health status of the U.S. population and to monitor trends over time. Nutritional status is evaluated through interviews and direct physician examinations. During interviews a 24-hour diet recall, a food frequency questionnaire, and a medical history are obtained. Physical examinations include anthropometric and biochemical measurements and physical and dental examinations.
A multistage probability cluster sampling plan is used to ensure a representative sample of U.S. households. For example, the second NHANES, (NHANES II; 1976 to 1980) had a sample size of 27,805, with 91 percent of these agreeing to participate in the survey. Oversampling of subgroups at high risk of malnutrition (such as households in areas with high levels of poverty) was conducted to ensure their adequate representation. In addition, a special survey of Hispanics in five southwestern states, Cubans in Dade County, Florida, and Puerto Ricans in the New York City metropolitan area was conducted from 1982 to 1984. To date, the routine collection of nutrition and health information has systematically excluded people in the military, institutionalized individuals, and individuals living on Native American reservations.
In addition to NHANES, NCHS administers the annual National Health Interview Survey (NHIS) to collect information on the health of civilian, noninstitutionalized Americans. NHIS collects information on the occurrence of injuries, acute illnesses, chronic conditions, and disabilities and the utilization of health care serves among people in the United States who are 17 years of age and older. NHIS consists of a two-part questionnaire that requests basic health and demographic information and a supplemental survey of several health related topics. For example, information on cancer-related risk factors such as participation in screening programs, diet, and family history of cancer were collected in 1987 and 1992.
African Americans and Hispanics have been oversampled in NHIS to improve estimates for these populations. However, Americans Indians and other ethnic minority groups with smaller numbers of individuals have historically been surveyed in insufficient numbers to draw conclusions about these populations. For example, 0.0006 percent (n = 135) of respondents to the 1987 Cancer Control Supplement were American or Alaska Natives. This percentage of respondents was poor in comparison to the proportions of indigenous peoples in the United States, which were estimated to be 0.6 percent in 1980 and 0.8 percent in 1990.