Investigating #vapingcessation in Twitter

Abstract Evidence suggests that an increasing number of e-cigarette users report intentions and attempts to quit vaping. Since exposure to e-cigarette-related content on social media may influence e-cigarette and other tobacco product use, including potentially e-cigarette cessation, we aimed to explore vaping cessation-related posts on Twitter by utilizing a mixed-methods approach. We collected tweets pertaining to vaping cessation for the time period between January 2022 and December 2022 using snscrape. Tweets were scraped for the following hashtags: #vapingcessation, #quitvaping, and #stopJuuling. Data were analysed using Azure Machine Learning and Nvivo 12 software. Sentiment analysis revealed that vaping cessation-related tweets typically embody positive sentiment and are mostly produced in the U.S. and Australia. Our qualitative analysis identified six emerging themes: vaping cessation support, promotion of vaping cessation, barriers and benefits to vaping cessation, personal vaping cessation, and usefulness of peer support for vaping cessation. Our findings imply that improved dissemination of evidence-based vaping cessation strategies to a broad audience through Twitter may promote vaping cessation at the population level.

smoking cessation tool is lacking . 1 Individuals who use nicotine (known addictive substance in tobacco product) containing e-cigarettes regularly often become dependent on e-cigarette use and intend to quit. 2 Furthermore, studies show that e-cigarette vapor contains carcinogenic carbonyl compounds. 3,4 Ecigarette use may be linked with lung and bladder cancer and respiratory disorders. 5,6 Thus, concerns for health may also motivate regular, exclusive e-cigarette users to quit e-cigarette use. 7 For the past several years, e-cigarette content on social media has been widely prevalent and, hence, understanding how the topic of vaping cessation is represented on social media is vital. 8 Prior research has yet to address how individuals search and share vaping cessation concerns on social media.
Twitter is a microblogging and social networking platform. Unstructured, free-text tweets relating to health care are frequently shared on Twitter. 9 Sentiment analysis is a natural language processing technique that enables the analysis text for the intensity of sentiment, which in turn facilitates the characterization of discussions pertaining to various health-related issues. 10,11 Sentiment analysis uses computational algorithms to extract subjective information from written text and to identify the strength of the text's positive or negative tone. [12][13][14] Sentiment analysis has been used to predict health behaviours through an increased understanding of how people feel with respect to speci c health topics or conditions. [15][16][17] A recent Twitter-based study on e-cigarettes found that there are more positive than negative sentiments about e-cigarettes prevalent on the platform. 18 However, there has been no research on the content related to vaping cessation on Twitter. Examining how vaping cessation is represented on Twitter is imperative because social media platforms such as Twitter may in uence an individual's decision to continue using e-cigarettes. [19][20][21][22] Twitter content may not only provide factual information about vaping (e.g., adverse consequences) but also express sentiments through tone or underlying message connotations. Gaining a deeper understanding of the sentiments prevalent on Twitter regarding vaping cessation may assist public health professionals and policymakers in better utilizing social media platforms to disseminate evidence-based vaping cessation strategies. For example, a nding that sentiments tied to vaping cessation are generally positive may suggest higher tendency among users on the platform to engage in discussions about vaping cessation and also higher receptivity towards vaping cessation messages. Thus, the objective of this study was to apply a mixed-methods approach to analyze the sentiments expressed in Twitter posts concerning vaping cessation. Speci cally, we examined Twitter posts under the following hashtags: #vapingcessation OR #quitvaping OR #stopJuuling.

