Commentary
The Brave New World-What Can We Realistically Expect to Achieve Through Cancer Control Early in the New Millenium?

Anthony B Miller

Abstract

Cancer control requires strategic planning, and thus knowledge about the trends of incidence and mortality associated with cancer as well as future projections, in order that appropriate decisions on priorities can be made. Cancer prevention requires tobacco control and dietary modification. Screening should utilize only effective strategies. The trends in Canada are expected to be mainly favourable by the year 2020, apart from rising trends in non-Hodgkin's lymphoma. Tobacco control policies in Canada are having an impact, while dietary modification is probably having an impact on colorectal cancer incidence. Screening for cancer of the cervix has achieved maximum impact with present levels of compliance, but as yet there is no evidence of an impact of breast screening. Priority for the future will have to be placed on prevention, especially on encouraging young adults to quit smoking and on dietary modification starting at young ages, and care should be taken with cost-effective application of screening.

Key words: cancer; control; prevention; projections; screening

Introduction

Cancer control comprises five components: prevention, early detection and screening, treatment, rehabilitation and palliative care. The World Health Organization (WHO) has developed the concept of national cancer control programs with the goals of preventing future cancers, diagnosing cancers early, providing curative therapy when available, ensuring freedom from suffering and reaching all members of the population.¹

National cancer control programs are intended to be based on strategic planning, involving a realistic assessment of need within each country, the resources available and the priorities for action. To facilitate such planning, a situation analysis should be done-that is, an assessment of the current incidence and mortality of cancer within the country, and of past and future trends. Estimating future trends is difficult and is usually based on a model that is dependent on what has gone on before. In the annual Canadian Cancer Statistics published by the National Cancer Institute of Canada, a useful feature has been the projections, although they extend for only a few years. However, for adequate cancer control planning, projections that last longer than three to five years are required, particularly to help us take stock and decide whether we can expect to see favourable or unfavourable trends in the future, and therefore to determine what our priorities should be as we undertake strategic planning.

In this paper, I first review available cancer control strategies for prevention and screening. Then I present the results of an exercise in which I have extended projections of cancer incidence and mortality in Canada through to the year 2020, in the hope that this will not only engender a productive debate but also facilitate more thought on what has been achieved in cancer control over the last 25 years in Canada, and what may be achieved over the next 20.

The opinions expressed in this paper are often personal, based on my own assessment of the evidence. Thus, when a statement that is not referenced is made on the interpretation of a trend, the reader should assume that this is my own view. If supporting references are available and relevant, however, I cite them, although I have not provided references for many well-recognized facts, which may be found elsewhere.²

Prevention in Cancer Control

Population attributable risks help in establishing priorities. The population attributable risk for tobacco-induced cancers is of the order of 30% in Western populations, but over the world as a whole it is closer to 20%; for dietary-associated cancers it is perhaps 30%; for cancers associated with infection, about 15%; and for occupational and environmental carcinogens, about 3-9%, depending on the prevalence and intensity of exposure in that particular population.² So, if we could apply all the knowledge we have now, we could prevent over half the cancers in the world.

Some years ago I estimated that if, in Canada, the effects of smoking could be eliminated, approximately 29% of cancer deaths might be prevented, and 22% of cancer cases. If dietary modification away from a diet of high fat, high energy and low plant content was successful, the effect would be roughly a 20-24% reduction. Breast cancer screening might prevent 25% of breast cancer deaths. Screening for cancer of the cervix should prevent 60% of cases and deaths from the disease.³

However, these actions will take time. Tobacco control will take at least 30 years. It has taken more than 30 years to build up the tobacco epidemic, and it is probably going to take a great deal longer than 30 years for control strategies to have a major impact.

Dietary modification, depending on the time of life when it takes effect, might take as little as 10 years or as long as 60 for any results to be apparent. A hint comes from the impact on cancer incidence of migration to high-risk areas: for colorectal cancer the change is fairly rapid, but for gastric and breast cancers the changes are much slower.

Hepatitis B vaccination can be expected to prevent liver cancer in high-risk countries with a high prevalence of infection. However, infants have to be vaccinated, and a major impact on liver cancer incidence cannot be expected for about 40 years, although there are other, more immediate non-cancer-related benefits from vaccination. Similarly, even if a vaccine for human papilloma virus (HPV) became available soon, it would take at least 30 years from vaccination before there was a clear effect on invasive cancer of the cervix. In contrast, screening and/or treatment, if effective, can have a rapid impact.

