The new breast cancer screening guidelines released last week, along with the supporting study, are among the most disputatious medical recommendations in recent memory. Critics on the right charge government rationing, while those on the left suspect an insurance company conspiracy to cut essential coverage. Adding gasoline to the fire is the even more heated health reform debate that has led its combatants to hijack the mammography issue to bolster their own particular views, pro and con.

To get past the politics, I read the study. It is hardly a page turner, but it’s a credible scientific analysis of the available data on breast cancer screening. A key question it addresses is how effective mammography is for women under age 50—an issue that has ping-ponged back and forth across the medical policy community for four decades. In this latest volley, the U.S. Preventive Services Task Force (USPSTF), citing the study, recommends “…against routine screening mammography in women aged 40 to 49 years.” But then it equivocates with, “The decision to start regular, biennial screening mammography before the age of 50 years should be an individual one and take patient context into account, including the patient’s values regarding specific benefits and harms.”

The recommendation is based on the relative benefits of testing women in their 40’s (one averted death per 1000 women) against the harm caused by overdiagnosis (finding and treating cancers that don’t shorten a woman’s life), by excessive radiation, and by false positives (inconvenience, psychological harm, and unnecessary follow-up imaging and biopsies).  Here are some of the study’s findings (which may surprise you):

  • Without screening, the average 40-year-old woman has a 12-15% lifetime breast cancer risk and a 3% mortality risk.
  • The advantage of commencing screening at age 40 (vs. 50) is very small—a 3% reduction in mortality (i.e. reducing the 3% probability of death to 2.9%).
  • Starting annual screening at age 40 (vs. 50) increases median life-years by 33 per 1000 women screened, or about 12 days per woman.
  • Slow-growing tumors are much more common than fast-growing ones, and the ratio of slow- to fast-growing tumors increases with age, so little survival benefit is lost between screening every year versus every other year.
  • For the small subset of women with aggressive, fast-growing tumors, even annual screening is not likely to confer a survival advantage.

These findings are quite interesting, but hardly controversial. The squabble arises from the panel’s using them to craft new, more restrictive recommendations on when women should seek mammography screening. To do that, it had to go beyond the science and apply human judgment as to whether the benefits of early screening outweigh the harm. This was further complicated—greatly, in my view—by the panel’s use of dissimilar measures for benefit and harm: death rates and life-years for the former, and false-positives and overdiagnoses for the latter. Comparing one with another is a lot like comparing avocadoes with chairs. It produces no objective score or benchmark against which the judgment part can then be applied. This practically guaranteed a major fracas when the result was to recommend restricted screening. That’s probably why the task force ended up waffling on its recommendation.

One oft-voiced fear is that these new guidelines will lead to rationing of mammographies by the government. But so far, no one has noticed the elephant in the living room: mammographies cost only $170 each (national average). How can they be denied to anyone when all but the poorest women can afford 25 cents a day for every-other-year screening if she and her doctor think it’s essential?

Wouldn’t it make more sense to simply drop insurance coverage for mammography altogether and let women decide for themselves when to spend their own money for it, based on whatever statistics, guidelines, desires, fears, and doctor recommendations they choose to heed? The purpose of insurance is to pay for things that are expensive and that you normally don’t expect to need. Mammography is neither. True, insurers love to cover it, because they get to add not only the screening cost to women’s premiums, but also nice margins for administration and profit. But why should American women, arguably the world’s most savvy consumers, overpay for that? And if mammographies were a normal consumer purchase, you can bet that enterprising medical entrepreneurs will figure out how to offer them for a lot less than $170, possibly even less than today’s copayments for the procedure.

Speaking of insurance, how do you suppose the current House and Senate health reform bills deal with mammography? Both the pending Senate bill and the recently passed House bill require insurers to cover preventive services that have been given an A or B grade by the USPSTF. Since the task force gave a C to under-age-50 mammography, (you heard it here first) it will not be a required benefit. So you can ignore all that political posturing by the usual suspects claiming that nothing will change. It just has—or will, if the bill becomes law.