METHODS
The process of extracting data from a website is commonly referred to as web scraping. In this study, tweets were manually collected using snscrape, 23 a Python based library that allows for the extraction of tweets without the need for personal Twitter API keys. The library provides a powerful search functionality to help lter tweets based on various conditions, such as date-time, language, and number of likes. For this study, we obtained English tweets related to vaping cessation by using the search query #vapingcessation OR #quitvaping OR #stopJuuling and setting the "since" and "until" ags to January 1, 2022 and December 31, 2022, respectively. Through this we obtained 405 publicly accessible vaping cessation-related full-text tweets with their metadata such as the number of followers and user geolocation. Identi able information was not accessible and was therefore not examined.
Data were downloaded in the form of CSV (comma separated value) le. The retrieved data were cleaned by removing the duplicate tweets, tweets containing URL(s) or hashtags only, and non-English tweets. The cleaning process resulted in an analytical sample of 231 tweets. We focused on the analysis of the textual information in the tweets and did not analyse photos and videos, as visual information is not integral to the communication in Twitter.
This study was conducted in two parts. First, to understand sentiments expressed towards vaping cessation, sentiment analysis was performed using Azure Machine Learning, 24 a popular open source tool designed to estimate the strength of positive and negative sentiment in short, informal text. It emits the text sentiment as a numerical rating known as a 'sentiment score' ranging between 0 and 1. A score closer to 0 it is indicative of a negative sentiment while a score closer to 1 is indicative of a positive sentiment. Positive numbers indicate favourable attitudes, while negative numbers indicate negative attitudes. 24 Intermediate values are tagged as neutral.
Thematic content analysis, performed using Nvivo 12, 25 was used to generate codes through an inductive approach to identify different themes represented in the data. Initially, the raw data (tweets) were reviewed to gain familiarity, and later, the tweets were inductively coded according to their meanings within the corresponding themes. For each theme, tweets were organized within categories according to emerging patterns. Themes and categories were not mutually exclusive. A tweet may be coded into one or more themes or categories. The themes' titles were adjusted according to the ndings.

RESULTS
According to our automated sentiment analysis pipeline, a total of 69% (160 of 231) tweets represented positive sentiment, 16% negative (n = 37), and 15% neutral (n = 34). Across accounts, the number of followers ranged from 0 to 21,338. The mean number of followers was 1,332 ± 2904. 86% (189 of 231) of tweets were posted from the U.S. Other locations represented in our curated dataset included Australia, Canada, Japan, Malaysia, the United Arab Emirates, and the United Kingdom. The distributions of the positive, negative, and neutral sentiments are shown in Fig. 1. Most of the positive and neutral sentiment tweets originated from the U.S. and Australia, respectively. The negative sentiment tweets' locations varied across the globe.
The qualitative ndings were grouped into 6 main emerging themes and several subthemes (Table 1) based on positive, negative, and neutral sentiments. The majority of the tweets corresponded to identi ed themes of vaping cessation support offers (43%) and promotion of vaping cessation (31%) (Fig. 2). The primary themes and subthemes are described below, and illustrative tweets related to each theme and subtheme are listed in Table 1. cessation. The role of the school in promoting vaping cessation was tweeted in terms of establishing vape disposal boxes or installation of vape detectors. A few evidences based videos were also posted to persuade individuals to quit vaping. Environment-related tweets that mentioned vaping devices as a detrimental source of lithium waste in land lls, and climate change due to exhaled vapor from ecigarettes, also encouraged quitting vaping.

Theme 3: Barriers to vaping cessation
The tweets with negative sentiments highlighted fear of anxiety, nicotine craving, withdrawal symptoms, and lack of motivation as barriers to quitting vaping. The primary reasons for the lack of motivation to stop vaping were lack of access to a personal vaping device or attempts at quitting without any professional support or assistance (i.e., quitting cold turkey). Numerous tweets mentioned intense nicotine craving as the cause of the failure to successfully stop vaping. Physical symptoms that were identi ed in the tweets as a barrier to quitting vaping were sore throats, short breaths, and lung scarring. Inability to manage anxiety was also another barrier to quitting vaping.
Theme 4: Personal vaping cessation experience A total of 7.7% (18 of 231) tweets shared personal timelines of remaining vape free, which ranged from 6 hours to 150 days. Some tweets mentioned the personal excitement and struggles of the vape-free journey. Some tweets shared individuals' information-seeking quit strategies.
Theme 5: Peer support for vaping cessation The importance of peers and family members (e.g., spouse, partner, friends, parents, grandparents, coworkers) in vaping cessation was another prominent theme. Tweets suggested that sharing a quit date with others in one's social network might be a great way to get support from them. Teens might get direct support from their parents, grandparents or school. Young adults might get support from their partners who intend to help them to quit vaping.
Theme 6: Bene ts to vaping cessation Tweet contents revealed various bene ts of quitting vaping such as reduced stroke risk, proper teen brain development, lower blood pressure, mode improvements, and improved heart rates and heart health. Few tweets offered helpful tips for a successful quit. The impact of second and third-hand vaping was also stated in a few tweets.

