Understanding Differences in Resident Loneliness/Isolation Across US Nursing Homes: A Multilevel Analysis of MDS 3.0 Data.

OBJECTIVES: Loneliness/isolation is known to result in negative health outcomes among older adults in US nursing homes. Understanding the variation in loneliness/isolation among long-stay nursing residents and the individual and facility-level characteristics associated with loneliness/isolation can help guide care planning and ultimately lead to the development of effective interventions. DESIGN: Cross-sectional. SETTING AND PARTICIPANTS: Long-stay residents aged ≥50 years (n = 732,808) living in US nursing homes (n = 13,491) between October 1, 2023 and December 31, 2023. METHODS: Data from the Minimum Data Set 3.0 provided resident-level characteristics. Residents able to self-report were asked: "How often do you feel lonely or isolated from those around you?" Responses were dichotomized as sometimes, often, always vs never or rarely. Facility-level variables came from the long-term care focus and Nursing Home Compare datasets. Multilevel logistic regression estimated proportional change in cluster variance and variance partitioning coefficients (VPCs) to assess contributions of resident-, facility-, and county-level factors, and the fully adjusted model identified characteristics associated with loneliness/isolation. RESULTS: The overall prevalence of loneliness/isolation was 20.7%, with most variation driven by differences between facilities rather than counties (VPCcounty = ∼2%; VPCfacility = ∼35%). Resident- and facility-level characteristics explained only a small portion, and county-level factors contributed minimally (proportional change in cluster variance = 1%-2%). Higher prevalence of loneliness/isolation was observed among women, those with mental health diagnoses or sensory impairments, and a lower prevalence was observed in older, Non-Hispanic Black residents; facility factors like for-profit status, rural location, racial/ethnic diversity, and more certified nursing assistant care were associated with reduced loneliness/isolation. CONCLUSIONS AND IMPLICATIONS: Interventions to reduce loneliness/isolation should focus on helping facilities address contextual factors. Staffing levels may be associated with loneliness/isolation and warrant further investigation. Future research should focus on the development, implementation, and evaluation of evidence-based practices to reduce resident loneliness/isolation.