Analysis of drug prescription patterns by doctors with machine learning algorithms
نویسندگان
1 Department of Industrial Engineering, Ankara Yıldırım Beyazıt University, Ankara, Turkey.
2 National Center for Health Insurance Research, Tehran, Iran.
doi
10.22105/jarie.2025.487148.1704چکیده
Access to high-quality medical and pharmaceutical services is essential for promoting community health and well-being. This study aims to explore and analyze physicians' prescribing behaviors across different healthcare settings to uncover key patterns and disparities. The primary objective is to investigate the factors influencing prescription patterns in governmental centers versus private clinics, focusing on variations driven by regulatory constraints, economic pressures, and clinical practices. The study's contribution lies in its data-driven approach to uncovering behavioral trends and in promoting evidence-based strategies to support resilient, efficient healthcare systems. This study applies data analysis techniques, including data extraction, normalization, and clustering using machine learning algorithms, to real-world data from an insurance company in Qom province, Iran. The results indicate significant differences in prescribing behaviors between physicians in governmental and private clinics, particularly in the number of prescribed items and the types of medications. Economic factors, regulatory frameworks, and the healthcare setting itself influence these differences. The findings also highlight distinct prescribing patterns and outliers, revealing potential inefficiencies in resource utilization and opportunities for improvement in healthcare service delivery. By identifying key drivers of prescribing behavior, the study offers actionable insights for policymakers and healthcare administrators to design targeted interventions that optimize resource allocation, improve healthcare quality, and enhance patient outcomes.