Validation of a Brief Dietary Assessment Tool for Estimating Dietary EPA and DHA Intake in Australian Adults at Risk of Cardiovascular Disease

Background: Short food frequency questionnaires (FFQs) are powerful screening instruments for estimating nutrient intakes and play an important role in risk stratification in vulnerable populations. Omega-3 fatty acids are of increasing importance in the prevention of chronic and degenerative disease, especially in older adults who are at higher risk of these chronic conditions. A short FFQ exists to rapidly assess omega-3 intake from marine sources, however it has not previously been validated for agreement with total omega-3 intake and ability to identify suboptimal omega-3 intakes in older adults or for use in Australia. Objective: The aim of the study was to validate a 9-item marine omega-3 FFQ (MFQ) for assessment of omega-3 intake against a validated 74-item semiquantitative FFQ. Methods: One hundred and eight participants (mean age 67 ± 10 years, 47% male) completed the MFQ designed to estimate omega-3 intake from marine sources and the 74-item FFQ designed to estimate usual omega-3 and total energy intake in addition to other nutrients. To test agreement between the two questionnaires for estimating total omega-3 intake, mean bias and 95% limits of agreement (LOA) were calculated. Sensitivity and specificity were assessed using 2 × 2 contingency tables based on whether participants did or did not meet National Heart Foundation (NHF) recommendations for omega-3 intake per day. Results: Mean intake of omega-3 estimated from the MFQ was 210 ± 235 mg/day compared with 295 ± 260 mg/day as estimated by the FFQ. Mean bias (95% LOA) for omega-3 assessed by the two questionnaires was 89 mg/day (−475 mg to 653 mg). The MFQ achieved 98% sensitivity and 31% specificity for the omega-3 cut-off of 500 mg/day. When nutrient composition of the marine products were replaced with Australian data, mean intake of omega-3 was 230 ± 253 mg and the mean bias improved to 64 mg (−681 mg to 553 mg) and achieved 93% sensitivity and 40% specificity. Conclusions: The MFQ shows promise as a rapid screening tool for identifying older adults with intakes of omega-3 fatty acids likely to be below recommendations for chronic and degenerative disease risk reduction. Given the clinically meaningful mean bias and wide LOA, it cannot be recommended as an appropriate tool for the purpose of reporting average intake of individuals. Use of Australian nutrient data improved the mean bias of the tool in estimating total omega-3 intake. The values should be replaced and the MFQ could then be a useful tool for research purposes at the population level.