Impact of informative censoring on the treatment effect estimate of disability worsening in multiple sclerosis clinical trials.

•Baseline characteristics affect risk of 12-week confirmed disability worsening.•Standard analytical method is to censor participants with missing confirmation status.•Censoring at time of disability worsening can bias the treatment effect estimate.•Censoring when confirmation is missing can bias the treatment effect estimate.•Observed rate multiple imputation method based on risk factors minimizes bias.