A serious problem in obtaining support for cancer prevention is that the future patient is unknown and under-represented. Cancer will develop in a third of the population, but for the majority we have no idea who will be affected, and so we do not know who the people are who should be pressing for the necessary actions to prevent their future cancers.

Tobacco control has to be the highest priority for cancer prevention. In Western populations, dietary modification involves encouraging people to eat more plant-based foods, whereas in countries where adequate amounts of plant-based foods are eaten, it is important to prevent people from changing over to high consumption of animal-based foods. Known carcinogens must be avoided, and this is particularly critical for many developing countries, as there is a tendency to import new industry and the relevant carcinogens. Hepatitis B vaccination has been available for many years, and yet in countries where children need vaccination probably only about 1% have been vaccinated. Health promotion programs therefore have to be reinforced, and we have to learn how to impart information in such a way that people will act to preserve their own health.

Screening in Cancer Control

When screening is proposed as a part of cancer control, only strategies known to be effective should be used in the general population. Effective programs can be run only on the basis of an educated target group. There have been many programs, particularly in developing countries, where the right educational background has not been provided to the public, for whom the concept of screening and prevention of death from cancer is foreign to their culture; thus, screening fails to reach those at greatest risk.

Programs have to be based on the natural history of the cancer. Part of the problem is that the natural history of the detectable precancerous phase may not be known until screening has been performed. Screening should be performed at the right ages and the right frequency. Screening for cancer of the cervix is not yet performed at the right frequency in Canada, so it is more expensive than it should be. Screening should always be conducted with attention to high quality, with adequate facilities and preferably in an organized setting. Fortunately, the Canadian breast cancer screening programs were set up in an organized way; the cervical cancer screening programs were not.

For breast cancer screening, there is continuing controversy over whether mammography screening is effective in women under the age of 50. However, it is now clear that in those for whom screening begins in their 40s there is no reduction in breast cancer deaths while they are in their 40s.^4-6 Two studies have suggested the effectiveness of breast self-examination in this age group.^7,8 These results, combined with evidence on the lack of benefit from adding mammography to breast physical examination and the teaching of breast self-examination⁹ as well as the early benefit found from screening women over the age of 50 in the first breast cancer screening trial,¹⁰ suggest that the major benefit from screening derives from the earlier detection of relatively advanced, not early, disease.¹¹

Even the best of the breast cancer screening trials comparing screening with no action did not reduce breast cancer deaths among women aged 50-69 by more than 30%.⁴ There is a possibility that in the period since the trials were completed showing benefit from screening versus no screening to the years when modern treatments became available, i.e. adjuvant chemotherapy and tamoxifen, the component that benefited from screening and early detection is now cured by modern treatment. So improved treatment could remove the opportunity to observe a benefit from screening.

Turning to cervical cancer screening, there is now better evidence on the advantages of conservative management of low-grade lesions.¹² In this study, the records of the largest cytology laboratory in Toronto were used, and women who had a cytologic label of dysplasia, whatever the grade, were identified. The records were subsequently linked to the Ontario Cancer Registry. About 11% of the cases of mild dysplasia were found to progress to moderate dysplasia or worse in the following two years, and about 20%, in five years; for progression to severe dysplasia or worse, the figures were only 2% after two years and 6% after five. Among cases of moderate dysplasia, only 16% progressed to severe dysplasia or worse in two years, and 25%, in five. So the proportion of women with mild or moderate dysplasia in whom there was progressive disease is relatively small. The majority of women with mild or moderate dysplasia showed regression to normal within five years.

The Bethesda classification puts the moderate and severe dysplasias in one category: high-grade lesions (HSIL).¹³ This is unfortunate, in that the line should have been drawn between moderate and severe dysplasia, and it results in a large number of women being overtreated at unnecessary cost. When young women are screened many will be found to have HPV effects, of which the large majority will regress untreated. To refer all women categorized as LSIL (mild dysplasia and HPV effects) is to risk substantial overtreatment; but even in the case of HSIL, many are also being overtreated, because the majority of the moderate dysplasias regress.

