Is same-hospital readmission rate a good surrogate for all-hospital readmission rate?

BACKGROUND: The Centers for Medicare & Medicaid Services (CMS) readmission measure is based on all-cause readmissions to any hospital within 30 days of discharge. Whether a measure based on same-hospital readmission, an outcome that is easier for hospitals and some systems to track, could serve as a proxy for the all-hospital measure is not known. OBJECTIVES: Evaluate whether same-hospital readmission rate is a good surrogate for all-hospital readmission rate. RESEARCH DESIGN: The study population was derived from the Medicare inpatient, outpatient, and carrier (physician) Standard Analytic Files. Thirty-day risk-standardized readmission rates (RSRRs) for heart failure (HF) for both all-hospital readmission and same-hospital readmission were assessed by using hierarchical logistic regression models. SUBJECTS: The sample consisted of 501,234 hospitalizations in 4674 hospitals with at least 1 hospitalization. MEASURES: Thirty-day readmission was defined as occurrence of at least 1 hospitalization in any US acute care hospital for any cause within 30 days of discharge after an index hospitalization. Same-hospital readmission was considered if the patient was admitted to the hospital that produced the original discharge within 30 days. RESULTS: Overall, 80.9% of all HF readmissions occurred in the same-hospital, whereas 19.1% of readmissions occurred in a different hospital. The mean difference between all-versus same-hospital RSRR was 4.7 ± 1.0%, ranging from 0.9% to 10.5% across these hospitals with 25th, 50th, and 75th percentiles of 4.1%, 4.7%, and 5.2%, respectively, and was variable across the range of average RSRR. CONCLUSION: Same-hospital readmission rate is an unreliable and biased indicator of all-hospital readmission rate with limited value as a benchmark for quality of care processes.