Predicting hospital readmissions in the Medicare population

Area of Science:

• Health Services Research
• Biostatistics
• Gerontology

Background:

• Hospital readmissions represent a significant cost and quality concern in healthcare.
• Predictive models are needed to identify patients at high risk for readmission.

Purpose of the Study:

• To develop and validate a multivariate logistic model identifying factors predicting hospital readmission within 60 days of discharge.
• To provide tools for healthcare professionals and organizations to improve patient outcomes.

Main Methods:

• Utilized a nationally random sample of Medicare beneficiaries.
• Employed multivariate logistic regression to identify significant predictors of readmission.

Main Results:

• Identified 10 statistically significant variables predicting 60-day hospital readmissions.
• Patient disease history and diagnosis emerged as the strongest predictors.

Conclusions:

• The developed model effectively predicts hospital readmissions.
• Findings can inform quality assessments by peer review organizations and enhance discharge planning by clinicians and social workers.