Development and validation of a photograph-based instrument to assess nutrition literacy: the NUTLY project.

Most instruments measuring nutrition literacy evaluate theoretical knowledge, not necessarily reflecting skills relevant to food choices. We aimed to develop and validate a photograph-based instrument to assess nutrition literacy (NUTLY) among adults in Portugal. NUTLY assesses the ability to distinguish foods with different nutritional profiles; from each of several combinations of three photographs (two foods with similar contents and one with higher content) participants are asked to identify the food with the highest energy/sodium content. The NUTLY version with 79 combinations, obtained after experts/lay people evaluations, was applied to a sample representing different age, gender and education groups (n=329). Dimensionality was evaluated through latent trait models. Combinations with negative or with positive small factor loadings were excluded after critical assessment. Internal consistency was measured using Cronbach's alpha and construct validity by comparing NUTLY scores with those obtained in the Medical Term Recognition Test and the Newest Vital Sign (NVS), and across education and training in nutrition/health groups. The cut-off to distinguish adequate/inadequate nutrition literacy was defined through ROC analysis using the Youden index criterion, after performing a Latent class analysis which identified a two-class model to have the best goodness of fit. Test-retest reliability was assessed after one month (n=158). The final NUTLY scale was unidimensional and included 48 combinations (energy: 33; sodium: 15; α=0.74). Mean scores (±standard deviation) were highest among nutritionists (39.9±4.4), followed by health professionals (38.5±4.1) and declined with decreasing education (p<0.001). Those with adequate nutrition literacy according to NVS showed higher NUTLY scores (37.9±4.3 vs. 33.9 ±6.9, p<0.001). Adequate nutrition literacy was defined as a NUTLY score≥35 (sensitivity: 89.3%; specificity: 93.7%). Test-retest reliability was high (ICC=0.77). NUTLY is a valid and reliable nutrition literacy measurement tool.