Background:
Reducing hospital readmission rates will decrease medical costs and improve public health. Moreover, the Centers for Medicare and Medicaid Services (CMS) will no longer reimburse hospitals for certain “preventable readmissions.” The objective of this study was to determine the predictors of 30‐day readmissions in patients discharged from the hospital internal medicine service.
Methods:
This was a retrospective cohort study in which 30‐day hospital readmissions were identified from billing data and medical record review at a tertiary‐care center, in Olmsted County, Minnesota. The subjects were patients discharged from the Division of Hospital Internal Medicine from January 1, 2008, to July 31, 2008. We measured the 30‐day hospital readmission rate and identified predictors of rehospitalization.
Results:
In total, 1576 patients were discharged by a hospital internal medicine physician from January 1, 2008, through July 31,2008. We excluded patients under 18 years of age, non‐Minnesota residents, and those who had not given prior research authorization or had died during their hospital stay. Among the remaining 1257 hospital visits, we identified 226 rehospitalizations within 30 days from discharge between January 1, 2008, and July 31, 2008. There were 289 deaths within the study period, and 61 patients died within 30 days of discharge. Patients were censored at death. This historical cohort of patients had an 18% cumulative probability of rehospitalization over the first 30 days from discharge. Univariate Cox proportional hazard models showed that male sex (HR 1.4, P = 0.01), more than 5 medications on admission (HR 1.5, P = 0.001), more than 5 medications at discharge (HR 1.6, P = 0.003), and discharge home with home health services (HR 1.9, P = 0.002), were significantly associated with an increased risk of rehospitalization. History of COPD (HR 1.5, P = 0.01), diabetes (HR 1.4, P = 0.02), renal disease (HR 1.6, P = 0.001), metastatic solid tumors (HR 1.9. P = 0.004), and the calculated Charlson index (HR 1.2, P = 0.001) were also associated with an increased risk of readmission. In multivariable analysis, male sex (HR 1.2. P = 0.01), Charison index (HR 1.1, P = 0.001), and discharge home with home health services (HR 1.2, P = 0.04) were strong predictors of readmission.
Conclusions:
This historical cohort of patients had an 18% cumulative probability of rehospitalization over the first 30 days from discharge. Male sex, Charlson index, and discharge home with home health services were the strongest predictors of rehospitalization and may be beneficial in early identification of patients at risk for readmission.
Author Disclosure:
A. Popa, none; K. Casey, none; E. Bergstralh, none; R. Gullerud, none; J. Huddleston, none.