Topics of Nicotine-Related Discussions on Twitter: Infoveillance Study.

BACKGROUND: Cultural trends in the United States, the nicotine consumer marketplace, and tobacco policies are changing. OBJECTIVE: The goal of this study was to identify and describe nicotine-related topics of conversation authored by the public and social bots on Twitter, including any misinformation or misconceptions that health education campaigns could potentially correct. METHODS: Twitter posts containing the term "nicotine" were obtained from September 30, 2018 to October 1, 2019. Methods were used to distinguish between posts from social bots and nonbots. Text classifiers were used to identify topics in posts (n=300,360). RESULTS: Prevalent topics of posts included vaping, smoking, addiction, withdrawal, nicotine health risks, and quit nicotine, with mentions of going "cold turkey" and needing help in quitting. Cessation was a common topic, with mentions of quitting and stopping smoking. Social bots discussed unsubstantiated health claims including how hypnotherapy, acupuncture, magnets worn on the ears, and time spent in the sauna can help in smoking cessation. CONCLUSIONS: Health education efforts are needed to correct unsubstantiated health claims on Twitter and ultimately direct individuals who want to quit smoking to evidence-based cessation strategies. Future interventions could be designed to follow these topics of discussions on Twitter and engage with members of the public about evidence-based cessation methods in near real time when people are contemplating cessation.

Characterizing Swisher Little Cigar-Related Posts on Twitter in 2018: Text Analysis.

BACKGROUND: Little cigars are growing in popularity in the United States, and Swisher is the market leader. The contexts and experiences associated with the use of Swisher-related products is understudied, but such information is available via publicly available posts on Twitter. OBJECTIVE: This study aimed to analyze Twitter posts to characterize Twitter users' recent experiences with Swisher-related products. METHODS: Twitter posts containing the term "swisher" were analyzed from January 1, 2018, to December 31, 2018. Text classifiers were used to identify topics in posts (n=81,333). RESULTS: The most prevalent topic was Person Tagging (mentioning a Twitter account in a post; 32.77%), followed by Flavors (eg, Grape and Strawberry; 20.96%) and Swisher use (eg, smoke swisher; 17.44%). Additional topics included Cannabis use (eg, blunt, roll, and gut swisher; 6.26%), Appeal (eg, like Swisher; 5.92%), Dislike (eg, posts that showed dissatisfaction with Swisher products; 3.53%), Purchases (eg, buy swisher; 1.90%), and Cigar comparison (eg, mentions of other cigar products including White-owl and Backwoods; 1.64%). CONCLUSIONS: This paper describes common contexts and experiences associated with the use of Swisher little cigars from the population posting on Twitter in 2018. These online messages may have offline consequences for tobacco-related behaviors, indicating the need for countering from public health officials. Findings should inform us about targets for surveillance, policy, and interventions addressing Swisher little cigars as well as communication planning and tobacco product counter messaging on Twitter.

Automated Identification of Hookahs (Waterpipes) on Instagram: An Application in Feature Extraction Using Convolutional Neural Network and Support Vector Machine Classification.

BACKGROUND: Instagram, with millions of posts per day, can be used to inform public health surveillance targets and policies. However, current research relying on image-based data often relies on hand coding of images, which is time-consuming and costly, ultimately limiting the scope of the study. Current best practices in automated image classification (eg, support vector machine (SVM), backpropagation neural network, and artificial neural network) are limited in their capacity to accurately distinguish between objects within images. OBJECTIVE: This study aimed to demonstrate how a convolutional neural network (CNN) can be used to extract unique features within an image and how SVM can then be used to classify the image. METHODS: Images of waterpipes or hookah (an emerging tobacco product possessing similar harms to that of cigarettes) were collected from Instagram and used in the analyses (N=840). A CNN was used to extract unique features from images identified to contain waterpipes. An SVM classifier was built to distinguish between images with and without waterpipes. Methods for image classification were then compared to show how a CNN+SVM classifier could improve accuracy. RESULTS: As the number of validated training images increased, the total number of extracted features increased. In addition, as the number of features learned by the SVM classifier increased, the average level of accuracy increased. Overall, 99.5% (418/420) of images classified were correctly identified as either hookah or nonhookah images. This level of accuracy was an improvement over earlier methods that used SVM, CNN, or bag-of-features alone. CONCLUSIONS: A CNN extracts more features of images, allowing an SVM classifier to be better informed, resulting in higher accuracy compared with methods that extract fewer features. Future research can use this method to grow the scope of image-based studies. The methods presented here might help detect increases in the popularity of certain tobacco products over time on social media. By taking images of waterpipes from Instagram, we place our methods in a context that can be utilized to inform health researchers analyzing social media to understand user experience with emerging tobacco products and inform public health surveillance targets and policies.

