Next generation media monitoring: Global coverage of electronic nicotine delivery systems (electronic cigarettes) on Bing, Google and Twitter, 2013-2018

Machine learning based assessment of preclinical health questionnaires

BACKGROUND

Within modern health systems, the possibility of accessing a large amount and a variety of data related to patients' health has increased significantly over the years. The source of this data could be mobile and wearable electronic systems used in everyday life, and specialized medical devices. In this study we aim to investigate the use of modern Machine Learning (ML) techniques for preclinical health assessment based on data collected from questionnaires filled out by patients.

METHOD

To identify the health conditions of pregnant women, we developed a questionnaire that was distributed in three maternity hospitals in the Mureș County, Romania. In this work we proposed and developed an ML model for pattern detection in common risk assessment based on data extracted from questionnaires.

RESULTS

Out of the 1278 women who answered the questionnaire, 381 smoked before pregnancy and only 216 quit smoking during the period in which they became pregnant. The performance of the model indicates the feasibility of the solution, with an accuracy of 98 % confirmed for the considered case study.

CONCLUSION

The proposed solution offers a simple and efficient way to digitize questionnaire data and to analyze the data through a reduced computational effort, both in terms of memory and computing power used.

Machine learning applications in tobacco research: a scoping review

OBJECTIVE

Identify and review the body of tobacco research literature that self-identified as using machine learning (ML) in the analysis.

DATA SOURCES

MEDLINE, EMABSE, PubMed, CINAHL Plus, APA PsycINFO and IEEE Xplore databases were searched up to September 2020. Studies were restricted to peer-reviewed, English-language journal articles, dissertations and conference papers comprising an empirical analysis where ML was identified to be the method used to examine human experience of tobacco. Studies of genomics and diagnostic imaging were excluded.

STUDY SELECTION

Two reviewers independently screened the titles and abstracts. The reference list of articles was also searched. In an iterative process, eligible studies were classified into domains based on their objectives and types of data used in the analysis.

DATA EXTRACTION

Using data charting forms, two reviewers independently extracted data from all studies. A narrative synthesis method was used to describe findings from each domain such as study design, objective, ML classes/algorithms, knowledge users and the presence of a data sharing statement. Trends of publication were visually depicted.

DATA SYNTHESIS

74 studies were grouped into four domains: ML-powered technology to assist smoking cessation (n=22); content analysis of tobacco on social media (n=32); smoker status classification from narrative clinical texts (n=6) and tobacco-related outcome prediction using administrative, survey or clinical trial data (n=14). Implications of these studies and future directions for ML researchers in tobacco control were discussed.

CONCLUSIONS

ML represents a powerful tool that could advance the research and policy decision-making of tobacco control. Further opportunities should be explored.

Evolution of tobacco products: recent history and future directions

Declines in cigarette smoking prevalence in many countries and the consolidation of the tobacco industry have prompted the introduction of other forms of nicotine delivery. These include electronic nicotine delivery systems (ENDS), heated tobacco products (HTPs) and oral nicotine products (ONPs). Evolving over time, some of these products now deliver nicotine at levels comparable to cigarettes and may serve as effective substitutes for smokers. However, certain products, especially ENDS like JUUL, have also appealed to youth and non-smokers, prompting concerns about expanding nicotine use (and potentially nicotine addiction). The tobacco industry could shift away from primarily promoting cigarettes to promoting ENDS, HTPs and/or ONPs, though at this time it continues to heavily promote cigarettes in low and middle-income countries. Differing regulatory regimes may place upward and downward pressures on both cigarettes and these newer products in terms of population use, and may ultimately drive the extent to which cigarettes are or are not displaced by ENDS, HTPs and/or ONPs in the coming decade.

From the Deeming Rule to JUUL-US News Coverage of Electronic Cigarettes, 2015-2018

INTRODUCTION

News media may influence public perceptions and attitudes about electronic cigarettes (e-cigarettes), which may influence product use and attitudes about their regulation. The purpose of this study is to describe trends in US news coverage of e-cigarettes during a period of evolving regulation, science, and trends in the use of e-cigarettes.

METHODS

We conducted a content analysis of e-cigarette topics and themes covered in US news articles from 2015 to 2018. Online news databases (Access World News, Factiva) were used to obtain US news articles from the top 34 circulating newspapers, four national wire services, and five leading online news sources.

RESULTS

The number of articles increased by 75.4% between 2015 and 2018 (n = 1609). Most articles focused on policy/regulation (43.5%) as a main topic, followed by health effects (22.3%) and prevalence/trends (17.9%). Discussion about flavor bans quadrupled (6.1% to 24.6%) and discussion of youth e-cigarette use was most prevalent (58.4%) in 2018, coinciding with an increase in coverage about JUUL. JUUL was mentioned in 50.8% of 2018 articles. Across years, articles more frequently mentioned e-cigarette risks (70%) than potential benefits (37.3%).

CONCLUSIONS

E-cigarettes continue to be a newsworthy topic, with coverage both reflecting numerous changes and events over time, and providing repeated opportunities for informing the public and policymakers about these novel products. Future research should continue to track how discourse changes over time and assess its potential influence on e-cigarette perceptions and policy changes.

