4th International Health Sciences and Management Conference , İstanbul, Türkiye, 20 - 23 Haziran 2019, cilt.1, sa.1, ss.457-460
Research Problem: Early diagnosis of diseases in health services is very important for individuals, society and country economy. Early diagnosis affects both the patient’s quality of life and the cost of treatment. However, it is known that every procedure applied on the behalf of early diagnosis increases the health costs and puts a significant burden on the country's economy. Aim of the Study: The aim of this study was to determine the direct diagnosis costs for disease groups. In addition, another aim of this study was to estimate the direct diagnose costs for the next years and to examine other variables (such as gender, age, seasons) that affect the direct diagnose costs. Methods: The population of this study consisted of 122.431 patients who applied to Family Medicine Outpatient Clinic of SBU Şişli Hamidiye Etfal Training and Research Hospital between 01.01.2016 - 15.08.2018. After removing missing, noisy and inconsistent data records, a total of 29.025 patient data including the patient's symptoms and direct diagnosis costs were included in our study. Our study is single centered, cross-sectional, retrospective, descriptive and analytical. In addition, naive bayes and decision trees were used in the classification of diseases and the prediction of diagnosis costs by using variables which have explanatory feature and affect the cost of diagnosis. The dataset is splitted into training and test sets, with a 75-25 ratio. Findings and Conclusion: When the direct diagnosis cost was examined for three years, the total cost was 1,509,761 TL and the average cost (cost per person) was 52 TL. While the total costs were 731,829 TL, 429,454 TL and 348,476 TL, the average costs were 53.9 TL (17.8 $), 48.7 TL (13.3 $) and 52.6 TL (10.9 $) in 2016, 2017 and 2018, respectively. The diseases have been classified by ICD-10 diagnostic groups and it was determined that the highest average cost belonged to the group of endocrine, nutritional and metabolic diseases with 63 TL and the lowest average cost was 33 TL with genitourinary diseases group. It was found that the direct diagnosis costs were significantly different by season, year, age, and gender (p <0.05). It was concluded that the highest accurate classification success is achieved by using decision trees. It was also shown that decision trees is an effective method for estimating diagnosis costs.
Note: 2016 USD average of selling rate = 3.02 TL; 2017 USD average of selling rate = 3.65 TL; 2018 USD
average of selling rate = 4,82 TL
Key Words: diagnostic groups, diagnose, diagnosis cost