Breast Cancer Screening in Women Younger Than 50 Years of Age: What's Next?

In this issue, a well-constructed cost-effectiveness analysis of screening mammography in women younger than 50 years of age [1] shows that screening in this age group is expensive, costing approximately $150 000 for a year of life saved. Those who do not like this result will no doubt challenge some details of the analysis, but they will not be able to shake its basic conclusion, which is determined by the low incidence of the disease, the relatively small beneficial effect of screening, and the long delay before any benefit appears. Unless the cost of mammography decreases dramatically, screening in younger women will always be expensive compared with the benefit it provides.

What does this mean? It means that in settings that face limits on resources, which is almost every setting, putting money into this activity instead of more cost-effective activities will do more harm than good. The amount of harm can be estimated by looking at other activities that would be pushed aside. For a nearby example, screening women 50 to 65 years of age delivers about seven times as much benefit per dollar as screening younger women [1], yet only about 60% of women older than age 50 years in the United States are screened regularly. Until that rate is increased to its practical limit of perhaps 90%, every year of life added by preferentially putting resources into screening women in their 40s will be accompanied by about 7 years of life lost because of suboptimal screening in women who develop the disease after 50 years of age. Even if budgets are increased to accommodate the additional costs of mammography in younger women, the net effect will still be more harm than good if the funds could otherwise go toward another activity that would have provided greater benefit.

Will this information slow down the push for breast cancer screening in younger women? Not if the deliberations continue as they have in the past. Screening mammography seems to have a life of its own, independent of any considerations of evidence, effectiveness, or cost-effectiveness. Indeed, the debate over its effectiveness has been one of the messiest of recent times [2, 3]. Those discussions are winding down as the trial results slowly approach statistical significance, but this does not mean that the appropriateness of mammography in younger women has been established. Even the staunchest supporters of screening mammography have to be disappointed by the small magnitude and delayed appearance of the benefit (about 1 death prevented in 1000 women after 10 years) and by the high probability of a false-positive result (about 25% over 10 years). Clearly, it is important now to ask whether this is a good way to use resources. As we try to answer that question, it will be important to determine what went wrong with the last debate and how to make the next one go right.

What went wrong with the debate over evidence is that the two sides had different ideas about how to approach evidence, but that disagreement was never brought into the open or resolved. On the surface, the last debate seemed to be about whether mammography in younger women was effective. In fact, the conflict was not about that conclusion but about the process we should use to draw that conclusion. Specifically, when we encounter a technology, such as mammography, that has both biological and social appeal, should we begin with a presumption that it is effective and appropriate and then proceed to promote it until evidence proves otherwise? Or should we make no initial presumption of benefit and resist promoting the technology until the evidence shows it to be effective?

This basic conflict not only separates the different participants in the debate but tears at each of us. My intuition and my heart have always been moved by the fact that breast cancer is bad; by a theory that screening should work; by an analogy with mammography in women older than age 50, which does work; and by a humanitarian urge that if it has any benefit, it should be used. But I also have a commitment to the scientific method, which begins with a hypothesis of no effect and requires evidence to change that position.

I can make a case for either. The case for my intuition and heart is straightforward: It just plain feels right. I am comfortable with it. And, on the flip side, any other stance makes me uncomfortable. In contrast, the case for the scientific method is much more cerebral, much less comfortable, and much less popular. But I find it compelling. First, it acknowledges that empirical evidence is a far better monitor of reality than my intuition and hope. Second, it is the basis for good clinical research and has made medicine and the other sciences what they are today. Third, it protects us from the one-way-street effect of premature diffusion. When a popular technology diffuses without evidence or analysis, calling it back is difficult. We tend to get stuck with our mistakes for a long time. But if an effective technology is held up while its effectiveness is being documented, it can always be disseminated once the evidence comes in (albeit with regret that an effective technology was withheld from some who could have benefited from its earlier application). Finally, the scientific method is more practical. Without it we have chaos, ignorance, and waste: chaos, because four different people can have four different beliefs; ignorance, because if we are content to rest on our personal beliefs, we have no incentive to do the hard work needed to determine whether we are actually right; and waste, because if we do not collect the necessary evidence, we lack the information we need to weed out harmful, ineffective, and inefficient practices.

But these are only my reasons. What went wrong with the debate over evidence of screening mammography is that we never argued these points or reached agreement on which approach should guide public policy. That is the lesson we need to learn from the past. Before we begin the next debate about a popular but questionable technology, we should reach an agreement-perhaps through a consensus conference-about the initial assumptions, the role of evidence, and the burden of proof.

Now let us turn to the next debate-the debate about cost-effectiveness. It is easy to predict what will happen. On the surface, this disagreement will seem to be about evidence and methods, with people picking away at analyses whose conclusions they don't like. Underlying that, however, will be a disagreement about a much more fundamental question: Should the cost-effectiveness of mammography be considered at all? The main factor that will distinguish the two sides will be the scope of concern. The essential purpose of a cost-effectiveness analysis is to calculate the net benefit or harm to a population if resources are put into one activity rather than another. But that question does not even arise if you do not look past the one activity that interests you. In this case, if one's scope is narrowed to women in their 40s who might get breast cancer, and if one is unconcerned about (or doesn't “see”) other populations, other diseases, or the effects of shifting resources, then the conclusion is straightforward-screen younger women. From this narrowed perspective, the results of a cost-effectiveness analysis are not only moot, they are an irritant. The scope of concern of a specialty society, an advocacy organization, or a disease-oriented charity might be narrowed in this way. Conversely, if one's scope of concern is broader, encompassing other patients and other diseases, and if one understands the implications to real people of limited resources and misplaced priorities, then the results of a cost-effectiveness analysis are critically important. To use the nearby example again, if one is just as concerned about the deaths of women who develop breast cancer after age 50 years as about deaths before age 50 years, then the need to give priority to screening older women is clear. With a truly comprehensive scope of concern, a cost-effectiveness analysis becomes an essential tool for clinical decision making.

The dysfunctional debates over mammography in younger women have gone on long enough. We cannot erase the contentious disagreements over evidence, but we can learn from them. To constructively address the coming controversy about cost-effectiveness-thereby enhancing the public's confidence rather than destroying it-the first step must be to address and reach explicit agreement on the principles that should guide those discussions. Specifically, we must ask ourselves whether, in settings with limited resources, we should focus just on the benefits to younger women or broaden our concern to include all patients with all diseases. If that issue is not addressed up front, we run a high risk that the coming debates about the appropriate role of mammography in younger women will get mired in the same kind of misinformation and acrimony that have nearly destroyed the discussion about evidence.

• Copyright ©2004 by the American College of Physicians