Prevention is better than cure? Not necessarily
ESSAY: A leading cancer expert says Gordon Brown’s disease-screening initiative could cause unnecessary distress and diminish autonomy.
On Monday 7 January, UK prime minister Gordon Brown launched a new NHS initiative in public health policy by repeating that wearisome cliché: ‘Prevention is better than cure.’ This was followed by his announcement of a raft of screening initiatives aimed at achieving immortality for all. Speaking on the occasion of the sixtieth anniversary of the founding of the NHS, Brown said his government would launch screening programmes (or ‘health MOTs’) to detect the risk of heart disease, stroke, diabetes, kidney disease and other ailments. It all seemed to take the medical profession by surprise. Who, if anyone, had been consulted?
I have served the National Health Service loyally for over 40 years and I feel that I deserve a campaign medal for every reorganisation or new initiative I’ve faced, taken by governments of all political hues. But this latest one I regard as Nature’s way of telling me to resign my commission without facing another battle. So by way of a swansong, let me explain to our prime minister, whose political party my family has actively supported through three generations, why he might just want to take further advice on these new proposals.
First of all, screening is not prevention of disease but detection of asymptomatic disease. Screening theory has it that the detection of the earliest stages of a disease before it creates symptoms, followed by prompt treatment, might postpone death, but of course not prevent death. Death is inevitable, like taxes and NHS reorganisations. Prevention literally means avoiding the onset of a specific disease and presupposes that we fully understand the aetiology of the disease in question. However, the fundamental problem with both the prevention of disease and the early asymptomatic diagnosis of disease is that the whole population is exposed to an intervention and its unwelcome side effects for the benefit of a minority. In contrast, cure of a disease is targeted at the individual who reaps the benefits and ‘pays the price’. And by ‘paying the price’, I’m talking about inconvenience and toxic side effects rather than cash up front.
In the practice of bedside medicine there is an ethical imperative of autonomy and informed consent. As a surgeon and an oncologist I have to inform the patient of the benefits versus the harms of my intervention whether it’s by surgery or the prescription of drugs. With the expansion of evidence-based medicine (EBM) in oncology, we can do this with increasing precision and individualisation. However, with the practice of public health interventions of screening and prevention, the evidence base tends to be less robust and the ethical model of respect for autonomy very poorly developed.
At one extreme, of course, advice on healthy lifestyles that include the avoidance of tobacco products, regular modest exercise, plenty of fruit and vegetables and a little red wine, can indeed extend life expectancy and enhance quality of life with little in the way of toxic side effects (1). Providing the government doesn’t become all sanctimonious about this, with thought police, exercise police, diet police, smoking police and by withholding healthcare for ‘self-inflicted’ sickness, I have no problem with promoting such good advice. The problem, though, becomes much more difficult when the intervention is a test or a ‘treatment’, or a ‘treatment’ as the consequence of a test.
Before we can sanction these kinds of interventions we must have robust evidence from randomised controlled trials of large and representative populations, which shows that the test or the treatment can reduce mortality from the disease with a favourable benefit/harm ratio. In addition, we need to be confident that the opportunity costs of such a programme don’t consume too many resources as to threaten the funding for treatment and cure of symptomatic disease. These are not the words of some maverick but can be taken as a summary of an excellent book, Screening, Evidence and Practice, by Angela Raffle, consultant in public health and the national screening programmes, and Sir Muir Gray, programmes director of the UK national screening committee 1996-2007 (2).
So, prevention is better than cure? I’m not so sure, Mr Brown - your programme of prevention could end up causing unnecessary stress amongst the population while also undermining our autonomy.
article continues after advertisement
‘Relative risk reduction’: is it worth it?
‘Whereof one cannot speak, thereof one must be silent’, stated Wittgenstein. Quite so Ludwig. But I can speak with authority about my own health and also about breast cancer, both of which subjects can shed some light on the current debate about prevention over cure.
To start with, my firsthand experience can inform and illustrate the discussion very nicely, as it describes the best of preventative action already available in the NHS, long before the PM’s announcement. My youngest brother David died of a massive heart attack at the age of 59 while in office as president of the Royal College of Paediatrics and Child Health. After a period in mourning I took myself off to see my local GP, Dr Chris Page, who counselled me on the harms and benefits of screening for risk of heart disease. He explained a risk assessment model factoring in family history, blood pressure, smoking history and blood lipid levels. He went on to explain how for each increasing increment of risk there were interventions of known toxicity that could modify this risk by a measurable quantity that could be described in absolute numbers rather than relative risk reductions. All of this was illustrated by coloured decision aid diagrams. I came to a well-informed decision that it would be in my best interest to have my blood pressure monitored and my serum lipids measured. (Smoking advice was not required, since I’d never taken up smoking.)
