Exploring patterns of multimorbidity in South Korea using exploratory factor analysis and non negative matrix factorization.

The increasing prevalence of multimorbidity and the co-occurrence of multiple chronic diseases presents a measurable challenge to public health, impacting healthcare strategies and planning. This study aimed to explore disease patterns and temporal clustering using data from South Korea's National Health Insurance Service, spanning 2002-2019. The dataset included approximately 1 million individuals, focusing on those with at least two chronic diseases while excluding individuals who died within five years of follow-up. We analyzed 126 non-communicable diseases, considering only those with a prevalence above 1%, and applied a wash-out period to determine incidence. Exploratory factor analysis (EFA) and non-negative matrix factorization (NMF) were used to identify disease clustering over time. Participants were divided into four groups: men and women in their 50 s and 60 s. EFA identified five patterns in men in their 50 s and seven in their 60 s, while four patterns emerged in women in their 50 s and five in their 60 s. NMF identified 10 clusters for men in their 50 s, 15 in their 60 s, and 16 clusters for women in both age groups. Our study confirms established comorbidity patterns and reveals previously unrecognized clusters, providing data-driven insights into multimorbidity mechanisms and supporting evidence-based healthcare strategies.