AN ASSESSMENT OF RELIABILITY OF 3 METHODS USED IN EVALUATING STATUS OF MULTIPLE SCLEROSIS PATIENTS

The reliability of three different evaluation methods used in a cooperative clinical trial of the efficacy of ACTH in multiple sclerosis patients was evaluated in a uniformity study that used an efficient statistical design requiring only 10 patients and 5 examiners. The methods were the standard neurologic examination, a scoring system for functional grades and disability status, and a 7-day symptom score. Each patient was examined only 3 times at the beginning of the study and 3 more times 6 days later. No significant differences among the 5 examiners were observed on 82 of the 87 items used to measure neurologic function. With the exception of 1 variable, there were no significant differences among the average values of the sequence of the 3 examinations, nor among the average increments of change in the numerical scores between the first and second trials. In an additional examination in which all 5 examiners simultaneously evaluated 3 patients 1 at a time, it was found that the 5 examiners observed uniformly in all of the neurologic tests. The results of this study indicate that, by and large, the three evaluation methods appear to be reliable in the evaluation of neurologic status when used in a cooperative clinical trial where several investigators contribute data. Furthermore, investigations of reliability in cooperative studies can be performed with the use of efficient statistical designs such as the incomplete Latin-square design. © 1969.