Relative validity of a web-based, self-administered, 24-h dietary recall to evaluate adherence to Canadian dietary guidelines

A new, web-based, 24-h recall (R24 W) was developed and programmed to assess adherence to Canadian dietary guidelines by automatically generating the Canadian Healthy Eating Index 2007 (C-HEI 2007). The aim of the present study was to determine the relative validity of the R24 W in terms of its ability to generate C-HEI 2007 scores that match those derived from a food record (FR). A total of 57 women and 50 men filled a 3-d FR and three 24-h food recalls with the R24 W. C-HEI 2007 was calculated with both dietary assessment tools and compared using de-attenuated correlations and kappa scores. Internal validity (correlations between total score and components) and external validity (comparison of the C-HEI 2007 by sex and age groups) were compared between the two dietary assessment methods. The average C-HEI 2007 score obtained with the R24 W (59.4 ± 11.8) was significantly lower than the one obtained with the FR (62.6 ± 11.1). However, the de-attenuated correlation coefficient between C-HEI 2007 obtained either with R24 W or FR was 0.80 (P < 0.01) and 91.6% of participants were classified in the same or adjacent quartile of C-HEI 2007 with both tools. Women obtained a higher C-HEI 2007 mean score (P < 0.01) with both methods and older adults (age: >51 y) received a higher score than younger adults (age: 18–30 y; P < 0.01) with the R24 W but not with the FR (P = 0.22). Although the R24 W yields lower C-HEI 2007 scores compared with data from FR, the classification of individuals was highly consistent between the two measurements.