Sooner or Later? Health Information Technology, Length of Stay, and Readmission Risk

Abstract

Adequate length of stay (LOS) during hospitalization is not only a critical determinant of quality of care, but can be a useful predictor of the risk of future readmissions. Recent studies have shown alarming evidence that the United States leads developed nations in terms of shorter hospital stays, rendering patients with greater risk of future readmissions. We focus on deviation between hospital LOS and the geometric mean LOS (GMLOS), a guideline for care delivery stipulated by the Centers for Medicare and Medicaid Services (CMS). Our objective is to establish the relationship between LOS and readmission risk, examine the role of health IT in reducing the deviation of LOS, and address cost trade‐offs between early discharges and readmissions. Based on a large panel of congestive heart failure (CHF) patients during a 4‐year period, we find that implementation of health IT applications is associated with a reduction in the deviation between LOS and GMLOS. This deviation is associated with a significant increase in future 30‐day readmission risk. Patients whose inpatient stays are shorter than the GMLOS guideline, by 1 day or more, are likely to exhibit a 1.1% greater risk of readmission, and even shorter stays are likely to further exacerbate their readmission risk. Further, we find that the total readmission costs are much higher than the costs incurred in keeping patients longer during their initial hospital visit. Our results have important policy implications for providers and hospital administrators with respect to their discharge policies of CHF patients.