Tracking population dietary habits is notoriously difficult, from cohort recruitment to the patchy recollections of what someone ate 24 hours ago. A recent article in Nature Communications approached diet studies via a freely available smartphone app, allowing a large cohort to be assessed with minimal commitment from the participants.
Data from over a million app users, who added on average nine entries to their digital food record each day for an average 197 days, was matched up with demographic and location data to understand the consumption habits of a US cohort.
Their results matched existing knowledge on food environments and dietary habits: high income, higher education, high supermarket access and low fast-food access (the latter two determined by location), all correlated with lower BMI, higher fruit and vegetable consumption, and lower fast-food consumption. One exception was a slight association between high income and high BMI.
The authors also matched their location data to the predominant ethnic group, which was possible due to the zip code level resolution of the data. Again, these results reinforced existing data on the prevalence of consumption of specific foods, and the prevalence of obesity, but across a broader area than previously possible.
This paper shows the power of repurposing existing digitalised data for nutrition research. Such large, long-term, detailed sampling of the US cohort would have been extremely challenging without the availability of an already popular app. Moreover, the privacy of individuals was protected, and the app developers donated the data from the research, facilitating a more refined understanding of their nutrition.