Back to Signals

Signal · Frailty

Newer composite frailty tools outperform guideline-endorsed scales, study finds

A large Swedish cohort study finds newer composite frailty indices predict institutionalisation, mortality and hospitalisation more accurately than commonly used guideline tools.

1 min read Primary Study NEW

A prospective Swedish cohort study (SNAC-K, n=3,108, followed for up to six years) compared seven geriatric assessment tools head-to-head for their ability to predict institutionalisation, mortality, dementia, falls and hospitalisation [Primary Study]. The Health Assessment Tool, the Intrinsic Capacity index and the SNAC-K Frailty Index all outperformed guideline-endorsed tools, including comorbidity- and function-based scales such as the G8, the Charlson Comorbidity Index and the Cumulative Illness Rating Scale. Tools incorporating gait speed and physical function performed best across outcomes.

NHS acute frailty pathways commonly use comorbidity- or function-based risk-stratification tools to identify patients for same-day emergency care or specialist geriatric assessment. This study, drawn from a large, long-followed community cohort, is among the more robust head-to-head comparisons yet published of such instruments against newer composite alternatives that incorporate physical performance measures directly. The source material contains no international comparator beyond the Swedish cohort itself, and none is asserted here.

For services designing or reviewing frailty case-finding and risk-stratification pathways, this evidence suggests that reliance on comorbidity-only tools may underperform relative to composite instruments incorporating physical function measures such as gait speed — a consideration for pathway design, subject to further validation before wholesale adoption.

Sources

  1. Abbadi, et al. Comparison of seven geriatric assessment tools for predicting institutionalisation, mortality, dementia, falls and hospitalisation. BMC Medicine. 2026. doi:10.1186/s12916-026-05008-2 PMID 42421064