Healthcare utilization and chronic condition clusters in multimorbidity patients using weighted k-means: a register-based study in Denmark

Abstract Background The growing burden of multimorbidity challenges the healthcare system due to increased healthcare utilization and uncoordinated care. Identifying patients with multimorbidity who have high healthcare utilization is essential to improve management and reduce pressure on the healthcare system. Objective This study aims to identify and characterize clusters of patients with multimorbidity based on both their chronic conditions and healthcare utilization patterns. Methods A weighted K-means method was applied to a population of 1,184,334 individuals with two or more out of 33 chronic conditions, defined using diagnostic algorithms based on ICD-10 and ATC codes. Sociodemographic variables were applied to describe the identified clusters. Results Four clusters were identified based on chronic conditions and healthcare utilization. Cluster 1 had the highest healthcare utilization and a high burden of both somatic and mental conditions, combined with low social status. Cluster 2, consisting primarily of younger women with mental conditions, showed high use of psychological services, and few somatic conditions. The largest cluster, cluster 3, had low healthcare utilization and consisted of individuals with common, manageable conditions, and relatively high social status. Cluster 4 was defined by older individuals with complex somatic conditions requiring frequent contact with general practitioners and specialists. Conclusion The identified clusters showed varying chronic condition patterns and levels of healthcare utilization. The findings underscore the importance of tailored strategies, particularly for multimorbidity patients with mental conditions taking social status into account, in order to improve care and manage resource use more effectively.