The Power of Storytelling in the Digital Age: A Multimodal Analysis of Persuasive Strategies and User Engagement in Tour Guide Influencer Content

The use of web-based search and social media can help identify epidemics, potentially earlier than clinical methods or even potentially identifying unreported outbreaks. Monitoring for eye-related epidemics, such as conjunctivitis outbreaks, can facilitate early public health intervention to reduce transmission and ocular comorbidities. However, monitoring social media content for conjunctivitis outbreaks is costly and laborious. Large language models (LLMs) could overcome these barriers by assessing the likelihood that real-world outbreaks are being described. However, public health actions for likely outbreaks could benefit more by knowing additional epidemiological characteristics, such as outbreak type, size, and severity. We aimed to assess whether and how well LLMs can classify epidemiological features from social media posts beyond conjunctivitis outbreak probability, including outbreak type, size, severity, etiology, and community setting. We used a validation framework comparing LLM classifications to those of other LLMs and human experts. We wrote code to generate synthetic conjunctivitis outbreak social media posts, embedded with specific preclassified epidemiological features to simulate various infectious eye disease outbreak and control scenarios. We used these posts to develop effective LLM prompts and test the capabilities of multiple LLMs. For top-performing LLMs, we gauged their practical utility in real-world epidemiological surveillance by comparing their assessments of Twitter/X, forum, and YouTube conjunctivitis posts. Finally, human raters also classified the posts, and we compared their classifications to those of a leading LLM for validation. Comparisons entailed correlation or sensitivity and specificity statistics. We assessed 7 LLMs for effectively classifying epidemiological data from 1152 synthetic posts, 370 Twitter/X posts, 290 forum posts, and 956 YouTube posts. Despite some discrepancies, the LLMs demonstrated a reliable capacity for nuanced epidemiological analysis across various data sources and compared to humans or between LLMs. Notably, GPT-4 and Mixtral 8x22b exhibited high performance, predicting conjunctivitis outbreak characteristics such as probability (GPT-4: correlation=0.73), size (Mixtral 8x22b: correlation=0.82), and type (infectious, allergic, or environmentally caused); however, there were notable exceptions. Assessing synthetic and real-world posts for etiological factors, infectious eye disease specialist validations revealed that GPT-4 had high specificity (0.83-1.00) but variable sensitivity (0.32-0.71). Interrater reliability analyses showed that LLM-expert agreement exceeded expert-expert agreement for severity assessment (intraclass correlation coefficient=0.69 vs 0.38), while agreement varied by condition type (κ=0.37-0.94). This investigation into the potential of LLMs for public health infoveillance suggests effectiveness in classifying key epidemiological characteristics from social media content about conjunctivitis outbreaks. Future studies should further explore LLMs' potential to support public health monitoring through the automated assessment and classification of potential infectious eye disease or other outbreaks. Their optimal role may be to act as a first line of documentation, alerting public health organizations for the follow-up of LLM-detected and -classified small, early outbreaks, with a focus on the most severe ones.

