High cost can be high value

A new article just came out in JAMA this morning. The authors look at the thirty day cost of care for Medicare beneficiaries dependent on whether or not they go to academic medical centers or community hospitals.

In this cross-sectional study of more than 1.2 million hospitalizations, major teaching hospitals had higher initial hospitalization costs than nonteaching hospitals, but the total costs of care at 30 days were lower at major teaching hospitals largely because of lower costs for post–acute care services and readmissions.

This is a big finding if it is generalizable to younger populations. Right now, quite a few insurers attempt to control costs by narrowing their networks. Narrowing networks can produce savings in one of two ways. One is the insurer leverages competetive hospital markets to get a price concession. In North Carolina, BCBS North Carolina put their ACA networks out for bid between Duke, UNC and Wake-Med. Those three systems are all high end, comprehensive systems. UNC won the ACA contract to be BCBS-NC’s sole tertiary care hospital system in the Research Triangle. UNC, from what I heard, put in a significantly lower bid than Duke. This lower bid helped BCBS-NC cut their ACA rates substantially in 2019.

The other way premiums can be cut is to screen out sick people from any product offered by a given insurer. Mark Shepard has an excellent paper that showed an incredible amount of adverse selection due to the presence of a single high end hospital in Massachusetts. Other research shows significant adverse selection by the presence of a sole specialty hospital. Excluding a high cost hospital lowers premiums for the people who don’t want access to that high cost hospital but it makes people who want/need access to a high cost and highly specialized care portfolio worse off.

This paper suggests that we need to re-evaluate how we think about attributing costs and benefits of academic medical centers. They may be expensive up front but over a reasonable short time frame, they could be quite cost effective at the value they provide.

Moving to state based exchanges and catastrophic reinsurance

Pennsylvania is in the process of moving their individual market from being hosted and run by Healthcare.gov to a state based exchange. This is a step that several other states are taking. However, Pennsylvania is being really creative in why they want to do this. First let’s get the details from WESA:

House Bill 3 would have Pennsylvania operate the online health insurance exchange that has been run by the federal government since 2014 through the Affordable Care Act.

The bill is being co-sponsored by Lancaster County Republican state Rep. Bryan Cutler and Allegheny County Democratic state Rep. Frank Dermody. Both are their respective parties’ floor leaders….

Pennsylvania pays the federal government about $94 million a year to run the marketplace, Cutler said. With the proposal, Pennsylvania would use technology other states have already shown to be effective to run the marketplace at a much lower cost, which he estimated to be about $35 million annually…

the measure would also take advantage of a federal Section 1332 “reinsurance” waiver that can help to lower costs and tailor the program for Pennsyvlanians.

That last sentence is the most interesting and unique part of the Pennsylvania plan.

Pennsylvania wants to take the difference between what they pay to Healthcare.gov and what they think they can run their own exchange for and apply that increment to a Section 1332 reinsurance waiver. This is nifty. 1332 waivers are fairly common for reinsurance purposes in red, blue and purple states. Waivers are required to be federal budget neutral against a “no change” baseline. Reinsurance waivers bring in a non-premium infusion of funds that creates a wedge between claims and premiums. That wedge reduces federal subsidies and those lower federal subsidies are then fed back into the reinsurance program. The end result is that non-subsidized premiums are lower and the relative spreads of subsidized premiums are compressed. This leads to higher non-subsidized enrollment and slightly lower subsidized enrollment.

Usually, the state reinsurance waivers have the non-premium cash infusion come from some state tax revenue. Pennsylvania is not doing that. Instead, they are betting that they can run the exchange cheaper and while charging insurers the same amount of money, there will be a wedge that can be diverted to paying off some catastrophic claims. This is different.

I am not sure how much rate relief reinsurance provides when the reinsurance is funded by premiums. It is not a new wedge of non-premium related cash. It does two things. First it slightly increases the amount of premiums paying claims so total premiums can go down. It is also a pool of money that might be worth one or two percent of state wide premium that can eat some catastrophic claims. This is valuable. It reduces extreme tail risk for any one insurer. Less variance means marginally lower rates as well.

More importantly from the goal of reducing non-subsidized gross premiums, removing some catastrophic claim risk marginally increases the incentives for current insurers to expand their footprint and minimize the number of monopoly counties. Iowa’s insurance markets had an extreme example of a hyper expensive individual with recurring million dollar claim months.

