Validity and reproducibility of an iodine‐specific food frequency questionnaire to estimate dietary iodine intake in older Australians

Abstract

Aim: Iodine deficiency, which has adverse effects on health has re-emerged in Australia. The present study aimed to develop and validate a novel iodine-specific food frequency questionnaire for use in older Australians. Methods: A 49-item food frequency questionnaire that included iodine-rich foods was constructed and administered in 84 men and women aged 60-95 years with normal cognitive function. Dietary iodine intake assessed by the food frequency questionnaire was validated against three repeated 24-hour dietary recalls. Urinary spot iodine concentrations were selected as iodine intake biomarker. Agreement between the two dietary methods was determined using a Bland-Altman plot and intra-class coefficients. Correlations between dietary and urinary iodine were assessed. Forty-three participants repeated the questionnaire after 9 months for reproducibility. Results: Mean iodine intake measured by the food frequency questionnaire and 24-hour dietary recalls did not differ significantly (P = 0.870). The two methods were moderately correlated (r = 0.377; P < 0.05) and the Bland-Altman plots demonstrated an acceptable level of agreement (P = 0.870). Despite an association (r = 0.230; P < 0.05) between urinary iodine concentrations and 24-hour dietary recalls, the food frequency questionnaire was not associated with urinary iodine concentration (r = 0.094; P = 0.40). The method of triads showed coefficients of 0.238 (urinary iodine), 0.953(food frequency questionnaire), 0.396 (24-hour dietary recall) with the unknown true value. Conclusion: A short food frequency questionnaire to assess habitual dietary iodine intake in older Australians has been shown to be valid at the group level with regard to categorising individuals according to their habitual iodine intake. Reproducibility of the food frequency questionnaire remains to be demonstrated.

Improvement in iodine status of pregnant Australian women 3years after introduction of a mandatory iodine fortification programme

“…In the 2012 study data were also collected on participants' perceived dietary adequacy, good food sources of iodine, changes to diet, supplement use, adequacy of dietary education received and awareness of the mandatory iodine fortification program. A validated iodine-specific food frequency questionnaire (FFQ) (Tan et al, 2012) was administered in 2011 and 2012, which assessed food intake over the previous month (including dairy, eggs, cereal products, fish and seafood, meat, vegetables, fruit and mixed dishes), as well as tap water consumption and use of iodised salt.…”

Dietary iodine exposure and brain structures and cognition in older people. Exploratory analysis in the Lothian Birth Cohort 1936

“…To the best of our knowledge, there are only two dietary questionnaires focusing on iodine intake specifically, one developed in younger UK female adults (25), and the other in older Australian adults (26). They also provide insight on the intake only on the months preceding their application, rather than intake throughout life.…”

Improved interpretation of studies comparing methods of dietary assessment: combining equivalence testing with the limits of agreement

“…Independent t test methods also exist for both normally distributed and non-normal data. It is often the case that dietary intake data are skewed as was the case with the initial data set on which the simulation used in the present analysis was based (8) . Non-normal data can be analysed using a similar approach for either paired or independent data based on the robust t test of Yuen (36) in the 'equivalence' package in R. Log transformation can also be considered.…”

“…This study was designed to assess methodological comparison studies based on agreement using an example based on our previous research. There are other methods for judging the usefulness of new dietary assessment tools, such as the method of triads, which we have used previously (8) , or missclassification, but they are not discussed here. Lombard et al (38) provided a recent review and recommendations on the use of other methods, specifically applicable to nutrient assessment.…”

“…To demonstrate equivalence testing and compare this with the BA method, a data set consisting of a random sample of eighty participants was simulated based on a previously published analysis using the means of I intake assessed, using the average of three repeated 24-HR and a FFQ (3 × 24HR 118·88 (SD 48·95) µg, FFQ 120·19 (SD 55·98) µg and correlation 0·614; P < 0·001) (8) . The data set was simulated using the matrix and drawnorm commands in STATA (version 12, STATA Inc.).…”