BACKGROUND The use of web-based search and social media can help identify epidemics, potentially earlier than clinical methods or even potentially identifying unreported outbreaks. Monitoring for eye-related epidemics, such as conjunctivitis outbreaks, can facilitate early public health intervention to reduce transmission and ocular comorbidities. However, monitoring social media content for conjunctivitis outbreaks is costly and laborious. Large language models (LLMs) could overcome these barriers by assessing the likelihood that real-world outbreaks are being described. However, public health actions for likely outbreaks could benefit more by knowing additional epidemiological characteristics, such as outbreak type, size, and severity. OBJECTIVE We aimed to assess whether and how well LLMs can classify epidemiological features from social media posts beyond conjunctivitis outbreak probability, including outbreak type, size, severity, etiology, and community setting. We used a validation framework comparing LLM classifications to those of other LLMs and human experts. METHODS We wrote code to generate synthetic conjunctivitis outbreak social media posts, embedded with specific preclassified epidemiological features to simulate various infectious eye disease outbreak and control scenarios. We used these posts to develop effective LLM prompts and test the capabilities of multiple LLMs. For top-performing LLMs, we gauged their practical utility in real-world epidemiological surveillance by comparing their assessments of Twitter/X, forum, and YouTube conjunctivitis posts. Finally, human raters also classified the posts, and we compared their classifications to those of a leading LLM for validation. Comparisons entailed correlation or sensitivity and specificity statistics. RESULTS We assessed 7 LLMs for effectively classifying epidemiological data from 1152 synthetic posts, 370 Twitter/X posts, 290 forum posts, and 956 YouTube posts. Despite some discrepancies, the LLMs demonstrated a reliable capacity for nuanced epidemiological analysis across various data sources and compared to humans or between LLMs. Notably, GPT-4 and Mixtral 8x22b exhibited high performance, predicting conjunctivitis outbreak characteristics such as probability (GPT-4: correlation=0.73), size (Mixtral 8x22b: correlation=0.82), and type (infectious, allergic, or environmentally caused); however, there were notable exceptions. Assessing synthetic and real-world posts for etiological factors, infectious eye disease specialist validations revealed that GPT-4 had high specificity (0.83-1.00) but variable sensitivity (0.32-0.71). Interrater reliability analyses showed that LLM-expert agreement exceeded expert-expert agreement for severity assessment (intraclass correlation coefficient=0.69 vs 0.38), while agreement varied by condition type (κ=0.37-0.94). CONCLUSIONS This investigation into the potential of LLMs for public health infoveillance suggests effectiveness in classifying key epidemiological characteristics from social media content about conjunctivitis outbreaks. Future studies should further explore LLMs’ potential to support public health monitoring through the automated assessment and classification of potential infectious eye disease or other outbreaks. Their optimal role may be to act as a first line of documentation, alerting public health organizations for the follow-up of LLM-detected and -classified small, early outbreaks, with a focus on the most severe ones.

Driven by the need to enhance user traffic on social media (SM) platforms for increasing their advertising revenues, SM platforms are experimenting with new content creation features. However, it is unclear if such initiatives are also beneficial for SM profile owners such as influencers, who are the prime content creators on the SM platforms who use SM posts to build their influence within their network of followers. Our study investigates the effect of introducing one such new SM feature: the “story” on the creation and consumption of SM posts. Leveraging social penetration theory, we hypothesize the influence of introducing story feature on (1) the frequency of SM post creation by profile owners and (2) the extent of follower engagement with SM posts. Employing a quasi-experimental design, we find that the introduction of the story feature reduces the frequency of SM post creation, but the enhanced self-disclosure through the story feature increases follower engagement with the SM posts. However, these effects are moderated by the situating culture of the SM communities: while low-power-distance cultures value profile owners’ self-disclosure, high-power-distance cultures exhibit a mixed influence. Advancing literature on social penetration theory and SM user engagement, our study demonstrates that new self-disclosive SM content creation features do not necessarily benefit all the concerned stakeholders and that the effectiveness of such features might vary from one community to another. Hence, the intended impact of introducing new SM features needs to be carefully evaluated by SM platforms in a holistic manner.

The prevalent trend of using memes and emoji on social media has changed the landscape of communication across online platforms. Memes and emoji, despite being a popular online language, have received limited attention in the social media marketing literature. This study aims to investigate how memes and emoji can enhance customer engagement when incorporated into a brand’s social media communications. In the first phase of the study, netnography was used to assess the frequency of emoji across three social media platforms: Facebook, Instagram, and X. A total of 2,800 social media posts and comments were analyzed to categorize the brands that use emoji more frequently. The results identified that the thumbs-up , heart , and smile were the most commonly used emoji among brand categories such as fashion and apparel, food, makeup, entertainment, and transportation. Netnography was followed by a thorough content analysis to explore customer engagement, in the form of reactions, comments, and shares, with the brands’ social media posts, which contained memes and emoji. In the second phase of the study, 20 in-depth, semi-structured interviews were conducted to validate the findings from the netnography and content analysis. The results suggest that the social media audience favors thumbs-up, and smile, over the heart , as they are perceived to convey positivity and warmth. Furthermore, emoji and memes are more likely to be accepted by the youth, especially when the brands’ communication resonates with their personality. Moreover, emoji can be used to express concern during complaint handling, resulting in a friendlier relationship with customers. Keywords: Meme Marketing, Emoji Marketing, Netnography, Customer Engagement https://zenodo.org/records/15857329

