Twitter posts containing both vaping-related (eg, “vape,” “ecig,” “e-juice,” “e-cigarette,” or “JUUL”; see Multimedia Appendix 1 for the complete list of keywords) and “Ice,” “Cool,” “Frost,” or “Arctic” terms were collected from January 1 to July 21, 2021, from Twitter’s streaming application programming interface. A total of 976,347 posts containing these terms were identified. Similar to previous studies, after excluding all retweets, we sampled out a random subset of 2001 posts for a content analysis [18,20]. Two coders (AG and EG) worked together with the last author (JPA) to become familiar with the data, created a codebook, and coded tweets into themes using an inductive approach. The unit of analysis was the text of the tweet. The purpose of the approach was to summarize the raw text-based data into a summary format and report the underlying themes evident in the data. Saturation was determined to be reached with 10 themes.

Vaping-related posts that contained the terms “Ice,” “Cool,” “Frost,” or “Arctic” but were determined to be unrelated to our research objectives were identified through manual content analysis and removed. For instance, irrelevant posts contained the following phrases: “cool vape trick” and “vaping is cool.” After excluding irrelevant and non-English posts, we were left with 590 (29.5%) tweets that were included in the content analysis.

The codebook consists of the following themes: personal testimonials (which were further divided into the following five subcategories: positive sentiment, negative sentiment, neutral sentiment, information seeking, and comparison to combustible tobacco), marketing, official (promotional) flavor description, disposable pod devices, JUUL, cannabidiol (CBD) products, and other. A tweet could be classified into multiple themes. To establish interrater reliability, the same subsample of posts (590 of the original 2001 posts) were double coded, with percent agreement ranging from 99.2% to 100% and Cohen κ ranging from 0.90 to 1.00. JPA served as an arbitrator resolving discrepancies between the coders.

Descriptive analyses were conducted first to show the prevalence of each theme. Further, pairwise co-occurrence analyses were used to evaluate connections across themes, since tweets often included more than one theme. R base package (version 4.0.2; R Foundation for Statistical Computing) was used to construct the co-occurrence matrix, and the igraph package (version 1.2.6) was used to visualize the results. Since the theme related to personal testimonials consisted of positive, negative, and neutral testimonies, Figure 1 does not include the overarching personal testimonials theme, and only includes its components: positive, negative, and neutral testimonies. The other category was also excluded since it had few co-occurrences with other themes.

All Twitter posts in this data set were publicly available and anonymized, and all analyses adhered to the terms and conditions, terms of use, and privacy policies of Twitter. To further protect privacy, posts exemplifying themes were paraphrased; no tweets are reported verbatim.

The protocol was approved by the university’s institutional review board (protocol HS-18-00697).

This study provides a summary of public Twitter posts collected over the course of a 7-month period, which includes mentions of both vaping-related and ice flavor–related terms. Posts in our corpus were related to ice-flavored e-cigarette product marketing, personal testimonials, and ice-flavored disposable pod devices. Theme co-occurrence in a single post was examined.

Marketing was a common theme in this study, while marketing and disposable pod devices represented a common theme co-occurrence in a single post. These findings are consistent with recent studies suggesting that ice-flavored products are often promoted in various pod-style cartridge-based disposable (eg, Puff Bar) and refillable (eg, PUFF Krush, PHIX) e-cigarette products [11]. These disposable and refillable pod-style products are among the fastest-growing segments of the e-cigarette market [3,9,10]. To reduce the appeal of e-cigarette products among youth, the US FDA took regulatory action (February 2020) to remove prefilled cartridge-based e-cigarettes in flavors other than menthol and tobacco, and issued warning letters to 10 e-cigarette companies instructing them to stop selling flavored disposable e-cigarettes by July 2020 [7,8]. Although some companies (eg, JUUL) voluntarily removed all flavors except menthol and tobacco, other companies (including those producing JUUL-compatible pods) continued to manufacture them. Given that ice hybrid–flavored products may contain both cooling and fruity flavors, it is unclear how these flavors fit into current regulatory policies. While the US FDA is moving toward removing menthol from combustible tobacco products, it is critical that regulatory actions consider restricting any constituents that taste/smell like menthol and produce a similar cooling sensory experience (ie, ice flavors) [12,24].

Personal testimonials was another common theme in this study, while positive sentiment was the most common category among testimonials. These findings are consistent with previous studies showing that Twitter users are exposed to positive messages about vaping and tobacco [25,26]. Positive messages about tobacco on social media may be linked with tobacco product experimentation [21]. Given that the messages about the ice-flavored products on Twitter are positive and that user conversations are not subject to federal policies, public health education campaigns informing the public on the harms of flavored e-cigarettes may be beneficial in countering these messages and reducing positive social norms about these products [21,26]. For instance, it could be possible to target individuals posting positive messages about ice-flavored e-cigarette products and deliver cessation messages directly to these individuals and their social networks. In addition, the US FDA regulations banning tobacco marketing on Twitter may help to limit the promotion of ice flavors to the public.

This study was limited to the analysis of discussions on ice-flavored e-cigarette products and may not pertain to other e-cigarette flavors. In addition, relatively few (n=2001, 0.2%) tweets among all identified ice tweets were analyzed (n=976,347), which limits the generalizability of our results. Moreover, the word “cool” was one of the four terms used to identify ice-flavored e-cigarette posts. “Cool” is a colloquialism and is commonly used to express the enjoyment of someone or something. The inclusion of this search term resulted in a collection of many posts that were unrelated to e-cigarette flavors; nonetheless, we found this limitation unavoidable and worked diligently to analyze pertinent tweets. This study focused on the text of the Twitter posts but did not code website links or accompanying images. Previous studies demonstrated that there is value in examining both text and image. In other words, it is possible that additional themes would have emerged had we coded images. Findings may not generalize to other time periods or other social media platforms. Our findings may not extend to all Twitter users or to the population of the United States.

Our findings may offer valuable insights into the public’s experience with and understanding of ice-flavored e-cigarette products. In this study, we found that Twitter discussions about ice-flavored e-cigarette products focused on marketing, personal testimonials, and ice-flavored disposable pod devices. Future studies should evaluate the relationship between exposure to personal testimonials of ice-flavored vaping products and curiosity, harm perceptions, and experimentation with these products among priority populations. Public health education campaigns informing the public on the harms of flavored e-cigarettes may be helpful in reducing positive social norms about ice-flavored products, while flavor regulatory actions may include ice flavors to prevent new products from circumventing current regulations. For instance, banning tobacco marketing posts on Twitter may reduce the exposure to ice-flavored e-cigarette product promotions. In addition, social media might be more influential than traditional marketing because participants can be actively engaged in the content they are viewing [21,22]. Close monitoring of ice flavor promotions on social media is needed to ensure that these promotions are not targeted toward minors and e-cigarette nonusers. Twitter has a policy prohibiting paid advertising of tobacco products to appear on its platform [27]. However, promotional posts related to ice-flavored products are still prevalent. Future studies should evaluate whether these promotions are sponsored by e-cigarette manufacturers or e-cigarette product distributors, or generated by general users.