Network adequacy standards and the ACA Enhancement bill

The House Democrats have released HR 1425, the Affordable Care Enhancement Act, as their bid for what ACA 2.0 could look like. With the proviso that it will not be enacted this year, it is a compilation of priorities. The two major elements are enhanced affordability through increased subsidization and throwing money at states that have not yet expanded Medicaid. There are significant but more modest policy changes in the law. I want to look at Sec 115. This section requires exchanges to have quantitative network adequacy regulations. It can be used to significantly limit narrow networks.

Network adequacy regulations require insurers to have a certain number or ratio of certain types of clinicians within certain geographies or populations. The intent of the regulation is to make sure that a plan that is being sold can actually offer sufficient clinical services that are reasonably conveniently located to the covered population. Rural regions tend to have have network adequacy standards that are looser than urban regions. An urban region might have a requirement that there are at least two endocrinologists within 20 miles of the center of each zip code while a rural region could require two endocrinologists within 50 miles of the center of each zip code.

Minimal viable network adequacy standards are necessary.

However, these metrics can be easily gamed.

I spent several years working at UPMC Health Plan on provider network data management and network design. After I left the network and provider team, I worked on risk adjustment revenue optimization. We could build networks that easily met Pennsylvania’s quantitative network adequacy standards that were actively repulsive to individuals with complex medical needs. We could build networks that were attractive to individuals with high cost conditions that would be, at best, questionable to meet regulatory approval without significant review.

Quantitative network adequacy standards can be gamed at two levels. First, in dense regions, there are a ton of providers. some providers have very high cost, net of any reinsurance and risk adjustment, patients. Carving out one or two docs from a potential network of several hundred docs could lead to a notable drop in per member per month spending as the super high cost patients won’t entertain switching networks if their particular doc is not in the network. The other major way to design repulsive networks that meet regulatory standards is to include generalists within a specialization rather than the sub and sub-sub-specialists.

For instance, rheumtology is a specialty that is in most quantitative standards for network adequacy measurement. There are massive variations in what a rheumotologist actually does. Most rheumatologists will take care of patients like my dad. His knees creak, his fingers are a bit stiff and he should lose a few pounds while maintaining or improving his functional strength. An insurer can meet network adequacy standards by only contracting with rheumatologists who take care of patients like my dad. Pediatric rheumatologists are far rarer. Network adequacy standards can be readily met even if an insurer never contracts with a pediatric rheumatologist. Any family that needs a pediatric rheumatologist will either never buy a plan with a personally inadequate network, or leave that network at the next opportunity.

Network adequacy standards that are based on either simple ratio or radii methods can and will be easily gamed in the face of imperfect risk adjustment. Different approaches to either measuring what is an adequate network, pricing of net premiums or risk adjustment are needed either instead of or in addition to quantitative network adequacy standards.

PS: I was on a team that had gotten a grant to look at an alternative approach this year but the project has been rescrambled and redirected to COVID related work. Hopefully we can find funding and interest for a 2021 attempt on this question.