• There’s No Magical Savings in Showing Prices to Doctors

      06/15/2017
      Aaron Carroll

    The following originally appeared on The Upshot (copyright 2017, The New York Times Company).

    Physicians are often unaware of the cost of a test, drug or scan that they order for their patients. If they were better informed, would they make different choices?

    Evidence shows that while this idea might make sense in theory, it doesn’t seem to bear out in practice.

    A recent study published in JAMA Internal Medicine involved almost 100,000 patients, more than 140,000 hospital admissions and a random distribution of laboratory tests. During the electronic ordering process, half the tests were presented to doctors alongside fees. While the cost to the patient might vary, these Medicare-allowable fees were what was reimbursed to the hospital for the test or tests being considered. The other half of the tests were presented without such data.

    The researchers suspected that in the group seeing the prices, there would be a decrease in the number of tests ordered each day per patient, and that spending on these tests would go down. This didn’t happen. Over the course of a year, there were no meaningful or consistent changes in ordering by the doctors; revealing the prices didn’t change what they did much at all.

    This isn’t the first time a study like this found that showing prices to doctors doesn’t make a difference. Earlier this year, a study published in Pediatricsreported on a similar randomized controlled trial on physicians caring for children. In this case, doctors were randomized to one of three groups. The first group saw the median price of a test when they ordered it. The second saw both the price (often lower) when obtained within the current health care system and outside it. The third group saw no price at all.

    Pediatric-focused clinicians showed no effect from price displays. Adult-focused clinicians actually ordered more tests when they saw the prices.

    A similarly designed study of more than 1,200 clinicians in an accountable care organization published earlier this year also found no effects from telling physicians prices.

    Some older studies have found that physicians might alter their behavior on individual tests, but in only five of the 27 they examined. Another found a small, but statistically significant, difference. Unfortunately, this study suffered from asymmetric randomization. Even before the intervention began, the tests chosen for the price-showing group were ordered more than three times as much as those chosen for the control group. More expensive tests appeared in the control group for some reason as well.

    Of course, any one study has the potential to be an outlier or subject to limitations that might warrant skepticism. These can be minimized by looking at the body of evidence in systematic reviews.

    One was published in 2015, and argued that in the majority of studies, giving physicians price information changes their ordering and prescribing behavior to lower the cost of care. A closer look, though, reveals that most of the studies in this analysis were more than a decade old. Many took place in other countries. And all were published before these latest, and largest, studies I discussed above. Another systematic review that looked at interventions focusing only on drug ordering found similar results, with similar caveats.

    I should be clear: We have good reason to want to believe that interventions focusing on giving physicians information about the prices of the things they order should make a difference. In 2007, a systematic review demonstrated that doctors were ignorant of the costs of prescription drugs. They underestimated the prices of expensive drugs, overestimated the prices of inexpensive ones, and did not understand the extent of the difference in price between those considered cheap and those considered pricey. Another, published in 2015, explored 79 studies, 14 of which were randomized controlled trials, that suggested that physicians could be educated to deliver “high-value, cost-conscious care.”

    But that education probably needs to be holistic. Flashing one point of data at a doctor does not get the job done; knowledge transmission needs to be accompanied by what this review called “reflective practice and a supportive environment.” Simply focusing on cost information may not be enough. The reasons that physicians order tests are more than financial, and efforts to influence their behavior most likely need to be more than informational.

    Additionally, it may be that issues of price transparency need to involve more than one component of the health care system. While focusing solely on physicians, or on patients, might not work well, trying to work on both simultaneously might. It’s also possible that intervening solely on one procedure, test or drug at a time may not be as powerful as trying to influence spending on care over all.

    Finally, trying to make physicians focus strictly on cost may be off base as well. Some care, even more expensive care, is worth it. What we really should attend to is value — the quality and impact relative to the cost. It is certainly harder to determine value than price, but that metric might make more of a difference to physicians, and to their patients.

