Randomly Assign Medicaid and Study?

My friend and Duke colleague Chris Conover suggests that states caught in the coverage gap due to their decision to not expand Medicaid (people too poor for exchange subsidies, but not eligible for Medicaid; WSJ says ~20-25% of the uninsured in North Carolina fall into this category) should consider a replication of the Oregon Medicaid experiment in which people were randomly assigned to Medicaid. States could then study the results.

There is merit to the idea in a State whose political leadership is uncertain about Medicaid expansion, and especially when some invoke the Oregon study as a reason to not expand Medicaid (I am not going to rehash that debate now). If multiple States undertook such an experiment,it would provide a great deal more evidence about the impact of coverage expansions, particularly given the idiosyncratic attributes of States. If North Carolina did this, we would need to very carefully design the study; for example, the Oregon study actually only included residents of Portland, as Chris notes in his piece, and we would need to make sure we invested enough resources in the study to provide definitive answers. North Carolina would especially need to make sure we could understand how such an experiment worked in both rural and urban areas.

My suggestion that North Carolina expand insurance coverage using a Basic Health Plan under Section 1331 of the ACA could certainly have an experimental component built into it. In fact, North Carolina could seek authority to do a BHP along side a Medicaid waiver, and randomly assign those below 100% of poverty to traditional Medicaid, or the private insurance/provider option that I proposed in the BHP. The comparison would then be to determine if the private coverage option differed from traditional Medicaid in terms of outcomes.

My white paper Don Taylor NC Health Reform Proposal 1 14 14 goes beyond health insurance expansion and calls for a demonstration/test of an alternative medical malpractice and patient safety approach among those newly covered, and efforts to expand the supply of health care providers by lessening regulation and expanding the practice authority of non-physician providers are also included. We should seek comprehensive reform efforts, and not only focus on coverage expansion.

The ACA has quite a lot of flexibility built into it for States, and the Obama administration has shown a willingness to allow States to experiment with different models and approaches. States like North Carolina have tremendous political leverage that we are now wasting. There are many potential approaches and models. North Carolina needs to pick one and move ahead with a coverage expansion that informs overall system reform, and commit to evaluating and learning from the results.

How North Carolina could (sorta) have universal coverage by 2016

Could North Carolina became the first State to achieve universal health insurance coverage? A student asked me recently what it would take financially to do so, and how it could most simply and quickly be done. Here is a quick estimate.

Using the Kaiser Family Foundations numbers, in 2016 there will be 1,216,000 uninsured persons in North Carolina if the Affordable Care Act were repealed. If implemented, they estimate that in 2016 398,000 persons will be covered by private insurance bought in exchanges who would otherwise be uninsured, and 377,000 (1) who otherwise would be uninsured will be covered by Medicaid if North Carolina undertakes expansion.

ScreenHunter_01 Dec. 12 11.54


That would leave 440,000 persons uninsured in 2016, or around 5% of the population, down from ~17-18% today. We could get to this point easily by expanding Medicaid, and the State developing its own exchange. Any state can achieve something like what is noted in first three rows of the table via the ACA; but what would it take to be the first state to cover everyone, with at least some modicum of insurance coverage?

The last row of the table shows how much it would take (I estimate $78 Million in 2016)  to achieve stop loss catastrophic hospital coverage, for the 440,000 left out by the ACA (didn’t sign up; ineligible) using the State Employees Health Plan infrastructure, and having a deductible of $10,000 for individuals, and $25,000 for families, with the state paying hospital costs above that. This is bare bones, hospital-only coverage to be sure, but it is the cheapest way to cover everyone with at least something that I can imagine.

A few details on how I got there.

Using the Hadley et al. (2008) estimate, the cost of care received by the uninsured in 2016 would be $1,100 per capita ($550 per capita actually paid by the uninsured, $550 unpaid, and therefore implicitly subsidized via the rest of the system) for a total cost of $484 Million in North Carolina in 2016 ($242 Million paid, $242 Million implicitly subsidized). Note that this cost will exist no matter what we do, but that half of it will be implicitly paid by those not receiving care by default.

