Email Post: Evaluation of bias and precision in methods of analysis for pragmatic trials with missing outcome data: a simulation study