Hookah-Related Posts to Twitter From 2017 to 2018: Thematic Analysis.

BACKGROUND: Hookah (or tobacco waterpipe) use has recently become prevalent in the United States. The contexts and experiences associated with hookah use are unclear, yet such information is abundant via publicly available hookah users' social media postings. OBJECTIVE: In this study, we utilized Twitter data to characterize Twitter users' recent experiences with hookah. METHODS: Twitter posts containing the term "hookah" were obtained from April 1, 2017 to 29 March, 2018. Text classifiers were used to identify clusters of topics that tended to co-occur in posts (n=176,706). RESULTS: The most prevalent topic cluster was Person Tagging (use of @username to tag another Twitter account in a post) at 21.58% (38,137/176,706) followed by Promotional or Social Events (eg, mentions of ladies' nights, parties, etc) at 20.20% (35,701/176,706) and Appeal or Abuse Liability (eg, craving, enjoying hookah) at 18.12% (32,013/176,706). Additional topics included Hookah Use Behavior (eg, mentions of taking a "hit" of hookah) at 11.67% (20,603/176,706), Polysubstance Use (eg, hookah use along with other substances) at 10.95% (19,353/176,706), Buying or Selling (eg, buy, order, purchase, sell) at 9.37% (16,552/176,706), and Flavors (eg, mint, cinnamon, watermelon) at 1.66% (2927/176,706). The topic Dislike of Hookah (eg, hate, quit, dislike) was rare at 0.59% (1043/176,706). CONCLUSIONS: Social events, appeal or abuse liability, flavors, and polysubstance use were the common contexts and experiences associated with Twitter discussions about hookah in 2017-2018. Considered in concert with traditional data sources about hookah, these results suggest that social events, appeal or abuse liability, flavors, and polysubstance use warrant consideration as targets in future surveillance, policy making, and interventions addressing hookah.

Images of Little Cigars and Cigarillos on Instagram Identified by the Hashtag #swisher: Thematic Analysis.

BACKGROUND: Little cigar and cigarillo use is becoming more prevalent in the United States and elsewhere, with implications for public health. As little cigar and cigarillo use grows in popularity, big social media data (eg, Instagram, Google Web Search, Twitter) can be used to capture and document the context in which individuals use, and are marketed, these tobacco products. Big social media data may allow people to organically demonstrate how and why they use little cigars and cigarillos, unprimed by a researcher, without instrument bias and at low costs. OBJECTIVE: This study characterized Swisher (the most popular brand of cigars in the United States, controlling over 75% of the market share) little cigar- and cigarillo-related posts on Instagram to inform the design of tobacco education campaigns and the development of future tobacco control efforts, and to demonstrate the utility in using big social media data in understanding health behaviors. METHODS: We collected images from Instagram, an image-based social media app allowing users to capture, customize, and post photos on the Internet with over 400 million active users. Inclusion criteria for this study consisted of an Instagram post with the hashtag "#swisher". We established rules for coding themes of images. RESULTS: Of 1967 images collected, 486 (24.71%) were marijuana related, 348 (17.69%) were of tobacco products or promotional material, 324 (16.47%) showed individuals smoking, 225 (11.44%) were memes, and 584 (29.69%) were classified as other (eg, selfies, food, sexually explicit images). Of the marijuana-related images, 157/486 (32.3%) contained a Swisher wrapper, indicating that a Swisher product was used in blunt making, which involves hollowing out a cigar and refilling it with marijuana. CONCLUSIONS: Images from Instagram may be used to complement and extend the study of health behaviors including tobacco use. Images may be as valuable as, or more valuable than, words from other social media platforms alone. Posts on Instagram showing Swisher products, including blunt making, could add to the normalization of little cigar and cigarillo use and is an area of future research. Tobacco control researchers should design social media campaigns to combat smoking imagery found on popular sites such as Instagram.

