Access to patient data required significant manual and error prone anonymization techniques that significantly impacted research efforts.
Automate anonymization of patient data for research use without staff intervention.
Every patient record requested by researchers required at least twenty minutes of staff time to review and redact to meet HIPAA requirements for anonymization. As a result, some studies requesting thousands of records were either significantly reduced in scope or completely canceled due to prohibitive time and cost requirements.
By leveraging our imaging analytics search engine, we were able to create a uniquely longitudinal solution to the problem. The ability to anonymize the patient record at the study level while maintaining the history from the entire patient jacket provided a solution that solved not only the problem at hand but several others we had not initially considered. The researchers and the administrative staff were able to now leverage a common platform for anonymization.
Using a variety of open source libraries and proprietary machine learning techniques, we were able to create automated workflows that anonymized patient data for research use that passed all HIPAA requirements and staff verifications. This enabled significant new large-scale research requests for data never before possible using manual techniques.