Abstract
BACKGROUND: The identification of contextual factors that modify associations between client frailty and their health and service use outcomes is essential for informed home health care and policy planning. Our objective was to examine variation in the associations between frailty and select 1-year health outcomes by caregiver distress and client sex among community-residing older care recipients. METHODS: We conducted a retrospective cohort study using linked population-based clinical and health administrative databases for all long-stay home care clients (n = 234,552) aged 66+ years assessed during April 2010-2013 in Ontario, Canada. Frailty was assessed using a previously validated 72-item frailty index (FI). Presence of caregiver distress was derived from clinical assessment items administered by trained home care assessors. Multivariable log-binomial regression models were used to examine variations in the associations between frailty and outcomes of interest (mortality, nursing home [NH] placement, all-cause and prolonged hospitalization) by caregiver distress, with further model stratification by client sex. RESULTS: Frailty prevalence varied little by sex (19.3% women, 19.9% men) despite significant sex-differences in clients' sociodemographic and health characteristics. In both sexes, frailty was significantly associated with all outcomes, particularly NH placement (RR = 3.84, 95%CI 3.75-3.93) and death (RR = 2.32, 95%CI 2.27-2.37), though risk ratios were greater for women. Caregiver distress was more common with increasing frailty and for male clients, and a significant independent predictor of NH placement and prolonged hospitalization in both sexes. The association between frailty and NH placement (but not other outcomes) varied by caregiver distress for both men and women (p < 0.001 interaction terms), showing a greater magnitude of association among clients without (vs. with) a distressed caregiver. CONCLUSIONS: As caregiver distress varies by client sex, represents a key driver of NH placement (even among relatively robust clients), and modifies the impact of other risk factors such as frailty, it should be routinely assessed. Further, sex-differences should be considered when developing and evaluating community-based services for older adults and their caregivers.