxiella). Previous administration of antibiotics reduces the likelihood of isolating an organism by 35%-40%.⁵ Patients meeting criteria for BCNE should prompt consideration of serologic testing. The most prevalent pathogens vary globally, and incidence data in the US is scarce. Worldwide, the majority of BCNE cases are caused by Coxiella, Bartonella, and Brucella species.^6,7

When clinical suspicion for IE remains high despite negative cultures, detailed history can uncover clues and guide additional testing. For example, contact with contaminated milk products or farm animals are associated with Brucella, Coxiella, and Erysipelothrix species IE.^7,8 Bartonella species are zoonotic gram-negative bacilli with a tropism for endothelial cells and are transmitted by arthropod vectors (ie, fleas, lice, ticks, and sandflies), cat scratches, or cat bites. Bartonella may account for 3%-4% of all cases of IE, most of which are due to B. henselae and B. quintana.^{7, 9} Underlying heart valve disease, alcoholism, cirrhosis, and homelessness are associated with B. henselae endocarditis.¹⁰

Diagnostic criteria are lacking for B. henselae IE, and the modified Duke criteria is of limited utility for diagnosing Bartonella IE because blood cultures are often negative and echocardiographic evidence of vegetation is not always apparent. Serology plays a critical role in the diagnosis of Bartonella infections. The addition of positive serology, Western blot or PCR for B. henselae and B. quintana as a major criterion in the modified Duke criteria for IE has been proposed but has not yet been formally accepted.⁹ For B. henselae IE, an IgG titer of ≥1:800 has been recommended as a cutoff for subacute IE because it combines a high specificity and positive predictive value along with reasonable sensitivity and negative predictive value in this situation.⁹ The humoral immune response rises over time, and thus acute IE due to Bartonella may not generate a substantial IgG titer. Interestingly, because of the indolent nature of this pathogen, most cases of IE present once IgG titers have begun to rise. Serum PCR testing has shown a sensitivity and specificity of 58% and 100%, respectively.¹¹ Isolation by blood culture requires specific growth media and prolonged incubation, with a sensitivity as low as 20% and 30% for blood and tissue, respectively.¹⁰ The microbiology laboratory should be notified of suspected Bartonella to intensify efforts to cultivate this organism. If infection with Coxiella or Brucella is suspected, the lab should also be informed, both to increase diagnostic yield and to trigger enhanced biosafety precautions when handling the specimens. Despite attempts to optimize the yield, up to 75% of Bartonella IE may remain culture negative,^12,13 making it difficult to meet the current major modified Duke criterion of positive blood cultures. H&E staining of valve tissue infected with Bartonella commonly reveals increased inflammation, fibrosis, and calcified granulomas relative to endocarditis from other causes.¹⁴ The Warthin-Starry silver stain can identify small, darkly staining bacteria in more than 75% of Bartonella endocarditis; however, this stain is not specific for Bartonella species.⁹

This case highlights the challenge of diagnosing subacute IE because this patient received antibiotics and steroids prior to presentation, clouding the clinical picture. Although he did not exhibit textbook signs of endocarditis, his symptoms (new onset heart failure and new regurgitant murmurs) prioritized the diagnosis. The combination of elevated serum titers, positive PCR, valve granulomas and abscesses on TEE, and pathology findings led the discussant to the correct diagnosis. Scratching beneath the surface revealed his penchant for cats, but this was only considered a key epidemiological feature later in his clinical course.

• Subacute IE typically presents with indolent constitutional symptoms over a course of weeks to months, whereas acute IE causes a rapid onset of fevers, rigors, and is more likely to exhibit embolic phenomena.
• Epidemiologic features specific to Bartonella species include alcoholism, cirrhosis, dog or cat exposure, homelessness, and body lice, and should be considered in suspected cases of BCNE.
• If suspicion for endocarditis remains high and animal exposure is elicited, then serologic and PCR testing for fastidious organisms should be strongly considered. The most common causes of BCNE include Coxiella, Bartonella, and Brucella species.
• The modified Duke criteria do not incorporate Bartonella within the diagnostic schema. Presentation is usually late and often requires valve replacement.

Acknowledgments

The authors thank Dr. Michael Pfeiffer from the Pennsylvania State Hershey Heart and Vascular Institute for providing his expertise in diagnostic echocardiography.

Disclosure

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

A 62-year-old man with severe chronic obstructive pulmonary disease (COPD; forced expiratory volume during the first second [FEV1] 40% predicted) and type 2 diabetes mellitus presented to a Veterans Affairs emergency department (ED) with a steadily worsening cough of 4-months’ duration. He also reported subjective fevers, sputum production, shortness of breath, and unintentional 20-pound weight loss. He denied chills, chest pain, nausea, or vomiting.

Cough is classified as acute, subacute, or chronic based on duration of less than 3 weeks, between 3-8 weeks, and greater than 8 weeks, respectively. Common causes of chronic cough include bronchitis, acid reflux, cough-variant asthma, and a side effect of angiotensin converting enzyme inhibitors. Unintentional weight loss suggests a serious disorder, including indolent infection, end-stage COPD, malignancy, and autoimmune causes. Among patients with chronic bronchitis, the microbiology of sputum is often mixed with commensal respiratory flora, including Streptococcus pneumoniae and Haemophilus species. When these organisms are not recovered in sputa, or when patients fail to respond to empiric treatment, the differential diagnosis should be broadened to include pulmonary tuberculosis, nontuberculous mycobacterial infection, lung abscess, pulmonary nocardiosis, or pertussis.

An exposure and social history can focus the differential. For example, coccidioidomycosis or histoplasmosis may present indolently, but have distinct geographic distributions. Bird fanciers may acquire hypersensitivity pneumonitis, psittacosis, or cryptococcosis. Risk factors including smoking history, corticosteroid use, uncontrolled diabetes, and ill contacts should be assessed.

He was discharged from the ED twice in the last 2 weeks after presenting with similar symptoms. On each occasion, he was treated for presumed COPD exacerbations with nebulized albuterol and ipratropium, methylprednisolone followed by oral prednisone, and azithromycin, which did not lead to improvement. Over the last 3 days, he developed lower extremity edema, orthopnea, and dyspnea at rest. He reported worsening fatigue, night sweats, and anorexia. He denied any sick contacts.

