NIH News Release
National Cancer Institute

Wednesday, December 8, 1999

Contact: NCI Press Office
(301) 496-6641

Questions and Answers for "Atlas of Cancer Mortality in the United States, 1950-94"

Description of the Atlas

1. What is the "Atlas of Cancer Mortality in the United States, 1950-94"?

The atlas is a book of maps, text, tables, and figures showing the geographic patterns of cancer death rates throughout the United States from 1950-94 for more than 40 cancers. The atlas was prepared and published by the National Cancer Institute (NCI).

Featured in the atlas are 254 color-coded maps that show the geographic variations in cancer mortality rates during 1970-94 compared to those during 1950-69. Different colors represent higher or lower rates. The color maps make it easy to pinpoint geographic areas with average, below average, or elevated rates.

2. What is the purpose of the atlas?

The geographic patterns of cancer may provide important clues to the causes of cancer. The atlas does not provide information about why death rates may be higher in certain localities than in others, but it can generate leads for in-depth epidemiologic studies that may shed light on factors contributing to cancer risks. Possible risk factors include tobacco use, occupational exposures, dietary habits, ethnic background, and environmental exposures from the air or water. In addition, geographic differences in mortality rates may reflect differences in access to medical care, such as screening, diagnosis, or treatment facilities.

Earlier cancer atlases published by NCI made it possible for researchers to identify factors that contributed to the high rates of certain cancers. For example, elevated death rates for lung cancer in coastal areas of Georgia, Virginia, northeastern Florida, and Louisiana were found to be linked to the exposures of shipyard workers to asbestos during World War II. Similarly, follow-up studies of elevated death rates of bladder cancer showed that they were linked to a variety of occupations, including truck drivers and other workers exposed to motor exhausts. In addition, the use of smokeless tobacco was found to cause the elevated rates of oral cancer observed since the 1950s among women in the rural Southeast.

Since it is estimated that the vast majority of cancers are linked to lifestyle and other environmental factors, it is hoped that many of the leads provided by the new atlas will guide further epidemiologic and public health activities aimed at preventing cancer.

3. What data are included in the atlas and on the atlas Web site?

Death rates are presented by race and gender for two time periods, 1950-69 and 1970-94. Rates were calculated for all cancers combined and for each cancer separately. Maps are presented at the level of county (3,055 counties) and state economic areas (508 SEAs). (SEAs are made up of individual or groups of counties within a state with similar economic and cultural characteristics.) Because some counties have very small populations, SEAs are used to ensure more accurate rates for the less common cancers. The rates are age-adjusted to the 1970 U.S. population. Each map also includes the national death rate. As a result of improvements in the disease classification scheme, maps for liver and biliary tract cancers are presented for the first time.

Mortality data are not available for all races because there are few annual population estimates on a county level for nonwhites. Mortality data for whites are available for both periods, 1950-69 and 1970-94, whereas data for blacks are available only for 1970-94. There are no county-specific mortality data specifically for Asians, Hispanics, or Native Americans for either time period. (Recent national mortality rates for other racial groups are available in NCI's Surveillance, Epidemiology, and End Results [SEER] Program monograph, Racial/Ethnic Patterns of Cancer in the United States 1988-1992; to order, call 1-800-4-CANCER or go to the SEER Web site,; click on 'publications.')

The entire contents of the atlas are available on NCI's Web site (see question #19). NCI researchers also included on the Web site the death rates for each of the 50 states and the District of Columbia as well as the downloadable tabulated data used to generate the atlas maps. Neither the state maps nor the tabulated data are part of the printed version of the atlas. This additional information would have added more than a thousand pages to the printed hard copy of the atlas.

Contents of the Atlas

4. How do the overall rates in men compare to the rates in women?

During 1970-94, most forms of cancer were more common among men than women, with the exception of breast, thyroid, and gallbladder cancers. The mortality rates for all cancers combined were 54 percent higher among white men than white women and 84 percent higher among black men than black women.

