Friend of the blog Bradley Flansbaum drew my attention to this MedPAC status report on Medicare Advantage (MA), presented by Scott Harrison and Carlos Zarabozo on December 8, 2016. It includes some MA facts I didn’t know, and I’m guessing you (or the vast majority of you) don’t know them either.
In 2017, the average, PMPM rebate to MA plans will be $89, higher than it has been since 2011. I don’t know what the figures are prior to that year. (The rebate is a percentage of the difference between an MA plan’s bid and the benchmark. The percentage, as well as the benchmark, varies by plan quality. Think of it as a kind of shared savings. Plans are required to use rebates to increase benefits or reduce cost sharing.)
For the 2017 plan year, the average MA bid was 90% of traditional Medicare spending (TM) for a comparable beneficiary. I do not recall it ever being that low, but could be wrong. Plans are still paid above their cost (hence the rebates), but in 2017, for the first time this century,* payments to plans are at parity with TM spending. This is something that MedPAC has advocated and is an ambition of the Affordable Care Act. Mission accomplished, with the caveat that MA coding intensity may still not be completely adjusted for. If so, plans could still be paid above a (properly diagnostically adjusted) TM rate.
MA payment rates are based on TM Part A and Part B costs. But — and this is something I never knew — this includes TM enrollees that only have Part A, and they spend less on Part A than those who also have Part B.** MA covers A and B, and 87% of TM beneficiaries have both (this proportion is shrinking over time). The upshot of this is that payment rates based on all TM beneficiaries are lower than they would be if they were calculated on Part A and Part B enrollees. Whereas, coding pushes payments up, this mismatch pushes payments down, by how much the status report doesn’t say.
MA enrollment growth is more rapid than Medicare as a whole, as it has been for many years.
There’s more in the report, though not a lot more. It’s a PDF of a brief PowerPoint presentation. But I thought the above were among the most interesting, and less widely known, facts.
* I could be wrong on this, but it’s certainly the first time since at least 2003.
** Just keep reading that sentence slowly. You’ll get it.
Republicans may have the votes pass a reconciliation bill that will hollow out the ACA in a few years. They pinkie swear that, at that point, they’ll pass some kind of replacement.
But for that to work, they’ll need Democratic support. Chuck Schumer is already saying they won’t get it: “We’re not going to do a replacement. If they repeal without a replacement, they will own it. Democrats will not then step up to the plate and come up with a half-baked solution that we will partially own. It’s all theirs.”
Maybe this is bluster. But I don’t think so: the 2016 election offered a lesson to Democrats about the political spoils of obstruction and brinksmanship. Repeal and delay could just be … repeal.
In that vein, Chris Koller has a must-read op-ed over at Politico. As the health insurance commissioner for Rhode Island, he oversaw a market for individual insurance that prohibited insurers from discriminating against the sick, but didn’t compel anyone to participate. It didn’t work so well:
To maintain the right balance between rates in the sick pool and the healthy pool, our office had to conduct extensive rate reviews and regulation, including forcing insurers to limit their rate hikes. The local Blue Cross plan, obligated by charter, was the only insurer to offer products; other carriers in Rhode Island found the market overregulated and too small. Only about 15,000 people bought coverage, and consumer dissatisfaction with this model was high. Approximately 100,000 people in the state—or about 10 percent of the population—either could not afford coverage or elected to go uninsured.
Fiddling with insurance regulations—a key Republican priority—wasn’t the answer.
With the legislature, we regularly considered less comprehensive benefit requirements. State-mandated benefits, however, were found to comprise only 10 percent of premium costs, and half of that amount was for mental and substance-use-disorder services now required by federal statute. Services comprising the other 90 percent were what most people consider essential — pharmacy, hospital and physician care. …
Wider rate bands — charging lower rates for young and healthy folks and higher rates for older ones — might have helped a little, but at the political cost of an irate and vocal older constituency. Taxing a broader base for our high-risk pool may have allowed for lower rates in our healthy pool, but historical experience with high-risk pools has shown that maintaining political support for an adequate tax base for a segregated pool of sick people is very challenging. …
The big lesson here is quite simple: Voluntary insurance is hard to do. There is a reason banks compel mortgage holders to buy homeowners insurance and states compel car owners to purchase liability insurance. Insurance pools cannot be comprised solely of those who think they will need to use coverage.
