Inaccuracies related to accepting new patient flags in Affordable Care Act plans’ provider directories

Whether or not providers in their network are or are not accepting new patients is a central piece of information for patients.¹ Patients generally rely on provider directories to learn about their in-network providers.² Existing studies have primarily assessed contact information, specialty, or network status for accuracy.^1,3 Studies have consistently identified inaccuracies across provider directories.^4-6 These errors can exacerbate health inequities and disparities.⁷ Despite being a requirement both under the Affordable Care Act and the No Suprises Act, the degree to which information related to the acceptance of new patients is listed correctly (or at all) in online provider directories remains unknown.

Data and methods

To fill this gap, we fielded a consumer-centric, large (N = 97 640 providers) secret shopper survey in 6 specialties (cardiology, endocrinology, gastroenterology, primary care, psychiatry, and psychology) in 7 states (Arizona, California, Colorado, Georgia, Ohio, Vermont, and Washington) across Affordable Care Act (ACA) Marketplace plans across all ACA pricing regions proportional to ACA enrollment (see Supplementary Appendix for details).

Callers were assigned a random address, which was used to search for providers in the carriers’ customer-facing online provider directories. Callers collected data on whether or not providers were listed in the carriers’ online provider directory as (1) accepting new patients or (2) not accepting new patients or (3) whether no such information was listed. Following a script, callers then called the listed provider to assess whether this information was correct by specifically asking whether the listed provider was accepting new patients.

Overall, callers conducted provider directory searches for 97 640 providers from May 16, 2022, to December 22, 2023. Callers were able to call and verify the accuracy of the accepting new patient information flag for 20 514 providers, with the remainder either unreachable or verified as having wrong contact, specialty, or inaccurate network status listed in the network directories.

Results

Based on the carriers’ provider directories, of the 20 514 providers callers verified, 17 488 (85.2%) indicated that they were accepting new patients, 949 (4.6%) indicated that they were not accepting new patients, and the remaining 2077 (10.1%) provider had no information listed.

We found that 26.8% (N = 5498) of the 20 514 provider directory listings exhibited errors, with substantial variation along various categories of interest (Table 1). Errors rates were highest for psychologists (44.3%, N = 816/1842) and lowest for gastroenterologists (19.6%, N = 769/3,925, P < 0.001). They were lowest in Vermont (18.1%, N = 55/304) and highest in Colorado (30.7%, N = 416/1,356, P < 0.001). We also found that PPOs (25.2%, 631/2503) fared better than EPOs (28.1%, N = 1151/4,089, P < 0.010). We found no differences based on provider rurality based on the classification used by the Federal Office of Rural Health Policy (FORHP). For carriers, the cumulative error rates ranged from 12.9% (Carrier 21, N = 18/140) to 69.1% (Carrier 17, N = 195/282, P < 0.001) (throughout, we restricted the carrier analyses to carriers with more than 100 observations. This reduced the number of carriers from 38 to 27).

The most common sources of errors where instances where provider directory listings incorrectly indicated that providers were accepting new patients when they were not (14.5%, N = 2976/20 514). This was followed by 1593 (7.8%, N = 1593/20 514) instances where the provider was accepting new patients, but no information was presented in the provider directory. We also found that in 2.4% of cases (N = 484/20 514), provider directories did not contain any information, but providers were not accepting new patients. Lastly, the percentage of providers who were listed as not accepting new patients when they were accepting new patients amounted to 2.2% (N = 445/20 514). Detailed analyses for each type of error are presented in the Supplementary Appendix.

Discussion

Inaccuracies in online provider directories are well established.^3,8,9 Previous studies have primarily focused on whether consumers encountered problems with the providers’ contact information, specialty, or network status.^1,3 Our findings here add to existing concerns about the information presented to consumers in online provider directories. We found that 1 in 4 provider directory listings exhibited errors, primarily focused on provider directory listings incorrectly indicated that providers were accepting new patients. Overall, we identified substantial differences across carriers, specialties, and states, but not plan type and rurality.

Our study has several limitations. First, our data are derived from ACA carriers in 7 states and 6 specialties. It is conceivable that other markets, states, and specialties could produce different findings. Moreover, while we conducted our study over the course of 19 months, provider directory accuracy may not be static over time. However, given previous work related to the persistence of inaccuracies even under the No Surprises Act, substantial improvements are unlikely.^9,10

Ultimately, our findings here indicate that the regulatory regimes of the ACA and the No Surprises Act are not fully protecting consumers and provide further evidence that corrective action and innovations to better protect consumers may be necessary.¹¹ Yet, information like whether or not a provider is or is not accepting new patients changes quickly as compared to more static items like provider specialty, making it challenging to keep updated in provider directories. However, vast discrepancies across carriers identified seem to indicate that higher rates of accuracy are achievable and may be associated with different administrative arrangements or dedicated resources.⁹

Supplementary material

Supplementary material is available at Health Affairs Scholar online.

Funding

This work was funded by the Robert Wood Johnson Foundation, Grant ID 80527.

Notes

Author notes