Oracle, John Snow helping FDA turn unstructured clinical notes into population health insights

The FDA-led Sentinel Initiative is tapping Oracle’s life-sciences division, Cerner Enviza, to spearhead development of AI tools for rendering EHR-bound clinical notes as usable pointers for clinicians and researchers.

Joining Cerner Enviza in the project, planned for a two-year run, is John Snow Labs, which will lend special expertise in natural language processing (NLP).

3 arms, 1 mission. The Sentinel Initiative has as its main mission improving how FDA evaluates the safety and efficacy of drugs and medical devices. The work is divided into three branchlets—Community Building and Outreach Center, Sentinel Operations Center and Sentinel Innovation Center. It’s the latter that’s running the project at hand.

The project’s primary thrust is using AI and NLP to better understand ways in which widely used medicines affect large populations. Up first is the asthma drug montelukast (popular name brand: Singulair), which is suspected of causing deleterious mental-health side effects.

The FDA selected Cerner Enviza thanks to its decades-long experience in research spanning commercial, clinical and regulatory realms, according to an announcement Oracle posted April 10. The relevant skills base includes working with “a broad range of Oracle provider networks to help accelerate the discovery, development and deployment of health insights and therapies.”

For its part, John Snow Labs is well regarded for its advances in AI and NLP, including healthcare-specific offerings, the announcement points out.

Project leaders hope advanced AI and NLP will prove a fast, scalable alternative to manual means of analyzing clinician notes embedded in EHRs. The traditional methods are understandably slow and questionably comprehensive. However, poring over aggregating such notes has been a must for capturing real-world patient experiences at both the individual and population levels.  

‘Key strategic’ priority. “Development and evaluation of tools that can enhance our ability to utilize unstructured EHR data is a key strategic priority for the Sentinel Innovation Center,” says Rishi Desai, PhD, assistant professor of medicine at Harvard Medical School and operations chief of the Sentinel Innovation Center. “We look forward to this new relationship and exciting initiative led by Cerner Enviza.”

Other names of note in the project, called the “Multi-source Observational Safety Study for Advanced Information Classification Using NLP (MOSAIC-NLP),” include Children’s Hospital of Orange County in California, National Jewish Health and Kaiser Permanente Washington Health Research Institute.

Oracle’s full announcement is here.