Using sequence and cluster analysis to characterize variables that unfold over time: implementation and practical considerations for epidemiologists

Characterizing longitudinal trajectories of variables that unfold over time (eg, social, health, or environmental variables) is a persistent challenge, but can be accomplished with sequence and cluster analysis, data-driven approaches that can differentiate timing, order, and duration of events. We present practical guidance on implementing sequence and cluster analysis for epidemiologists with the goal of providing clear advice on decision points and tradeoffs. We introduce the three main steps of sequence and cluster analysis: (1) coding trajectories of ordered events (data cleaning); (2) measuring dissimilarity between trajectories (sequence analysis); and (3) grouping similar trajectories (cluster analysis). Each of these steps presents researchers with several decision points, such as data cleaning rules, options for evaluating sequence dissimilarity, and choices of clustering algorithms. After outlining each of the sequence analysis steps, we provide an applied example of sequence analysis in which we create and group transition-to-retirement trajectories from age 51 to 75 years for a sample of 9189 Health and Retirement Study participants using self-reported employment information, then estimate the association between transition-to-retirement groups and self-rated health. We seek to provide an initial guide for epidemiologists through analytic decisions and implementation challenges of sequence analysis as this approach is increasingly implemented and undergoes methodological advances. This article is part of a Special Collection on Methods in Social Epidemiology.