# Adaptation of the comprehensive rheumatologic assessment of frailty (CRAF) as a multidimensional frailty screening tool in patients with rheumatoid arthritis in Vietnam

## TL;DR

This study adapts and validates the CRAF tool for assessing frailty in rheumatoid arthritis patients in Vietnam, showing it is reliable and effective.

## Contribution

The study provides the first validated Vietnamese version of the CRAF tool for frailty assessment in rheumatoid arthritis patients.

## Key findings

- CRAF showed strong convergent validity with the Fried phenotype (p < 0.001).
- CRAF had higher discriminatory power than the Fried phenotype (AUC 0.947).
- Comorbidity, medication, BMI, DAS28-CRP, and age were significant predictors of frailty.

## Abstract

In recent times, there has been acknowledgment of the prevalence of frailty and pre-frailty among individuals with rheumatoid arthritis (RA). Comprehensive Rheumatologic Assessment of Frailty (CRAF) stands out as a dependable tool grounded in synthesis and clinical judgment. Despite this, a validated Vietnamese rendition of the CRAF is currently unavailable. This study seeks to assess the reliability and validity of the CRAF in a patient with RA in Vietnam.

A cross-sectional investigation was carried out with 402 patients diagnosed with rheumatoid arthritis, encompassing both inpatients and outpatients at the Centre for Rheumatology at Bach Mai Hospital in Hanoi, Vietnam. CRAF was employed to gauge the extent of frailty. To establish convergent validity, the scores from the CRAF were correlated with those from the Fried phenotype. Discriminant validity was ascertained through the utilization of receiver operating characteristic (ROC) curve analysis. Additionally, a multivariate logistic regression model was applied to evaluate the individual determinants’ relative impact on the CRAF.

In testing for convergent validity, a significant correlation was found between CRAF and Fried phenotype (p < 0.001). The discriminatory power of CRAF was higher than those of the Fried phenotype (difference between areas under the ROC curves = 0.947 (95% CI: 0.927–0.967). Variables associated with frailty at the multivariate analysis were comorbitidy, medication intake, BMI, DAS28-CRP, and age (all at p < 0.01).

CRAF exhibited strong validity and accurate discrimination. Incorporating frailty assessment into regular rheumatological practices could signify a significant advancement in the care of rheumatoid arthritis.