A Comparison of Frailty Assessment Instruments in Different Clinical and Social Care Settings: The Frailtools Project

Abstract

Objective: To determine which of 8 commonly employed frailty assessment tools demonstrate the most appropriate characteristics to be employed in different clinical and social settings.

Design: Cross-sectional multicenter European-based study.

Setting and participants: 1440 patients aged ≥75 years evaluated in geriatric inpatient wards, geriatric outpatient clinics, primary care clinics, and nursing homes.

Methods: The frailty instruments used were Frailty Phenotype, SHARE-FI, 3-item Frailty Trait Scale (FTS-3), 5-item Frailty Trait Scale (FTS-5), FRAIL, 35-item Frailty Index (FI-35), Gérontopôle Frailty Screening Tool (GFST), and Clinical Frailty Scale (CFS). The settings were geriatrics wards, outpatient clinics, primary care, and nursing homes. Suitability was evaluated by considering the feasibility (patients with the test fully completed), administration time (time spent for administering the test), and interscale agreement (Cohen kappa index among instruments to detect frailty).

Results: The prevalence of frailty varied across settings and adopted tests. The scales with the mean highest feasibility were the FRAIL scale (99.4%), SHARE-FI (98.3%), and GFST (95.0%). The mean shortest administration times were obtained with CFS (24 seconds), GFST (72 seconds), and FRAIL scale (90 seconds). The interscale agreement between most of the tests was fair. CFS followed by FTS-5 agreed at least moderately with a greater number of scales overall and in almost all settings.

Conclusions and implications: Based on feasibility, time to undertake the tool, and agreement with other scales, different scales would be recommended according to the setting considered. Our findings suggest that most of the tools evaluated are actually assessing different frailty constructs.

Keywords: Frailty instruments; clinical and social settings; feasibility; intertest agreement; prevalence.