How It Works

Our industry-leading precision research software uses artificial intelligence (AI) concepts such as natural language processing (NLP) and machine learning (ML) on both structured and unstructured data across multiple systems to precision match patients to clinical research trials.

Structured data includes coded data in medical records such as ICD-10 and LOINC codes as well other fields such as age or gender that follow a standardized format. It’s fairly easy to extract information from these fields, and most systems stop here, but structured data only tell part of the patient’s story. 

There is a wealth of patient information in unstructured data sources — free-form clinical notes, lab results, imaging findings, older forms of structured fields, and more. This information is crucial to accurately evaluate a patient’s eligibility against a trial’s inclusion/exclusion criteria.  

And we don’t stop there. Our technology maps all the data to ICD-10 codes and 120+ ontologies as “concepts” in the backend and uses these maps to create dynamic, holistic patient views — what we call patient graphs. These graphs provide broader clinical context and truly allow patient characteristics to be precision matched to trial criteria, resulting in fewer false positives and less wasted time.