College students experience new pressures and choices as they transition to independent living and can easily develop unhealthy eating habits, resulting in obesity and obesity-related chronic diseases in later life. This study aimed to test the hypothesis that nutrition literacy (NL) mediated the relationship between multi-level factors influencing healthy eating behavior identified from the social-ecological model and healthy eating behavior of college students. A four-part questionnaire was completed by 412 participants recruited from six different four-year universities in Taiwan (effective response rate = 85.8%). Data were analyzed using descriptive statistics, an independent samples t-test, hierarchical multiple regression, and mediation analysis. The results indicated that the students’ mean nutrition literacy score was 4.32 (SD = 0.78, range = 1–6). In the social-ecological framework, nutrition literacy significantly predicted healthy eating behavior (β = 0.28, p < 0.001; ΔF = 32.54, p < 0.001; ΔR² = 0.05) with control variables of background, intrapersonal, interpersonal, environmental, and macrosystem factors. Nutrition literacy mediated the effects of seven factors on healthy eating behavior across four levels. These findings suggested that strengthening influential multi-level factors associated with healthy eating behavior not only enhanced NL, but also improved individuals’ healthy eating behavior.

Background: As ChatGPT becomes a primary information source for college students, its performance in providing dietary advice is under scrutiny. This study assessed ChatGPT’s performance in providing nutritional guidance to college students. Methods: ChatGPT’s performance on dietary advice was evaluated by 30 experienced dietitians and assessed using an objective nutrition literacy (NL) test. The dietitians were recruited to assess the quality of ChatGPT’s dietary advice, including its NL achievement and response quality. Results: The results indicate that ChatGPT’s performance varies across scenarios and is suboptimal for achieving NL with full achievement rates from 7.50% to 37.56%. While the responses excelled in readability, they lacked understandability, practicality, and completeness. In the NL test, ChatGPT showed an 84.38% accuracy rate, surpassing the NL level of Taiwanese college students. The top concern among the dietitians, cited 52 times in 242 feedback entries, was that the “response information lacks thoroughness or rigor, leading to misunderstandings or misuse”. Despite the potential of ChatGPT as a supplementary educational tool, significant gaps must be addressed, especially in detailed dietary inquiries. Conclusion: This study highlights the need for improved AI educational approaches and suggests the potential for developing ChatGPT teaching guides or usage instructions to train college students and support dietitians.

Background: Advancements in digital technologies seek to promote health and access to services. However, people lacking abilities and confidence to use technology are likely to be left behind, leading to health disparities. In providing digital health services, health care providers need to be aware of users’ diverse electronic health (eHealth) literacy to address their particular needs and ensure equitable uptake and use of digital services. To understand such needs, an instrument that captures users’ knowledge, skills, trust, motivation, and experiences in relation to technology is required. The eHealth Literacy Questionnaire (eHLQ) is a multidimensional tool with 7 scales covering diverse dimensions of eHealth literacy. The tool was simultaneously developed in English and Danish using a grounded and validity-driven approach and has been shown to have strong psychometric properties. Objective: This study aims to translate and culturally adapt the eHLQ for application among Mandarin-speaking people with chronic diseases in Taiwan and then undertake a rigorous set of validity-testing procedures. Methods: The cross-cultural adaptation of the eHLQ included translation and evaluation of the translations. The measurement properties were assessed using classical test theory and item response theory (IRT) approaches. Content validity, known-group validity, and internal consistency were explored, as well as item characteristic curves (ICCs), item discrimination, and item location/difficulty. Results: The adapted version was reviewed, and a recommended forward translation was confirmed through consensus. The tool exhibited good content validity. A total of 420 people with 1 or more chronic diseases participated in a validity-testing survey. The eHLQ showed good internal consistency (Cronbach α=.75-.95). For known-group validity, all 7 eHLQ scales showed strong associations with education. Unidimensionality and local independence assumptions were met except for scale 2. IRT analysis showed that all items demonstrated good discrimination (range 0.27-12.15) and a good range of difficulty (range 0.59-1.67) except for 2 items in scale 7. Conclusions: Using a rigorous process, the eHLQ was translated from English into a culturally appropriate tool for use in the Mandarin language. Validity testing provided evidence of satisfactory-to-strong psychometric properties of the eHLQ. The 7 scales are likely to be useful research tools for evaluating digital health interventions and for informing the development of health technology products and interventions that equitably suit diverse users’ needs.