The number of uninsured people would rise from 28.9 million to 58.7 million in 2019, an increase of 29.8 million people (103 percent). The share of nonelderly people without insurance would increase from 11 percent to 21 percent, a higher rate of uninsurance than before the ACA because of the disruption to the nongroup insurance market.
Of the 29.8 million newly uninsured, 22.5 million people become uninsured as a result of eliminating the premium tax credits, the Medicaid expansion, and the individual mandate. The additional 7.3 million people become uninsured because of the near collapse of the nongroup insurance market.
Maybe these consequences are so politically intolerable that they’ll never come to pass. Maybe Republicans and Democrats really will come together, sing kumbayah, and replace the ACA.
But can you be sure? Repeal and delay lights a fuse that’s attached to a bomb. No one should be surprised if it explodes.
A co-worker struggling to make ends meet comes to you with a problem. The price of admission to a dear colleague’s retirement party at an upscale establishment is beyond her means, though not yours.
You both feel obliged to attend. She’d rather bring some refreshments to a conference room than spend what she cannot afford on a lavish event. You like the idea of a grand send-off for your retiring colleague. There can be only one party.
A similar, underrecognized conundrum arises in health insurance. Both you and your less fortunate co-worker are obliged under the law to obtain coverage (which you both want anyway).
But you differ in what you’d prefer to pay for. A high-level manager who makes, say, $200,000 per year is probably willing and able to paymore for health care than someone who makes $50,000.
Unfortunately, neither person really has a choice because the plans all cover “medically necessary” care, meaning any care that offers a clinical benefit. That includes lots of expensive and technologically sophisticated care that is no better, or only slightly better, than cheaper alternatives. You may be just fine paying for high-tech care of marginal value. For your colleague of more modest means, it’s a stretch.
Consider, for example, treating prostate cancer with proton-beam therapy. It’s more expensive than alternatives like intensity-modulated radiation therapy, but isn’t proven to be any better. If given the choice, many people — especially those with lower incomes — might rather buy health insurance plans that exclude high-cost, low-value treatments.
This one-size-fits-all approach to insurance coverage disproportionately hurts low-income people, many of whom might reasonably prefer to devote their scarce dollars to housing or their children’s education. To some extent, subsidies and other monetary adjustments can mitigate this problem. Medicare and Medicaid, for example, are financed in large part out of federal income taxes. And within the Affordable Care Act marketplaces, lower-income people receive subsidies that cover some of their costs.
People who receive coverage through their employers, however, don’t get that kind of help. Perversely, employer-sponsored health insurance is more highly subsidized for the rich than the poor. The subsidy comes in the form of an exclusion of health-insurance premiums from taxation. Since income tax rates are progressive — that is, the rich pay a higher rate of income tax than the poor — lower-income families get less of a benefit.
But both high-wage and low-wage workers at the same company are effectively forced into the same plans. To qualify for the tax exclusion, federal law requires that companies offer the same plans to all or most of their employees, with no consideration for the variable demand for health care. Employees then pay for their fringe benefits by taking home lower wages — and a flat, across-the-board cut in wages burdens low-wage workers disproportionately.
In theory, the labor market could adjust in ways that might lessen the problem: Low-income workers, for example, could demand higher wages for being forced into plans that are more expensive than they’d prefer. These would have to be made up by reducing the wages of high-income workers, something it’s not clear they would accept. There’s no evidence that labor markets actually work this way.
“The notion that labor markets perfectly offset the varying preferences for health insurance among workers by giving higher wages to those who value health insurance less is a comforting but crazy idea,” said Amitabh Chandra, a Harvard economist.
The problems with one-size-fits-all insurance run deeper. In some insurance markets, like those for small businesses in Massachusetts, employees across companies are pooled together and pay the same premium. A recent report from the Massachusetts attorney general showed that workers in companies that receive health care at less expensive hospitals effectively subsidize those at comparable companies who receive care at more expensive ones.
The uniformity of insurance plans also affects the pace and composition of technological innovation. To extend our party metaphor, if everyone — even those who preferred simpler events — were effectively forced to pay for any retirement party, regardless of how lavish, we’d see, and pay for, retirement parties of ever-escalating extravagance.
