The emergence of vaping, or e-cigarette technology, whereby a liquid solution is heated in a portable device until aerosolized and then inhaled, has been greeted with both promise and alarm [1-4]. In response to concerns about lung injury associated with vaping [5] and the increasing prevalence of youth vaping [6,7], Apple removed previously-approved vaping-related apps from its App Store (for iOS) in November 2019 [8], but the competing Google Play Store (for Android) did not remove vaping-related apps. Vaping-related apps may negatively influence youth by glamorizing vaping; however, some apps may benefit adults who use tobacco by providing cessation assistance or harm reduction information. The net balance of benefits and drawbacks from vaping-related apps likely depends on app content.

Vaping nicotine may be less harmful than smoking combustible cigarettes [9] and may be a potential tool in helping people quit smoking [4,10,11]. Yet, alongside this potential is limited scientific evidence about the efficacy of nicotine vaping in promoting and maintaining tobacco-smoking cessation, as well as mounting concerns about a youth vaping epidemic [12,13]. Evidence to date appears to be mixed and context-dependent regarding whether e-cigarettes help adults quit [10] with concerns that the amount of nicotine in popular vaping products is too high and that people who quit smoking by vaping may return to smoking [1].

E-cigarettes are currently the most popular tobacco product among US youth, with an estimated 1 in 5 US high-school seniors reporting having vaped nicotine in the previous month, and 1 in 4 reporting having vaped nicotine- or flavor-containing e-liquids in 2018 [6]. A higher past-month prevalence of vaping was reported in the 2019 National Youth Tobacco Survey, with more than 1 in 4 US high-school students using e-cigarettes [7]. Although adolescents are smoking fewer cigarettes per day [14], and smoking has become less prevalent [15] in recent years, there are also reports of greater nicotine dependence among adolescents and young adults who vape prefilled pods (cartridges) [16], and a body of research indicating harm to the developing brain from nicotine [17]. In response, national and local governments have taken steps to propose or enact restrictions (or bans) on e-cigarette sales [18,19] and flavored vaping products [20].

Similarly, vaping cannabis extracts or flower is believed to be less harmful than inhaling combusted cannabis [21-23]. Vaping is an increasingly popular method of cannabis consumption in the United States [24-26] although cannabis remains illegal in many US states and most countries. Both formal and informal cannabis product markets now offer a wide array of ways to consume tetrahydrocannabinol (THC), including vaping products [2,27]. The legalization of cannabidiol from hemp products in the United States in 2018 has led to an increase in interest and availability of cannabidiol-containing products [28] some of which can be vaped and do not need to be legally purchased from a licensed dispensary. Many vaping products with both THC and cannabidiol are sold online with limited regulation, and published product testing studies have found that tested cannabinoid content was often higher or lower than labeled content [29,30]. While Canada legalized cannabis nationally in October 2018, extracts and edibles were not legal until October 2019, due to additional difficulties in regulation.

Against this background, in the spring and summer of 2019, cases of acute lung injury related to vaping began to be reported in the United States. This new illness, named electronic or vaping product use associated lung injury (EVALI), ultimately caused at least 52 deaths and over 2600 hospitalizations by December 2019 [5]. Most of these patients were young adults who were previously healthy and reported a gradual onset of respiratory, constitutional, and gastrointestinal symptoms. Most—though not all—patients reported having used e-cigarette or vaping products containing THC. The most likely culprit was identified as vitamin E acetate, used as a bulking agent in primarily THC-containing vaping products, often procured in a state where cannabis was still illegal or from an unlicensed seller [31-34]. People who vape cannabis or nicotine may turn to mobile apps for guidance on safe use.

Mobile phones and mobile apps are increasingly used for accessing health and safety information about many topics [35,36], including substances. Apps are most used by younger populations with higher income and education [37]. The Google Play Store for Android and the Apple App Store for iOS are the 2 major mobile app platforms and marketplaces for digital apps, with 2.6 million available apps on Google Play and 1.8 million apps available on the Apple App Store in early 2020 [38]. Many popular apps have versions for both platforms. Both app stores have content, technical, and stylistic guidelines that developers must follow in order to have their app approved; violations can be grounds for removal of a previously approved app. It is generally believed that getting approval from the Apple App Store is more difficult than getting approval from the Google Play Store [39], with Apple App Store apps often viewed as higher quality and less likely to be free to use [40].