Two diagnostic issues have emerged. His edema, orthopnea, and dyspnea at rest suggest a new cause of hypervolemia, perhaps caused by sodium retention from corticosteroids, pulmonary edema from valvular or myocardial disease, or renal failure. More concerning is that he has been treated with azithromycin twice recently but still has night sweats, fatigue, and anorexia. The presence of weight loss despite extracellular volume accumulation suggests an indolent systemic illness. Infection with macrolide-resistant organisms, such as nocardia, mycobacteria, or endemic mycoses, remains high on the differential diagnosis.

His past medical history included hypertension, untreated chronic hepatitis C, tobacco dependence, alcohol use disorder, and extraction of 8 decayed teeth 2 months earlier. He served in a noncombat role during the Vietnam War. He consumed 12 beers weekly with a remote history of alcoholism which required rehabilitation, reported a 50 pack-year smoking history, and denied intravenous (IV) drug use. He lived with an appropriately vaccinated dog and denied recent insect or animal exposures. He had a cat that passed away from an unknown illness 3 years prior. He was in a monogamous relationship with his girlfriend of 35 years. His father had coronary disease. His medications included glyburide, hydrochlorothiazide, lisinopril, theophylline, and meloxicam. Chronic cough, weight loss, diabetes, alcoholism, and history of dental disease raise concern for lung abscess. Oral microbiota such as Streptococcus viridans and Actinomycetes are usually harmless, but when aspirated repeatedly, such as during alcohol intoxication, may evolve into a lung abscess via bronchogenic spread. The combination of unintentional weight loss and smoking history raises concern for lung malignancy. Small cell lung cancer can present with paraneoplastic Cushing’s syndrome and could explain the patient’s volume overload. Finally, human immunodeficiency virus (HIV) serostatus should be determined in all adult patients.

His temperature was 37 °C, blood pressure 161/69 mm Hg, pulse 104 beats per minute, respiratory rate 20 breaths per minute, and oxygen saturation was 95% on room air. On examination, he was an unkempt, ill-appearing man. He had poor dentition, but no oral ulcers or petechiae. Pulmonary exam revealed diffuse rhonchi and scattered wheezes. He developed dyspnea after speaking 2 sentences. Cardiovascular exam showed regular tachycardia, normal S1 and S2 heart sounds, and both an S3 and S4 gallop. A grade III/VI holosystolic murmur at the left lower sternal border with apical radiation, and an early, grade III/IV diastolic murmur at the right upper sternal border were present. Neck exam showed jugular venous distention (JVD) 8 cm above the right clavicle. Lower extremities showed symmetric 3+ pitting edema to the knees. His abdomen was soft, nondistended, and without hepatosplenomegaly. There was no lymphadenopathy. Skin exam showed small, healed excoriations on his anterior shins, forearms, and knuckles. There were no petechiae, Janeway lesions, or Osler’s nodes.

These exam findings change the differential substantially. New regurgitant murmurs strongly suggest infective endocarditis (IE). A diastolic murmur is never normal and suggests aortic regurgitation. The holosystolic murmur with apical radiation suggests mitral regurgitation. Cutaneous stigmata should always be sought, but are found in fewer than half of cases of subacute IE, and their absence does not rule out this diagnosis. Disheveled hygiene and excoriations suggest a skin source of infection, and poor dentition is concerning for an oral source. For the moment, the source does not matter. His clinical condition is serious: tachycardia, JVD, edema, and two-sentence dyspnea indicate congestive heart failure. Even before labs and imaging return, inpatient admission is warranted.

Serum sodium concentration was 140 mEq/L, potassium 3.7 mEq/L, chloride 103 mEq/L, bicarbonate 30 mEq/L, blood urea nitrogen (BUN) 26 mg/dL, creatinine 0.8 mg/dL, glucose 120 mg/dL, and calcium 9.0 mg/dL. The white blood cell count was 7100/µL, hemoglobin 11.8 g/dL, and platelet count 101 K/µL. Brain natriuretic peptide (BNP) was 785 pg/mL (reference range 0-100 pg/mL), aspartate aminotransferase 77 U/L, alanine aminotransferase 57 U/L, alkaline phosphatase 125 U/L, total bilirubin 0.8 mg/dL, total protein 7.7 g/dL, and albumin 3.7 g/dL. Erythrocyte sedimentation (ESR) rate was 38 mm/hour (reference range 0-25 mm/hour) and C-reactive protein (CRP) 0.62 mg/dL (reference range <1.0 mg/dL). Cardiac troponins were 0.03 ng/mL (reference range <0.04 ng/mL). Screening for HIV was negative. Urinalysis showed trace blood by dipstick, but no glucose, protein, dysmorphic red blood cells, or casts. Two sets of peripheral blood cultures were drawn. Two sets of blood cultures from his previous ED visits were negative (drawn 6 and 14 days prior).

These laboratory values are nonspecific, and the differential remains unchanged, with top concern for IE, then lung abscess. Ideally, 3 sets of cultures drawn greater than 12 hours apart should be obtained because the likelihood of pathogen detection rises with the volume of blood tested. Thrombocytopenia and microscopic hematuria suggest microangiopathic hemolytic anemia, and a peripheral blood smear should be examined for schistocytes. Glomerulonephritis from immune complex deposition can occur in IE, but is unlikely with a normal serum creatinine and lack of proteinuria, dysmorphic red blood cells, or casts. The elevated BNP suggests cardiac strain due to a regurgitant valve. ESR and CRP are rarely helpful in this situation, and perhaps previous treatment with azithromycin and steroids prevented significant elevation.

Due to concern for IE, blood cultures were repeated, and IV vancomycin, IV ceftriaxone, and IV gentamicin were initiated. Azithromycin and prednisone were discontinued. His respiratory status continued to improve with IV furosemide, albuterol, ipratropium, and supportive care.

TTE inadequately visualizes the mitral valve, but is useful for tricuspid valve assessment because the right ventricle is closer to the chest wall. Transesophageal echocardiography (TEE) is indicated for a more detailed assessment of the left heart valves for vegetations and perivalvar abscesses. The new regurgitant murmurs satisfy a major criterion of the modified Duke criteria, and valvar vegetations suggests IE. He does not yet fulfill the other major modified Duke criterion for IE, nor does he satisfy enough minor criteria because there are no diagnostic vascular, microbiologic, or immunologic phenomena. However, no diagnostic rubric is perfect, and these results should not supersede clinical judgment. Despite the absence of positive cultures, the concern for bacterial IE remains high. The absence of embolic phenomena fits best with subacute rather than acute IE. Three negative blood cultures to date suggest a fastidious organism is responsible, although oral flora remain on the differential.

