BACKGROUND: The focus of health care reform is shifting from all-cause to potentially preventable readmissions. Potentially preventable within-stay readmission rates is a measure recently adopted by the Centers for Medicare and Medicaid Services for the Inpatient Rehabilitation Facility Quality Reporting Program.
OBJECTIVE: We examined the patient-level predictors of potentially preventable within-stay readmissions among Medicare beneficiaries receiving care in inpatient rehabilitation facilities. We also studied the reasons for readmissions and the risk-standardized variation across states.
DESIGN: Retrospective cohort study.
SETTING: Inpatient rehabilitation facilities.
PATIENTS: Medicare fee-for-service beneficiaries receiving inpatient rehabilitation after hospitalization in 2012-2013 (N = 345,697).
METHODS: Medicare claims were reviewed to identify potentially preventable readmissions occurring during inpatient rehabilitation.
MAIN OUTCOME MEASURES: (1) Observed rates and odds of potentially preventable within-stay readmissions by patient sociodemographic and clinical characteristics, (2) risk-standardized state rates, and (3) primary diagnoses for hospital readmissions.
RESULTS: The overall rate of potentially preventable within-stay readmissions was 3.5% (n = 11,945). Older age, male gender, hospitalizations during the previous 6 months, longer hospital lengths of stay, intensive care unit use, and number of comorbidities were associated with increased odds. Dual eligibility and disability status were not associated with increased odds. Greater functional scores at rehabilitation admission were associated with lower odds. Rates and odds varied across rehabilitation impairment groups. Risk-standardized state rates ranged from 3.1% to 4.1%. Readmissions for conditions reflecting inadequate management of infections (36.8%) were the most frequent and readmissions for inadequate injury prevention (6.1%) least frequent.
CONCLUSIONS: Potentially preventable within-stay readmissions may represent a target for inpatient rehabilitation care improvement. Our findings highlight the need for care coordination across providers. Future research should focus on care processes that reduce patients' risk of these potentially preventable rehospitalizations.
LEVEL OF EVIDENCE: III.