Associations between Social Determinants of Health and Posthospitalization Rehabilitation among Critically Ill Older Adults

Rationale: New or worsened disability often develops in older adults who survive hospitalization with a stay in the intensive care unit (ICU); those with social vulnerabilities experience a greater burden of disability. It is unknown whether social determinants of health (SDOHs) are associated with disparities in delivery of posthospitalization rehabilitation. Objectives: To evaluate whether SDOHs are associated with disparities in delivery of postdischarge rehabilitation services to older adults in skilled nursing facilities (SNFs), in home health, and in outpatient rehabilitation facilities in the 100 days after an ICU hospitalization. Methods: In this observational cohort study, we used data from the National Health and Aging Trends Study with links to Medicare inpatient claims (2011–2019), the Minimum Data Set for SNFs, the Outcomes and Assessment Information Set for home health, and outpatient claims files. We identified older adults with an ICU hospitalization who survived to discharge. The main outcome was rehabilitation delivery in the first 100 days after discharge, which was ascertained as minutes of physical and/or occupational therapy in SNFs and receipt of any physical and/or occupational therapy in home health and outpatient settings. We constructed multivariable regression models to evaluate the association between SDOHs (socioeconomic disadvantage, race and ethnicity, education, and limited English proficiency) and rehabilitation delivery, adjusting for demographic characteristics, hospitalization, and posthospitalization characteristics. Results: We identified 295 ICU hospitalizations with SNF stays, 205 with home health assessments, and 784 for which patients were alive for >7 days after discharge and thereby eligible for outpatient rehabilitation. Socioeconomic disadvantage, non-White race or Hispanic ethnicity, and lower level of education were associated with reduced delivery of in-home rehabilitation (adjusted odds ratios, 0.43 [95% confidence interval, 0.23–0.81], 0.39 [0.16–0.93], and 0.42 [0.18–0.99], respectively). Although the sample sizes in the respective cohorts were small, SDOHs were not associated with rehabilitation delivery in SNFs or outpatient settings. Conclusions: We found socioeconomic disadvantage, minoritized race and ethnicity, and lower education level were associated with reduced delivery of rehabilitation at home among older ICU survivors. Given its important role in promoting functional recovery, our findings warrant interventions to mitigate inequities in rehabilitation delivery at home after hospital discharge.