Commentary: Dietary diaries versus food frequency questionnaires—a case of undigestible data

Abstract

In conducting studies of diet and disease risk, methods of measuring diet with sufficient validity to detect important associations are essential. Cost is also a critical factor because prospective studies, which are necessarily large, are desirable to avoid problems of selection and recall bias. Most investigators have converged to use some form of a food frequency questionnaire (FFQ) for this purpose, and the validity of this approach has been documented repeatedly by comparisons with more detailed methods, correlations with biochemical indicators of dietary factors, and the ability to predict risk of future disease.^1,2 However, all methods of dietary assessment are imperfect, and quantification of measurement error is desirable both to help in the interpretation of findings from epidemiological studies, and to correct relative risks and confidence intervals for this source of error.