Validation of an index to estimate the prevalence of frailty among community-dwelling seniors

Abstract

Background: This study validates cut-points for a frailty index (FI) to identify seniors at risk of a hospital-related event and estimates the number of frail seniors living in the community. The FI developed by Rockwood and Mitnitski defines levels of frailty based on scores of 0 to 1.0.

Data and methods: The cut-point validation was conducted using Stratum-Specific Likelihood Ratios applied to combined 2003 and 2005 Canadian Community Health Survey (CCHS) data, linked to hospital records from the Discharge Abstract Database (2002 to 2007). Based on the validated cut-points, frailty prevalence was estimated using 2009/2010 CCHS data.

Results: Seniors scoring more than 0.21 on the FI were considered to be at elevated risk of hospital-related events. Four additional frailty levels were identified: non-frail (0 to ≤0.1), pre-frail (>0.1 to ≤0.21), more frail (>0.30 to ≤0.35) (women only), and most frail (frail-group subset) (0.45 or more). The number of community-dwelling seniors considered to be frail was estimated at about 1 million (24%) in 2009/2010; another 1.4 million (32%) could be considered pre-frail. Frailty prevalence rose with age; was higher among women than among men; and varied by geographic location.

Interpretation: A cut-point of more than 0.21 can be used to identify frail seniors living in the community.

Keywords: activities of daily living; health status indicators; hospital records; probabilistic linkage; validation studies.