Comparison of results from different imputation techniques of missing data in an anti-obesity drug trial

Tags: Poster
Jørgensen} A1, Lundstrøm L2, Wetterslev J2, Astrup A3, Gøtzsche P1
1The Nordic Cohrane Centre, Denmark, 2Copenhagen Trial Unit, Centre of Clinical Intervention Research, Denmark, 3Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen, Denmark

Background: In randomised trials of medical interventions the most reliable analysis follows the intention-to-treat (ITT) principle. However, the ITT analysis requires that missing outcome data have to be imputed. Different imputation techniques may give different results and some may lead to bias. In anti-obesity drug trials, many data are missing and the most used imputation method is last observation carried forward (LOCF). LOCF is generally considered conservative, but there are more reliable methods such as multiple imputation (MI).

Objectives: To compare four different methods of handling missing data in a 60-week placebo controlled anti-obesity drug trial.

Methods: We compared an analysis of complete cases with datasets where missing body weight measurements had been replaced using three different imputation methods: LOCF, baseline carried forward (BOCF) and MI.

Results: 561 participants were randomised. At the end of the trial, 85% of the participants were lost to follow up. Compared to placebo, there was a significantly greater weight loss with the drug in all analyses: 9.5kg (SE 1.17) in the complete case analysis (N = 86), 6.8kg (SE 0.66) using LOCF (N = 561), 6.4kg (SE 0.90) using MI (N = 561) and 1.5kg (SE 0.28) using BOCF (N = 561).

Conclusions: The different imputation methods gave very different results. LOCF was not conservative compared to the more reliable MI, in fact LOCF had a lower SE. Therefore, meta-analyses on anti-obesity drugs based on an ITT analysis using LOCF are likely to be biased.