This week the US Preventive Services Task Force (USPSTF) revised their recommendations for screening mammograms. Their recommendations have ignited much controversy and have weakened the credibility of a formerly objective scientific body.
This post is longer than usual. It deals with an important subject in some detail. For the readers who like to delve into the details and see the data, set this aside for when you can give it some time, follow the links, and check out the articles yourself. If you just want an executive summary, skim for the bold face type and read the conclusion.
What did the USPSTF recommend?
My regular readers know that the USPSTF is a national body of scientists who periodically assess the medical literature and provide recommendations about preventive tests and treatments. Because they are unaffiliated with any specific interest group, the USPSTF developed a reputation as the most objective and unbiased source of medical recommendations. For every test or therapy they reviewed, they weighed patient benefit against the potential for patient harm (regardless of cost) and reported whether the intervention was beneficial, harmful or that there is insufficient evidence to decide. I personally looked to their recommendations and wrote about them frequently.
The USPSTF’s last review of the literature regarding mammograms was in 2002. At that time they recommended a mammogram every one to two years for women aged 40 and older.
This week, in the Annals of Internal Medicine they published their new recommendations, supported by two articles detailing the scientific evidence that was reviewed to reach their conclusions. Their current recommendation (link 1 below) for mammograms is
The USPSTF recommends against routine screening mammography in women aged 40 to 49 years. The decision to start regular, biennial screening mammography before the age of 50 years should be an individual one and take into account patient context, including the patient’s values regarding specific benefits and harms.
The USPSTF recommends biennial screening mammography for women between the ages of 50 and 74 years.
The USPSTF concludes that the current evidence is insufficient to assess the additional benefits and harms of screening mammography in women 75 years or older.
So the major changes from 2002 are that
- mammograms are no longer recommended between the ages of 40 and 49,
- between 50 and 74 mammograms are recommended every two years, rather than every one to two years, and
- there is insufficient evidence about the benefits and harms of mammograms in women 75 years and older.
Shockingly, the first two points are entirely contradicted by the scientific evidence on which the recommendations are based.
What are the benefits and harms of mammograms for women in their 40s?
The article which examines the harms and benefits of mammograms in different age groups can be found at link 2, below. Table 1 (Pooled Risk Ratios for Breast Cancer Mortality From Mammography Screening Trials for All Ages) is a very helpful review of the pooled data from all the randomized trials for mammography in different age groups. The table shows how many women are needed to be invited to undergo mammograms in each age group to prevent 1 death from breast cancer. The numbers for women in their 60s are terrific. 377 women in their 60s need to be offered mammography to save 1 life from breast cancer. (The lower the number of women who must undergo mammograms to save one life, the more beneficial the intervention.) In their 50s 1,339 women must be screened to save a life, and in their 40s 1,904 must be screened. So mammography is still life-saving for women in their 40s it just saves fewer lives than for women in their 50s and 60s. In fact, the abstract of the article concludes
Mammography screening reduces breast cancer mortality for women aged 39 to 69 years.
So if mammograms have a life-saving benefit for women in their 40s, albeit a much smaller one than for women in their 50s or 60s, why would they not be recommended? Perhaps the harms caused by mammograms in that age group outweigh the benefits.
The article also methodically reviewed harms caused by mammograms. The risk posed by radiation exposure due to mammograms was studied without conclusive evidence of significant harm. The pain, anxiety and distress associated with undergoing mammograms and being told about potentially abnormal results were also studied. Most women surveyed reported that the pain and anxiety would not deter them from future mammograms. The most important harms that may result from mammography is a biopsy to determine if an abnormality is benign or malignant, additional imaging that may be required, and overdiagnosis, which means the diagnosis of breast cancer that is so slow-growing it is unlikely to shorten the woman’s lifespan.
