Türkiye Kamu Hastanelerinde Tıbbi Cihaz Kullanımının Etkinliğinin Veri Zarflama Analizi Kullanılarak Tespit Edilmesi ve Yapay Sinir Ağları ile Modellenmesi

Gamze ALDEMİR

Türkiye Kamu Hastanelerinde Tıbbi Cihaz Kullanımının Etkinliğinin Veri Zarflama Analizi Kullanılarak Tespit Edilmesi ve Yapay Sinir Ağları ile Modellenmesi

Thesis Type: Postgraduate

Institution Of The Thesis: Erciyes University, Fen Bilimleri Enstitüsü, Turkey

Approval Date: 2019

Thesis Language: Turkish

Student: Gamze ALDEMİR

Supervisor: Hatice Erkekoğlu

Abstract:

The investments made in the field of health in Turkey have ensured a significant increase parallel to the technological developments during the recent years. The reforms made in service delivery have created a competition environment between the hospitals. The developing competition environment allowed hospitals to work more tediously by effective usage of their respective current sources such as budget, workforce, medical devices, technology. The purpose of this study is to calculate the efficiency of public hospitals bound to Ministry of Health with the help of Data Envelopment Analysis (VZA) and develop an Artificial Neural Network (YSA) model for calculation of efficiency for the current and future hospitals. The data used for efficiency analysis was acquired from Public Hospitals Associations General Service Information Report Bulletin for 2015 issued annually by the General Directorate of Public Hospitals (KHGM). The medical device information at hospital level used for the study was procured with an application made to Ministry of Health. The study was limited to 117 hospitals of A1 and A2 groups. The branch hospitals and hospitals of group B, C, D, E were not included into the study as, respectively, they served for a specific group of patients and diseases and they were affected by the population areas they served. The efficiency analysis was conducted with two different methods, namely CCR and BCC. 49 and 73 of 117 hospitals were considered to be efficient according to CCR model and BCC model respectively. The artificial neural network model was developed to estimate the efficiency after Data Envelopment Analysis. Different neuron combinations were tried in YSA model and the tests were limited with 5000 iterations. The network model was determined with minimal error margin. The test data was examined in this model and estimation and real values were compared and high ratio of success was proven for the model.

Keywords:Efficiency, Data Envelopment Analysis, Artificial Neural Networks, Public Hospitals, Effectiveness