My blood pressure was mildly elevated, as was my total cholesterol. So for the past two years I’ve been taking a statin and a mild diuretic without toxic side effects. For this inconvenience, I’ve perhaps reduced my risk of a myocardial infarct by about 1 in 10 (3). This experience shows the best of evidence-based prevention within the NHS; I now wish to compare this with my knowledge of what’s on offer for women by way of prevention and screening for breast cancer, where the serious problems with prevention-as-policy become clear for all to see.
Most of my career has been devoted to the research and treatment of breast cancer, and I have been in the front line in the development of screening and prevention of the disease. Furthermore, my family has been afflicted with breast cancer, making the challenge up front and personal (4).
Let’s start with prevention. Most of the risk factors for breast cancer are beyond our control - that is, sex, age, race and genetic inheritance. Some risk reduction might be possible by adopting the healthy lifestyle behaviours described above together with the avoidance of binge drinking and, if there is a choice in the matter, starting a family before the age of 30. However, we are still searching for the holy grail of chemoprevention of breast cancer. In 1985, Jack Cuzick and I published a very important observation in the Lancet (5), describing how women with breast cancer who were treated with adjuvant tamoxifen demonstrated a significant reduction in the risk of a new breast cancer of the other breast. This observation was confirmed by other studies and led to the launch of the IBIS 1 trial for the prevention of breast cancer with tamoxifen amongst women at a high risk. This and similar trials confirmed that tamoxifen could lead to a relative risk reduction (RRR) in the incidence of breast cancer by more than 30 per cent at the cost of some significant side effects (6). That study provided an excellent example of how to calculate benefit/harm analyses that might inform public policy and allow the individual subject to judge for themselves.
So, what is ‘relative risk reduction’, and how useful is it? Most statisticians and epidemiologists describe risks in relative terms such as 50 per cent increase or 25 per cent reduction. To translate this into simple numbers demands an understanding of the background risk. Let me provide a simple illustration. Say a young student is involved in field-work in Siberia and is keen to come home for Christmas, but he only has £300 to spend on air fares. A flight on BA would cost £500 but a flight on Air Uzbluchistan (AU) is within his budget. Just before he travels he learns that AU has a less than perfect safety record with a 50 per cent increase in the chance of crashing compared with BA. To understand that risk, which sounds very alarming, he needs to know BA’s safety record. He then learns that BA only crashes one in every 2,000,000 flights (ie, risk is 0.000002 per cent); thus, a 50 per cent increase in that risk is an extra 1:1,000,000 (ie, an absolute risk of 0.000003 per cent). In other words, a 50 per cent increase of a very small risk may be a risk worth taking; if our student in Siberia was sensible, he would take the cheaper and ‘riskier’ flight. The reverse is also true: a 50 per cent decrease of a very small risk is a very small gain.
Now let us compute the absolute benefits of chemoprevention of breast cancer. The normal risk for women is about 2:1,000 a year. The women in IBIS 1 had a background risk of about three times that, 6:1,000 a year or 6 per cent in a decade. A 30 per cent relative reduction risk (RRR) would mean that about 2 per cent might avoid breast cancer over a decade but at the cost of increased hazard of thrombosis, endometrial cancer and other gynaecological problems (6). In the USA it is commonplace to accept this trade-off, but as for myself, a co-author of the study, I don’t think the gain is worth the pain. Yet we still continue the search. The IBIS 2 trial is of a similar design, but instead of using tamoxifen it is using the aromatase inhibitor, anastrozole, which has been shown to be twice as effective as tamoxifen in preventing contralateral breast cancer with a much better safety profile (7). Who knows, that might be a preventative with a sufficient benefit/harm ratio to win over the informed woman who is at increased risk of breast cancer.
The dangers of breast cancer screening
Now let us adopt the same principle for mammographic screening for breast cancer. Proponents of screening often parade women who claim that ‘screening saved my life’. Most lay people and screening zealots think that is the killer argument. Well, in a way it is - it kills off the debate at a stroke, not because it can’t be rebutted but because people like me find it difficult to be unkind. But I believe the time is long past when we have to patronise womenfolk in order to retain our popularity.