If the 440,000 uninsured persons used care uniformly, the problem would be relatively easy to fix, but that is not the case. A recent U.S. Census Analysis shows per capita out of pocket spending by the uninsured ages 0-64 of $446 (compared to my estimate above of $550), but with 63.7% of such persons having expenditures of $0 (see Table 1). Similarly, a MEPS-based analysis showed that 2002 spending by the uninsured had a mean of $1,491, but the median was only $396, again showing that the distribution includes many $0 or very low users, and a few very large ones. The very large costs are catastrophic to families who have no way of paying them, and are a burden to hospitals and other providers who have to figure out how to cross subsidize such care, some of which they must provide under longstanding law, in what is essentially an unfunded mandate.

How could we achieve what I outline in the table above?

  • Expand Medicaid under the ACA, perhaps using the Arkansas “private expansion” Medicaid model in which persons between 100%-138% of poverty receive private insurance purchased on the exchange that North Carolina could set up. Those under 100% of poverty could be covered by traditional Medicaid initially. Perhaps over time, more of them could be moved to the exchange. I have long been in favor of seeking increased flexibility for our Medicaid program so that the state can focus on acute health care, and shifting more of the responsibility for long term care to the federal government.
  • Use the N.C. State Employees health plan to create a stop-loss catastrophic hospital insurance payment mechanism, through which the State would pay for the cost of annual hospital care above $10,000 for individuals, and above $25,000 for families. I estimate that this would cost North Carolina $78 Million in 2016, and this would be an annual recurring cost (for magnitude comparison, the annual budget for running the North Carolina General Assembly is around $51 Million in 2013).
  • Why use the States Employees Health Plan (SEHP)? The SEHP has members in all 100 counties, and is certainly paying bills in every hospital in North Carolina, which means they have negotiated rates. I got the $10,000/$25,000 figures by comparing them to the SEHP maximum coinsurance (cost share + deductible) amount for their traditional 70/30 plan (in 2014 this will be $7,586 for an individual; $22,758 for a family; looking for round numbers for for simplicity and since we are talking about 2016, I rounded up to $10,000 for individuals, and $25,000 for families). This would mean hospitals would also have to charge uninsured persons what the State Employees Health Plan pays for care, a step that UNC Hospitals has already taken, which is a consumer protection of sorts that at least means the uninsured aren’t charged a rate higher than what any insurance company pays.
  • Using a recent ASPE study on the distribution of hospital bills incurred by the uninsured, 56% of such hospital bills are larger than $10,000, and 28% of those larger than $25,000 (see Table 3). Without knowing the distribution of single uninsured v those in families, I assumed the cost of the program to be $78 Million in 2016, which is 40% of the half of the uncompensated care that is currently “implicitly financed” by hospitals not receiving payment.

I acknowledge that being uninsured is not only a financial problem and that there are human consequences. This is not the same as providing the 440,000 residual uninsured with health insurance, which would be expected to increase their use of care substantially. And it could correctly be said that this policy will most directly benefit hospitals who are now providing such care but not getting reimbursed. However, if our state expands Medicaid and sets up a vigorous health care exchange to market those policies, we will have done a great deal to address the issue of uninsurance. However, we know that some will not comply, and others will not be eligible for the ACA, yet they will still be living and working here, and in some cases using large amounts of health care. It would be preferable to develop a straightforward means of paying for this care.

The plan above in neither a liberal, nor a a conservative, dream solution. And the ideas above that could be tweaked in many ways. However, what I have written could plausibly be implemented by 2016, and by doing so, North Carolina could prove itself to be a leader in state-based health reform, and address the challenges of health reform, head on.


(1) Kaiser estimates that the number covered by Medicaid in 2016 under expansion will increase by 478,000, but the net reduction in the uninsured will be 377,000. 478k-377k=101k estimate of woodwork effect (those now eligible but not signed up) + crowd out from private and non-group insurance to Medicaid.