In much the same way, medical innovators respond to the size of the market for new technologies. The fact that health plans routinely cover all medically necessary care sends an “if you build it, we will pay for it” signal. Innovators are not getting the right signals, the right incentives, to develop high-value or cost-saving treatments. It’s more lucrative, instead, to develop pricey new therapies, even if they offer only marginal clinical benefits. The result is lots of new treatments that don’t provide much bang for the buck.
Those new treatments pose a continuing challenge to efforts to bend the cost curve. Economists have long known that technology is the primary driver of escalating health expenditures. Indeed, in 2012 medical technologies that were not offered a decade earlier accounted for almost a third of Medicare spending delivered by physicians and outpatient hospital departments.
What’s to be done? Managing technological innovation would require us to consider policy changes that would have been unthinkable a generation ago. “The tax code could be restructured to make extravagant health insurance less appealing,” Mr. Chandra suggested. Employers might then offer health plans that appealed more to low-income workers.
The Cadillac tax on expensive health plans, which is scheduled to go into effect in 2020, is a step in that direction, but according to Mr. Chandra doesn’t go far enough. And it is unpopular across the political spectrum. Other ideas — like incorporating cost-effectiveness criteria into Medicare and private plan coverage criteria — are sure to prompt disagreement.
Contentious solutions they may be. But as the cost of health care and health insurance rises, it won’t be a party for politicians feeling more pressure from consumers.
After campaigning for years on a plan of “repeal and replace Obamacare,” Republicans finally have the means within their grasp to make much of that possible. They control the presidency, the House, and the Senate. The filibuster still poses some potential threats to their plans, but it’s also within their means to abolish its widespread use in such a way that they could both repeal the Affordable Care Act and replace it with something of their own design.
What would that be? In contrast to what many say, there are Republican plans out there to consider.
Maybe the person working near you, the one who dragged himself to work and is now coughing and sneezing, couldn’t afford to stay home. Each week about 1.5 million Americans without paid sick leave go to work despite feeling ill. At least half of employees of restaurants and hospitals — two settings where disease is easily spread — go to work when they have a cold or the flu, according to a recent poll.
Shortly after his nomination as Treasury Secretary was announced, Steven Mnuchin went on television to offer some tax-reform principles that would guide the new administration:
Any reductions we have in upper income taxes will be offset by less deductions so that there will be no absolute tax cut for the upper class. There will be a big tax cut for the middle class, but any tax cuts we have for the upper class will be offset by less deductions that pay for it.
This appears to be false, at least if Republicans follow through on their plan to repeal-and-delay the ACA. Although the bill that’s under discussion would only gut the ACA in 2019, it would immediately cut taxes on the very rich.
How big is the tax break? According to CBO, the reconciliation bill that Republicans are using as a template would cut taxes by $623 billion over ten years. Of that, $123 billion would come from the repeal of the Medicare tax surcharge and $223 billion from the repeal of the tax on investment income.
That $346 billion represents about $1,000 for every man, woman, and child in the United States. Every cent will go into the pockets of people making more than $200,000 per year—the “upper class” that Mnuchin says won’t be getting any tax cuts. To my knowledge, there has been no discussion of offsetting those cuts in the repeal-and-delay bill. Maybe offsets will come later, but you’ll forgive me for some skepticism.
Also, where does this leave Republicans when they look to finance their version of health reform? If the party remains steadfast in its anti-tax commitment, the only options are savage budgets cuts or deficit spending. Neither option is appealing, which will complicate negotiations over any “terrific” replacement.
Look, Republicans won the election. They can go ahead and cut taxes if they want. But they don’t get a freebie just because this particular cut is part of a health-care bill. Call repeal-and-delay what it is: a big tax cut for the wealthy coupled with a vague commitment to Republican-style health reform in a couple of years.
In a recent Health Affairs article, the Commonwealth Foundation conducted their periodic survey of eleven countries to see how access issues might have improved or worsened. I’ve covered these data before. I talk about what’s changed, and why out-of-pocket costs are still a problem in my latest post over at the AcademyHealth blog. Go read it!
When I was a resident, I always marvelled at how on nights and weekends, remarkably fewer people were “required” to take care of patients than on weekdays during work hours. Logic dictated that either people must be receiving substandard care on off hours, because there were fewer personnel, or else perhaps those extra people weren’t necessary.
Was I right? To the research! This is Healthcare Triage News.
