Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales.

BACKGROUND Patient-reported outcome (PRO) scales are often used in multiple sclerosis (MS) research. Full understanding of items can be influenced by disease worsening, mood disturbances and cognitive problems of the MS patient. Earlier research with the Multiple Sclerosis Impact Scale (MSIS-29) showed that proxy respondents (i.e. partners of patients) can provide useful information. OBJECTIVE To determine agreement between patients and proxy respondents on different MS PRO scales. METHODS 139 Patients and partners completed the MSIS-29 (Physical and Psychological scale), Multiple Sclerosis Walking Scale (MSWS-12), Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ) and Guy's Neurological Disability Scale (GNDS). We calculated the mean difference and intra-class correlation coefficients (ICC) on scale level and weighted kappas (κ(w)) on item level. RESULTS On all scales, except MSNQ, the partner score was higher. ICCs were good for MSWS, GNDS and MSIS Physical, and moderate for MSNQ and MSIS Psychological. κ(w) was excellent for MSWS items, fair to good for GNDS, MSIS Physical and MSIS Psychological items, and poor for MSNQ items. CONCLUSION Partners of patients with MS can be a useful source of information for several PRO scales, especially when the focus is on physical functioning. For psychological functioning this seems to be less reliable.