PRIVACY-PRESERVING STATISTICAL ANALYSIS ON HEALTH DATA

Electronic Health Information (EHI) is a very high demanding resource for every researcher in different health related areas. However, privacy acts prevent direct access to this information without patient's consent. Therefore, different solutions have been proposed such as de-identification, on-site analysis, and limited remote access, to preserve the data owner's privacy. Each of those approaches has different drawbacks and/or limitations. For instance, de-identification will reduce data utility because of low precision of the final released data, and also it has a risk of data re-identification. On-site analysis has some physical limitations and time consuming procedures like background checks. Remote access increases security risks, and when data has to be pulled from multiple data resources, it requires patient consent for data disclosure. In this paper, we have proposed a set of privacy-preserving methods for popular health statistical analysis. Using this set of secure protocols, health researchers, as data users, are able to receive the results of their queries from the data owners, while each data custodian can keep their sensitive data private. Proposed methods have been tested using sample data to illustrate the performance of the results in terms of computational and communication complexities.