Instagram Posts Related to Backwoods Cigarillo Blunts: Content Analysis.

BACKGROUND: Instagram, one of the most popular social media platforms among youth, offers a unique opportunity to examine blunts-partially or fully hollowed-out large cigars, little cigars, and cigarillos that are filled with marijuana. Cigarillo brands like Backwoods (Imperial Tobacco Group Brands LLC) have product features that facilitate blunt making, including a variety of brand-specific flavors that enhance the smoking experience (eg, honey, dark stout). Backwoods has an active online presence with a user-friendly website. OBJECTIVE: This study examined the extent to which Backwoods cigarillo-related posts on Instagram showed blunt making. Instagram offers a unique opportunity to examine blunt making as Instagram accounts will contain images reflective of behavior occurring without the prime of a researcher. METHODS: Data consisted of publicly available Instagram posts with the hashtag #backwoods collected from August 30 to September 12, 2018. Inclusion criteria for this study included an Instagram post with the hashtag "#backwoods". Rules were established to content analyze posts. Categories included Type of post (ie, photo, video, or both); Blunt-related hashtags (ie, the corresponding post caption contained one or more hashtags like #blunts, #cannabis, and #weed that were identified in previous social media research); Rolling blunts (ie, the post contained an image of one or more individuals rolling a Backwoods cigarillo visibly containing marijuana); and Smoking blunts (ie, the post contained an image of one or more individuals blowing smoke or holding a lit blunt). We coded images for Product flavor reference, where a code of 1 showed a Backwoods cigarillo pack with a brand-specific flavor (eg, honey, dark stout, Russian crème) visible in the blunt-related image, and a code of 0 indicated that it was not visible anywhere in the image. RESULTS: Among all posts (N=1206), 871 (72.2%) were coded as Blunt-related hashtags. A total of 125 (10.4%) images were coded as Smoking blunts, and 25 (2.1%) were coded as Rolling blunts (ie, Backwoods cigarillo explicitly used to roll blunts). Among blunt images, 434 of 836 (51.9%) were coded as Product flavor (ie, a Backwoods pack with a brand-specific flavor was visible). CONCLUSIONS: Most Backwoods cigarillo-related Instagram images were blunt-related, and these blunt-related images showed Backwoods packages indicating flavor preference. Continued monitoring and surveillance of blunt-related posts on Instagram is needed to inform policies and interventions that reduce the risk that youth may experiment with blunts. Specific policies could include restrictions on product features (eg, flavors, perforated lines, attractive resealable foil pouches, sale as singles) that facilitate blunt making.

Sociodemographic differences in young adults' recall of tobacco and cannabis marketing online and in television/film.

Young adults (YA) who report viewing pro-tobacco and cannabis marketing are at increased risk for using tobacco and cannabis. However, there is a growing diversity of tobacco and cannabis products on the market, as well as methods for marketing them. Prevalence of, and sociodemographic differences in, YA's recall of various types of tobacco and cannabis marketing is not well-characterized. Data were from a cohort of YA (mean age: 19.8) from Southern California in 2019. Respondents were asked whether they recalled having seen two types of marketing (online advertisements and portrayals of product use in TV/movies) for 5 tobacco and 3 cannabis products among never-users of tobacco (N = 954) and cannabis (N = 1,046), respectively. Sociodemographic differences in marketing recall were subsequently assessed. Among tobacco-naïve respondents, 31.3% and 49.3% recalled seeing online advertisements and tobacco use in TV/movies, respectively. Among cannabis-naïve respondents, 18.7% and 31.0% recalled seeing online advertisements and cannabis use in TV/movies, respectively. Overall, respondents recalled seeing tobacco and cannabis products on TV/movies at higher rates than seeing online advertisements, with the exception of electronic cigarettes, for which online advertisements were seen at higher rates. Women (vs. men) had higher odds of seeing tobacco (aOR = 1.9) and cannabis use in TV/movies (aOR = 1.4) and cannabis marketing online (aOR = 1.4). LGB (vs. straight) respondents had higher odds of seeing cannabis marketing online (aOR = 1.7). Efforts to regulate exposure to tobacco and cannabis marketing among young women and LGB people merit further consideration.

