Relative validity of three different dietary assessment tools as a part of a food-based clinical trial for weight loss
Clinical trials require standardised dietary practices to allow prescribed dietary targets to be achieved and monitored. The methods of dietary intake assessment play a vital role in this process, though new innovative methods may help to streamline the process further. This cross-sectional study aims to statistically compare the nutrient data obtained at baseline from diet history (DH), food record (FR) and self-administered computerised dietary assessments (DA) conducted in a food-based clinical trial for weight-loss. Data for n=71 (n=15 male and n=56 female participants, 45.4±7.9yrs, BMI 31.6±3.5kg/m²) was obtained of which 69 matching DH and FR, 32 matching DH and DA and 30 matching FR and DA were available. Significant differences were found for all macronutrients for all matching assessment methods. The relationship between assessment methods varied widely: DH-DA (r=0.123-0.704), DH-FR (r=0.241-0.504), DA-FR (r=0.250-0.596). For energy (kJ) DA had the strongest correlations (DH r=0.704, FR r=0.596) compared with DH and FR (r=0.425). The nutrient with the weakest correlations for all assessment methods were the monounsaturated fatty acids (DH-FR r=0.320, DH-DA r=0.123, DA-FR r=0.250). Interestingly, Bland-Altman analyses for DA-FR and DA-DH showed identical scatter patterns for energy (kJ). The results suggest that the food record and computerised dietary assessments are most comparable at baseline possibly due to the method of administration (participant completed). Though the sample size is small, these results suggest that new computer-based technologies for dietetics may provide added benefits in the clinical trials setting. Funding source: The work of this project was funded under an NHMRC project grant and a University of Wollongong Early Career Researcher Health and Behavioural Sciences grant.