There is rarely a need to “hold” blood cultures for prolonged periods because modern instruments typically yield positive results within 7 days for most bacteria, including the HACEK group. Blood culture-negative endocarditis (BCNE) is considered when 3 sets of cultures are negative for at least 5 days. In this situation, one should consider other microorganisms based on the patient’s exposure history. Only certain species with complex growth requirements, such as Brucella and Bartonella, require prolonged holds. Revisiting his exposure history would be helpful in deciding whether serologic testing warranted. If he recalls exposure to parturient animals, then Coxiella is worth pursuing; if he has been bitten by lice, then B. quintana rises as a possibility; if the scratches on his limbs are from recent cat scratches, then B. henselae becomes more likely. Both C. burnetti and Bartonella endocarditis might be partially treated by his courses of azithromycin, confounding the picture.

If the infectious work-up is ultimately negative, one could then consider other etiologies of endocarditis, such as nonbacterial thrombotic endocarditis, which is seen in the context of malignancy and systemic lupus erythematosus (Libman-Sacks endocarditis). Other mimickers of IE include myxomatous valve degeneration, ruptured mitral chordae, and eosinophilic heart disease (Löffler’s endocarditis).

A transesophageal echocardiogram confirmed the presence of small echodensities on the aortic valve’s right and left coronary cusps, consistent with vegetations. The vegetation on the anterior leaflet of the mitral valve from the TTE also showed an aneurysm with a small perforation (Figure 2).

The combination of aortic regurgitation and the mitral valve aneurysm supports IE, because the aortic regurgitant jet directly strikes the anterior mitral valve leaflet, seeding the valve with infection from the aortic cusps. A positive serum PCR is diagnostic, but if it had been negative or unavailable, the serology would remain very helpful. In this context, the elevated IgG titer implicates B. henselae, the agent responsible for cat scratch disease (CSD). Out of context, these titers would not be diagnostic, because anti-Bartonella IgG may be increased due to a prior subclinical episode of CSD. Anti-Bartonella IgM is an unreliable indicator of recent infection because it may wane within weeks, and this IgG titer is higher than what is observed with most remote infections.

Revisiting previous cat exposure is warranted. He lost his cat to an illness 3 years prior, however it would be appropriate to inquire about other animals, such as a stray kitten with fleas, which his skin scratches suggest. Up to 50% of all cats in flea endemic regions harbor Bartonella and are asymptomatic. Rarely, dogs can serve as reservoirs of this organism, with a presumed transmission route via flea, louse, or tick. Regardless of the route of infection, treatment should be focused on B. henselae IE.

Azithromycin can treat CSD, and its use for his presumed COPD exacerbation may have temporized his infection. However, azithromycin monotherapy is not recommended for B. henselae IE. Treatment is usually with 2 antibiotics, including an aminoglycoside (gentamicin) for the first 2 weeks, combined with either a tetracycline, a macrolide, or a beta-lactam for a minimum of 4-6 weeks. Oral rifampin can be considered if gentamicin is not tolerated. After completing IV treatment, an additional 6 months of oral doxycycline or azithromycin should be considered, especially for those who have not undergone valve surgery.

His symptoms are compatible with subacute IE, which is typically more difficult to diagnose than acute IE due to its insidious onset. He meets criteria for blood culture negative IE based on 3 sets of negative blood cultures for greater than 5 days and major criteria for IE. The pathologic changes are consistent with B. henselae infection.

The incidence of IE in the United States is 40,000 cases per year¹ with an in-hospital mortality of 15%-20% and a 1-year mortality of up to 40%.^2,3 Five to 20% of patients with IE never develop positive blood cultures⁴ due to receipt of antibiotics prior to culture, inadequate microbiologic testing, or infection caused by noncultivable bacteria (eg, Tropheryma whipplei), fastidious extracellular bacteria (eg, HACEK group and nutritionally variant streptococci), or by intracellular pathogens with complex nutrient requirements (eg, Bartonella, Chlamydia, Brucella, or Coxiella). Previous administration of antibiotics reduces the likelihood of isolating an organism by 35%-40%.⁵ Patients meeting criteria for BCNE should prompt consideration of serologic testing. The most prevalent pathogens vary globally, and incidence data in the US is scarce. Worldwide, the majority of BCNE cases are caused by Coxiella, Bartonella, and Brucella species.^6,7

When clinical suspicion for IE remains high despite negative cultures, detailed history can uncover clues and guide additional testing. For example, contact with contaminated milk products or farm animals are associated with Brucella, Coxiella, and Erysipelothrix species IE.^7,8 Bartonella species are zoonotic gram-negative bacilli with a tropism for endothelial cells and are transmitted by arthropod vectors (ie, fleas, lice, ticks, and sandflies), cat scratches, or cat bites. Bartonella may account for 3%-4% of all cases of IE, most of which are due to B. henselae and B. quintana.^{7, 9} Underlying heart valve disease, alcoholism, cirrhosis, and homelessness are associated with B. henselae endocarditis.¹⁰

Diagnostic criteria are lacking for B. henselae IE, and the modified Duke criteria is of limited utility for diagnosing Bartonella IE because blood cultures are often negative and echocardiographic evidence of vegetation is not always apparent. Serology plays a critical role in the diagnosis of Bartonella infections. The addition of positive serology, Western blot or PCR for B. henselae and B. quintana as a major criterion in the modified Duke criteria for IE has been proposed but has not yet been formally accepted.⁹ For B. henselae IE, an IgG titer of ≥1:800 has been recommended as a cutoff for subacute IE because it combines a high specificity and positive predictive value along with reasonable sensitivity and negative predictive value in this situation.⁹ The humoral immune response rises over time, and thus acute IE due to Bartonella may not generate a substantial IgG titer. Interestingly, because of the indolent nature of this pathogen, most cases of IE present once IgG titers have begun to rise. Serum PCR testing has shown a sensitivity and specificity of 58% and 100%, respectively.¹¹ Isolation by blood culture requires specific growth media and prolonged incubation, with a sensitivity as low as 20% and 30% for blood and tissue, respectively.¹⁰ The microbiology laboratory should be notified of suspected Bartonella to intensify efforts to cultivate this organism. If infection with Coxiella or Brucella is suspected, the lab should also be informed, both to increase diagnostic yield and to trigger enhanced biosafety precautions when handling the specimens. Despite attempts to optimize the yield, up to 75% of Bartonella IE may remain culture negative,^12,13 making it difficult to meet the current major modified Duke criterion of positive blood cultures. H&E staining of valve tissue infected with Bartonella commonly reveals increased inflammation, fibrosis, and calcified granulomas relative to endocarditis from other causes.¹⁴ The Warthin-Starry silver stain can identify small, darkly staining bacteria in more than 75% of Bartonella endocarditis; however, this stain is not specific for Bartonella species.⁹

