Ninety-seven percent of providers agree that strong data strategies are essential for value-based care, yet most VBC programs still operate on claims data riddled with coding inconsistencies, missing fields, and stale records. Without a trust layer that scores every record before it enters a payment model, value-based care becomes volume-based guessing with extra steps.
Value-based care has a data problem it refuses to name
Value-based care models tie reimbursement to outcomes. Outcomes depend on data. And most of that data is unscored, unvalidated, and stitched together from sources that no one has reconciled.
A June 2024 House panel heard experts tell CMS it needs to do a better job with value-based care, citing fragmented reporting requirements and inconsistent quality metrics across programs. The underlying issue is structural: VBC payment models assume data integrity that does not exist. When 97% of providers and 96% of payers agree that strong data strategies matter for VBC, but fewer than half have implemented formal data governance, the gap between aspiration and execution is enormous.
This is not a technology deficit. It is a trust deficit.
Why claims data fails value-based care
Claims data was designed for fee-for-service billing. It captures what was billed, not what happened. Diagnosis codes get selected for reimbursement optimization, not clinical accuracy. Procedure timestamps reflect submission dates, not service dates.
When VBC programs use claims as their primary data source, they inherit every distortion baked into the fee-for-service system they claim to replace. Risk adjustment scores inflate. Quality metrics drift. Attribution models assign patients to providers they have not seen in months.
The result: providers game metrics they do not trust, payers build models on data they cannot verify, and patients see no improvement in care coordination.
The eight dimensions VBC data must pass
At SuperTruth, we score every health data record 0 to 100 using the Data Trust Index across eight dimensions. For VBC programs, each dimension maps directly to a payment integrity requirement.
Provenance (25% weight) answers whether the record originated from a verified clinical source or was transcribed, aggregated, or inferred. Consent (20%) confirms whether the patient authorized this specific use of their data. Recency (15%) flags whether the record reflects the patient's current state or a snapshot from 18 months ago.
Quality (10%), Concordance (10%), and Validation (10%) assess whether the data is complete, consistent across sources, and independently verified. Breadth (5%) and Stability (5%) measure coverage across relevant clinical domains and consistency over time.
A record scoring below 40 on the DTI should never enter a risk adjustment calculation. A record scoring above 80 can support shared savings determinations with confidence. Most VBC programs have no way to make this distinction.
Key statistics
SuperTruth's work with imaware standardized 105,000 diagnostic records, reducing processing time from three weeks to two hours. That is a 95% reduction in time spent reconciling data before it could be used.
The same engagement saved over 200 hours per month in manual data validation and identified a patient segment driving 20% of revenue that had been invisible in raw data.
Across the industry, CMS operates more than 40 active alternative payment models, each with distinct reporting and quality measurement requirements. The MACRA framework alone requires physicians to track and submit data across four performance categories, with payment adjustments of up to 9% riding on data accuracy.
Meanwhile, a 2023 industry survey found that data integration challenges are the single largest barrier to VBC adoption, cited by 78% of health system executives.
What AI does to unscored VBC data
Health plans and ACOs are deploying AI models for risk stratification, utilization prediction, and care gap identification. Every one of these models inherits the quality of its training data.
When an AI model trains on VBC data that has never been scored for provenance or recency, it learns patterns from stale records and misattributed encounters. The model then makes confident predictions based on unreliable inputs. Clinicians receive alerts they ignore. Care managers chase patients who already received interventions. The program's total cost of care rises, and no one can explain why.
Scoring data before it enters any AI pipeline is not optional for VBC programs. It is the difference between a model that improves outcomes and one that generates expensive noise.
How Geisinger's experience proves the point
Geisinger Health System, frequently cited as a VBC bellwether, spent years building integrated data infrastructure before its payment models could function. As MedPage Today reported in August 2024, the system's ability to weather COVID and maintain value-based contracts depended on data systems that connected clinical, claims, and social determinants information in near real time.
Most health systems lack Geisinger's decade-long head start. They need a trust layer they can deploy now, not a multi-year data warehouse project.
The fix: score before you pay
Every record that enters a VBC quality calculation, risk adjustment model, or shared savings determination should carry a trust score. Not a completeness flag. Not a duplicate check. A composite score that reflects provenance, consent, recency, concordance, and validation simultaneously.
SuperTruth's DTI Engine does exactly this. It scores records at ingestion, flags records below configurable thresholds, and provides the audit trail that CMS and commercial payers increasingly require.
For community-based organizations participating in Medicaid VBC programs, data trust scoring addresses the specific challenge of integrating social determinants data from non-clinical sources. For large ACOs, it eliminates months of retrospective data cleaning that currently delays shared savings reconciliation.
The cost of waiting
VBC programs that defer data integrity investments pay twice. They pay once in inaccurate quality scores that reduce reimbursement. They pay again when retrospective audits claw back shared savings based on records that cannot withstand scrutiny.
The long-term care sector, as a January 2025 MedPage Today opinion piece argued, faces the same challenge at an even larger scale: operationalizing VBC requires data infrastructure that most facilities simply do not have.
Scoring data is cheaper than defending it after the fact.
To evaluate how DTI scoring applies to your VBC program's data, contact Louis Simeonidis, SVP of Commercial Operations, at louis@supertruth.ai or (215) 918-4140.
Further reading:
Jason Alan Snyder
Co-founder of SuperTruth and Artists & Robots, and an inventor on the Data Trust Index patents. Twenty-plus years building technology inside Interpublic Group. He writes here nearly every day on data trust, provenance, and what AI should be allowed to act on, and publishes essays on his Substack.
About SuperTruth · LinkedIn · Substack · jasonalansnyder.com
See it in practice
DTI scores the record, not the patient.
8 dimensions. 0 to 100. Travels with every record permanently.