Big Data firm tackles “regulatory grade” real-world evidence (RWE)

AI tools enable Verantos to provide more accurate RWE data for FDA approvals


By focusing on electronic health-record (EHR) data to generate actionable clinical results, Verantos is putting out a marker that gathering the clinical data necessary for drug approvals is going to be different, going forward. The company has now launched a commercial service to gather data from EHRs, process the data in its system that includes natural language processing and machine learning, and produce observational (or other types) of clinical studies of sufficient quality to, for example, justify approval of new indications for drugs.

According to Dr. Dan Riskin, CEO, today’s use of claims data in attempts to get at RWE is so poor as to be unusable; claims data is at best 50% accurate (which can affect study conclusions). There are two significant problems that Verantos seeks to address in using EHR data: the effects of comorbidities on therapy outcomes; and the difficulty in identifying subgroups, within one disease class, that are more amenable to therapy than the group as a whole. “Precision medicine rests, in part, on a more accurate identification of patients,” says Riskin, while noting that recent FDA guidance on RWE shows the agency is getting “smart about data validity.”

Riskin and his company have experience in adopting AI tools to health records; the company he previously founded, Health Fidelity, is successfully serving insurers to quantify risk assessments based on patient populations, using similar AI tools. He adds that Verantos has already performed a number of studies for Top 10 biopharma companies, and is now offering its services on a commercial scale.