When Good Intentions Do Bad Things
Breast cancer is a common disease – 1 in 8 women and a small number of men will be afflicted over their lifetimes. This year alone, approximately a quarter of a million new cases will be diagnosed and 40,450 women will die. That means everyone is more and more likely to have a wife, mother, daughter or friend with the disease.
Naturally, we want to “do something” about it. A just released study in the New England Journal of Medicine by Dartmouth researchers suggests that mammography screening, as presently practiced, may be the wrong prescription.
Despite widespread mischaracterization and misconception, mammography screening is not breast cancer prevention. It is an attempt to diagnose the cancer early when the prognosis is better, treatments are more effective, and before the cancer can grow into a larger tumor. But what if a large percentage of the cancers diagnosed by screening are small, indolent cancers that do not require treatment because they do not grow, spread and threaten women’s lives. This is the problem of over-diagnosis – finding cancers that would never become clinically apparent or cause death.
Assuming the incidence of breast cancer has stayed steady over time, treating small, early cancers picked by screening should eventually lead to a commensurate reduction in diagnosis of larger, more advanced cancers. The Dartmouth study looked at breast cancer incidence and size at diagnosis in 1975-1979, just before the implementation of widespread mammography screening, and cancers diagnosed in 2008-2012. They found a substantial increase in the number of small cancers discovered (162 extra cases per 100,000 women in the period after screening became prevalent) but a much smaller decrease in the number of large cancers discovered (30 fewer per 100,000 women). They posit that the difference – the 132 small cancers that screening should have prevented from progressing to large cancers – represents over-diagnosis. They also found that while breast cancer mortality decreased over the study period, screening was not the primary reason; Two thirds of the reduction was due to improved treatments, which decrease the value of early detection .
Critics claim the incidence of breast cancer has risen so that far more large cancers were prevented than the study found. In addition, they claim the study is gainsaid by multiple trials showing that mammography saves lives.
Whether the underlying incidence of breast cancer changed is unknown, but there is little evidence it increased. The American Cancer Society reports that breast cancer incidence declined after 2000, as women stopped taking post-menopausal hormone replacement therapy in response to the 2002 publication of the Women’s Health Initiative, which linked hormone therapy to increased risk of breast cancer and heart disease. The highly respected Cochrane Review found that the most reliable (well conducted) studies showed that mammography screening had little or no effect on breast cancer mortality. Reduced breast cancer mortality was only documented in trials with statistical deficiencies.
This new Dartmouth study confirms earlier studies that reached the same conclusions. A 2015 study in JAMA Internal Medicine compared different rates of screening, breast cancer incidence, and breast cancer mortality in 16 million women residing in 547 U.S. counties who were screened in 2000. It found a positive correlation between the extent of screening and breast cancer incidence— more screening finds more cancers—but no correlation between the extent of screening and ten-year breast cancer mortality. Screening discovered mostly small, early-stage cancers. There was no reduction in larger cancers or cancers that had already spread through distant metastases. Similarly, a 2012 New England Journal of Medicine review of 30 years of mammography screening showed a 109 percent increase in the incidence of small, early-stage breast cancer but only an 8 percent decrease in the incidence of advanced cancers, with virtually no reduction in the most advanced, metastatic cases.
Over-diagnosis has financial, physical and psychological costs. A Health Affairs study estimated that over-diagnosis and false positives in screening women aged 40-59 costs $4 billion annually—half the annual expenditure on mammogram screening. Over-diagnosis leads to over-treatment with surgery (with young women increasingly selecting unilateral or even bilateral mastectomies rather than lumpectomies), radiotherapy (which adversely affects the heart and lungs), and toxic chemotherapy. And multiple studies have documented the severe anxiety and depression that follows a cancer diagnosis.
If we want to help women with breast cancer we should not subject them to screening protocols that, no matter how well intentioned, end up harming them. Over-diagnosis, along with the other side-effects of screening (false positives and false negatives), can be minimized by adopting the evidence-based recommendations of the U.S. Preventative Services Task Force to start mammography screening later (age 50) and do it less frequently (every two years).