More on mammograms

Emotions, as they say, run high, especially about “health.” There is a great deal of superfluous, extraneous, and irrelevant discussion taking place with regard to the new mammogram guidelines. Yesterday, many organizations rushed to the microphone to say, “Don’t worry! We still recommend every women gets mammograms as frequently as possible.” They were worried lest women think they did not care. Let’s examine this phenomenon.

There is a range in how much people care about their own physical being, from hypochondria to indifference. I’ll use the term hypochondriac to mean not just those who suffer from the mental disease of ridiculous concern, but also for those who worry disproportionately to the actual risk of any malady. Antibacterial-goo-carrying hand washers are in this category, as are mothers who fearfully drag their tot to doc every time it sneezes. The indifferent eat and drink whatever the hell they want and say, “A short life, but a merry one.”

The reactions of those at either caring extreme to those at the opposite end is not symmetric. That is, hypochondriacs regard the indifferent in a manner that is wildly different than how the indifferent think about hypochondriacs.

The indifferent think hypochondriacs are nuts, and that they’d be better off if they just wouldn’t worry so much. But the indifferent are inclined to think that if the hypochondriacs want to run to the doctor for every perceived symptom, let them; as long as they leave us alone.

Hypochondriacs—and their close cousins the worrywarts, hand-wringers, and fretters—think the indifferent are nuts, too. But in general the hypochondriacs cannot abide the indifferent and want to change their behavior. Misery loves company, and the misery caused by constant concern should be enforced, even mandated, for all. Hypochondriacs rally under the banner, “Something might be wrong with you!”, which is a true statement, but an empty one, because it contains no information beyond the fact that some people get sick and some don’t. It doesn’t tell you the odds.

Very well. Hypochondriac females desire, as early as they are allowed, to have a mammogram, and have them often, “so that they know.” The indifferent look at these females and say, “Go ahead, but I don’t want to take the chance of a false positive.” But the hypos can’t stomach this, don’t care at all about false positives—they might as well not even exist—and so they browbeat the indifferent until they agree, at least verbally, with the hypos’ concerns.

But there is middle ground.

All the new mammogram recommendations did was to seek a balance between the admitted inaccuracies of mammograms—and they are very inaccurate—with the chance of developing breast cancer (which is small), all while attempting to account for the average person’s feelings about false positives and false negatives. The average person is not a hypochondriac nor is she indifferent. Further, there are many more in the middle than at the extremes.

The study—and my own work—highlighted four things which were previously unemphasized, and which every woman thinking about a mammogram should consider.

  1. False positives exist and are common. And they have serious consequences. The costs involved with the mammogram’s mistakes should, and almost never are, considered before deciding to have the test. Women should be better advised.
  2. The doctor’s costs are not the same as the woman’s. Obviously, the doctor doesn’t suffer at all if the mammogram gives a false positive. In fact, his billing increases as he calls the woman back for further tests. It’s not his breast being squeezed under glass, nor his breast being cut open to extract a possible unnecessary biopsy, nor his breast being exposed to additional radition.
  3. The age at which to begin mammograms is not set in stone. To those who condemn the new guidelines, I ask—seriously—why not recommend mammograms at 13 or 14? Why not get them twice yearly, or even once every quarter? The answer you give cannot be “Don’t be silly.” You must base your response on quantitative data. That is what the study (and my work) tries to do.
  4. Not all cancers are equal. Some breast disease can be lived with and won’t kill, and some of these, sometimes, are better left alone. Read the original paper (linked here) for a more thorough discussion.

That last point need emphasis. It might not be that 50 is the right age for healthy, otherwise not-at-risk women (at risk women include those with family histories of cancer, etc.) to commence screening, it might be that it is 52 or 48. It also might not be best for tests to be annual or semi-annual. What is best and how do we tell? To explain that requires the math I pointed to yesterday. The point to take away is that this is a question that can be answered empirically and quantitatively—and not politically. The correct age and timing will be different for each woman—and there are tools to help her decide what these numbers are.

One more thing: I heard some docs say, “The data used in the study were not all current. Treatment methods have changed.” This is true, but largely irrelevant. Breast cancer treatment is not at issue: the accuracy and usefulness of mammograms is, and these fallible instruments have not improved much through time.


  1. Ok, now you’ve gone too far–by not going far enough.

    All this emphasis on false positives, while well reasoned & factual, is missing its counterpart — mention of false negatives. Presumably those happen too? Some remarks on that might be in order (even if not so applicable to this particular topic, this is a factor in other areas — the analysis presented here is readily transferred to other topics after minor editorial adjustment).

  2. This post is more educational than the previous one. Yes, there is middle ground. Let hypochondriacs and the indifferent be.

