Glossary
What is the Data Trust Index?
Updated September 14, 2026
The Data Trust Index (DTI) is an eight-dimensional scoring framework, developed by SuperTruth Inc., that scores any record on a 0 to 100 scale; the published weights were calibrated in health. It is the authorization step: per record, in the moment, the score that decides whether data can be trusted for a given use. The score travels permanently with every record, so any model, clinician, or auditor downstream can see how much the record deserves to be trusted.
The DTI methodology is published and citable: "The Data Trust Index: A Multidimensional Framework for Evaluating Health Data Integrity in AI Systems," SuperTruth Inc., 16 April 2026, DOI 10.5281/zenodo.19601616.
DTI scores the record, not the patient. Provenance, consent, recency, and concordance are properties of a data artifact, not attributes of a person. Nothing in the score judges, rates, or follows a human being; it tells the person acting on a record how much the record itself deserves to be trusted.
The eight dimensions and their default weights are Provenance (25%), Consent (20%), Recency (15%), Quality (10%), Concordance (10%), Validation (10%), Breadth (5%), and Stability (5%). Institutions can configure the weights for a given study or compliance context; the defaults are the patented weights.
Scores map to four trust tiers (Bronze, Silver, Gold, Platinum) plus a fail band below 55: Bronze (55 to 69) is exploratory, Silver (70 to 79) is operational-grade, Gold (80 to 89) is clinical-grade, and Platinum (90 and above) is regulatory-grade. A record below 55 is held for remediation.
DTI scoring is context-dependent. The same record may score differently depending on the study, the use case, and the governing compliance body, and every score is replayable: the inputs and weights that produced it are part of the record's audit trail.
The closest analogy is Visa, which decides in the moment whether a transaction can be trusted without ever holding your account. DTI is that authorization decision, made per record and per use.
What the score is not
Three confusions come up often enough to settle first. The score is not a judgment about a person: DTI scores the record, not the patient, and nothing in it rates or follows a human being. It is not a data quality score: quality asks whether a record is complete and correctly formatted, and a record can pass every format check and still have no defensible origin, no consent for the use at hand, and a value that expired two years ago (the distinction in full on /compare). And it is not a judgment about an institution: the same laboratory can produce a Platinum record and a Bronze one on the same day, because the score describes the artifact in front of you and the use you intend for it.
DTI scores the record. DataSpine, run by SuperTruth, supplies the sourced public record of the place beside it, and VIGIL scores the agent that acts on the record after the score is set.
The eight dimensions and their default weights
Every record is scored on eight dimensions. Each has a default weight, and the eight weights sum to 100. The descriptors below are the ones the DTI Engine page uses; the weights are the published defaults from the DTI white paper (2026). Institutions can configure the weights for a study or a compliance context, and the configuration in force is recorded with every score.
Provenance, 25 percent. Source pedigree, device attestation, and chain-of-custody. Where did this data come from, and can you prove it? Measures: CLIA/CAP lab source, device attestation, chain-of-custody documentation.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Consent, 20 percent. Explicitness, scope, duration, and revocation hygiene. Was the person asked, and does that consent still hold? Measures: consent explicitness, scope match, duration, revocation status.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Recency, 15 percent. Time-decay calibrated by modality. A CBC result ages differently than a continuous HRV stream. Measures: age relative to modality-specific decay curves, last-updated timestamp.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Quality, 10 percent. Resolution, completeness, missingness, and noise floor. How clean and complete is the record itself? Measures: field completeness, formatting, noise floor, missingness patterns.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Concordance, 10 percent. Corroboration across independent sources. Does the lab result agree with the wearable? Does the EHR match the claim? Measures: cross-source agreement (lab, wearable, EHR), temporal consistency.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Validation, 10 percent. Clinical linkage to outcomes and peer-reviewed evidence. Is this data type established as clinically meaningful? Measures: outcome linkage, peer-review status, clinical evidence grade.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Breadth, 5 percent. Dimensional richness across biomarkers, biometrics, and contextual signals. A richer record is a more trustworthy one. Measures: biomarker count, biometric coverage, contextual signal richness.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Stability, 5 percent. Variance control and test-retest reliability. Does this record produce consistent measurements over time? Measures: variance across repeated measurements, test-retest reliability coefficient.
Weight: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor and measures: supertruth.ai/products/dti-engine.
Provenance and consent together carry 45 percent of the default weight (DTI white paper, 2026). That is deliberate: a record with an unprovable origin or an unverified consent can be complete, current and internally consistent and still be unusable for the purpose at hand.
How a score is computed
Scoring is deterministic. Each dimension is computed against the record's provenance graph and its own inputs, weighted, and combined into one number from 0 to 100. There is no model inference in scoring, so the same inputs under the same weights always produce the same score. The score is computed at ingestion and sealed with the record, and the weights in force can be configured per engagement. Because the inputs and the weights are part of the record's audit trail, every score is replayable: an auditor can recompute it and get the same number. (Technical architecture, supertruth.ai/products/dti-engine.)
