This Article Abstract Full Text (PDF) eLetters: Submit a response to this article Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services E-mail this article link to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Andersen, L. D. Articles by Reeves, M. Search for Related Content PubMed PubMed Citation Articles by Andersen, L. D. Articles by Reeves, M.

Assessing a Decade of Progress in Cancer Control

^a Department of Population Health Sciences, University of Wisconsin Medical School, Madison, Wisconsin, USA; ^b Comprehensive Cancer Center; University of Wisconsin Medical School, Madison, Wisconsin, USA; ^c Michigan State University, East Lansing, Michigan, USA

Correspondence: Patrick Remington, M.D., M.P.H., Department of Population Health Sciences, 610 North Walnut Street, Madison, Wisconsin 53705, USA. Telephone: 608-263-6294; Fax: 608-263-2820; e-mail: plreming{at}facstaff.wisc.edu

	ABSTRACT

Top Abstract Introduction Methods Results Discussion Conclusions References

 
The age-adjusted death rate from cancer peaked in the U.S. in 1990, and has declined steadily since then. We assess reasons for this progress by examining trends in cancer mortality by age, gender, and cause, using underlying cause mortality data from the Centers for Disease Control. Mortality rates for 2000 were estimated using models based on 1979 through 1997 mortality data. Indirect standardization was used to calculate the expected number of cancer deaths in 2000, by age, gender, and cause, assuming that the rates in 1990 had not changed. In the U.S. in 2000, there were an estimated 500,000 deaths from cancer; 64,000 (12.7%) fewer than expected, with 51,900 fewer cancer deaths among men and 12,200 fewer deaths among women. The decline in deaths among men resulted from fewer deaths from lung cancer (20,800), colon cancer (6,700), and prostate cancer (12,900). The decline in deaths among women resulted from fewer deaths from breast cancer (11,100) and colon cancer (4,200), but there were more deaths from lung cancer (6,500). Among women over the age of 75, 5,000 more died of cancer than expected. Declines in lung, prostate, and colon cancer deaths among men and breast and colon cancer among women account for 86% of the recent decline in cancer deaths over the past decade.

Key Words. Cancer • Epidemiology • Surveillance • Trends • Cancer control • Public health

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion Conclusions References

 
For most of the 20th century, cancer rates have increased steadily in the U.S. [1]. Recently, the American Cancer Society (ACS), National Cancer Institute, and Centers for Disease Control and Prevention (CDC) have collaborated on the publication of "report cards" to the nation on progress related to cancer prevention and control in the U.S. [1, 2]. These reports show a decline in age-adjusted cancer mortality rates since 1990, after increasing steadily from 1973 to 1990 [1, 3]. The purpose of this paper is to summarize the contributions to trends in cancer mortality by age and gender and the progress in the war on cancer over the past decade.

	METHODS

Top Abstract Introduction Methods Results Discussion Conclusions References

 
Mortality data were obtained from the National Centers for Health Statistics, through the CDC online database (CDC Wonder; available at: http://wonder.cdc.gov). We examined trends in cancer deaths by year from 1979 to 1997, by age and gender, for 23 different anatomical sites. Analyses were not conducted by race, since the data were only available for Caucasian, African American, and other.

Using methods similar to those used by the ACS to estimate mortality rates based on time trends, we estimated the number of deaths that occurred in the year 2000 for the various types of cancer in each age and gender group [1]. In order to assess progress in reducing deaths from cancer over the past decade, we compared these estimates with the number of deaths that would have occurred in the year 2000 if cancer mortality rates had not changed during the past decade.

We modeled age and gender-specific mortality rates for each cancer site over time, using MINITAB statistical software to perform regressions. Each model was evaluated by examining residuals, R-squared values, and t tests for the statistical significance of regression coefficients. The best models were obtained using polynomial regressions with up to three terms on log transformed mortality rates. Most leading regression coefficients were statistically significant at an alpha level of 0.05. These models allowed us to estimate mortality rates for 1990 with reduced variability, and to estimate mortality rates for the year 2000.

To test the reliability of this method for estimating mortality rates 3 years beyond the available mortality data, we used the same method to estimate mortality rates for 1997, using only data prior to 1994. These estimated rates corresponded with actual 1997 data to within 2% for each cancer type, and less than 0.1% overall.

