# Estimating the Number of Latent Ranks of the Fugl-Meyer Assessment Score for the Affected Upper Extremity After Stroke

## TL;DR

This study evaluates the Fugl-Meyer Assessment score for stroke patients to determine how many underlying levels of motor recovery it can detect.

## Contribution

The study introduces a new method using latent rank theory to determine the optimal number of latent ranks in FMA-UE scores.

## Key findings

- Seven latent ranks were found to be most appropriate for both proximal and distal FMA-UE items.
- The FMA-UE was shown to have high construct validity based on latent rank analysis.
- The study proposes a novel interpretation of FMA-UE scores that has not been previously identified.

## Abstract

Many clinical stroke rehabilitation studies have adopted the upper extremity motor section of the Fugl-Meyer Assessment (FMA-UE). In addition, some clinical studies use specific FMA-UE scores as inclusion criteria. However, it remains unclear whether it is appropriate to determine the criterion based on the total score of FMA-UE. This study aimed to determine a highly valid criterion using the latent rank theory (LRT) that can estimate the number of latent ranks of FMA-UE. This was a multicenter cross-sectional study; patients with stroke were recruited from 25 hospitals between March 2018 and April 2022. For all patients, FMA-UE results and participant information were collected. The collected FMA-UE data were divided into proximal and distal items and verified the dimensionality of the data. After that, the LRT was used to determine the latent ranks. Seven ranks were considered the most appropriate for proximal and distal items when estimating the number of latent ranks. These results suggest that FMA-UE has high construct validity. Furthermore, we recommend the novel interpretability of FMA-UE, which previous studies have yet to find. Although this cross-sectional study cannot directly guide stroke patients' recovery processes, it may be practical for optimizing the difficulty of stroke rehabilitation.