The non-linear relationship between self-reported cognitive concerns and performance-based cognitive dysfunction in people with multiple sclerosis: a cross-sectional study.

Existing research suggests a weak, inconsistent link between self-reported cognitive concerns (SCC) and performance-based measures of cognitive dysfunction in people with multiple sclerosis (pwMS). This study aimed to evaluate for a non-linear relationship between SCC and cognitive impairment in pwMS. A consecutive sample of 933 pwMS routinely completed the Minimal Assessment of Cognitive Function in MS (MACFIMS) battery at a Canadian neuropsychiatry clinic between 2020 and 2025. Covariates included age, sex, Expanded Disability Status Scale scores, disease duration, disease course, Wechsler Test of Adult Reading scores, and Hospital Anxiety and Depression Scale sub-scale scores. Cognitive impairment was defined as ≥ 3 MACFIMS raw scores ≥ 1.5 SD below matched normative data. Using a grid search to identify a breakpoint in Perceived Deficits Questionnaire (PDQ) scores (to assess SCC), a segmented logistic regression analysis was undertaken to predict cognitive impairment from PDQ scores, controlling for covariates. Mean participant age was 43.34 years (SD 10.89), 75.12% were female, median EDSS was 2.00 (IQR 1.50-3.50), and 46.73% had cognitive impairment. PDQ scores ≥ 50 independently predicted increased odds of cognitive impairment (OR = 1.92, 95% CI [1.22-3.02], p = 0.0052), while PDQ scores < 50 were not associated with cognitive impairment (OR = 1.03, 95% CI [0.87-1.21], p = 0.75). In conclusion, above a critical threshold, SCC indicate increased odds of cognitive impairment.