Background: To identify personal and clinical variables that predict post-acute care (PAC) versus home discharge among hospitalized adults. A secondary objective examined differences based on patients’ cognitive status.
Methods: This cross-sectional, prospective, observational study collected electronic health record (EHR) and REDCap survey data to examine variables that predicted discharge to PAC among hospitalized patients. The primary outcome was the discharge disposition (Institutional PAC vs. home) confirmed via EHR documentation after discharge. Multivariate logistic regression analyses identified independent predictors. To assess for potential effect modification by cognitive status, patients were stratified into those with and without cognitive impairments and separate adjusted models were performed for each subgroup. Setting: Reported findings from an urban safety-net hospital located in Boston, MA. Participants: A total of 629 participants admitted between February 2024 and January 2025 aged 18 and older were recruited for the study. Demographic characteristics: age (mean = 72.1 years, SD = ±10.2, range = 26-92); biological sex = 50.9% Female; race = 51.5% Black or African American, 37.8% White, 0.5% Asian, 0.2% Indian American or Alaskan Native, 7.3% other, and 2.7% unspecified. Recruitment aimed to ensure representation of patients with cognitive impairments (determined through ICD-10 codes), and the final sample included 40.9% with and 59.1% without cognitive impairment.
Results: Results: PAC discharge was significantly associated with need for assistance (independent to total), length of stay (LOS), mobility impairment (AM-PAC 6 Clicks), and depression (PHQ2 scores) Compared with patients identified as independent, those requiring assistance had progressively greater odds of discharge to institutional PAC: minimal assistance (OR, 4.05; 95% CI, 2.35–7.01) to total assistance (OR, 13.9; 95% CI, 4.50–42.9). Longer LOS was also associated with increased odds of institutional PAC discharge (OR, 1.13; 95% CI, 1.08 – 1.18). Patients with a mobility impairment ≥50% (AM-PAC 12-17) had higher odds (OR, 3.82; 95% CI, 1.67–8.74) of being discharged to institutional PAC. A positive depression screen (PHQ-2 ≥3) increased the PAC discharge odds (OR, 1.72; 95% CI, 1.10-2.70).Stratified analyses showed variable associations for PAC discharge among patients with and without cognitive impairments. For example, patients without cognitive impairments who required moderate and maximum assistance had increased odds for PAC discharge, compared to those with cognitive impairments.
Conclusions: Assistance needs, mobility impairment, and depression predicted PAC discharge, with stronger associations among patients without cognitive impairment. Future work will examine findings comparing discharge to different institutional PAC settings.