Characterizing #Backwoods on Instagram: "The Number One Selling All Natural Cigar".

We sought to assess the proportion of Backwoods (Imperial Tobacco Group Brands LLC) cigar-related posts to Instagram that may contain misleading claims, nature-evoking imagery, and appealing flavors. Inclusion criteria for this study included an Instagram post with the hashtag "#backwoods" from 30 August to 12 September 2018. Rules were established to content analyze (n = 1206) posts. Categories included misleading packaging (i.e., the post contained an image of a Backwoods product with the descriptor "natural" on the packaging), misleading promo (i.e., the corresponding caption to the post contained hashtag(s) like "#natural", "#authentic", "#alwaystrue"), nature-evoking imagery (i.e., the post contained images of grass, water, and pastural views along with a Backwoods product), flavors (i.e., the post contained a Backwoods product with brand-specific flavors on the packaging), flavor promo (i.e., the corresponding caption to the post contained hashtag(s) of Backwoods' brand-specific flavors), marijuana-related (i.e., the post contained an image of marijuana next to a Backwoods pack, rolled cigars visibly contained marijuana, or hollowed-out cigars next to marijuana), smoking (the post contained an image of smoke or a lit cigar), brand-specific promo (i.e., the post contained an image of a Backwoods t-shirt, sweatshirt, hat, etc.), and perceived gender. Among the posts analyzed, 645 (53.5%) were marijuana-related, 564 (46.8%) were flavors, 463 (38.4%) were misleading packaging, 335 (27.8%) were flavor promo, 309 (25.6%) were misleading promo, 188 (15.6%) were nature-evoking imagery, 165 (13.7%) were smoking, 157 (13.0%) were brand-specific promo, and 239 (19.8%) were perceived male gender. Backwoods cigar-related posts to Instagram often contained misleading images and promotions of a "natural" tobacco product, images of marijuana use (in the form of blunt-making), brand-specific flavors, smoking, and promotional merchandise. Misleading images and the depictions of marijuana use in addition to the variety of flavor options may increase product appeal to consumers. These results underscore the need for comprehensive regulation of cigar products similar to cigarettes.

Risk factors associated with subsequent initiation of cigarettes and e-cigarettes in adolescence: A structural equation modeling approach.

BACKGROUND: Previous youth tobacco research has identified multiple correlated risk factors for initiation of cigarette and e-cigarette use; whether these factors are independently associated with initiation is not known, due to challenges with disentangling the independent effects of these correlated risk factors. METHODS: Students in 11th/12th grade enrolled in the Southern California Children's Health Study were surveyed in 2014 (baseline) and again in 2015 (N = 1553). Structural equation models (SEM) were developed to investigate associations of susceptibility, marketing, and the social environment (as latent factors), and other tobacco use at baseline with cigarette or e-cigarette initiation between baseline and follow-up. Analyses were restricted to baseline never cigarette users (N = 1293) for models evaluating cigarette initiation, and to never e-cigarette users (N = 1197) for models evaluating e-cigarette initiation. RESULTS: In fully-adjusted prospective SEM models, latent factors for cigarette susceptibility, marketing, and the social environment, along with ever e-cigarette use and ever hookah use at baseline were independently associated with cigarette initiation between baseline and follow-up (P < 0.05). Similarly, latent factors for e-cigarette susceptibility, marketing, and the social environment, along with ever hookah use at baseline were associated with e-cigarette initiation between baseline and follow-up (P < 0.05); however, cigarette use at baseline was not associated with e-cigarette initiation in SEM models (P = 0.16). CONCLUSIONS: We identified independent effects of multiple risk factors in SEM models on initiation of cigarettes and e-cigarettes. E-cigarette use was associated with cigarette initiation, but cigarette use was not associated with e-cigarette initiation in fully adjusted models. Research to identify underlying causal mechanisms is warranted.

The Why We Retweet scale.

