Hospital-at-home versus hospital admission for acute care in Singapore: a prospective quasi-experimental study on cost, utilisation and clinical outcomes

BACKGROUND: Hospital-at-home (HaH) is increasingly adopted to expand hospital capacity, however, data from Asian healthcare remain limited. The study compares cost, utilisation, and clinical outcomes of HaH program compared to traditional hospital care in Singapore. METHODS: This prospective quasi-experimental study compared patients enrolled in the HaH programme with a comparator group of clinically eligible patients who received ward-based care because they lived outside the programme's catchment area or when HaH capacity was full. Adults admitted to two hospitals within an academic health system between January 2021 and May 2023 screened for eligibility based on clinical, social and functional criteria. The primary outcome was cost per bed day, with secondary outcomes including total episode cost, clinical outcomes, and patient-reported measures. Inverse probability treatment weighting was applied to balance baseline confounders, and regression analyses were used to evaluate outcomes. RESULTS: Among 302 patients (151 HaH, 151 ward), HaH patients were older (mean [SD] age, 58 [21] vs 51 [21] years) with higher Charlson Comorbidity Index (3 [3] vs 2 [3]). The top three diagnoses were cellulitis (43.0% HaH, 36.4% ward), urinary tract infection (17.9% HaH, 15.9% ward) and rhabdomyolysis (15.9% HaH, 23.8% ward). HaH had lower median cost per day S$676 (95% CI S$610 to S$742) and median cost per episode S$1,665 (95% CI S$765 to S$2,566) than ward group. No significant difference in weighted 30-day readmission risk (RR 1.2 [95% CI 0.56 to 2.58]; p-value = 0.65), 2 weeks post-discharge mean improvement in EQ-5D utility value (0.03 [95% CI -0.06 to 0.11; p-value = 0.54]) and EQ-VAS (1.0 [95% CI -3.7 to 5.6]; p-value = 0.69). There were two expected deaths in HaH and none in ward group. Weighted median lengths of stay in HaH and ward group were 5 and 4 days respectively (difference 1.0 [95% CI 0.2 to 1.8]; p-value = 0.02); HaH overall saved 682 hospital bed-days. CONCLUSION: For suitable patients, HaH care costs less than ward-based care, with no significant difference in clinical outcomes. TRIAL REGISTRATION: This study is registered at ClincialTrials.gov (NCT04330378, registration date 31 May 2020). It is funded by the National Medical Research Council (NMRC) Health Services Research Grant (HSRG) (MOH-000738).