Health AI Has Been Trained on Unscored Data. SuperTruth Just Published the Standard That Fixes It.

The Data Trust Index™ scores every health record across eight dimensions before it reaches a model, a claim, or a clinical decision. Validated on 105,000 records with imaware. Now available on Zenodo.

PHILADELPHIA, April 27, 2026 /PRNewswire/ -- A denied insurance claim built on a six-year-old lab value. A failed clinical trial whose enrollment cohort was poisoned by duplicate records. A diagnostic AI that surfaced a finding traced back to data no one could vouch for. These are not edge cases. They are the operating conditions of an industry that has scaled artificial intelligence faster than it has scaled trust.

SuperTruth Inc. today published the first formal framework designed to fix that. The Data Trust Index™ (DTI™) scores health data records across eight dimensions, including provenance, consent, recency, and concordance, and produces a single composite score from 0 to 100. The full methodology is available on Zenodo with a permanent DOI: 10.5281/zenodo.19601616.

Most data quality tools report on data after it has been copied into a warehouse. The DTI™ scores at the source, in SQL, and the score travels with the record permanently. Quality tells you what is in a record. Trust tells you whether you should act on it.

"Every model deployed in healthcare today inherits the trust posture of its weakest input," said Jason Alan Snyder, Co-Founder and Chief AI Officer of SuperTruth. "The DTI is the first attempt to make that posture visible, scored, and auditable at the level of the individual record. We are publishing it openly because the industry will not converge on a standard it cannot inspect."

The framework is already extending. SuperTruth has filed a provisional patent covering a ninth DTI dimension, the Behavioral Integrity Index, which scores the conduct of AI agents operating against trusted data. The DTI is designed as a living standard rather than a fixed schema.

Three places where trust is the deciding factor:

Healthcare AI and health systems. The FDA's AI/ML SaMD action plan and HHS data-provenance rules both require developers to document the characteristics of training data. The DTI™ provides the per-record signal that those frameworks demand.

Pharmaceutical and clinical research. Failed or delayed trial enrollment costs the industry an estimated $2.4 billion annually. The DTI™ establishes a minimum trust floor, Gold (80 or above), so only defensible records enter cohort matching.

Insurance and risk underwriting. The DTI™ enables a new class of actuarial input: scored population health data with documented provenance and consent. Underwriting is built on evidence rather than assumptions.

Validated at scale. A deployment with imaware, an oncology diagnostics company, standardized 105,000 diagnostic records using the DTI™ methodology. Analysis time was reduced by 95 percent, eliminating more than 200 hours of monthly manual processing.

"The lab industry has never had a trust standard. DTI created one."
— Brodie Flanders, CEO, imaware

A live implementation of the pipeline is accessible at supertruth.ai/dti. Three synthetic patient records can be scored in real time without a login.

About SuperTruth

SuperTruth is the trust layer underneath health data and AI. The company's platform scores every health record across eight dimensions before it reaches a model, a claim, or a clinical event. imaware is a launch partner. The DTI™ and Data Trust Index are trademarks of SuperTruth Inc., protected by a multi-patent portfolio covering the AI Orchestration Engine, the DTI scoring methodology, and the Behavioral Integrity Index.

Media Contact:

Rheanna Crescenzo
Director of Marketing, SuperTruth Inc.
(215) 918-4140
[email protected]

SOURCE SuperTruth, Inc.