There is now good evidence that screening for colorectal cancer with the fecal occult blood test will reduce colorectal cancer mortality by about 20%.^14-17 Whether the same can be achieved by flexible sigmoidoscopy is still under investigation. However, other cancer control strategies may be having an impact on colorectal cancer. So to initiate screening now, when incidence and mortality from the disease is already on the downturn, is to risk a great deal of expenditure for a small impact.

Evaluating new screening technologies in the future is not likely to be easy, as the rapid dissemination of information and the pressures induced by commercial interests may appear to preclude the use of the preferred evaluation method: the large-scale, randomized screening trial. This is precisely the situation that is beginning to emerge over spiral computed tomography scanning for lung cancer, in which enthusiastic early reports¹⁸ are fuelling demand from high-risk subjects, especially in the US. The window of opportunity for evaluating this approach may be very small.

This situation is similar to the confusing one that still exists over screening for prostate cancer using prostate-specific antigen (PSA) testing. Interpreting the rapid changes in prostate cancer incidence and mortality that have occurred is not easy, although it seems clear that there are many artifacts that could be responsible for the changes in mortality.^19-21 Only randomized trials of screening, however, will determine the precise causes of the changes and whether screening is effective. Fortunately, two large trials have nearly completed their intake.²²

Past Trends in Cancer Incidence and Mortality

Comparison of trends in both the incidence and mortality of the major cancer sites in Canada and other countries is informative. Incidence data are derived from the successive volumes of Cancer Incidence in Five Continents (published by the International Agency for Research on Cancer), and mortality data are from the WHO cancer mortality database. Canadian national data are available only from 1969, although for some provincial registries data were included from the first volume of Cancer Incidence in Five Continents, which relates to the period 1960-1962. In general, I include data from the United States, and sometimes from the United Kingdom and Japan in these comparisons, but if there are remarkable changes occurring in other countries that appear to facilitate interpretation of the trends (e.g. some of the Mediterranean countries in the case of diet-associated cancers), data from these are added also.

Colorectal Cancer

The incidence of colorectal cancer has been rising in many countries (Figure 1), but in Canada and New Zealand there has been a recent decline among females. Declines have also been noted in the US.²³

With regard to mortality, however, there have been declines in many countries over a considerable period. One exception is Japan (Figure 2), where the increase in mortality reflects the underlying increases in incidence. Similar trends are occurring in Spain and Greece, and to a lesser extent in Italy. France, however, is showing a reduction in colorectal cancer mortality.

Figure 3 shows these trends together for Canada, based on data from Canadian Cancer Statistics. Decreases in both incidence and mortality can be attributed to the effects of primary prevention. The earlier rise in incidence in Canada could be artifactual, since there were improvements in cancer registration during this period, particularly in Quebec but also in some of the Atlantic provinces and probably in Alberta and Ontario. The other problem is that when cases (rather than people) are registered as having a disease in which multiple primary tumours are common, incidence will remain elevated as a result of duplicate or, in some instances, triplicate registrations.

FIGURE 1 Trends in colorectal cancer incidence, New Zealand, Canada, Denmark and Finland

FIGURE 2 Trends in colorectal cancer mortality, New Zealand, United Kingdom, Canada and Japan

FIGURE 3 Trends in colorectal cancer incidence and mortality, Canada

Among the changes in diet that have occurred in the last two decades in Canada was one related to the standards of fat in beef. What was once regarded as of high quality, i.e. marbled beef with a lot of fat in it, came to be perceived as lower-quality beef; high-quality beef became lean beef. Subsequently, it is likely that many people made changes in their diet, initially induced by concern about heart disease, and then by the Canadian Cancer Society's campaign that began in the 1980s. Therefore, perhaps at least some, if not most, of the decline in both colorectal cancer mortality and incidence has been due to dietary modification.

Lung Cancer

In the case of lung cancer we can restrict our attention to trends in mortality, as it so closely reflects incidence. Figure 4 contrasts the trends in lung cancer mortality in Canada, the UK and the US. Both in the US and Canada, there was a rise in lung cancer mortality among men and then a fall; neither reached the peak mortality in the UK, which occurred about 15 years earlier. Currently, however, lung cancer mortality among men in Canada is very similar to that in the UK, and lower than in the US. These trends in age-standardized rates are based on remarkable changes that have been largely cohort-related,²⁴ with the effect that rates among younger males have declined for several years, whereas those among older males have plateaued and declined only slightly.

