Privacy Techniques in Health Data Mining

Hospitals and Doctors collect and store a myriad of health data in the course of their work. This data is often siloed and inaccessible to data scientists and medical researchers mainly because of the complex web of privacy regulations and harsh penalties for exposing private information. This review aims to review the current state of research into privacy techniques which enable data mining on health data by; analyzing recent journal papers, comparing the input and output of the different methods, reviewing the processes that researchers undertook and identifying gaps in the research. Electronic Health Records (EHRs) are complex, and the information included can be a mix of structured and unstructured data. This review found that it was common for researchers to deal with either structured or unstructured data, but usually not both in their work. The research methodology and testing rigour of the various papers made it quite challenging to compare apples to apples. The most beneficial contribution at this stage is a recommendation to; create standards for data collection and minimum standards for research that will enable the field to move forward from a solid foundation.