This case highlights the challenge of diagnosing subacute IE because this patient received antibiotics and steroids prior to presentation, clouding the clinical picture. Although he did not exhibit textbook signs of endocarditis, his symptoms (new onset heart failure and new regurgitant murmurs) prioritized the diagnosis. The combination of elevated serum titers, positive PCR, valve granulomas and abscesses on TEE, and pathology findings led the discussant to the correct diagnosis. Scratching beneath the surface revealed his penchant for cats, but this was only considered a key epidemiological feature later in his clinical course.

• Subacute IE typically presents with indolent constitutional symptoms over a course of weeks to months, whereas acute IE causes a rapid onset of fevers, rigors, and is more likely to exhibit embolic phenomena.
• Epidemiologic features specific to Bartonella species include alcoholism, cirrhosis, dog or cat exposure, homelessness, and body lice, and should be considered in suspected cases of BCNE.
• If suspicion for endocarditis remains high and animal exposure is elicited, then serologic and PCR testing for fastidious organisms should be strongly considered. The most common causes of BCNE include Coxiella, Bartonella, and Brucella species.
• The modified Duke criteria do not incorporate Bartonella within the diagnostic schema. Presentation is usually late and often requires valve replacement.

Acknowledgments

The authors thank Dr. Michael Pfeiffer from the Pennsylvania State Hershey Heart and Vascular Institute for providing his expertise in diagnostic echocardiography.

Disclosure

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

A 62-year-old man with severe chronic obstructive pulmonary disease (COPD; forced expiratory volume during the first second [FEV1] 40% predicted) and type 2 diabetes mellitus presented to a Veterans Affairs emergency department (ED) with a steadily worsening cough of 4-months’ duration. He also reported subjective fevers, sputum production, shortness of breath, and unintentional 20-pound weight loss. He denied chills, chest pain, nausea, or vomiting.

Cough is classified as acute, subacute, or chronic based on duration of less than 3 weeks, between 3-8 weeks, and greater than 8 weeks, respectively. Common causes of chronic cough include bronchitis, acid reflux, cough-variant asthma, and a side effect of angiotensin converting enzyme inhibitors. Unintentional weight loss suggests a serious disorder, including indolent infection, end-stage COPD, malignancy, and autoimmune causes. Among patients with chronic bronchitis, the microbiology of sputum is often mixed with commensal respiratory flora, including Streptococcus pneumoniae and Haemophilus species. When these organisms are not recovered in sputa, or when patients fail to respond to empiric treatment, the differential diagnosis should be broadened to include pulmonary tuberculosis, nontuberculous mycobacterial infection, lung abscess, pulmonary nocardiosis, or pertussis.

An exposure and social history can focus the differential. For example, coccidioidomycosis or histoplasmosis may present indolently, but have distinct geographic distributions. Bird fanciers may acquire hypersensitivity pneumonitis, psittacosis, or cryptococcosis. Risk factors including smoking history, corticosteroid use, uncontrolled diabetes, and ill contacts should be assessed.

He was discharged from the ED twice in the last 2 weeks after presenting with similar symptoms. On each occasion, he was treated for presumed COPD exacerbations with nebulized albuterol and ipratropium, methylprednisolone followed by oral prednisone, and azithromycin, which did not lead to improvement. Over the last 3 days, he developed lower extremity edema, orthopnea, and dyspnea at rest. He reported worsening fatigue, night sweats, and anorexia. He denied any sick contacts.

Two diagnostic issues have emerged. His edema, orthopnea, and dyspnea at rest suggest a new cause of hypervolemia, perhaps caused by sodium retention from corticosteroids, pulmonary edema from valvular or myocardial disease, or renal failure. More concerning is that he has been treated with azithromycin twice recently but still has night sweats, fatigue, and anorexia. The presence of weight loss despite extracellular volume accumulation suggests an indolent systemic illness. Infection with macrolide-resistant organisms, such as nocardia, mycobacteria, or endemic mycoses, remains high on the differential diagnosis.

His past medical history included hypertension, untreated chronic hepatitis C, tobacco dependence, alcohol use disorder, and extraction of 8 decayed teeth 2 months earlier. He served in a noncombat role during the Vietnam War. He consumed 12 beers weekly with a remote history of alcoholism which required rehabilitation, reported a 50 pack-year smoking history, and denied intravenous (IV) drug use. He lived with an appropriately vaccinated dog and denied recent insect or animal exposures. He had a cat that passed away from an unknown illness 3 years prior. He was in a monogamous relationship with his girlfriend of 35 years. His father had coronary disease. His medications included glyburide, hydrochlorothiazide, lisinopril, theophylline, and meloxicam. Chronic cough, weight loss, diabetes, alcoholism, and history of dental disease raise concern for lung abscess. Oral microbiota such as Streptococcus viridans and Actinomycetes are usually harmless, but when aspirated repeatedly, such as during alcohol intoxication, may evolve into a lung abscess via bronchogenic spread. The combination of unintentional weight loss and smoking history raises concern for lung malignancy. Small cell lung cancer can present with paraneoplastic Cushing’s syndrome and could explain the patient’s volume overload. Finally, human immunodeficiency virus (HIV) serostatus should be determined in all adult patients.

His temperature was 37 °C, blood pressure 161/69 mm Hg, pulse 104 beats per minute, respiratory rate 20 breaths per minute, and oxygen saturation was 95% on room air. On examination, he was an unkempt, ill-appearing man. He had poor dentition, but no oral ulcers or petechiae. Pulmonary exam revealed diffuse rhonchi and scattered wheezes. He developed dyspnea after speaking 2 sentences. Cardiovascular exam showed regular tachycardia, normal S1 and S2 heart sounds, and both an S3 and S4 gallop. A grade III/VI holosystolic murmur at the left lower sternal border with apical radiation, and an early, grade III/IV diastolic murmur at the right upper sternal border were present. Neck exam showed jugular venous distention (JVD) 8 cm above the right clavicle. Lower extremities showed symmetric 3+ pitting edema to the knees. His abdomen was soft, nondistended, and without hepatosplenomegaly. There was no lymphadenopathy. Skin exam showed small, healed excoriations on his anterior shins, forearms, and knuckles. There were no petechiae, Janeway lesions, or Osler’s nodes.

