[Application and case study of group-based multi-trajectory model in longitudinal data research]

The development of longitudinal cohorts has made the identification and surveillance of multiple biological markers and behavioral factors which influence disease course or health status become possible. However, traditional statistical methods typically use univariate longitudinal data for research, failing to fully exploit the information from multivariate longitudinal data. The group-based multi-trajectory model (GBMTM) emerged as a method to study the developmental trajectory of multivariate data in recent years. GBMTM has distinct advantages in analyzing multivariate longitudinal data by identifying potential subgroups of populations following similar trajectories by multiple indicators that influence the outcome of interest. In this study, we introduced the application of GBMTM by explaining the fundamental principles and using the data from a health management study in the elderly by using smart wearing equipment to investigate the relationship between multiple life-related variables and hypertension to promote the wider use of GBMTM in longitudinal cohort studies.