Participate Now
WP 2 and WP 3 aims to specify the best statistical approach for estimating — point estimate and confidence interval — an adjusted risk difference in case of a two-parallel group cluster-randomised trial (WP 2), or a cross-over or stepped-wedge cluster-randomized trial (WP 3).
Methods will be compared by using a Monte Carlo simulation study. Realistic scenarios (number of clusters, cluster sizes, level of intracluster correlation, prevalence of the outcome, effect size) will be defined by discussion with members of the scientific advisory board. The number of covariates, the levels of correlations between covariates, and their potential association with the outcome will also be defined with the scientific advisory board. These covariates will be simulated in such a way as to retain the pre-specified intra-class correlation structure for the outcome.
Simulation studies will be performed in agreement with the Aims, Data-generating mechanisms, Estimands, Methods and Performance measures (ADEMP) structured approach.