Validation of a frailty phenotype screening questionnaire for rural Chinese older adults: a cross-sectional study

Frailty represents a significant public health issue among older adults in rural China. While the Fried Frailty Phenotype (FFP) is one of the most widely used and well-validated tools for assessing physical frailty, its reliance on physical performance measures and specialized devices limits its feasibility for large-scale screening in such settings. Therefore, a simple and feasible screening instrument based on the FFP framework is needed.

To validate the Chinese Frailty Phenotype Questionnaire (CFPQ) for rural Chinese older adults.

This cross-sectional study enrolled 1590 rural-dwelling adults (≥ 65 years) in northern China. The 5-item dichotomous questionnaire operationalized the criteria of the Fried Frailty Phenotype (FFP) endorsed by the Chinese Medical Association (CMA) and included five items: Fatigue, Resistance, Ambulation, Inactivity and Loss of weight. The psychometric evaluation assessed reliability (including internal consistency and test-retest stability), construct validity (including convergent and discriminant validity) and criterion validity. Criterion validity was evaluated through diagnostic accuracy (ROC analysis) and agreement with the criterion standard (kappa statistic).

The CFPQ showed a test-retest intraclass correlation coefficient of 0.89. The internal consistency, measured using Kuder-Richardson formula 20, was 0.49. Construct validity was assessed by calculating correlation coefficients between CFPQ items and cross-sectional outcomes, which ranged from - 0.33 to 0.23. Convergent validity was supported by theoretically consistent correlations: Ambulation was negatively associated with gait speed (r = -0.33, p < 0.001); Resistance with the 30-second chair stand test (r = -0.23, p < 0.001); Fatigue with the EQ-5D Visual Analogue Scale (r = -0.20, p < 0.001), and Loss of weight with body mass index (r = -0.11, p < 0.001). Discriminant validity was evidenced by statistically distinct patterns of correlations across domains. Against the Fried Frailty Phenotype (FFP-CMA) criterion, the CFPQ achieved an area under the curve of 0.93 (95% CI: 0.92-0.95). A score of ≥ 3 was identified as the optimal cut-off, with a specificity of 97.3%, a negative predictive value of 98.9%, and a positive predictive value of 34.4%. The kappa statistic indicating agreement between the CFPQ and FFP-CMA was 0.41.

The CFPQ demonstrates acceptable validity and reliability as a practical tool for frailty phenotype screening among rural-dwelling older adults in China.

Not applicable.