These exam findings change the differential substantially. New regurgitant murmurs strongly suggest infective endocarditis (IE). A diastolic murmur is never normal and suggests aortic regurgitation. The holosystolic murmur with apical radiation suggests mitral regurgitation. Cutaneous stigmata should always be sought, but are found in fewer than half of cases of subacute IE, and their absence does not rule out this diagnosis. Disheveled hygiene and excoriations suggest a skin source of infection, and poor dentition is concerning for an oral source. For the moment, the source does not matter. His clinical condition is serious: tachycardia, JVD, edema, and two-sentence dyspnea indicate congestive heart failure. Even before labs and imaging return, inpatient admission is warranted.

Serum sodium concentration was 140 mEq/L, potassium 3.7 mEq/L, chloride 103 mEq/L, bicarbonate 30 mEq/L, blood urea nitrogen (BUN) 26 mg/dL, creatinine 0.8 mg/dL, glucose 120 mg/dL, and calcium 9.0 mg/dL. The white blood cell count was 7100/µL, hemoglobin 11.8 g/dL, and platelet count 101 K/µL. Brain natriuretic peptide (BNP) was 785 pg/mL (reference range 0-100 pg/mL), aspartate aminotransferase 77 U/L, alanine aminotransferase 57 U/L, alkaline phosphatase 125 U/L, total bilirubin 0.8 mg/dL, total protein 7.7 g/dL, and albumin 3.7 g/dL. Erythrocyte sedimentation (ESR) rate was 38 mm/hour (reference range 0-25 mm/hour) and C-reactive protein (CRP) 0.62 mg/dL (reference range <1.0 mg/dL). Cardiac troponins were 0.03 ng/mL (reference range <0.04 ng/mL). Screening for HIV was negative. Urinalysis showed trace blood by dipstick, but no glucose, protein, dysmorphic red blood cells, or casts. Two sets of peripheral blood cultures were drawn. Two sets of blood cultures from his previous ED visits were negative (drawn 6 and 14 days prior).

These laboratory values are nonspecific, and the differential remains unchanged, with top concern for IE, then lung abscess. Ideally, 3 sets of cultures drawn greater than 12 hours apart should be obtained because the likelihood of pathogen detection rises with the volume of blood tested. Thrombocytopenia and microscopic hematuria suggest microangiopathic hemolytic anemia, and a peripheral blood smear should be examined for schistocytes. Glomerulonephritis from immune complex deposition can occur in IE, but is unlikely with a normal serum creatinine and lack of proteinuria, dysmorphic red blood cells, or casts. The elevated BNP suggests cardiac strain due to a regurgitant valve. ESR and CRP are rarely helpful in this situation, and perhaps previous treatment with azithromycin and steroids prevented significant elevation.

Due to concern for IE, blood cultures were repeated, and IV vancomycin, IV ceftriaxone, and IV gentamicin were initiated. Azithromycin and prednisone were discontinued. His respiratory status continued to improve with IV furosemide, albuterol, ipratropium, and supportive care.

TTE inadequately visualizes the mitral valve, but is useful for tricuspid valve assessment because the right ventricle is closer to the chest wall. Transesophageal echocardiography (TEE) is indicated for a more detailed assessment of the left heart valves for vegetations and perivalvar abscesses. The new regurgitant murmurs satisfy a major criterion of the modified Duke criteria, and valvar vegetations suggests IE. He does not yet fulfill the other major modified Duke criterion for IE, nor does he satisfy enough minor criteria because there are no diagnostic vascular, microbiologic, or immunologic phenomena. However, no diagnostic rubric is perfect, and these results should not supersede clinical judgment. Despite the absence of positive cultures, the concern for bacterial IE remains high. The absence of embolic phenomena fits best with subacute rather than acute IE. Three negative blood cultures to date suggest a fastidious organism is responsible, although oral flora remain on the differential.

There is rarely a need to “hold” blood cultures for prolonged periods because modern instruments typically yield positive results within 7 days for most bacteria, including the HACEK group. Blood culture-negative endocarditis (BCNE) is considered when 3 sets of cultures are negative for at least 5 days. In this situation, one should consider other microorganisms based on the patient’s exposure history. Only certain species with complex growth requirements, such as Brucella and Bartonella, require prolonged holds. Revisiting his exposure history would be helpful in deciding whether serologic testing warranted. If he recalls exposure to parturient animals, then Coxiella is worth pursuing; if he has been bitten by lice, then B. quintana rises as a possibility; if the scratches on his limbs are from recent cat scratches, then B. henselae becomes more likely. Both C. burnetti and Bartonella endocarditis might be partially treated by his courses of azithromycin, confounding the picture.

If the infectious work-up is ultimately negative, one could then consider other etiologies of endocarditis, such as nonbacterial thrombotic endocarditis, which is seen in the context of malignancy and systemic lupus erythematosus (Libman-Sacks endocarditis). Other mimickers of IE include myxomatous valve degeneration, ruptured mitral chordae, and eosinophilic heart disease (Löffler’s endocarditis).

A transesophageal echocardiogram confirmed the presence of small echodensities on the aortic valve’s right and left coronary cusps, consistent with vegetations. The vegetation on the anterior leaflet of the mitral valve from the TTE also showed an aneurysm with a small perforation (Figure 2).

The combination of aortic regurgitation and the mitral valve aneurysm supports IE, because the aortic regurgitant jet directly strikes the anterior mitral valve leaflet, seeding the valve with infection from the aortic cusps. A positive serum PCR is diagnostic, but if it had been negative or unavailable, the serology would remain very helpful. In this context, the elevated IgG titer implicates B. henselae, the agent responsible for cat scratch disease (CSD). Out of context, these titers would not be diagnostic, because anti-Bartonella IgG may be increased due to a prior subclinical episode of CSD. Anti-Bartonella IgM is an unreliable indicator of recent infection because it may wane within weeks, and this IgG titer is higher than what is observed with most remote infections.

