Abstract
Separating personal health information (PHI) from research data is a critical issue. In our work the Central CodeBook (CCB) functions as an ‘honest broker’ by mapping between internally meaningful, subject codes (IDs) and sensitive external identifiers like medical record numbers. The CCB is a web service enabled database-backed web application that brokers communications with our evolving EMR via the Clinical Information Network Exchange (CNEX).
Introduction
Separating personal health information from research data is a critical current issue. HIPAA regulations and stricter IRB policy have mandated that research databases not contain personnel identifiers. However many research databases (particularly prospective studies) are works-in-progress, meaning that the data need in principle be re identifiable in order to correctly associate new incoming data within the research DB. The common practice is to keep a codebook on paper, locked away, or in an isolated spreadsheet that maps research DB subject identifiers to sensitive medical and personal data. When an update is needed in the research DB, the keeper of the codebook looks up the medical record number in order to find the assigned subject code, or vice versa. One problem that manifests at our institution related to this practice is that the knowledge of which patients are participating in which research study or clinical trial is dispersed and therefore not available. The Central Codebook at the David Geffen School of Medicine at UCLA is one component of an interface to our federated electronic medical record system that was developed to address this set of problems. One function of a ‘codebook’ in a research context is to map between identifiers. In our work it maps between internally meaningful identifiers, subject codes, and sensitive external identifiers like medical record numbers and brokers communications with our EMR. Our poster will present the architecture of the Codebook and the larger framework in which it exists and how it interoperates with other components of the Clinical Network Exchange (CNEX). Being a part of CNEX the Codebook acts as a gateway into the EMR for research systems. The Codebook has the capability to request lab and other patient related data from the EMR. In the case of lab data, the Codebook returns to the research database data identified by subject code and not medical record number. In this sense it serves the role of an honest broker.
Methods
The CCB was initially developed within the open source framework called the OpenACS. The application server is the AOLserver, a multi-threaded http server with an embedded TCL interpreter. The business logic is implemented in TCL and PL/SQL, and the data store is built in an Oracle 9i database instance. Our current effort is to re-architect the CCB in Java around the caCORE framework and the enterprise controlled vocabularies at the heart of the caBIG efforts.
Results
The CCB is utilized by the UCLA Tissue MicroArrray, Prostate SPORE, and Inflammatory Bowel Disease research systems. Combined there are over 5300 patient records managed by the CCB.