Validity and reliability of the Thai version of the simple frailty questionnaire (T-FRAIL) with modifications to improve its diagnostic properties in the preoperative setting

BACKGROUND: Several methods are available for identifying frailty, but limited tools have been validated in Thai context. Our objective was to evaluate the validity and reliability of the Thai version of the Simple Frailty Questionnaire (T- FRAIL) compared to the Thai Frailty Index (TFI) and to explore modifications to improve its diagnostic properties. METHODS: The T-FRAIL was translated with permission using a standardized protocol, that included forward and back-translation. Content validity analysis was performed using input from 5 geriatricians. Test-retest reliability, concurrent validity, diagnostic properties, and options to increase the sensitivity of the questionnaire were explored. A cross-sectional study for evaluation validity and reliability was carried out among 3 hundred patients aged 60 or more undergoing elective surgery at a university hospital. RESULTS: The item content validity index (I-CVI) showed 1.0 for each questionnaire item. Test-retest reliability within a 7-day interval was done in 30 patients with a good intraclass correlation coefficient of 0.880. Compared with the TFI, the T-FRAIL yielded an excellent accuracy (area under the curve = 0.882). The identification of frailty using a score of 2 points or more provided the best Youden’s index at 63.1 with a sensitivity of 77.5% (95% CI 69.0–84.6) and a specificity of 85.6% (95% CI 79.6–90.3). A cutoff point of 1 out of 5 items for original T-FRAIL provided a sensitivity of 93.3% and a specificity of 61.1%. The modified T-FRAIL (T-FRAIL_M1), by reducing the “illnesses” criterion to 4 or more diseases, at a cutoff point at 1 had a sensitivity of 94.2% and a specificity of 57.8%. Another modified T-FRAIL (T-FRAIL_M2), by combining three components, at a cutoff point at 1 yielded a sensitivity of 85.8% and a specificity of 80.6%. CONCLUSION: The T-FRAIL and its modification demonstrated satisfactory validity and reliability to identify frailty in elderly patients. The cutoff score of 1 point from 5 items from the original version of T-FRAIL and T-FRAIL_M1 provides a highly sensitive screening tool. T-FRAIL_M1 with a cutoff point of 2 and T-FRAIL_M2 yields reasonable sensitivity and specificity for practical use.