Revisiting previous cat exposure is warranted. He lost his cat to an illness 3 years prior, however it would be appropriate to inquire about other animals, such as a stray kitten with fleas, which his skin scratches suggest. Up to 50% of all cats in flea endemic regions harbor Bartonella and are asymptomatic. Rarely, dogs can serve as reservoirs of this organism, with a presumed transmission route via flea, louse, or tick. Regardless of the route of infection, treatment should be focused on B. henselae IE.

Azithromycin can treat CSD, and its use for his presumed COPD exacerbation may have temporized his infection. However, azithromycin monotherapy is not recommended for B. henselae IE. Treatment is usually with 2 antibiotics, including an aminoglycoside (gentamicin) for the first 2 weeks, combined with either a tetracycline, a macrolide, or a beta-lactam for a minimum of 4-6 weeks. Oral rifampin can be considered if gentamicin is not tolerated. After completing IV treatment, an additional 6 months of oral doxycycline or azithromycin should be considered, especially for those who have not undergone valve surgery.

His symptoms are compatible with subacute IE, which is typically more difficult to diagnose than acute IE due to its insidious onset. He meets criteria for blood culture negative IE based on 3 sets of negative blood cultures for greater than 5 days and major criteria for IE. The pathologic changes are consistent with B. henselae infection.

The incidence of IE in the United States is 40,000 cases per year¹ with an in-hospital mortality of 15%-20% and a 1-year mortality of up to 40%.^2,3 Five to 20% of patients with IE never develop positive blood cultures⁴ due to receipt of antibiotics prior to culture, inadequate microbiologic testing, or infection caused by noncultivable bacteria (eg, Tropheryma whipplei), fastidious extracellular bacteria (eg, HACEK group and nutritionally variant streptococci), or by intracellular pathogens with complex nutrient requirements (eg, Bartonella, Chlamydia, Brucella, or Coxiella). Previous administration of antibiotics reduces the likelihood of isolating an organism by 35%-40%.⁵ Patients meeting criteria for BCNE should prompt consideration of serologic testing. The most prevalent pathogens vary globally, and incidence data in the US is scarce. Worldwide, the majority of BCNE cases are caused by Coxiella, Bartonella, and Brucella species.^6,7

When clinical suspicion for IE remains high despite negative cultures, detailed history can uncover clues and guide additional testing. For example, contact with contaminated milk products or farm animals are associated with Brucella, Coxiella, and Erysipelothrix species IE.^7,8 Bartonella species are zoonotic gram-negative bacilli with a tropism for endothelial cells and are transmitted by arthropod vectors (ie, fleas, lice, ticks, and sandflies), cat scratches, or cat bites. Bartonella may account for 3%-4% of all cases of IE, most of which are due to B. henselae and B. quintana.^{7, 9} Underlying heart valve disease, alcoholism, cirrhosis, and homelessness are associated with B. henselae endocarditis.¹⁰

Diagnostic criteria are lacking for B. henselae IE, and the modified Duke criteria is of limited utility for diagnosing Bartonella IE because blood cultures are often negative and echocardiographic evidence of vegetation is not always apparent. Serology plays a critical role in the diagnosis of Bartonella infections. The addition of positive serology, Western blot or PCR for B. henselae and B. quintana as a major criterion in the modified Duke criteria for IE has been proposed but has not yet been formally accepted.⁹ For B. henselae IE, an IgG titer of ≥1:800 has been recommended as a cutoff for subacute IE because it combines a high specificity and positive predictive value along with reasonable sensitivity and negative predictive value in this situation.⁹ The humoral immune response rises over time, and thus acute IE due to Bartonella may not generate a substantial IgG titer. Interestingly, because of the indolent nature of this pathogen, most cases of IE present once IgG titers have begun to rise. Serum PCR testing has shown a sensitivity and specificity of 58% and 100%, respectively.¹¹ Isolation by blood culture requires specific growth media and prolonged incubation, with a sensitivity as low as 20% and 30% for blood and tissue, respectively.¹⁰ The microbiology laboratory should be notified of suspected Bartonella to intensify efforts to cultivate this organism. If infection with Coxiella or Brucella is suspected, the lab should also be informed, both to increase diagnostic yield and to trigger enhanced biosafety precautions when handling the specimens. Despite attempts to optimize the yield, up to 75% of Bartonella IE may remain culture negative,^12,13 making it difficult to meet the current major modified Duke criterion of positive blood cultures. H&E staining of valve tissue infected with Bartonella commonly reveals increased inflammation, fibrosis, and calcified granulomas relative to endocarditis from other causes.¹⁴ The Warthin-Starry silver stain can identify small, darkly staining bacteria in more than 75% of Bartonella endocarditis; however, this stain is not specific for Bartonella species.⁹

This case highlights the challenge of diagnosing subacute IE because this patient received antibiotics and steroids prior to presentation, clouding the clinical picture. Although he did not exhibit textbook signs of endocarditis, his symptoms (new onset heart failure and new regurgitant murmurs) prioritized the diagnosis. The combination of elevated serum titers, positive PCR, valve granulomas and abscesses on TEE, and pathology findings led the discussant to the correct diagnosis. Scratching beneath the surface revealed his penchant for cats, but this was only considered a key epidemiological feature later in his clinical course.

• Subacute IE typically presents with indolent constitutional symptoms over a course of weeks to months, whereas acute IE causes a rapid onset of fevers, rigors, and is more likely to exhibit embolic phenomena.
• Epidemiologic features specific to Bartonella species include alcoholism, cirrhosis, dog or cat exposure, homelessness, and body lice, and should be considered in suspected cases of BCNE.
• If suspicion for endocarditis remains high and animal exposure is elicited, then serologic and PCR testing for fastidious organisms should be strongly considered. The most common causes of BCNE include Coxiella, Bartonella, and Brucella species.
• The modified Duke criteria do not incorporate Bartonella within the diagnostic schema. Presentation is usually late and often requires valve replacement.

Acknowledgments

The authors thank Dr. Michael Pfeiffer from the Pennsylvania State Hershey Heart and Vascular Institute for providing his expertise in diagnostic echocardiography.