Both app stores specifically address tobacco and cannabis content in their developer guidelines (Table 1). Google Play Store’s substance-related content guidelines prohibit apps “facilitating the sale” of tobacco, marijuana, alcohol, or illegal drugs, or those “depicting or encouraging” use by minors [41]. The Apple App Store guidelines prohibit apps that “encourage consumption of tobacco and vape products, illegal drugs, or excessive amounts of alcohol,” particularly those encouraging minors [42]. Facilitating sale of these substances is also not allowed [42].

In response to health concerns about vaping devices (eg, youth nicotine use, lung injury) and calls for bans on nicotine flavors and vaping devices, Apple prohibited new vaping-related mobile apps from its iOS App Store in June 2019 and removed 181 previously approved vaping-related apps on November 15, 2019 [8,43]. Articles published on technology, vaping, and cannabis industry–affiliated websites in the following weeks decried this move by Apple, stating that the ban impacted device users’ ability to safely use their products [44], removed a resource that assists people with quitting smoking [45], and would have detrimental effects on innovation [45], and that the app store approval process was inconsistent [46].

At a time of escalating concern about vaping, Apple’s policy change may benefit public health by lessening youth exposure to content that glamorizes vaping. Exposure to vaping-related content online has been associated with greater intentions to vape [47] and greater likelihood of vaping [48] among youth. However, Apple’s restrictions may also block important sources of information and vaping device control for adults seeking to use vaping devices safely.

The content and evidence base of vaping-related apps is currently unknown, though previous reviews and content analyses have examined apps related to smoking cessation, cannabis, alcohol [49], and other substance use [50]. Smoking cessation apps are abundant, though few have been demonstrated to be evidence-based [51-55], and many have limitations with usability [56]. A review [57] of cannabis apps conducted in 2014 found that the most common content areas largely consisted of cannabis strain classification guides, factoids about cannabis, and games, but few apps addressed negative health effects of cannabis use. To gain insight into the potential positive and negative effects of removing vaping-related apps from app stores, characterization of the remaining vaping-related apps in the Google Play Store, the main competing source for mobile apps, is an imperative first step.

We analyzed the top vaping-related mobile apps on the Google Play Store with respect to app characteristics, intended purpose, and provision of features and information for limiting the potential harms of smoking and vaping. To our knowledge, the Google Play Store did not enact any restrictions on vaping-related apps in response to EVALI. Within the context of government and industry regulatory policy changes surrounding vaping devices, understanding the types of apps available for mobile phones and tablets can shed light on how this form of digital media may reflect and influence information about nicotine and cannabis vaping available to consumers. Depending on their purposes and features, vaping-related apps could encourage or discourage vaping. Some apps may provide resources for adults seeking to switch to vaping from smoking or to use vaping devices safely, while others may negatively impact youth by glamorizing vaping. As such, the removal of vaping-related apps may have mixed effects on population health, depending on which types of content predominate.

A search for apps using the keywords vape and vaping was performed on the Google Play Store on December 17, 2019 from an IP address in the United States. The names and order of the first 100 apps displayed in response to the search query were recorded. Similar to previous studies reviewing apps, only the first 100 apps of 250 search results were examined, because it is unlikely that individuals would browse further for a desired app without refining their search terms [56-58]. We chose not to include e-cigarettes or e-cigs as search terms because of their infrequent use outside of academic and regulatory discourses.

Resulting vaping apps available in the Google Play Store served as a proxy for apps that may have been available in the Apple App Store prior to November 2019, as a list of removed apps was not publicly available. Thus, coding vaping-related apps in the Google Play Store provided insight into the types of the apps both currently available to Google Play Store customers and previously available to Apple App Store customers. A search for vaping and vape in the Apple App Store in December 2019, performed on both a Mac and an iPhone, yielded zero results on the Mac Store and yielded only apps related to quitting vaping or to vapor-like image effects in the iPhone app store.

