Intracluster correlation measures in cluster randomized trials with binary outcomes

Ariane M Mbekwe Yepnang (1) , Agnès Caille (1, 2) , Sandra M Eldridge (3) , Bruno Giraudeau (1, 2)

1 SPHERE - MethodS in Patients-centered outcomes and HEalth ResEarch
2 CIC - Centre d’Investigation Clinique [Tours] CIC 1415
3 QMUL - Queen Mary University of London

In cluster randomized trials, a measure of intracluster correlation such as the intraclass correlation coefficient (ICC) should be reported for each primary outcome. Providing intracluster correlation estimates may help in calculating sample size of future cluster randomized trials and also in interpreting the results of the trial from which they are derived. For a binary outcome, the ICC is known to be associated with its prevalence, which raises at least two issues. First, it questions the use of ICC estimates obtained on a binary outcome in a trial for sample size calculations in a subsequent trial in which the same binary outcome is expected to have a different prevalence. Second, it challenges the interpretation of ICC estimates because they do not solely depend on clustering level. Other intracluster correlation measures have been proposed for clustered binary data settings including the variance partition coefficient, the median odds ratio and the tetrachoric correlation coefficient. Under certain assumptions, the theoretical maximum possible value for an ICC associated with a binary outcome can be derived and we propose to consider the relative deviation of an ICC estimate to this maximum value, as another measure of the intracluster correlation. We conducted a simulation study to explore the dependence of these intracluster correlation measures on outcome prevalence and found that all these measures are associated with prevalence. Even if some were slightly less dependent than the ICC in some scenarios, in general, none differs from the ICC regarding the dependence on prevalence.

Mots clés: Intraclass correlation coefficient; binary outcome; prevalence; cluster; variance partition coefficient; median odds ratio; tetrachoric correlation coefficient

Domaines: Médecine humaine et pathologie

Ariane M Mbekwe Yepnang, Agnès Caille, Sandra M Eldridge, Bruno Giraudeau. Intracluster correlation measures in cluster randomized trials with binary outcomes. Statistical Methods in Medical Research, 2021, 30 (8), pp.1988-2003. ⟨10.1177/09622802211026004⟩. ⟨hal-03405804⟩