Measuring food intake in studies of obesity

Abstract

The problem of how to measure habitual food intake in studies of obesity remains an enigma in nutritional research. The existence of obesity-specific underreporting was rather controversial until the advent of the doubly labelled water technique gave credence to previously anecdotal evidence that such a bias does in fact exist. This paper reviews a number of issues relevant to interpreting dietary data in studies involving obesity. Topics covered include: participation biases, normative biases, importance of matching method to study, selective underreporting, and a brief discussion of the potential implications of generalised and selective underreporting in analytical epidemiology. It is concluded that selective underreporting of certain food types by obese individuals would produce consequences in analytical epidemiological studies that are both unpredictable and complex. Since it is becoming increasingly acknowledged that selective reporting error does occur, it is important to emphasise that correction for energy intake is not sufficient to eliminate the biases from this type of error. This is true both for obesity-related selective reporting errors and more universal types of selective underreporting, e.g. foods of low social desirability. Additional research is urgently required to examine the consequences of this type of error.

Recommendations for a trans-European dietary assessment method in children between 4 and 14 years

“…In general, the studies show a tendency to an increase in underreporting with increasing age, the reason for that is thoroughly discussed in papers by Livingstone and Robson, 2000;Livingstone et al, 2004. Moreover, it was identified that obese adolescents and children underreported energy intake to a larger extent than normal-weight subjects, which is comparable to results found among adults (Lissner, 2002;Livingstone et al, 2004). Most of the validation studies were evaluating diet records with varying length.…”

Measuring dietary intake in children and adolescents in the context of overweight and obesity

“…It is commonly assumed that correction by energy-adjustment methods will in some way rectify the problem, without having to exclude data. 26,27 However, this may only be the case if under-reporting has occurred uniformly across foods at the level of the whole diet 26 and this is not likely to be so. Energy-adjustment methods cannot eliminate bias because of selective reporting of foods or beverages with the net effect being that nutrients are not under-reported in direct proportion to energy.…”

“…22 Differences in error and biases between populations further highlight the need to validate dietary assessment methods in the population for which their use is intended. 26 Additional issues to consider include the purpose of the dietary assessment, subject burden, resources and time available to collect and analyze the dietary data, expertize required to interpret the data and whether group-or individual-level data is required.…”

“…27 Hence, correction for EI does not eliminate this bias from selective reporting error. 26 A research priority is to improve understanding of the errors associated with reported dietary intakes and to develop statistical approaches to data analysis that attempt to deal this error. 22 Differences in error and biases between populations further highlight the need to validate dietary assessment methods in the population for which their use is intended.…”

Soft drinks consumption, diet quality and BMI in a Mediterranean population

“…P for linear trend 5 0?001 Unfortunately, energy under-reporting is a common problem for the analysis of associations between diet and health outcomes, particularly BMI (16,17) . It has been shown that obese subjects tend to selectively under-or overestimate selected foods (16,17) . In the present study, prevalence of low energy reporting decreased with frequent soft drinks consumption.…”