Abstract
Mammography is the only proven frontline screening method for breast cancer. Following the demonstration of a reduction in breast cancer mortality with mammography, population mammographic screening services have been instituted, and there has been discussion in the medical literature of how to convey the pros and cons of screening to invited women. Much of the discussion has focused on the negative aspects of screening, such as false-positive and negative screens, overdiagnosis and anxiety. Also, some commentators have advocated rather cumbersome amounts of quantitative information. In this article we review the original evidence on the positive and negative aspects of screening, and show that the latter may have been exaggerated in the past. We suggest a few simple and clear points that should be made to the invited women, summarizing the positive and negative aspects without a mass of confusing statistics.
In recent years there have been a number of publications focusing generally on the information provided to women invited for screening mammography [1–5]. For the most part, these have concentrated on the negative effects of screening, such as the small short-term probability of false-positive mammograms, anxiety and so on. One aspect that tends to be overlooked is the inevitable fact that the healthcare providers would not be offering screening, or the referring physicians would not be recommending it, had they not already decided that on balance the screening was likely to be beneficial. Thus, it would seem to be illogical and unethical to decide that an intervention was beneficial and should be provided, and then to appear to attempt to dissuade the potential recipients from taking up the offer.
Therefore, when providing information to women offered mammographic screening it is incumbent on the provider to acknowledge that it has, along with a number of authoritative bodies [6–8], decided that breast screening is potentially beneficial, and to state why this decision was made. Thereafter, there is a duty to be open about the potential benefits and the potential harms of screening. It is also very important that such information is provided with clarity. There is no point in bombarding the invited women with statistics, nor in emphasizing obscure scientific concepts that they are too busy to read and assimilate. This also implies a need for brevity. Such brevity should not compromise the frankness or the clarity of the information. In addition, there should be a source of further information for those women who require it.
Women offered screening may be given information in different formats, depending on the medicolegal environment or the policy of the provider of screening. For example, the information may be on the provider's website, in an ad hoc information leaflet or in an xtra few paragraphs on an invitation letter. For brevity, all such formats will be referred to as the information sheet below.
Here we consider the specific positive and negative aspects of screening that are sufficiently important to appear on an information sheet for invited women. We will discuss the scientific issues and background in considerably more detail here than would be appropriate on such an information sheet. The implications for the latter are discussed later.
Which items of information are important?
Positive aspects
Prevention of breast cancer death
It is clear from the randomized, controlled trials that the policy of offering breast cancer screening is associated with a 20–30% reduction in breast cancer mortality [6,9]. The benefit of actually attending screening, particularly in modern, organized programs, is likely to be of the order of a 40% breast cancer mortality reduction [10,11]. The mortality benefit, i.e., the reduction in risk of dying from breast cancer associated with screening, is the reason for offering screening and should therefore have prominence in the information provided to invited women. Some commentators have argued for an emphasis on the small absolute benefit [1,2]. It is reasonable to acknowledge that for the majority of women attending there is no mortality benefit, as they will not develop breast cancer in their lifetime. However, by analogy with insurance for the minority who are unfortunate enough to develop the disease, the benefit is substantial. In addition, it should be pointed out in the information sheet that screen-detected cancers have very high survival [14]. If the screenee is among the unlucky few who have a breast cancer diagnosed as a result of the screening, the overwhelming probability is that the tumor will be curable. In addition, recent work suggests that the absolute benefit is not as small as previously thought, with an estimated saving of 3 lives/1000 women screened regularly for 10 years [12].
Example of an information sheet to be included with an invitation to attend mammographic screening.
The Department of Health, in common with many other public health bodies around the world, has determined that screening with mammography (x-ray of the breasts) saves lives from breast cancer. Only one in 16 women will have breast cancer during the age range for screening, but for those unfortunate enough to do so, being screened makes it very likely that the cancer will be curable.
The majority of women will not develop breast cancer, and so the only effect of screening for these women will be to confirm that they do not have the disease. Screening all women is necessary because we do not know who will develop breast cancer and who will not. This may be thought of as similar to paying an insurance premium.
Screening requires an x-ray of the breasts. This involves a dose of radiation. However, the dose is very small and there is no serious risk of disease or side effects from the radiation. The breasts have to be held between two plates for the x-ray to be taken, and the compression may be uncomfortable. It only lasts a few seconds, however.
You should receive the results of your mammogram in 2 or 3 weeks. For more than 99% of women attending for a screening test, no cancer is diagnosed. Around 5% of women are called back for further tests to investigate possible abnormalities on the x-ray. This can be an anxious time. You should know that for these women who are called back, most are found not to have breast cancer.