In a competitive market where the subsidies are tied to the second least expensive Silver and there is one super-outlier who can not be re-insured against, every carrier lives in fear of being chosen by the one outlier. If they set their rates low enough to be attractive to healthy people, they lose money on the catastrophic expected claims. If they set the rates high enough to cover a $12,000,000 claim, no one buys their product.

No one wants to catch a spinning, falling knife.

Reinsurance funded by a fixed surcharge on premiums among all individual market insurers in the state smooths the pool and dramatically reduces risk.

If the Pennsylvania proposal goes through and the waiver gets approved, this is an interesting experiment to increase competition.

Performative instead of performing marketplace protection

There are many ways a state can improve its own individual marketplace. Some are effective at lowering premiums and increasing enrollment and others merely are full of sound and fury and contain no real money flows. We need to differentiate these actions as some states do a bit of both but have rules and cultural/political norms that hobble the markets more than anything that they have done since January 20, 2017.

New Jersey is a good example of both performing protection of their marketplace and performative projections.

New Jersey has a Democratic trifecta. It has passed a state based individual mandate that went into effect on January 1, 2019. It uses the revenue from the state based mandate to partially fund reinsurance. It is in the process of transitioning to a state based marketplace. It can free ride on New York and Maryland state based exchange awareness advertising.

These steps all actively improve the guaranteed issue, community rated individual market. There are also some performative protections as Andrew Sprung at Xpostfactoid lays out:

On May 31, New Jersey legislators introduced, with Governor Murphy’s support, a raft of bills* that  codify in state law the ACA’s coverage rules in the individual and small group health insurance markets, including protections for people with pre-existing conditions

Separate bills maintain a ban on medical underwriting or exclusion of pre-existing conditions (S626), mandate coverage of the ACA’s Essential Health Benefits (S562) and a set of preventive services (S3803), and limit age rating — the degree to which the oldest enrollees can be charged more than the youngest adult enrollees — to the ACA’s 3-to-1 ratio (S3810).**

Many other states with Democratic governors and legislatures have passed or have in progress similar laws that duplicate the ACA’s federal standards. Such laws are redundant by definition; they are designed as protection against future further Republican action to undermine the ACA.

Some of these are typical state insurance regulations of varying degrees of wisdom. But there is a fundamental challenge to a state trying to protect the ACA benefits and regulations without federal funding — money matters for the ACA market to by functional:

State-based ACA-mirroring laws would not mitigate the damage if the Supreme Court strikes down the entire ACA, however — including the marketplace subsidies and the Medicaid expansion. Guaranteed issue, modified community rating and Essential Health Benefits together would render coverage unaffordable for the majority of current enrollees without the federally funded subsidies, which the states could not afford to replicate.  Pre-ACA, states like New York and New Jersey that had enacted guaranteed issue were prohibitively expensive. New Jersey enacted guaranteed issue in the individual market in 1993; by 2003, enrollment had been halved and stood at 78,000, compared to about 300,000 today…

It is a statement of values and intent but without the money, it is also a statement of unaffordability.

New Jersey has another step that it could take that would increase coverage in the state, improve affordability for both on and off-exchange individuals and make the markets more functional but so far they have not done anything about a state based policy that makes the market smaller and more expensive than it could be.

New Jersey has fairly strong requirements of allowable cost sharing within a metal band. Silver plans in New Jersey are all at or over 70% actuarial value and Bronze plans must hover around 64% actuarial value. The possible spread in New Jersey is between 6 and 8 actuarial value points. The Center for Medicare and Medicaid Services (CMS) allows for Bronze plans to range from 58% to 65% actuarial value and Silver plans to range from 66% to 72% points. The maximum allowable spread is 14 points. As a rule of thumb, the bigger the spread in actuarial value points for the same insurer and network between the benchmark plan and the least expensive plan, the cheaper the plan is for subsidized buyers.

We have talked about premium spread strategies on this blog for years. I have several papers under review/accepted that play with this idea that should be out sometime in the second half of 2019 or early 2020. This is not a new idea.

New Jersey has the ability to maximize the spread. It could mandate all Silver plans to have actuarial values above 70% while mandating all insurers offer a low actuarial value Bronze plan and then anything else that the insurer wants. It could do that. Doing so could potentially double the spread which would dramatically increase affordability for subsidized buyers. Improved affordability for subsidized buyers brings in a healthier and cheaper, on average, risk mix which lowers non-subsidized premiums.

