Validation of a buffet meal design in an experimental restaurant

We assessed the reproducibility of intakes and meal mechanics parameters (cumulative energy intake (CEI), number of bites, bite rate, mean energy content per bite) during a buffet meal designed in a natural setting, and their sensitivity to food deprivation. Fourteen men were invited to three lunch sessions in an experimental restaurant. Subjects ate their regular breakfast before sessions A and B. They skipped breakfast before session FAST. The same ad libitum buffet was offered each time. Energy intakes and meal mechanics were assessed by foods weighing and video recording. Intrasubject reproducibility was evaluated by determining intraclass correlation coefficients (ICC). Mixed-models were used to assess the effects of the sessions on CEI. We found a good reproducibility between A and B for total energy (ICC = 0.82), carbohydrate (ICC = 0.83), lipid (ICC = 0.81) and protein intake (ICC = 0.79) and for meal mechanics parameters. Total energy, lipid and carbohydrate intake were higher in FAST than in A and B. CEI were found sensitive to differences in hunger level while the other meal mechanics parameters were stable between sessions. In conclusion, a buffet meal in a normal eating environment is a valid tool for assessing the effects of interventions on intakes.