Among women, mortality from lung cancer has been rising in both the US and Canada; the rates in Canada are a little lower than in the US. In the UK the rates were initially higher, but reached a peak about the middle of the 1980s and then fell, so that they are now below those in the US and Canada. There is a suggestion in both the US and Canada that rates may have peaked recently, but it is too early to be certain.

Figure 5 includes data from selected countries to illustrate some of the contrasting trends that are occurring. Finland shows similar reductions to the UK, but Hungary has now achieved the remarkable distinction of having the highest reported lung cancer mortality in the world. France and Sweden seem to have avoided the major increases among males that occurred in North America and many other countries of Northern and Western Europe. Japan has overtaken Sweden in lung cancer rates among males but not females, and Hungary shows important increases among females.

FIGURE 4 Trends in lung cancer mortality, United Kingdom, USA and Canada

FIGURE 5 Trends in lung cancer mortality, Hungary, Finland, France, Japan and Sweden

Breast Cancer

Most countries have been recording increases in the incidence of breast cancer. Canada is no exception to this trend (Figure 6). Much of this increase is likely to be the effect of early detection programs of various sorts, greater awareness among the public and professionals working in breast cancer and the introduction of mammography screening.

Canada has slightly higher breast cancer mortality rates than the US. For most of the period reviewed, rates have been stable, but recently there has been a fall (Figure 7). The UK displays a rise and then a fall. Some other countries have shown different trends, for example, increases in both Finland and Japan. There has been a tendency in the UK to claim that at least half of the decline has been due to the breast screening program. However, the time relationships of the declines do not suggest that they are related to screening. It is important to note that Sweden has not yet experienced anything like the decreases reported in the UK, Canada and the US, yet there were several major screening trials in Sweden, and Sweden was the first country to decide on organized breast cancer screening.

It takes nearly nine years for the impact of breast cancer screening on mortality in a population to be detectable, because most of the deaths that occur in the initial years after introduction of screening are among women whose breast cancers were diagnosed before the program was introduced. With the lack of a reduction in breast cancer mortality in Sweden, it seems unlikely that the reductions in the UK, the US and Canada are due to screening. Rather, they are probably due to the introduction of adjuvant chemotherapy for premenopausal women and tamoxifen for post-menopausal women. So we are probably seeing an impact of treatment on mortality, with the impact of screening, if any, yet to follow.

FIGURE 6 Trends in breast cancer incidence, cancer registries of USA (Connecticut), Canada (Saskatchewan), Denmark, Finland and Japan (Miyagi)

FIGURE 7 Trends in breast cancer mortality, United Kingdom, Canada, USA, Sweden, Finland and Japan

Cancer of the Cervix

For some time, the incidence and mortality associated with cancer of the cervix has been falling in Canada. Much of the recent decline is attributable to screening, but possibly some of the earlier decline was due to improvements in early detection and treatment, as in Sweden.²⁵ The incidence of cervical cancer in Canada has fallen to a similar extent as in most of the Nordic countries, where major impacts of screening were observed except for Norway, which failed to introduce organized programs (Figure 8). In Canada this resulted largely from annual screening, yet the impact of five-year screening in Finland and screening every three or four years in Sweden has been almost identical to our own. So cervical cancer screening has been successful in Canada, but at a cost- an opportunity cost.

Figure 9 depicts trends in cervical cancer mortality for selected countries. The UK, in spite of showing lower mortality than the US and Canada in the 1950s, had much slower downward trends, and only recently has this begun to accelerate. Much of the earlier fall in the UK was probably not due to screening.²⁶ The mortality trends in Denmark and Finland are as expected from the incidence trends in Figure 8. After a rise Japan shows a slow decline, and Israel's trend is stable, although there was a suggestion of a rise in the 1970s, which may have been aborted by the screening that was introduced.

