Abstract
BACKGROUND: The Veterans Affairs Frailty Index (VA-FI) has been implemented in subgroups of U.S. Veterans. The objectives were to operationalize the VA-FI in long-term care residents and examine its associations with mortality, falls, and fractures.
METHODS: This retrospective cohort study included Veterans ≥65 years who stayed ≥90 days in Community Living Centers (CLCs). We calculated the VA-FI by dividing the count of health deficits by 31. We categorized the residents in 4 groups: non/pre-frail, mildly, moderately and severely frail. Mortality, the first events of falls and fractures were examined with a 1-year follow-up. Cox regression and Fine-Gray competing-risk models were fit to assess associations between frailty and outcomes.
RESULTS: Among 45 183 CLC residents, 12.2% were non/pre-frail, 24.2%, 28.8%, and 34.8% were mildly, moderately and severely frail. With increasing frailty severity, rates of mortality, falls and fractures ranged respectively from 65.3 to 59.0, 13.1 to 18.8, and 1.9 to 2.2 per 100 person-years. Relative to the non/pre-frail, residents with mild, moderate and severe frailty had lower rates of death (HRs [95% CIs]: 0.88 [0.84-0.92], 0.87 [0.83-0.91], 0.90 [0.85-0.94]), higher rates of falls (SHRs [95% CI]: 1.26 [1.10-1.43], 1.32 [1.17-1.50], 1.29 [1.14-1.46]) and fractures (1.29 [0.92-1.81], 1.47 [1.06-2.04], 1.45 [1.04-2.02]).
CONCLUSIONS: Frailty was highly prevalent in Veterans receiving long-term care in CLCs and was associated with higher rates of falls and fractures. Frail residents had a lower rate of mortality compared with their non/pre-frail counterparts. Length of stay, confounded by unobserved factors, and collider bias could potentially serve as explanations for this counterintuitive finding.