COMPREHENSIVE REDACTION FOR NEUROLOGICAL IMAGING -
DESCRIPTION (provided by applicant): In neuroscience, the potential for large scale collaboration and data sharing is seriously under- mined by concerns over the management and handling of personal health information (PHI) in neuroimagery data sets. In particular, HIPAA and HITECH rules mandate substantial measures for the preservation of subject privacy. For the researcher, whose focus is on the science, the management of information privacy is a wholly burdensome task. The vast potential for sharing and meta analysis of neuroimagery in research goes unfulfilled for two reasons: (i) systematic sanitization is hampered by the dearth of tools to systematically expunge PHI from DICOM images, and (ii) an established workflow that integrates sanitization into the data sharing process remains conspicuously absent. This proposal seeks to address both of these challenges by augmenting an established and popular software framework for managing and sharing neuroimagery - the Extensible Neuroimaging Archive Toolkit (XNAT) - with a toolset and integrated workflow for redaction of PHI in DICOM image files. The redaction process to be engaged by this effort differs from naive sanitization tech- niques in that it uses pseudonymous identifiers in a Privacy Mapping Database to disambiguate between subjects without tracking their identities. This approach gives researchers the maximum power and flexibility in sharing neuroimagery datasets, while transparently coping with PHI considerations in a standardized data curation process. The proposed redaction toolset complements the data curation and management tools within XNAT and folds neatly within existing XNAT operational workflows and lab processes. The toolset and workflow adhere to the secure design principles of psychological acceptability and least privilege, promoting broad user adoption and ensuring subject confidentiality. The sanitization process itself is built upon principles of legal redaction and rules of evidence, providing heightened levels of assurance for scientists and investigators for HIPAA compliance. The long term objectives of this project are to create a comprehensive and practical infrastructure for managing PHI in neuroimagery datasets, and to relieve the burden of the investigator from the technical aspects of data sanitization and redaction. In so doing, this effort will remove substantial obstacles to large-scale collaboration and data sharing in neuroscience.
PUBLIC HEALTH RELEVANCE: In neuroscience, the potential for large scale collaboration and data sharing is seriously under- mined by concerns over the management and handling of personal health information (PHI) in neuroimagery data sets. This proposal seeks to address this problem by augmenting the Extensi- ble Neuroimaging Archive Toolkit (XNAT) with a toolset and integrated workflow for pseudonymous redaction of PHI in DICOM image files. The approach adopted by this effort gives researchers the maximum power and flexibility in sharing neuroimagery datasets, while transparently coping with PHI considerations in a standardized data curation process.