FIGURE 8 Trends in cervical cancer incidence, Denmark, Canada, Norway, Sweden and Finland

FIGURE 9 Trends in cervical cancer mortality, Denmark, United Kingdom, USA, Canada, Finland, Japan and Israel

Future Canadian Trends in Cancer Incidence and Mortality

Figures 10-13 illustrate my projections of cancer incidence and mortality for Canada through to the year 2020. These have, in a few instances, been derived from a simple linear projection of recent trends, but for most, I have superimposed my understanding of recent changes and what I expect for the future. These are not based on any statistical model, but can be regarded as "educated guesses" on my part, with all the problems of validity that such guesses imply. Table 1 summarizes past trends and my projections.

TABLE 1 Actual (1970-1996) and projected (1996-2020) changes in cancer incidence and mortality in Canada, by sex
Site		1970-1996	1996-2020
Stomach
Males:	Incidence	-43%	-8%
	Mortality	-59%	-9%
Females:	Incidence	-44%	-5%
	Mortality	-60%	-7%
Colorectum
Males:	Incidence	+28%	-19%
	Mortality	-21%	-27%
Females:	Incidence	-6%	-30%
	Mortality	-38%	-41%
Melanoma
Males:	Incidence	+206%	+12%
	Mortality	+108%	+8%
Females:	Incidence	+115%	+4%
	Mortality	+45%	+7%
Lung
Males:	Incidence	+44%	-31%	(-14%)a
	Mortality	+31%	-32%	(-15%)a
Females:	Incidence	+344%	+22%
	Mortality	+270%	+23%
Breast
Females:	Incidence	+27%	+2%
	Mortality	-8%	-18%
Cervix
Females:	Incidence	-53%	-2%
	Mortality	-67%	-4%
Body of uterus
Females:	Incidence	0%	-12%
	Mortality	-32%	-12%
Ovary
Females:	Incidence	-3%	0%
	Mortality	-22%	-11%
Prostate
Males:	Incidence	+123%	-3%
	Mortality	+19%	-5%
Bladder
Males:	Incidence	+2%	-23%
	Mortality	-22%	-14%
Non-Hodgkin's lymphoma
Males:	Incidence	+138%	+44%
	Mortality	+55%	-9%
Females:	Incidence	+113%	+29%
	Mortality	+49%	+12%
a Pessimistic estimate

FIGURE 10 Projections of trends in incidence of the major cancers in Canada, in females

FIGURE 11 Projections of trends in mortality from the major cancers in Canada, in females

FIGURE 12 Projections of trends in incidence of the major cancers in Canada, in males

FIGURE 13 Projections of trends in mortality from the major cancers in Canada, in males

Similar to my projection for females, I predict a continued drop in male colorectal cancer and a continuing rise in non-Hodgkin's lymphoma, but to a higher level than in females. Better understanding the reasons for the growth in non-Hodgkin's lymphoma (possibly environmental factors²⁷) and introducing corrective actions, if possible, are now major cancer control research priorities.

Both incidence and mortality for those cancers not included in my calculations will probably remain relatively stable. There will be some reductions, especially for the tobacco-related cancers. It is quite likely that the incidence of testis cancer will continue to rise, a situation for which we have no certain explanation, and that mesothelioma will continue to increase in both incidence and mortality.

Discussion

There are clearly question marks over many of my projections. Will we continue to see a beneficial effect on tobacco-induced cancers? Will the current reversal of smoking rates in teenagers have a major adverse effect? Is the effect of diet as strong as I indicate? What will the effect of screening really be?

I am not too pessimistic about teenage smoking, in that it may be possible to influence behaviour after the teenage years in sufficient time to avoid major increases in cancer incidence and mortality. The studies on the effects of quitting smoking show that if people smoke only to the age of 30 or 35, there will be hardly any increase in their lung cancer risk. Quitting at the age of 40 will not have as strong an impact. If quitting does not occur until the age of 50, there is a major elevation in lifetime risk, since duration of smoking is so important.²

Thus there are dual challenges for tobacco control: prevention in uptake and acceleration in dealing with addiction, with a big challenge to ensure that young adults recognize the major benefits they will achieve if they stop smoking.

