The Journal of Nutrition, Health and Aging (Apr 2025)

Comparative analysis of frailty identification tools in community services across the Asia-Pacific: A systematic review and meta-analysis

  • Yi-Chen Wu,
  • Chia-Te Chen,
  • Shu-Fen Shen,
  • Liang-Kung Chen,
  • Li-Ning Peng,
  • Heng-Hsin Tung

Journal volume & issue
Vol. 29, no. 4
p. 100496

Abstract

Read online

Objectives: This study aimed to compare various frailty screening and assessment tools with the Frailty Phenotype (FP), Frailty Index (FI), and Comprehensive Geriatric Assessment (CGA), which are considered the current gold standards, among the Asia-Pacific population in community settings. Design: Systematic review and meta-analysis. Setting and participants: This review included studies evaluating frailty identification tools based on the criteria of population, index and reference tests, and diagnosis of interest. Methods: A diagnostic test accuracy review was conducted to assess frailty instruments recommended by the Asia-Pacific Clinical Practice Guidelines. Comprehensive electronic database searches and manual searches were conducted up to August 20, 2024. Study quality, including risks of bias and applicability, was assessed using the QUADAS-2 tool. Hierarchical analysis and Youden’s index were employed to identify the optimal tool and cutoff points, and pooled frailty prevalence was calculated. Results: Fourteen studies were included: 10 for the FRAIL scale, 3 for TUG, and 2 for the SOF index (screening tools), and 2 each for the CFS and KCL, and 1 for the REFS (assessment tools). All studies demonstrated a high risk of bias. The pooled sensitivity and specificity for screening tools were 0.63 and 0.89, respectively, whereas for assessment tools, they were 0.79 and 0.85. The pooled prevalence of frailty and pre-frailty was 19.7% and 31.7%, respectively. The pooled diagnostic odds ratios were highest for the FRAIL scale (15.72) and CFS (35.03) among the screening and assessment tools. The subgroup analysis revealed that the setting had no significant impact on screening tool performance (p = 0.58), but a borderline significant effect was observed for assessment tools (p = 0.06), although this result is limited by the small number of studies, with only one conducted in a community setting. The FRAIL scale, with a cutoff of 2, had a Youden’s index of 0.60, signifying optimal screening performance. Conclusion: Among the frailty instruments recommended by the Asia-Pacific Clinical Practice Guidelines, this meta-analysis identifies the FRAIL scale as the most robust tool for distinguishing frailty, with a cutoff of 2 significantly enhancing diagnostic accuracy. Furthermore, the estimated prevalence of frailty in the Asia-Pacific region is 19.7% across various community settings, underscoring the need for further research and the development of validated assessment tools tailored to this population.

Keywords