    @aaronecarroll

     
      Health Policy
      costs, spending, transparency, Upshot
    item.php

  • Amgen’s Money-Back Guarantee: Simply Too Good To Be True

      05/09/2017
      guest contributor

    The following is a guest post by Rodney Hayward, Director of the Robert Wood Johnson Foundation Clinical Scholars, and the National Clinician Scholars at IHPI. He’s a Professor of Medicine and Public Health at the University of Michigan. Follow him on Twitter at @ProfHayward.

    Multiple small clinical trials suggested that PCSK9is (a new class of cholesterol-lowering medications) might be the next super drug-class. These early trials suggested that PCSK9is dramatically decreased cardiovascular disease (CVD) events (strokes and heart attacks) in those who had very high CVD risk and could not take statins (the current standard for treating cholesterol).

    But most people with very high CVD risk tolerate statins just fine, greatly limiting the market for PCSK9is. A potential solution was to prove that the drugs worked better even for those on statins already. Unfortunately for the main PCSK9i manufacturer (Amgen), the first large randomized trial examining PCKS9-abs benefits in those on statins was highly disappointing. In the much larger market of high CVD risk people on statins, the relative reduction in non-fatal CVD events was much less than expected (25%) with no effect on mortality.

    Although no one wants to have a heart attack or stroke, the lack of a mortality benefit – and the ~$14,500/yr price tag – led most to conclude that the expense was not worth the gain. Some pundits called for a price cut, but last week Amgen came up with a better idea – a full Money-Back Guarantee!

    What A Deal!
    In response to their falling stock price, Amgen announced last week that they would refund the cost of the drug in full if an individual taking it regularly has a major CVD event. This was portrayed by some media sources as an advance for pay-for-performance, and it certainly makes great intuitive sense. Yes, it may be a small consolation if the PCSK9i didn’t prevent your stroke, but “you get your money back if the medication doesn’t work for you.”

    The problem is that the italicized sales pitch above is completely untrue.

    What A Scam!
    Modern psychology has demonstrated that we humans are terrible intuitive statisticians and that it’s quite easy to fool us with statistical illusions. So let’s briefly go through the math to understand why Amgen’s sales pitch is untrue.

    A Best-case Scenario: A sizable group (though a small proportion of all adults) continues to be at very high CVD risk (~5%  5-year risk) after taking a statin, a daily aspirin, and blood pressure medications as indicated (cost ≈ $30-$150/yr). If you give a PCSK9i to 1000 such individuals, current evidence suggests that you would prevent, on average, 12 to 13 non-fatal CVD events over a five year period (0.05 [5-year risk] * 0.25 [PCKS9i’s relative risk reduction] = 0.0125 = 12.5 in 1000) at a cost of ~$72.5 million. We know that more than 30% of these events will be quite minor, but let’s ignore that so as to not be accused of nitpicking. Over 5 years, Amgen would need to return ~$1.4 million to ~37.5 of the 1000 people (2.5yrs [average years to CVD event in the 5-year period] * $14,500/yr * 37.5 [number of people who had a CVD event) ≈ $1.4 million).

    But if only 12 to 13 people in 1000 benefited and only 37 to 38 people received refunds, who are the remaining 950 people who did not benefit? They are the 95% of people who were not destined to have a heart attack or stroke regardless.

    So the claim, “If the PCSK9i doesn’t work for you, you get your money back!” is a complete scam.  Over 98% (950 of ~962.5) of those who did not have a stroke or heart attack over 5-years received no benefit from the PCSK9i because they were not destined to have a CVD event regardless of treatment.

    Ironically, the less likely a person is to benefit from the PCSK9i (ie, the lower their CVD risk), the less likely Amgen will need to refund their money.

    So let’s restate Amgen’s Money-Back guarantee more accurately.

