Background

Dietary assessment based on self-reported or proxy-reported data is often prone to misreporting, including both under- and over-reporting. Evaluation of this bias is crucial for correct interpretation of dietary exposure.

Methods

The prevalence and characteristics of misreporting of energy intake was investigated among 3096 participants (3-64 years old) in the Belgian National Food Consumption Survey 2014 (FCS2014). Information on food intake was collected with two non-consecutive 24-hour dietary recalls (GloboDiet®). Objective measures were obtained for anthropometric parameters. Misreporting was assessed using the Goldberg method. Additionally, the prevalence and size of underreporting was compared with FCS2004 for participants aged 15-64.

Results

Overall in 2014, 24.3% of participants underreported, 75.2% were plausible reporter and 0.5% overreported. Females underreported more often than males (OR 1.37; 95% CI 1.16-1.63). Under-reporting was less common in children (3-9 years) where parents reported dietary intake, than in adolescents (10-17 years) (OR 0.17; CI 0.12-0.26) and adults (18-64 years) (OR 0.49; CI 0.26-0.95) who self-reported. Obese participants underreported more often than overweight (OR 1.92; CI 1.45-2.55), normal weight (OR 3.71; CI 2.84-4.84) or underweight participants (OR 5.57; CI 3.37-9.21). Low educational level (OR 1.44; CI 1.21-1.71) and energy-restricted diet (OR 3.81; CI 2.40-6.03) increased the odds of under-reporting. The prevalence of underreporting was higher in 2014 (34%) than in 2004 (28%), but excluding under-reporters resulted in both FCS in an increase of the mean habitual energy intake of 300 kcal/day at population level. Further comparison with FCS2004 will be presented.

Conclusions

Underreporting of energy intake was associated with individual characteristics such as sex, age, BMI, educational level and diet. Proxy-reporting by parents resulted in less underreporting than self-reporting in adolescents and adults.