(2) Using AHRQ estimate of 32% of uninsured expenses accruing to inpatient hospitals plus hospital outpatient, yields estimate of $155 Million of the $484 Million total uninsured cost estimate in 2016 being hospital care. Using ASPE estimate that 28% of the uncompensated hospital bills incurred by the uninsured are greater than $25,000, and that 56% of them are larger than $10,000, I assume that paying for a weighted average of the individual and family uninsured hospital bills above $10k and $25k would account for half of the uncompensated hospital care, or $77.5 Million, in 2016.

(3) Note: The North Carolina Hospital Association claims that hospitals delivered $1.5 Billion charity care and bad debt in 2010 (see p. 41 of the pdf, page 37 of the document); I am not going to get into it, but they are discounting for a wildly unrealistic initial number, whereas the ASPE work asserts costs; note the markdown from charges to costs in table 3 row 1, is 25 fold.

ACA redistribution via Medicaid: what it means for future reform

The self imposed redistribution from mostly poor (mostly red) states, to mostly rich (mostly blue) states via the ACA Medicaid expansion is a direct result of the June 2012 Supreme Court ruling that made it voluntary.

That 7-2 court decision, and the subsequent state decisions, mean that the primary liberal/progressive health reform goal of expanding insurance coverage is being thwarted in some of the most needy states. Liberals/progressives have two choices: fight out the state-by-state Medicaid expansion decisions, or seek a health reform deal with conservatives that would be more likely to expand coverage in the non-expanding, poorest states, sooner. The first is not a pleasing outcome, and the second seems like a political impossibility.

This result was likely inevitable given the SCOTUS decision coupled with the re-election of President Obama; as I said in my post the day the decision was released:

…in the Medicaid aspect of the ruling the court identified the penalty of losing all of your states’ Medicaid funding if you don’t undertake the prescribed Medicaid expansion, to be something that the Federal Government could not do because it would be coercive to states. While this may seem to Conservatives a bit like the question “other than that Mrs. Lincoln, how was the play?” in the long run I suspect this precedent will be important going forward in policy debates.

Leaving the Medicaid expansion in place, while allowing states to not undertake the expansion without losing all medicaid funding has set up a fascinating test of ideology v. financial self interest for Conservative states. People’s lives are at stake here and I don’t mean to minimize that, but again, elections are important and I suspect what State politicians plan to do about the Medicaid expansions will be a key question in some states this Fall.

Reihan Salam has persuasively noted that a default insurance option is needed for health reform, motivated at least in part by the difficulties of healthcare.gov. I agree with him–if I could do just one thing to the ACA, it would be to add such a default option.

However, I have long felt that a political deal on health reform was needed, and such a deal was at the heart of a book I put out in September 2011 that claimed to identify a health reform deal between Democrats and Republicans, that had at its heart replacing the individual mandate with a default insurance option in the form of universal catastrophic health insurance implemented via the Medicare program. My overriding political point in Fall 2011 was that a Super Committee deal that made the SCOTUS case go away could have removed the doomsday outcome for both sides. We didn’t get such a deal, and also got a mixed SCOTUS decision, that has lead directly to an uneven Medicaid expansion.

My proposed deal is not a liberal/progressive dream, but then neither is the uneven Medicaid expansion.

Paradoxically, a SCOTUS ruling that had struck down the individual mandate as unconstitutional and invalidated premium supported private insurance sold in exchanges but that left the Medicaid expansion untouched would have produced what would have seemed like a bigger loss for Democrats at the time, but that would have at least resulted in all persons up to 133% of the poverty level being guaranteed health insurance. From such a base, the parties could fight another day and the red states, especially in the South, could have continued their tradition of saying they hated the federal government (all the way to the bank). But that is not where we are, so where do we go from here?

