Validity and reproducibility of a self-administered food frequency questionnaire in older people

Abstract

This study assesses the validity and reproducibility of a 145-item self- This change has seen a need for dietary assessment tools which are able to measure long-term habitual nutrient intake. The cost efficiency, together with ease of both implementation and of data analysis, has lead to the popularity of the food frequency questionnaire (FFQ) as a tool for use in large epidemiological studies.2 Studies involving investigation of nutrients or foods and age-related disease associations rely on dietary information obtained from older subjects. The FFQ offers advantages for application to elderly populations as it does not rely on shortterm memory and involves a relatively low burden to respondent^.^ Despite the fact that the population of interest is often elderly, few studies have investigated the validity of self-administered food frequency questionnaires among older subject^.^-^^ Of these, only a handful have involved comparison with weighed food records and fewer have used self-administered FFQs.A valid and reliable dietary assessment instrument that can be self-administered by older lay persons and analysed using software that requires minimal coding input from dietitians or nutritionists, may enable largescale epidemiological studies involving food intake as an exposure variable without significant increase in study costs.This study aimed to assess both the concurrent validity for nutrient estimates of a self-administered 145-item semi-quantitative FFQ compared to three, four-day weighed food records (WFRs) spaced evenly over a year among respondents in a representative older Australian population. WFRs collected over a year are generally acknowledged as the best standard available, but measurement errors and changes to normal food consumption habits during recording periods are likely to be sufficiently great that the WFR cannot be regarded as a true gold standard. So, instead of assessing criterion validity, we can only really assess concurrent validity against a different type of dietary measure.A secondary aim was to assess the FFQ's short-term and longer-term repeatability. The questionnaire was used in the populationbased Blue Mountains Eye Study. Methods Study PopulationThe Blue Mountains Eye Study (BMES) is a population-based survey of vision and common eye diseases among older urban residents (aged 49 and older) in two post- 456

Food and nutrient consumption trends in older Australians: a 10-year cohort study

“…We acknowledge the potential limitations of an FFQ for measuring absolute intakes of food and nutrient intakes. However, our work on the validity and reliability of the FFQ shows that it has good repeatability and validity, and it ranks and classifies people relatively accurately into quintiles of the distribution, and has Spearman's correlation coefficients 40.5 for most nutrients (Flood et al, 2004;Smith et al, 1998b). Thus, although the point estimates of food and nutrient intakes may be over-or underestimated among individuals, with a tendency to overestimate at the group level, the observed changes, measured consistently by the same method over three time points, are likely to be real rather than the effects of measurement error associated with the FFQ.…”

“…The FFQ was tested for validity comparing nutrient estimates with those obtained from three separate 4-day weighed food records, collected over the period of 1 year to account for seasonal variation, in a randomly selected subsample of the BMES cohort (n ¼ 79) and was found to generally rank people well (Smith et al, 1998b;Flood et al, 2004).…”

Consumption of ‘extra’ foods (energy-dense, nutrient-poor) among children aged 16–24 months from western Sydney, Australia

“…For example, the FFQ used in the Blue Mountains Eye Study of older people lists the reference size of soft drink, cake, hot chips and ice cream as 1 can, 1 slice, 1 cup and 1/2 cup. These equate to 375 ml, 45 g, 113 g and 70 g, respectively, compared with the median 39,40 .…”

Assessment of dietary fat intake and innate immune activation as risk factors for impaired lung function

“…Dietary intake was assessed at baseline using a self-administered, semiquantitative food questionnaire, validated as previously described (Smith et al, 1998). Nutrient intakes were determined by a custom-made nutrient analysis programme based on NUTTAB 2006 database (Food Standards Australia New Zealand, 2006).…”