Validation of a dietary history questionnaire for American Indian and Alaska Native people

We assessed reliability and relative validity of a self-administered computer-assisted dietary history questionnaire (DHQ) for use in a prospective study of diet, lifestyle, and chronic disease in American Indians in the Dakotas and Southwestern US and Alaska Native people.

Reliability was assessed by one-month test-retest of the dietary history questionnaire. Validity was assessed by comparison of the weighted average of up to 12 monthly 24-hour recalls collected prospectively and a dietary history questionnaire completed in the 13th month.

Participants were recruited at the baseline visit of the Education and Research Toward Health Study in Alaska, the Northern Plains and the Dakotas.

Reliability (Pearson correlation) of the DHQ ranged from r = 0.43 for vitamin A density to r = 0.90 for energy intake. The association of nutrient and food estimates assessed by 24-hour recalls and the DHQ completed at the end of the year reflected no bias towards recent intake. Macronutrients expressed as density (nutrients per 1000 calories) did appear to be valid (r = 0.50-0.71) as did several micronutrients (range r = .22 to 0.59), fiber (r = 0.51), and servings of red meat (r = 0.67). However, the DHQ overestimated intake and gross amounts of nutrients were not strongly associated with the weighted average of the 24-hour recalls.

The DHQ developed for estimation of dietary intake in American Indians and Native people in Alaska is reliable. Estimates of nutrient density appeared to have acceptable relative validity for use in epidemiologic studies.