The Centers for Medicare & Medicaid Services (CMS) uses 30-day readmission rates to measure quality of care and guide pay-for-performance. The CMS tracks readmissions following index hospitalizations for acute myocardial infarction (AMI), heart failure, chronic obstructive pulmonary disease (COPD), and pneumonia because hospitalizations for these conditions are frequent and account for a large proportion of readmissions.1 The proportion and cost of unplanned readmissions following sepsis hospitalizations are not known. We hypothesized that sepsis hospitalizations account for a higher proportion of unplanned 30-day readmissions than hospitalizations for AMI, heart failure, COPD, and pneumonia in the United States.
Methods ∣
We analyzed data from the 2013 Nationwide Readmissions Database, which aggregates acute care hospitalizations from 21 states and represents inpatient use for 49% of the US population.2 This study was deemed exempt by the University of Pittsburgh institutional review board. For adults aged 18 years or older, we identified index admissions for medical reasons (using diagnosis-related group codes) that were followed by an unplanned hospital readmission within 30 days of discharge. We excluded index admissions if patients were not discharged alive, left against medical advice, were transferred to another acute care hospital, or occurred during December (because data for 30-day readmissions were not available). Patients with more than 1 admission-readmission episode could be included.
For these patients, we determined the weighted proportions of index admissions due to sepsis, AMI, COPD, heart failure, and pneumonia. Because patients hospitalized for AMI, heart failure, COPD, and pneumonia may have sepsis, we calculated the proportion of sepsis cases that overlapped with other conditions. We conducted primary analyses using International Classification of Diseases, Ninth Revision, Clinical Modification codes in 10 discharge diagnoses fields to identify sepsis and used CMS methods to identify the remaining 4 conditions, which use only the primary discharge diagnosis field.3,4 We conducted sensitivity analyses using stringent criteria used by CMS for sepsis and extended the CMS criteria to identify the remaining 4 conditions to 10 discharge diagnoses fields. We estimated costs for readmissions using previous approaches.5
We performed pairwise comparisons of proportions of index admissions, length of stay, and cost for each of the 5 conditions using multinomial logistic, negative binomial, and γ regression, respectively. For all analyses, robust standard errors were used, and 2-sided P values less than .005 were considered significant to account for multiple comparisons. All statistical analyses were performed using SAS (SAS Institute), version 9.3, and Stata (StataCorp), version 13.1.
Results ∣
Among 14 325 172 hospitalizations, we identified 1 187 697 index admissions for medical reasons that were associated with an unplanned 30-day readmission. Of those, 147 084 (12.2%; 95% CI, 11.9%-12.4%) had a diagnosis of sepsis, 15 001 (1.3%; 95% CI, 1.2%-1.3%) AMI, 79 480 (6.7%; 95% CI, 6.5%-6.8%) heart failure, 54 396 (4.6%; 95% CI, 4.5%-4.8%) COPD, and 59 378 (5.0%; 95% CI, 5.0%-5.3%) pneumonia. Among sepsis index admissions, 1061 (0.7%) also had diagnostic codes that met CMS criteria for AMI, 5063 (3.4%) heart failure, 4829 (3.3%) COPD, and 11 093 (7.5%) pneumonia.
The mean length of stay for unplanned readmissions following sepsis hospitalization was longer than readmissions following AMI, heart failure, COPD, and pneumonia (Table). The estimated mean cost per readmission was highest for sepsis compared with the other diagnoses ($10 070 [95% CI, $10 021-$10 119] for sepsis, $8417 [95% CI, $8355-$8480] for COPD, $9051 [95% CI, $8990-$9113] for heart failure, $9424 [95% CI, $9279-$9571] for AMI, and $9533 [95% CI, $9466-$9600] for pneumonia; P < .005 for all pairwise comparisons). Sepsis remained a leading cause of readmissions and cost in sensitivity analyses using the CMS sepsis criteria and extending the CMS criteria for AMI, heart failure, COPD, and pneumonia to 10 discharge diagnoses fields (Table).
Table.