It is quite possible that the effects that nutrition has on some of the risk factors for breast cancer may result in a major future impact on breast cancer incidence. We know that Western-type nutrition has an impact on age of menarche and that age of menarche is related to breast cancer risk; this type of nutrition also has the effect of postponing age at menopause. The longer the ovaries function from menarche to menopause, the greater the breast cancer risk. Nutrition also has an impact on height. Height is clearly related to breast cancer risk, particularly in premenopausal women. Thus, the mechanism by which Western-type nutrition affects breast cancer is probably indirect, in that it has an impact on factors that themselves affect ovarian activity.

For colorectal cancer, the mechanism is probably completely different. Some of the evidence suggests that the impact of fibre or bulky foods is on the metabolism of carcinogens in the bowel. There may also be an indirect effect of fat through bile acids on colon mucosa, with necrosis of the superficial layers increasing the turnover of the cells in the base of the crypts, thus increasing the chance of errors occurring and cancers occurring. There is also some evidence that there is an effect mediated through insulin.

So, although we do not know enough about the mechanisms, we should not assume that they are identical for breast cancer and colorectal cancer. However, the expectation in terms of when changes in nutrition will show a beneficial effect are different: the effect on colorectal cancer incidence has probably already started; the effect on breast cancer incidence could be delayed, unless it is already being seen in the reduced risk among baby boomers.

Many are hoping for major advances from the human genome project. The proportion of cancers of our common sites caused by dominant genes of the BRCA1 and 2 type is likely to be low, however. For breast cancer, the proportion is probably about 5%, as it probably is for colorectal cancer, but for other cancers the proportion will be much less. Thus, in terms of these dominant genetic effects, there will not be a major population impact.

What seems to be increasingly likely is that genetic polymorphisms of various sorts will be identified that will begin to make it possible to get a better handle on an individual's susceptibility to various cancers. Whether that will then result in an improvement in our ability to control cancer generally is unclear. If a particular subgroup that is at increased risk of a specific cancer can be identified, it may be worthwhile to use certain drugs or other agents to prevent cancer in them, or to concentrate screening on them. That may make certain types of cancer control actions more cost-effective, but may not necessarily result in a greater impact in the population.

For screening, the great unknown with current tests is the extent to which the effects will be complementary to either prevention (for colorectal cancer) or treatment (for breast cancer). Only long-term monitoring of trends will reveal this. For prostate cancer, if the effect of early diagnosis is to improve the results of treatment of advanced disease (as is probably true for breast cancer), then it may be far more immediate than one might anticipate from the apparent long lead times gained by PSA screening and could become detectable very soon from downturns in prostate cancer mortality. For cancer of the cervix the challenge is, as it has been for over two decades, to reach those at risk and avoid overscreening those not at risk. With regard to lung cancer, there is increasing interest in helical CT scanning,¹⁸ in part because of concern over the increasing proportion of cases of lung cancer occurring in North America in ex-smokers. However, large randomized trials are essential to decide whether that approach really works, and these will take some time to conduct.

In conclusion, therefore, there has been major success in Canada over the last 25 years in controlling stomach and cervical cancers, and in setting the stage for the eventual control of tobacco-associated cancers and probably some of the diet-associated cancers such as colorectal cancer. Apart from cervical cancer, this success comes from prevention, which deserves continuing high priority in terms of both research and application. We need to be cautious in our approach to screening, however. Success in cervical cancer screening was achieved at a greater cost than was probably necessary; the role of screening is uncertain for colorectal cancer, and as yet it is unknown for prostate cancer or lung cancer.

References

1. World Health Organization. National cancer control programmes. Policies and managerial guidelines. Geneva: WHO, 1995.

2. Tomatis L, Aitio A, Day NE, Heseltine E, Kaldor J, Miller AB, et al, eds. Cancer: causes, occurrence and control. Lyon: International Agency for Research on Cancer, 1990; IARC Scientific Publications No 100.

3. Miller AB. Planning cancer control strategies. Chronic Dis Can 1992;13(1 Suppl): S1-S40.

4. Kerlikowske K, Grady D, Rubin SM, et al. Efficacy of screening mammography. A meta-analysis. JAMA 1995;273:149-54.

5. Alexander FE, Anderson TJ, Brown HK, et al. 14 years of follow-up from the Edinburgh randomised trial of breast cancer screening. Lancet 1999;353:1903-8.

6. Miller AB, To T, Baines CJ, Wall C. The Canadian national breast screening study: update on breast cancer mortality. J Natl Cancer Inst Monographs 1997;22:37-41.

