Estimating the impact of discharge to nursing home on readmission and mortality: a propensity score matched analysis

Background The hospital discharge process entails substantial challenges, particularly among older multimorbid patients with a high risk of subsequent readmission and/or death. Discharging hospitals often face a decision regarding whether such patients should be discharged to a short- or long-term stay at a nursing home, or sent home with community-based support services. Therefore, the aim of this study was to investigate the impact of these forms of post-hospital care on readmissions and mortality. Methods We used data from national Swedish health-, and social care registries from 2015 to 2019 to perform a retrospective analysis using a target trial emulation design. The study included patients above 65 years of age with multimorbidity who were community dwelling prior to hospitalization and subsequently discharged to either a nursing home or back to their own residence with formal home care interventions. To estimate the average treatment effect on the treated of nursing home vs. in-home care, each hospital discharge to a nursing home was matched to a discharge to home care using a propensity score generated by a gradient boosting model based on the patient’s previous medical and social care history. Outcomes were assessed using a double robust approach consisting of survival analysis with statistical adjustment for residual confounding. Results Hospital discharges to nursing homes had a lower risk of hospital readmission across 7-, 30- and 90-day endpoints, with hazard ratios of 0.64, 0.77, and 0.85 respectively. Short term (7 day) mortality was similar between hospital discharges to home care and nursing homes (HR 0.92), but higher at 30 and 90 days among nursing home discharges (HR 1.32 and 1.54, respectively). Discussion Our findings suggest benefits of nursing home care in terms of reduced readmission, especially in the short-term. While short term mortality was similar between groups, longer term mortality was higher for nursing home discharges, which may be the result of differences in care practices or unmeasured confounding. Taken together, the study suggests that the strongest impact of NH care is in the short term, while further research is needed to validate the results, particularly regarding longer-term mortality.