The problem starts before the algorithm ever runs. Incomplete patient histories. Consent forms that vanish somewhere in the data pipeline. Diagnostic results that land on a researcher's desk three weeks after they needed them. For developers building healthcare AI, these aren't outlier scenarios—they're Tuesday.
SuperTruth, a Philadelphia-based startup that's been operating mostly under the radar, thinks the industry has been attacking the wrong part of the problem. Instead of building better models, the company argues, developers should be scoring the data before it ever reaches training systems. On April 27, the startup published its Data Trust Index methodology—an open framework that evaluates health records across eight dimensions of reliability—and validated it on 105,000 diagnostic records. The claim: a 95% reduction in data preparation time, collapsing what once took three weeks into a two-hour sprint.
Bold numbers. But in healthcare AI, where hype tends to run a few steps ahead of reality, the details matter more than the pitch deck.
Eight Scores, One Record
The DTI methodology assigns each health record a 0-to-100 score by evaluating provenance, consent, recency, quality, concordance, validation, breadth, and stability. SuperTruth has positioned the system as a patented standard and released the methodology on Zenodo, the open-access repository, for industry review.
Jason Alan Snyder, SuperTruth's co-founder and former Global Chief AI Officer at Momentum Worldwide, has been emphatic about the company's zero-copy architecture: data stays where it lives. The scoring engine evaluates records in place, never centralizing them into a single repository. For healthcare organizations navigating HIPAA compliance and fragmented consent frameworks, that's not a technical nicety. It's the difference between feasible and impossible.
"We're not moving data around," Snyder told industry audiences repeatedly over the past year. "We're scoring it where it sits."
Whether that architecture holds up at true enterprise scale—think multi-hospital systems with tens of millions of records—remains an open question.
The imaware Test Case
The clearest evidence of how the system performs comes from imaware, the at-home testing company SuperTruth acquired in December 2024. SuperTruth says it ran the DTI against 105,000 oncology diagnostic records, flagging sub-threshold data before training began and eliminating more than 200 hours of monthly manual processing. Analysis time reportedly dropped from three weeks to two hours.
Those figures are vendor-reported, not independently verified—a caveat that matters in an industry where pilot successes don't always translate into production deployments. Still, the imaware validation offers something concrete in a market crowded with white papers and proof-of-concept demos that never leave the lab.
Bobby Hill, SuperTruth's co-founder and CEO, has framed the DTI as infrastructure for federated research, pharma real-world evidence studies, and government use cases—environments where data quality can't be assumed and audit trails aren't optional. The company's website targets AI developers, research networks, health plans, and VA systems, all built on the premise that scoring must happen upstream, not after models start misbehaving.
A Terminology Collision
SuperTruth published the DTI framework openly, inviting scrutiny and, perhaps, adoption. But the name itself isn't exactly exclusive intellectual property. Ataccama, a data management platform, rolled out its own Data Trust Index feature on June 2, exposing governance signals for AI tools running on Snowflake. Earlier references to "Data Trust Index" pop up in luxury brand research from 2021. The terminology overlap creates confusion in a market where trust and provenance are already slippery concepts.
That said, SuperTruth's insistence on scoring consent and provenance—not just data quality—does carve out distinct ground. The company operates from 24 S. 24th St. in Philadelphia, lists somewhere between 11 and 50 employees on LinkedIn (a wide range that suggests either rapid hiring or LinkedIn's imprecise categorization), and brought on Dustin Raney as Chief Strategy Officer in November 2025 to lead partnerships.
What Comes Next—and What's Missing
SuperTruth emerged from stealth with the imaware acquisition, a deal Axios valued at $100 million for the combined entity. The company has publicly committed to never profiting from children's data, launching the Sean's Friends donor-advised fund in May to formalize that stance—a rare ethical line in an industry where data monetization is standard operating procedure.
What's less clear: who's funding all this. Dealroom lists "no known external funding" as of its last crawl. No Series A. No institutional investors disclosed in public filings or press releases. The company hasn't clarified whether the imaware acquisition brought capital structure beyond the deal itself, or if this is bootstrapped ambition on an ambitious scale.
For now, SuperTruth's bet is straightforward: healthcare AI developers will adopt an open scoring standard if it demonstrably cuts prep time and keeps regulators off their backs. Whether the DTI becomes industry infrastructure—something like FICO scores for health data—or remains a proprietary differentiator depends on what happens outside SuperTruth's own walls. Independent audits. Implementations at organizations with no acquisition ties to the company. Validation from researchers who didn't build the system.
The methodology is open. The proof, as always, will be in how others use it.
