Cancer Patterns in Peru's Health Insurance System: Data Science with K-Means

Authors

  • Wildon Rojas Paucar
  • Alberto Octavio Carranza López
  • Elena Miriam Chávez Garcés
  • Elizabeth Luisa Medina Soto

Keywords:

Cancer Patterns, K-Means Algorithm, Health Services, Data Science

Abstract

Introduction: Cancer has become one of the leading causes of mortality in Latin America. This research analyzes the varying healthcare utilization behaviors among cancer patients affiliated with Peru's Comprehensive Health Insurance System (SIS), using data obtained from regional hospital centers through the K-means algorithm.

Objective: To identify patterns or clusters within data from the National Registry of Oncology Care for SIS patients using the K-means clustering algorithm.

Methods: The study adopted a non-experimental, descriptive, and exploratory research approach, analyzing 307,145 oncology care records from the Intangible Solidarity Health Fund (FISSAL) for the year 2023, involving 29,962 patients nationwide.

Results: Six distinct clusters were identified based on profiles of cost and frequency of oncology services authorized by SIS in the affected regions of Peru. Additionally, the performance of the algorithm was scalable and efficient in processing the data, providing valuable and meaningful insights into the segmentation characteristics of each cluster.

Conclusion: Proper application of the K-means algorithm enabled the identification of key behaviors and significant patterns, enhancing the understanding of oncology data dynamics within FISSAL, thus informing the redesign of care strategies and improving health outcomes for cancer patients.

Downloads

Download data is not yet available.

Published

2025-03-20

How to Cite

1.
Rojas Paucar W, Carranza López AO, Chávez Garcés EM, Medina Soto EL. Cancer Patterns in Peru’s Health Insurance System: Data Science with K-Means. Rev Cubana Inv Bioméd [Internet]. 2025 Mar. 20 [cited 2025 Jul. 31];44. Available from: https://revibiomedica.sld.cu/index.php/ibi/article/view/3777

Issue

Section

ARTÍCULOS ORIGINALES