Adoption and Effectiveness of AI-Based Anomaly Detection for Cross Provider Health Data Exchange
Summary: arXiv:2604.09630v1 Announce Type: cross
Abstract
This study investigates the adoption and effectiveness of AI-based anomaly detection in cross-provider electronic health record (EHR) environments. It aims to:
- Identify the organisational and digital capabilities required for successful implementation.
- Evaluate the performance and interpretability of lightweight anomaly detection approaches using contextual audit data.
Methodology
A semi-systematic scoping synthesis is conducted to derive a four-pillar readiness framework. This framework covers:
- Governance
- Infrastructure/Interoperability
- Workforce
- AI Integration
The framework is operationalised as a 10-item checklist with measurable indicators. Additionally, a simulation of cross-provider audit logs is performed, incorporating contextual features such as:
- Provider mismatch
- Time of access
- Days since discharge
- Session duration
- Access frequency
Performance Evaluation
A rule-based approach is benchmarked against an Isolation Forest model, with SHAP (SHapley Additive exPlanations) used to explain model behaviour. The results show that:
- Rule-based methods achieve high recall but generate a higher volume of alerts.
- Isolation Forest reduces alert burden at the cost of lower sensitivity.
Key Findings
SHAP analysis reveals that provider mismatch and off-hours access are dominant drivers of anomalies. The study proposes a staged deployment strategy that combines:
- Rules for comprehensive coverage
- Machine learning techniques for prioritisation
This strategy is supported by explainability and continuous monitoring, ensuring that the implementation of AI-based anomaly detection is both effective and interpretable.
Conclusion
The findings contribute a practical readiness framework and empirical insights to guide the implementation of AI-based anomaly detection in multi-provider healthcare environments. By understanding the organisational and digital capabilities necessary for successful deployment, healthcare providers can enhance the security and efficiency of EHR systems, ultimately improving patient care and safety.