When patients come to see me, they normally have a problem that physicians call the “chief complaint.” In investigating and treating that complaint, I must obtain informed consent—communicating the risks and benefits of any proposed tests, procedures, treatments, and alternative treatments to ensure that my patient makes an informed decision about how to proceed. Gone are the days that a physician could dictate care without such consent.
In contrast, when I recommend a screening test, I’m not treating a known problem. Rather, I’m seeking out a healthy person and proposing a medical procedure to discover unsuspected disease before the patient develops symptoms. To be effective, the screening test must be sensitive enough to detect the presence of the disease and specific enough to identify people who don’t have it. There must also be effective treatments, so that catching the disease early, before it becomes clinically apparent, leads to better outcomes. Thus, screening is not directed at individuals with specific complaints but rather at large populations in an attempt to lessen their mortality risk from specific diseases.
Screening differs from traditional treatment in another important way. Unlike patients with specific problems, who must give informed consent to treatment, healthy patients are often pressured to undergo screening and are rarely provided with the knowledge needed to make an informed decision. Many medical organizations and patient-advocacy groups oversell the benefits of screening but almost never mention that screening tests, like other medical procedures, have potential harms. One-third of men undergoing prostate-specific antigen (PSA) screening for prostate cancer, for example, don’t recall being told that the test was being performed, let alone hearing about the risks and benefits.
Yet Americans have enthusiastically embraced screening. The United States performs screening, particularly for cancer, more commonly than other advanced countries. We spend tens of billions annually on cancer screening, including $8 billion for breast cancer, over $6 billion for cervical cancer, and $3 billion for prostate cancer. For some cancers, screening expenditures nearly equal the amount spent on treatment.
Why do we encourage large numbers of healthy people to undergo testing that may reveal some future problem, without informing them of the efficacy of these tests and their possible harms, while insisting that people who actually need medical intervention give informed consent? The reasons are more political than scientific. Patient-advocacy groups attract donations, gain prestige, and accumulate power by pushing screening for particular diseases. Physician-specialty groups promote screening in their areas of expertise. Most important, politicians of both parties garner support by funding screening programs for important constituent groups. Yet the effectiveness of screening is overrated: many screening programs mislead the public, waste scarce medical resources, and harm the people they are supposed to help.
The Stanford Prevention Research Center, in a comprehensive review of 39 screening tests for 19 diseases, found that reductions in death due to screening are rare. Nevertheless, patients routinely undergo overly intensive, low-value tests. About half of women who had a hysterectomy and no longer have a cervix, for example, are still getting PAP tests for cervical cancer screening. Despite the absence of evidence supporting screening in these older populations, cervical and breast cancer screenings are performed in 38 percent and 50 percent, respectively, of women over 80; and more than 50 percent of men over 75 report that their physicians continue to recommend PSA screening. It is nearly certain that these older men and women will die of some cause other than occult breast, cervical, or prostate cancer. One study found that significant numbers of Medicare patients with a known advanced cancer and a short life expectancy were still being screened for other cancers.
Most patients are unaware of screening’s side effects. False-positive results lead to additional testing, anxiety, and unnecessary and invasive diagnostic procedures. Screening can overdiagnose disease and lead to needless treatment by uncovering cases that would never have become clinically apparent, let alone life-threatening. Finally, screening generates false negatives as well—deluding patients into a sense of security and encouraging them to ignore signs of disease that they would otherwise bring to medical attention.
The contentious debate over mammography screening for breast cancer illustrates the problems surrounding screening. Following publication of several randomized trials assessing breast cancer screening over 30 years ago, U.S. medical groups recommended starting annual mammograms at age 40. Forty-two states enacted laws between 1987 and 2000 requiring insurers to cover mammography screenings in their insurance plans, and patient-advocacy groups such as the Susan Komen Foundation sprang up, supporting screening. Today, barely a week passes without a celebrity revealing her breast cancer diagnosis and urging women to get screened. Male and female politicians pledge unwavering support for mammography programs.
