The Bambui health and aging study: is calibration of dietary intake necessary among older adults?

ABSTRACT

Accurate measurements of food intake in surveys are difficult to be obtained especially in the older adults. Calibration, using two instruments, has been adopted in order to improve the information. This study was carried out in a random sample of older adults (> or =60) living in a Brazilian town. Food intake was obtained by the semi-quantitative food frequency questionnaire associated with photos (SFFQ-P) and the 24 hours recall (R24) with or without food models. Statistical analysis included tests to compare averages, Pearson's correlation coefficient and multiple linear regression. For all nutrients, the average intake obtained through SFFQ-P was significantly higher compared to the obtained by the R24 (p < or =0.05). SFFQ-P correlated better with R24 with food models, which was chosen as the reference method. In the multiple linear regression, vitamin C, fatty acid, protein and zinc intakes were influenced by age, especially when such nutrients were more concentrated in food not frequently reported. For protein and zinc, the significantly interaction between age with the 24 hours recall may have happened because of a differential selection of food sources as a function of age. This fact is probable related to chewing difficulties, reduction in the income and, a diet simplification associated with less complicated preparations, such as meat. This study reinforces the need for calibration of SFFQ in nutritional surveys among older adults, especially in etiological studies where the exposure assessment has to be accurate. In this case, R24 with food models should be used as a reference method to best estimate the true intake.