Self-Reported Frailty Screening Tools: Comparing Construct Validity of the Frailty Phenotype Questionnaire and FRAIL

We examined the construct validity of 2 self-reported frailty questionnaires, the Frailty Phenotype Questionnaire (FPQ) and FRAIL, against the Cardiovascular Health Study frailty phenotype (CHS-FP).

Cross-sectional data analysis of longitudinal prospective cohort study.

We included data from 230 older adults (mean age: 67.2 ± 7.4 years) from the "Longitudinal Assessment of Biomarkers for characterization of early Sarcopenia and Osteosarcopenic Obesity in predicting frailty and functional decline in community-dwelling Asian older adults Study" (GeriLABS 2) recruited between December 2017 and March 2019.

We compared area under receiver operating characteristic curves (AUC), agreement, correlation, and predictive validity against outcome measures [Short Physical Performance Battery, 5 times repeat chair stand (RCS-5), Frenchay activities index, International Physical Activity Questionnaire, life-space assessment, Social Functioning Scale 8 (SFS-8), EuroQol-5 dimensions (utility value)] using logistic regression adjusted for age, gender, and vascular risk factors. We examined concurrent validity across robust versus prefrail/frail for inflammatory blood biomarkers [tumor necrosis factor receptor 1 and C-reactive protein (CRP)] and dual-energy x-ray absorptiometry body composition [bone mineral density (BMD); appendicular lean mass index (ALMI), and fat mass index (FMI)].

Prevalence of prefrail/frail was 25.7%, 14.8%, and 48.3% for FPQ, FRAIL, and CHS-FP, respectively. Compared with FRAIL, FPQ had better diagnostic performance (AUC = 0.617 vs 0.531, P = .002; sensitivity = 37.8% vs 18.0%; specificity = 85.6% vs 88.2%) and agreement (AC1-Stat = 0.303 vs 0.197). FPQ showed good predictive validity [RCS-5: odds ratio (OR) 2.38; 95% CI: 1.17-4.86; International Physical Activity Questionnaire: OR 3.62; 95% CI:1.78-7.34; SFS-8: OR 2.11; 95% CI: 1.64-5.89 vs FRAIL: all P > .05]. Only FRAIL showed concurrent validity for CRP, compared with both FPQ and FRAIL for TNF-R1. FRAIL showed better concurrent validity for BMD, FMI, and possibly ALMI, unlike FPQ (all P > .05).

Our results support complementary validity of FPQ and FRAIL in independent community-dwelling older adults. FPQ has increased case detection sensitivity with good predictive validity, whereas FRAIL demonstrates concurrent validity for inflammation and body composition. With better diagnostic performance and validity for blood biomarkers and clinical outcomes, FPQ has utility for early frailty detection in the community setting.