[Using sequence and cluster analysis to characterize variables that unfold over time in life-course epidemiology...

The life-course epidemiology emphasizes that the timing, duration, and interactions of exposures can have life-time health effects, yet it remains challenging to characterize complex dynamic trajectories in epidemiology studies. This study presents practice guidance for implementing sequence and cluster analysis in epidemiological studies to identify long-term trajectories, with an example of constructing a female-specific reproductive life sequence. By ordering the occurrence of reproductive events over women's lifespans, we encoded multiple reproductive events and their timing into ordered-state sequences, then using sequence dissimilarity measures and cluster analysis to identify typical reproductive patterns, describe heterogeneity across patterns, and assess their values for type 2 diabetes risk stratification. This study seeks to provide a guide for researchers on how sequence analysis can be used to reconstruct key life-course event trajectories from multi-source, non-longitudinal data, providing a novel methodological perspective for life-course epidemiology, as well as offering a useful reference for characterizing population health trajectories and examining the impact of cumulative exposures across the life-course.