Background: A pilot study was carried out to evaluate the relative validity and reproducibility of a dietary history questionnaire (DH) designed to be used in Spain as part of a prospective European research project on diet and cancer.
Methods: A total of 91 volunteers (46 men and 45 women) aged 35-60 years, and of various occupational categories, participated in the study. The DH was structured according to occasions of intake and was administered twice by interview, with a year's interval. Twelve 24-hour diet recalls, administered monthly between the two administrations of the DH, were used as a reference method. Natural units, geometric figures and a set of photographs were used to quantity portion size.
Results: Of a total of 17 food groups, the Pearson correlation coefficients between the daily intake of food groups, based on the mean of 12 24-hour diet recalls, and the second DH showed values between 0.65 and 0.90 in 14 groups (potatoes, vegetables, fruits, cheese, other dairy products, bread, other cereals, meat and poultry, fish and shellfish, butter, margarine and oils, sugar and sweets, other desserts, alcohol and coffee and tea) in males and in 11 groups (potatoes, other vegetables, fruits, cheese, other dairy products, bread, meat and poultry, fish and shellfish, sugar and sweets, alcohol and coffee and tea) in females. In comparison with the mean of 12 24-hour diet recalls the DH questionnaire over-estimated (in both sexes) the intake of vegetables, cereals, butter, margarine and oils and underestimated the intake of sugar and sweets, and other desserts.
Conclusions: The DH method is more expensive and more difficult to apply than other instruments but it provides good information on the habitual food intake of the individual.