Forecasting Medical Inflation in Relation to Medical Cost Components: A Comparative Approach Between Traditional Econometric and Machine Learning Models

Abstract

Medical inflation has become a critical concern in modern economies due to its profound impact on healthcare affordability and accessibility. This study examines relationship between medical inflation and its underlying cost components, including Medical Care, Medical Care Commodities, Prescription Drugs, Medical Care Services, Professional Services, Physicians' Services, Dental Services, and Hospital Services. This research aims to forecast future medical inflation rates by employing a comparative modelling framework that integrates both traditional econometric models and modern machine learning techniques. This research applies Vector Autoregression (VAR) and Autoregressive Distributed Lag (ARDL) as traditional econometric models, alongside Random Forest (RF) and Artificial Neural Network (ANN), as the representative machine learning models, to evaluate their effectiveness in forecasting. The research involves analysing time series data and identifying the influence of each medical cost component on overall medical inflation. Both short-term and long-term relationships are assessed using econometric models. In contrast, the machine learning model captures complex, non-linear patterns in the data. Forecasting accuracy is compared across models using standard performance metrics to determine the most reliable method for predicting medical inflation from 2025 to 2030. By combining econometric theory with modern data science techniques, this study offers insights into the drivers of medical inflation. It presents a comprehensive evaluation of forecasting methods. The findings contribute to better healthcare planning and cost management by identifying which components most significantly affect inflation, and which modelling approach yields the most accurate forecasts.