This module will expand upon the earlier presentation on use of linear mixed models and marginal models for analysis of binary correlated outcomes to longitudinal studies where a binary or categorical outcome is measured in one or both eyes over time. Topics to be covered include: advantages of longitudinal analysis over cross-sectional analysis focusing on a single key time point; exploring the longitudinal data structure; choosing a model; model assumptions; choosing a correlation structure; model fitting and checking; interpretation of results; non-convergence issues; and missing data. All concepts will be illustrated using examples of longitudinal analyses from clinical trials or longitudinal observational studies in ophthalmology.