A critical bottleneck in translational and clinical research is access to large volumes of high-quality clinical data. While structured data exist in medical EHR systems, a large portion of patient information including patient status, treatments, and outcomes is contained in unstructured text fields. Research in Natural Language Processing (NLP) aims to unlock this hidden and often inaccessible information. However, numerous challenges exist in developing and evaluating NLP methods, much of it centered on having “gold-standard” metrics for evaluation, and access to data that may contain personal health information (PHI).
This DREAM Challenge will focus on the development and evaluation of of NLP algorithms that can improve clinical trial matching and recruitment.