De-Identification as a Critical Success Factor in Clinical and Population Research
Dee Lang RHIT; De-ID Data Corp; Steven Merahn MD; De-ID Data Corp;
Content:
Protecting the patient privacy and confidentiality is a major issue when planning data repositories for quality, clinical and population research and developing searchable databases and rapid learning health systems. Regardless of primary data sources -- whether transcribedreports, databases or electronic medical records with free-text fields-- de-identification of medical records and patient data is required. Whether you are a clinical or public health researcher, an informatics professional or a quality manager seeking a HIPAA- compliant solution tocaccess aggregated data, rapid, reliable, accurate and cost-efficient de-identification is a new mandatory to advance research and rapid learning systems while ensuring patient privacy. This presentation will explore experience with choosing and using de-identification systems to expand data access and leverage and manage data assets for cancer and pathology research and the creation of data warehouses and repositories that assure patient privacy. Hear about the process of choosing a de-identification methodologies and how they are integrated into clinical and research workflow. The presenters will present specific examples of how de-identification methodologies have been applied to create concept-based access to patient records and reports.
Technology:
De-ID software, working of a Microsoft operating system, uses heuristics and rule sets to identify HIPAA safe harbor patient identification elements and replace them with proxies and offsets; a JAVA wrapper is utilized for in-line connectivity or automated data flow
Design:
DE-ID is inserted in the data management workflow and generates a second, de-identified record which can then be fed to a data repository with less restrictive access
Results:
Over 10 major academic institutions are use DE-ID as part of their patient privacy methodologies within a data management workflow, permitting new levels of data aggregation and sharing among individuals and institutions for clinical performance, quality and population research, especially in oncology.
Conclusion:
Automated deidentification tools have a place in the overall patient privacy methods associated with large data management projects and can enhance and advance the development of multi-institution data repositories
