Meaningful change in cognition in multiple sclerosis: method matters
L A S Walker1, P D Mendella, A Stewart
1Neuropsychology Service, The Ottawa Hospital Faculty of Medicine, Division of Neurology, University of Ottawa School of Psychology, University of Ottawa, Ottawa, Ontario, Canada.
Objective:
To determine if different methods of evaluating cognitive change over time yield measurably different outcomes.
Methods:
Twelve cognitively impaired patients with clinically definite Multiple sclerosis (10 relapsing-remitting, 2 secondary progressive) underwent neuropsychological testing (baseline, 6, 12 months). Data was analysed using: t-tests evaluating group differences on individual tests, group differences in composite scores, reliable change analyses at the level of the individual, and comparisons regarding number of tests failed at each time point.
Results:
Group t-tests on individual tests yielded no change. When tests were grouped according to theoretical constructs, analyses revealed change in processing speed. Reliable change estimates revealed that 16% of the sample deteriorated. When change was measured with respect to the number of domains affected at each time point, 58% of the sample deteriorated on at least one subtest.
Conclusions:
Methodology has a significant impact on interpretation of longitudinal data. In the same group of subjects, traditional group analyses documented no change in individual test scores or change on a single composite score. Analyses of individual results documented change from 16 to 58% of the sample. Advantages and disadvantages of each method were discussed. Findings have implications for interpretation of longitudinal studies.
