Food intake and usual dietary intake are among the key determinants of health to be assessed in medical research and important confounding factors to be accounted for in clinical studies. Although various methods are available for gathering dietary data, those based on innovative technologies are particularly promising. With combined cost-effectiveness and ease of use, it is safe to assume that mobile technologies can now optimize tracking of eating occasions and dietary behaviors. Yet, choosing a dietary assessment tool that meets research objectives and data quality standards remains challenging. In this paper, we describe the purposes of collecting dietary data in medical research and outline the main considerations for using mobile dietary assessment tools based on participant and researcher expectations.

Objectives Eating habits have been shown to impact the gut microbiome. Here we aimed to define several types of dietary patterns in a U.S. adult cohort and test their associations with the gut microbiome. Methods Using supervised and unsupervised approaches, we built dietary patterns based on a food frequency questionnaire of the American Gut Project database. Focusing on 1800 adult participants living in the United States, we defined patterns as partitions (groups of participants) or factors (combinations of food variables) driven by specific dietary criteria: fibers, proteins, Healthy Eating Index (HEI 2010), food items, food groups and micronutrients. We then associated these patterns with 16S gut microbiome data for 744 participants, excluding those reporting antibiotic intake in the last year or specific diseases. Analyses were adjusted for age, sex and BMI. Results Compared to individual features like fibers and proteins, or to factors representing reduced numbers of features, five unsupervised partitions based on food groups were best associated with gut microbiome beta-diversity. Two partitions presented a lower consumption of animal products, with one being almost completely exclusive and the other, close to a flexitarian diet, presenting the best diet quality as measured by HEI. A third one consisted mostly of participants under low carbohydrate diets with nearly no consumption of starchy foods or sweet products. Finally, the last two partitions presented Western-like diets with increased consumption of mixed dishes, sweet products and refined cereals, one of them being more diverse with increased nuts and whole cereals. Gut microbiome alpha-diversity was slightly increased in the flexitarian partition compared to the most westernized one. Strikingly, the low carbohydrate partition was associated with low levels of the Bifidobacterium genus. Conclusions We showed in a U.S. adult cohort that a global diet may be more associated with gut microbiome variations than individual features like fibers or proteins. Five diet partitions were identified and their specific associations with gut microbiome were studied. These results confirm the importance to consider diet as a whole when studying gut microbiota diversity. Funding Sources Danone Research.

UNSTRUCTURED Food intake and usual dietary intake are among the key determinants of health to be assessed in medical research and important confounding factors to be accounted for in clinical studies. Although various methods are available for gathering dietary data, those based on innovative technologies are particularly promising. With combined cost-effectiveness and ease of use, it is safe to assume that mobile technologies can now optimize tracking of eating occasions and dietary behaviors. Yet, choosing a dietary assessment tool that meets research objectives and data quality standards remains challenging. In this paper, we describe the purposes of collecting dietary data in medical research and outline the main considerations for using mobile dietary assessment tools based on participant and researcher expectations.