A computerized diet history questionnaire for epidemiologic studies.

OBJECTIVE: To describe methods used to computerize the diet history questionnaire developed for the Coronary Artery Risk Factor Development in Young Adult (CARDIA) study and to describe quality-control procedures used in conjunction with dietary assessment.
DESIGN: The computerized diet history is being used in a case-control study. Because of the computerized nature of the questionnaire, we developed quality-control procedures that incorporate listening to an audiotape of the interview while visually reviewing recorded data.
SETTING: Three centers involved in a population-based epidemiologic study of colon cancer.
SUBJECTS: Men and women between the ages of 30 and 79 years.
RESULTS: Quality-control results showed that 100% of the computerized forms would be free of errors if the data were subjected to visual review only. Probing errors, which accounted for 47.3% of all errors, were the most commonly encountered errors. In probing errors the interviewer did not probe in a nondirective manner, or the interviewer failed to verify responses that might be considered questionable. APPLICATIONS: The CARDIA diet history was computerized for use in epidemiologic studies of the association between diet and disease. Review of the audiotapes of the interviews showed that most errors made in obtaining a dietary assessment were not detectable from visual review of the data. Although the quality-control procedures were developed for a computerized diet history questionnaire, they are applicable to other dietary assessment methods.