Disclosure

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

A proposed metric to quantify the impact of an antimicrobial stewardship program (ASP) is using changes in the antibiotic days of therapy (DOT) per 1000 patient-days, which is the total number of days any dose of an antibiotic is administered during a specified time period, standardized by the number of patient-days.¹ Although DOT is useful for comparing antibiotic use among hospitals or time periods, this metric is a composite result of an ASP’s often multifaceted approach to improving antibiotic use. Thus, DOT provides a loose estimate of the direct impact of specific ASP activities and does not quantify the amount of antibiotics directly avoided or direct cost savings on the patient level. To ameliorate this, we reviewed our institution’s ASP prospective audit and feedback (PAF) and applied a novel metric, days of therapy avoided (DOTA), to calculate the number of antibiotic days avoided that directly result from our ASP’s actions targeting antibiotic stoppage. From DOTA, we also calculate attributable cost savings.

To quantify the direct impact of PAF, DOTA (Table) was calculated. Antibiotic costs avoided were calculated by multiplying the average wholesale price (AWP) per day (range: $0.44-$534; mean: $67.85) by DOTA. This calculation was done twice under 2 assumptions: that PAF led to the prevention of (1) 1 more day of antibiotic prescription and (2) the remainder of the documented or assumed LOT.

Over 4 years, the ASP made 1594 interventions to stop antibiotics. Accepted interventions totaled 1151 (72%): 513 (44.5%) for NI and 638 (55.4%) for TC, involving 431 and 575 unique patients, respectively. Nearly half (45.8%) of the NI interventions targeted asymptomatic bacteriuria, whereas respiratory tract infections were the most common (42.2%) indication for the TC intervention.

Under the most conservative assumption that each accepted PAF recommendation avoided 1 day of unnecessary antibiotics, we estimated a total of 1151 DOTA; 690 (59.9%) were intravenous antibiotics. The average DOT on which the PAF note was written was 3.07 ± 1.69 for NI and 6.38 ± 2.73 for TC. A planned LOT was documented for only 36.7% of the courses. On the basis of documented or assumed LOT, we estimate that the NI and TC interventions led to between 1077 and 2826 DOTA and between 397 and 1598 DOTA, respectively. Potential fluoroquinolone DOTA ranged from 300 to 1126; for third- and fourth-generation cephalosporins, there were 314 to 1017 DOTA.

Using the conservative estimate of 1151 DOTA, the costs avoided totaled $16,700, which includes $10,700 for intravenous antibiotics. When the AWP per day of each antibiotic was applied to the remaining LOTs avoided, the maximum potential cost savings was $67,100. Additional cost savings may have been realized if indirect expenses, such as pharmacy preparation and nursing administration time or costs of medical supplies, were evaluated.

We investigated DOTA as a measure of the direct patient-level and intervention-specific impact of an ASP’s PAF. DOTA may be useful for ASPs with limited access to an electronic record or electronically generated DOT reports because DOTA and cost savings can be tracked manually and prospectively with each accepted intervention. DOTA can also help ASPs identify which clinical conditions are responsible for the most antibiotic overuse, and thus may benefit from the development of clinical treatment guidelines. We found that the highest yield areas for DOTA were targeting asymptomatic bacteriuria (NI) and respiratory infections (TC). In doing so, these have also succeeded in reducing high-risk, broad-spectrum antimicrobials, such as fluoroquinolones and advanced-generation cephalosporins. Further research is needed to assess if DOTA correlates with other ASP metrics and clinical outcomes; however, current evidence supports that reducing unnecessary antibiotic use is fundamental to reducing antibiotic resistance and adverse events.¹⁰

The limitations of measuring DOTA include time consumption, particularly if not collected prospectively. However, we make several conclusions. ASP PAF stopping antibiotics was well accepted and reduced antibiotic use. Second, calculating DOTA requires little technology and only knowledge of the planned LOT and drug costs. DOTA also identifies which infectious indications to focus PAF efforts on and gain the greatest impact. Overall, DOTA is a simple, useful, and promising measurement of the direct antibiotic and economic impacts of specific ASP PAF and warrants further investigation as an ASP metric.

Acknowledgments

The authors thank the patients and RGH staff, particularly the departments of infectious diseases, pharmacy, and internal medicine, for their support.

Disclosure

The authors declare no conflicts of interest. This study was previously presented in poster form at the Society for Healthcare Epidemiology of America Spring Conference in St. Louis, Missouri (March 29-31, 2017).

A proposed metric to quantify the impact of an antimicrobial stewardship program (ASP) is using changes in the antibiotic days of therapy (DOT) per 1000 patient-days, which is the total number of days any dose of an antibiotic is administered during a specified time period, standardized by the number of patient-days.¹ Although DOT is useful for comparing antibiotic use among hospitals or time periods, this metric is a composite result of an ASP’s often multifaceted approach to improving antibiotic use. Thus, DOT provides a loose estimate of the direct impact of specific ASP activities and does not quantify the amount of antibiotics directly avoided or direct cost savings on the patient level. To ameliorate this, we reviewed our institution’s ASP prospective audit and feedback (PAF) and applied a novel metric, days of therapy avoided (DOTA), to calculate the number of antibiotic days avoided that directly result from our ASP’s actions targeting antibiotic stoppage. From DOTA, we also calculate attributable cost savings.

To quantify the direct impact of PAF, DOTA (Table) was calculated. Antibiotic costs avoided were calculated by multiplying the average wholesale price (AWP) per day (range: $0.44-$534; mean: $67.85) by DOTA. This calculation was done twice under 2 assumptions: that PAF led to the prevention of (1) 1 more day of antibiotic prescription and (2) the remainder of the documented or assumed LOT.

Over 4 years, the ASP made 1594 interventions to stop antibiotics. Accepted interventions totaled 1151 (72%): 513 (44.5%) for NI and 638 (55.4%) for TC, involving 431 and 575 unique patients, respectively. Nearly half (45.8%) of the NI interventions targeted asymptomatic bacteriuria, whereas respiratory tract infections were the most common (42.2%) indication for the TC intervention.

Under the most conservative assumption that each accepted PAF recommendation avoided 1 day of unnecessary antibiotics, we estimated a total of 1151 DOTA; 690 (59.9%) were intravenous antibiotics. The average DOT on which the PAF note was written was 3.07 ± 1.69 for NI and 6.38 ± 2.73 for TC. A planned LOT was documented for only 36.7% of the courses. On the basis of documented or assumed LOT, we estimate that the NI and TC interventions led to between 1077 and 2826 DOTA and between 397 and 1598 DOTA, respectively. Potential fluoroquinolone DOTA ranged from 300 to 1126; for third- and fourth-generation cephalosporins, there were 314 to 1017 DOTA.

