Healthy food diversity in an Italian population

Introduction: Diet diversity is often considered when assessing overall diet quality. It would be important to critically evaluate diet diversity in relation to recommendations of consumption of different food groups. The Healthy Food Diversity Index (HFD-I) by Drescher et al. (2007) combines diet diversity, assessed by the Berry Index (BI), and adherence to national guidelines. Objectives: We developed the HFD-I for an Italian population, to assess diet both diversity and adherence to CREA dietary guidelines. Methods: Data were derived from controls of an Italian case-control study on colorectal cancer (n = 4154), where food intake was assessed via a valid and reproducible food frequency questionnaire (FFQ) including 78 food items. Food frequencies were standardized to CREA portion sizes, considering a 2000kcal diet. We excluded 5 food items that did not contribute to the energy intake (e.g. coffee). Food items were categorized into 5 food groups (FGs) based on CREA guidelines: (FG1) cereals/potatoes, (FG2) fruits/vegetables, (FG3) meat/fish/eggs/legumes, (FG4) dairy, (FG5) discretionary foods; with further subgroup divisions. According to CREA guidelines, we assigned to FGs and their subgroups optimal proportions on the total diet. Nutritional equivalence within subgroups (e.g., milk/yogurt) was also considered. For each participant, diet diversity was assessed using the BI, which accounts for proportion of each food item within the total diet. Food Health Values (HVs) were obtained considering the consumption of each FFQ food item in relation to the corresponding optimal proportion. HFD-I was calculated as the product of the sum of HVs with the BI. Pearson correlations between FGs and subgroups and both BI and HFD-I were computed, adjusting for multiple comparisons with a false discovery rate method. Results: BI was inversely correlated with FG1 (-0.45) and with FG5 (-0.32). HFD-I was strongly directly correlated with FG2 (0.76) and inversely correlated with FG1 (-0.25) and FG5 (-0.21). For subgroups, BI was correlated with bread substitutes (-0.46) and soft drinks/juices (-0.39). HFD-I was directly correlated with fruit (0.75), vegetables (0.21), legumes (0.22) and inversely with cold cuts (-0.24). For other FGs and subgroups, for both BI and HFD-I absolute values of correlations were <0.2. Conclusions: The HFD-I provides a comprehensive measure of diet quality, stressing the importance of increased fruit and vegetable consumption