    • Individual Perspective: If you are a very high CVD risk person on a statin and you take a PCSK9i for five years, you have about a 1.25% chance of avoiding a non-fatal heart attack or stroke at a cost of ~$70,000, about a 3.75% chance of having a stroke or heart attack and getting your money back, and about a 95% chance of paying ~$70,000 and receiving no benefit.
    • Payer’s Perspective: For every 1000 high CVD risk beneficiaries on statins, paying for a PCSK9i for 5 years will prevent 12-13 non-fatal heart attacks or strokes, on average, and Amgen has generously offered to reduce the price to you by ~ 3.75%, from ~$72.5 million to ~$70 million.

    @ProfHayward 

     
      Health Policy
      costs, drugs, guarantee, pharma
    item.php

  • Healthcare Triage: Orphan Drug Costs

      04/17/2017
      Aaron Carroll

    Last week, we argued that the benefits of the Orphan Drug Act are more elusive than commonly assumed. But what are its costs? That’s the topic of this week’s Healthcare Triage.

    Sources can be found in the posts Nick has written on the subject! Special thanks to him!

    @aaronecarroll

     
      Health Policy
      costs, Healthcare Triage, orphan drugs
    item.php

  • AcademyHealth: How should we pay for gene therapy?

      05/27/2016
      Aaron Carroll

    For many, many years we’ve been hearing about gene therapy – the chance that we can get into people’s DNA and fix it to resolve problems and fix disease. In a recent piece in Science, Stuart Orkin and Philip Reilly discuss what finally achieving success might mean:

    Go read about that, and what I think about it, over at the AcademyHealth blog! Go!

    @aaronecarroll

     
      Health Policy
      AcademyHealth, costs, drugs, gene therapy
    item.php

  • AcademyHealth: Now people have coverage. How well is it serving them?

      05/05/2016
      Aaron Carroll

    Insurance is just a first step towards improving access. Also important is whether that insurance helps people to get the care that they need. In a recent report from the Urban Institute, John Holahan, Michael Karpman, and Stephen Zuckerman probed that question.

    Go read my thoughts on their work in my latest post over at the AcademyHealth blog. Seriously, go read it!

    @aaronecarroll

     
      Health Policy
      AcadcemyHealth, access, choice, costs, insurance exchange, Medicaid expansion
    item.php

  • What are the health care costs of obesity?

      07/15/2015
      Aaron Carroll

    “Savings in Medical Expenditures Associated with Reductions in Body Mass Index Among US Adults with Obesity, by Diabetes Status“:

    Background: The prevalence of obesity has more than doubled in the USA in the past 30 years. Obesity is a significant risk factor for diabetes, cardiovascular disease, and other clinically significant co-morbidities. This paper estimates the medical care cost savings that can be achieved from a given amount of weight loss by people with different starting values of body mass index (BMI), for those with and without diabetes. This information is an important input into analyses of the cost effectiveness of obesity treatments and prevention programs.

    Methods: Two-part models of instrumental variables were estimated using data from the Medical Expenditure Panel Survey (MEPS) for 2000–2010. Models were estimated for all adults as well as separately for those with and without diabetes. We calculated the causal impact of changes in BMI on medical care expenditures, cost savings for specific changes in BMI (5, 10, 15, and 20 %) from starting BMI levels ranging from 30 to 45 kg/m2, as well as the total excess medical care expenditures caused by obesity.

    This study was an attempt to use MEPS data to see how much losing weight might save people in health care costs, both with and without diabetes. The topline results showed that, on average, obesity raised health care costs by just more than $3500 per person. That means that obesity cost the US more than $315 billion a year.

    But the paper goes further, and it’s important to dig down. The costs are NOT equally spread over obese individuals. People with class 1 obesity, or those whose BMI is greater than 30 but less than 35, pretty much have no elevated health care costs. When we say lots of people in the US are obese or overweight, many of them still have BMIs less than 35.

