MEASURES OF FOOD CHOICE BEHAVIOR RELATED TO INTERVENTION MESSAGES IN WORKSITE HEALTH PROMOTION

Worksite nutrition educators and researchers need not only nutrient co mposition data but also information on employees' food choices to crea te intervention messages that are food-focused and tailored to specifi c target audiences. This paper describes a method of calculating measu res of food choice behavior related to intervention messages and repor ts relationships between workers' food choices and demographic charact eristics. This work was conducted as part of the Working Well Trial, a 5-year worksite cancer prevention randomized, controlled prospective field experiment funded by the National Cancer Institute. It was imple mented in 114 worksites employing 37,291 workers who were engaged in a variety of businesses. In the fall of 1990, 20,801 respondents comple ted and returned a self-administered baseline survey. A modified Block 88-item food frequency questionnaire (FFQ) with portion sizes was use d for the quantitative assessment both of nutrient intake and workers' food choices. The worksite mean response rate was 71.6%. Responses to behavioral items regarding meat were used to measure meat preparation behaviors not captured on the FFQ. We found that higher education, gr eater age, and female gender were associated with food choices closest to the recommendations to increase fiber, fruits, and vegetables and to reduce fat. An exception to this pattern was the recommendation to increase the consumption of beans and lentils, which was associated wi th lower education, greater age, and male gender. The relationship of job status to the food choice variables was inconsistent across foods and study centers and there were differences between study centers in the magnitude of associations between food choices and demographic cha racteristics.