Diabetes is epidemic and many diabetics lose their vision. Diabetic retinopathy is treatable, so it makes sense to screen for it. But screening doesn’t happen as much as it should, in part because there aren’t enough people who can do it. So it would be great to have an inexpensive and accurate screening procedure for diabetic retinopathy.
In JAMA, Varun Gulshan and colleagues report on a neural network that can identify diabetic retinopathy from retinal images. They used a deep learning algorithm that taught itself to identify retinopathy by analyzing a lot of images that had been previously classified as cases or non-cases by experts. Then they tested the algorithm on fresh images and found that it accurately identified retinopathy. This may be the beginning of an important change in medicine.
Objective To apply deep learning to create an algorithm for automated detection of diabetic retinopathy and diabetic macular edema in retinal fundus photographs.
Design and Setting A specific type of neural network optimized for image classification called a deep convolutional neural network was trained using a retrospective development data set of 128 175 retinal images, which were graded 3 to 7 times for diabetic retinopathy, diabetic macular edema, and image gradability by a panel of 54 US licensed ophthalmologists and ophthalmology senior residents between May and December 2015. The resultant algorithm was validated in January and February 2016 using 2 separate data sets, both graded by at least 7 US board-certified ophthalmologists with high intragrader consistency.
Exposure Deep learning–trained algorithm.
Main Outcomes and Measures The sensitivity and specificity of the algorithm for detecting referable diabetic retinopathy (RDR), defined as moderate and worse diabetic retinopathy, referable diabetic macular edema, or both, were generated based on the reference standard of the majority decision of the ophthalmologist panel…
Results* The EyePACS-1 data set consisted of 9963 images from 4997 patients…; (prevalence of RDR = 7.8%); the Messidor-2 data set had 1748 images from 874 patients (prevalence of RDR = 14.6%]). For detecting RDR, the algorithm had an area under the receiver operating curve [AUC] of 0.991 (95% CI, 0.988-0.993) for EyePACS-1 and 0.990 (95% CI, 0.986-0.995) for Messidor-2. Using the first operating cut point with high specificity, for EyePACS-1, the sensitivity was 90.3% (95% CI, 87.5%-92.7%) and the specificity was 98.1% (95% CI, 97.8%-98.5%). For Messidor-2, the sensitivity was 87.0% (95% CI, 81.1%-91.0%) and the specificity was 98.5% (95% CI, 97.7%-99.1%). Using a second operating point with high sensitivity in the development set, for EyePACS-1 the sensitivity was 97.5% and specificity was 93.4% and for Messidor-2 the sensitivity was 96.1% and specificity was 93.9%.
This, people, is what accurate classification looks like. It’s biometric porn.
A recent Cochrane Review summarized evidence for another method for detecting diabetic macular edema (one of the conditions Gulshan et al. screened for) and found that
In nine studies providing data on CSMO (759 participants, 1303 eyes), pooled sensitivity was 0.78 (95% confidence interval (CI) 0.72 to 0.83) and specificity was 0.86 (95% CI 0.76 to 0.93).
Gulshan et al.’s neural network was substantially more accurate than that. However, the Results do not mean that the neural network is ready for deployment. JAMA published several commentaries, by people better informed than me, which discuss reasons why this technology may need further evaluation.
Let’s stipulate, however, that screening for diabetic retinopathy by a machine learning algorithm is going to work. (And if it does, it’s likely that machine learning algorithms will also succeed at other important diagnostic tasks.) What will that mean for medicine?
First, the machine learners will screen better than humans do. The authors argue that
This automated system for the detection of diabetic retinopathy offers several advantages, including consistency of interpretation (because a machine will make the same prediction on a specific image every time), high sensitivity and specificity, and near instantaneous reporting of results.
What happens to total health care spending? Automation makes things cheaper, right? Beam and Kohane observe that
Once a [neural network] has been “trained,” it can be deployed on a relatively modest budget. Deep learning uses a… graphics processing unit costing approximately $1000… [that] can process about 3000 images per second… This translates to an image processing capacity of almost 260 million images per day (because these devices can work around the clock), all for the cost of approximately $1000.
Does this mean that screening by neural network will reduce health care spending? That’s unlikely. If the price of an accurate screen for diabetic retinopathy falls, we will screen more. If we screen more, we’ll uncover lots of retinopathy. Finding more people who can’t see will increase the demand for expensive ophthalmalogic treatments. Therefore, unless we also find a way to make those treatments less costly — or find a cheap way to prevent diabetes — automated screening will likely increase total health care spending.
