JMIR Publications

ABSTRACT

Background:

Potentially inappropriate antipsychotic use (APU), physical restraint use (PRU), and polypharmacy are common and potentially harmful care practices in long-term care (LTC) homes. Evidence on the social variables associated with these practices is predominantly US-based or pre-COVID-19, with limited contemporary Canadian research. We examined whether resident- and home-level social characteristics, including area-based deprivation, were associated with these practices across four Canadian provinces.

Objective:

To identify resident social characteristics and LTC home contextual factors associated with potentially inappropriate APU, PRU, and polypharmacy among LTC home residents in the Canadian provinces of Alberta, British Columbia, Manitoba, and Ontario.

Methods:

We used cross-sectional data from the Counting What Counts study, including 2,779 residents with cognitive impairment across 62 LTC homes in the four provinces. Homes were stratified by health region, bed size, and ownership model, with a random 10% drawn per stratum. Outcomes were derived from Resident Assessment Instrument – Minimum Data set 2.0 (RAI-MDS 2.0) records. Resident social characteristics were collected via staff proxy surveys. Area deprivation was measured using the Canadian Index of Multiple Deprivation linked by dissemination area. We estimated multivariable multilevel logistic regression models with random home-level intercepts for each outcome. Group-mean centring separated within-home from between-home effects of resident social characteristics. Sensitivity analyses tested alternative outcome definitions.

Results:

Potentially inappropriate APU prevalence was 18.9% (526/2,779), daily PRU was 3.5% (98/2,779), polypharmacy (≥9 medications) was 54.2% (1,507/2,779). In adjusted models, older age was associated with lower odds of APU (odd ratio [OR] 0.81, 95% confidence interval [CI] 0.73–0.89) and polypharmacy (OR 0.91, 95% CI 0.82–1.01). Residents without family or friend visitors had lower odds of polypharmacy (OR 0.67, 95% CI 0.49–0.93). In group-mean centred models, a higher proportion of residents who were not White in a home was associated with higher odds of PRU (OR 124.9, 95% CI 1.52–10,245.29), and a higher proportion speaking English as an additional language was associated with higher odds of APU (OR 16.2, 95% CI 3.3–79.1). Area-based economic dependency was associated with higher odds of APU (OR 1.19, 95% CI 1.02–1.38), and situational vulnerability with higher odds of PRU (OR 1.79, 95% CI 1.19–2.70).

Conclusions:

Beyond clinical characteristics, social and structural factors, including area-level deprivation and language composition, were associated with potentially inappropriate care practices. These findings support structural interventions addressing staffing and resource allocation, alongside culturally safe, language-concordant care and caregiver support.