Evaluation of Methods for Dealing with Outcome Missing Data in Random Controlled Trials

• Activity Number: 251
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Breda Munoz*+ and Virginia M. Lesser and James Hersey
• Companies: RTI International and Oregon State University and RTI International
• Keywords: random controlled desugbs ; missing data ; imputation

Abstract

As in most clinical trials, random controlled trials are also affected by missing data as the result of participants discontinuing the trial. Methods used to account for missing data include analyzing the observed data, using the last information carried forward approach, and applying complex model based approaches for longitudinal data that incorporate drop out models. We focused on single and multiple imputation approaches where donors for the imputations are selected using propensity score models, model based approaches, and nearest neighborhood approaches. We used simulations to evaluate the properties of these approaches under different rates of missing data. Finally, we illustrate the methods using data from a three arm weight management random controlled trial.