    The paper further reports that a person who has a starting BMI of 40, and can lose 5% of their weight, might expect to see reductions in health care costs of $2137. But only about 6% of adults have a BMI that high. Losing 5% of weight if you have a starting BMI of 35 would save you $528. Losing that weight if you’re starting with a BMI of 30 would save you $69.

    The savings also depend on whether you have diabetes. For instance, the savings that might be achieved by a 5% weight loss for a person who has  BMI of 40 is $2123 if they have type 2 diabetes, but only $643 if they don’t.

    My point is this: Talking about the “costs of obesity” in the aggregate makes it sound like if everyone in the US would just lose a little weight, we’d save massive money. But it’s not that simple. The vast majority of the costs of obesity are located in a relatively small group of people, those who are extremely obese and have diabetes.

    If our goal is to reduce the cost to the US for obesity, we should likely focus on that small group of people, who are also likely the hardest to reach and influence. Trying to get lots of people who have a BMI of 30-35 to lose some weight might make them feel better, but won’t influence health care spending in the US much at all.

    @aaronecarroll

     
      Health Policy
      costs, diabetes, obesity, spending
    item.php

  • Newly eligible for Medicaid spending more than we hoped – ctd.

      07/13/2015
      Aaron Carroll

    I have obviously touched a nerve in many TIE readers. So let’s take a step back and revisit what I wrote and what is going on here.

    There were some studies that made predictions about the health and composition of the newly eligible for Medicaid. Those studies could be interpreted to mean the Medicaid expansion would cost less than predicted, because people entering would be healthier than those already enrolled. I WAS ONE OF THE PEOPLE WHO THOUGHT THAT INTERPRETATION MIGHT BE POSSIBLE. A recent report from CMS suggests that this lower-than-expected expense didn’t happen. I wrote that we need to figure out why and keep watching.

    None of this was to suggest that the prior research was incorrect or poorly performed. It appears, upon further looking at the report, that there are some reasons that the differences are higher than estimated:

    There are several explanations for the difference between the estimates in this year’s report and those in previous reports. First, most of the States that implemented the eligibility expansion are covering newly eligible adults in Medicaid managed care programs, and on average the capitation rates for the newly eligible adult enrollees were significantly greater than the projected average costs previously calculated.

    Austin flagged this, by the way. In fact, some of the capitation rates may have been raised because states predicted pent-up demand because they assumed people entering might be sicker (in spite of the research I talked about).

    And there’s this:

    Data for newly eligible adults are still limited. While CMS has reported some enrollment and expenditure data for this group, data on claims and managed care encounters, along with data on the health status and demographics of these enrollees, are not yet available. Thus, there is still considerable uncertainty about the health care costs of newly eligible adults in 2014, as well as for future years.

    Medicaid still expects newly eligible adults to cost less in the future (Figure 6). It’s entirely possible they are correct.

    But here’s the thing. It may be that this research turns out to be incorrect. I doubt it, in that I think the work is correct in that the newly eligible appear to be less sick than those already covered. It may be that this research is correct, and that still spending goes up for other reasons. Maybe people just spend more. Maybe docs find new ways to bill more. Maybe policy makers set capitation rates too high.* But if I report on research, and make some guesses as to the future, and they turn out to be incorrect, I will write that my predictions were wrong, that we need to figure out why, and that we should keep looking into it.

    That’s what I did. That’s what I will continue to do.

    @aaronecarroll

    *This appears to be the best hope for an answer right now – but we have to make sure.

