Suzanne W Fletcher
Screening for breast cancer
Harvard Medical School, BostOIl, MA, USA
E vidence about screening for breast cancer dem onstrates how principles of clinical epidem iology can be used to aid in decisions about clinical care and health care policy. R egard-less of condition, three key questions m ust be asked and answ e-red before undertaking screening: 1) H ow great is the burden of suffering caused by the condition? 2) H ow effective is early treatm ent follow ing screening (as opposed to results of treat-m ent w hen the condition w ould be diagnosed w ithout a screen-ing program )? and 3) H ow good is the screenning test in term s of sensitivity, specificity, predictive value, sim plicity, cost, safety, acceptability and labeling effects?
B reast cancer incidence and m ortality varies by coun-try. In the U nited S tates, breast cancer is the m ost com m on non-skin cancer and second m ost deadly cancer in w om en. T here are three m ain tests used, either alone or in com bina-tion, for breast cancer screening: m am m ography, clinical breast exam ination by a health professional, and breast self-exam ination.
M am m ography, w ith or w ithout clinical breast exam i-nation clearly reduces m O ltality in w om en aged 50-69, but the effect in younger w om en rem ains unclear even after trials involving approxim ately 170,000 w om en. E ight random ized controlled trials all show ed m ortality reductions for w om en over 50 years of age; at seven years of follow -up, the com -bined result show ed a 34% m ortality reduction that w as sta-tistically significant. T hese sam e studies w ere inconsistent in finding m ortality reductions in w om en ages 40-49, com bined
analyses show ed no effect at seven years of follow -up ahd a 10% to 17% statistically insignific Iant reduction at 10 to 12 years. T his delayed and sm aller trend in younger w om en m ay be due to the fact that m any w om en entering trials in their forties had their cancers diagnosed w hen they w ere in their fifties, at an age w hen m am m ography is know n to be effec-tive. O nly one random ized controlled trial has been done com -paring m ortality differences after either m am m ography or a very through clinical breast exam ination. A t seven years of follow -up, there w as no difference in m ortality rates. N o ran-dom ized trials of breast self-exam ination have reported m or-tality data.
M am m ography is the m ost sensitive screening test cur-rently available for breast cancer screening. T he high false-positive rate of m am m ography can lead to substantial num -bers of w om en experiencing anxiety about breast cancer, the possibility of overdiagnosis w ith "pseudocancers"is real, and the high cost of m am m ography in the U nited S tates m eans already alm ost $3 billion is spent annualy on m am m ography, about half or w hich is for w om en under age 50.
W hen preventive policies are prom ulgated, strong evi-dence of benefit is essential because so m any people m ust be screened to potentialy help one person. W hile m uch m ore needs to be done to ensure that all w om en aged 50-69 recei ve routine screening in countries w here breast cancer on an im -portant health risk, m ore research is necessary before insti-tuting routine breast cancer screening in younger w om en.