Abstract
OBJECTIVE: To identify predictors of functional improvement and home discharge across a wide age spectrum in a real-world inpatient neurorehabilitation cohort, and to assess whether age independently influences rehabilitation outcomes. DESIGN: Retrospective observational cohort study. SETTING: Two inpatient neurorehabilitation units within a university hospital: a high-intensity program (Neurorehabilitation Unit A) and a geriatric-adapted, less intensive program (Neurorehabilitation Unit B). PARTICIPANTS: A total of 694 patients (N=694) admitted for neurorehabilitation between January 2018 and April 2020. Mean age was 66.6±17.5 years; 47.1% were women, and 60.1% were admitted poststroke. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Home discharge; functional improvement defined as ΔFIM≥10 and Montebello Rehabilitation Factor Score (MRFS)≥0.5. RESULTS: At discharge, 74.9% of patients returned home. Functional improvement was achieved in 32.5% (ΔFIM≥10) and 18.3% (MRFS≥0.5). In multivariable models, age was not independently associated with any outcome. Positive predictors of home discharge included higher FIM score at discharge (odds ratio, 1.05; 95% CI, 1.03-1.07), lower FIM score at admission, and greater therapy intensity. Functional improvement was associated with longer length of stay and lower comorbidity burden. Use of antipsychotics (home discharge) and antidepressants (MRFS) were negatively associated with outcomes. Hospitalization in the geriatric unit (Neurorehabilitation Unit B) was associated with lower odds of recovery and discharge home, likely reflecting increased frailty and complexity. CONCLUSIONS: Chronological age was not an independent predictor of home discharge or functional improvement in this mixed-age cohort; generalization to the oldest-old and markedly frail populations should be cautious. Functional status, comorbidities, and therapy factors were more relevant for prognosis. These findings support individualized, age-inclusive rehabilitation strategies that focus on clinical complexity rather than age alone.