For those unlucky enough to have breast cancer, early detection by screening makes it much easier to cure with less extreme surgery and other aggressive treatments. With the coming of screening, the proportion of breast cancer patients having mastectomy (removal of the breast) has fallen to 30%.
It is possible that some breast cancers detected by screening would never have come to clinical attention if the screening had not happened. However, these represent a very small proportion of breast cancers.
Screening does not pick up 100% of breast cancers. If at any time after a normal screen, you find a lump or any unusual change in either of your breasts, you should see your doctor immediately. Do not wait until your next screening test.
Reassurance of negative results
The fact that the majority of women will not develop breast cancer, and therefore do not derive a mortality benefit, can be considered a positive as well as a negative feature. In many developed countries, there is a high level of awareness of breast cancer and many women attend for screening in order to be reassured that they do not have the disease. Around 95% of those attending will receive such reassurance from the initial mammogram result; after assessment of those with suspicious mammograms, more than 99% of attenders will have the reassurance [13].
Avoidance of more aggressive therapy
With very few exceptions, small cancers that are screen-detected have excellent prognosis with local surgery and without adjuvant chemotherapy [15,16]. With the introduction of screening, rates of mastectomy have fallen and rates of wide local excision have increased [17]. Thus, attending for screening, for those who do develop breast cancer, makes it much more likely that both radical surgery and aggressive adjuvant therapies that significantly diminished quality of life can be avoided. This should also be conveyed to the invited population.
Negative aspects
Effects of the screening process, particularly on those who do not have breast cancer
As noted above, the majority of women screened do not gain any mortality or therapeutic benefit because they do not have breast cancer on the occasion of screening, nor will they develop the disease in their lifetime. This fact should be featured on the information sheet. In addition, potential screenees should be told that the screening involves compression of the breast, which may cause discomfort or pain, and that during the examination they will receive a small dose of radiation. The discomfort from compression is very brief, and the radiation dose is at a level associated only with theoretical risks, and not with any observed morbidity, but the invited women need to be informed of both. They should also be reassured that the radiation dose is small and does not confer serious risk of disease, to avoid causing undue anxiety.
This brings us to the question of anxiety. Anxiety associated with screening is transient, but it is real. Do women need to be told that being tested for a serious chronic disease may cause them anxiety? Our opinion is that they do not, and that such a statement on an information sheet might be considered an insult to their intelligence. Indeed, in some cases of ‘worried well’, the screening may alleviate anxiety. It is, however, appropriate to mention the anxiety that may arise in the case of women recalled for further investigation (see below).
Recall for suspicious features that ultimately prove benign
It should be made clear that a screening mammogram is not a definitive diagnostic test – it simply indicates whether further tests are needed or not. Invited women should know in advance that a minority (5–10% in many programs, varying by program and by whether it is a first or subsequent screen) of screenees will be called back for further examinations. They should also be informed that only a minority of those called back (5–15% in most programs) will actually have breast cancer. The information sheet should acknowledge that this may be an anxious time for those called for assessment [18], but the knowledge that most of these women do not have breast cancer and that the few who do usually have excellent prognosis may counter this.
Screening may miss some cancers & cause false reassurance
The invited women, and more importantly, the women attending for screening, should be told that some tumors are not seen on a mammogram. Women should seek medical advice for any unusual breast changes, even if they have recently had a negative screen. Invited women should be warned on the information sheet not to delay seeking advice about such changes until their next screening appointment. Sensitivity of mammography is generally good, but even if it were perfect, some new tumors will arise after a negative screen and before the next scheduled screen.
Possibility of overdiagnosis
This is perhaps the most difficult issue, and a matter of considerable conjecture and debate with respect to the information to be provided to potential screenees, especially in the context of ductal carcinoma in situ (DCIS). It is suspected that at least some screen-detected breast tumors would not have been diagnosed within the lifetime of the women in the absence of screening, but at present, there is no way of knowing which few tumors may truly be indolent. Since the majority of breast tumors do progress and will eventually be life-threatening, all cancers diagnosed must be treated. In the past, rather sweeping assertions have been made regarding overdiagnosis, suggesting proportions of the order of 20–30% of tumors overdiagnosed within screening programs [3,19]. Such assertions are simplistic and incorrect, as they are usually based on observed increases in incidence of breast cancer in conjunction with the initiation of screening programs. However, these increases in incidence are largely made up of cases that would have occurred later, but whose diagnosis has been advanced by lead time, which is a precondition to any benefit of screening. In addition, other factors, such as pre-existing trends in increasing incidence of the disease, and the necessity to identify screened or invited cohorts when estimating overdiagnosis, are often neglected. Studies that take such complexities into account tend to arrive at much more modest estimates of overdiagnosis, of the order of 10% or less [20–22].
