Implementation of a K-Means-Based Intelligent Patient Complaint Clustering System to Identify Handling Priorities

Patient complaints are the body’s response to health disturbances, triggered by internal factors such as genetics or external ones like the living environment. Understanding these causes allows community health centers (puskesmas) to take more effective preventive measures and design more targeted services. This study utilizes patient complaint data sourced from medical records, which include biodata and medical history, as well as complaint details that form the research subject. The main goal of this study is to develop an intelligent system that can generate clusters of patient complaints using the K-Means Clustering algorithm. The system is developed using the Research and Development (RnD) method. The clustering process applies a data mining approach, producing clusters based on patient complaints. A total of 600 complaint records, categorized into 72 distinct types, were used. The output consists of three clusters: C1 (high intensity) with 24 categories, C2 (moderate intensity) with 14 categories, and C3 (low intensity) with 34 categories. A practicality test yielded a score of 0.81, indicating the system is highly practical, while an effectiveness test by medical staff scored 0.88, showing the system is highly effective. This system enables health centers to identify trending complaints in the community and develop more focused prevention and treatment strategies. The clustering results also serve as a valuable foundation for strategic decision-making in disease control.