Investigation of the data mining methods optimum use of physics and human resources in hospitals

In 2008, the ministry of health implemented SağlıkNet project and began to store data produced by SHS (Health Service Providers) according to the HL7 (Healt Level Seven) standards in ministry servers. The SağlıkNet project aims to standardize health data, accelerate data flow among health facilities, support scientific studies and develop proactive approaches with decision support systems. In this way, the data obtained will provide health service managers with operational decisions for their daily activities, as well as it will support the proactive management approach. In this study, SHS aims to use the existing resources more efficiently by processing the data of physical structure, equipment, patient, administrative and health personnel by using data mining methods in order to increase the service quality. In this way, providing adequate staff support at the right time, determining the appropriate policlinic rooms according to the branches, reducing the workload of physicians and will be able to record the produced data faster. The data used in this study were obtained from Afyonkarahisar State Hospital between 2015 and 2016. The collected data were analyzed by data mining methods, time series, market basket analysis and association rules. As a result of the study, it has been seen that the hospital has increased from November to March and it has decreased regularly from April to September. On a weekly basis, most applications were seen on Mondays and decreased until Friday. When patients who applied to more than one outpatient clinic were analyzed by association rules, it was observed that the patients who applied to internal diseases applied to another clinic. It has been determined that the women have more applications than men and the patients in the age group of 50-54 are in the majority. It was seen that 75% of the total application was made until 12:00 and 65% of the total application was examined. In order to help the diagnosis of the patients, it was seen that most of the tests were requested as blood tests. When the market basket analysis of the requested examinations was performed, it was seen that the blood and direct graphical tests were performed together. Keywords: Data Mining, Hospital Resources, Optimization, Health Services Management