Facilitating patient and administrator analyses of electronic health record accesses

Electronic Health Records;Software Engineering;Audit Logs;Anomaly Detection;Healthcare Security

Duffy, Eric ; Gunter ; Carl A.

关键词: Electronic Health Records;    Software Engineering;    Audit Logs;    Anomaly Detection;    Healthcare Security;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/45545/Eric_Duffy.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

The past two decades in the United States have ushered in an era of increasing ubiquity of digitized healthcareas the speed and sophistication of technology follows an ever-growing trend. Electronic health records (EHRs)are an integral part of the growing healthcare industry which offers ease of access and new functionality whilesimultaneously causing worries over their privacy and security. In an effort to address these concerns, muchlegislation has been enacted in order to tighten the oversight and requirements for accessing protected healthinformation (PHI). Most recently, the Department of Health and Human Services has released rulemakingwhich requires providers utilizing EHRs to comply with patients’ requests for logs of the accesses to theirrecords.In this work, we outline our system for complying with this regulation while easing the burden of compli-ance for providers and simultaneously providing patients with informative and satisfying information aboutwhy their accounts were accessed. We implement a system called the Multiview Audit Interface (MAI) whichutilizes recent research in the data mining and anomaly detection communities to provide a unified interfacefor conveniently using these algorithms for patients and administrators. We then test this system on a de-identified access log from Northwestern Memorial Hospital containing months of audit data. We constructa framework for implementing these algorithms as modules, thereby recycling existing code, encouragingmulti-faceted comprehensions of their results, and offering an easy-to-use interface that administrators andpatients can use alike. We demonstrate the the power of three modules currently implemented and showhow the extensibility of the framework can be harvested to develop modules in the future.