MACHINE LEARNING TOOLS FOR HOSPITAL PHARMACY SUPPLY CHAIN: INVENTORY MANAGEMENT TASKS

Abstract

The pharmaceutical supply chain (PSC) plays a critical role in ensuring the availability, quality, and timely delivery of pharmaceutical products to patients. However, the PSC faces numerous chal lenges, including maintaining product integrity, combating counterfeit drugs, and addressing supply shortages, which can impact healthcare system efficiency and patient access to medications. This study focuses on improving the PSC, specifically within the Central Pharmacy of Hospitals (PCH) in Algeria. The PCH faced challenges related to medication calculations, efficiency, paperwork, and inventory management. To address these challenges, an intelligent supply chain system is devel oped with the objectives of improving efficiency, reducing errors, simplifying documentation, and optimizing inventory management.The developed system comprises modules for inventory man agement, analytics and reporting, and workflow automation. This system utilizes object detection techniques, specifically YOLO and Faster R-CNN algorithms, for medication detection and classifi cation. A custom dataset is generated, encompassing diverse images of medications from various angles and lighting conditions. The dataset is annotated and augmented to enhance performance. Limitations and challenges such as data quality, technical expertise, implementation costs, privacy, security, and system integration are considered. Despite these challenges, the proposed intelligent supply chain system has the potential to simplify operations and enhance the effectiveness of phar maceutical supply chains, eventually benefiting healthcare establishments and improving patient access to essential pharmaceutical products.