Development and Evaluation of the National Cancer Institute's Dietary Screener Questionnaire Scoring Algorithms

Methods for improving the utility of short dietary assessment instruments are needed. We sought to describe the development of the NHANES Dietary Screener Questionnaire (DSQ) and its scoring algorithms and performance. The 19-item DSQ assesses intakes of fruits and vegetables, whole grains, added sugars, dairy, fiber, and calcium. Two nonconsecutive 24-h dietary recalls and the DSQ were administered in NHANES 2009-2010 to respondents aged 2-69 y ( = 7588). The DSQ frequency responses, coupled with sex- and age-specific portion size information, were regressed on intake from 24-h recalls by using the National Cancer Institute usual intake method to obtain scoring algorithms to estimate mean and prevalences of reaching 2 a priori threshold levels. The resulting scoring algorithms were applied to the DSQ and compared with intakes estimated with the 24-h recall data only. The stability of the derived scoring algorithms was evaluated in repeated sampling. Finally, scoring algorithms were applied to screener data, and these estimates were compared with those from multiple 24-h recalls in 3 external studies. The DSQ and its scoring algorithms produced estimates of mean intake and prevalence that agreed closely with those from multiple 24-h recalls. The scoring algorithms were stable in repeated sampling. Differences in the means were