Measuring Frailty in Old Age: Rasch Analysis of the FRAIL Scale

Objective. To evaluate by Rasch analysis the validity of the FRAIL scale for measuring frailty in older people in Colombia. Methods. Cross-sectional study including 2506 people aged ≥60 years living in Bucaramanga, Medellín, Pereira, Popayán, and Santa Marta in 2021. Guidelines for analysis of the FRAIL scale were followed and the Raschmodel was used with adjustment of response categories, items and people, differential functioning of the items, dimensionality, local independence of the items, Wright reliability, and adjustments of the infit and outfit mean squares. Results. Overfitting of the weight loss item was identified although it did not compromise the unidimensionality or the total score. Wright reliability was 0.80; the measure explained 45.2% of the variance in raw scores. Conclusions. TheFRAIL scale is a valid tool for assessing frailty in elderly people. It is unidimensional, reliable and unbiased by age forthe frail state but not for the prefrail condition. Inclusion of the gender variable and categorization of the age variable with 70 years as cut-off point are suggested.