Predictors for Unplanned Readmissions within 18 Days after Hospital Discharge: a Retrospective Cohort Study.

ABSTRACT

Since the introduction of the reimbursement system based on diagnosis-related groups (DRG) in Swiss hospitals in 2012, most readmissions occurring within 18 days and appertaining to the same major diagnostic category (MDC) are merged and thus often reimbursed to a lesser extent. While readmissions reflect increased distress for patients and their relatives, the causes are mainly patient-related and difficult to influence. However, it may be possible to identify cases at higher risk for readmission. Therefore, the aim of this study was to find predictors for early readmissions in the same MDC, to identify high-risk index hospitalizations and possibly prevent unnecessary readmissions. The data of all patients admitted to the Clinic of Internal Medicine at the University Hospital of Basel, Switzerland, hospitalized for longer than 24 hours during the pre-DRG period between October 2009 and September 2010 were retrospectively collected. Data were examined for predictors of unplanned readmission within 18 days under the same MDC ('relevant readmission') by means of logistic regression. 7479 patients (median age 67.8 years, 56% male) were admitted to the Clinic of Internal Medicine, with 232 patients (3.1%) being readmitted at least once. Logistic regression revealed male sex (p =0.035) and a high number of prescribed drugs at discharge (p <0.005) as patient-related predictors. The MDCs respiratory system, cardiovascular system, and gastrointestinal/hepatobiliary system were identified as high-risk categories (each p <0.005). Age and length of index hospital stay added no significant explanatory value to the regression model. Unplanned readmissions under the same MDC within 18 days were infrequent and not related to patients' age or length of hospital stay. Overall, multimorbid patients, and hospitalizations regarding the cardiovascular, respiratory, or gastrointestinal system appear to be most at risk and should therefore be specifically targeted in the prevention of early readmissions.