But is this rationing? Not really. Almost everybody can already afford it as a normal consumer expense. And giving poor women the money to buy it (or something else) would be a trivial expense for the taxpayer.

What do you think? Please let me know in the comment space below.

This entry was posted in Health Costs, Health Insurance, Health Reform Goals and tagged , , , , . Bookmark the permalink.


  1. Lindsay says:

    I particularly like the portion of your analysis that notes that the comparison of benefits to harm of mammograms in the 40s as set forth in the study is really an apples-to-oranges comparison. No doubt some of this is occasioned by an unwillingness to put a dollar value on the average number of days of life saved by having a mammogram in the 40s. To do so would make the study even more controversial, so it seems the apples-to-oranges comparison can’t be avoided.

    But not so fast – maybe we can avoid the apples-to-oranges comparison or at least most of it. To recap the harms, they are (1) over diagnosis, (2) some additional radiation, and (3) false positives (inconvenience, psychological harm, and unnecessary follow-up imaging and biopsies). Let’s analyze each of these items one at a time. “Over diagnosis” means two things: the expenditure of money on treatments which are without benefits, the cost of which can be measured in dollars, and the cost of the woman’s time undergoing such treatment, which can be measured in hours and days of wasted and unproductive time taken away from the woman’s life. “Additional radiation” means radiation that results in additional harm such as cancers that otherwise wouldn’t have occurred or cancers which are aggravated by the radiation, which again can be measured in the expected number of days of life lost. “Inconvenience” means the amount of time spent getting a mammogram, including the time to set up the appointment, the time to travel to and from the health center, and the time spent in the health center getting the mammogram and filling out forms, all of which can be measured in hours and days of time taken away from women. “Psychological harm”, the psychological harm of overdiagnosis and false positives, is more difficult to deal with, but let’s assume it can be offset because there is similar psychological harm in not getting a mammogram and worrying about whether one has breast cancer that might be detected but isn’t. Lastly, “unnecessary follow-up imaging and biopsies” can be reduced to time out of woman’s life taken away by these procedures as well as the dollar cost of these procedures. In sum, all the harm can be condensed into (1) an expected number of days of a woman’s life taken from her and (2) a dollar cost. Suppose a further analysis showed the harm to be 7 days of woman’s life lost on average and $800 in additional medical costs. Since the benefit is 12 days, the net benefit must be 5 days and the net harm must be $800. The comparison, then, boils down to comparing one day of life with $160.

    This doesn’t answer the question, but it simplifies it tremendously and quantifies it and may help us make at least some decisions. If, for example, a similar analysis of other medical tests that are currently approved shows that, in general, we as a society, are willing to pay more than $160 a day for these other medical tests, it would make sense to pay for mammogram testing of women in their 40s. Similarly, if the other medical tests that are currently approved demonstrate we are not willing to pay anything close to $160 a day, then it would not make sense to pay for mammograms for women in their 40s. Obviously, this analysis doesn’t answer the more fundamental question of what is the right tradeoff between days of life and dollar costs, but that is a question to which we will never get an answer. If the U.S. Preventive Services Task Force had supported its recommendation by comparison with norms of days saved vs. dollar costs incurred, perhaps the recommendation would have been easier to understand and accept.

    Furthermore this simple comparison of days of life save vs. dollar cost may also help individual women answer the question of whether to get mammograms in their 40s for themselves. Some women might say to themselves, I have a very busy life and the small benefit isn’t worth all the personal hassle today to have a controversial test performed that has only limited benefit in the distant future. Other women might say to themselves, I’m terribly worried about getting breast cancer and the additional time and hassle involved in testing and the follow-up is well worth it to alleviate my concern and lengthen my expected life.

    In summary, the benefit-harm analysis done by the study could, with further work, be put into a form that would make it easier to make policy decisions and would make the decisions of individual women easier.

