Measurement of cancer incidence in the United States: sources and uses of data.

Since 1937, investigations have compiled a vast amount of information on cancer incidence in the United States. Three national cancer surveys were conducted in 1937, 1947, and 1969-71; each used somewhat different methods but covered populations who overlapped greatly. Concurrently, population-based tumor registries were developed that were capable of measuring cancer incidence. Since 1973, the Surveillance, Epidemiology, and End Results Program, an amalgam of such registries, had generated cancer incidence data for the United States. Available data provide a rough measure of the size of the cancer problem and may help those who direct prevention programs and those who plan studies of cancer etiology. Because the completeness and validity of existing incidence data are not known with certainty, comparisons among these incidence data are not known with certainty, comparisons among these incidence rates require cautious interpretation.