Reliability and Validity of a Culture-Specific Dietary Assessment Questionnaire in the Adult Population of Trinidad and Tobago

Lesley Ann Foster-Nicholas,David Shavlik,Patricia Dyett,Celine Heskey,Gina Segovia-Siapco
DOI: https://doi.org/10.1093/cdn/nzaa061_027
2020-05-29
Current Developments in Nutrition
Abstract:Abstract Objectives To evaluate the reliability and validity of a culture-specific semi-quantitative food frequency questionnaire (FFQ) in assessing nutrient intake estimates of the adult population of Trinidad and Tobago. Methods An electronic 154-item semi-quantitative FFQ was created using Google forms to reflect the culture-specific and street foods consumed by persons living in Trinidad and Tobago. The self-administered FFQ was emailed twice, 3 months apart, to 91 participants ≥18 years. The first FFQ was validated against 4 one-day food records with digital images which served as the reference method. Using SPSS Version 26, the validity and reproducibility of the FFQ's nutrient intake estimates were determined using paired t-tests, bivariate correlations and cross-classifications. Correlations were calculated on energy-adjusted nutrients. Cross-classification was used to show agreement between the FFQ and weighted average food records. Results Participants consisted of 22% male and 78% females, median age 38 ± 10 years. Intake estimates of nutrients in both administrations of the FFQ were highly reproducible: Correlations ranged from medium (r = 0.58, P ≤ 0.0001, for vitamin C) to high (r = 0.83, P ≤ .0001, for carbohydrates); cross-classification agreements within ±1 quintile ranged from 69% (cholesterol) to 89% (fiber and vitamin A). For validity, the energy-adjusted correlations average was r = 0.37, with correlations (r) ranging from 0.22 (P ≤ .050, polyunsaturated fatty acids (PUFA)) to 0.67 (P ≤ .0001, cholesterol). Cross-classification showed that an average of 61% of the FFQ estimates were correctly classified within ±1 quintile of exact agreement in the food records. Agreements between the FFQ and the food records ranged from 48% for calcium to 71% for cholesterol, all significant at P ≤ 0.0001. Conclusions The culture-specific FFQ has a good level of reliability and validity for use in future studies to assess nutrient intake of adults living in Trinidad and Tobago. Funding Sources N/A.
What problem does this paper attempt to address?