005-23 – What can relatives tell us about MS patients' cognitive function? MSNQ-I Rasch analysis and relation to patient demographic and disease variables.

Introduction: Formal cognitive assessment in multiple sclerosis (MS) is expensive. Informal evaluations of health professionals and patients are unreliable. Relative report is more accurate. The MS Neuropsychological Questionnaire for informants (MSNQ-I) is a possible solution to quantify MS patient cognition. A Rasch-transformed scale has not been derived from a large sample & it is unknown how patient variables affect relatives‘ perception of patient cognition.

Aims: To use Rasch-transformed measurements of the MSNQ-I to quantify MS patient cognitive status and determine their relation to other variables.
Methods: MSNQ-I scores from relatives of 2039 MS patients, from the Trajectory of Neurological Conditions database, with relevant patient variables, were analysed using the Rasch model and multiple linear regression. The Rasch analysis had two random n=500 samples (exploration and validation).

Results: The Rasch analysis revealed 3 items related to executive function, which were omitted from the final solution, resulting in a 12-item scale (MSNQ-I-12) with excellent model fit, unidimensional and reliability (Cronbach’s Alpha=.958). In the regression, 7 variables accounted for a significant amount of variance on the MSNQ-I-12 (F(7,1798)=99.43, p<.001; R2=.28, adjusted R2=.28). The partial regression coefficients showed that variables which had a significant unique relationship to the MSNQ-I-12 were: gender (t(1799)=4.35, p<.001), Expanded Disability Status Scale band (t(1799) = -5.17, p<.001), Hospital Anxiety and Depression Scale–Anxiety (t(1799)=2.63, p=.009), Neurological Fatigue Index (t(1799) = 7.81, p<.001), London Handicap Scale (t(1799)=10.17, p<.001) and the EuroQoL Instrument (t(1799)=2.13, p=.034).

Conclusion: The MSNQ-I-12 generated interval level measurement. When using the MSNQ-I 12, consideration should be given to other patient variables which relate to relatives’ perception of patient cognitio