The ‘Carrot Test’ : An approach to characterize individual differences in oral processing behaviour and eating rate

Resúmen

Background: Eating rate is a modifiable risk factor for obesity and efficient methods to objectively characterise an individual's oral processing behaviours could help better identify people at risk of increased energy consumption. Many previous approaches to characterise oral processing and eating rate have relied on specialised equipment or wearable devices that are time consuming, expensive or require expertise to administer. The current trial used video-coding of the consumption of a standardised test food (the ‘carrot test’) to measure oral processing. Objective: We sought (i) to test whether self-reported eating rate (SRER) is predictive of food oral processing derived from coded eating behaviours captured in the laboratory with a standardised test food, and (ii) to test whether differences in SRER are predictive of oral processing behaviours, eating rate and intake of a test meal. Methods: Two hundred and fifty-three volunteers (86 male and 167 female, mean age 39.5 ± 13.6 years, mean BMI 22.2 ± 3.4 kg/m²) provided their SRER and anthropometric measurements of height, weight and dual-energy X-ray absorptiometry (DEXA) percentage fat mass. Participants were also video recorded eating a fixed 50 g portion of carrot and an ad libitum lunch meal of fried rice. Average eating rate (g/min), bite size (g) and number of chews per bite for the carrot and lunch were derived through behavioural coding of the videos. Energy intake (kcal) was recorded at lunch and a later afternoon snack. Results: Faster SRER significantly predicted faster eating rate, larger bite size and more chews per bite observed during intake of the carrot (ß = −0.26–0.21, p ≤ 0.001) and the lunch (ß = −0.26–0.35, p ≤ 0.014). SRER did not significantly predict intake at lunch or during the afternoon snack (ß = 0.05–0.07, p ≥ 0.265). Participants’ oral processing of the carrot significantly predicted oral processing of the lunch (ß = −0.25–0.40, p ≤ 0.047) and faster eating rate of the carrot significantly predicted increased lunch intake (ß = 0.119, p = 0.045). None of the oral processing behaviours predicted afternoon snack intake (ß = −0.01–0.05, p ≥ 0.496). None of these associations were moderated by BMI or body composition. Conclusion: We confirm that SRER is a valid measure of group level differences in individual oral processing behaviours, but did not predict an individual's energy intake at a lunch-time meal. With this approach, it is possible to characterise differences in eating rate by coding eating behaviours for a standardized test food (in this case, a fixed portion of raw carrot). This approach could be used to provide an objective measure of a person's habitual oral processing behaviour, and was shown to be a significant predictor of eating rate and energy intake for a later test meal.