Hospital-at-home with point-of-care ultrasound for infections in nursing home residents: feasibility, safety, and care-setting shift in a Taiwanese service evaluation

Background Nursing-home residents with acute infection are commonly transferred to hospital, where they are exposed to delirium, functional decline and iatrogenic harm. In October 2024 our nursing home implemented Hospital-at-Home (HaH) under Taiwan’s Acute Care At Home program, using point-of-care ultrasound (POCUS) to treat pneumonia, urinary tract infection (UTI) and soft-tissue infection (SSTI) on site. We asked whether hospital-level infection care could be delivered safely within the nursing home. Methods Retrospective service evaluation at one central-Taiwan nursing home (HaH period October 2024–December 2025), with a before-and-after description of infection-related hospital use (pre-implementation January–September 2024). Residents entered HaH directly from the nursing home (Model B) or after emergency-department assessment (Model C). Outcomes were completion without escalation to inpatient care, rehospitalization and mortality; the route comparison is descriptive. Results There were 130 HaH episodes (81 UTI, 42 pneumonia, 7 SSTI) in 87 residents (episode-level median age 79.6 years; Barthel Index < 20 in 100 episodes, 76.9%). HaH was completed without escalation in 116 of 130 episodes (89.2%, 95% CI 82.6–94.0). No death occurred during a HaH episode; 30-day mortality was 5 of 130 (3.8%, 1.3–8.7), a lower bound because ascertainment used study-hospital records only. Among the 116 completers, 6 (5.2%, 1.9–10.9) were rehospitalized within 7 days. Within 30 days of treatment ending, 18 of 130 (13.8%) were admitted for any cause and 10 (7.7%) for the same infection. Post-/pre- inpatient rate ratios per 1,000 resident-days were 0.68 (0.41–1.11) for UTI, 0.33 (0.10–1.01) for SSTI and 1.06 (0.73–1.55) for pneumonia; total hospital-level infection care (inpatient plus HaH) rose from 3.15 to 5.22 care contacts per 1,000 resident-days (5.02, ≈+59%, after de-duplication). Completion was 95.9% in Model B and 80.7% in Model C, a difference largely confined to pneumonia, where the routes’ diagnostic standards differed. Conclusions In a frail, highly dependent nursing-home population, POCUS-supported HaH enabled most residents with pneumonia, UTI or SSTI to complete hospital-level treatment without inpatient admission, with no deaths during treatment. The hospitalization data show a shift in where infection care was delivered, not a reduction in infection burden; whether HaH reduces admissions, cost or harm requires controlled prospective study.