Baseline risk factors associated with all-cause early hospitalization of older patients following admission to Danish municipal temporary stays

Background Transitions from hospital to community are high-risk for older adults. In Denmark, municipal temporary stay (TS) facilities provide short-term, bed-based post-acute support, but determinants of early (re)hospitalization after TS admission are not well described. We estimated baseline risk factors for 30-day and 180-day hospitalization among TS patients. 
Methods We performed a register-based cohort study that includes adults with TS admission in 14 municipalities (2016-2023). Individual-level linkages captured demographics, diagnosis history, healthcare-utilization markers, and characteristics of recent hospitalization episodes. Outcomes were all-cause hospitalization within 30 and 180 days after the index TS admission, with death treated as a competing event. We estimated cumulative incidence using the Aalen-Johansen method and fitted additive competing-risk regression with inverse failure probability weighting, which allows us to directly model absolute risk ratios (ARRs). Discrimination for 30-day risk was assessed with time-dependent c-index and Brier score using 3-fold cross-validation. 
Results Among 11,284 patients (median age 81 years), 26.1% were hospitalized, and 7.6% died within 30 days without prior hospitalization. In adjusted models, male sex (ARR 1.16, 95% CI 1.09-1.24), higher multimorbidity (1-2 vs 0: 1.17, 1.04-1.31; ≥3 vs 0: 1.43, 1.27-1.61), and recent hospitalization (1.24, 1.14-1.34) increased 30-day risk, whereas older age decreased it per 10 years (0.96, 0.93-0.98). Several morbidities were associated with higher 30-day risk (cancer-related morbidities, cirrhosis, chronic kidney disease, chronic heart failure, atrial fibrillation, chronic pulmonary disease, diabetes), while dementia and prior stroke/TIA were associated with lower risk. Healthcare-utilization markers showed dose-response relations (≥4 prior hospitalizations: 1.58; ≥10 medications: 1.28; ≥3 procedures: 1.34). In the recently hospitalized subgroup, a fall-injury primary diagnosis reduced 30-day risk (0.88), recent surgery increased it (1.09), and hospital stays > 14 days conferred higher risk (1.31). The best 30-day model yielded a c-index of 0.623 and Brier score of 0.186. 
Conclusions Early (re)hospitalization after TS admission is common and patterned by sex, multimorbidity, intensive prior healthcare use, and selected morbidities. Although model discrimination was modest, the identified risk factors can inform targeted interventions in transitional care delivered at TS settings. 
Clinical trial number Not applicable.