Association of Palliative Care Initiation With Acute Healthcare Utilization Among Community‐Dwelling Adults Living With Dementia

OBJECTIVES: To measure the association between initiating palliative care using the combined criteria of short survival expectations and increased care needs with acute healthcare utilization among community-dwelling people living with dementia. DESIGN: A population-level propensity-based overlap-weighted cohort study. SETTING: Ontario, Canada. PARTICIPANTS: 50,961 community-dwelling people living with dementia between 2010 and 2023, with 1-year follow-up until 2024. We used moderate to severe dementia and the initiation of homecare services as proxies for short survival expectations and increased care needs, respectively. We used linked health administrative data with propensity-based overlap weighting on sociodemographic and clinical factors to address confounding by the indication to receive palliative care. EXPOSURES: Receipt of home-based palliative care within 30 days of homecare initiation. MAIN OUTCOME MEASURES: Individual 1-year cause-specific hazards of emergency department (ED) use, hospitalization, and intensive care unit (ICU) admission. RESULTS: Palliative care was associated with a higher risk of hospitalization in the first 90 days (e.g., hazard ratio (HR) 1.43 at 30 days; 95% confidence interval (CI) 1.25-1.64), which was driven predominantly by hospitalization with palliative intent, but was no longer significantly different beyond 180 days, compared to not receiving palliative care. [Correction added on 31 January 2026, after first online publication: The preceding sentence has been revised in this version.] At 90 days, the cumulative incidence of hospitalization, death, or admission to a nursing home was approximately 20% among those receiving palliative care, compared to 15% in those not receiving palliative care. Although the overall absolute rate of ICU admission was low (0.01 ± 0.1 admissions per month), palliative care initiation was associated with a 186% higher risk in the first 30 days (HR 2.86; 95% CI 1.75-4.69) but not beyond 90 days. Palliative care was not associated with ED use over the following year (HR 1.02; 95% CI 0.72-1.45). CONCLUSIONS: The current approach to delivering palliative care services for community-dwelling people living with dementia in Canada may paradoxically increase early acute care use, likely reflecting system gaps and unmet needs. Given the high prevalence of dementia, there is an urgent need to re-examine homecare delivery models.