Application of the K-Means Method for Disease Clustering in Medical Records at Puskesmas Jatiwangi

Medical record data is often underutilized, limiting opportunities to analyze disease distribution patterns. At Puskesmas Jatiwangi, such data is primarily used for documentation purposes, providing minimal insights to support data-driven decision-making. This study applies the K-Means Clustering method to group disease types based on specific similarities, enabling better identification of disease distribution patterns. The study utilized 556 patient medical records from October 2024, including attributes such as age, gender, and disease diagnosis. The analysis followed the Knowledge Discovery in Databases (KDD) process, which involves data selection, preprocessing, data transformation, K-Means algorithm implementation, and clustering evaluation using the Davies-Bouldin Index (DBI). Testing was conducted with varying k values from 2 to 10 to determine the optimal number of clusters. The results indicated that the best DBI value of 0.847 was achieved at k = 2, forming two main clusters. The first cluster represented common diseases such as acute respiratory infections (ARIs) and toothaches, while the second cluster included specific conditions like sciatica and acute lymphadenitis. This study demonstrates that the K-Means method is effective for clustering medical record data, providing valuable insights into disease distribution patterns and aiding in the development of targeted health policies. Keywords: K-Means, Medical Records, Clustering, Davies-Bouldin Index, Puskesmas Jatiwangi