Predictive Detection of Healthcare Billing Fraud Through Explainable AI and Secure Data Governance

Source abstract

As healthcare systems become increasingly digital and claims are processed electronically, healthcare billing fraud has become a major problem for healthcare providers, insurance companies, and government healthcare agencies. Traditional fraud detection methods relying on manual auditing and rule-based techniques are often ineffective against complex and evolving fraud schemes. This study introduces a predictive healthcare billing fraud detection framework integrating Explainable Artificial Intelligence (XAI) and secure data governance. Using the NHIS Healthcare Claims and Fraud Dataset, machine learning algorithms including XGBoost, Random Forest, and Logistic Regression are applied to classify fraudulent and legitimate healthcare claims. SHAP and LIME are incorporated to improve transparency and interpretability by explaining the influence of individual features on fraud predictions. The framework also integrates privacy protection, audit logging, role-based access control, and data encryption to safeguard sensitive healthcare information. Performance is evaluated using accuracy, precision, recall, F1-score, and ROC-AUC. The results demonstrate that combining predictive analytics, XAI, and secure data governance can improve healthcare billing fraud detection while supporting transparent, trustworthy, and secure healthcare financial management systems.