Background: Youth e-cigarette use rose sharply between 2013 and 2024 in the United States, prompting widespread prevention campaigns at national, state, and local levels. However, many campaigns encountered online opposition, sometimes leading to message distortion or campaign withdrawal. While previous studies have examined individual campaigns, little is known about how oppositional dynamics differ across social media platforms with distinct architectures. Objective: This study aimed to conduct a retrospective, cross-platform surveillance study of oppositional responses to US youth e-cigarette prevention campaigns, comparing tactics, themes, and engagement on Twitter (now X) and TikTok to inform platform-specific public health strategies. Methods: We collected Twitter (2014‐2020) and TikTok (2020‐2023) posts related to major US e-cigarette prevention campaigns using 15 campaign-specific hashtags and 4 verified prevention campaign handles. We included public, English-language posts from geographic regions allowed by the platforms. Machine learning classification and human coding were used to detect oppositional content, characterize narrative frames, and classify user types. Engagement was assessed using post-level metrics, including likes, shares, comments, and retweets. We analyzed message prevalence, engagement patterns, and oppositional themes. Results: On Twitter, opposition comprised 26.8% (83,074/310,207) of campaign-related posts overall but dominated certain campaigns (eg, : 6052/6113, 99%). A small cluster of advocacy and commercial accounts generated 57.8% of opposition retweets. Dominant narratives included questioning the credibility of health authorities, claims that prevention advertisements backfired, vaping rights, and product promotion. In contrast, TikTok opposition constituted only 3.5% (108/3127, 95% CI 3.1%-3.9%) of posts and was characterized by humor (71/108, 65.7%), mockery (48/108, 44.4%), and ironic portrayals of vaping (30/108, 27.8%). Individual creators comprised 76.1% (153/201) of accounts sharing prevention posts, and opposition videos used the visibility-boosting hashtag #fyp significantly more than prevention posts (51.9% vs 32.2%; <.001). Despite inconsistent hashtag use, prevention posts achieved higher average engagement than oppositional content. Conclusions: This novel cross-platform, multicampaign analysis of opposition responses to e-cigarette prevention campaigns revealed how opposition reflects distinct platform architectures. Twitter opposition was highly coordinated and amplified by commercial and advocacy accounts, especially during regional campaigns. TikTok opposition was decentralized and humor-driven, aligning with the platform’s entertainment-oriented algorithm. The findings strengthen health communication by introducing a framework for evaluating platform-specific vulnerabilities and informing evidence-based campaign design. On Twitter, effective countermeasures may require real-time monitoring of social media discourse to support rapid responses to coordinated opposition. On TikTok, leveraging creator partnerships and remix-friendly content may help public health messages compete with entertainment-dominated discourse. Consistent hashtag use can strengthen engagement by minimizing the fragmentation of content visibility, and credible health sources should increasingly reinforce prevention narratives on both platforms. Greater platform accountability and transparency are needed to ensure that prevention content is not systematically deprioritized by algorithms relative to commercial promotion.
Depression subtype classification from social media posts: few-shot prompting vs. fine-tuning of large language models
BackgroundSocial media provides timely proxy signals of mental health, but reliable tweet-level classification of depression subtypes remains challenging due to short, noisy text, overlapping symptomatology,