PREDICTING CYBERSECURITY RISK IN HEALTHCARE PHARMACY INFRASTRUCTURES

Abstract

In an era of increasing cyber threats against healthcare institutions, medical pharmacies are emerging as critical yet vulnerable components of the digital health ecosystem. This study presents a comprehensive machine learning–based framework for predicting cybersecurity risk in pharmacy environments using operational, threat, and control-related features. We evaluated the predictive performance of three regression models Linear Regression, Support Vector Regressor (SVR), and Random Forest Regressor, using metrics such as R² score, RMSE, and MAE. Random Forest outperformed all models with an R² of 0.91, RMSE of 0.42, and MAE of 0.28, confirming its superiority in capturing non-linear relationships within pharmacy operations. For binary risk classification, the Random Forest Classifier achieved an AUC of 1.00, with a confusion matrix showing high precision (91.4%) and recall (87.6%). Feature importance analysis revealed that control effectiveness, threat probability, and asset value were the most influential factors affecting cybersecurity risk scores. These insights provide actionable guidance for pharmacy security planning and risk mitigation. The proposed framework is scalable, interpretable, and compatible with real-time security monitoring systems. Our findings support the integration of AI into pharmacy IT governance and regulatory compliance strategies to transform risk management from reactive defense to proactive threat anticipation.