Hospital readmissions following physician call system change: a comparison of concentrated and distributed schedules

Background — Physician call schedules are a critical element for medical practice and hospital efficiency. We compared readmission rates before and after a change in physician call system at Sunnybrook Health Sciences Centre.

Methods — We studied patients discharged over a decade (2004 through 2013) and identified whether or not each patient was readmitted within the subsequent 28 days. We excluded patients discharged for a surgical, obstetrical or psychiatric diagnosis. We used time-to-event analysis and time-series analysis to compare rates of readmission before and after the physician call system change (January 1, 2009).

Results — A total of 89,697 patients were discharged of whom 10,001 (11%) were subsequently readmitted and 4,280 died. The risk of readmission was increased by about 26% following physician call system change (9.7% vs 12.2%, p < 0.001). Time-series analysis confirmed a 26% increase in the readmission rate after call system change (95% confidence interval: 22% to 31%, p < 0.001). The increase in readmission rate after call system change persisted across patients with diverse ages, estimated readmission risks, and medical diagnoses. The net effect was equal to 7,240 additional patient-days in hospital following call system change. A modest increase was observed at a nearby acute care hospital that did not change physician call system and no increase in risk of death was observed with increased hospital readmissions.

Conclusion — We suggest that changes in physician call systems sometimes increase subsequent hospital readmission rates. Further reductions in readmissions may instead require additional resources or ingenuity.