Abstract
A nested case-control is a case-control study within a cohort. An advantage of the nested case-control design is the number of subjects for whom outcome measures are needed are small compared to the cohort size, which is ideal when acquisition of outcomes is costly and time consuming and prevalence of cases is small. Existing selection schemes such as 1:1, 1:m, and m:n have been implemented in various studies. Upon choice of scheme, implementation requires consistent implementation across all strata, which is usually too restrictive in nested case-control studies. To maximize the usage of the available information, we propose a flexible cluster matching algorithm, in which multiple cases are matched to a group of selected controls. This may lead to unbalanced designs and induce complicated correlation structures. Mixed Effects models provide a flexible modeling framework to account for the clustering structure together with other study designs, such as repeated measures. An illustration from a nested case-control study consisting of longitudinal data is presented.