    I don’t know whether politics is involved in the recommendation, and it’s irrelevant to me. A recommendation is simply a guideline or an option, and you can still make your own decision. Be informed, be aware of the risk and our own bodily changes, do your own research and take actions as early as possible. “The best protection is early detection.” Sometimes, we are our own best doctors.

  3. Ken,

    Right, false negatives. Here the doc’s comments about better treatment work in the favor of fewer screenings. Why? Because, presumably, if a deadly cancer goes undedicated (benign cancers can go undetected harmlessly) longer, the better treatments he touted can cure or manage it easier once it is detected.


    Amen, sister. Heal thyself.

  4. I have seen posts that imply a gender angle to all of this. I think it bears repeating that similar analysis applies to prostate cancer. There is much debate about the appropriate response to rising PSA levels, for example. As with the mammography, much depends on age, family history, etc.

  5. This is sure getting a lot of attention and is generating more heat than light in the MSM. In part this is because the media lead with the drama – potentially no insurance funding for mammograms until 50. Left as an afterthought are two critically important caveats. First, if you are in a designated high risk group then insurance should pay for your mammograms regardless of age. Second, if you want to pay for a mammogram yourself before the designated age you are free to do so.

    For 25 years I was responsible for buying health insurance for my small company (N ~ 40 employees). The major issue I was concerned with was eliminating wasteful use of medical resources such as emergency room visits versus primary care physician visits, generics versus name brand prescriptions plus discretionary services. Many with insurance treat medical services as a free good and over-use or mis-use or abuse the services. There are few if any incentives for people or service providers or insurance companies to behave in a more judicious fashion. Even as the purchaser our premiums did not reflect our discretionary usage but a general class of usage. The mammography or prostate testing examples and the reactions to any change in “free” access constitute a microcosm of the whole health reform debate. There are many things wrong with the current behavior of health insurance providers but there are equally many things wrong with the behavior of users and potential users of medical services. Divorcing use from the cost of using the system is always problematic and generally leads to a misallocation and waste of services. Most recently we went to an insurance program with a much larger deductible coupled with a Health Savings Account. We essentially designed it so that we would put the money saved in premiums into the employees’ HSAs which essentially covered the change in the deductible. Now employees have a significant incentive to be more careful in their consumption of medical services because we created more of a payment nexus: You actually had to receive and pay a bill. The money accumulated in your HSA could be used for any medicl procedure in the future. Thus if you wanted fancy eyeglasses or contacts you could pay for them from your HSA. At the same time, the coverage was sufficient to protect everyone from potentially ruinous medical procedures such as lung transplants and heart by-pass operations that should be the primary targets of insurance (i.e., risk sharing as opposed to cost shifting).

    The latter point, i.e., risk sharing versus cost shifting, is one of the reasons why I never bought Dental Insurance for the company – most such programs are pure cost shifting exercises and are very poor insurance buys.

  6. I like Briggs’ honest and rational viewpoint. But I don’t like a group of government wogs, no oncologist in the mix, changing a policy after a major decline in mortality rates coincident with the introdcution and widespread use of mammograms; I don’t like the timing which is so bad as to suggest intent; and I don’t like the largely absent voice of women in the discussion.

  7. Noblesse Oblige,

    I do not know what you mean by “government wogs” but I think that being no oncologists might be a good thing. Don’t forget that it’s sort of taking your car to the mechanic and expect to be told that no work is needed.

    The decline in mortality -which I did not check so I take it for granted- could very well due to an increase in awareness and general preventative medicine, not just mammograms. So to me you’re attributing the merit to a single procedure that we already know is faulty (Briggs already wrote on how better -i.e. less errors- it would be if you just assume you do not have breast cancer).

    As far as timing I can confidently say that for a paper such as this no timing can be concerted so precisely since it takes quite some time from start to finish.

    Finally what kind of voices you would like to hear? I am assuming your question is rhetoric and you would like to know from women who are “saved”. However you’ll almost never hear from those who endured suffering due to false positives and they are the only one that can balance the view.

    And in any case I always warn (for what my opinion is worth) people to be skeptical of decisions taken in a distress/emotional condition since they might not be all that objective. And if you want to talk to women who faced the issue I think you’ll likely add a lot of emotion and little information useful for doing the right thing (see the first comment to the last update column in this blog).

    Just my opinion.


  8. Am highly suspicious about any guideline that uses the age of 50 as a cutoff between no mammograms at all before and then one every year (or two). Artifacts of study design and picking simplistic round numbers are the likely culprits here.

    Guidelines suggesting an earlier baseline mammogram and then a decreasing interval between followup mammograms, as the risk increases in a non-stepwise fashion, would be far more plausable and, until proven otherwise, more scientifically based.

    Not only that, they would make more common sense and be more credible to both women and the medical community. They might even get acted upon!

Leave a Comment

Your email address will not be published. Required fields are marked *