Quality and safety in long-term care in Ontario: the impact of language discordance

Objectives — This study compared quality indicators across linguistic groups and sought to determine whether disparities are influenced by resident-facility language discordance in long-term care.

Design — Population-based retrospective cohort study using linked databases.

Setting and Participants — Retrospective cohort of newly admitted residents of long-term care facilities in Ontario, Canada, between 2010 and 2016 (N=47,727). Individual residents' information was obtained from the Resident Assessment Instrument Minimum Data Set (RAI-MDS) to determine resident's primary language, clinical characteristics, and healthcare indicators.

Measures — Main covariates of interest were primary language of the resident and predominant language of the long-term care facility, which was determined using the French designation status as defined in the French Language Services Act. Primary outcomes were a set of quality and safety indicators related to long-term care: worsening of depression, falls, moderate-severe pain, use of antipsychotic medication, and physical restraints. Multivariable logistic regression models were used to assess the impact of resident's primary language, facility language, and resident-facility language discordance on each quality indicator.

Results — Overall, there were few differences between francophones and anglophones for quality and safety indicators. Francophones were more likely to report pain (10.9% vs 9.9%; P = .001) and be physically restrained (7.3% vs 5.2%; P < .001), whereas a greater proportion of anglophones experienced worsening of depressive symptoms (24.0% vs 22.9%; P = .001). However, quality indicators were generally worse for francophones in Non-Designated facilities, except for pain, which was more commonly reported by francophones in French-Designated facilities. Anglophones were more likely to be physically restrained in French-Designated facilities (6.7% vs 5.1%; P < .001).

Conclusions and Implications — For francophones, quality indicators tended to be worse in the presence of resident-facility language discordance. However, these findings did not persist after adjusting for individual- and facility-level characteristics, suggesting that the disparities observed at the population level cannot be attributed to linguistic factors alone.