Insurers, researchers, and consumers are struggling with the price transparency files required under the No Surprises Act. This post reviews how a public network ID could make these files smaller and more accurate. We’ve previously written about the many benefits of a public identifier for networks (here, and here).
Background
Under the No Surprises Act, insurers are required to post machine readable files online in a specified format. These files often called the Transparency in Coverage or TIC files. The technical specifications for format TIC files are found in this Github repository.
Medical service prices are established within contractual agreements between providers and insurers; the prices themselves are often contained in a fee schedule annexed to the contract. These provider contracts are bundled into a provider network, and in the large majority of cases the plan simply references the coverage and rates for all providers set within the network. Put another way, prices rarely vary according to which health plan or employer uses this contract for coverage of this provider under the plan.
Problems with Current Structure of Machine-Readable Files
Currently the regulations at 45 CFR 147.212(b)(1)(iii)(A) call for the use of HIOS ID of the plan (the plan identification scheme used for ACA plans) or the EIN of the employer for plans that do not have a HIOS ID. This way of identifying to which plans the stated prices apply is necessary because there are no nationwide identifiers for health plans, and has many negative effects:
- Provider networks do not have any specific identifiers akin to the HIOS ID assigned to each ACA insurance plan. Numerous plans offered by an insurer use the same network of providers; the negotiated rates for medical services are most often set at the level of the network and therefore do not vary between these plans.
- Lacking a network identifier, TIC files repeat the same contractually established price over and over – the price is listed separately in the file for each plan that uses the same provider-insurer contract (i.e. the same network). This repetition of that same data makes the TIC files very large without adding any new information.
- Repetition of the same data multiple times reduces data quality. Prices that are repeated but don’t agree with one another could be due to an error or a price arrangement that is specific to one employer or plan. There is current no way to distinguish an error from a truly distinct plan-specific price.
- Repeated data makes it easy for missing data to be overlooked. Prices that are missing are likely due to an error, but as above this cannot be distinguished from a plan-specific price.
- Because of the lack of clear identifiers (plan and network), the applicability of the prices for a given consumer is muddy.
This duplication of rates is not needed for correct representation of prices within each plan, rather it is an artifact arising from the lack of a network identifier in the defined format for reporting the price data. A file format which more closely represented the original network and its source contracts would be much smaller and have far fewer errors.
Repeated data also makes the files difficult to work with on a strictly technical level. Because the repeated data and the fact that they are generally posted in file format called JSON, these files are huge: large insurers publish files that are terabytes in size (for comparison a terabyte is the size of the entire hard drive of a moderately priced laptop), and even the smaller files. These large JSON files cannot be opened and manipulated like an Excel spreadsheet, and for large files even statistical programs like R or Python run into problems with memory. Companies and organizations that access and use these files often have to turn to sophisticated technologies like Spark and use cloud service providers to unpack and transform these files into database tables that are more usable.
In addition, it is difficult for consumers to be certain that a particular price stated in the file is in fact the one that is relevant for them. For the majority of Americans who get their insurance through their employer, the only identifier that can relates data in the TIC files to a consumer’s own plan is their employer’s Employer Identification Number (EIN) assigned by the IRS.
To find a price that is relevant for their own health plan, an employee would first have to look up their EIN. In doing so they might find multiple EINs if their company has a complicated ownership structure, and ascertaining which EIN actually goes with their particular health plan is almost impossible. While there are lists of which insurance company each employer uses to offer their employees coverage, these two are vague – the insurer may use a different name or there may be distinct plan arrangements that make it unclear which EIN applies.
Benefits of Using a Network ID
A public network ID would make the files far more useful for consumers, researchers and many others. If a consumer’s insurance card listed their network ID, it would be far easier for them to look up the network ID and go straight to the price information for that network.
Using a Network ID would allow the prices stipulated in a single provider-insurer contract to be listed only once (or at worst, far fewer times than it is listed now). This would also address the duplication and data quality problems described above.
We suggest that CMS, regulators, and others involved in producing and using TIC files adopt a public network ID.