I think Liberal/Progressive reformers need a health reform deal because of the uneven Medicaid expansion, but we are in a difficult position because we have no control over the the biggest block to a health reform deal: the fact that elected Republicans do not hold any coherent health reform position(s) for which they are willing to vote (old posts here, here, here, here, here, etc)

It takes two sides to make a deal. This doesn’t mean there aren’t conservative intellectuals with reform ideas–Capretta, Moffit, Ponnuru, Roy, Salam, Douthat and others–these are thoughtful people with ideas that I think are reasonable to differing degrees. But whatever I think of them, some amalgamation of their ideas desperately needs to meet the Republican-controlled Commerce Committee in the House of Representatives, ground zero for any actual health reform effort; and then the CBO.

I think that all of these intellectuals realize that the Republican party is the only way for their ideas to reach legislative fruition, and they know that eventually the Party will have to be for something in health reform. And I believe they are quietly working towards making this case within the Republican Party. The entire country, but especially Liberals/Progressives who know that more must be done on health reform, should be rooting for them to succeed.

Update: while I wasn’t attempting an exhaustive list of conservative public intellectuals with reform ideas, I should definitely have included Yuval Levin.

ACA: self imposed redistribution from poor to rich states

At present, 24 States (and DC) have decided to move ahead with the Medicaid expansion provided for in Obamacare, and 21 have rejected expansion, while 6 are still considering their options. If the current decisions hold, it will result in a self-imposed redistribution of money from poorer (and typically Red states), to richer (and typically Blue ones).

According to an analysis I have done using Kaiser Family Foundation data–in 2016 alone–the 24 expanding states will receive $30.3 Billion additional federal dollars, while those not expanding will forego an additional $35.0 Billion they could have had (the fence sitters have an aggregate $15.2 Billion at stake in 2016). This represents a huge redistribution of federal money from non-expanding to expanding states. The table below highlights the biggest self imposed losers, and winners, again for 2016 alone (there are predictable impacts on state uninsured rates).

ScreenHunter_04 Oct. 25 14.45

Note: total is for all states in foregoing & gaining group.

Most states will also have increased spending if they expand Medicaid; for North Carolina, in 2016 Kaiser estimates that the state will have to spend $390 Million to leverage around $4 Billion in extra federal money, and reduce the ranks of the uninsured by around 375,000 persons (about 475,000 more would be covered by Medicaid). To put the foregone $4 Billion in context, North Carolina’s total Medicaid budget in fiscal year 2014 is around $14 Billion, and there is certainly no alternative proposal as impactful on the uninsured in my state at any cost.

States that are not expanding Medicaid have historically received more in federal spending per dollar of federal taxes paid by the state ($2.18) as compared to States that are expanding ($1.85) and those that are considering expansion ($1.53), all in 2009, a year with a very large federal deficit. In year 2000, the last year of a federal surplus, those states rejecting expansion received $1.36 in federal spending per tax dollar paid as compared to $1.10 for those undertaking expansion (the fence sitters were net donor states, $0.87). Similar patterns held in both 1994 and 2004 (other years shown in this table I put together using IRS & Kaiser sources Tax Flows Table.10.25.13_blog).

While the Medicaid program is not the only means through which richer states have cross subsidized poorer ones, it has been a large and consistent source of such flows. By choosing not to expand Medicaid, the poorer, mostly politically “red” states are redistributing money toward the richer, mostly politically “blue” ones (there are exceptions; red Kentucky is both expanding Medicaid and has one of the best functioning State exchanges). Further, those States that are expanding Medicaid have also tended to set up state-based insurance exchanges, which are currently operating much better than the federal one, meaning that income based subsidies associated with the purchase of private health insurance may flow less freely to poorer states, at least in the short term. And there is a court case that could stop the flow of such subsidies to states not operating their own exchange all together. I have not tried to estimate the magnitude of these sources of redistribution from poor to rich states under different scenarios because things are so fluid, but the Medicaid numbers outlined are potentially just the start.

The bottom line is that if the current State Medicaid expansion decisions persist, the unintended story of the ACA will turn out to be the redistribution of money from poorer States, to richer ones, an outcome imposed by the poorer states, upon themselves. I will write more about what I think this means for the future of health reform over the next few days.