Every woman can interpret her screening experience in a way that reinforces her decision to accept the invitation. The result was negative: ‘Thank God for the reassurance.’ The result was a false alarm: ‘Well you can’t be too careful.’ The result showed duct carcinoma in situ (DCIS): ‘Thank God it was caught before it had a chance to spread.’ The result showed invasive cancer: ‘Thank God it was caught early.’ All of these sound like sensible and of course understandable reactions to screening programme results. But let us examine these four scenarios, and ask what they really reveal about the problem of prevention in public health.
Just how much reassurance is a negative result worth? For a start, most women overestimate their risk of developing breast cancer (8). If we stick with post-menopausal women, then the annual risk of developing breast cancer in a normal population is 2:1,000 a year, and if, as in the UK, screening is at three yearly intervals, then the accumulated risk over three years is 6:1,000. Therefore in each period 994:1,000 women might expect not to develop the disease. Assuming that the interval cancer rate (ie, those cancers that develop between the first and second round of screening) is about a third of the screen detected rate (9), then only an additional 4:1,000 woman over three years win extra reassurance.
In the second scenario - the false alarm - no woman has a gain. It’s like thanking the fireman for rescuing you from the fire when he threw you in to begin with! This false alarm and unnecessary surgery is of no value to the woman. It can even be damaging in a subtle way: if as a chance finding the benign nodule or scar is close to an area of atypical ductal hyperlasia (ADH) or in situ lobular carcinoma, then the woman is classified as being at an increased risk. What does she do then apart from worry?
The third scenario - ‘thank God it was caught before it could spread’ - follows the detection of duct carcinoma DCIS. In about 30-40 per cent of such cases the disease is multi-focal and the woman is advised mastectomy for this ‘early’ breast cancer (10). Yet if left undetected some might regress and others might co-exist with the woman for the rest of her life (11). Whenever a screening programme starts, 20 per cent of the screen-detected cancers are DCIS and screening theory would suggest that their detection would ultimately lead to a fall in invasive cancers in the population. Not so; there is also a long- term increase in the incidence of invasive cancers (12, 13).
Finally the fourth scenario, where invasive cancer is detected. Yes indeed, a life might have been saved. As I have described, and also have been backed up by many others (14,15), the estimate of the number of lives saved is about 1:1,000 per 10 years of screening. (Note this is two orders of magnitude less than the benefits I described for the prevention of heart disease.) Yet it does not follow that the screen detection of each cancer represents a life saved. If left to nature, the cancer might have progressed slowly, become clinically obvious and then have been cured by treatment on presentation. After all, we are close to curing 75 per cent of cases with modern therapy (16), and screening is good at detecting ‘good’ cancers. It’s the bad cancers that slip through the net and appear as interval cases (17). Furthermore some of the cancers are so slow-growing that if undetected they would never appear in a woman’s lifetime. These are what American health writer H Gilbert Welch describes as pseudo-cancers (18). Last of all, we never know for sure if indeed a life has been saved when mammography detects a cancer. A small grade III cancer has a lethal capacity, even if caught ‘early’. The biology of a tumour is more important than its chronology.
For too long the mantra of screening - ‘catch it early and we will save your life and save your breast’ - has been allowed to unchallenged. Breast cancer is too complex a problem for such a facile solution. For too long women have been patronised and coerced into screening. And now the government wants to extend screening for breast cancer to women below the age of 50, in the face of the recently published UK trial that showed no significant advantage for this age group (19).
I trust, or at least hope, that the UK Department of Health will learn from the errors of the past and not be rushed into programmes of screening for prostate cancer, bowel cancer, ovarian cancer and other pathology without having evidence-based harm/benefit analyses to offer the individual to allow them an informed choice. Though if Brown’s new focus for the NHS is anything to go by, the DoH seems to be embracing screening rather than questioning its benefits and efficacy.
In discussing the ethical issues of screening we must accept a tension that exists between ‘Utilitarian’ principles and those of ‘Autonomy’. Utilitarianism involves social engineering for the ‘greatest good of the greatest number’, whereas autonomy assumes the individual has an informed choice when health interventions for ‘their own good’ are considered. Social engineering and coercion might be acceptable for hygiene and substance abuse, but when the balance of benefit versus harm is a close call then surely the right to self-determination trumps the principle of utilitarianism. Nowhere is this truer than in the area of screening for cancer. We screen for cancer to reduce cause-specific mortality without an increase in all-cause mortality and at an acceptable cost in terms of medical morbidity. Even where there is level-one evidence from RCTs of a reduction in cancer-specific mortality, the benefit in absolute risk reduction may be so small that the individual should have the right to make a personal trade-off against the undoubted harms of false alarms, over-diagnosis and radical treatments for diseases that, if left to nature, would never announce themselves in a lifetime.