Three investigators (MM, EV, JT) each extracted app characteristics directly from the Google Play Store for one-third of the apps (33-34 apps per investigator). Extracted app characteristics included the developer name, content and age ratings, cost, average star ratings, number of reviewers, number of installations, last date updated, and URL. An app purpose coding guide assessed whether the app was relevant to vaping (yes or no), the intended substance (nicotine/tobacco, cannabis/THC, or unclear), and the purpose of the app.

The app purpose coding guide was initiated by MM and developed iteratively among the investigators. Any information on the app store webpage, including description and screenshots, was used for coding the app purpose and content. Apps could have multiple purposes. After initial coding, the investigators discussed tabulated results and aspects of the coding guide that were unclear, then refined the coding guide based on common patterns and any confusing aspects of the coding guide.

The second coding guide was then applied to 20% of the apps, which were triple coded. Each investigator coded 21 apps (7 recoded from their first pass, with 7 each from the 2 other coders). Unanimous agreement across all 3 investigators was then assessed for the app’s relevance to vaping (yes or no), intended substance (nicotine/cannabis/unclear), and each of 10 potential purpose categories (coils, e-liquids, mods, shopping, games, social, device, quitting smoking, quitting vaping, other). If an app was determined to be about something other than vaping, its purpose was not coded. Agreement before consensus was 19/21 (90%) for relevance to vaping, 12/17 (71%) intended substance, and 14-17/17 (82%-100%) for the purpose categories. Individual apps for which at least one investigator had a discordant code were discussed until consensus was reached.

A third and final coding guide was applied to the full list of apps. Of note, the coding guide was clarified so that apps referencing heat-not-burn were coded as vaping-related. While tobacco heat-not-burn devices are considered distinct from vaping devices, heat-not-burn of cannabis flower is often considered vaping, and neither involves combustion. Additionally, the group could not determine the purpose of 6 apps from the Google Play Store descriptions. These apps were downloaded and evaluated using additional information from the downloaded app itself.

The final 8 purpose categories were do-it-yourself (DIY) coils, DIY e-liquids, shopping, entertainment, social, device, quitting smoking, and quitting vaping. Summary percentages and means for app metadata and purposes were calculated.

Next, apps within each of the 8 purpose categories were ranked by total number of installs, and the top 2 to 5 apps per category were downloaded for review. Instead of ranking the most popular apps, we ranked apps by popularity within categories, so that apps with less overall popularity but potentially important purposes would be included. The number of apps selected per category varied due to ties in the reported number of installs. Because apps could have multiple categories, this procedure resulted in a list of 18 apps with 10 to 1 million downloads each. Three of these apps disappeared from the store before they could be downloaded for review; one could be replaced with a premium version of the same app. A total of 16 apps were downloaded onto 2 Samsung Galaxy Tab A tablets and a Google Pixel 2 smartphone.

A random selection (6/16, 38%) was reviewed by all 3 investigators. Discrepancies were discussed, and the coding guide was updated. The final coding guide included evaluations of whether the content of the downloaded app matched the purpose category (yes or no) and whether it had the following types of information (yes or no): information about harms of vaping (eg, lung injury, nicotine dependence), information about safer vaping or DIY device use (eg, how to prevent explosions), and information about harms of smoking (eg, nicotine dependence, cardiovascular harms, cancer risk). For apps coded as being intended for quitting smoking or quitting vaping, the presence of a tracking feature in the downloaded app (eg, tracking days without smoking or vaping, tracking money saved) was noted (yes or no). For apps coded as pairing with devices, the presence of features for tracking temperature, dosage, or device locking was noted (yes or no). Differences between investigators were discussed until consensus was reached.

Finally, the Mobile App Rating Scale (MARS) was applied to all 16 downloaded apps. The MARS is 23-item multidimensional measure for rating the quality of mobile health apps, with 5 subscales in the areas of engagement (5 items), functionality (4 items), aesthetics (3 items), information (7 items, including affiliation of developer), and subjective rating (4 items) [59]. Subjective rating was not applied as these items involve hypothetical personal use. Each item was rated on a scale of 1 to 5, and ratings that differed by more than 2 points between investigators (eg, ratings of 1 vs 4 or 2 vs 5) were discussed. Investigators could adjust their ratings after discussion before averaging scores and did so 7 times across the 5 discussed apps. Some information items were rated as not applicable (N/A) and were excluded from average score calculations. For example, the item about meeting goals would be rated as N/A if the app purpose was not related to quitting smoking or vaping, and the visual information item would be rated as N/A if the app only contained text. Other smaller discrepancies were averaged without further discussion. Average ratings were calculated for each of the 4 subscales (engagement, functionality, aesthetics, information) and then averaged across the 3 investigator ratings. A final score was averaged for all downloaded apps and within each category.

