Resident- and Facility-Level Predictors of Quality of Life in Long-Term Care

Abstract

Although there is substantial research on quality of care in nursing homes (NH), less is known about what contributes to quality of life (QOL) for NH residents. This study assesses multiple domains of QOL and examines facility- and resident-level correlates for different domains. Data come from (a) self-reported resident interviews using a multidimensional measure of QOL; (b) resident clinical data from the Minimum Data Set; and (c) facility-level characteristics from Minnesota Department of Human Services. We used factor analysis to confirm domains of QOL, and then employed cross-sectional hierarchical linear modeling to identify significant resident- and facility-level predictors of each domain. We examined six unique domains of QOL: environment, personal attention, food, engagement, negative mood, and positive mood. In multilevel models, resident-level characteristics were more reliable correlates of QOL than facility characteristics. Among resident characteristics, gender, age, marital status, activities of daily living, mood disorders, cognitive limitations, and length of stay consistently predicted QOL domains. Among facility characteristics, size, staff hours, quality of care, and percent of residents on Medicaid predicted multiple QOL domains. Examining separate domains rather than a single summary score makes associations with predictors more accurate. Resident characteristics account for the majority of variability in resident QOL. Helping residents maintain functional abilities, and providing an engaging social environment may be particularly important in improving QOL.