After a career as an incident investigator at risk management firms such as Kroll and FTI Consulting, Aaron Narva worked with large international banking clients at compliance software maker Exiger. He was responsible for monitoring the client's compliance after it made headlines a decade ago in a money laundering scandal.
“During my time at Exigar, we acquired several software businesses, including an AI software tool that helps extract risk from unstructured public documents. We've built a tool to help identify risks,” Narva told TechCrunch.
That work gave him the idea for Conflixis. Hospitals and other large healthcare organizations face similar corruption risks as banks. Pharmaceutical companies and device makers are notoriously cozy with doctors, who are required to disclose conflicts of interest such as junkets, consulting fees and sponsorship of research grants.
Many studies have shown that people who get along too well are more likely to prescribe those drugs and devices, regardless of whether they produce better outcomes for the patient. The risks are so great that the government runs a database called OpenPaymentsData.com, where anyone can view conflict of interest disclosures.
But publicizing these conflicts doesn't stop the problems and exposes hospitals to legal risks. Numerous laws prohibit such actions by physicians, from the Stark Law to the Anti-Kickback Statute (AKS).
At the same time, commercial interests need to collaborate with doctors, or medical professionals, to help research new drugs and build devices. Therefore, not all interactions are prohibited.
Narva envisioned an AI-powered software-as-a-service for hospitals and large healthcare organizations that would identify real-world situations that put the hospital, if not the patient, at risk.
“In a large health system, there can be 200,000 relationships between physicians and vendors and suppliers,” Narva says. “Out of those 200,000 relationships, which one of those six or so risks is impacting you?”
Risks range from legal violations to unfavorable medical outcomes. The federal government also provides a database that provides public information on the quality of hospital care.
Narva called a friend he'd known since eighth grade, Joseph Bergen, then director of engineering at Buzzfeed, to get his opinion on the idea. Bergen liked it so much that he quit his job and became a co-founder.
Conflixis works by ingesting data from OpenPaymentsData.com, hospital procurement data, claims data, patient outcome records, conflict of interest forms, and other sources. Analyze all discrepancies to identify issues for the hospital to investigate.
“Okay, we looked at all 5,000 or 10,000 relationships. [and] Here are seven things you really need to look at,” Narva explains by way of example. “It’s like boiling the ocean and there are seven people here.”
Conflixis can also go a step further and predict hospital spending and suggest ways to reduce spending. For example, is a hospital purchasing more expensive equipment rather than less expensive equipment based on recommendations from physicians who have relationships with that vendor?
“We can empower hospitals to significantly reduce regulatory risk, increase trust and transparency with patients, as well as make better operational decisions about what to purchase. ” he says.
The company was founded in 2023 and already has a small number of customers with just under $5 million in revenue, Narva said. The company just announced a $4.2 million seed round co-led by Lerer Hippeau (a fund founded by former BuzzFeed chairman Kenneth Lerer) and Origin Ventures, with additional funding from Mark VC, Springtime Ventures, and pre-seed investor Cretiv Capital. Participating.
Conflixis joins a field crowded with healthcare industry compliance software companies like Compliatric and Symplr, some of which are more focused on protecting patient data than on corruption and procurement.
Narva says what sets Conflixis apart is the way it combines employees' careers in research work with LLMs. He said they modified an off-the-shelf model to find patterns in the data based on “our background in transaction surveillance and corruption in big data investigations.”