Existing studies of hospital readmissions typically focus on specific diagnoses, age groups, discharge dispositions, payer classes, or hospitals, and often use small samples. It is not clear how predictive models generated from such studies generalize across diseases, hospitals, or time periods. In this study, a logistic regression model of readmission risk within 30 days based on hospital administrative data was constructed and validated across hospitals and time periods. The hospitals included both general and specialty hospitals such as long-term care, women’s, and children’s hospitals. The administrative data included information on patient’s demographics, diagnoses, procedures, and discharge disposition. Derivation and validation samples for the cross-hospital analysis yielded C-statistics of 0.722 and 0.706, respectively. The cross-time period analysis yielded C-statistics from 0.736 to 0.755 for five derivation samples, and from 0.681 to 0.701 for fifteen validation samples. The findings indicate that a prediction model can be used with relative success to extrapolate beyond the estimation sample both in terms of hospital and time period. Such risk estimates can be used to inform discharge intervention decisions and increase care coordination.
30-day all-cause hospital readmissions, predictive analytics, logistic regression
Date of this Version
Operations Research, Systems Engineering and Industrial Engineering Commons, Statistics and Probability Commons