• All the premium-wage trade-off literature

    In light of the Cadillac tax kerfuffle, it seems like I should carefully go through all the (important) econ literature on the premium-wage trade-off.* Sure, I’ve posted about about it several times, and those posts cite some studies (Sarah Kliff’s recent post cites some of them too). But I have never systematically tried to make sense of all of them.

    I don’t have time to do so quickly, but it’ll be quite a while before the Cadillac tax is repealed or modified, so slow and steady is just fine. First question: Does the list below include all the important literature? If you think I’m overlooking something worthwhile, tell me. Note that the first two entries below review literature through the mid-1990s, so I will make my life easier and not reexamine all that they cover.

    Another round up of the literature, organized by finding, is here.

    My plan: First, I will identify which studies examine public sector employees (teachers, government workers). In that setting, we should not expect a 1:1 premium-wage trade-off because the public sector is not in competition to the same extent as private firms. Though this has been pointed out to me several times, I’ve yet to see it mentioned in a study’s discussion. I’m curious to see if I can find a reference.

    Next, I will highlight the few papers that do find a 1:1 trade-off. I know there are at least two (Baicker and Chandra 2006, Kolstad and Kowalski 2012). Are there others? Then, I will try to comment on why the others do not find a 1:1 trade-off where it is expected. This is probably for methodological or data reasons, though perhaps there are settings in which competition isn’t sufficient to produce it (?). I would hope the paper authors discuss these issues.

    Finally, I will decide whether it’s important that there’s no examination of a setting in which premiums actually go down (if true). It may not be important because they may still not go down under the Cadillac tax, but just rise less quickly. In any case, is there a good reason to expect asymmetry? I’m not aware of one.

    * The idea that workers pay their full employer-sponsored health insurance premiums from lower wages. When premiums go up, wages go down and vice versa. What’s held steady, all other things being equal, is total compensation.


    Comments closed
  • Why are vaccine prices so low? (continued)

    Why are (private) vaccine prices far below the prices for drugs that are much less effective?

    When I first posed and attempted to answer this question last week I said I had some more reading to do. I’ve now done it, but I didn’t find a complete answer — nothing more substantial than what I provided in my post. As best I can tell, this is not a question fully addressed in the scholarly literature. (If you know otherwise, point me to a source.)

    I did learn a few thing from my additional reading, however:

    • Though childhood vaccination may be cost saving — and is certainly cost-effective — newer vaccines for adolescents are not all as cost-effective and not cost saving.
    • The number of vaccine suppliers in the US has dwindled over the years, from 26 in 1967 to 12 in 2002. Of the eight childhood vaccines recommended in 2005, five had only one source. This contributes to supply shortages and suggests that the prices of vaccines are too low. Concentrated demand from government programs, which purchases over half of vaccine doses at discounted prices, explains the low level of competition in the market.
    • On the one hand, vaccine markets are limited in size to the number of kids. They’re not like maintenance drugs, taken regularly for years, and their target population cannot be expanded (as, for example, can cholesterol-lowering drugs through guideline changes). Due to licensing regulations and varying guidelines, vaccine markets tend also to be limited geographically, with different suppliers in different countries. Lower prices outside the US do not offer market-expanding opportunities for US-focused manufacturers. On the other hand, vaccine market size is highly predictable. The revenue constraint is price.
    • Vaccines represent 1.5% of all pharmaceutical revenue.
    • Relatively low profit potential, as compared to other drug products, limits investment in vaccines. The concern about potential litigation if a vaccine (or batch of one) should cause harm may also deter investment.
    • Apparently, employers, who manage most of the structure of private health insurance benefits for the vaccine-relevant population, will not pay for vaccines at their social value. Or, perhaps its the consumers (workers) who won’t do so.
    • One explanation for low flu vaccine prices is the prospect that, should a shortage arise, the government will direct limited supply toward the highest risk patients. Therefore, a purchaser willing to pay more in response to insufficient supply (or in anticipation of it) is thwarted. The price signal is broken.

    As you can see from these facts and ideas, many others have identified issues in vaccine markets stemming from low prices (e.g., too little entry, supply disruptions). It’s well understood that government programs play a role in reducing (government) prices and concentrating demand. But why are private prices low too?

    The government backstop is my best answer. If providers of vaccines raise private prices too high, insurers may stop covering them; their enrollees who cannot afford them could obtain them through government programs. The Affordable Care Act breaks this logic, since vaccines must now be covered. Unless my explanation is wrong, we should see their private prices rise.


