Dynamic and Personalized Access Control to Electronic Health Records

Efficient access control methods are popular, as electronically advised medical services demand the patients’ private medical information. In critical cases, where the patients’ life is in danger, different medical professionals linked to critical situations must be permitted access to patients’ Electronic Health Records (EHRs). The research objective of this work is to create machine learning methods dependent on patients’ medical information and integrate them with an Attribute Based Access Control (ABAC) paradigm. We introduce an ABAC mechanism which is able to permit access to sensitive EHR-systems by using diagnostic context handlers, where contextual information, is utilized to characterize critical situations and grant access to health records. Particularly, we utilize patients’ current medical metrics to estimate the risk of suffering from cerebral infarction, cerebrovascular diseases, or hypertension, by leveraging Artificial Neural Networks (ANNs). The predicted risks of these diseases are evaluated by our machine learning-based context handlers, to predict the criticality of patients’ health state. The developed access control method protects the patients’ heath, and simultaneously contributes to secure access for emergency healthcare professionals to private data. Integrating this predictive mechanism with machine-learning based context handlers proved to be an effective tool to strengthen the access control mechanism’s performance to EHR-systems.