Note: I was assisted by excellent research assistance from Callie Gable, a Duke Undergrad. Any errors are my responsibility, however.

The cost of States not expanding Medicaid. Kaiser Family Foundation, July 2013.

Internal Revenue Service (IRS) report Internal Revenue Gross Collections, by Type of Tax and State, for Fiscal Year 2009

Internal Revenue Service (IRS) Federal Funds – Summary Distribution, by State and Island Areas : 2009

Talking North Carolina Medicaid on Plain Talk Politics

I was a guest along with Suzanne Buckley of NARAL NC on Plain Talk Politics hosted by Jeanne  Milliken Bonds last week. The show focused on Medicaid and Reproductive rights in North Carolina, and was taped last week before the General Assembly adjourned. Here is the audio tape (about 40 minutes).

Follow @PlaintalkPol on twitter and Jeanne Bonds if you are interested in North Carolina politics, viewed through a policy angle.

Republican Disconnect: Abortion and Medicaid in North Carolina

There is a disconnect between the North Carolina Republican parties’ passionate commitment to protecting the health of women who would choose to have an abortion, and their cool, immediate opposition to the Medicaid expansion available under that ACA. Proponents of S353 spoke in terms of no stone being left unturned to improve and reform abortion clinics on the floor of the House of Representatives during the debate on Thursday, July 11, but they blithely turned their backs on 500,000 of our most vulnerable fellow citizens early this year when they rejected Medicaid expansion.

As the Republican Conference Leader, Rep. Ruth Samuelson (R-Mecklenburg) said in response to questions about the cost of the proposed regulation of abortion clinics (as recorded in Rose Hoban’s live blog of the debate):

“The point that has caused me the most befuddlement, the issue of this being cost-prohibitive, you don’t raise safety because its cost prohibitive. Do we want someone to die in an abortion clinic in North Carolina before raising the standards?” Samuelson said. [she was saying you can’t put a cost on safety/someone’s life]

Rep. Pat McElrath (R-Emerald Isle) offered a similar sentiment that poor women should not be denied access to care because they could not afford to pay for it:

“I’ve listened about safe and healthy access, and that’s exactly what this bill is about,” McElrath said. “Poor women should not have any less access to clean conditions to sanitary conditions than those who can afford to go to that one facility[that now meets the proposed regulations-my insertion] in Asheville.”

I am going to grant the benefit of the doubt to House Republicans that their singular goal is to protect the health of women to drive home the the disconnect between their advocacy of S353–where cost was not even viewed as a legitimate line of inquiry by House Republicans; who could put a price on health and safety!–and their rejection of the Medicaid expansion that could have provided insurance to 500,000 poor North Carolinians, protecting them from financial hardship, helping to ensure access to care, and improving their health.

Cost has often been stated as a reason to not expand Medicaid, even though the terms of the expansion are financially advantageous to North Carolina, and our State could later cancel the expansion for any reason.

Others have said that being uninsured is better than being covered by Medicaid, a politically motivated sentiment that does not hold up to a broad and careful look at the research evidence. All that can be said about Medicaid (and insurance generally) and health cannot fit onto a bumper sticker, but expanding Medicaid would extend access to our State’s health care system in a manner that is supported by our hospitals, and which does not forestall future health care reforms. (Lots on the impact of insurance and Medicaid on health here, with more on the recent Oregon Medicaid study here).

I actually don’t know if there is a particular quality of care problem in abortion clinics in North Carolina, but I do know that there is a general quality of care problem in our nation, with medical errors having been identified as the 7th or 8th leading cause of death in the United States, leading to between 48,000-98,000 deaths in the U.S. in 1999 alone; a decade later, progress had been made, including more focus on quality in outpatient settings, but quality/safety problems remain persistent, including this study showing medical errors remaining a problem in North Carolina hospitals a decade after the Institute of Medicine’s To Err is Human was published. So, I don’t take quality concerns lightly, and efforts to improve quality and improve safety are important.

However, zeal to address these issues shouldn’t be reserved only for women who choose to receive an abortion.