BACKGROUND: Twitter offers a platform for rapid diffusion of information and its users' attitudes and behaviors. Insights about information propagation via retweets (the message forwarding function) offer observable explanations of ways in which modern human interactions get organized in the form of online networks, and contextualized in the form of public health, policy decisions, disaster management, and civic participation. This study conceptualized and validated the Why We Retweet Scale to contextualize retweeting behavior. OBJECTIVE: Twitter users were identified using clustering algorithms that consider a users' position in their network and invited for an online survey. Participants (N = 1433) responded to 19 questions about why they retweet. Exploratory factor Analysis (EFA) was conducted on a scale development sample (70% of original sample), which informed the Confirmatory Factor Analysis (CFA) on a scale testing sample (30% of the original sample). Varimax rotation was used to obtain a rotated factor solution, which resulted in interpretable factors. Demographic differences among scale factors were analyzed using one-way ANOVA or independent samples t-tests. RESULTS: The final model (χ221 = 28, RMSEA = .03 [90% CI, 0.00-0.06], CFA = .99, TLI = 0.99) represented a parsimonious solution with 4 factors, measured by 2-3 items each, creating a final scale consisting of 9 items. Factor labels and definitions were: (1) Show approval, "Show support to the tweeter"; (2) Argue, "To argue against a tweet that I disagree with"; (3) Gain attention, "Add followers or gain attention"; and (4) Entertain, "Create humor/amusement". Demographic differences were also reported. CONCLUSIONS: The Why We Retweet Scale offers a useful conceptualization and assessment of motivations for retweeting. In the future, communication strategists might consider the factors associated with information propagation when designing campaign messages to maximize message reach and engagement on Twitter.

Identifying Sentiment of Hookah-Related Posts on Twitter.

BACKGROUND: The increasing popularity of hookah (or waterpipe) use in the United States and elsewhere has consequences for public health because it has similar health risks to that of combustible cigarettes. While hookah use rapidly increases in popularity, social media data (Twitter, Instagram) can be used to capture and describe the social and environmental contexts in which individuals use, perceive, discuss, and are marketed this tobacco product. These data may allow people to organically report on their sentiment toward tobacco products like hookah unprimed by a researcher, without instrument bias, and at low costs. OBJECTIVE: This study describes the sentiment of hookah-related posts on Twitter and describes the importance of debiasing Twitter data when attempting to understand attitudes. METHODS: Hookah-related posts on Twitter (N=986,320) were collected from March 24, 2015, to December 2, 2016. Machine learning models were used to describe sentiment on 20 different emotions and to debias the data so that Twitter posts reflected sentiment of legitimate human users and not of social bots or marketing-oriented accounts that would possibly provide overly positive or overly negative sentiment of hookah. RESULTS: From the analytical sample, 352,116 tweets (59.50%) were classified as positive while 177,537 (30.00%) were classified as negative, and 62,139 (10.50%) neutral. Among all positive tweets, 218,312 (62.00%) were classified as highly positive emotions (eg, active, alert, excited, elated, happy, and pleasant), while 133,804 (38.00%) positive tweets were classified as passive positive emotions (eg, contented, serene, calm, relaxed, and subdued). Among all negative tweets, 95,870 (54.00%) were classified as subdued negative emotions (eg, sad, unhappy, depressed, and bored) while the remaining 81,667 (46.00%) negative tweets were classified as highly negative emotions (eg, tense, nervous, stressed, upset, and unpleasant). Sentiment changed drastically when comparing a corpus of tweets with social bots to one without. For example, the probability of any one tweet reflecting joy was 61.30% from the debiased (or bot free) corpus of tweets. In contrast, the probability of any one tweet reflecting joy was 16.40% from the biased corpus. CONCLUSIONS: Social media data provide researchers the ability to understand public sentiment and attitudes by listening to what people are saying in their own words. Tobacco control programmers in charge of risk communication may consider targeting individuals posting positive messages about hookah on Twitter or designing messages that amplify the negative sentiments. Posts on Twitter communicating positive sentiment toward hookah could add to the normalization of hookah use and is an area of future research. Findings from this study demonstrated the importance of debiasing data when attempting to understand attitudes from Twitter data.