Data mining: A new opportunity to support the solution of public health issues in Colombia

This paper presents an approach to support the solution of some of the current public health issues in Colombia. This paper tackles two problems in the healthcare sector. The first verifies the proper provision of health services. The second defines the epidemiological profile of the critical patients to analyze the elapsed time between the detection of their diseases and their evolution to a chronic phase. The CRISP-DM (Cross Industry Standart Process for Data Mining) methodology was used to confront these problems. We support the analysis method with the IRHC data (Individual Registers of Health Care). Our proposal for the solution of the first problem used both the sequential patterns and sequential clustering techniques. Also, we used clustering and association rules to solve the second problem. The results show that it is possible to analyze different aspects related to the quality of service of the medical attention. These results could be used to improve the decision making about the medical services in Colombia.