That’s not necessarily a bad thing. I’d rather have a world in which we spend more on health care but many fewer people lose their vision.
And what happens to radiologists in a world of cheap, quick, and accurate neural network screeners? Saurabh Jha and Eric Topol think radiologists won’t necessarily disappear.
The primary purpose of radiologists is the provision of medical information; the image is only a means to information. Radiologists are more aptly considered “information specialists” specializing in medical imaging.
What needs to be added to a screening result to transform it into medical information? One thing is an interpretation of the clinical significance of the finding. Right now, radiologists (and other physicians) can make these interpretations, while computers cannot. So if screening gets cheap and becomes more common, the demand for interpretation increases. Clinicians therefore remain in demand, so long as they are able to add value to the screening result.
Which is to say, until machines also learn to judge the clinical significance of findings better than physicians can.
Finally, automated screening could also improve access to global health care. According to WHO,
the number of adults living with diabetes has almost quadrupled since 1980 to 422 million adults. This dramatic rise is largely due to the rise in type 2 diabetes and factors driving it include overweight and obesity.
Moreover, these factors driving the diabetes epidemic are increasing. But 85% of the world makes $20 USD / day or less. There is no serious near-term prospect of deploying physicians to screen the vision of most of humanity. However, artificially-intelligent systems served by less expensive staff could be deployed globally.** Thus, machine learning could screen many populations that currently have little or no access to specialized care.
The catch, of course, is that it would be pointless to screen the other 85% unless we were prepared to also offer them treatment. The development of cheap, quick, and accurate will confront us with the question of whether we believe in one global standard of health care.
*To interpret the Results it is helpful to know that:
Sensitivity = Proportion of Cases Correctly Identified (range 0.0 to 1.0).
Specificity = Proportion of Non-Cases Correctly Identified (range 0.0 to 1.0).
Area Under the Receiver Operating Curve [AUC] is a statistic that summarizes information about Sensitivity and Specificity (range 0.5 to 1.0).
**Machine learners get better with practice. Deploying the neural network globally would accelerate the rate at which they learn by giving them access to many more images.
The big risk of repeal-and-delay (well, one big risk) is that the individual insurance market will unravel before repeal takes effect. As Robert Laszewski tartly noted, “Republicans are being awfully naïve. They seem to be ignoring the risks in the transition period, particularly because they need insurance companies to provide insurance during the transition.”
Well, not all Republicans. Christopher Condeluci, former counsel to the Senate Finance Committee and respected Republican health-care wonk, offers some insight:
I recognize this will spur questions, many of which I am unable to answer at this point, but I wanted to clarify the following: Republicans (Congressional Republicans and the Trump Administration) have certainly argued that the individual market is deteriorating, but Republicans have not announced that they want to let the individual market die. Actually, Republicans recognize that they have to fix the markets. As a result, Republicans are currently determining what steps need to be taken to improve the regulatory environment. In addition, discussions are ongoing regarding the funding for the cost-sharing subsidies, as well as payments under the ACA’s risk stabilization programs.
Some of the issues being discussed: requiring pre-verification for [special enrollment periods, or SEPs], rigorous enforcement of SEP enrollment, changing the 90-day grace period in cases where a policyholder fails to pay their premiums, modifying the age bands to 5 to 1, and fixing the risk adjustment formula (note, I recognize there are other important issues to be discussed). Republicans also recognize the insurance carrier’s desire to fund the cost-sharing subsidies, but there are political and procedural issues that need to be worked through here. While it is highly unlikely that payments under the risk corridor program will be provided to carriers, the conflict over the reinsurance payments could be resolved in a way that works for both the carriers and Republicans.
So again, Republicans recognize there is a problem, they want to fix the problem, and they are currently working to figure out what can be done, and balancing the “must dos” to stabilize the markets with the prevailing political concerns that will no doubt come from certain corners of the Republican Conference.
In the end, will it be enough? That is, if Republicans fund the cost-sharing subsidies, provide a majority of the reinsurance payments for, for example, 2016, and fix the current regulatory environment, will insurance carriers stay in the market? It’s unclear at this point. But I—like many, regardless of whether you are a Republican or a Democrat—am hopeful that the markets will be improved in short order.
Austin and Aaron are participants in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to amazon.com.