    Comments closed
  • People are bad at choosing heath plans, and they don’t even know it

    The following originally appeared on The Upshot (copyright 2015, The New York Times Company).

    It’s open enrollment season for almost every kind of health insurance in America. Millions of Americans using Medicare plans, employer-sponsored health insurance or Affordable Care Act marketplaces select health plans each fall. Many consumers face numerous options, and research shows that they make many mistakes, often paying more than they need to.

    Some err by selecting deductibles that are too low. Lower deductibles can be a fine choice for some consumers, but trying to save money with a lower deductible can be a poor choice if a person pays even more in premiums. For instance, at one large American company in 2010, employees could reduce their deductible by $250 — to $750 from $1,000 — by paying $500 more in premiums. Trading $500 for $250 is clearly a bad deal for the consumer.

    Yet a majority of the firm’s workers made bad deals like this, according to a study by Saurabh Bhargava, a Carnegie Mellon economist, and his colleagues. Workers were offered a choice of 48 plans that were identical except in cost sharing and premiums. Though no plan would have been optimal for every employee, a $1,000 deductible plan would have been better for many and at least a no-worse choice for 97 percent of employees who chose a lower deductible.

    People make mistakes like this for a variety of reasons. Some don’t understand basic health insurance concepts. In an experiment accompanying Mr. Bhargava’s study, 71 percent of people couldn’t identify fundamental cost-sharing features of health insurance plans. This type of illiteracy was highly predictive of mistakes like overpaying for a lower deductible.

    Another study, led by George Lowenstein, a professor of economics and psychology at Carnegie Mellon, found that people misunderstood plan features and costs. Even with plan details right in front of them, only 40 percent of privately insured Americans could identify how much they’d have to pay for an M.R.I. scan. Only 11 percent could report what a four-day hospital stay would cost them. Yet study subjects were overconfident. All said they understood what a “co-pay” was, but 28 percent could not correctly answer a question testing their understanding of the term; only 7 percent would admit to not knowing what “maximum out-of-pocket” meant, but 41 percent couldn’t define it.

    Another study found that less than a third of respondents could correctly answer questions about coverage features of their own plan. Yet another found that only a minority of workers at a large firm could answer questions about plan characteristics or their own, recent health care spending.

    Without a doubt, choosing a plan can be daunting. A shopper in the Affordable Care Act marketplace can choose from 40 plans, on average. A typical Medicare beneficiary can choose from among nearly 20 Medicare Advantage plans and 30 stand-alone prescription drug plans.

    In selecting plans, consumers are prone to mental shortcuts that often lead to poor choices. Plan labels — like the “gold,” “silver” or “bronze” — can fool people. To some, “gold” sounds better than “bronze,” even if it isn’t. In one study, people were asked to select hypothetical plans with these labels, but the researchers reversed the meaning of “gold” and “bronze” for half of them. It didn’t matter. Most people picked “gold” anyway.

    The ordering of choices also matters. Consumers tend to select plans near the top of a list, a phenomenon that arises in other contexts: Economists download more papers from the tops of lists of new studies, as my colleague Neil Irwin reported; politicians at the top of ballots receive more votes.

    Eric Johnson, a Columbia business professor, led a study that found that without substantial additional assistance, a consumer’s likelihood of selecting the lowest-cost plan is no better than chance. The researchers conducted a series of experiments on people similar to those who would shop for marketplace coverage. Each study participant was asked to presume he’d use a certain amount of health care and, based on that, to choose the lowest-cost plan from among eight choices, which varied by premium, doctor co-pay and deductible. Only 21 percent could accomplish this task, a figure not statistically different from chance. The annual cost of errors was about $250.

    A separate analysis showed that participants had a stronger aversion to an increase in costs in deductible or co-pay than to the same increase in premium. Because a dollar is a dollar, no matter how you spend it, this is another indication of irrational decision making.

    But when study subjects were provided with a tutorial or with a calculator that revealed the full cost of each plan, or if they were placed in the lowest-cost plan by default (from which they could voluntarily switch), their chance of selecting the cheapest plan was much higher, upward of 75 percent in some experiments.

    Though some Medicare beneficiaries switch to lower-cost drug plans over time, another way consumers get stuck with bad deals is by staying in plans as their premiums increase, a status quo bias. One study found that New Jersey enrollees in Medicare prescription drug plans paid an average of $536 more over three years because of this kind of inertia. Some insurers strategically enter markets with low prices and increase them over time, exploiting consumers’ inertia. This “invest then harvest” pricing strategy has been observed in markets for Medicare Advantage plans, commercial health insurance and others.

