The identification of the cardiovascular risk factor (CVRF) profile of individual patients is key to the prevention of cardiovascular disease (CVD), and the development of personalized preventive approaches. Using data from annual medical examinations in a cohort of workers, the aim of the study was to characterize the evolution of CVRFs and the CVD risk score (SCORE) over three time points between 2009 and 2017. For descriptive analyses, mean, standard deviation, and quartile values were used for quantitative variables, and percentages for categorical ones. Cluster analysis was performed using the Kml3D package in R software. This algorithm, which creates distinct groups based on similarities in the evolution of variables of interest measured at different time points, divided the cohort into 2 clusters. Cluster 1 comprised younger workers with lower mean body mass index, waist circumference, blood glucose values, and SCORE, and higher mean HDL cholesterol values. Cluster 2 had the opposite characteristics. In conclusion, it was found that, over time, subjects in cluster 1 showed a higher improvement in CVRF control and a lower increase in their SCORE, compared with cluster 2. The identification of subjects included in these profiles could facilitate the development of better personalized medical approaches to CVD preventive measures.
Keywords: cluster analysis; longitudinal study; real-world data.