     
      Health Policy
      costs, Medicaid, Medicaid expansion, spending
    item.php

  • Vlogbrothers: Is Obamacare Succeeding

      03/26/2015
      Aaron Carroll

    Almost two years ago, John Green did a Vlogbrothers video on why health care costs so much in the United States. It relied heavily on TIE’s series on the same topic, and if you haven’t seen it, then you’re odd, because it’s had almost 6.5 million views:

    Needless to say, this video – more than any other thing – led to Healthcare Triage being possible. Anyway, this week, John did a video on “whether Obamacare is working” 5 years later, and it also relies on some Upshot and other TIE-related material. It’s worth a watch as well:

    @aaronecarroll

     
      Health Policy
      ACA, costs, Healthcare Triage, Obamacare, Vlogbrothers
    item.php

  • Let me know when to start panicking about health care spending again

      04/21/2014
      Aaron Carroll

    For some time, we’ve seen a slowdown in health care spending. Some argued that this was due to the ACA. Others thought it was the slowed-down economy. I’ve been one of the people happy to say “I don’t know”. Regardless, per Jonathan Cohn, things are looking bad again:

    Now the respite may be ending. You can see it in the latest monthly reports from the Altarum Institute, an Ann Arbor-based think-tank that monitors national health care spending. These reports, based on government data, are the equivalent of an early warning system for medical costs. They are one of the first places a spending spike would show up. According to Altarum, expenditures started to rise more quickly in the middle of 2013. Since then, the rate has gone up even more. 

    He also has a chart to ruin your day:

    altarum_cost_graph

    As always, past performance is not a guarantee of future results. But we’ve all benefited from reduced spending growth in the last few years. If that changes, things could get messy fast.

    Read Jon’s whole piece.

    @aaronecarroll

     

     
      Health Policy
      ACA, costs, economy, spending
    item.php

  • If a 10 cent coupon incentive works, is anyone surprised that $120 does?

      04/14/2014
      Aaron Carroll

    Great editorial today in BloombergView on Medicare’s payment practices for drug administration:

    Here’s how the system works: When a doctor administers a drug in his or her office, Medicare pays 106 percent of its average selling price. The doctor keeps the extra as compensation for administering the injection.

    What has this got to do with eye doctors? The drug Lucentis, used to treat macular degeneration, cost Medicare almost $2,000 a shot in 2012. Another drug, Avastin, which works just as well, costs about $50. If you were the doctor, faced with a system that pays you 6 percent of the drug’s cost, which would you choose? That Medicare spent a total of about $1 billion on Lucentis in 2012 suggests most ophthalmologists went with the more expensive one.

    Get that? You’re an ophthalmologist. You see a patient with macular degeneration. You have two options. One pays you $3. The other pays you $120. Which are you going to choose?

    Recognize that we’re not talking about costs to the patients. We’re talking about payments to doctors. Medicare is going to cover the cost of the drug no matter what (which is also crazy, since one costs them 3900% more than the other). We’re talking about the “profit” the physician or practice makes. They get $3 or $120. Do we really think that doctors might not be influenced by this?

    Medicare spent about a billion dollars on Lucentis in 2012. It’s not alone:

    This problem goes beyond a single drug. Of the $20 billion Medicare spent on drugs administered by doctors in 2010, 85 percent went to the 55 most expensive ones. In what seems unlikely to be a coincidence, 42 of those drugs also showed an increase in use from 2008 to 2010.

    So why is it this way?

    The Centers for Medicare & Medicaid Services, the agency that runs Medicare, says it’s required to pay for treatment that a doctor deems medically necessary, and it lacks the authority to direct treatment based on cost. All Medicare can do to control costs is tell doctors the price of what they’re prescribing, as well as the alternatives. Which is to say, almost nothing.

    Evidently, the administration’s latest budget request dropped the fee paid to doctors from 6% to 3%. In other words, it would change the math on this to $1.50 for Avastin and $60 for Lucentis. Think this will fix it?

    We could set a fixed amount for drug administration, period. That would end the incentive for choosing the more expensive drug. Or, we could change the whole model, to only reimburse the cost of the cheapest drug if they’re equivalent. I’m not holding my breath.

    @aaronecarroll

    UPDATE: Loren Adler informs me that there are proposals to set the administrative payment to a flat fee. Here’s one. Here’s another.

     
      Health Policy
      costs, drugs, Medicare, spending
    item.php