As open enrollment is creeping up on us, I’ll be spending some time doing retrospectives on interesting policy experiments. The Arkansas model of premium support for most of their Medicaid population is a fascinating experiment. The basis of the experiment was for the Democrats to find someway to get more than half of the Republican caucus onboard for a Medicaid expansion. So they did not do expansion, they did premium support for private insurance on the Exchanges with supersized subsidies for the Medicaid eligibile population. The first year results are mixed. Access to care, utilization and enrollment are great. The problem will be cost, and that problem is an underlying issue with the Arkansas 1115 waiver.
First, the good news:
MT @PolicyRx: After Arkansas adopted private option… • ED use ↓ 2% • uninsured ED use ↓ 24% • uninsured hospital admits ↓ 30% #nashpconf14
— Dan Diamond (@ddiamond) October 7, 2014
A decline in uninsured ED usage would have met everyone’s expectations as the population of the uninsured plummetted. A smaller population, means, all else being equal, less utilization even if the declining population cohort is cherry picked as low utilizers. I can’t say anything about the characteristics of the newly covered population. The same logic applies to uninsured hospital visits, fewer uninsured individuals should lead to fewer uninsured hospital admissions. These were expected results.
Now the total decline in emergency room visits was unexpected. Massachusetts saw ER usage increase after insurance expansion. This is utterly fascinating and meaningful as a 2% decline on a population of a couple million people is statistically significant and more importantly financially significant. It may mean that there is successful diversion of acute care cases to lower, more appropriate levels of care such as urgent care clinics or primary care provider office visits. More research is needed for that set of claims to be made. If Arkansas can replicate the ER diversion and minimization for next year and figure out why it is working, this is something that should be taken to other states for replication.
Enrollment is good as well. Slightly more than 200,000 people are in the program. The net uninsured rate in Arkansas has been cut in half
Access to care is pretty good as well as Arkansas is using full commercial reimbursement (more on this later) for broad commercial networks. Any doc who wants in is probably in.
And that is where there are problems lie.
Arkansas did not do a straight up expansion like Kentucky; instead they decided to shovel money to Blue Cross and Blue Shield and pay commercial rates for Medicaid. That is expensive as hell compared to paying Medicaid rates for Medicaid or even Medicare rates for Medicaid. This is a problem because the 1115 waiver process requires the states that apply for a waiver for an experiment to do better care at the same cost to the Federal government as regular Medicaid or the same care at lower cost to the Federal government. The process should be budget neutral.
Commercial rates are usually based on Medicare plus a significant kicker. Mayhew Insurance basic commercial contract for providers in common specialities in densely populated and well servied regions is Medicare plus 25%. Medicare rates are theoretically based on two factors, the average cost of a service in a region with all the back-end costs amortized into the rate plus several sweetners (graduate medical education is the most notable). Commercial rates are where providers make their money. Medicaid rates tend to be calculated at the marginal cost of providing a service.
The Kaiser Family Foundation looked at the rate changes that PPACA mandated for Medicaid primary care providers in calendar year 2013 and 2014. For a certain set of billable codes, primary care providers in Medicaid would see a rate bump from Medicaid rates to Medicare rates. Arkansas paid roughly average national rates for its Medicaid program in 2012, and the gap between its Medicaid and local Medicare rates was between 25% and 50%. Assuming with absolutely no evidence that BCBS in Arkansas has similar reimbursement models as Mayhew Insurance does in a very different state, for the same service to a Medicaid recipient as a commercially insured recipient, Arkansas would be paying at least 60% more per unit of service under a commercial contract than under a standard Medicaid contract.
There are a couple of ways to argue that the very high reimbursement rate is worth it from a budget neutrality view point. Arkansas adapted all of the arguments to some degree. The first is that the transition from a fee for service to managed care system will align the insurer’s incentives to correctly stamp down on utilization. The much higher prices would be paid for far fewer services as unneccessary services are avoided and people are directed to lower cost but still appropriate levels of care. We might be seeing some of that in the ER utilization data. But traditionally the shift from Fee for Service to HMO/Medicaid Managed Care models won’t produce savings anywhere near the magnitude needed for budget neutrality.
