Assessing change in diet-intervention research.

In nutrition-intervention research, it is important to consider the sensitivity of dietary assessment instruments to the changes in nutrient intake or dietary behavior under study. This presentation describes a measure called "responsiveness," an index of an instrument's sensitivity to change. Illustrations of this measure are from two randomized dietary-intervention trials that targeted reductions in fat intake: the Women's Health Trial (WHT), a trial to test whether fat reduction would reduce the risk of breast cancer, and the Eating Patterns Study (EPS), a trial to evaluate a self-help booklet to promote dietary change. In the WHT, a 4-d diet record (FDDR) was only slightly more responsive to dietary change than was a food-frequency questionnaire (FFQ). In the EPS, a fat-related diet-habits questionnaire was most responsive, followed by an FDDR and an FFQ. These data suggest that short, inexpensive measures such as FFQs or questionnaires that assess dietary habits can be as responsive as multiple-day diet records. More research is needed on the relative responsiveness of dietary assessment tools. Intervention studies should include at least two types of dietary assessment tools and the relative validity, reliability, and responsiveness of these tools should be reported as part of the study outcome.