Information needed to identify good candidates for clinical trials is often contained in dictated or typed notes in the patient's electronic medical chart. For example, history of hospitalization for a specific condition (e.g., heart failure) may be an important trial enrollment criterion. This information is typically not available in structured form but may be contained in the text of the report as shown in Figure 1 below.
Figure 1. A sample clinical note.
We provide state of the art information retrieval technology to extract useful information from the text of clinical reports. Our technology identifies key terms for disorders, symptoms, medications and any other data elements relevant to the trial enrollment criteria (e.g., blood pressure, lab values). For example, Figure 2 below shows some of the terms identified in the sample note (Figure 1). The term "orthopnea" occurs in a negated context. We automatically determine that this term is negated and use only non-negated terms to compile a list of candidates that meet the inclusion and do not meet the exclusion criteria.
Figure 2. Sample clinical note automatically processed with key terms highlighted.
