[Reliability and factorial structure of a rating scale for amyotrophic lateral sclerosis]

Abstract

The Modified Norris Scale is a rating scale for amyotrophic lateral sclerosis (ALS), which consists of two parts, the Limb Norris Scale and the Norris Bulbar Scale. The Limb Scale has 21 items to evaluate extremity function and the Bulbar Scale has 13 items to evaluate bulbar function. Each item is rated in 4 ordinal categories. Considering the habitual difference, we translated the English scale into Japanese one with minor modification, and added more detailed explanations for all categories of each item. Then we examined reliability and factorial structure of the translated scale. The subjects were 23 patients with motor disturbance and each subject was rated twice by 2-4 neurologists. As a measure of reliability, the Kappa coefficient proposed by Cohen (1960) and Kraemer (1980) was calculated for each item and the intraclass correlation coefficient (ICC) was evaluated for total scores of each of two scales. To analyze the factorial structure, the factor analysis was carried out. The minimum and the maximum Kappa values were .70 and .97 for intra-rater reliability of the Limb Scale's items, .60 and .83 for inter-rater reliability of the Limb Scale's items, .41 and 1.00 for intra-rater reliability of the Bulbar Scale's items and .26 and .81 for inter-rater reliability of the Bulbar Scale's items, respectively. Concerning the factorial structure, the contribution of the first factor was 83.6% for the Limb Scale and that for the Bulbar Scale was 66.7%. This indicates unidimensionality of both Scales. The ICCs for the total scores were .97 (95%C.I. .95-.99) for the Limb Scale and .86 (.73-.93) for the Bulbar Scale, respectively. On the basis of these results, the Scale has unidimensionality and high reliability enough for practical use.