DISCUSSION
This may be one of the rst studies to analyze the sentiments expressed on Twitter posts about quitting e-cigarette use or vaping cessation. The study focused speci cally on the following hashtags: #vapingcessation, #quitvaping, and #stopJuuling. The sentiment analysis revealed that tweets with vaping cessation hashtags typically expressed positive sentiments and were mostly tweeted from the U.S. and Australia. The prevalence of negative and neutral sentiments proportion was almost similar (16% and 14% respectively) and substantially lower than positive sentiments. The qualitative analysis identi ed 6 emergent themes: vaping cessation support offers, promotion of vaping cessation, barriers to vaping cessation, bene ts of vaping cessation, personal vaping cessation experience, and peer support for vaping cessation. Our ndings suggest that Twitter is a useful tool to identify vaping cessation-related sentiments and themes commonly expressed in the public.
Our ndings suggest that the majority of the posts related to vaping cessation hashtags were tweeted in the United States and Australia, which indicates a higher interest in vaping cessation in these geographical locations. Regulations on nicotine-containing e-cigarettes differ globally. 26 Across all Australian States and Territories, it is illegal to sell nicotine-containing e-cigarettes because liquid nicotine is classi ed as a 'Schedule 7-Dangerous Poison'; however, users can legally import nicotine-containing ecigarettes through the Personal Importation Scheme which states that users must obtain a prescription from a physician. 27 Australia has relatively low rates of e-cigarette use compared to other countries, 28 which is likely a result of this regulatory policy. However, anyone 21 years of age or older in the U.S. can purchase nicotine-containing e-cigarettes, 29 which are highly addictive. 30 Given the higher prevalence of e-cigarette use in the U.S. and relatively freer access to nicotine-containing e-cigarette products, a multipronged approach to counter e-cigarette use through prevention and cessation may be particularly useful in the U.S. Such an approach may bene t from strategically utilizing social media platforms such as Twitter.
Our current study presents a timely analysis of Twitter handles. Thus, Twitter might be a particularly promising digital platform for delivering vaping cessation interventions to a broader community as prior Twitter-delivered interventions for smoking demonstrate promise. 31 Additionally, Twitter can allow private groups to be created, making it ideal for delivering and privatizing a vaping cessation intervention to those who will seek such quitting support. Evidence also suggests that Twitter-based social media interventions that combine traditional online social support with daily auto-messages is novel for tobacco product use cessation. 31 Hence, Twitter is an affordable tool with the potential to circulate evidencebased information on vaping cessation to the broader community on a global scale. Future studies are encouraged to explore how to integrate other social media platforms such as Instagram or TikTok with vaping cessation interventions.
The emergent themes related to vaping cessation hashtags on Twitter as highlighted by the present data are consistent with previous qualitative Twitter-based studies which have described similar themes related to barriers, promotion, and motivation to quit smoking or e-cigarette use. [32][33][34][35][36] However, the current qualitative study based on sentiment analysis may be the rst study to focus on the unique themes associated with vaping cessation. The current qualitative ndings may provide a "baseline" for future social media-based vaping cessation intervention programs.
There are several limitations to this study, including the use of hashtags for data collection. For example, many traditional smoking-related posts (such as #quitsmoking or #smokingcessation) might also include vaping cessation hashtags. Data pertaining to such hashtags might have been excluded from the current dataset, reducing the potentially eligible pool of tweets included in the current data analysis. Additionally, we might have missed relevant messages by limiting our search strategy to #vapingcessation, #quitvaping, and #stopJuuling. Many accounts did not publicly declare the locations of their origin, which precluded them from being included in our analysis. Further, we did not consider retweets of the eligible posts. Future studies should provide a better understanding of retweeting behaviour because retweeting is one of the primary modes of information dissemination on Twitter. Finally, we did not include media (pictures, videos, or emoji) related to #vapingcessation OR #quitvaping OR #stopJuuling on Twitter. Future studies should also consider analysing such media posts as prior studies have reported that video or emoji containing posts express positive sentiments and engage the users most in the social media. 37 Despite the above-mentioned limitations, our ndings demonstrate that vaping cessation hashtags on Twitter may have a lot of potential to promote cessation at the population level. Future research and policy efforts need to focus on increasing vapers' motivation to quit vaping using social media platforms. Additionally, those who are attempting to stop vaping may bene t from tailored Twitter-based treatment interventions. Moreover, to be effective, Twitter-based interventions may need to utilize peer or other social networks dynamics and address the physical and psychological consequences of vaping withdrawal.

Declarations
Ethics approval and consent to participate: Not applicable Consent for publication: Not applicable (No Human participants) Availability of data and materials: The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Competing interests: None
Funding: This research was supported by funds from the National Cancer Institute (R01CA228905) Figure 1 Distribution of tweet sentiments geographical locations Figure 2 The proportion of tweets and emergent themes