Table 2 (Age-Specific Screening Results From the Breast Cancer Surveillance Consortium) summarizes the potential harms by age per 1,000 women screened. We would expect the harms in women in their 40s to outbalance the benefits and lead to the recommendation against mammograms at that age. The left column in the table deals with women in their 40s. Recall that it takes 1,904 women to be screened to save one life from breast cancer in that age group. Since the table shows the harms per 1,000 women screened, if we multiply each number by 1,904/1,000 we discover the harms that result from saving one life in this age group. So to save one life from breast cancer by screening women in their 40s with mammograms we would have to inflict 186 women with false positive mammogram results (abnormal mammograms without any cancer present), 161 women would have to undergo additional imaging and 18 women would have to undergo biopsies.
Now, I don’t want to minimize the misery involved in an unnecessary breast biopsy (unnecessary only in hindsight, of course). The procedure itself is uncomfortable and waiting for results is torture. Still, 18 biopsies to save one life strikes me as a wonderful trade-off, especially given that saving a life in the 40s yields many years of additional living. How can the USPSTF possibly judge that these harms outweigh the benefits? They hedge their recommendation by saying that the decision should be individualized for each patient, but that’s true for all medical decisions. Why then recommend against it?
Because it’s sometimes difficult to understand the numbers involved, I ask you to do the following thought experiment. You are a physician addressing an auditorium of 1,904 women in their 40s. You tell them “If you all have mammograms before you turn 50 one of your lives will be saved from breast cancer compared to just waiting until you all turn 50. Some of you will get breast cancer and die of it whether you get mammograms or not, and most of you won’t get breast cancer. If you start mammograms now, 18 of you will get biopsies, and a lot of you will get incorrect initial mammogram results requiring more mammograms or ultrasounds or MRIs. But only one of you will have her life saved from breast cancer.” Each woman can and should decide for herself but they’d like a recommendation from you.
I can’t imagine not recommending mammograms under this circumstance. Of course some women may choose against mammography, but to me the harms seem minor compared to the benefits.
How frequently should women have mammograms?
The decision to recommend biennial mammograms is based on a study published in an article at link 3, below. The study uses various models to compare the harms and benefits from different mammogram frequencies. But this article does something that has generally been outside the purview of the USPSTF. It considers cost.
The results state that having a mammogram every two years had 81% of the benefit of annual mammograms with far fewer false-positive results, unnecessary biopsies and additional imaging. And they state explicitly that annual mammograms consume more resources than mammograms every other year.
There are two important objections to this. First of all, getting only 81% of the life-saving benefits of annual mammograms may not strike many women as a victory. Why not capture all the benefits by having the test annually? The answer might be the additional harms, but as we’ve seen in the previous section, the harms are quite modest and most women would not be deterred by them. The second objection is that previously the unique value of the USPSTF was that it advised physicians about the value of tests regardless of costs. That allowed us to give patients the best advice we could, and allowed patients to decide if the cost was worth the benefit. Now, the USPSTF is deciding for us that annual mammograms are too expensive.
So, I and many other physicians will advise women to have annual mammograms starting at age 40. And in my opinion they should continue to do so until they are ill or frail enough that a new diagnosis of breast cancer would be unlikely to shorten their lives. For some women this might be in their 70s, for others, in their 80s.
But the broader outrage is that a group that I used to turn to without hesitation for objective review of the scientific data, a group that informed my daily advice to patients, has provided recommendations inconsistent with their own data and motivated by economics. Patients and physicians have neither the skills nor the time to scour the world literature on important topics. Who will fill the role that the USPSTF abandoned?
(3) The Annals of Internal Medicine article studying how frequently mammograms should be done: Effects of Mammography Screening Under Different Screening Schedules: Model Estimates of Potential Benefits and Harm
New York Times article: Panel Urges Mammograms at 50, Not 40
LA Times article: Mammogram guidelines spark heated debate
Wall Street Journal Editorial: A Breast Cancer Preview
Important legal mumbo jumbo:
Anything you read on the web should be used to supplement, not replace, your doctor’s advice. Anything that I write is no exception. I’m a doctor, but I’m not your doctor despite the fact that you read or comment on my posts. Leaving a comment on a post is a wonderful way to enter into a discussion with other readers, but I will not respond to comments (just because of time constraints).