Facility-Level Variation of Resident Loneliness in Assisted Living and Associated Organizational Context Factors: A Repeated Cross-Sectional Study

Background and Objectives

Loneliness is common among nursing home residents, and it is also thought to be a problem in assisted living (AL). However, we lack research on loneliness in AL. Our objectives were to assess changes in risk-adjusted prevalence of loneliness in AL, and facility-level variations in loneliness before and during the COVID-19 pandemic, and facility-level factors associated with AL resident loneliness during the pandemic.

Research Design and Methods

This population-based, repeated cross-sectional study used Resident Assessment Instrument—Home Care (RAI-HC) data (01/2017–12/2021) from Alberta, Canada. On a system-level, we estimated quarterly, risk-adjusted loneliness prevalence, and used segmented regressions to assess whether loneliness changed after the start of the pandemic. For risk adjustment, we used resident-covariates known to be associated with loneliness, but out the health system's or AL home's control (e.g., age or cognitive impairment) to enable fair comparisons over time. Linking AL home surveys, collected in COVID-19 waves 1 (March–June 2020) and 2 (October 2020–February 2021) to RAI-HC records, we used covariate-adjusted general estimating equations (GEE) to assess AL home factors (e.g., staffing shortages, social distancing measures) associated with resident-level loneliness during the pandemic.

Results

Quarterly samples included 2026–2721 residents. Loneliness [95% confidence interval] fluctuated between 13.6% [11.5%–15.7%], and 16.8% [14.4%–19.2%], with no statistically significant increase during the pandemic. Facility-level median [inter-quartile range] loneliness prevalence varied considerably before (14.9% [8.3%–21.1%) and during the pandemic (13.5% [6.9%–21.3%]). GEEs included 985 residents in 41 facilities (wave 1), and 1134 residents in 42 facilities (wave 2). Facility-factors associated with decreased odds of loneliness included: facilitating caregiver involvement (odds ratio = 0.531 [95% confidence interval: 0.286–0.986]), essential visitor policies (0.672 [0.454–0.994]), and video calls with volunteers or religious/spiritual leaders (0.603 [0.435–0.836]). Facilitating outdoor activities/visits (2.486 [1.561–3.961], and providing hallway-based activities (1.645 [1.183–2.288]) were associated with increased odds of loneliness.

Discussion and Implications

Loneliness did not change during COVID-19 in AL on a health system level, but varied considerably between facilities before and during the pandemic. Modifiable facility-level factors explained variations in loneliness within facilities, suggesting important targets for policies and improvement interventions.