Of the top 100 apps captured by our search in December 2019, 79 (79%) were determined to be about vaping and were coded for purpose and other features. It was determined that 15 of the 21 apps not about vaping referred to cigarette smoking or only smoking cessation. An additional 6 apps could not be found again during content coding in January 2020 and February 2020 and were removed from the analysis. There were 4 sets of apps with both free and pro versions. The pro version typically cost money to download and did not have advertisements or had additional features. All app descriptions were in English, though some apps appeared to have content in other languages.

Of the 79 apps, most apps were free (60, 76%) or free with in-app purchases (13, 16%); 6 apps cost between US $0.99 and US $6.99 to download. Over half the apps had in-app advertisements (45, 57%). Only one-third of apps (36, 33%) were rated as Mature 17+, while the rest were rated as Teen (13, 16%) or Everyone (40, 51%) (Table 2).

There were 15 different Google Play Store–provided categories displayed with the app description, with the most common being tools and lifestyle, followed by health and fitness, simulation, and social. As one indicator of popularity, app downloads or installations ranged from at least 10 installs to over 1 million installs, with the first 10 apps presented in the search having at least 10,000 installs. As another set of popularity indicators, the average for the 68 apps with ratings was 4.0 stars, with a range of 2.5-5.0 stars by an average of 770 raters.

Investigator-coded app purposes are described in Table 2. The most common investigator-coded app purposes were creating DIY vaping e-liquids (28/79, 35%) and coils (25/79, 32%); 16 apps (16/79, 20%) were coded as for both creating coils and e-liquids. The next most common purposes were entertainment or games (19/79, 24%), social networking with other app users (16/79, 20%), and shopping (15/79, 19%). Social and shopping also tended to be co-occurring purposes (10/79, %). Finally, apps to help people quit smoking (8/79, 10%), directly control vaping devices (4/79, 5%), and quit vaping were less common (2/79, 3%). Out of 8 apps with quitting smoking features (typically a “cigarettes avoided” widget), 7 were also coded as DIY e-liquid or coil purposes. The majority referred to vaping with nicotine (43/79, 54%). Few apps were intended for cannabis (3/79, 4%), and 1 app referred to both nicotine and cannabidiol from cannabis (1/79, 1%). The intended substance was unclear in the remaining apps (31/79, 39%) (Table 3).

Because apps could fall into multiple categories, we did not statistically test differences in star ratings. However, we noted that the highest rated apps by users were in the DIY e-liquids and quitting smoking and vaping categories, while the lowest were in the device category. The most popular apps by number of installations were in the DIY e-liquids, DIY coils, and entertainment categories.

The majority of downloaded apps (14/16, 88%) matched the descriptions in the app store (Table 4). The exceptions were that 1 app coded as having quitting smoking features did not have any such features, and 1 app that was coded as not having shopping features did, in fact, have links to shopping through the app. Few apps had information about harms of vaping (4/16, 25%), safer vaping (3/16, 19%), or harms of smoking (2/16, 13%). When they did have such information, it was often difficult to find. Of note, 2 of the entertainment apps had a feature where an avatar would cough when vaping too much, which may normalize moderation in vaping; 5 apps had tracking features, which mainly recorded days passed since a user-provided quit-smoking date; and 2 of the apps with tracking features also displayed money saved and health benefits. Both device apps appeared to have temperature controls, but did not have dosage or locking settings visible, though these may have become apparent once paired with a device.

Overall, the 16 downloaded apps had a mean MARS score of 3.63, with a highest mean subscore for functionality (MARS score: mean 4.13) and lowest mean subscore for engagement (MARS score: mean 3.36). Within the subtypes of purposes, the highest mean MARS scores were for social, shopping, and device apps, and the lowest mean scores were for the quitting smoking and DIY coils apps.