How APIs Can Help with US Healthcare Price Transparency: Part 1


Can APIs help bring new hope to the United States healthcare system? The Postman Open Technologies team recently investigated the state of data around healthcare price transparency initiatives in the US. This first blog post in our series on the topic provides background information and shares our initial findings as to how APIs, technology, and best practices could support this important work.

Hospital price transparency

In January 2021, a federal Hospital Price Transparency rule, issued by the Center for Medicare and Medicaid Services (CMS) under the Affordable Care Act (ACA), took effect, requiring hospitals to publicly post their service prices on the web in both machine-readable and user-friendly formats. Despite a significant pushback by the industry, the rule was upheld by US courts.

Over a year and a half down the road, it has been widely publicized that compliance is very low and the data quality is not up to expectations. In February 2022, for example, the Patient Rights Advocate group published a report with these findings:

  • Out of 1,000 hospitals surveyed, only 14.3% were complying with the transparency rule
  • Out of the 37.9% of hospitals that posted a sufficient number of negotiated rates, over half were not compliant with other criteria of the rule
  • Only 0.5% of hospitals owned by the three largest hospital systems in the country were compliant

More technical insights on the state of this data were recently provided by Alec Stein from DoltHub on his blog, where he shared his experience analyzing 1835 hospital price lists. The process, described in detail in a Python notebook, corroborates low compliance and highlights various data quality issues. A fascinating aspect of this project is also how the data was actually put together through a bounty hunting system. There is actually no easy way to discover the files published by hospitals, and thanks to community efforts, at least a subset of the data is now available in a public SQL database.

Transparency in coverage

In January 2022, a similar ruling came into effect, but this time focusing on group health plans and health insurance issuers. The Transparency in Coverage ruling addresses the public availability of in-network rates and allowed amount for out-of-network providers. It is a three-stage process, with the first focusing on the release of machine-readable files containing the price data. Phases two and three, planned for 2023 and 2024, call for internet-based price comparison tools that allow an individual to receive an estimate of their cost-sharing responsibility for a specific item or service. For phase one, CMS has published JSON and XML schemas as well as file-naming conventions on GitHub, and clarified that Excel files, which were widely used by hospitals, would not qualify as machine-readable. A Docker-based JSON schema validation tool is also available, but such operations can be performed using common JSON tools or in Postman.

CMS allowed a bit of extra time for insurers to prepare and started enforcing this ruling as of July 1, 2022. In light of this deadline, an initial article in the Wall Street Journal reported that some of the large providers started releasing files on that day. According to Maximilian Pany, a graduate student and researcher associated with the Healthcare Markets and Regulation Lab at Harvard Medical School, an early evaluation revealed that some files lacked disclosures and were missing specific pricing information.

This new trove of data is expected to be massive, and it is more complex than the hospital releases, so it will naturally take time to process and analyze. Like the hospital data, there is no easy way to know where these files are being published. This Red Cross web page provides a useful entry point to some providers. Navigation to sites such as United Healthcare or Optum leads to long and sometimes slow-to-load lists of files to download. Most of these pages are neither human nor machine friendly, and there are no APIs on the horizon.

Our thoughts and actions

It is clear from our initial scan that, despite the CMS rulings, the price transparency data remains difficult to discover, access, and use, and suffers from various quality issues. While some of the reasons may be non-technical in nature, we are interested in exploring how APIs, combined with better tools, standards, and practices, can be applied to improve this situation by facilitating compliance or enabling the development of web applications and tools.

We envision this as a collaborative effort and have created a Postman Open Technologies – U.S. CMS Transparency in Coverage public workspace and GitHub repository to be used as a shared environment for research and development purposes. If you are a data user, a data provider, from CMS, or simply interested in this challenge, we encourage you to contribute to the workspace.

At this point, we have seeded the workspace with a copy of the JSON schemas from CMS and started to draft web services specifications that we feel would be beneficial to all. The idea is that some of these APIs would be implemented and deployed by various agencies or communities involved in this space. Prototypes could also be implemented as proofs of concept or for R&D purposes.

Some of the areas of improvement we have identified so far include:

  • Registration and discovery: Both hospital and insurance data files are hard to find. Having the ability to register reporting entities and associated URLs where the information is located would be very useful to support discovery, harvesting, or assessing basic compliance.
  • Data access: Clearly, this is a major need. Having the ability to query these datasets through APIs would go a long way in improving the situation. Upgrading the data to  digital knowledge by wrapping it with metadata would further increase quality and machine actionability. This seems essential to enable the implementation of price comparison tools for phases two and three of coverage transparency.
  • Transformation and publication: A common requirement for both rulings is the availability of this pricing information in user-friendly formats. Data transformation services could help achieve this objective, as well as convert the information across machine-actionable formats.
  • Validation and QA: Data quality is a significant challenge. While this has not been thoroughly assessed for the recently released insurer data, it is well documented for the hospital data. Services to assess quality and validate files would help improve across the board. Tools such as schema-level validation are not enough—second-level validation is necessary for robust testing.
  • Coding: A particular issue that was identified for the hospital data relates to the use of correct codes for the various services and associated prices. Industry-standard classifications, such as Current Procedural Terminology (CPT), Healthcare Common Procedural Coding System (HCPSC), and International Classification of Diseases (ICD), must be used with the data. But numerous issues, such as invalid or malformed codes, have been reported. Services to look up and validate codes would strengthen the validation/QA layer (and would also benefit anyone else using these code lists).

There are other issues that need to be addressed (such as harvesting), but the above is a good and challenging starting point. We hope this initial blog post inspires you to join us in this endeavor. We also still have a lot to learn about this, and so we welcome others’ expertise in this area. Look for Part 2 of this blog series coming soon!

What do you think about this topic? Tell us in a comment below.


Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.