Narrativa Raises $1.3 Million To Speed Up Clinical Trials Using AI
Keerthi Vedantam is a bioscience reporter at dot.LA. She cut her teeth covering everything from cloud computing to 5G in San Francisco and Seattle. Before she covered tech, Keerthi reported on tribal lands and congressional policy in Washington, D.C. Connect with her on Twitter, Clubhouse (@keerthivedantam) or Signal at 408-470-0776.
Clinical trials, often the one of the most expensive parts of bringing a drug to market, can last as long as a decade.
Los Angeles-based Narrativa uses AI and natural language processors to write reports. It already boasts clients from Dow Jones and Chevron to Cisco, and is making a play in biotech, promising to speed up the clinical trial process with its software called "Gabrielle."
The startup uses artificial intelligence to automate medical reports on clinical studies, compressing a process that can take weeks into hours. It announced Wednesday that it's raised $1.3 million from undisclosed investors.
"It takes teams and teams of data analysts to comb through the data and actually produce reports that the clinical scientists and then the decision makers actually can use and say, 'OK, is this ready to go to market? Can this be approved by the FDA or do we need to go back and adapt some things?'" Narrativa President Jennifer Bittinger said.
AI and machine learning, the company argues, can help clinical trial conductors process large chunks of data at scale and with more precision and accuracy than human beings.
"We're not trying to replace people and their jobs," said Bittinger. "What we do is we say 'give our software your lower value tasks, the things you don't really want to do so that you can focus on your higher value tasks'."
Narrativa joins a slew of companies ready to tackle the end part of a clinical trial, where contract research organizations tasked with finding patients and conducting clinical trials have to draw up thorough, accurate and regulatory-compliant reports.
Nference, a Massachusetts startup that uses AI to host a suite of products to accelerate drug discovery and clinical trial research, raised $60 million in December. Yseop, a Paris-based startup, also uses natural language processing to help companies write medical reports.
Companies can feed its platform "Gabrielle" data, ask specific questions about what kind of information it wants to glean from the numbers — for example adverse reactions in the first weeks of the trial or how different age groups reacted to the drugs — and get reports in a matter of hours.
But while it speeds up the process, the results may be marginal for clinical trials.
"The report writing is probably one of the least time-consuming processes," said Dr. Eunjoo Pacifici, a professor at the USC School of Pharmacy. "The most time-taking process is really the conducting of the clinical trial itself."
AI's reach into clinical reports is part of a long trend of AI and machine learning in the clinical trial pipeline, which is plagued with inefficiencies and regulatory compliance issues. AI companies have popped up to fix different parts of this.
For example, Deep 6 AI was created to find potential patients to participate in trials by scanning patient records for symptoms that match up with a drug's targeted symptoms. Another, AliveCor, helps monitor and collect data on a patient's heart rhythms through the duration of a clinical trial. A Q1 2020 report by Pitchbook analyst Beendan Burke predicted healthcare AI will be a $1.6 billion industry in 2023.
"This has to still be a human activity," Pacifici said. "And using artificial intelligence is another tool in the toolbox."Keerthi Vedantam is a bioscience reporter at dot.LA. She cut her teeth covering everything from cloud computing to 5G in San Francisco and Seattle. Before she covered tech, Keerthi reported on tribal lands and congressional policy in Washington, D.C. Connect with her on Twitter, Clubhouse (@keerthivedantam) or Signal at 408-470-0776.
Image Source: Salt AI - Aber Whitcomb
Image Source: Snap
Image Source: Rivian
Image Source: ŌURA