7. Gastrin G, Miller AB, To T, et al. Incidence and mortality from breast cancer in the Mama program for breast screening in Finland, 1973-1986. Cancer 1994;73:2168-74.

8. Harvey BJ, Miller AB, Baines CJ, Corey PN. Effect of breast self-examination techniques on the risk of death from breast cancer. Can Med Assoc J 1997;157:1205-12.

9. Miller AB, Baines CJ, To T, Wall C, et al. Canadian national breast screening study: 2. Breast cancer detection and death rates among women age 50-59 years. Can Med Assoc J 1992;147:1477-88.

10. Shapiro S, Strax P, Venet L. Periodic breast cancer screening in reducing mortality from breast cancer. JAMA 1971;215:1777-85.

11. Miller AB. Screening for cancer: is it time for a paradigm shift? Ann R Coll Physicians Surg Can, 1994;27:353-5.

12. Holowaty P, Miller AB, Rohan T, To T. The natural history of dysplasia of the uterine cervix. J Natl Cancer Inst 1999;91:252-8.

13. National Cancer Institute Workshop. The 1998 Bethesda system for reporting cervical/vaginal cytological diagnoses. JAMA 1989;262:931-4.

14. Hardcastle JD, Chamberlain JO, Robinson MH, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348:1472-7.

15. Kronborg O, Fenger C, Olsen J, et al. Randomized study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996;348:1467-71.

16. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993;328:1365-71.

17. Mandel JS, Church TR, Ederer F, Bond JH. Colorectal cancer mortality: effectiveness of biennial screening for fecal occult blood. J Natl Cancer Inst 1999;91:434-7.

18. Henschke CI, McCauley DI, Yankelevitz DF, et al. Early lung cancer action project: overall design and findings from baseline screening. Lancet 1999;354:99-105.

19. Hankey BF, Feuer EJ, Limin XC, et al. Cancer surveillance series: interpreting trends in prostate cancer-part I: evidence of the effects of screening in recent prostate cancer incidence, mortality, and survival rates. J Natl Cancer Inst 1999;91:1017-24.

20. Feuer EJ, Merrill RM, Hankey BF. Cancer surveillance series: interpreting trends in prostate cancer-part II: cause of death misclassification and the recent rise and fall in prostate cancer mortality. J Natl Cancer Inst 1999;91:1025-32.

21. Etzioni R, Legler JM, Feuer EJ, et al. Cancer surveillance series: interpreting trends in prostate cancer-part III: quantifying the link between population prostate-specific antigen testing and recent declines in prostate cancer mortality. J Natl Cancer Inst 1999;91:1033-9.

22. The International Prostate Screening Trial Evaluation Group. Rationale for randomised trials of prostate cancer screening. Eur J Cancer 1999;35:262-71.

23. Chu KC, Tarone RE, Chow W-H, et al. Temporal patterns in colorectal cancer incidence, survival, and mortality from 1950 through 1990. J Natl Cancer Inst 1994;86:997-1006.

24. Miller AB. Recent trends of lung cancer mortality in Canada. Can Med Assoc J 1977;116:28-30.

25. Ponten J, Adami HO, Bergstrom R, et al. Strategies for global control of cervical cancer. Int J Cancer 1995;60:1-26.

26. Quinn M, Babb P, Jones J, et al. Effect of screening on incidence of and mortality from cancer of cervix in England: evaluation based on routinely collected statistics. Br Med J 1999;318;904-8.

27. Rothman N, Cantor KP, Blair A, et al. A nested case-control study of non-Hodgkin lymphoma and serum organochlorine residues. Lancet 1997;350:240-4.

Anthony B Miller, Head, Division of Clinical Epidemiology, Deutsches Krebsforschungszentrum; and Senior Epidemiologist, International Agency for Research on Cancer; and Professor Emeritus, Department of Public Health Sciences, University of Toronto

Correspondence: Anthony B Miller, Division of Clinical Epidemiology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D - 69120 Heidelberg, Postfach 101949 .  69009 Heidelberg, Germany; E-mail: A.Miller@DKFZ-Heidelberg.de

This article is based on a seminar presented at the Laboratory Centre for Disease Control on April 27, 1999.

[Table of Contents [Next]