For each cancer type, the modeled death rate in 1990 was multiplied by the population in 2000 (projected by the U.S. Census Bureau [4]) for each age/gender group, to determine the "expected" number of deaths if the cancer rate had not changed over the past decade. This expected number was then compared with the number of deaths estimated to have occurred in 2000 (the product of the modeled death rate in 2000 times the projected population in 2000), to determine the lives "lost" or "saved." The Standard Mortality Ratio (SMR), a measure of percent change in mortality rates, was calculated for each cancer site by dividing the estimated number of deaths in 2000 by the expected number of deaths in 2000.

	RESULTS

Top Abstract Introduction Methods Results Discussion Conclusions References

 
We estimated that 64,100 fewer lives were lost to cancer in 2000 in the U.S. than would have been expected based on 1990 rates (Table 1). This decline in cancer mortality was greater among men than among women: 51,900 fewer men died than expected, while 12,200 fewer women died than expected. Eighty-one percent of the decline in overall cancer mortality was attributable to the decline in male mortality (Fig. 1).

View larger version (17K): [in this window] [in a new window]   Figure 1. The number of lives saved (lost) from cancer by age group and gender in the U.S. in 2000 (compared with the rate in 1990, using indirect age adjustment).

The absolute numbers of lives saved or lost for each cancer site since 1990 are listed in Table 1 and shown graphically in Figure 2. The largest absolute decreases in mortality were found in prostate and lung cancer in men, breast cancer in women, and colon cancer in both sexes. These four cancer types account for over 55,000 lives saved, or 86% of the total.

View larger version (18K): [in this window] [in a new window]   Figure 2. The number of cancer lives saved (lost) by cause in the U.S. in 2000 (compared with the rate in 1990, using indirect age adjustment).

Breast, prostate, and colorectal cancers also remain leading causes of cancer mortality; these three types of cancer accounted for about 125,000 deaths in 2000. Since its peak in 1990, breast cancer mortality has declined by about 19%. This decrease occurred among all age groups under 85 years. Prostate cancer mortality rates peaked in 1991 and have since decreased by about 33%, with all age groups contributing to the decline. Colorectal cancer mortality rates peaked in 1980 and have since fallen by a total of about 21%, with about half of this decline occurring in the past decade. Colorectal cancer mortality rates have declined in all age groups. Because of the declines in these three cancer types, about 35,000 fewer deaths than expected occurred in 2000.

Liver cancer and non-Hodgkin's lymphoma caused the largest mortality increases during the 1990s, aside from lung cancer in women. About 14,000 Americans died of liver cancer in 2000, and about 25,000 died of non-Hodgkin's lymphoma—a total of about 4,200 more deaths than expected based on 1990 rates. These increases occurred in all age groups in liver cancer and in people over 45 in non-Hodgkin's lymphoma.

Relative changes in site-specific cancer rates were assessed using SMRs, calculated by dividing the number of deaths based on the model by the expected number of deaths based on 1990 rates (Table 1). The SMR of liver cancer (139) was significantly higher than that of any other cancer, indicating a larger relative increase in liver cancer mortality compared with other cancer types. Lung cancer in females, non-Hodgkin's lymphoma, sarcoma, and melanoma also had relatively large SMRs (ranging from 112 to 116). Prostate, testicular, and stomach cancer had the smallest SMRs (ranging from 68 to 72), indicating large relative decreases in number of deaths.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion Conclusions References

 
Cancer mortality rates peaked in the U.S. in 1990, and have declined steadily since then. By 2000, cancer mortality had declined by about 11%, leading to an estimated 64,000 fewer deaths annually than expected, compared with 1990. Most of this decline was due to relatively small mortality rate reductions in the four leading causes of cancer mortality—lung, colon, prostate, and breast cancers.

The reduction in the number of lung cancer deaths among men was the major contributor to the overall decline in the number of cancer deaths in the U.S., accounting for almost one-third of the lives saved. Trends in lung cancer mortality mirror trends in smoking in the population decades later. Thus, the progress in lung cancer mortality is due to the slow, but steady, decline in smoking rates among men of all ages during the 1970s and 1980s [5, 6]. In contrast, the lack of progress in reducing smoking among women has led to continued increases in lung cancer deaths among women—much greater than any other cancer [7]. Lung cancer now far exceeds breast cancer as the leading cause of cancer death. If the trends of the past decade continue, more women will die of lung cancer than men by the end of the next decade.