Most of the budget neutrality argument for Arkansas is coming from comparing two counterfactuals. The first counterfactual is what would it take to expand the 2013 Arkansas Medicaid network to handle an additional 200,000 or more people. The second is what would it take to handle the entire reasonably healthy portion of the Arkansas Legacy and Expansion Medicaid population on a commercial network. Arkansas was not calculating the cost differential on the basis of a 2014 private network comparison to a 2013 Medicaid network. Arkansas was comparing a hypothetical 2014 expanded Medicaid network with higher reimbursement to a pre-built Commercial network with standard commercial reimbursement.
This set of assumption carries a lot of water for Arkansas as it allows them to assume away most of the observed cost differnetial as they upcharged their Medicaid network to nearly equalize it to commercial rates. From here, cost sharing, utilization review, and risk management could come into play to get the last bit of budget neutrality.
I have doubts about the need to assume equalized rates for Medicaid Expansion and traditional broad scale commercial networks as there are numerous other states that have been able to expand their Medicaid enrollment without seeing providers flee the networks when their rates did not increase by 50%. Some states have increased rates slightly for Medicaid, but no one else has seen dramatic network drops for the lack of higher reimbursement.
The fate of pure private option programs like Arkansas will need another year or two of evidence during the demonstration phase to see if budget neutrality assumptions are truly violated but my inkling is that they will be. At that point, even a favorable federal Health and Human Services Department will have a hard time granting replication waivers for the Arkansas experiment. Instead HHS will probably go with either HIP 2.0 from Indiana where a current program is massively expanded without massive reimbursement hikes or Healthy PA from Pennsylvania where the state is effectively reinventing the Medicaid Managed Care wheel with Medicaid like reimbursement rates in order to expand Medicaid without calling it a Medicaid expansion.
“Enrollment is good as well. Slightly more than 200,000 people are in the program. The net uninsured rate in Arkansas has been cut in half”
More good news for the ACA, even though Arkansas is not fully expanding.
So, I fortunately am able to get health insurance through my husband’s work, as I’m self-employed. But you pay full freight, which is over $400 extra for me for some fairly unimpressive coverage. So I wanted to shop around to compare the two. Last year at this time (open enrollment for group health), the website was borked all to hell and I didn’t get anywhere. I re-enrolled with his, thinking I could drop it later. I couldn’t, they wouldn’t let me unless I was picking up another group health policy at my own employer or something odd like that. Anyway. I’m trying it again this year, open enrollment closes on Oct. 31. I can’t price a policy out under the exchange until November 15. I realize I’m lucky to have a couple of options here, but are my choices really to stick with my husband’s underwhelming plan or to take a big gamble that the prices on a self-pay policy will be favorable?
big ole hound
The premise that logical thought will prevail in Arkansas is faulty. ACA is hated by many there because a black man engineered it. Impossible. I would like to see the racial make-up of the 200k in the program. After spending time there a few years ago and seeing racist hatred at it’s best, I moved.
@jibeaux: Look at HealthSherpa with your age/zip code for a rough idea.
Or if you have some time on your hands and you live in Washington State, look at
As this has all the 2015 information of what plans are being offered in what counties at what rate. You should be able to get a very good idea of what type of coverage that you want to get for yourself would be available on the Exchange for a price that you are comfortable with.
As always thanks so much for these posts Richard.
I’m wondering whether the difference between the MA and AR admit rates is the explosion of urgent care facilities in the time between.I don’t know if they are everywhere the way they are here in NYC, but they would seem to offer a much better patient experience than an ER visit for easily treated acute issues.
We obviously can’t know whether that is true, but the proliferation of walk in treatment of all kinds is an unexpected (to me anyway) consequence of the PPACA.
@jayackroyd: Jay — I don’t know, as other states with 1/1/14 Medicaid expansion dates are seeing temporary ER surges and declines once people get used to the new system… so the variable of more frequent/more accessible urgent cares in 2014 v. 2007 Massachusetts probably does not have a lot of explanatory power:
Something interesting is going on where flat ER utilization would be a massive win, and a decline is interestingly, puzzingly, benefically bizarre