Narrow network thoughts

Health Affairs has another good article by Dafney, Handel, Marone, and Ody examining narrow networks on the ACA exchanges. I like their methodology and the conclusions jibe with what others have published.

The hospital discharge data, which we obtained for a single recent year for all eight states, enabled us to construct a utilization-weighted measure of breadth, instead of relying on simple counts of hospitals. We defined hospital network breadth for a network in a given rating area as the number of discharges for patients living in the rating area that occurred at in-network hospitals, divided by the total number of discharges for patients living in the rating area.

And they found that narrow networks are less expensive:

We found that if all plans offered these broader provider networks, then the population-weighted average premium for the second-lowest-price silver plan would increase by 10.8 percent, or $330 per year. Our estimate implies that narrow networks lowered premium subsidies by roughly $2.4 billion in 2014.3

There are a couple of stories that the authors acknowledge as to how narrow networks save money. I want to recap and comment on those potential stories and then extend on them.

Outlier reduction

Here, insurers exclude outlier providers. There are two different types of outliers that can be excluded. The first type are providers (hospitals and doctors) who have very expensive contracts. They might be specialists, they might be high prestige providers, they might be the CFO’s kids, they might have a special payment rate enshrined in state law. Trimming these providers from the network brings down the average rate per service paid.

The other outlier trimming strategy is to focus on unusual utilization patterns. An identical patient presented to different physicians/hospitals could have significantly different courses of treatment. One group of providers may watch and wait far more often than others who are willing to order advanced imagining or engage in early surgery. The long term outcomes may be the same but the cost profiles are very different. Insurers who are in a market for a while tend to know which docs go all in as soon as the patient shows up. Cutting those providers out of the network will reduce incremental utilization.

Negotiating Leverage

Being able to say no means insurers can get better rates. Insurers can approach a dozen doctors and say that they need three and here is the rate. Some doctors will take that rate and some won’t. This is an urban strategy that works in markets where failing to meet network adequacy standards is nearly impossible. Rural areas won’t see this strategy as much because the market dynamics of payers and providers flip. Providers have power through network adequacy as it is either them or drive forty miles for the next specialist.

These two major stories are good stories to tell from a public policy point of view. Markets work, competition works. The next story is not as favorable.

Win ugly

Mark Shepard at Harvard has an excellent paper on narrow networks and adverse selection in Massachusetts.

substantial adverse selection against plans covering the most prestigious and expensive “star” hospitals

Here, networks can be inexpensive because they exclude providers who are likely to attract very high cost individuals. This can be entire facilities such as a prestigious sub-specialty hospital, or cancer centers or provider practice groups. Network adequacy in urban areas can be easily met even as providers who sub-specialize in coagulation disorders or cystic fibrosis or genetic defects are excluded from the network. This creates a self-selected covered population that is comparatively cheap to cover.

Under this story, risk adjustment should play a role in sending funds from the low cost network insurer to the insurers covering the people who are very sick in the broad network. The ACA risk adjustment formula is based on average state wide premium. This means high cost networks may be underpaid relative to the cost of care that they provide. This is a problem.

These are the major mechanisms through which narrow networks drive down premiums. Every narrow network is unique in its own special way. The networks I helped build at UPMC relied mostly on negotiating leverage to get good rates. Other networks, at carriers with a large fraction of their premiums going out the door as risk adjustment payments, were designed to cherry pick a healthy population. But almost all of the narrow networks will tell at least one of these stories.