Background: Training of machine learning algorithms on dish images collected in other countries requires possible sources of systematic discrepancies, including country-specific food composition databases (FCDBs), to be tackled. The US Nutrition5k project provides for ~5000 dish images and related dish- and ingredient-level information on mass, energy, and macronutrients from the US FCDB. The aim of this study is to (1) identify challenges/solutions in linking the nutritional composition of Italian foods with food images from Nutrition5k and (2) assess potential differences in nutrient content estimated across the Italian and US FCDBs and their determinants. Methods: After food matching, expert data curation, and handling of missing values, dish-level ingredients from Nutrition5k were integrated with the Italian-FCDB-specific nutritional composition (86 components); dish-specific nutrient content was calculated by summing the corresponding ingredient-specific nutritional values. Measures of agreement/difference were calculated between Italian- and US-FCDB-specific content of energy and macronutrients. Potential determinants of identified differences were investigated with multiple robust regression models. Results: Dishes showed a median mass of 145 g and included three ingredients in median. Energy, proteins, fats, and carbohydrates showed moderate-to-strong agreement between Italian- and US-FCDB-specific content; carbohydrates showed the worst performance, with the Italian FCDB providing smaller median values (median raw difference between the Italian and US FCDBs: −2.10 g). Regression models on dishes suggested a role for mass, number of ingredients, and presence of recreated recipes, alone or jointly with differential use of raw/cooked ingredients across the two FCDBs. Conclusions: In the era of machine learning approaches for food image recognition, manual data curation in the alignment of FCDBs is worth the effort.
16 October 2025 in Nutrition, Metabolism and Cardiovascular Diseases
Background: Hospital canteens provide an effective setting to improve users' dietary habits. The study evaluates users' food choices—considering nutritional composition and environmental impact -after an educational and environmental intervention, comparing the results with pre-intervention choices. Methods: A pre-post intervention study was conducted in three hospital canteens (C1, C2, C3) in northeastern Italy, during two index weeks in September 2022 (T0) and 2023 (T1). Canteen users were recruited before and after an intervention consisting of posters on healthy eating, descriptive norm messages, and environmental changes regarding fruit and vegetables. Photos of participants' lunch trays were collected, and food choices were analyzed for nutritional composition and sustainability. Descriptive statistics and appropriate non-parametric and chi-square tests were used to compare outcomes between assessment periods and across canteens. Results: 2,851 trays were analyzed: 1,227 at T0 (798 in C1, 228 in C2 and 201 in C3) and 1,624 at T1 (1,005 in C1, 348 in C2, 271 in C3). In C1 and C3, there was an increase in median energy (+30 kcal; +135 kcal) compared to pre-intervention meals, while in C2 there was a decrease (−118 kcal). Despite a slight improvement in macronutrient composition, at T1 meals in all canteens were still high in lipids (30%E; 39%E; 35%E) and low in carbohydrates (44%E; 39%E; 41%E). The fiber value fell within the reference range only in C1 and C3. The median carbon (CF) and water (WF) footprints of meals in all canteens remained high: at T1 CF ranged from 966 g CO 2 eq. to 1,227 g CO 2 eq. and WF from 1,025 L H 2 O to 1,207 L H 2 O. Conclusion: The intervention has led to partial improvements in food choices. To achieve more significant results, it may be necessary to implement a parallel intervention on food offer.
24 September 2026 in European Journal of Nutrition
Purpose: To assess knowledge of lactose intolerance (LI) in Italian adults, identifying major knowledge gaps and factors associated with awareness. Methods: A cross-sectional study was conducted between December 2024 and March 2025 among Italian adults (≥ 18 years). A newly developed 39-item online survey assessed knowledge of LI across four modules (definition, symptoms, diagnosis, and management). The knowledge score (range: − 39; +39) was calculated and participants were classified into three knowledge levels based on the tertile distribution of scores: “low knowledge”, “intermediate knowledge”, and “high knowledge”’. Univariable and multivariable ordinal logistic regression were carried out to investigate the association between respondents’ characteristics and LI knowledge. Results: A total of 2025 respondents (85.4% female) were included in the analysis. The mean knowledge score was 15.4 ± 7.6. Overall, 36.6% of respondents showed low, 32.2% intermediate, and 31.2% high knowledge. Common misconceptions concern the belief that trace amounts of lactose (e.g., in lactose-free products or medications) can trigger symptoms and that a lactose-free diet does not pose a risk for nutritional deficiencies. Questions on unvalidated diagnostic methods elicited many “I don’t know” responses, reflecting uncertainty about their reliability. Female gender, confirmed LI diagnosis, higher education and nutrition training were associated with higher knowledge, with the latter showing the strongest association (Odds Ratio 10.8, 95% CI: 8.30–14.2). Conclusion: These findings highlight persistent misconceptions about LI and underscore the need for targeted educational strategies to improve public understanding and reduce inappropriate diagnostic and dietary practices.