An empirical study of the spillover effects of workload on patient length of stay

We use two years of inpatient data from 243 California hospitals to quantify the relationship between hospital-level workload and patient length of stay (LOS), and its "spillover" effects across patient types. Patients are categorized as medical or surgical, and the effects of same type patient workload (occupancy) on LOS are analyzed. The analysis is repeated with workload replaced by other type patient occupancy, providing a "spillover" effect. We find that the effects of inpatient workload on LOS spillover across patient types, which we theorize results from most inpatients, regardless of type, utilizing the same shared resources (e.g. pharmacy and laboratory). These spillover effects remain even while we find that the effects of workload vary at different time points during a patient's stay: LOS increases as inpatient workload on the day of admission increases, while inpatient workload at the end of the stay has a U-shaped effect on LOS.