RESUMO
BACKGROUND: Automation of dietary assessments allows participant behaviour to be captured by video observation. It also allows clinicians to identify areas that effect reporting accuracy. The present observational study describes the differences in behaviour according to the type of foods selected by participants using a dietary assessment website encompassing diet history methodology. METHODS: Eleven free-living adults with type 2 diabetes mellitus were video recorded while using a dietary assessment website in a laboratory setting. Significant observable events were identified using TRANSANA software and matched for time of occurrence with grouped food item selections from the website. The frequency and proportion of behaviour types per food group were calculated and trends of associations between food groups and observed behaviour type were determined using weighted chi-square analyses. RESULTS: Sixteen categories of observed behaviour type were constructed from 7724 significant observable events. The food groups with the highest percentage of observed behaviour types were savoury sauces (7.66%), takeaways (6.47%), sugary foods (6.27%), fats and oils (5.93%), and alcoholic beverages (5.93%), which are all socially undesirable food groups. Self-touching of the face, head movement, postural movement and movement in the chair were observed significantly more often than other observed behaviour types across all food groups. CONCLUSIONS: Behaviours related to the reporting of foods during a dietary assessment may have significant implications for dietary practice.