Reliability of outcome measures in clinical trials in secondary progressive multiple sclerosis.

OBJECTIVE To investigate the reliability of clinical outcomes in secondary progressive multiple sclerosis (SPMS) trials, we compared the frequency of progression and "improvement" events on different clinical outcome measures in the placebo arms of 2 large randomized controlled trial (RCT) datasets. METHODS Using original trial data from the placebo arms of IMPACT and ASCEND, 2 large RCTs in SPMS, we compared disability progression and similarly defined "improvement" with and without 3 or 6 month confirmation on the outcome measures EDSS, timed 25 foot walk (T25FW), nine-hole peg test (9HPT), and their combinations. RESULTS In both datasets, the EDSS showed the highest rates of "improvement" over time, and the smallest difference between progression and "improvement" rates, followed by the T25FW and the 9HPT. For the T25FW and 9HPT, "improvement" rates were fairly stable over time and remained at below or around the 10% level. For the EDSS, "improvement" rates increased in parallel with disability progression rates. CONCLUSIONS All investigated outcome measures in SPMS showed some evidence of random variation and measurement error, the T25FW and 9HPT less so than the more established outcome EDSS. Our findings are relevant for the design and critical appraisal of trials in SPMS.