A better way to measure healthcare costs in closed claims data
This is the second post in our four-part taXonomy X blog series (part 1 here), and here we’ll discuss one of the most common pain points in closed claims analytics: costs. Specifically, we’ll discuss why traditional allowed and billed amounts can be incomplete or inconsistent, and how taXonomy’s Market Costs offering addresses these limitations by offering a more consistent, validated, and study-ready cost signal to support high-quality real-world evidence (RWE) generation.
Accurately quantifying the costs of services and therapies has become an increasingly critical endpoint in RWE studies. However, in closed claims, “costs” are often oversimplified especially when billed or payer allowed amounts are used as a proxy and assumed to be complete and broadly representative.
In real-world data (RWD), payer-allowed and billed amounts may be missing, obfuscated or unevenly distributed, and closed claims datasets can be concentrated to specific insurance plans or payer types, both of which can limit generalizability in cost-focused studies. Even when allowed or billed amounts are present, the cost details and what they represent are often inconsistent across suppliers, making it challenging to establish normalization and ensure accurate comparability. Further, some sources may offer more granular cost metrics (e.g., separating payer-paid and patient-paid components such as deductibles and copays), while others may offer only a single rolled-up amount, further exacerbating the challenges of accurately applying these values to cost-based use cases.
Heterogeneity in supplier cost fields is a well-established limitation. When present in closed claims, it extends beyond simple missingness. This can bias conclusions, constrain study feasibility and reproducibility, and complicate comparisons across cohorts and markets.
These known challenges inspired HealthVerity to develop taXonomy’s Market Costs offering: a novel, proprietary methodology that generates consistent, accurate, and standardized cost values, addressing gaps in incomplete or heterogeneous supplier cost fields.
How taXonomy market costs work
Medical Market Costs (835 remittance-based):
For medical services, HealthVerity Medical Market Costs are built from raw remittance data captured on Electronic Remittance Advice (ERA) 835 forms. The data on these forms document the actual payment exchange among patients, payers, and providers. Because the 835 remittance form includes very detailed payment details and adjustments, it enables the creation of standardized allowed amounts, payer paid, and patient paid benchmarks at the individual procedure code level.
HealthVerity aggregates these transactional values into unit-cost benchmarks that are categorized at a highly granular level via a unique combination of procedure code, procedure modifier(s), revenue code(s), units billed, payer type, U.S. Census region, and calendar year. The distribution of these values is provided for every unique combination of these fields, granting researchers the ability to isolate factors that most directly influence reimbursement within a given region and/or across payer types, for example. National level benchmarks are also readily available when this level of granularity is not required.
Pharmacy Market Costs (open pharmacy claims-based):
For prescriptions, taXonomy Pharmacy Market Costs are built from open pharmacy claims, which reflect the real-world adjudication and payment dynamics occurring between payers and pharmacies for dispensed prescriptions.
Pharmacy benchmarking is built at the “base product key” level, which captures a product configuration inclusive of name, form, strength, quantity dispensed, and generic indicator. Similarly, a base product key’s cost is captured across every unique combination of quantity dispensed, payer type, U.S. Census region, and calendar year. These benchmarks are then used to populate market cost fields on prescription records including allowed amount as well as plan-vs-patient paid components.
Putting taXonomy market costs to the test
In our recently-accepted ISPOR 2026 abstract, we evaluated whether integrating 835 remittance data and open pharmacy claims improves the generalizability of cost estimates.
Specifically, payer cost was defined as the allowed amount captured in closed claims, while market cost applied closed-claims utilization to a median unit cost benchmark stratified by service type, payer, geography, and time period.
A key question is whether market-based estimates reproduce the same spending signal observed using payer costs when available. To assess this, we evaluated the concordance between market and payer cost estimates by calculating total annual direct medical spending among taXonomy patients with continuous commercial medical and pharmacy enrollment throughout calendar year 2024, restricted to individuals with both cost measures available.Using statistical techniques commonly used in clinical agreement studies, we evaluated alignment between the two measures using linear regression and the Equivariant Passing–Bablok method to estimate translation between measures.
What we found
The validation exercise used a random sample of 10,409,060 continuously-enrolled commercially insured individuals with both sets of costs available. The results showed strong alignment overall. Linear regression produced a slope of 0.996 (95% CI: 0.982–1.01), indicating accurate prediction of payer-derived spending from market-derived spending. Passing–Bablok produced a slope of 0.956 (0.944–0.968), providing a translation from market-derived to payer-derived spending, the inverse of which can be used to translate from payer-derived to market-derived spending.
The overall conclusion is clear. Market-derived spending demonstrates strong agreement with payer-derived spending at the person-year level, supporting market cost measures as an appropriate proxy for payer costs in RWE studies.
Why this matters
This expands what’s feasible for cost-based analyses using closed claims data. Rather than restricting analyses only to segments where payer allowed amounts are fully populated, or dropping records when allowed is missing or obfuscated, market costs provide a standardized alternative that tracks closely with payer-derived spending.
In turn, the taXonomy Market Cost methodology supports high quality RWE generation for life sciences organizations and the researchers working on their behalf by making cost measurements more comprehensive and consistent, resulting in improved study design and findings that are more reliable and generalizable across markets, greater reproducibility, and enhanced robustness to variability in payer-cost completeness.
