Applying CHAID algorithm to investigate critical attributes of secured interoperable health data exchange

With enhanced interoperability in healthcare environment the exchange of electronic health records (EHRs), both intra and inter organisations, has increased manifold. Sharing of the EHR creates room for illegal disclosures and confidentiality breaches. Interoperable healthcare is a complex system with many independent components. To design a secured framework for such a system, one need to understand the most important security attributes and predict various dependencies among them. The security attributes selected for statistical analysis are taken from the real-time study of patient-doctor relationship existing in any hospital or clinic. Hospitals with functional EHR-systems are the prerequisite of this study. The dependencies in the obtained data are generated through classification technique, chi-squared automatic interaction detection (CHAID). The decision tree obtained is analysed and verified using regression. The paper enabled the identification of the salient feature controlling which would maximally reduce security threats while sharing EHRs in interoperable healthcare environment.