Background
Dietary patterns are associated with health outcomes and environment sustainability, having socioeconomic drivers. This ecological study aims to identify dietary patterns in food availability data at the global level using multivariate statistical methodology, to associate the identified dietary patterns with socioeconomic data and to analyze the adequacy of the applied multivariate statistical methods for this purpose.
Methods
Principal Component Analysis was applied to median values of times series of food availability data of 172 UN registered countries available at FAOSTAT database in Food Balance Sheets section in a sectional manner, after calculating median values of time series for each food group and country. Principal Components were associated with socioeconomic data available from the World Bank database. Sensitivity analyses were realized to verify the stability of dietary patterns through five different times.
Results
Five principal components were identified in the median values of each time series, each characterizing a possible dietary pattern. The first one, a westernized dietary pattern, was composed of energy-dense and processed foods, foods of animal origin, alcoholic beverages, but also, albeit less, by vegetables, fruits and nuts, being associated with income, urbanization and trade liberalization. This westernized pattern was characterized more animal origin and processed foods, such as vegetable oils, alcoholic beverages and stimulants yet preserving unprocessed and regional foods. The other dietary patterns were three agricultural patterns characterized more by regional foods, especially starchy staples, and one coastal dietary pattern composed of fish and seafoods, being associated with GINI index, poverty, and female labor force. Sensitivity analyses demonstrated the stability of dietary patterns.
Conclusions
Principal Component Analysis was adequate to identify dietary patterns in food availability data. A westernized dietary pattern was identified, being associated with income, urbanization, and trade liberalization. This association did not occur for the remain of the dietary patterns identified, these being less driven by economic development.