Scoring does not stop at intake. DTI runs at intake, at query time and continuously in the background, so a source that degrades over time degrades in score, and a consent revocation re-scores the record everywhere it is used. A score below the threshold opens a structured review rather than passing silently.
The four tiers and the fail band
Scores map to four trust tiers (Bronze, Silver, Gold, Platinum) plus a fail band below 55. The ranges are the white paper's; the fitness descriptors are the DTI Engine page's own.
Platinum, 90 and above. Regulatory-grade. Meets the highest standards across all eight dimensions. Suitable for regulatory submission and clinical decision support.
Range: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor: supertruth.ai/products/dti-engine.
Gold, 80 to 89. Clinical-grade. Strong across all dimensions. Suitable for clinical research, patient-facing products, and institutional AI models.
Range: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor: supertruth.ai/products/dti-engine.
Silver, 70 to 79. Operational-grade. Fit for operational analytics and population health. Flag dimensional gaps before clinical use.
Range: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor: supertruth.ai/products/dti-engine.
Bronze, 55 to 69. Exploratory. Useful for discovery and hypothesis generation. Remediation required before operational or clinical deployment.
Range: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor: supertruth.ai/products/dti-engine.
Below 55: held. Does not meet the minimum trust threshold. The record is quarantined pending remediation or rejection and does not enter a tiered use.
Range: DTI white paper, DOI 10.5281/zenodo.19601616, 2026. Descriptor: supertruth.ai/products/dti-engine.
A floor is a use decision, not a property of the record. A research cohort may admit Silver; a regulatory submission may admit only Platinum. The tier tells the person or the model acting on the record which uses it can defend.
Three records, three scores
The DTI Engine page runs three synthetic records through the pipeline so the score can be watched moving. None describes a real person.
- Maria Gonzalez, 90, Platinum. Seven source systems, five verified provider identifiers, an active consent with full scope, 47 coded medications. The record is cleared for AI-assisted clinical use; three care gaps and two drug-interaction flags travel with it.
- Sarah Chen, 85, Gold. Two sources and two verified providers, a narrow consent, and a conflict: one clinician records Metformin at 500 mg twice daily, another at 1,000 mg twice daily. The score holds Gold; the conflict flag sends the record to a human before clinical use.
- James Wilson, 65, Bronze. Partial documentation from two channels, no primary source connection, no verified provider, no formal authorization on file. Six flags. The record is marked incomplete and cannot drive an automated clinical decision.
Run the pipeline yourself on the DTI Engine page, or assemble a record source by source in the Record Remix and watch the score recompute.
Where the weights came from
The methodology is published and citable: "The Data Trust Index: a scoring framework for health data trust," SuperTruth Inc., 2026, DOI 10.5281/zenodo.19601616. The paper sets out the eight dimensions, the default weights and the tiers. The weights were calibrated in health, where SuperTruth started because a wrong record there costs the most. The production use documented on this site is imaware, its diagnostics company: SuperTruth standardized 105,000 diagnostic records and cut analysis time from three weeks to two hours (case study, published April 2026).
DTI scores any record; the published weights were calibrated in health.
DTI beyond health
The dimensions are properties of any record: origin, permission, age, completeness, agreement, validation, coverage and stability apply to a payment record or a media log as much as to a lab result. What changes by sector is the weighting. As the DTI Engine page puts it, a healthcare AI model weighs consent and provenance differently than a financial compliance engine weighing accuracy and regulatory alignment. The worked examples on this site are health records; this section states the scope and does not add one.
DTI versus a data quality score
A data quality score answers: what does this record contain? Is it complete, correctly formatted, and does it pass validation rules? A data integrity check answers: has this record been tampered with, and is the chain of custody intact? DTI answers a third question: should I act on this record? It asks where the record came from, whether consent was obtained for this use, whether it is current, whether it agrees with other sources and whether the data type is validated against outcomes.
The comparison page sets DTI beside the data quality tools teams already run, Great Expectations and Monte Carlo among them, which check completeness, format and consistency: three of the eight dimensions at best. Those companies are named so you can compare; they are not affiliated with SuperTruth and have not endorsed it, and their names are trademarks of their owners.
How to cite the Data Trust Index
Cite the DOI rather than this website for the methodology; the paper is versioned and permanent, the website is not. The citation line: "The Data Trust Index: a scoring framework for health data trust," SuperTruth Inc., 2026, DOI 10.5281/zenodo.19601616. Every figure on this page has an anchor (for example #dimension-provenance or #tier-gold) so a citation can point at one fact, and the dated facts about the company and the platform live on /cite.