A sensitivity analysis of missing data methods for planning a study: a statistician perspective

Judith Anzures-Cabrera, Statistical Scientist, Roche Products Ltd UK

The statistician working on study X was asked the question “which method are you going to use for the analysis of missing data on HDL cholesterol, the primary endpoint?” At that point the statistician did not know how to answer the question. Luckily for her she had access to study Y, a previous study from the same program, where the same endpoint had been collected at similar times. Using data from study Y, the statistician generated three datasets assuming different missing data mechanisms: MCAR, MAR and MNAR. By carefully considering the question that she wanted to answer with the missing data model, the statistician decided to undertake a sensitivity analysis of missing data methods based on a MAR assumption (multiple imputation, MMRM and BOCF). At the end of this talk if you will find out which one was the method that the statistician selected for her statistical analysis plan.