Prostate cancer mortality rates have declined since 1990, returning to a level similar to the 1980 rate, accounting for about one-sixth of lives saved. It is difficult to determine the reasons for the decline in prostate cancer deaths. Trends in incidence and age-specific mortality rates are consistent with reduction of mortality due to earlier detection by prostate-specific antigen (PSA) screening [8]. Alternatively, the recent fall in incidence and mortality may be due to a reduction in risk behaviors, such as overconsumption of saturated fats [8]. It has also been suggested that newer treatments, such as androgen blockade, while usually not curative, may postpone disease progression long enough for a substantial proportion of patients to die of unrelated causes, thus reducing mortality due to prostate cancer [8]. Others have argued that mortality rates artificially rose in the 1980s and fell in the 1990s in response to the concurrent rise and fall in incidence, since the incidence may affect the number of deaths mistakenly attributed to prostate cancer [9, 10].

The decline in breast cancer mortality also accounted for about one-sixth of the reduction in cancer mortality. Improvements in breast cancer treatment in the 1980s (such as tamoxifen and adjuvant chemotherapy for node-positive disease) are thought to be partly responsible for the decline in breast cancer mortality in the early 1990s, while the widespread use of mammography beginning in the 1980s has contributed to the continued decline in mortality later in the decade [11]. However, some authors caution that the decline may be largely attributable to birth cohort effects, since women who are now 50 to 80 years of age were younger, on average, at first childbirth than previous cohorts [12].

Colorectal cancer, the third leading cause of cancer mortality in the U.S. by gender, accounted for another one-sixth of lives saved. National colorectal cancer mortality rates began slowly declining prior to widespread screening, perhaps due in part to improvements in diet [1]. While the role of dietary fiber is controversial, some have estimated that colorectal cancer incidence rates could be reduced by one-half though dietary changes, such as lowering fat intake and increasing vegetable intake [13, 14]. An increase in the intake of aspirin and other nonsteroidal anti-inflammatory drugs may also have contributed to the decline [1]. Finally, screening may have contributed to the steeper declines in colorectal cancer mortality observed during the past decade [1]. The 5-year survival rate for colorectal cancer patients has improved from 50% to nearly 90% when the cancer is diagnosed at the localized stage [15]. Despite the availability of effective screening techniques, such as the fecal occult blood test and sigmoidoscopy, adherence to screening is still below 50% [16].

Aside from lung cancer in women, liver cancer and non-Hodgkin's lymphoma showed the largest increases in mortality during the past decade. Part of the rise in non-Hodgkin's lymphoma mortality can be attributed to the HIV epidemic and increased survival of AIDS patients, since patients with AIDS have a 1% annual risk of developing non-Hodgkin's lymphoma. However, this cannot entirely account for the increase in mortality. Rises in other viruses, such as Epstein-Barr and hepatitis C, may be involved, as well as increased exposure to toxic chemicals, especially those used in farming [17]. The increase in liver cancer deaths may similarly be related to hepatitis infections and increasing exposure to some chemicals and toxins [18].

Several factors limit the accuracy of the results of our model. Cause of death misclassification diminishes the quality of the underlying mortality data [19]. The validity of trend extrapolation depends upon the stability of factors affecting mortality, such as environmental factors. However, our model estimates for cancer deaths in 2000 agree with those of the ACS to within 1% for all cancers and to within 10% for each subtype [7]. Because we adjusted our model to provide the best fit for each cancer type, our estimates may be more reliable than the simple linear models used by the ACS.

Finally, our analysis did not examine trends in cancer mortality by race. However, the CDC recently examined trends in cancer mortality from 1990 to 1998 for the major cancer sites (lung, colorectal, prostate, and breast) [20]. That study found that cancer rates are generally declining, but that progress is greater among Caucasians than among African Americans, American Indians, Asians, and Hispanics. In particular, cancer rates continue to be higher among African Americans and are increasing among American Indians.

	CONCLUSIONS

Top Abstract Introduction Methods Results Discussion Conclusions References

 
The modest decline in the age-adjusted cancer mortality rate is a sign of progress in the "War on Cancer" [21, 22]. Declines in the number of deaths from lung cancer among men, prostate, breast, and colon cancer account for almost all of the reduction in the public health burden from cancer. These results underscore the importance of preventive measures, such as smoking cessation, screening for early tumor detection, and appropriate treatment, to reduce mortality from common cancers. Even small percent reductions in mortality rates of common cancers have a large impact on the number of cancer deaths.

	ACKNOWLEDGMENT

Top Abstract Introduction Methods Results Discussion Conclusions References

 
This work was supported by a grant from the National Cancer Institute and the National Institute of Environmental Health Sciences, U01 CA82004.

