How Presentation of Recurrence Risk Influences Decision-Making

[gdpawel]

I've always known about the pervasive way clinical trials focus on the relative risk (which powerfully exaggerates the benefits of drugs) and drug companies frame the question in terms of relative risks (systematically inflates their value), and absolute risk.

The number needed to treat (NNT), developed in 1988 to avoid the confusing distinction between "relative" and "absolute" reduction of risk, is perhaps one of the most important, least recognized, and most emblematic distortions you can find.

Some years back, the NCI issued a clinical alert to oncologists announcing the results of several clinical trials showing that women with node negative breast cancer benefited from chemotherapy. According to "number needed to treat" analysis, one hundred women would have to undergo chemotherapy for 10 to benefit.

Ninety women would risk toxicities but get none of the benefits. So what is the harm? The toxicities included not only those that can end your life like heart failure and leukemia, but some of those that can ruin you life like loss of cognitive function, loss of libido, severe arthritis and risk of bone fractures. These harms are usually ignored or understated. One of the reasons is because they are understudied.

So it began the "standard" practice to administer chemotherapy to women with node negative breast cancer that still exists today. Treat everyone to improve the survival chances of a small minority. How will the new gene profiling tests for prognosis be used in the real world today? Will women choose chemotherapy even though they have only a small chance of a recurrence? The bias towards chemotherapy and its overuse still permeates our society and will affect how these profile tests are used.

Many women will opt for chemotherapy even for a one or two percent benefit. Will women consider a low risk result low enought to forgo chemotherapy, or will they persue it anyway because of historic bias?

The clinical alert mentioned above was issued in 1987, a year before the NNT was developed to avoid the confusing distinction between "relative" and "absolute" reduction of risk.

A more honest use of NNT is not just an issue of forthrightness, it is also cost-effective.

According to a previous NYT article, physicians, like Dr. Eric P. Winer, who directs the breast oncology center at the Dana-Farber Cancer Institute in Boston, are taking their own best shot at figuring out who really benefits from chemotherapy. He asks how sensitive the tumor is to estrogen, how aggressive a pathologist believes it is, how big it is, how much has spread to the lymph nodes and whether its surface has a type of protein, HER2, that is associated with a better response to chemotherapy. After talking through the decision with his patients, he is comfortable omitting chemotherapy in some who would have had it not long ago.

A statistically small reduction in risk my be very important to some women, while for others chemotherapy is not worth it. Is it a tuff decision to take something potentially toxic when you have a 90 percent chance of being cured without it? Studies like this reaffirm the way many oncologists still practice. It should help understanding to base cancer therapies on the "specific" characteristics of their patients.