Disease case is a diverse and complex problem which faced by health insurance institutions. BPJS Kesehatan, Indonesian national health security, has undoubtedly faced this problem, especially because of the socioeconomic gap between 34 provinces in Indonesia. To describe the diverse and complex conditions of resource needs for all hospital patients, this paper will calculate the Case Mix Index (CMI) for each province and conduct a cluster analysis to group provinces based on the CMI and total health facility visits similarity. Once the CMI for each province is identified, the cluster analysis is executed by using K-Means and Hierarchical clustering method to compare each result. The first step of cluster analysis is to identify the optimal number of clusters. In this article, several K-value selection techniques is used to find out the optimal number of clusters. By using K-value selection methods, the optimal number of clusters is two province clusters. The first cluster, namely Cluster 1, consists of four provinces which are DKI Jakarta, Jawa Barat, Jawa Tengah, and Jawa Timur. The second cluster, denoted as Cluster 2, consists of the rest provinces which are not included in Cluster 1. Although the optimal number of clusters are identified, this paper also adjust the cluster analysis result to provide comparison with the current INACBG regionalization. The result of this paper can be utilized as a recommendation for INA-CBGS tariff regionalization since using CMI as one of the clustering variables could depict the diverse condition in 34 provinces in Indonesia.