Background and Data Sources

The impact of cancer in a population is measured and described by looking at a combination of three elements: (1) the number of new cases per year per 100,000 persons (incidence rate), (2) the number of deaths per 100,000 persons per year (mortality rate), and (3) a determination of the proportion of patients alive at some point after their diagnosis of cancer (survival rate). Cancer occurs throughout the world, but the risk of cancer varies from region to region, suggesting geographic, environmental, and cultural implications. Cancer incidence is monitored by population-based tumor registries around the world. Not all countries maintain population-based tumor registries and in many countries, including the United States, these registries monitor considerably less than the entire population. Incidence data from existing population-based registries from around the world are compiled by the International Agency for Research on Cancer (IARC), a part of the World Health Organization. The international incidence rates for the cancer sites presented in this section are taken from their publication, Cancer Incidence in Five Continents (CI5), Volume VI (Parkin, 1992). To make meaningful comparisons among different countries, the rates are age-adjusted to the world standard population. The effect of age-adjustment is to eliminate differences in rates when the population of one country has a different age distribution from that of another country. Many countries have vital statistics departments which keep track of cancer deaths for the entire country, providing a broad base for comparing cancer mortality rates across countries. For many specific cancers, and for all cancers combined, there are wide variations in death rates among different countries. The international mortality data are based on statistical analyses by the American Cancer Society of cancer mortality data provided by the World Health Organization (Boring, 1992). These rates are age-adjusted to the world standard (Parkin, 1992). Death rates have been used to illustrate differences in cancer risk from one country to another. However, the diversity in mortality rates may be caused by a combination of contributing factors: different incidence rates, dissimilar distributions of prognostic factors such as stage, and differences in survival rates. For some sites such as lung cancer, where survival is poor, mortality is probably a good surrogate for incidence, but for cancer sites with good prognosis, mortality is not a good surrogate for incidence. When available, incidence rates, in addition to mortality rates, should be used to evaluate risk differences. Today, in the United States, there are many statewide cancer registries and some regional registries based on groups of counties, many of which surround large metropolitan areas. Some of these population-based registries keep track of cancer incidence in their geographic areas only; others also collect follow-up information in order to calculate survival. In 1973, the National Cancer Institute (NCI) began the Surveillance, Epidemiology, and End Results (SEER) Program in order to estimate cancer incidence and patient survival in the United States. SEER collects cancer incidence data in nine geographic areas with a combined population of approximately 9.6 percent of the entire U.S. population. In this chapter, data from SEER are used to show cancer incidence in the United States by primary cancer site, race, sex, age, and year of diagnosis. The incidence rates presented are age-adjusted to the U.S. 1970 standard population. An exception is when SEER rates are compared to international incidence rates, in which case they are adjusted to the world standard. Incidence trends are presented for 1973 to 1991. Two measures to evaluate trends are given: (1) the estimated annual percent change (EAPC) and (2) the percent change. The EAPC is calculated by linear regression fit through the logarithms of the annual rates. The percent change is calculated between the average of the 1973 and 1974 rates and the average of the 1990 and 1991 rates. Survival data from SEER are presented for patients diagnosed in 1974­76, 1977­79, 1980­82, and 1983­90. Survival data from the End Results Program of NCI are provided for the periods 1960­63 and 1970­73. Because the earlier data are not population-based (except in Connecticut), these data and SEER are not strictly comparable, but each represents the best data available for that time period. Relative survival rates are expressed as percents, representing the proportion of patients who had not died from causes associated with their cancer 5 years after diagnosis. U.S. mortality rates over time are presented by age from 1973 to 1991 and by cancer site and sex from 1950 to 1991. The primary data source is the National Center for Health Statistics (NCHS) public use files as analyzed by the National Cancer Institute. Early data analyses are from NCI Monograph No. 59 (McKay, 1982); those for more recent years are analyses by SEER of the NCHS mortality files. The U.S. mortality rates are age-adjusted to the U.S. standard population in 1970. When U.S. rates are compared to international mortality rates, they are adjusted to the world standard.

continue