Do We Really Want to Know?

Recently I had the opportunity to write an editorial for an article on screening mammography in the elderly. Some of you know that I have had a longstanding interest in this topic. After all, we (clinicians, academicians and policy-makers, collectively) have spent years agonizing over whether to do mammograms in 40 year olds, but have spent little time studying and thinking about older women. As you may know, if we followed current NCI and ACS recommendations, we would be ordering screening mammograms without regard to advanced age, even for centenarians!

My editorial ("Older women, mammography, and mortality from breast cancer." American Journal of Medicine vol.108:174-175) was about a retrospective study that found that the effectiveness of mammography among women aged 65 to 79 years of age is likely pretty good. Among other things, I pointed out that a randomized trial of screening mammography will probably never be done in older women. I cited some other studies showing that there is about a 5-year delay between the time a screening mammogram is done and the time that a benefit in terms of breast cancer mortality can be measured, even in women aged 50-65 years. This led me to suggest that concluding mammographic screening for most women at about age 80 or so seems reasonable, since the average life expectancy at that age only modestly exceeds the 5-year delay before a screening mammogram would be expected to benefit the patient. A reasonable cutoff might be even earlier for those with major health problems, such as congestive heart failure.

Although I thought my editorial well-written, concise, and supported by facts, someone wrote a letter to the editor to complain. Allow me to digress for a moment, for any of you contemplating an academic career. I greatly enjoy my career, and strongly encourage academic careers for those with interest and motivation. But you do have to get used to the fact that someone will always disagree with you, no matter how well thought out your arguments might be. So I thought I would share with you the crux of the argument made by the author of the letter-to-the-editor, and my own subsequent reflections.

In summary, the letter-writer's argument was that I shouldn't have recommended any upper age limit for screening mammography, as older women vary a great deal in terms of disease burden and functional status. Recommending an upper age limit for routine use of screening mammography was failing to take into account the diversity of older women. Instead, we should try to estimate the remaining life expectancy of our patients and use an age cutoff that is customized for each woman.

I thought that this writer made some good points. I certainly agree that older women exhibit a great diversity of health and functional status. I believe in a customized approach to health care. I had actually alluded to the idea of customizing a cutoff age for screening mammography when I mentioned having an earlier cutoff for women with major health problems.

But is the idea of calculating an estimate of life expectancy to make routine screening decisions an idea that is ready for prime time? For major therapeutic decisions, perhaps so. For a decision about a routine screening test, I am not so sure. It is presently much easier to estimate life expectancy for large groups than it is for individuals. It is particularly difficult to estimate who will live longer than the average. For example, if someone has a low ejection fraction, their life expectancy is probably lower than average for that age, but if someone has a normal ejection fraction, their life expectancy is not necessarily longer than average for that age.

Let's say for the purpose of discussion that someone worked out a reasonably accurate, easy algorithm for estimating remaining life expectancy. We could program a clinical calculator for it on our Web site. We could have the medical assistant calculate life expectancy at the time of those routine office visits at which screening issues are typically discussed, and write it on our flow sheets, right next to the weight and blood pressure.

Such a possibility would undoubtedly be instructive, particularly for our trainees. I have been amazed at the number of people who don't realize that a healthy 65-year old woman has on average 20 years or so of remaining life expectancy. But is this information that our patients (especially our 80 year-old patients) are looking for as part of periodic office visits? I am not so sure. Imagine the counseling part of the visit. "Well, Mrs. Jones, your blood pressure is marginal, you are otherwise doing well, and we estimate you will live 48 more months, so you can stop having mammograms now." This is the kind of information that people may need to confront when making an important therapeutic decision, say whether or how to treat a cancer. But I am not sure that many patients are looking for this information just to decide about having a mammogram or fecal occult blood testing. Some might even consider it more a death sentence than a useful statistical estimate.

And so, I am back to thinking that perhaps we don't have to try to know everything. In the case of screening decisions, maybe it is not so bad to have an age cutoff based on average life expectancy, while recognizing that any general guideline may require modification due to patient preference and special circumstances. Some year, hopefully quite a few years from now, I hope that my doctor will just stop recommending mammograms without telling me more than I want to hear about it.

Ann Butler Nattinger, MD
Professor and Chief of General Internal Medicine
Medical College of Wisconsin

"Reflections" is a collection of essays by the health professionals of the Medical College of Wisconsin.