Yet the success of mammogram screening in preventing breast cancer deaths is greatly exaggerated. If such screening were effective, it would pick up cancers early, leading to a reduction in cancers diagnosed at advanced stages (those most likely to kill) and reduced breast cancer mortality. A recent study in JAMA Internal Medicine compared different rates of screening, breast cancer incidence, and breast cancer mortality in 16 million woman 40 years and older residing in 547 U.S. counties. The women were screened for breast cancer in 2000 and tracked for ten years. The JAMA study found a positive correlation between the extent of screening and breast cancer incidence—that is, more screening finds more cancers—but no correlation between the extent of screening and ten-year breast cancer mortality. Screening mostly discovered small cancers (less than two centimeters) and early-stage cancers. The study found no reduction in the numbers of larger cancers or in the incidence of more dangerous, locally advanced cancers and cancers that had already spread through distant metastases. These new findings mirror a 2012 New England Journal of Medicine review of 30 years of mammography screening showing 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.
What explains the lack of success? We have been screening the wrong women—those younger than 50 and older than 75—and screening all women too intensely. Breast cancer incidence rises with age, with the highest rates found in women over age 60. Mammography is less accurate in younger women, whose breasts are denser. As a result, the net benefit of annual screening is less for younger women, who have a lower absolute risk of breast cancer, a greater risk of false positive (and false negative) findings on mammography, and an increased risk of overdiagnosis. Among women who get regular mammograms over ten years, the number whose lives will be saved because of screening will be five of 10,000 40- to 49-year-old women; ten of 10,000 50- to 59-year-old women; and 42 of 10,000 60- to 69-year-old women. Over those same ten years, 61 percent of 40- to 50-year-old women who have annual mammograms will get a false positive, leading to more tests and anxiety and, for about 20 percent of these women, an unneeded biopsy. Mammograms will also miss about 20 percent of breast cancers, falsely reassuring many women and some physicians, who will disregard signs and symptoms of cancer, such as a breast mass. Finally, an estimated 19 percent to 31 percent of the cancers found prove to be overdiagnoses—early cancers that would never become life-threatening—leading to overtreatment 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.
A recent study in Health Affairs estimates that false positives and overdiagnosis in women 40 to 59 costs $4 billion annually—half the annual expenditure on mammogram screening. A report from Yale researchers in JAMA Internal Medicine found that 40 percent of Medicare’s annual billion-dollar expenditure on mammogram screening pays for screening women over 75, for whom nearly all cancers detected likely represent overdiagnosis. These older women are far more likely to die of some other ailment than they are of an early, mammographically detected breast cancer.
Much of this information is not new, yet mammogram screening remains popular. In fact, questioning its efficacy is a political third rail.
When the U.S. Preventive Services Task Force (USPSTF), a respected, independent panel of primary-care and preventive-medicine experts, advised in 2009 that mammography no longer be routinely recommended for 40- to 49-year-old women and that women between 50 and 74 have mammograms every other year rather than annually, the reaction from official medicine and women’s groups was swift and fierce. Legislators, the American Cancer Society, the American College of Radiology, the American College of Obstetrics and Gynecology, and various women’s groups condemned the recommendations as dangerous. President Obama’s Affordable Care Act (ACA), which was being negotiated during this time, requires health plans to cover preventive services, including screening, without any cost sharing (such as co-payments, deductibles, or coinsurance), where the USPSTF has found high certainty of substantial or moderate net benefit (grades A or B). In response to political pressure, the ACA was amended to cover annual mammogram screening at no cost, starting at age 40, though the USPSTF had given only a C grade (moderate certainty of small net benefit) to screening in the 40–49 age range.
Updating its guidance in 2015, the USPSTF issued draft screening recommendations (finalized in January 2016) that echo its 2009 call to raise the starting age for mammographies and decrease their frequency. The task force acknowledged that screening reduces the risk of dying from breast cancer but found that the benefit for the 40–49 age group is small, while the risks of overdiagnosis and false-positive tests resulting in unnecessary procedures are high. Balancing the risks and benefits, the USPSTF recommended against routine screening for average-risk, 40–49-year-old women and suggested that women consult with their physicians to determine if they bear greater-than-average risk of cancer that would warrant earlier screening. As it had in 2009, the USPSTF again advised biennial screening for women aged 50 to 74, rather than the annual screening that most cancer groups propose, because published randomized trials indicated no clear trend for greater benefit from annual screening versus biennial screening. Modeling studies predicted minimal benefit when moving from biennial to annual mammography but a large increase in harms (false positives, unnecessary breast biopsies, and overdiagnosed breast cancers).