	REFERENCES

Top Abstract Introduction Methods Results Discussion Conclusions References

 

1. Ries LA, Wingo PA, Miller DS et al. The annual report to the nation on the status of cancer, 1973-1997, with a special section on colorectal cancer. Cancer 2000;88:2398–2424.[CrossRef][Medline]
2. Wingo PA, Ries LA, Rosenberg HM et al. Cancer incidence and mortality, 1973-1995: a report card for the U.S. Cancer 1998;82:1197–1207.[CrossRef][Medline]
3. Chu KC, Baker SG, Tarone RE. A method for identifying abrupt changes in U.S. cancer mortality trends. Cancer 1999;86:157–169.[CrossRef][Medline]
4. U.S. Census Bureau Projections of the Total Resident Population by 5-Year Age Groups, and Sex with Special Age Categories: Middle Series, 1999 to 2000. US Census Bureau Population Projections Program. Available at: http://www.census.gov/population/www/projections/popproj.html. Accessed September 15, 2000.
5. Weiss W. Cigarette smoking and lung cancer trends. A light at the end of the tunnel? Chest 1997;111:1414–1416.[Abstract/Free Full Text]
6. Wingo PA, Ries LA, Giovino GA et al. Annual report to the nation on the status of cancer, 1973-1996, with a special section on lung cancer and tobacco smoking. J Natl Cancer Inst 1999;91:675–690.[Abstract/Free Full Text]
7. Greenlee RT, Murray T, Bolden S et al. Cancer statistics, 2000. CA Cancer J Clin 2000;50:7–33.[Abstract]
8. Hankey BF, Feuer EJ, Clegg LX 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;91:1017–1024.[Abstract/Free Full Text]
9. Dennis LK, Resnick MI. Analysis of recent trends in prostate cancer incidence and mortality. Prostate 2000;42:247–252.[CrossRef][Medline]
10. 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;91:1025–1032.[Abstract/Free Full Text]
11. Reynolds T. Declining breast cancer mortality: what's behind it? J Natl Cancer Inst 1999;91:750–753.[Free Full Text]
12. Feigelson HS, Henderson BE, Pike MC. Re: Recent trends in U.S. breast cancer incidence, survival, and mortality rates. J Natl Cancer Inst 1997;89:1810–1812.[Free Full Text]
13. Rumm PD, Foote M, Quenan L. Colorectal cancer in Wisconsin: trends in mortality, prevention, and early detection strategies. WMJ 2000;99:30–33.
14. Brownson R, Remington P, Davis J, eds. Chronic Disease Epidemiology and Control, Second Edition. Washington DC: American Public Health Association, 1998:191-214.
15. Winawer SJ, Flehinger BJ, Schottenfeld D et al. Screening for colorectal cancer with fecal occult blood testing and sigmoidoscopy. J Natl Cancer Inst 1993;85:1311–1318.[Abstract/Free Full Text]
16. Kuwada SK. Colorectal cancer 2000. Education and screening are essential if outcomes are to improve. Postgrad Med 2000;107:96–107.[CrossRef]
17. Bosly A, Coiffier B. [Recent data on the epidemiology of non-Hodgkin lymphoma. Groupe d'Etudes des Lymphomes de l'Adulte (GELA)]. Pathol Biol (Paris) 1997;45:449–452.[Medline]
18. McKean-Cowdin R, Feigelson HS, Ross RK et al. Declining cancer rates in the 1990s. J Clin Oncol 2000;18:2258–2268.[Abstract/Free Full Text]
19. Percy C, Stanek 3rd E, Gloeckler L. Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am J Public Health 1981;71:242–250.[Abstract/Free Full Text]
20. Centers for Disease Control. Recent trends in mortality rates for four major cancers, by sex and race/ethnicity—United States, 1990-1998. MMWR. Morb Mortal Wkly Rep 2002;51:49–53.[Medline]
21. Bailar 3rd JC, Smith EM. Progress against cancer? N Engl J Med 1986;314:1226–1232.[Abstract]
22. Bailar 3rd JC, Gornik HL. Cancer undefeated. N Engl J Med 1997;336:1569–1574.[Abstract/Free Full Text]

This article has been cited by other articles:

				P. L. Remington and A. Trentham-Dietz Measuring Progress in Cancer Control: A Bird's Eye View Oncologist, December 1, 2003; 8(6): 539 - 540. [Full Text] [PDF]

This Article Abstract Full Text (PDF) eLetters: Submit a response to this article Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services E-mail this article link to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Andersen, L. D. Articles by Reeves, M. Search for Related Content PubMed PubMed Citation Articles by Andersen, L. D. Articles by Reeves, M.