
5
Sample Size Determination for Correlated
Outcomes from Two-Level Randomized
Clinical Trials
5.1 Introduction
Two-level data structures arise frequently in randomized clinical trials. For ex-
ample, in cross-sectional studies which enroll clusters or groups of subjects and
evaluate them at a single point in time, the subjects nested within the groups
would be considered the first level and the organizational entity that forms
the groups (e.g., households or clinics) are the second level. This two-level
structure also arises in longitudinal trials in which repeated measurements
are obtained on study subjects during follow-up. In this case, the repeated ...