Development, validation, and reproducibility of food group-based frequency questionnaires for clinical use in Brazil : a pre-hypertension and hypertension diet assessment

Abstract

The Blood pressure control diet is well described; however, it has not been implemented in clinical care, possibly due to the impracticability of the diet assessment in these contexts. In order to facilitate the dietary assessment, we developed and assessed the validity and reproducibility of two food group-based food frequency questionnaires (FG-FFQs), with a one-week (7-day FG-FFQ) and a one-month (30-day FG-FFQ) period of coverage for patients with pre-hypertension or hypertension. In 2010, 155 men and women, 30–70 years old, were invited to participate in a prospective study in two outpatient clinics in Porto Alegre, southern Brazil. The participants responded to two 30-day, two 7-day FG-FFQ, four 24-h dietary recalls, and underwent demographic, anthropometric, and blood pressure assessments. The validity and reproducibility were assessed using partial correlation coefficients adjusted for sex and age, and the internal validity was tested using the intra-class correlation coefficient. The participants were aged 61 ( 10) years and 60% were women. The validity correlation coefficient was higher than r = 0.80 in the 30-day FG-FFQ for whole bread (r = 0.81) and the 7-day FG-FFQ for diet/light/zero soda and industrialized juices (r = 0.84) in comparison to the 24-h dietary recalls. The global internal validity was = 0.59, but it increased to = 0.76 when 19 redundant food groups were excluded. The reproducibility was higher than r = 0.80 for pasta, potatoes and manioc, bakery goods, sugar and cocoa, and beans for both versions. The 30-day had a slightly higher validity, both had good internal validity, and the 7-day FG-FFQ had a higher reproducibility.