Derivation of a frailty index from the resident assessment instrument – home care adapted for Switzerland: a study based on retrospective data analysis

Catherine Ludwig, Catherine Busnel

The screening of frail individuals at risk for functional health decline and adverse health outcomes lies in the evolving agenda of home care providers. Such a screening can be based on a frailty index (FI) derived from data collected with interRAI instruments used in clinical routines to define care plans. The objective of this study was to assess the feasibility of deriving an FI from the Resident Assessment Instrument - Home Care adapted for Switzerland (Swiss RAI-HC). Data were collected by the Geneva Institution for Homecare and Assistance in clinical routines. The sample consisted of 3714 individuals aged 65 or older (67.7% females) who had each received a Swiss RAI-HC upon admission in the year of 2015. The FI was derived from 52 variables identified and scored according to published guidelines. Adverse health outcomes were either assessed during follow-up assessments (falls, hospitalizations) or documented from administrative records (mortality). The results showed that the FI was distributed normally, with a mean of 0.24 (± 0.13), an interquartile range of 0.16, and values of 0.04 at percentile 1 and 0.63 at percentile 99. The effect of Age was significant (R(2) = 0.011) with a slope of β = 0.002, 95% CI = [0.001-0.002]. Sex as well as the Age × Sex interaction were not significant. The FI predicted deaths (OR = 9.99, 95% CI = [3.20-29.99]), hospitalizations (OR = 3.40, 95% CI = [1.78-6.32]), and falls (OR = 5.00, 95% CI = [2.68-9.38]). The results support the feasibility of an FI derivation from the Swiss RAI-HC, hence replicating previous demonstrations based on interRAI instruments. The results also replicated findings showing that the FI is a good predictor of adverse health outcomes. Yet, the results suggest that home care recipients demonstrate a frailty pattern different from the one reported in community dwellers but comparable to clinical samples. Further work is needed to assess the characteristics of the proposed index in community-dwelling, non-clinical samples for comparability with the existing literature and external validation TRIAL REGISTRATION: ClinicalTrials.gov NCT03139162 . Retrospectively registered May 2, 2017.