This brings us to the issue of overdiagnosis in the context of DCIS. The proportion of screen-detected DCIS that would have progressed to invasive disease had it gone undetected and untreated is not certain. It has been suggested that only a minority of DCIS cases would progress to invasive disease if left untreated, on the basis of long-term observation for subsequent invasive cancer in cases previously misdiagnosed as benign disease and therefore not treated [23]. However, DCIS that is misdiagnosed as benign is not a representative sample of all DCIS. Recent research based on rates of DCIS and invasive cancer [24], and on the pathological features of DCIS [25], suggests that the majority of screen-detected DCIS has the potential to progress if left untreated.
In total, the best evidence indicates that over-diagnosis is at worst a minor phenomenon. Our own research suggests that less than 5% of tumors at a first screen are overdiagnosed, in keeping with the logic that overdiagnosis is an extreme form of length bias, and that at subsequent screens overdiagnosis rates are close to zero [21]. Even if we assume pessimistically that the figure at first screen is 10%, a woman attending her first screen has a less than one in a thousand chance of an overdiagnosed breast cancer.
Unnecessary treatment or overtreatment
This is closely related to overdiagnosis, as tautologically any treatment of an overdiagnosed lesion is overtreatment, but the best quality evidence indicates that overdiagnosed lesions are rare. While the issue of overtreatment of (nonoverdiagnosed) screen-detected cancers is relevant in the larger context of screening, it is essentially a matter for multidisciplinary diagnosis and treatment decisions, and not the screening information sheet. What is needed is to avoid applying treatments more suitable for advanced disease to early stage, screen-detected cancers. It is clear that all but a small and identifiable subgroup of very small breast cancers have excellent prognosis without aggressive adjuvant therapies [15]. It is important to tailor the treatment to the characteristics of the tumor.
Future perspective
In the next few years, various technical innovations are likely, including the more widespread use of digital mammography, the development of automated ultrasound and the referral of groups at high genetic or familial risk for magnetic resonance imaging. All of these are likely to change false-positive and false-negative rates. The qualitative advantages and disadvantages of screening will remain, but the magnitude of the benefits is likely to increase, while that of the disadvantages will probably decline.
Conclusion
From the above it can be seen that there are a number of positive and negative aspects of mammographic screening that the provider has a duty to convey to the women offered screening. It appears that whatever has been provided on information sheets in the past, the harms are probably overemphasized in papers in the medical literature. For purposes of information to the public, the benefits and harms can be conveyed in a nonpatronizing but comprehensible manner, but it is probably necessary to avoid excessive numerical detail. In addition, if detailed quantitative estimates were provided, there would arguably be an ethical obligation to provide ranges of uncertainty on these. An example of the sort of information that might be provided with an invitation to screening is shown in Box 1. The numbers quoted pertain to the UK Breast Screening Programme, but the points can readily be adapted to other programs.
Executive summary
Introduction
There is a need for clear and simple information for women invited to mammographic breast screening.
The medical literature in the past has tended to emphasize the disadvantages of screening.
There is uncertainty about the appropriate quantitative information to convey in order to inform decision making.
Prevention of breast cancer deaths - this is the most important point.
Reassurance of negative results - women attend screening not to be told they have breast cancer, but to be told they have not.
Avoidance of more aggressive therapies - mastectomy rates have fallen sharply in the epoch of mammography.
Effects of the screening process, particularly on those who do not have breast cancer - screening involves compression and radiation exposure for all screenees, but only the few who actually have breast cancer benefit as a result.
Recall for suspicious features that ultimately prove benign - a small percentage of women without breast cancer will undergo further diagnostic tests as a result of suspicious features on the mammogram.
Screening may miss some cancers and cause false reassurance - women should be told that no screening test is 100% sensitive.
Possibility of overdiagnosis - there may be a small number of cancers diagnosed at screening that would never have arisen clinically in the absence of screening.
Unnecessary treatment or overtreatment - there is a need to apply less radical treatment to very early tumors with limited aggressive potential.
Advances in screening technology will mean that information provided to invited women will have to evolve over time.
Harms of screening have probably been exaggerated in the past. A general, nontechnical summary of the benefits and disadvantages of screening is indicated, with minimal and simple quantitative information.