I believe the passion stated by proponents of S353 is real and heartfelt. I also think many of the supporters of S353 think abortion should be outlawed, but I give them credit for stating publicly that access to abortion is a constitutionally guaranteed right. And there may be a particular quality issue with abortion clinics in North Carolina, or there might not be–I truly don’t know.

I would like to respectfully ask the Republican majority that now rules our State to reconsider their decision to not expand Medicaid, and to apply some of the passion and the ‘no cost is too high sentiment’ displayed during the S353 debate toward expanding access to health care to poor people who aren’t seeking an abortion, and then set about developing a North Carolina specific plan for health reform that focuses on access, cost, quality and safety problems wherever they exist. There are many North Carolinians of different political persuasions ready to join them on that journey.

Key Difference in Senate v House Tax Reform hints at Medicaid expansion

This is a handy comparison of HB 998 v. Senate modifications to it, released yesterday (h/t Elizabeth Malm via twitter). Several points following up on yesterday’s post:

  • The Senate version reduces General Fund tax receipts by $684 Million from FY 2013-15; about a $300 Million larger tax cut than the House over this time period. Will the total collected add up to planned spending? (impt question, not my focus here)
  • One of the larger, and most consequential differences from a health policy perspective is that the Senate version caps the sales tax refund (state and local) that a 501-c-3 corporation (Non Profit) receive from the state. Below is a cut from page 5 of the comparison document (column to the left is the House proposal, the one of the right is the Senate’s).

ScreenHunter_01 Jun. 12 10.17

Currently, non profits are exempt from state and local sales taxes, and apply every 6 months to have them refunded by the State. Here is a 2011 North Carolina Tax Expenditure report laying out the cost to the state of that year.

ScreenHunter_02 Jun. 12 10.33

I am trying to explicitly nail down how much of the $228.2 Million in 2011 flowed to hospitals, but I suspect it to be a very large proportion based on estimates of what several hospitals think the proposed change will mean for them. Update: this report shows that hospitals received between 75-81% of the total (h/t Ed McLenaghan).

Back to the Senate proposal. In 2014, only the first $5 Million in purchases will be exempt from sales tax (all are now exempt, which would be continued by the House), but by 2016 this will drop to a $100,000 exemption. Most NP are small operations, but hospitals are big operations with many purchases that are subjected to sales tax, and most of them are organized as nonprofits. In short, for a place like Duke Health System, UNC, WakeMed, etc. this will result in a huge tax increase (the aggregate tax increase on all now exempt nonprofits implied by the Senate proposal is ~ $730 Million from 2014-18). I have the increase under a proposal like this as quite an eye popping number for Duke University Health System, but I am not going to give the number because I cannot confirm it.  In one sense, what has been proposed by the Senate is better that what it might have been when there was discussion of the full sales tax on professional-to-professional services in addition to this, which under some interpretations might have meant every time a Duke physician consulted in Duke’s hospital, a sales tax would be levied (because of how they were organized).

Anyway, the line of folks in North Carolina mourning for the woes of Duke is pretty short, but my main point is this. When the Senate moved from its tax reform that heavily relied on sales tax increases toward what they proposed yesterday, they kept a very large sales tax increase on nonprofit corporations in place, the most consequential of which are hospitals. I have been saying all along it was only a matter of time until N.C. expanded Medicaid, but this looks like a placeholder for the Medicaid expansion in North Carolina, with the hospitals essentially setting up to pay a large proportion of North Carolina’s cost of the expansion (~$900 Million in 2014-19 time frame; the magnitudes of the sales tax increase discussed above and State Medicaid outlay for expansion are quite close), when the politics allow it to occur. Most likely next year, to take effect in 2015. However, as the revised Senate tax reform released yesterday points out, the General Assembly is undertaking all sorts of huge changes with very little notice as their session winds down.

Update: live blog from Tax Foundation of Senate debate; about 1:20pm there was an amendment to delay 1 year and add a tier for the sales tax exemption of nonprofits.