Limited utility of price transparency data for drugs

Abstract

Since 2021, hospitals have been required to report the prices of common, shoppable services, including drugs, as per the Hospital Price Transparency Rule. In this paper, we used Hospital Price Transparency data aggregated by Turquoise Health to investigate the usability of price transparency data to evaluate variation in reimbursement for provider-administered drugs. We extracted records for 30 procedure codes corresponding to provider-administered drugs reported by at least 1,000 National Provider Identifiers (NPIs) and evaluated variability in rates reported for each procedure code. Among 3,321,502 records extracted, 65% had missing NPI, reimbursement rate, or National Drug Code (NDC) information. The remaining 35% of entries reporting NDC information did not necessarily report negotiated rates in the quantity of the NDC. Instead, they contained a combination of prices expressed in at least 3 different quantities: the quantity in which the procedure code is expressed, the unit of the NDC, and the total quantity of drugs administered to the patient. Until providers follow standardized requirements for the reporting of the data, the Hospital Price Transparency data should be used with caution, as the inability to correctly identify the unit in which prices are expressed can lead to incorrect inferences about drug prices.

Plain language summary

Price transparency data allow for comparison of negotiated rates or prices for common health care services and drugs. In this paper, we show that these data in their current form is not useful for comparing drug prices because there are no requirements to provide the unit of measurement. This can lead to incorrect conclusions about price variation. Upcoming requirements might solve this, but there will need to be strong enforcement and standardized reporting.

Implications for managed care pharmacy

Price transparency data hold promise for shedding light on provider-administered drug prices. However, currently, providers are not required to report the unit and type of measurement, limiting the usability of the data to calculate price variation or estimate price markups, for example. Updated requirements from the Centers for Medicare & Medicaid Services in 2025 may solve this issue, but strict enforcement and standardized reporting will be needed for transparency datasets to be useful for analyzing drug spending.

Introduction

Provider-administered drugs are given to patients in conjunction with a provider service and represent an increasing share of drug spending in the United States. For instance, in Medicare, spending on provider-administered drugs, which are covered under Part B, has risen to represent 27% of total Medicare drug spending.¹ Spending on provider-administered drugs is projected to continue to increase given the rising prevalence of biologic medications in the pharmaceutical pipeline.

Recent research aiming to bring transparency to drug reimbursement has focused on drugs dispensed in pharmacies, which are covered by an insurance policy’s pharmacy benefit.² However, little is known about the reimbursement of provider-administered drugs in the commercial market. This is partially related to the lack of data available on commercial reimbursement rates for provider-administered products. The recent availability of data generated by Price Transparency initiatives implemented in recent years generates an opportunity to shed light into this opacity. Data generated by the Hospital Price Transparency (HPT) rule and the Transparency in Coverage rule^3,4 hold promise for bringing much-needed transparency into commercial reimbursement rates for provider-administered drugs. These data can also allow payers to understand what their competition is paying, enabling them to bargain more effectively.

Prior research evaluating the quality of Hospital Transparency data revealed important limitations associated with completeness and usability.^5,6 These investigations, however, have not devoted attention to the usability of drug price data. The evaluation of negotiated rates for provider-administered drug products presents certain complexities compared with data for medical services, particularly as it relates to the quantity of drug in which the negotiated rate is reported. Specifically, negotiated rates for drugs could be expressed in the unit of the procedure code, the unit of the specific drug product selected, or the average amount of the drug administered to patients.

To comprehend the differences between the 3 possible quantities that negotiated rates could potentially be reported in, it is important to understand the way provider-administered products are billed. Provider-administered products are billed using procedure codes. For example, brand-name trastuzumab (Herceptin) is billed using procedure code J9355, which has an average sales price expressed per 10 mg. Trastuzumab is available in different formulations with different quantities: vials of 150 mg and vials of 440 mg. Each of these formulations has a unique identifier, called National Drug Code (NDC) numbers. Providers can opt to use different combinations of each NDC number to provide the total dose to a patient (estimated at 560 mg for a 70-kg patient). Because the pre-2025 regulations did not require providers to specify the reported rate in the quantity defined by the procedure code, reported rates could potentially be expressed per 10 mg (the unit of the procedure code), 150 or 440 mg (the units of the NDC numbers), or 560 mg (the total amount administered to the patient in this case).

ASSESSMENT OF HPT DATA FOR ANALYZING DRUG PRICES

We investigated the usability of HPT data for provider-administered drugs using files obtained from Turquoise Health, a data vendor that has aggregated HPT data (Supplementary Methods ^{(759.5KB, pdf)} , available in online article).⁷ Specifically, we obtained files reporting entries most recently updated in Q1 2024. The sample included 30 procedure codes corresponding to provider-administered drugs reported by at least 1,000 National Provider Identifiers (NPIs), as listed in Supplementary Table 1 ^{(759.5KB, pdf)} . The data included a total of 3,321,502 records for the selected sample of products (Figure 1). After excluding records with missing NPIs or for which payment rates were missing or zero, the sample included 3,302,180 records.

