Relative Validity of a Short Food Frequency Questionnaire to Assess Intake of Non-Essential Energy-Dense Nutritionally-Deficient (NEEDNT) Food Items

by Hayley Jane Maxwell

Institution: University of Otago
Year: 0
Keywords: NEEDNT; FFQ; Food Frequency Questionnaire; Validation; Obesity; Weight management
Record ID: 1301957
Full text PDF: http://hdl.handle.net/10523/5527


Abstract Background: Obesity is increasing in prevalence around the world, posing a huge burden to society due to associations with non-communicable disease. The energy density of our diet is one factor that contributes to obesity. To reduce the prevalence of obesity, it has been suggested that lowering energy density should be prioritised over increasing physical activity and/or increasing consumption of fruit and vegetables. Energy dense foods often lack essential nutrients, and therefore these foods are not required in a healthy diet. The Non-essential Energy-Dense Nutritionally-Deficient Food Frequency Questionnaire (NEEDNT-FFQ) was based on the Non-Essential Energy-Dense Nutritionally-Deficient (NEEDNT) Food List, and tested for reliability. However, the relative validity of an FFQ must also be determined before it can be used to assess intake of non-essential energy-dense nutritionally deficient foods and beverages. Objective: The aim of the present study was to examine the relative validity of the NEEDNT-FFQ, a short non-quantitative FFQ based on the NEEDNT Food List. Design: This was a cross sectional study in adults (18-65years) with a BMI>30m/kg2. Participants completed the NEEDNT-FFQ around seven days after the completion of an estimated 7DDR. Results: Five men and 29 women participated in the study. Participants’ ages ranged from 33 to 64 years, and BMI ranged from 30.7-50.5kg/m2. Gender was significantly associated with NEEDNT-FFQ total scores, with men scoring 14 points higher than females. BMI was not significantly associated with NEEDNT-FFQ total scores. The mean percentage correctly classified and total percentage correctly and adjacently classified for all NEEDNT-FFQ items were 64.5% and 91.7% respectively. The mean percentage of NEEDNT-FFQ items misclassified into beyond adjacent categories was 8.3%. Five NEEDNT-FFQ items did not produce Kappa values, Spearman’s correlation coefficients (SCC) or intraclass correlation coefficients (ICC) as they were eaten so infrequently. Thirty-one (72.1%) NEEDNT-FFQ items obtained Kappa values above 0.2 (p<0.05), the mean Kappa value was 0.352. SCCs above 0.3 (p<0.05) were observed for 29 (67.4%) of NEEDNT-FFQ items, the mean SCC was 0.423. SCCs ranged from -0.072 for popcorn to 0.849 for alcoholic drinks. ICCs above 0.4 (p<0.05) were observed for 31 (72.1%) NEEDNT-FFQ items. The mean ICC was 0.550. ICC ranged from -0.455 for donuts and sweet breads to 0.956 for regular powdered drinks. NEEDNT-FFQ and 7DDR total scores produced a weighted kappa value of 0.288 (p=0.000), indicating a fair level of chance-adjusted agreement between the two methods, SCC of 0.656 (p=0.000) and ICC of 0.742 (p=0.000), indicating a strong positive relationship between the NEEDNT-FFQ and 7DDR. The paired t-test and Bland Altman plot for NEEDNT-FFQ and 7DDR total scores indicated that the NEEDNT-FFQ overestimated the total score by 7.9 points compared to the 7DDR. The paired t-test also indicated that 14 (27.7%) of NEEDNT-FFQ items were overestimated by the NEEDNT-FFQ relative to…