  2. Randy Dipner says:

    My wife and I have been discussing this report since it was released. We know too many young women who had cancers detectd early to fall for a simple statistical argument. Rather, the idea of making the procedure self paid really addresses the core issue. Let the individual decide (with the consultation of her doctor) when it is right for her.

    In addition, the illumination concerning the House bill is very interesting and deserves widespread disseminatiom.

  3. John Anders says:

    Steve Hyde’s analysis once again cuts through and clarifies a tortured medical issue. As ever, it should be the right of the individual to make choices regarding his or her’s medical care. My wife and I continue to consult this blog for timely advise and opinion. We come from the left side of the spectrum but are often amazed at how Hyde coaxes us out of our position — and up to the middle ground. It’s hard to argue with SH’s clear but passionate logic.

  4. Dr. Bob Browne says:

    Stephen, nice analysis. I have several comments since I have dealt with similar economic/medical/ethical issues many times in the past.

    First, congratulations to your first commenter. His/her analysis is a form of what health economists call cost-utility analysis. It is widely used in other countries but is controversial in this country on philosophical grounds (as I will explain later). This type of analysis attempts to quantify all attributable costs and compare them to a standardized benefit measure. The most common measure is the QALY (quality adjusted life year). The tacit assumption in the QALY is that not all life years are created equal, that a year of life with some disability is worth less. The use of such cost and benefit measurements raises several interesting issues.

    First, ideally the cost side of the equation should measure ALL costs, both direct (e.g. medical expenses, costs of adverse events) and indirect costs (e.g. lost productivity). Rarely do such analyses capture all the costs, many of which are indeed hard to quantify.

    Second, these analyses rarely do a good job of quantifying all the costs of inaction. For example, in this case what are the total individual and societal costs of patients not diagnosed and therefor harmed. Again both direct and indirect costs must be quantified.

    Third the methods for calculating QALYs vary and none is universally accepted, partially because quality of life is a subjective and very personal based on individual values (a fact expressly admitted in the “weasel out” statements in the recommendation.

    Fourth, I always have trouble with these analyses when the decision rules and thresholds are not expressed clearly up front before the analysis. In many cases the methodology suggest that they agreed “We can’t tell you what the right answer will be but we’ll know it when we see it!”

    That being said, in general other governments and societies have generally accepted that a technology that costs over $35-50 thousand/QALY is not of significant value to “society” (i.e. the government budget). The irony of such figures is that many other societal mandates and initiatives that are widely “accepted” (e.g. air pollution controls, air bags, seat belts) easily fail to meet this threshold.

    However there is another factor that is even more important coloring this debate, the dilemma of societal vs. individual good. The current recommendation rides the fence here and decides to take neither position. Their general recommendation suggest that across the entire population the use of mammography in women under age 50 may not make sense economically in their opinion(by whatever decision criteria they used – the specific decision rules were not published). However they then offer the ultimate waffle and state that individual circumstances and values may dictate different decisions!

    This situation is not unique and it always occurs when population decisions (one size fits all) inevitably conflict with individual realities. It is more inflammatory in our society with our historical emphasis on individual decision making and independence. But individual decisions must not be cushioned from all financial responsibility for those decisions. The only answer is your approach that puts both the decision and at least some of the costs squarely on the consumer.

    Finally It is another compelling reason why we must get government out of their current role in our health care system. The “independence” of the USPSTF or any body appointed by any individual or agency associated with HHS is a myth. I have spent the last 25 years dealing with such agencies and “independent” groups and they are clearly influenced by their appointer, starting with whom they choose for the group. The public will NEVER trust the decisions of such a group that is not truly free from undue private or public sector influence. The current health bills are replete with an incredible network of new Federal agencies, task groups etc. whose sole role will be to make these “recommendations”. Yet the government is also the country’s largest health payer. Talk about conflict of interest.


Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>