Length of Stay and Cost for Unplanned 30-Day Readmissions After an Index Admission for Sepsis, Acute Myocardial Infarction, Heart Failure, Pneumonia, and Chronic Obstructive Pulmonary Disease
| National Readmission Dataa | Weighted Proportion of Cases in the United States | ||||
|---|---|---|---|---|---|
| No. of All Index Admissions Readmitted Within 30 Days | Estimated Mean Length of Stay (95% CI), db | Estimated Mean Cost per Readmission (95% CI), $b | Percentage of Index Admissions Readmitted Within 30 Days (95% CI) | Percentage of Total Estimated Cost of All Readmissions (95% CI) | |
| Admissions associated with 30 d readmission | 1 187 697 | 6.4 (6.4-6.5) | 8242 (8225-8258) | NA | 100.0 |
| Primary Analyses c | |||||
| Sepsis | 147 084 | 7.4 (7.3-7.4) | 10 070 (10 021-10 119) | 12.2 (11.9-12.4) | 14.5 (14.2-14.8) |
| Acute myocardial infarction | 15 001 | 5.7 (5.6-5.8) | 9424 (9279-9571) | 1.2 (1.2-1.3) | 1.4 (1.3-1.5) |
| Heart failure | 79 480 | 6.4 (6.4-6.5) | 9051 (8990-9113) | 6.7 (6.5-6.8) | 7.5 (7.3-7.7) |
| Pneumonia | 59 378 | 6.7 (6.6-6.7) | 9533 (9466-9600) | 5.2 (5.0-5.3) | 5.5 (5.4-5.7) |
| Chronic obstructive pulmonary disease | 54 396 | 6.0 (5.9-6.0) | 8417 (8355-8480) | 4.6 (4.5-4.8) | 4.3 (4.1-4.4) |
| Sensitivity Analyses d | |||||
| Sepsis | 89 800 | 7.6 (7.6-7.7) | 10 828 (10 760-10 897) | 7.3 (7.1-7.5) | 9.1 (8.8-9.4) |
| Acute myocardial infarction | 21 281 | 6.0 (5.9-6.1) | 9530 (9408-9654) | 1.8 (1.7-1.8) | 2.0 (1.9-2.1) |
| Heart failure | 236 636 | 6.5 (6.5-6.5) | 9248 (9211-9285) | 20.0 (19.6-20.4) | 22.1 (21.6-22.6) |
| Pneumonia | 130 904 | 6.9 (6.9-7.0) | 9749 (9700-9797) | 11.1 (10.9-11.4) | 12.5 (12.2-12.8) |
| Chronic obstructive pulmonary disease | 201 867 | 6.3 (6.3-6.4) | 8677 (8641-8713) | 17.4 (17-17.7) | 17.2 (16.7-17.7) |
Abbreviation: NA, not applicable.
All pairwise comparisons of proportions, length of stay, and cost for sepsis and each of the 4 conditions are significant at the P < .005 level.
Calculated by applying the inverse log function to the mean log cost, which was estimated using γ regression.
Performed using principal discharge fields for acute myocardial infarction, heart failure, pneumonia, chronic obstructive pulmonary disease, and 10 discharge diagnosis fields for sepsis.
Includes Centers for Medicare & Medicaid Services (CMS) approach to identify sepsis for the sepsis quality core measure denominator population6 and expansion of the CMS definition from the principal discharge diagnosis field to the first 10 discharge fields for acute myocardial infarction, heart failure, pneumonia, and chronic obstructive pulmonary disease.
Discussion ∣
Among medical conditions, sepsis is a leading cause of readmissions and associated costs. Adding sepsis to the Hospital Readmission Reduction Program may lead to development of new interventions to reduce unplanned readmissions and associated costs. This study is limited in that the National Readmissions Database uses state specific identifiers that cannot follow-up patients across states, which may underestimate readmission rates. In addition, readmission rates and cost estimates may vary based on different sepsis definitions.
Funding/Support:
Dr Yende was supported by grants R01GM097471 and R34GM107650 from the National Institute of General Medical Sciences.
Role of the Funder/Sponsor:
The National Institute of General Medical Sciences had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Footnotes
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Yende reported receiving grants from Bristol-Myers Squibb. No other disclosures were reported.
Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the US government.
References
- 1.Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med. 2009;360(14):1418–1428. [DOI] [PubMed] [Google Scholar]
- 2.Agency for Healthcare Research and Quality. HCUP Nationwide Readmissions Database. https://www.hcup-us.ahrq.gov/nrdoverview.jsp. Accessed August 8, 2016.
- 3.Centers for Medicare & Medicaid Services. 2016. Condition-specific measures updates and specifications report hospital-level 30-day risk-standardized readmission measures. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/HospitalQualityInits/Measure-Methodology.html. Accessed December 22, 2016.
- 4.Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29(7):1303–1310. [DOI] [PubMed] [Google Scholar]
- 5.Kahn JM, Le T, Angus DC, et al. ; ProVent Study Group Investigators. The epidemiology of chronic critical illness in the United States. Crit Care Med. 2015;43(2):282–287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.National Quality Forum. Severe sepsis and septic shock: management bundle. http://www.qualityforum.org/Projects/i-m/Infectious_Disease_Endorsement_Maintenance_2012/0500.aspx. Accessed December 7, 2016. [Google Scholar]