Identifying dietary patterns across age, educational level and physical activity level in a cross-sectional study: the Tromsø Study 2015 - 2016

Background A healthy diet can decrease the risk of several lifestyle diseases. From studying the health effects of single foods, research now focuses on examining complete diets and dietary patterns reflecting the combined intake of different foods. The main goals of the current study were to identify dietary patterns and then investigate how these differ in terms of sex, age, educational level and physical activity level (PAL) in a general Nordic population. Methods We used data from the seventh survey of the population-based Tromsø Study in Norway, conducted in 2015-2016. The study included 21,083 participants aged 40−99 years, of which 72% completed a comprehensive food frequency questionnaire (FFQ). After exclusion, the study sample included 10,899 participants with valid FFQ data. First, to cluster food variables, the participants were partitioned in homogeneous cohorts according to sex, age, educational level and PAL. Non-overlapping diet groups were then identified using repeated hierarchical cluster analysis on the food variables. Second, average standardized diet intake scores were calculated for all individuals for each diet group. The individual diet (intake) scores were then modelled in terms of age, education and PAL using regression models. Differences in diet scores according to education and PAL were investigated by pairwise hypothesis tests, controlling the nominal significance level using Tukey’s method. Results The cluster analysis revealed three dietary patterns, here named the Meat and Sweets diet, the Traditional diet, and the Plant-based- and Tea diet. Women had a lower intake of the Traditional diet and a higher preference for the Plant-based- and Tea diet compared to men. Preference for the Meat and Sweets diet and Traditional diet showed significant negative and positive trends as function of age, respectively. Adjusting for age, the group having high education and high PAL compared favourably with the group having low education and low PAL, having a significant lower intake of the Meat and Sweets and the Traditional diets and a significant higher intake of the Plant-based- and Tea diet. Conclusions Three dietary patterns (Meat and Sweets, Traditional, and Plant-based- and Tea) were found by repeated clustering of randomly sampled homogeneous cohorts of individuals. Diet preferences depended significantly on sex, age, education and PAL, showing a more unhealthy dietary pattern with lower age, low education and low PAL