Yet, the decision New Jersey has made is that it wants to lower top end exposure for people who are insured at the trade-off of having more people uninsured and facing an infinite deductible. That is a viable trade-off to make, but it is one that weakens the market.

We need to separate signal from noise and make trade-offs explicit when we look at what states are doing to their individual health insurance markets.

A bill heading nowhere that health economists could love

Every legislature has hundreds of bills that go nowhere. They are filed, and then they are referred to a subcommittee where they are then politely ignored. Some of these bills are messaging bills that a representative is “DOING SOMETHING” other bills are perpetual hobby horses, while some bills are filed with the expectation that they die as a writing exercise for future policy development.

Oregon has an interesting bill that is a combination of a first draft for future policy and a DOING SOMETHING bill that identifies a real problem and proposes a solution that only a health economist could love.

The problem identified in HB-2009 is that insurance for the non-subsidized population that earns between 401-600% Federal Poverty Level is often too expensive. This is a real problem.

This bill links two policy proposals that have been floating out there. The first is a coverage expansion proposal. The state would create a Medicaid buy-in program for folks who are neither Medicaid nor subsidy eligible and earn between 138% to 600% FPL. They would be able to get an insurance policy that pays very providers significantly less than commercial rates which means the non-subsidized premiums would be at a significant discount compared to most/all other plans.

This is not too adventurous of a policy. New Mexico is investigating an off-exchange only Medicaid buy-in and several other states are thinking about this as well.

The financing side is also fairly straightforward. The bill proposed a state based individual mandate. The mandate would apply to individuals who don’t maintain affordable coverage for at least nine months a year. The revenue would partially fund the Medicaid buy-in administrative expenses.

Again, this is not too unusual. Several states have their own individual mandate that funds various health coverage related programs in a variety of ways.

The interesting portion is that the Oregon bill had to have been written by a health economist:

Except as provided in subsection (3) of this section… The penalty is an amount equal to nine percent of the individual’s taxable income as reported on the individual’s income tax return…
An individual is not subject to the penalty under subsection (2) of this section if:
(a) The out-of-pocket costs for the minimum essential coverage available to the individual exceed nine percent of the taxable income reported on the individual’s income tax return;

Now that is an individual mandate with teeth.

The goal of this mandate proposal is not raise revenue. The goal of this mandate as written is to make being uninsured more expensive than being insured. Some people will opt to do that. But most people will say that if they have to pay 9% of their income irrespective of whether or not they have insurance, they would rather at least get something instead of nothing for the 9% of income.

This bill is going nowhere, but it is an interesting proposal that is the logical (and completely implausible) set of solutions to a real problem in the ACA. It would lower premiums by offering a very low premium plan to a significant segment of the non-subsidized population, and it improves the risk pool by hip checking a good chunk of low risk folks into the risk pools thus lowering average claims and premiums.

Reality bites

I’ve been conferencing up and down the Eastern seaboard for the past week and many of my colleagues and collaborators are at their specialty major conferences as well. I found this from the American Society of Clinical Oncologists (ASCO) to be fascinating.

This study looks at the survival improvement results reported from the Phase 3 clinical trials of major (and expensive) new cancer drugs and then compares the results from a major cancer registry.

Ideally, the results from the clinical trial are the same as the real world results or perhaps slightly worse as new technologies, techniques, and learning by doing will have occurred to get a little bit of an incremental improvement wedge. That is not the usual case. Instead, the real world results are a bit worse than the clinical trial results.

Why does this matter beyond the obvious that lower survival times are less desirable than longer survival times?

As we move towards value based and outcome based contracts, we need to figure out which evidence is reliable and what things we consider in the creation of the contracts. The trial evidence is scientifically rigorous but for some reason, the translation suffers some degradation. Should initial contract phases be based on unadjusted clinical trial outcomes? Should there be a discount rate with some type of upside kicker to account for the likely case of lower pragmatic performance while adding some space for gains in the case of a happy surprise? How long should contracts run until they are renegotiated or re-benchmarked against pragmatic, real world evidence?

I don’t know the answers to any of those questions, but this study raises these thorny implications.