In or out: network construction thoughts

Good take by @bjdickmayhew, but I'd adjust the editorial bias to consider that by excluding certain providers from networks, insurers are as or more interested in reducing the risk of certain physician practice patterns (cause) as they are avoiding patient utilization (effect). https://t.co/2rdI2LYAGs

— Marshall Votta (@ydeologi) October 24, 2018

This is a reference to a post earlier this week regarding how insurers can legally dodge high cost individuals if there were to be no risk adjustment payments that at least partially compensates them for risk. I mentioned network games:

only offer narrow networks with donut benefit design and only in certain counties… If their data says that there are a dozen individuals with hemophilia, they won’t offer coverage of the regional hemophilia treatment center.

This reply (rightly) calls me out for excluding some of the other very valid and far less nefarious reasons why an insurer may decide to exclude a provider or set of providers (hospitals or docs) from a narrow network product. There are several different types of decisions that can lead to an exclusion from a narrow network. Let’s assume that the narrow network that is being crafted is in an area with reasonably high provider density. Therefore adding or dropping any particular provider won’t run into any network adequacy (regulatory or marketing) concerns.

1) Practice patterns — if identical patients on all observables go to two different doctors and one doctor consistently orders far more intensive and expensive treatments than the other doctor without any validated improvements in quality, care or patient experience, then cutting out the high expense provider may make sense.

2) Quality — if one doctor has a consistent history of much higher than risk adjusted expected re-admissions or avoidable complications for their patients then carving out a bad doc improves both the financial performance of the network and the quality of the network.

3) Price level — Assuming a doctor has perfectly acceptable but not exceptional quality but has a contract that pays them twice what other, similar doctors are paid for the same procedure lists. Here exclusion is a cost control measure on the insurer. It would not be surprising if a renegoatiated contract at a lower price level would bring that doctor back into the narrow network.

4) Strange contracting reasons — Insurer-Provider contracts are a black box to me. There are frequently clauses that state if a narrow network is offered all providers under one contract must be in it OR narrow networks could be a take it or leave proposition between several individuals offering similar services in a micro-market.

5) Risk management/risk dodging — this is the hemophilia center example. If a practice is the only group that treats a very well defined, miniscule and expensive population, and risk adjustment/reinsurance is inadequate to cover costs, then cutting out the particular practice is attractive.

6) Halo effects leading to higher pricing — Academic Medical Centers (AMC) and dedicated Childrens Hospitals tend to be expensive on a per unit basis for regular services and really, really good at unusual services. There is a quality argument to pay very high prices for unusual services but unbundling the contract is tough. We saw this with ACA plans in Washington State with Seattle Childrens in 2014:

Washington State starkly illustrates this trade-off as the Seattle Children’s Hospital was excluded from most of the narrow network plans on the Washington State Exchange initially.

Mark Shepard has found that high prestige hospitals create significant selection effects.

7) Data infrastructure for population health management — as we move away from Fee for Service (FFS) towards some variant of alternative payment models (APM) and population health management, the data and infrastructure demands placed on practices have changed. Some practices might not be able/willing to be involved in this different care financing model or the insurer wants to keep the APM in a very small universe for testing and management control purposes.

8) Corporate walls — I came out of the integrated delivery network world of UPMC where corporate owned both the insurer and a huge hospital system. Some narrow networks were fundamentally build on UPMC owned and employed clinicians plus a miniscule smattering of other providers that were needed for regulatory purposes. A perfectly competent, reasonably priced doc that was on their own or in another system might never be invited into the narrow network.

9) Just because — weird stuff happens.