We developed a measurement scale for assessment of impairment in MS patients (MSIS) in accordance with the recommendations of WHO. The items were kept close to a standard neurologic examination, and a short battery of cognitive tests was added. Normality was assigned to the value, zero, and the theoretical maximum score was 204. Two-hundred and ten multiple sclerosis (MS) patients were rated by one neurologist on the MS impairment Scale (MSIS), the Extended Disability Status Scale (EDSS), and the Ambulation Index (AI). The median MSIS score was 52 (5-147), the median EDSS 6.5 (1-9.5), and the median AI 5 (1-5). The relation between the MSIS scores and the EDSS was best described by an exponential function (non-linear regression coefficient, R = 0.87). Sixty-two of the patients were reexamined and rated by another neurologist. The interrater reliability coefficient (R) of the MSIS was 0.95, of the EDSS 0.91, and of the AI 0.94. Forty patients were examined twice by the first neurologist. The intrarater reliability coefficient was 0.97 for the MSIS, 0.95 for the EDSS, and 0.98 for the AI. The MSIS is easy to use and is robust to observer dissimilarities. It has a monomodal univariate distribution and has a better discriminatory power than the EDSS, especially in the EDSS range 6-9, while the interrater reliability of the MSIS is at least as good as that of the EDSS and the AI.