Tiotropium in Combination with Placebo, Salmeterol, or Fluticasone–Salmeterol for Chronic Obstructive Pulmonary Disease: Possible Confounding Effect of Treatment Withdrawal?

Abstract

In the updated systematic review on the effects ofscreening mammography in women age 40 to 49 years, Armstrongand colleagues (1) used relative effect measures to express the mostimportant benefits of screening mammography, while expressing themost important harms of screening mammography in terms of ab-solute effect measures. Notably, they estimated that screening mam-mography is associated with relative risk reduction in breast cancermortality of 7% to 23%, whereas 30 to 200 women age 40 to 49years will die after annual screening mammography as a result ofradiation-induced breast cancer (1). It would be more appropriate,however, to express benefits and harms using the same effect mea-sures.It is particularly important to contrast net benefits and net risksin patients with and without screening mammography. Of crucialimportance for any woman contemplating screening mammographyis to understand what outcome she can expect with and withoutscreening mammography. If individualized recommendations calledby the American College of Physicians (2) is to become reality, esti-mates of net benefits and net harms of screening mammography inwomen with different baseline risks for death from breast cancerwithout undergoing screening mammography should have been in-cluded in the recent guidelines and systematic review (1, 2). If Arm-strong and colleagues had presented data in such a way, they mayhave calculated that, in the best-case scenario for women at “averagerisk” for breast cancer, screening mammography may save 46 livesper 100 000 women screened over 10 years, at the cost of 30 per100 000 lives lost. In the worst-case scenario, the net benefit is ac-tually negative, and more lives will be lost than saved (Appendix,available at www.annals.org).We believe that presenting data in this way may better helpwomen understand benefits and harms of screening mammography,which need to be compared with their own individual risk for breastcancer. Only in this way can a woman truly exercise her preferencesand values and preserve her autonomy consistent with ethical andrational decision making (3).