Beyond Sepsis Survival: Post‐Acute Vulnerability Profiles in Older Home Health Patients

Abstract

ABSTRACT Background Older sepsis survivors face elevated rehospitalization risk shaped by multimorbidity and post‐sepsis vulnerability, yet traditional risk models evaluate these factors in isolation. To identify clinically meaningful profiles of coexisting rehospitalization risk factors among older sepsis survivors receiving home health care (HHC), we used clustering to characterize risk subgroups and examined their association with 30‐day rehospitalization. Methods A retrospective cohort study of Medicare fee‐for‐service beneficiaries aged 65 years or older who transitioned to HHC after hospitalization for sepsis between January 1, 2021 and December 31, 2022 in the United States. Partitioning around medoids clustering with Gower distance was used to identify risk clusters based on risk factors available during the index hospitalization and first post‐discharge HHC assessment. Multivariable logistic regression examined associations between cluster membership and all‐cause 30‐day rehospitalization, adjusting for contextual, HHC agency, and service delivery factors. Results Among 21,888 index stays, two rehospitalization risk clusters were identified. Both clusters demonstrated post‐acute vulnerability after sepsis hospitalization but differed in intensity and breadth of impairments. The higher vulnerability cluster had greater functional dependency, cognitive and behavioral impairment, urinary and bowel incontinence, difficulty following medical instructions, and prior healthcare utilization. Compared with the lower post‐acute vulnerability cluster, the higher post‐acute vulnerability cluster had a higher 30‐day rehospitalization rate (27.8% vs. 21.1%; p < 0.001) and significantly higher odds of 30‐day rehospitalization in unadjusted (OR 1.492; 95% CI 1.400–1.589; p < 0.001) and adjusted analyses (aOR 1.480; 95% CI 1.384–1.582; p < 0.001). Conclusion Among older sepsis survivors receiving HHC, rehospitalization risk clustered into two post‐acute vulnerability profiles that primarily reflected differences in the overall extent of multidimensional vulnerability. These profiles may help characterize co‐occurring post‐acute needs and support future evaluation of whether cluster‐informed care coordination or stratified post‐acute care can improve outcomes.