    Providing consumers with easier access to cost comparison information can help. A study published in The Journal of Economics in 2012 found that when a random sample of Medicare beneficiaries got letters that compared prescription drug plan costs, they were more likely to switch plans and to save money, relative to nonrecipients. When pharmacy students helped California Medicare beneficiaries understand drug plan costs, 60 percent switched plans.

    Few consumers get this much help. When researchers at the University of Pennsylvania examined the Obamacare online marketplaces last year, they found only a few that provided some of the tools consumers need. Most marketplaces presented plans in order of the cost of the premium, which doesn’t take other cost sharing into account. (However, California ranked plans according to total cost, Kentucky listed them randomly, and Minnesota ranked them based on best match according to a series of preference questions, similar in spirit to an approach recommended by University of California, Berkeley economists in a recent Brookings policy paper.)

    Only three states offered cost estimators. The federal government’s site, HealthCare.gov, will offer more information about plans — like which physicians are in plan networks — and cost comparison tools.

    If last year is any guide, once again few consumers will actively shop for a more beneficial plan. An analysis by the Department of Health and Human Services showed that more than 70 percent could have found a cheaper plan. The research is clear: Most won’t find that cheaper plan without a great deal more help.


    Comments closed
  • AJMC: Implications of state decisions to expand or not expand Medicaid

    Earlier this fall, I participated in a series of discussions hosted by the American Journal of Managed Care about health reform and the changing health insurance and delivery landscape. The video below is one exchange from the series, focused on implications of state decisions to expand Medicaid, or not.

    I was joined by

    • Leah Binder, President and CEO of The Leapfrog Group
    • Margaret O’Kane, President of the National Committee for Quality Assurance
    • Matt Salo, Executive Director of the National Association of Medicaid Directors
    • Dennis Scanlon (moderator), Professor of Health Policy and Administration and Director of the Center for Healthcare and Policy Research, College of Health and Human Development, The Pennsylvania State University

    I’ll post other videos from the discussion series, but if you can’t wait, you’ll find a couple more here, with more to come.


    Comments closed
  • Middle-aged, white Americans’ death rate is up. Alcohol and drugs are to blame.

    As Gina Kolata reported, the death rate for middle-aged, white Americans has risen in the last 15 years or so, while that of many other groups here and in other wealthy nations fell. The rising death rates are concentrated in less educated, middle-aged whites.

    That finding was reported Monday by two Princeton economists, Angus Deaton, who last month won the 2015 Nobel Memorial Prize in for Economic Science, and Anne Case. Analyzing health and mortality data from the Centers for Disease Control and Prevention and from other sources, they concluded that rising annual death rates among this group are being driven not by the big killers like heart disease and diabetes but by an epidemic of suicides and afflictions stemming from substance abuse: alcoholic liver disease and overdoses of heroin and prescription opioids.

    The original study is here.

    (Here on TIE, we’ve posted before about rising mortality rates for women in certain counties, for white women without a high school diploma, and for white men and women with fewer years of education. We’ve also posted about a source of bias that could explain some or all of the phenomenon.)

    And, here’s your periodic reminder that, at the moment, Medicaid and Medicare records pertaining to substance use disorder are not included in certain research files, hampering our ability to study these issues.

    white American deaths


    Comments closed
  • AcademyHealth: Access to opioid treatment programs, part 2

    The Obama Administration is making a strong push to increase access to medication-assisted treatment (MAT) for addiction to opioids. In the second of two, related posts on the AcademyHealth blog (posted last week), I reviewed evidence of prior MAT access expansion measures.



    Comments closed
  • Time travel

    Via Michael Younger:time travel

    I worry about recursion until the start of time, a very bad ending.


    Comments closed
  • Why are vaccine prices so low?

    Vaccines are biologic drugs. But compared to many other biologics, they’re cheap. I took to Twitter to try to find out why.

    This led to a few Twitter-conversations, emails, and some article reading. Here’s what I learned so far:

    To an insurer, the combined price of all the childhood vaccines, put together, is about $2000, tops. Even at the high end — and even though that price tag was a lot lower in years past — it’s money well spent. Vaccination is considered one of the most, if not the most, valuable and effective, clinical preventive service. Vaccination is highly cost-effective. In fact, a 2005 article in JAMA Pediatrics concluded that “the current routine childhood immunization schedule results in substantial cost savings” with direct and societal cost-benefit ratios of 5.3 and 16.5, respectively. A 2014 article in Pediatrics concurs.