Solace in Social Media: Women Unite Under COVID-19

Ethical standards for cardiothoracic surgeons' participation in social media

Health risks due to preventable infections such as human papillomavirus (HPV) are exacerbated by persistent vaccine hesitancy. Due to limited sample sizes and the time needed to roll out, traditional methodologies like surveys and interviews offer restricted insights into quickly evolving vaccine concerns. Social media platforms can serve as fertile ground for monitoring vaccine-related conversations and detecting emerging concerns in a scalable and dynamic manner. Using state-of-the-art large language models, we propose a minimally supervised end-to-end approach to identify concerns against HPV vaccination from social media posts. We detect and characterize the concerns against HPV vaccination pre- and post-2020 to understand the evolution of HPV vaccine discourse. Upon analyzing 653 k HPV-related post-2020 tweets, adverse effects, personal anecdotes, and vaccine mandates emerged as the dominant themes. Compared to pre-2020, there is a shift towards personal anecdotes of vaccine injury with a growing call for parental consent and transparency. The proposed approach provides an end-to-end system, i.e. given a collection of tweets, a list of prevalent concerns is returned, providing critical insights for crafting targeted interventions, debunking messages, and informing public health campaigns.

The high prevalence of noncommunicable diseases and the growing importance of social media have prompted health care professionals (HCPs) to use social media to deliver health information aimed at reducing lifestyle risk factors. Previous studies have acknowledged that the identification of elements that influence user engagement metrics could help HCPs in creating engaging posts toward effective health promotion on social media. Nevertheless, few studies have attempted to comprehensively identify a list of elements in social media posts that could influence user engagement metrics. This systematic review aimed to identify elements influencing user engagement metrics in social media posts by HCPs aimed to reduce lifestyle risk factors. Relevant studies in English, published between January 2006 and June 2023 were identified from MEDLINE or OVID, Scopus, Web of Science, and CINAHL databases. Included studies were those that examined social media posts by HCPs aimed at reducing the 4 key lifestyle risk factors. Additionally, the studies also outlined elements in social media posts that influenced user engagement metrics. The titles, abstracts, and full papers were screened and reviewed for eligibility. Following data extraction, narrative synthesis was performed. All investigated elements in the included studies were categorized. The elements in social media posts that influenced user engagement metrics were identified. A total of 19 studies were included in this review. Investigated elements were grouped into 9 categories, with 35 elements found to influence user engagement. The 3 predominant categories of elements influencing user engagement were communication using supportive or emotive elements, communication aimed toward behavioral changes, and the appearance of posts. In contrast, the source of post content, social media platform, and timing of post had less than 3 studies with elements influencing user engagement. Findings demonstrated that supportive or emotive communication toward behavioral changes and post appearance could increase postlevel interactions, indicating a favorable response from the users toward posts made by HCPs. As social media continues to evolve, these elements should be constantly evaluated through further research.

BackgroundApproximately 80,000 primary brain tumors are diagnosed annually. Social media provides a source of information and support for patients diagnosed with brain tumors; however, use of this forum for dissemination of information about brain tumors has not been evaluated. The objective of this study was to evaluate social media utilization and content related to brain tumors with an emphasis on patients’ trends in usage.MethodsSocial media platforms were systematically evaluated using two search methods: systematic manual inquiry and a keyword-based social media tracker. The search terms included brain tumor, glioblastoma, glioma, and glioblastoma multiforme. Social media content (which includes Facebook pages and groups, YouTube videos, and Twitter or Instagram accounts) and posts were assessed for activity (as quantified by views of posts) and analyzed using a categorization framework.ResultsThe manual and keyword searches identified 946 sources of social media content, with a total count of 7,184,846 points of engagement. Social media platforms had significant variations in content type. YouTube was the largest social media platform for sharing content related to brain tumors overall, with an emphasis on surgical videos and documented patient experiences. Facebook accounted for the majority of patient-to-patient support, and Twitter was the most common platform for scientific dissemination. Overall social media content was mostly focused on treatment overviews and patient experience. When evaluated by search term, most social media posts by the “brain tumor” community shared illness narratives, and searches specific to “glioma” and “glioblastoma” demonstrated a higher proportion of educational and treatment posts.ConclusionsThis study presents novel observations of the characteristics of social media utilization for the online brain tumor community. A robust patient community exists online, with an emphasis on sharing personal narratives, treatment information, patient-to-patient support, treatment options, and fundraising events. This study provides a window to the role of social media utilization by patients, their families, and health professionals. These findings demonstrate the different roles of Facebook, YouTube, and Twitter in the rapidly changing era of social media and its relationship with neurosurgery and neuro-oncology.

