Dietary adherence in feeding studies

Nutrition Research Newsletter, Sept, 2005 by Denise Hall, Marlene Most

In nutrition research, well-controlled feeding studies are valued for their ability to examine physiological response to diet with high precision. Participants in these studies must consume foods and beverages that are prepared in a research kitchen, usually without eating other foods. Participant adherence to the diet is essential. Knowing the level of adherence to the study diet is important for understanding the true influence of the intervention diet on study endpoint measurements. Researchers must also consider aspects of the study design itself that may encourage or discourage participant adherence. A recent study in JADA examined how study design aspects influenced diet adherence in a variety of studies.

All participants surveyed had completed one or more of eight controlled feeding studies where all meals were provided at the Pennington Biomedical Research Center in Baton Rouge, LA. One hundred and fifty four people responded to the post study anonymous questionnaire. The questionnaire contained 21 questions concerning the participants' dietary adherence during the study and factors that may have influenced adherence.

Total adherence to the diet (consumed all study foods/ beverages plus refrained from consuming nonstudy foods/ beverages) was reported on 48% of the questionnaires, and poor adherence (more than four deviations) was reported on 7%. Adherence did not differ with regard to sex, marital status, educational level, or personal income level. Eating all of the foods provided was a greater challenge than refraining from eating foods not allowed. Diet assignment, the allowance of alcohol, and a self-selected Saturday meal affected adherence. The number of repeated days in a menu cycle did not influence adherence.

The results showed that more than 90% of the respondents either always or usually adhered (deviating less than four times) to the diet during the study period. The anonymous nature of the survey encouraged honest answers. Of course, the adherence of those who did not respond to this study is unknown and the responses may reflect a biased sample, with the most adherent participants completing the survey. However, results from well-executed feedings studies may be viewed with confidence and because deviations are few they are unlikely to jeopardize study results.

Denise Hall, and Marlene Most. Dietary adherence in well-controlled feeding studies. JADA; 105:1285-1288 (August 2005) [Address correspondence to: Marlene Most, PhD, RD, FADA, Associate Professor of Research, Metabolic Kitchen, Pennington Biomedical Research Center, Louisiana State University System, 6400 Perkins Rd, Baton Rouge, LA 70808. E-mail: mostmm@pbrc.edu]