Date Published: April 18, 2019
Publisher: Public Library of Science
Author(s): Anne Marte Wetting Johansen, Jannicke Borch Myhre, Anette Hjartåker, Lene Frost Andersen, Michele Drehmer.
The aim of this study was to validate energy intake (EI) reported by a pre-coded food diary (PFD) against energy expenditure (EE) measured by the ActiReg system consisting of an activity and position monitor and a calculation program (ActiCalc). Dietary intake was recorded by the PFD and EE was measured by the ActiReg system over a 7-day period. One hundred and twenty adult participants completed the study, 42 men and 78 women. The average group EI was 17% lower compared to measured EE. The 95% limits of agreement were 6.7 and -2.9 MJ/day. Of all participants, 68% were classified as acceptable reporters, 29% as underreporters and only 3% as overreporters. Fifty percent of the men and 30% of the women were classified into the same quartile for EI and EE, whereas 5% of both men and women were ranked in the opposite quartile by the two methods (weighted kappa coefficient = 0.29). Pearson correlation coefficient between reported EI and measured EE was 0.49 (p<0.001). High BMI was related to larger underreporting when EE was low. Furthermore, this study found that PFD underestimates EI on the group level with an average of 17% and showed large variation in the validity of the PFD on the individual level.
For several decades nutritional researches have worked on improving dietary assessment methods. A valid dietary assessment method measures the true dietary intake over a defined time period. Errors in diet-report instruments can result in important diet-disease relationships being overlooked . For instance, Bingham et al. suggested that dietary measurement error might explain the absence of a significant association between dietary fat intake and breast cancer risk in cohort studies . Underreporting of energy intake (EI) is a common phenomenon in nutrition research . Since EI is closely linked to the intake of other nutrients, underreporting of EI typically results in underreporting of nutrient intakes. Underreporting of EI is either due to under-eating, underrecording or a combination of the two. Factors that have been associated with underreporting are obesity, weight consciousness, restrained eating, socio-economic status and social desirability [4, 5].
Characteristics of the participants and the main results are presented in Table 1. Sixty-five percent of participants were women and mean (SD) age and BMI were 35 (13) years and 24.1 (3.4) kg/m2 among all participants. Mean age and BMI did not differ significantly between men and women. EE was significantly higher than EI for the group of participants as a whole (mean difference 1.9 MJ/day, p<0.01). The difference between EE and EI was significantly larger for men (mean difference 2.8 MJ/day) than for women (1.4 MJ/day, p<0.01). When EE-EI was calculated as the percentage of EE, the gender difference was no longer statistically significant (20.3% versus 13.3%, p = 0.07). Results from the present study showed that the PFD underreported EI by an average of 17% compared to EE measured with the ActiReg system. Participants with high EE underreported more than participants with low EE, independent of BMI. Among participants with low EE, high BMI was associated with larger underreporting. In summary, the PFD underreported EI by 17% compared to EE measured with the ActiReg in a group of motivated adults in Norway. This finding is in line with what similar validation studies of prospective dietary assessment methods have found previously. The degree of underreporting increased with increasing EE, also when we stratified by BMI. Thus the PFD seems to be less useful in groups of persons with high EE. The PFD showed a moderate ability to rank individuals according to EI. Source: http://doi.org/10.1371/journal.pone.0215638