    In comparison, Tarceva, a biologic drug used to treat lung and pancreatic cancer, costs tens of thousands of dollars for a few weeks to a few months of survival. It is borderline (and arguably not) cost-effective for the former and nowhere near so for the latter.

    Given their high value, their necessity for school enrollment (in most places), endorsement by the medical establishment, and widespread insurance coverage for them, you’d think the high demand for vaccines would push their prices up to something like that for Tarceva and other high-priced biologic drugs. Why not?

    One lead on an answer came from my friend John Methot, who has a decade of experience working on early drug discovery. He focused on development costs. Before you argue that “costs” don’t inform “prices,” bear with me (and him):

    Biologic drugs are intended to modify the mechanism of a disease process. This is incredibly hard. We have to understand the disease biology well enough to identify potential target genes, then spend years “validating” a target: proving that if we modulate its product we can modify the phenotype in (only) the desired way. Vaccine discovery is conceptually simpler: isolate the pathogen of interest and then figure out the minimal antigen portion you can introduce to cause an adaptive immune response. That’s still hard (e.g. decades of failure inventing vaccines for HIV and Malaria), but it’s a more well-defined problem with more established approaches.

    Development: With a vaccine, you are not trying to deliver a therapeutically active protein to a specific tissue with all the associated pharmacodynamic complexity, off-target effects, exposure and other issues to deal with. You’re only trying to introduce (part of) an antigen to immune cells in circulation. That is a simpler, albeit still complex, proposition. So the development time/cost of a vaccine is likely substantially lower than a biologic drug.

    Production: Probably not as big a difference, but vaccine production is also very well-established and routine as compared to biologic drug production. In fact I bet many big pharma companies that make vaccines contract out the actual production.

    To these observations, we might add that some vaccines are developed largely on government funds (see polio). The earlier ones weren’t patented, which might have increased competition, at least in the past.

    So, vaccine development and production may be easier than for other biologics. That could mean its costs are lower. Lower costs impose a lower barrier to entry. More entrants, or potential entrants, means more competition. More competition can lead to lower prices. Voilà!

    Other evidence supporting this theory:

    • There is price variation. Some larger physician practices can negotiate vaccine price discounts, something that should not be possible if there were no vaccine competition.
    • Though considered inferior and not available in the U.S., Synflorix is something of a competitor with Prevnar 13.
    • Danzon and Periera made the case that vaccines markets are dynamically competitive: Superior products push out older ones. This probably is price increasing over time, but I wonder if it limits the rate of increase. For instance, if vaccine A is superior to B, it can be priced higher, but not infinitely higher. There’s some limit, at least in the short term, because use of A and B overlap while B is still considered acceptable. B’s price would also go down in this setting.
    • Danzon and Periera also wrote that new vaccine classes initially attract multiple entrants, though not all stay in the market. Still, that might imply a threat of future entry. On the whole, it might put brake on price growth.
    • Danzon and Periera think competition among vaccines pushes their prices below our willingness to pay.

    Not so fast, says Matt Davis, Professor of Pediatrics, Internal Medicine, Public Policy, and Health Management & Policy Deputy Director, Institute for Healthcare Policy and Innovation at the University of Michigan. By email, he conveyed the following:

    • Mark Pauly suggested that pharma has invested less in vaccines because they’re more, not less, difficult and costly to discover.
    • Davis’s work examining over 132 vaccines and 4235 non-vaccine products over a 16-year period showed indistinguishable failure rates between the two groups. The times spent in the phases of development was slightly longer for vaccines. If vaccines were easier to develop, their failure rates should be higher and time spent in development phases should be shorter than other drugs. Caveat: This analysis did not directly compare vaccines to other biologic drugs.
    • The most competition any vaccine manufacturer faces is from one other producer. There are effectively no generic vaccines in the US. But, maybe it’s enough to have two actual entrants and the threat of more to keep prices low though (?).

    Over half of all childhood vaccine doses are publicly purchased at discount. (That discount is much smaller today than it once was, Matt Davis told me). The two, big vaccine-providing public programs are Vaccines for Children (VFC) and Section 317. Together they provide vaccines for children who cannot otherwise finance them. The former requires no congressional action for expansion: once an advisory committee deems that a vaccine must be provided, the federal government must establish a contract with a manufacturer to do so.

