becomes the basis for discussions about progress. For those who are intimidated by statistics, presentation of the same trends in graphical form may be important.
Another alternative is to present the data in terms of "potential reduction of deaths." Various ethnic groups experience different rates of mortality from cancer at different sites. The reasons for these differences are not always clear, but one could construct a hypothetical U.S. population based on the lowest rates of each of the leading causes of cancer in any ethnic group. Using this as a reference population, one could compare the number of deaths that now occur in each ethnic group with the number of deaths that would have occurred in the hypothetical population with the best rates at major sites drawn from any ethnic group. The difference between the number of deaths that occurred and the number that would have occurred in the hypothetical population would be the potential reduction. This could also be expressed as the percent potential reduction, and it would be the goal of the cancer program to approach the "all American best rate" for all ethnic groups.
The committee asked NCHS to do such calculations, and their findings are significant (Figure 7-1). The total observed annual numbers of deaths from cancer among men for the period covered (1990 to 1995) was 276,146, but if the rate of cancer among all Americans was the lowest rate of cancer for any ethnic group, the expected number of deaths among men would have been 137,476. This is a potential reduction of 138,670 deaths, or 50 percent. The potential reduction for women was comparable. The white population, because of its majority size, would have experienced the greatest reduction in the number of deaths. The African-American population would have experienced the greatest percent reduction because of its greater burden of cancer. A more rigorous study of cancer among the various ethnic groups of the United States would be of great help in achieving the potential reduction in the numbers of deaths from cancer. Each ethnic group could set its own target on the basis of its own potential reduction and participate fully in the effort to reach the target. In the same manner, targets can be set for the medically underserved population when data are collected for this group.
The entire report (National Center for Health Statistics, 1998) of NCHS is included in Appendix F, but the caveats deserve special consideration, and are fully stated here.
Quality of data varies by "race" and ethnicity in both death certificate and Census data. The quality of data for the white and the African-American populations are good in contrast to that for other groups. American Indian data are estimated to be underreported by about 10 percent. With respect to population data, perhaps the most serious problem is the