CIDER Version Upgrade to Launch at Washington University St. Louis
This month version 3.0 of the Clinical Investigation Data Exploration Repository (CIDER) will be launched. This milestone marks the end of a years-long development process, the availability of a robust new resource for basic and clinical researchers at Washington University in St. Louis (WUSTL), and represents a clear integration of caBIG® principles into new working models.
CIDER is a comprehensive warehouse of in-patient and out-patient clinical information that provides qualified investigators with access to complete medical records—including patient demographic data, procedures, clinical laboratory data, outcomes information, medical images, and large collections of scanned paper reports—for approximately 5 million patients seen at 13 hospitals across the BJC Healthcare system, which includes the academic hospitals associated with WUSTL. CIDER developers have adhered to the principles of standard language and interoperability that have guided caBIG® programs; much of the more recently entered data has been indexed and is searchable, and future accessibility through caGrid is planned.
CIDER was originally conceived as a means to mine the medical record for secondary research purposes and to enhance patient recruitment to active clinical studies, yet the system will also feature a clinical decision support engine that can automatically alert physicians to critical issues regarding patients, based on the system's analysis of clinical laboratory data.
Connecting Multiple Resources
This massive initiative began with an in-house system, developed in 1993, called the Clinical Data Repository that stores all in-patient medical information. CIDER reorganizes data from this resource with two external databases that collect information on out-patient data such as problem lists, medications, and progress notes. Linking these systems—which used different vocabularies and formats—was no small task. Over the course of 18 months, ontology experts sifted through more than 80,000 terms across all the disparate data vocabularies in the various systems to ensure the terms are fully harmonized with the primary industry standard formats for each type of data. As a result, the data contained in CIDER is aimed to be fully compatible with caBIG® principles.
A number of research projects are already lining up to take advantage of this vast storehouse of clinical data, and more than 40 researchers are registered the first training courses. These users will employ a simple Web interface to find patient cohorts and download relevant clinical information that is currently stored in the system. And, in the next two to three years, additional development is planned so that this single portal will allow users to search for all data available in the university's clinical data management system, stored in ClinPortal, biospecimen repositories, stored in caTissue, and microarray data, available through caArray. Additionally, the development team plans to connect CIDER to the WUSTL instance of caGrid and enable users to leverage caB2B as a single query interface for all research systems.
