The use of computerised 24 h dietary recalls in the French SU.VI.MAX study: Number of recalls required

Abstract

Objective: The 24 h dietary recall is a widely used method to estimate nutritional intakes in epidemiological studies. The objective of the present study was to estimate the number of recalls necessary for an accurate estimation of nutrient intake in French adults followed for 4y. Subjects and methods: Participants of the SU.VI.MAX study (intervention study on the effects of antioxidant supplementation on chronic diseases) who completed a 24 h dietary recall every 2 months for at least 1 y. Inter- and intra-individual variance ratios Sw/Sb) were calculated by analysis of variance for two time periods: year 1 and 2 (n=4955) and year 3 and 4 (n=1458). The number of recalls necessary was calculated using an accuracy of 0.9. Results: The highest intra-individual/inter-individual variance ratio in the first period was seen for β-carotene and the lowest for carbohydrate. The number of recalls necessary was five for carbohydrate and calcium intake and 16 for β-carotene. For proteins, total and saturated fat, fibre, vitamin C and iron eight recalls were required, while nine, 11 and 10 recalls were necessary for mono- and polyunsaturated fat and vitamin E, respectively. The variance ratios in the second period were all lower and fewer recalls were therefore required. The same difference in number of recalls required between the two time periods was observed when only those subjects were included who completed at least 18 recalls (n=727). Conclusion: These results indicate that for an accurate estimation of carbohydrate intake only, already five recalls are necessary. Fewer recalls may be needed during long-time follow-up.