Adapting eHEALS for Social Media and Generative AI: Cross-Sectional Study among Older Patients (Preprint)

Le résumé fourni par la source

BACKGROUND The eHealth Literacy Scale (eHEALS) is the most common eHealth literacy (eHL) measurement instrument in empirical research. Published in 2006, the eHEALS does not explicitly mention newer use cases such as social media and generative AI (genAI). Nevertheless, skills addressed related to seeking and evaluating electronic health information may be relevant for the use of social media and genAI. However, participants may not consider social media and genAI when answering the eHEALS, and mean scores of eHEALS versions adapted to social media and genAI contexts may differ, although positive correlations may exist. OBJECTIVE This study aims to investigate whether (1) patients consider social media and genAI when answering the German eHEALS (GR-eHEALS), (2) there are correlations between the GR-eHEALS and 2 versions adapted to social media (GR-eHEALS social media) and genAI (GR-eHEALS genAI), as well as between these adapted versions, and (3) significant differences exist between the 3 versions and each of the pairs. METHODS A cross-sectional study was conducted based on data from a survey of patients aged ≥67 years at an outpatient facility in northern Germany and an inpatient and outpatient facility in southern Germany. eHL was assessed using the German eHEALS (GR-eHEALS) and two adapted versions for the social media and genAI contexts. Internal consistency was assessed using Cronbach α and McDonald ω. Two-tailed Pearson correlations were calculated to investigate the relationships between the GR-eHEALS versions. Repeated-measures ANOVAs were performed to examine differences between the 3 versions, and Bonferroni post hoc analyses for pairwise comparisons. RESULTS A total of 202 participants aged between 67 and 96 years (mean 75.73, SD 6.65) were included. Internal consistency for the whole scales was high across the 3 GR-eHEALS versions (Cronbach α=0.93–0.96, McDonald ω=0.93–0.96). Mean scores for GR-eHEALS (mean 26.43, SD 9.38) were higher than for GR-eHEALS genAI (mean 17.10, SD 9.59) and GR-eHEALS social media (mean 15.07, SD 8.00). Few participants considered genAI (19.00%–20.79%) and social media (8.42%–12.00%) when answering the GR-eHEALS items. Significant positive correlations were found between each pair of the 3 GR-eHEALS versions, and significant mean score differences were observed between the GR-eHEALS and the GR-eHEALS social media, as well as the GR-eHEALS and the GR-eHEALS genAI. Mean scores were also significantly different between the GR-eHEALS social media and GR-eHEALS genAI overall, as well as for the information seeking subscale, but not for the information appraisal subscale. CONCLUSIONS To our knowledge, this is the first study examining the association between the GR-eHEALS and 2 versions adapted for social media and genAI contexts. The findings indicate that the GR-eHEALS provides only limited insights into eHL related to social media and genAI. Therefore, both the GR-eHEALS social media and genAI may be used as context-specific alternatives.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.