The Eating Behavior Patterns Questionnaire predicts dietary fat intake in African American women DAVID G. SCHLUNDT, PhD; MARGARET K. HARGREAVES, PhD; MACIEJ S. BUCHOWSKI, PhD ABSTRACT Objective To develop a behavioral assessment of eating that would be predictive of fat intake in African American women. Design Questionnaires were developed using a three-stage design, involving item generation, item refinement, and ques- tionnaire validation. Subjects Focus groups sessions were conducted with 40 Af- rican American women, initial questionnaire development employed 80 African American women, and questionnaire validation involved 310 African American women from di- verse socioeconomic backgrounds. Statistical analyses Transcripts of focus groups were used to generate 113 behavioral questionnaire items. The initial questionnaire was administered along with a food frequency questionnaire, and the item pool was reduced to 51 items. Factor analysis was used to create subscales. Correlation (r) and multiple regression analysis (R) were used to evaluate construct validity. Results Factor analysis revealed six subscales: low-fat eat- ing, emotional eating, snacking on sweets, cultural/ethnic, haphazard planning, and meal skipping. The scales are signifi- cant predictors of micronutrient (R values from 0.22 to 0.47) and macronutrient intakes (R values from 0.33 to 0.58) as- sessed using a food frequency questionnaire and show con- struct validity in relationship to other measures of eating be- havior (r values from 0.22 to 0.65). Applications The Eating Behavior Patterns Questionnaire (EBPQ) may be a useful tool for clinical assessment, clinical and community nutrition intervention studies, and epidemio- logic research with African American women. J Am Diet Assoc. 2003;103:338-345. H igh-fat and low-fiber intakes have been suggested as pos- sible etiologic factors in breast, colon, prostate, and other cancers (1-4) as well as heart disease, obesity, and noninsulin-dependent diabetes (5-8). Prevalence of these chronic diseases is higher in minority groups (9,10), es- pecially African American women (11,12). Because popula- tion-wide reductions in consumption of dietary fat could poten- tially have a substantial impact on morbidity and mortality (8), there is a need for a dietary assessment tool that can be readily applied to the problem of changing behavior. Three methodol- ogies have traditionally been used for quantifying fat intake: (a) dietary records, (b) dietary recalls, and (c) food frequency questionnaires (13,14), but they are cumbersome to apply (13- 20), and they do not capture behaviors associated with at- tempts to increase or decrease a specific food or nutrient (21). The Kristal Eating Patterns Questionnaire was developed to measure behavioral attempts to restrict fat and increase fiber intakes (22,23). It was validated against the traditional meth- ods of dietary assessment. Hargreaves and colleagues devel- oped and evaluated a 16-item Eating Styles Questionnaire (ESQ), designed specifically for use with African American women, to measure dietary fat intake associated with their stages of behavior change (24). Neither our ESQ nor the Kristal Eating Patterns Questionnaire, however, included important behaviors, such as emotional eating and impulsive snacking, that appear to influence fat and fiber intakes (25). Failure to consider both healthy and unhealthy behaviors may affect the D. G. Schlundt is an associate professor, Department of Psychology and Diabetes Research and Training Center, Vanderbilt University, Nashville, TN; M. K. Hargreaves is an associate professor with the Drew-Meharry-Morehouse Consortium Cancer Center and the Department of Internal Medicine, Meharry Medical College; and M. S. Buchowski is a professor, Department of Family and Community Medicine, Me- harry Medical College, Nashville, TN. Address correspondence to: David G. Schlundt, PhD, Department of Psychology, Vanderbilt University, 301 Wilson Hall, Nashville, TN 37203. E-mail: D.Schlundt@ Vanderbilt.Edu Copyright © 2003 by the American Dietetic Association. 0002-8223/03/10303-0004$35.00/0 doi: 10.1053/jada.2003.50049 RESEARCH 338 / March 2003 Volume 103 Number 3