Enhancing Clinical Data for CBTN through Automated EHR extraction

Clinical data, especially high-quality longitudinal data, has challenging to obtain and utilize as it currently requires primarily manual extraction and different tools to enable ongoing flows of clinical data to the community. Additionally, raw EHR data are often not of immediate use for researchers as related EHR data elements may be stored disparately, and data may not be readily human readable. Experience has shown that cleaning and initial curation is needed to process data for clinical understandability, especially in multi-modal data contexts where the clinical data are utilized as analysis features. This cleaning and curation process is contextual and this project will focus on clinical data of most utility to CBTN. We will leverage ExtractEHR, a tool originally developed by Drs. Tamara Miller and Richard Aplenc, that extracts, collects, and synthesizes electronic health record (EHR) data for patients identified for inclusion over a specified period of time. The ExtractEHR package includes laboratory and microbiology result data and extensive, non-laboratory data including demographics, vital signs, flowsheet data, medications ordered or administered, radiology results, procedure, and physician notes. A key goal of this project is to provide this data to NCI Childhood Cancer Data Initiative for use by the wider childhood cancer research community.