# Differences in Item Discrimination of the 25‐Question Geriatric Locomotive Function Scale Between Younger and Middle‐Aged Adults and Older Adults: An Analysis Using the Item Response Theory

## TL;DR

This study shows how a health assessment tool works differently in younger/middle-aged and older adults, suggesting ways to improve age-specific evaluations.

## Contribution

The paper identifies age-specific differences in item discrimination of a locomotive function scale using item response theory.

## Key findings

- Pain- and social participation-related items were more discriminative in younger/middle-aged adults.
- Mobility- and anxiety-related items were more discriminative in older adults.
- The 65-year age cutoff showed robust results in item discrimination differences.

## Abstract

The 25‐item Geriatric Locomotive Function Scale is widely used to identify locomotive syndrome, a condition characterized by progressive decline in motor function. Although its reliability in older adults has been established, item‐level properties across age groups remain unclear. This study examined item discrimination and difficulty using an item response theory approach and compared item characteristics between younger and middle‐aged adults and older adults.

Community‐dwelling adults aged 18 years or older were classified as younger and middle‐aged (< 65 years) or older (≥ 65 years). The graded response model estimated item discrimination and difficulty. Discrimination was interpreted using Baker's classification, while values between 0.5 and 2.5 were considered empirically acceptable. Difficulty parameters ranged approximately from −3 to +3. Model fit was assessed using the root mean square error of approximation. Wald tests compared item parameters between age groups, and sensitivity analyses used age thresholds of 60, 70, and 75 years.

Among 866 participants, most items showed acceptable discrimination and difficulty. Pain‐ and social participation‐related items were more discriminative in younger and middle‐aged adults, whereas mobility‐ and anxiety‐related items were more discriminative in older adults. Model fit was good (root mean square error of approximation = 0.04–0.05), and Wald tests confirmed significant item‐level differences between groups. Sensitivity analyses supported the robustness of the 65‐year cutoff.

The Geriatric Locomotive Function Scale demonstrated statistically validated age‐related item patterns. These findings support developing age‐tailored short forms and adaptive screening tools for locomotive syndrome.