5. What are the cancers that caused the most deaths?

During 1970-94, almost 60 percent of all cancer deaths among men were due to four primary sites: lung, prostate, colon, and pancreas. Among women, nearly 60 percent of all cancer deaths were due to cancers of the breast, lung, colon, ovary, and pancreas.

6. Has there been a change in overall geographic patterns?

The patterns observed in 1950-69 for several cancers have generally persisted in 1970-94, such as the broad stretches of high rates for cancers of the breast, colon, and rectum in the Northeast. For cancers of the corpus uteri, prostate, and bladder, the geographic clustering of areas with elevated rates has become more pronounced. The patterns for some tumors, such as lung cancer, have changed dramatically.

7. What are the geographic patterns for lung cancer?

The greatest changes in geographic patterns are seen with lung cancer, with recent elevated mortality rates among white men across the South, among white women in the far West, and among blacks in southern urban areas.

Among whites, patterns changed substantially over time. Among white men in the 1950s and 1960s, high rates were observed in urban areas of the Northeast and North Central states and in areas along the Southeast and Gulf coasts. By the 1980s to the mid-1990s, clustering of elevated rates was prominent across the Southeast and South Central areas, with relatively low rates throughout much of the Northeast. For white women, little geographic variation was evident in the 1950s, but by the 1980s and 1990s high rates began to appear in clusters along the Atlantic and Pacific coasts. For both sexes, consistently low rates were seen in the mountain and plains states. Among blacks, rates were consistently low across the South.

The changing patterns for lung cancer generally follow the regional and time trends in cigarette smoking.

8. What are the geographic patterns for colon cancer deaths?

Colon cancer mortality has been elevated in the Northeast for at least four decades. Mortality rates for colon cancer are highest in the Northeast, including parts of New England, the mid-Atlantic, and Midwestern states. Rates across the South and West have remained relatively low. Dietary and nutritional factors are likely to be involved, but further studies are needed to identify the specific nutrients.

9. What are the geographic patterns for breast cancer deaths?

High rates of breast cancer for both 1950-69 and 1970-94 are seen in the Northeast and North Central regions of the country and in scattered areas of the far West. Rates are low across the South and in the Rocky Mountain areas. Differences between the North and South are more pronounced among white women than among black women.

The high rates in urban centers in the Northeast have been seen for four decades. Regional variations in breast cancer mortality have been partially attributed to established risk factors, including late age at first birth, early menarche and late menopause, and to certain other factors, including education and mammography history.

10. What are the geographic patterns for prostate cancer deaths?

High rates for prostate cancer occur in the Northwest, Rocky Mountain, and North Central areas of the United States, with low rates in the South Central areas. Overall, black men have especially high mortality from prostate cancer, with pockets of elevated rates in the Southeastern part of the country. These patterns may be related in part to screening and treatment practices. However, there is some evidence that agricultural exposures may contribute to the geographic variation.

11. What current and future studies are suggested by the mortality data?

Studies in progress:

Possible areas to study in the future:

Interpretation of the Atlas

12. What are some principles that are important to keep in mind when interpreting the atlas maps?

13. What are the strengths and limitations of the data?

By publishing the atlas, NCI's epidemiology program has provided a unique resource to help researchers and state health departments across the country identify patterns of mortality at the county and SEA level, where the population is small enough to be relatively homogeneous, yet large enough to provide reliable data and stable rates. By using the county and SEA rates, it is possible to uncover patterns of cancer that previously escaped notice when larger areas, such as regions or states, were evaluated.

For cancers with poor survival rates and clear-cut diagnoses, mortality data closely reflect incidence data. This is particularly helpful for geographic areas where no incidence data are available. (NCI's SEER program began collecting cancer incidence data in 1973 from certain geographic areas that today comprise about 14 percent of the U.S. population.)

Also, information from death certificates regarding the underlying cause of death is likely to be fairly accurate, because cancers are generally more accurately reported than other causes of death.