Using the conservative estimate of 1151 DOTA, the costs avoided totaled $16,700, which includes $10,700 for intravenous antibiotics. When the AWP per day of each antibiotic was applied to the remaining LOTs avoided, the maximum potential cost savings was $67,100. Additional cost savings may have been realized if indirect expenses, such as pharmacy preparation and nursing administration time or costs of medical supplies, were evaluated.

We investigated DOTA as a measure of the direct patient-level and intervention-specific impact of an ASP’s PAF. DOTA may be useful for ASPs with limited access to an electronic record or electronically generated DOT reports because DOTA and cost savings can be tracked manually and prospectively with each accepted intervention. DOTA can also help ASPs identify which clinical conditions are responsible for the most antibiotic overuse, and thus may benefit from the development of clinical treatment guidelines. We found that the highest yield areas for DOTA were targeting asymptomatic bacteriuria (NI) and respiratory infections (TC). In doing so, these have also succeeded in reducing high-risk, broad-spectrum antimicrobials, such as fluoroquinolones and advanced-generation cephalosporins. Further research is needed to assess if DOTA correlates with other ASP metrics and clinical outcomes; however, current evidence supports that reducing unnecessary antibiotic use is fundamental to reducing antibiotic resistance and adverse events.¹⁰

The limitations of measuring DOTA include time consumption, particularly if not collected prospectively. However, we make several conclusions. ASP PAF stopping antibiotics was well accepted and reduced antibiotic use. Second, calculating DOTA requires little technology and only knowledge of the planned LOT and drug costs. DOTA also identifies which infectious indications to focus PAF efforts on and gain the greatest impact. Overall, DOTA is a simple, useful, and promising measurement of the direct antibiotic and economic impacts of specific ASP PAF and warrants further investigation as an ASP metric.

Acknowledgments

The authors thank the patients and RGH staff, particularly the departments of infectious diseases, pharmacy, and internal medicine, for their support.

Disclosure

The authors declare no conflicts of interest. This study was previously presented in poster form at the Society for Healthcare Epidemiology of America Spring Conference in St. Louis, Missouri (March 29-31, 2017).

A proposed metric to quantify the impact of an antimicrobial stewardship program (ASP) is using changes in the antibiotic days of therapy (DOT) per 1000 patient-days, which is the total number of days any dose of an antibiotic is administered during a specified time period, standardized by the number of patient-days.¹ Although DOT is useful for comparing antibiotic use among hospitals or time periods, this metric is a composite result of an ASP’s often multifaceted approach to improving antibiotic use. Thus, DOT provides a loose estimate of the direct impact of specific ASP activities and does not quantify the amount of antibiotics directly avoided or direct cost savings on the patient level. To ameliorate this, we reviewed our institution’s ASP prospective audit and feedback (PAF) and applied a novel metric, days of therapy avoided (DOTA), to calculate the number of antibiotic days avoided that directly result from our ASP’s actions targeting antibiotic stoppage. From DOTA, we also calculate attributable cost savings.

To quantify the direct impact of PAF, DOTA (Table) was calculated. Antibiotic costs avoided were calculated by multiplying the average wholesale price (AWP) per day (range: $0.44-$534; mean: $67.85) by DOTA. This calculation was done twice under 2 assumptions: that PAF led to the prevention of (1) 1 more day of antibiotic prescription and (2) the remainder of the documented or assumed LOT.

Over 4 years, the ASP made 1594 interventions to stop antibiotics. Accepted interventions totaled 1151 (72%): 513 (44.5%) for NI and 638 (55.4%) for TC, involving 431 and 575 unique patients, respectively. Nearly half (45.8%) of the NI interventions targeted asymptomatic bacteriuria, whereas respiratory tract infections were the most common (42.2%) indication for the TC intervention.

Under the most conservative assumption that each accepted PAF recommendation avoided 1 day of unnecessary antibiotics, we estimated a total of 1151 DOTA; 690 (59.9%) were intravenous antibiotics. The average DOT on which the PAF note was written was 3.07 ± 1.69 for NI and 6.38 ± 2.73 for TC. A planned LOT was documented for only 36.7% of the courses. On the basis of documented or assumed LOT, we estimate that the NI and TC interventions led to between 1077 and 2826 DOTA and between 397 and 1598 DOTA, respectively. Potential fluoroquinolone DOTA ranged from 300 to 1126; for third- and fourth-generation cephalosporins, there were 314 to 1017 DOTA.

Using the conservative estimate of 1151 DOTA, the costs avoided totaled $16,700, which includes $10,700 for intravenous antibiotics. When the AWP per day of each antibiotic was applied to the remaining LOTs avoided, the maximum potential cost savings was $67,100. Additional cost savings may have been realized if indirect expenses, such as pharmacy preparation and nursing administration time or costs of medical supplies, were evaluated.

We investigated DOTA as a measure of the direct patient-level and intervention-specific impact of an ASP’s PAF. DOTA may be useful for ASPs with limited access to an electronic record or electronically generated DOT reports because DOTA and cost savings can be tracked manually and prospectively with each accepted intervention. DOTA can also help ASPs identify which clinical conditions are responsible for the most antibiotic overuse, and thus may benefit from the development of clinical treatment guidelines. We found that the highest yield areas for DOTA were targeting asymptomatic bacteriuria (NI) and respiratory infections (TC). In doing so, these have also succeeded in reducing high-risk, broad-spectrum antimicrobials, such as fluoroquinolones and advanced-generation cephalosporins. Further research is needed to assess if DOTA correlates with other ASP metrics and clinical outcomes; however, current evidence supports that reducing unnecessary antibiotic use is fundamental to reducing antibiotic resistance and adverse events.¹⁰

The limitations of measuring DOTA include time consumption, particularly if not collected prospectively. However, we make several conclusions. ASP PAF stopping antibiotics was well accepted and reduced antibiotic use. Second, calculating DOTA requires little technology and only knowledge of the planned LOT and drug costs. DOTA also identifies which infectious indications to focus PAF efforts on and gain the greatest impact. Overall, DOTA is a simple, useful, and promising measurement of the direct antibiotic and economic impacts of specific ASP PAF and warrants further investigation as an ASP metric.

Acknowledgments

The authors thank the patients and RGH staff, particularly the departments of infectious diseases, pharmacy, and internal medicine, for their support.

Disclosure

The authors declare no conflicts of interest. This study was previously presented in poster form at the Society for Healthcare Epidemiology of America Spring Conference in St. Louis, Missouri (March 29-31, 2017).