Machine Learning Analysis of Consolidated Purchasing: A Case Study of Antiretroviral Medication 2019 Pricing Trends in Mexico

Abstract

This paper investigates trends in antiretroviral medication prices and their impact on public health in Mexico during 2019. Using three machine learning models developed in Python (logistic regression, random forest, and K-Nearest Neighbors or KNN), this study discerns increasing or decreasing patterns in antiretroviral (ARV) drug price fluctuations using a dataset comprising 15,220 observations of ARV drugs acquired between 2016 and 2019. Results indicate that random forests exhibited the highest precision in predicting price changes, followed by KNN and logistic regression. Significant factors affecting acquisition prices, such as drug type and duration of procurement strategy, were identified. In addition to analyzing price trends, the paper explores the budgetary considerations associated with these fluctuations, providing insights into the financial implications for healthcare systems and stakeholders. It is important to note that this paper focuses on a specific ARV pharmaceutical purchasing scheme. Moreover, the study emphasizes the creation of a unified and detailed medication price database, highlighting the significant effort invested in compiling complete and comprehensive information from various sources. This study’s findings underscore the effectiveness of initiatives such as consolidated purchasing approaches and the integration of newer, cost-effective medications into treatment protocols. These initiatives have led to significant cost savings in antiretroviral medication procurement, contributing to improved access for individuals living with HIV/AIDS. Overall, the research highlights the importance of datadriven approaches and strategic planning in optimizing pharmaceutical purchasing processes and ensuring sustainable access to essential medications for public health interventions.