    These programs could moderate price growth in the pre-Affordable Care Act era as follows. If vaccine prices went up too high, perhaps an insurer would drop them from coverage or require higher cost sharing. That would make them less affordable for families, and more would obtain them through the public programs, like VFC. Manufacturers get a lower price from those programs — and perhaps there’s a threat federal prices could go lower still (?) — so they have an incentive to keep prices from rising too much in the commercial market. In fact,

    Concern has been raised that failure to cover vaccine administration could result in shifting children from private providers to health department clinics [28]. Previous experience showed that low administration fee reimbursement rates [29, 30] and high out-of-pocket costs to parents resulted in referrals to health department clinics [31, 32, 33].

    However, now that the ACA mandates zero-cost sharing vaccine coverage from insurers, the threat of dropping or reducing that coverage is gone. If my theory is right, that should push vaccine prices up, particularly for those with no competition. This is empirically testable.

    I’ve got more reading to do on this subject and a forthcoming plane ride on which to do it. I’ll post more when I’ve done that. Meanwhile, got any other explanations for why vaccines are so cheap? Email them to or tweet them at me.


    Comments closed
  • AcademyHealth: Access to opioid treatment programs, part 1

    The Obama Administration is making a strong push to increase access to medication-assisted treatment (MAT) for addiction to opioids. In the first of two, related posts on the AcademyHealth blog, I review evidence of prior MAT access expansion measures. Read it here.



    Comments closed
  • Care about cost effectiveness? Take a meeting!

    I very rarely recommend anyone watch or participate in a meeting, so this is a BFD: If you care about cost effectiveness, I recommend that you spend some time watching parts of one or two webcast meetings this week. Here’s why:

    Whenever I raise the application of cost effectiveness analysis (or comparative effectiveness) in health care, I get a lot of reasonable push back such as:

    • What about heterogeneity of treatment effects?
    • What about special populations? Vulnerable populations? Special and vulnerable populations?
    • What if this is the only treatment for a rare disease that kills babies? Special, vulnerable populations of babies?

    It’s issues like these that lead reasonable people to propose processes that consider cost effectiveness alongside other factors. (It’s my view that it leads unreasonable people to give up entirely on cost- and comparative-effectiveness. Your mileage may vary.)

    Consideration of cost effectiveness and other issues, such as those above, is what NICE does. And, here in the U.S., that’s what the Comparative Effectiveness Public Advisory Council (CEPAC) and the California Technology Assessment Forum (CTAF) do too.

    People who ask questions like those above are absolutely right to imply that weighing costs, effectiveness, and issues of justice and fairness is exceedingly challenging. Yet, these things can and do get weighed, at least implicitly, by various components of our health system—and likely in no consistent fashion geographically, over time, or across populations or treatments.

    CTAF and CEPAC members (of which I am one) struggle mightily to weigh these things in a more deliberative and consistent manner. It’s exceedingly difficult, but worth your time to watch.

    So, before you conclude that we just can’t work all this out in some kind of sensible process, take a look at what CEPAC and CTAF do. This week offers two opportunities:

    The first will be a meeting of the New England Comparative Effectiveness Public Advisory Council (CEPAC) that will review the report on the PCSK9 inhibitors Praluent® and RepathaTM to treat high cholesterol.  The second will be a meeting of the California Technology Assessment Forum (CTAF) that will review the report on treatments for congestive heart failure (CHF) (Entresto™ and CardioMEMS™). […]

    On Tuesday, October 27, 2015, the New England CEPAC will meet in Boston, MA to review the PCSK9 report.  […] An agenda of the meeting can be found here. […] Attendance for the in-person meeting is at capacity, so those wishing to hear the meeting should register for the live webcast of the event by clicking here. A video recording will be available on the website the week after the meeting.

    On Thursday, October 29, 2015, CTAF will meet in Oakland, CA to discuss the CHF report. […] An agenda of the meeting can be found here.  […]

    A limited number of seats are still available for members of the public who wish to attend the CTAF meeting […].  Members of the public may register by clicking here, those unable to attend the meeting in-person may register for a live webcast of the event by clicking here. A video recording will be available on the website the week after the meeting.

    If you can only spend a short amount of time watching one (or both) of these meetings, I recommend you focus it on the evidence review, deliberations, and voting. I don’t mean to short change the policy discussion, but I’d add that only after you see the other parts of the process.




    Comments closed