FIGURE 1.

Overview of Claims Selection

Our evaluation of these records unraveled 2 critical limitations:

• 1. Most entries are missing NDC numbers: A total of 64.5% of observations did not report an NDC number (Figure 1). This high proportion of records with missing data are problematic for analysis because, in the absence of additional information on quantity and type of unit, one cannot infer the quantity reported because a procedure code often corresponds to multiple different NDC numbers.
• 2. Reporting of drug prices representing inconsistent quantities of product: We observed that entries reporting NDC numbers did not necessarily report negotiated rates in the quantity of the NDC. Specifically, based on the distribution of negotiated rates, we inferred that reported rates likely represent at least 3 different quantities: the quantity in which the average sales price is expressed, the unit of the NDC, and the total quantity of drugs administered to the patient. We illustrate this ambiguous reporting using NDC number 55513-0924-10 (infliximab 100-mg vial) as an example (Figure 2), which corresponds to 10 units of procedure code J1745. The Medicare payment level for this procedure code was $32.16 per 10 mg in Q1 2024. This would correspond to a Medicare payment of $321.60 per vial of 100 mg, or $2,572.80 per 8 vials needed to treat an 80-kg patient at a dose of 10 mg/kg. The shape of the histogram suggests a mix in terms of the units used by providers in reporting. Specifically, the presence of a large mass of NPIs on the left-hand side of the axis likely represents providers correctly reporting rates per 10 mg (in the quantity of the procedure code). The peak around $300 likely represents NPIs reporting payment rates in the quantity of the NDC. Finally, larger observations probably correspond to providers reporting negotiated rates for the total amount provided to the patient, which changes with the indication and the weight of the patient, thus explaining the wide distribution. A similar distribution is noted for other products in Supplementary Figures 1-3 ^{(759.5KB, pdf)} , with a combination of rates likely expressed in the quantity in which the procedure code is expressed, the unit of the NDC, and the total quantity of drugs administered to the patient.

FIGURE 2.

Distribution of Payment Rates for NDC Number 57894-0030-01

Although one can make informed guesses on the likely unit of reporting based on the amount, it is not possible to fully infer the quantity represented by the reported data, which hampers the usability of the data. Concerningly, the combination of payment rates that represent differing amounts can lead to incorrect inferences of the mark-up on the drug. For instance, if a user of the data assumed that payment rates were consistently reported in the unit of the Healthcare Common Procedure Coding System procedure code (10 mg in our example), they would estimate a 1,000% mark-up on infliximab for providers reporting data per 100 mg and even a more extreme estimate for those reporting rates for the total amount provided.

Discussion

In recognition of these shortcomings, starting in January 2025, the Centers for Medicare & Medicaid Services (CMS) will require hospitals to provide information on drug unit and type of measurement. Specifically, this means that if a standard charge has been established for the drug, hospitals are required to indicate the drug unit and type of measurement as separate data elements. The drug type of measurement should correspond to the standard charge for drugs as either defined by the NDC or National Council for Prescription Drug Program.⁸ Although this is a step in the right direction that could resolve the previously mentioned limitations of the data, a major impediment to data usability could be provider compliance with this new regulation. Compliance is concerning because, at the start of 2024, after 3 years of the Hospital Provider Transparency rule being in effect, 17% of providers still have not made their negotiated rates public.⁹ Thus, by raising awareness on the usability of Hospital Provider Transparency data, this manuscript will hold relevance despite the new CMS requirement, as data quality may not drastically improve immediately in the absence of strong enforcement.

In addition to the enforcement recommendations put forth by Jiang et al,¹⁰ requirements to report drug unit and type of measurement in a uniform and standardized manner should be enforced. Currently, providers are required to report the negotiated rates in a machine-readable file, but in practice the format of these files can vary widely from provider to provider, and this leads to data inconsistencies and reduces the ease of access to this data. Requiring providers to submit the data through the same form and with consistent fields would enhance data quality significantly. CMS currently enforces noncompliance with Price Transparency requirements through monetary penalties. After being issued a warning, hospitals must submit a corrective action plan and become compliant by a given deadline. Failure to do so results in a penalty. CMS should enforce the updated requirements related to drug unit and type of measurement. Standardized reporting and availability of drug units will enable data users to identify the quantity of drug in which the reported rate is expressed, thereby reducing the likelihood of wrong inferences.

In summary, in its current format, Hospital Provider Transparency data do not allow for an accurate assessment of reimbursement rates nego-tiated between providers and payers for provider-administered drugs. The combination of negotiated rates expressed in different quantities across providers can easily result in wrong inferences. We hope that our description of these important limitations of the data is helpful for investigators using Hospital Provider Transparency to evaluate reimbursement of provider-administered drugs. It is important to note that the payer data generated by the Transparency in Coverage rule would benefit from the updated requirements being made to the Hospital Transparency Data in 2025, as well as our recommendations for standardized reporting. Any future improvements to the data will be affected by provider compliance and the method in which the drug unit and type of measurement is reported.