A culture-sensitive quantitative food frequency questionnaire used in an African population: 1. Development and reproducibility

Abstract

AbstractObjective:To develop and assess the reproducibility of a quantitative food frequency questionnaire (QFFQ) sensitive to the culture of the African population of the North West Province, South Africa.Design:A cross-sectional study.Setting:A community-based field study in a population stratified according to level of urbanization from deep rural to urban upper class.Subjects:A total of 144 (99 women and 45 men) residents of the North West Province, aged between 15 and 65 years, participated in the study.Methods:A culture-sensitive, 145-item interviewer-administered QFFQ was designed to cover the whole diet. Portion sizes were estimated from a food portion photograph book (FPPB) showing foods in three portion sizes. The QFFQ was administered twice, 6–12 weeks apart.Results:Spearman rank correlation coefficients between the two administrations varied from 0.14 for calcium to 0.75 for alcohol. The mean percentage difference between intakes was 8.5 (standard deviation = 9.9). Energy, protein, carbohydrate and calcium gave differences within 10%. Few significant differences among correlation coefficients or percentage difference for gender, age group or strata of urbanization were present. Bland–Altman plots showed significant proportional bias for protein, fibre and vitamin C. More than 70% of the participants were classified into adjacent quintiles for all nutrients. For food groups, correlation coefficients ranged from 0.25 for milk to 0.45 for vegetable and maize meal groups and 80% of participants were classified into adjacent quintiles.Conclusions:The QFFQ appeared to be a reproducible dietary intake assessment instrument.