Predicting medication wastage using machine learning based on patient beliefs.

Area of Science:

• Health economics
• Artificial intelligence in healthcare
• Pharmaceutical management

Background:

• Medication wastage poses a significant financial and resource challenge to subsidised healthcare systems in Southeast Asia.
• Addressing medication wastage is crucial for the sustainability of healthcare services in resource-constrained regions.

Purpose of the Study:

• To develop and evaluate a machine learning (ML) model for predicting medication wastage in Malaysia.
• To identify key predictors of medication wastage using patient data.

Main Methods:

• A cross-sectional survey of 734 patients in Malaysian public healthcare facilities.
• Data collection included demographics, medication history, and beliefs about medicines via validated questionnaires.
• Evaluation of multiple ML regression models, with performance measured by root mean squared error (RMSE).

Main Results:

• The XGBoost ML model demonstrated superior performance in predicting medication wastage (RMSE = 4.67).
• A parsimonious model using seven features, identified through sequential backward elimination, proved practical for clinical use.
• Significant predictors of medication wastage included beliefs about medicines, age, ethnicity, region, and monthly income.

Conclusions:

• This study pioneers the application of ML to combat medication wastage in Southeast Asia.
• The developed ML model offers a foundation for targeted interventions to reduce waste and optimize medication allocation.
• Findings provide valuable insights for policymakers and healthcare providers in Malaysia and similar healthcare systems globally.