I therefore propose that the uncritical promotion of screening is unethical by modern ethical standards and reflects a paternalistic attitude that would be unacceptable for treatment aimed at curing established disease. Gordon Brown and his government should seriously rethink making screening and prevention their big idea for public health policy in the coming years. In the area of established disease, focusing on prevention can be dangerous, and very often cure is better than prevention.
Michael Baum worked for 30 years as a surgeon specialising in breast cancer, and is now professor emeritus of surgery at University College London.
(1) Khaw KT, Wareham N, Combined Impact of Health Behaviours and Mortality in Men and Women: The EPIC-Norfolk Prospective Population Study, PLoS Med. 2008 Jan 8;5(1):e12
(2) Screening Evidence and Practice, Angela Raffle and Muir Gray, Oxford University Press, Oxford 2007
(3) Fidan D, Economic analysis of treatments reducing coronary heart disease mortality in England and Wales, 2000-2010, QJM 2007 May;100(5):277-89
(4) Baum M, ‘A lifetime in breast cancer research’, Eur J Cancer. 2007 Jul;43(10):1496-7
(5) Cuzick J, Baum M., ‘Tamoxifen and contralateral breast cancer’, Lancet 1985;ii:282
(6) Cuzick, J., Forbes, J., Edwards, R., Baum, M., Cawthorn, S., Coates, A., et al., ‘First results from the International Breast Cancer Intervention Study (IBIS-I): a randomised prevention trial’, Lancet, 2002. 360(9336): p. 817-24
(7) Howell, A., Cuzick, J., Baum, M., et al., ‘Results of the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial after completion of 5 years’ adjuvant treatment for breast cancer’, Lancet, 2005. 365(9453): p. 60-2
(8) Black W.C., Nease R.F. Jr., Tosteson A.N. ‘Perceptions of breast cancer risk and screening effectiveness in women younger than 50 years of age’. Journal of the National Cancer Institute 1995;87:720-731
(9) Blanks RG, Moss SM, McGahan CE, Quinn MJ, Babb PJ. ‘Effect of NHS breast screening programme on mortality from breast cancer in England and Wales, 1990-8: Comparison of observed with predicted mortality’. BMJ 2000; 321:665-9
(10) NHS cancer screening programmes. NHS Breast Screening Programme & British Association of Surgical Oncology Breast Group. ‘An audit of screen detected breast cancers for the year of screening April 1999 to March 2000’.16-5-2001
(11) Collins LC, Tamimi RM, Baer HJ, Connolly JL, Colditz GA, Schnitt SJ Cancer. Outcome of patients with ductal carcinoma in situ untreated after diagnostic biopsy. Cancer 2005 May 1;103(9):1778-84
(12) Zackrisson S, Andersson I, Manjer J and Garne JP, ‘Rate of over-diagnosis of breast cancer 15 years after end of Malmö mammographic screening trial: follow up study’. BMJ 2006;332:689-92
(13) Møller H, Davies E, ‘Over-diagnosis in breast cancer screening’. BMJ 2006; 332: 691-692
(14) Miller AB. The costs and benefits of breast cancer screening. Am J Prev Med 1993;9:175-80
(15) Sarfati D. Howden-Chapman P. Woodward A. Salmond C. ‘Does the frame affect the picture? A study into how attitudes to screening for cancer care are affected by the way benefits are expressed’, Journal of Medical Screening 1998; 5(3):137-140
(16) Vervoort MM, Draisma G, Frachebaud J, van de Poll-Franse, de koning HJ (2004). ‘Trends in the usage of adjuvant systemic therapy for breast cancer in the Netherlands and its effect on mortality’ Br J Cancer 91: 241-247
(17) Watmough D.J. Interval Breast Cancers. American J Roentgenology July 1993, 161, 3.
(18) H.Gilbert Welch, Should I be tested for Cancer?, University of California Press, 2004
 Moss SM, Cuckle H, Evans A, Johns L, Waller M, Bobrow L, for the Trial Management Group. ‘Effect of mammographic screening from age 40 years on breast cancer mortality at 10 years’ follow-up: a randomised controlled trial’. Lancet 2006;368:2053-60.