One limitation, however, is that often it is difficult to evaluate whether high mortality rates for less fatal cancers point to risk factors or indicate poor quality of care. For example, geographic variations in cervical cancer, which frequently is associated with long-term survival, may reflect variations in screening practices or access to quality medical care, as well as risk factors.

In addition, certain cancers, such as liver, brain, lung, and bone, which are often the site of secondary metastases, may be incorrectly specified as the primary cause of death.

Another limitation of the mortality data is that it is not possible to evaluate the effect that moving from one part of the country to another has on the death rates because residential histories are not included on death certificates.

14. Does the atlas provide information about cancer clusters?

No. A "cluster" is generally defined as the occurrence of a greater than expected number of cancer cases or deaths over a short period of time in a small area, such as a neighborhood, a workplace, or medical practice. In such instances, the atlas may be helpful in providing background information about how the overall cancer mortality rates in that particular county or SEA compare to the surrounding counties or SEAs, or to the state or national rates, but it cannot provide information at the level of town or neighborhood. These case clusters are different from the clustering of high-rate counties and SEAs that are seen on the cancer maps.

State/local health departments are responsible for conducting case cluster studies in a town or neighborhood and are the agencies to which a suspected cluster should be reported. If there is a need for further evaluation after an initial report, the health department attempts to verify the reported diagnoses by contacting patients and relatives and obtaining medical records. The researchers then compare the number of cases in the suspected cancer cluster with information in cancer registries and census data, and review the scientific literature to see if previous studies have reported a link between the reported cancer and any exposure. Most state health departments report that fewer than five percent of cancer cluster investigations go beyond this stage to require a comprehensive field investigation.

Background and Resources

15. Who prepared the atlas?

The atlas was prepared by NCI scientists, Susan S. Devesa, Ph.D., Dan J. Grauman, M.A., Robert N. Hoover, M.D., and Joseph F. Fraumeni, Jr., M.D. from the Division of Cancer Epidemiology and Genetics in Bethesda, Md. Two former NCI scientists--William J. Blot, Ph.D., from the International Epidemiology Institute, Ltd., Rockville, Md., and Gene Pennello, Ph.D., from the Food and Drug Administration, Rockville, Md. — were also authors.

16. How long did it take to prepare the atlas, and what costs were involved in its preparation?

The authors began work on the atlas in the early 1990s. The costs include $262,000 for publishing 20,000 copies and $136,000 for creating the Web sites, as well as the salaries of the authors and staff.

17. What resources were used to prepare the atlas?

NCI scientists used death certificate data and annual population estimates to calculate the cancer death rates. Data from death certificates were provided by the National Center for Health Statistics (NCHS). The state health departments collect death certificates which include information on the cause of death and residence of the deceased at the time of death. These data are then forwarded to the NCHS. The annual population estimates by age, sex, and race were based on data from the Bureau of the Census.

With this information, NCI scientists developed the statistical methods and computerized mapping technology necessary to determine the geographic patterns of cancer deaths.

18. Have NCI or any other groups prepared similar atlases?

19. How can I get a copy of the new atlas?

Single copies can be ordered without charge by calling NCI's Cancer Information Service at 1-800-4-CANCER (1-800-422-6237). It may also be ordered on NCI's on-line Publications Locator Service The number for callers with TTY equipment is 1-800-332-8615.

NCI's Web site contains the printed version of the atlas; files can be downloaded as needed. In addition, the tabulated data used to generate the atlas maps as well as national mortality rates are available on-line.

Another feature of the atlas Web site is that the user can create customized maps. For example, the user can compare rates in different time periods, look at rates for any cancer in any county, zoom and pan different areas of the country, and make color selections.

For more information about cancer visit NCI's Web site for patients, public, and the mass media at

Related Press Releases:
The National Cancer Institute Publishes New Atlas of Cancer Mortality

Research Contributions from Earlier Atlases