In May 2015, the American College of Physicians (ACP) High Value Care Task Force published identical recommendations. Similarly, the International Agency for Research on Cancer has reaffirmed its 2002 handbook recommending mammography screening only for women between the ages of 50 and 69. These recommendations follow most other countries’ practices, where screening starts later, ends earlier, and gets done at less frequent intervals than in the United States. European women under 50 are rarely offered screening. The United Kingdom screens women 50 to 70 years old every three years. Last year, the Swiss Medical Board recommended that no new mammography screening programs be initiated and that existing ones be phased out.
In 2015, as in 2009, the reaction to the USPSTF was immediate and intense. The medical-specialty societies again condemned the USPSTF guidelines and reiterated their support for annual screening starting at 40. Only the American Cancer Society, which had initially rejected the USPSTF update, has budged: on October 20, 2015, it recommended splitting the difference between its old guidelines and the USPSTF’s, suggesting that annual screening start between ages 45 and 54, with biennial screening thereafter. Writing in the Washington Post, Democratic congresswoman Debbie Wasserman Schultz claimed that deferring mammograms “until after age 50 is dangerous” and will result in “needless deaths.” More than 60 lawmakers of both parties wrote to the Department of Health and Human Services, urging that it “not jeopardize access to these screenings.” Republican senator David Vitter warned that “we shouldn’t let bureaucrats in Washington limit access to prevention and early detection resources. These decisions should be left to women and their doctors.”
Leaving the decision to “women and their doctors” is precisely what the USPSTF was proposing. Both the USPSTF and the ACP emphasized that younger women should make the decision to screen only after consultation with their doctors, weighing the potential benefit against the possible harms, and noted that women who have a first-degree relative with breast or ovarian cancer or other risk factors, such as a known genetic predisposition, should consider early testing. The suggestion that the guidelines will cause needed mammograms to be deferred is simply incorrect. On the contrary, the unneeded, sometimes harmful, mammograms currently being performed on younger women are the ones that will be deferred. It’s the politicians who, despite all the evidence, are urging younger women to undergo screening without discussing their individual health needs with their physicians. Women should be informed of the significant risks of overdiagnosis, false positives, and false negatives that most independent review bodies feel outweigh the benefits of mammogram screening in the 40- to 49-year-old age group. Instead, a bipartisan group of lawmakers inserted the Protecting Access to Lifesaving Screenings (PALS) Act—placing a two-year moratorium on implementing the USPSTF recommendations—into the December 2015 Omnibus Spending Bill. The PALS Act ensures that women will continue to be routinely subjected to screening that starts too early and is conducted too frequently, without having the discussions they need to make informed decisions.
Screening advocates also claim that the USPSTF’s C grade for screening younger women and the recommendation for biennial rather than annual screening for women over 50 will lead insurers to stop covering screening mammograms without out-of-pocket costs—causing women to put off testing. This is unlikely to prove a serious problem. For women who don’t need these tests, limiting coverage is a good thing. For women whose risk factors justify screening before age 50 and annually thereafter, insurers will probably continue to cover screening without cost sharing. Even if some women become responsible for some out-of-pocket costs, these costs are unlikely to be a barrier to indicated screening. Poorer women will be covered without expense by Medicaid or the National Breast and Cervical Cancer Early Detection Program, funded by the Centers for Disease Control and administered by the states. Wealthier women already pay some out-of-pocket costs for nearly all their medical care. Few of these women will defer a mammogram that their physicians suggested because they have to share the relatively modest cost.
The anomaly in this debate is that screening advocates insist that tests of dubious value and known harms for healthy people be provided free of charge and with minimal or no informed consent, while tests and treatments for people with known illnesses are subject to out-of-pocket expenses. The advocates’ policy prescription seems designed to lead to overuse of screening and underuse of therapeutic testing and treatment.
Scientific evidence, not political considerations, should determine health-insurance coverage and payment policies. If we’re serious about getting greater value for our health-care dollar, we should rethink our obsession with cancer screening. The $4 billion expended annually on breast cancer screening that leads to false positives and overdiagnoses and the billions more that could be saved by curtailing spending on dubious screening for other diseases (e.g., the USPSTF recommends stopping PSA screening for prostate cancer at any age) would be better directed to cancer research and treatment—and to providing the information that people need to make informed decisions about cancer screening.
Research has found no correlation between the extent of screening and ten-year breast cancer mortality. (GARO/PHANIE/THE IMAGE WORKS)