Reply to “Dermatologists in social media: A study on top influencers, posts, and user engagement”: Dermatologist influencers on TikTok

How Can <i>IJDS</i> Authors, Reviewers, and Editors Use (and Misuse) Generative AI?

In the digital age, social media has been a go-to platform for stress-related discussions, yielding valuable data to advance the understanding and detection of stress. Swift identification of stress indicators in these online conversations is essential in enabling immediate support and helping to avert subsequent severe mental and physical health issues, especially during global crises such as pandemics and conflicts. Detecting stress in social media posts automatically poses a formidable challenge. While techniques such as supervised Pretrained Language Models (PLMs) and zero-shot Large Language Models (LLMs) based classifiers have demonstrated significant performance, they exhibit limitations, especially on platforms like Reddit. For example, on Reddit, users tend to write lengthy, expressive posts, which causes these methods to often fail to consider the entire context, leading to incomplete or inaccurate assessments of a user's mental health or stress status. To overcome these limitations, we present a new approach to identifying and classifying stress-related discourse on social media. Our approach involves analyzing condensed versions of user posts, such as user-provided summaries or the "Too Long Didn’t Read" (TLDR) portion of the original post. We question whether these abridged texts can yield a more accurate classification of stress. In this paper, we make the following contributions. First, we investigate the relationship between the performance of the model's perceived textual context and the length of social media posts. Second, we present a novel approach to use the summarized texts for stress detection. We experiment with different classifiers to evaluate their performance on stress detection accuracy using summarized versus full-length posts. Furthermore, by examining the emotional and linguistic features of the original posts and their summaries, we suggest improvements to current state-of-the-art LLM-based stress classifier prompts, thereby enhancing stress detection capabilities. Finally, when user summaries are absent, we synthetically generate meaningful user post summaries by incorporating the power of LLMs. Our results show that the stress detection performance deteriorates for longer posts, and utilizing the TLDR and summaries improves classification outcomes. We also provide augmented datasets containing human and AI-generated summaries for future research in stress detection on social media.

Alts and Automediality: Compartmentalising the Self through Multiple Social Media Profiles

BackgroundHazardous drinking among college students persists, despite ongoing university alcohol education and alcohol intervention programs. College students often post comments or pictures of drinking episodes on social media platforms.ObjectiveThis study aimed to understand one university’s student attitudes toward alcohol use by examining student posts about drinking on social media platforms and to identify opportunities to reduce alcohol-related harm and inform novel alcohol interventions.MethodsWe analyzed social media posts from 7 social media platforms using qualitative inductive coding based on grounded theory to identify the contexts of student drinking and the attitudes and behaviors of students and peers during drinking episodes. We reviewed publicly available social media posts that referenced alcohol, collaborating with undergraduate students to select their most used platforms and develop locally relevant search terms; all posts in our data set were generated by students associated with a specific university. From the codes, we derived themes about student culture regarding alcohol use.ResultsIn total, 1151 social media posts were included in this study. These included 809 Twitter tweets, 113 Instagram posts, 100 Greekrank posts, 64 Reddit posts, 34 College Confidential posts, 23 Facebook posts, and 8 YouTube posts. Posts included both implicit and explicit portrayals of alcohol use. Across all types of posts reviewed, positive drinking attitudes were most common, followed by negative and then neutral attitudes, but valence varied by platform. Posts that portrayed drinking positively received positive peer feedback and indicate that drinking is viewed by students as an essential and positive part of university student culture.ConclusionsSocial media provide a real-time picture of students’ behavior during their own and others’ heavy drinking. Posts portray heavy drinking as a normal part of student culture, reinforced by peers’ positive feedback on posts. Interventions for college drinking should help students manage alcohol intake in real time, provide safety information during alcohol use episodes, and raise student awareness of web-based privacy concerns and reputation management. Additional interventions for students, alumni, and parents are needed to address positive attitudes about and traditions of drinking.

Persuasion strategies of misinformation-containing posts in the social media