Hospital readmissions and quality of care

Med Care. 1999 May;37(5):490-501. doi: 10.1097/00005650-199905000-00008.


Background: Readmission rates are often proposed as markers for quality of care. However, a consistent link between readmissions and quality has not been established.

Objective: To test the relation of readmission to quality and the utility of readmissions as hospital quality measures.

Subjects: One thousand, seven hundred and fifty-eight Medicare patients hospitalized in four states between 1991 to 1992 with pneumonia or congestive heart failure (CHF).

Design: Case control.

Measures: Related adverse readmissions (RARs), defined as readmissions that indicate potentially sub-optimal care during initial hospitalization, were identified from administrative data using readmission diagnoses and intervening time periods designated by physician panels. We used linear regression to estimate the association between implicit and explicit quality measures and readmission status (RARs, non-RAR readmissions, and nonreadmissions), adjusting for severity. We tested whether RARs were associated with inferior care and performed simulations to determine whether RARs discriminated between hospitals on the basis of quality.

Results: Compared with nonreadmitted pneumonia patients, patients with RARs had lower adjusted quality measured both by explicit (0.25 standardized units, P = 0.004) and implicit methods (0.17, P = 0.047). Adjusted differences for CHF patients were 0.17 (P = 0.048) and 0.20 (P = 0.017), respectively. In some analyses, patients with non-RAR readmissions also experienced lower quality. However, rates of inferior quality care did not differ significantly by readmission status, and simulations identified no meaningful relationship between RARs and hospital quality of care.

Conclusions: RARs are statistically associated with lower quality of care. However, neither RARs nor other readmissions appear to be useful tools for identifying patients who experience inferior care or for comparing quality among hospitals.

Publication types

  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Aged
  • Aged, 80 and over
  • Case-Control Studies
  • Feasibility Studies
  • Female
  • Heart Failure / therapy
  • Hospitals / standards*
  • Humans
  • Logistic Models
  • Male
  • Medicare / statistics & numerical data
  • Patient Readmission / statistics & numerical data*
  • Pneumonia / therapy
  • Quality Indicators, Health Care*
  • Reproducibility of Results
  • Risk Factors
  • Sensitivity and Specificity
  • United States