How to evaluate cost data for HEOR and real-world evidence studies

Most teams planning a total cost of care or burden of illness study assume the cost data will be there when they need it. Then they build their cohort, apply a few attrition steps and find the allowed amount field is empty for a third of their patients. Or it’s populated, but defined differently by every supplier. Or populated as a single number when the question is what the patient paid.

Before a researcher can answer the scientific question, they have to answer a data question: Do I have enough comprehensive, trustworthy cost data to proceed?

It's a question our data experts hear from clients constantly, and it was the focus of a recent HealthVerity webinar. Below are the key takeaways, from what to check before you trust a cost field to how market costs fill the gaps.

Five questions to ask before using a cost field in claims data

There are no strict industry definitions for cost fields. Two suppliers can both label a column "allowed amount" and mean different things. Before using any cost data, you should ask:

1. What does the field actually measure? Charges (list price), allowed amounts, net payments after patient responsibility or underlying production costs are all different numbers. Sources like HCUP report cost-to-charge ratios that reflect what it costs a provider to deliver care, not what they were paid.
2. What is the fill rate, and who is missing? Check the share of enrollees with claims, including non-users, then check how well each field is populated among those who do have claims.
3. Does cost per capita pass a face validity check? Sum costs per person across original claims and adjustments, then compare against a benchmark like the Medical Expenditure Panel Survey (MEPS). Are you even in the right order of magnitude?
4. Has imputation been applied? Imputation is a well-established method, but you need to know where it was used, what method was applied and whether it's documented.
5. Are amounts comparable across sources? Coordination of benefits, adjustment logic, payer type, geography and inflation adjustment all affect whether values can be combined.

It’s also worth reviewing our cost methodology whitepaper, which details how HealthVerity market costs show alignment with established U.S. spending benchmarks and display strong agreement with payer-provided allowed amounts.

Common problems with cost data in claims-based studies

Three problems come up time and time again:

Fill rates shrink as cohorts narrow. A cost field that looks usable across the full dataset can become unusable once you filter to a specific therapy area, subpopulation or line of therapy. Every attrition step can cut the share of patients with populated cost values, which directly affects study feasibility.

Populated vs. consistent. In a deduplicated closed claims dataset built from multiple suppliers, some provide only allowed amounts, some break out payer and patient responsibility, some report actual payments and some report imputed values with methods that aren't disclosed. Treat them as interchangeable and you put comparability, interpretation and validity at risk.

A total isn't always the right number. HEOR, RWE and market access teams increasingly need to know who bore the cost. A single rolled-up allowed amount is directionally useful, but it can't answer questions about affordability, patient out-of-pocket burden or cost sharing.

How HealthVerity market costs work

To address these gaps, HealthVerity developed a market cost methodology that combines the strengths of two data types:

• Closed claims provide enrollment-controlled utilization: the procedure code, modifiers, revenue code and units for medical claims, and quantity dispensed by product form and strength for pharmacy claims.
• Open claims provide market unit costs: allowed and paid amounts from 835 remittances on the medical side and paid pharmacy claims on the pharmacy side, which also make it possible to derive out-of-pocket amounts.

Market unit costs are stratified by payer type, census region, calendar year and claim-level attributes, then applied to the actual utilization observed in closed claims. For rare procedures without enough sample at full granularity, the methodology steps back hierarchically: first holding payer type constant, then moving to an overall benchmark.

The result is a standardized cost measure with roughly 97% coverage, compared with closed claims sources whose allowed amount fill rates range anywhere from 0% to 100% depending on the supplier. Market cost fields include allowed, paid and out-of-pocket amounts, so teams get the payer and patient breakout even where the supplier only reported a total.

Validating market costs against supplier-reported claims costs

The team first ran a soft validation, comparing median supplier-reported closed claims costs to median market costs once every relevant stratification criterion was aligned.

The values aren't identical cell by cell, and they aren't expected to be. What matters is that they land in the same range and move in the same direction once the comparison is normalized. The team saw similar alignment for advanced therapies and anti-nausea medications, with alignment tightening further when stratified by indication, such as Crohn's disease.

HealthVerity then tested agreement more formally using methods from clinical agreement studies, including linear regression and Passing-Bablok regression, on 2024 total costs for patients with both supplier and market costs. The findings, presented at ISPOR 2026 in the poster: Do Data Sources Matter? Comparing Market Cost and Payer Cost Approaches to Calculating Medical Spending Measures in Real-World Data, showed strong agreement between market-derived and payer-derived spending at the person-year level.

When to use market costs alongside supplier-reported costs

Market costs aren't meant to replace observed supplier costs. When actual payments are available and sufficient, researchers can prioritize them. When attrition erodes fill rates, or when the study needs a payer versus patient split the supplier didn't provide, market costs fill the gap. The two approaches are designed to work side by side.

Using cost data for market access, brand and patient services teams

Cost questions don't stop at HEOR. Market access teams preparing payer conversations, brand teams evaluating affordability barriers and patient services teams sizing copay and support programs all need to understand what patients actually pay, by payer type and region, for their therapy and its competitors.

That's where market costs become even more relevant following HealthVerity's combination with Symphony Health. Market costs are built on open claims, and the combined ecosystem now spans 100+ linked, privacy-certified sources under one HIPAA certification with deep open claims coverage. Teams no longer have to choose between the enrollment controls researchers need and the market-level cost signal commercial teams rely on. They can understand the true financial experience of their patients across the market, not just the slice one supplier happens to report.

Know where to find market costs

Market costs are available in taXonomy, HealthVerity's closed claims data model. That's intentional: cost studies depend on capturing every prescription and procedure a patient has over the study period, which requires enrollment controls. With 300M+ covered lives, taXonomy is the largest closed claims product available, and taXonomy pairs enrollment-controlled utilization with standardized market costs.