IMPLICATIONS

E-cigarette news coverage in the United States increased between 2015 and 2018 and predominantly focused on policy and regulation. Notable spikes in volume were associated with some but not all major e-cigarette events, including the FDA's deeming rule, Surgeon General's report, and release of the National Youth Tobacco Survey data in 2018. Coverage of the 2018 National Academy of Medicine, Engineering, and Sciences report on the Public Health Consequences of E-cigarettes received minimal news coverage. The high volume in 2018 was driven in large part by coverage of the e-cigarette brand JUUL; over half of news articles in 2018 referenced JUUL specifically.

Social Media as a Research Tool (SMaaRT) for Risky Behavior Analytics: Methodological Review

BACKGROUND

Modifiable risky health behaviors, such as tobacco use, excessive alcohol use, being overweight, lack of physical activity, and unhealthy eating habits, are some of the major factors for developing chronic health conditions. Social media platforms have become indispensable means of communication in the digital era. They provide an opportunity for individuals to express themselves, as well as share their health-related concerns with peers and health care providers, with respect to risky behaviors. Such peer interactions can be utilized as valuable data sources to better understand inter-and intrapersonal psychosocial mediators and the mechanisms of social influence that drive behavior change.

OBJECTIVE

The objective of this review is to summarize computational and quantitative techniques facilitating the analysis of data generated through peer interactions pertaining to risky health behaviors on social media platforms.

METHODS

We performed a systematic review of the literature in September 2020 by searching three databases-PubMed, Web of Science, and Scopus-using relevant keywords, such as "social media," "online health communities," "machine learning," "data mining," etc. The reporting of the studies was directed by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Two reviewers independently assessed the eligibility of studies based on the inclusion and exclusion criteria. We extracted the required information from the selected studies.

RESULTS

The initial search returned a total of 1554 studies, and after careful analysis of titles, abstracts, and full texts, a total of 64 studies were included in this review. We extracted the following key characteristics from all of the studies: social media platform used for conducting the study, risky health behavior studied, the number of posts analyzed, study focus, key methodological functions and tools used for data analysis, evaluation metrics used, and summary of the key findings. The most commonly used social media platform was Twitter, followed by Facebook, QuitNet, and Reddit. The most commonly studied risky health behavior was nicotine use, followed by drug or substance abuse and alcohol use. Various supervised and unsupervised machine learning approaches were used for analyzing textual data generated from online peer interactions. Few studies utilized deep learning methods for analyzing textual data as well as image or video data. Social network analysis was also performed, as reported in some studies.

CONCLUSIONS

Our review consolidates the methodological underpinnings for analyzing risky health behaviors and has enhanced our understanding of how social media can be leveraged for nuanced behavioral modeling and representation. The knowledge gained from our review can serve as a foundational component for the development of persuasive health communication and effective behavior modification technologies aimed at the individual and population levels.

E-Cigarette Advocates on Twitter: Content Analysis of Vaping-Related Tweets

Background

As the majority of Twitter content is publicly available, the platform has become a rich data source for public health surveillance, providing insights into emergent phenomena, such as vaping. Although there is a growing body of literature that has examined the content of vaping-related tweets, less is known about the people who generate and disseminate these messages and the role of e-cigarette advocates in the promotion of these devices.

Objective

This study aimed to identify key conversation trends and patterns over time, and discern the core voices, message frames, and sentiment surrounding e-cigarette discussions on Twitter.

Methods

A random sample of data were collected from Australian Twitter users who referenced at least one of 15 identified e-cigarette related keywords during 2012, 2014, 2016, or 2018. Data collection was facilitated by TrISMA (Tracking Infrastructure for Social Media Analysis) and analyzed by content analysis.

Results

A sample of 4432 vaping-related tweets posted and retweeted by Australian users was analyzed. Positive sentiment (3754/4432, 84.70%) dominated the discourse surrounding e-cigarettes, and vape retailers and manufacturers (1161/4432, 26.20%), the general public (1079/4432, 24.35%), and e-cigarette advocates (1038/4432, 23.42%) were the most prominent posters. Several tactics were used by e-cigarette advocates to communicate their beliefs, including attempts to frame e-cigarettes as safer than traditional cigarettes, imply that federal government agencies lack sufficient competence or evidence for the policies they endorse about vaping, and denounce as propaganda “gateway” claims of youth progressing from e-cigarettes to combustible tobacco. Some of the most common themes presented in tweets were advertising or promoting e-cigarette products (2040/4432, 46.03%), promoting e-cigarette use or intent to use (970/4432, 21.89%), and discussing the potential of e-cigarettes to be used as a smoking cessation aid or tobacco alternative (716/4432, 16.16%), as well as the perceived health and safety benefits and consequences of e-cigarette use (681/4432, 15.37%).

Conclusions

Australian Twitter content does not reflect the country’s current regulatory approach to e-cigarettes. Rather, the conversation on Twitter generally encourages e-cigarette use, promotes vaping as a socially acceptable practice, discredits scientific evidence of health risks, and rallies around the idea that e-cigarettes should largely be outside the bounds of health policy. The one-sided nature of the discussion is concerning, as is the lack of disclosure and transparency, especially among vaping enthusiasts who dominate the majority of e-cigarette discussions on Twitter, where it is unclear if comments are endorsed, sanctioned, or even supported by the industry.

News coverage of the E-cigarette, or Vaping, product use Associated Lung Injury (EVALI) outbreak and internet searches for vaping cessation

BACKGROUND

In the latter half of 2019, an outbreak of pulmonary disease in the USA resulted in 2807 hospitalisations and 68 deaths, as of 18 February 2020. Given the severity of the outbreak, we assessed whether articles during the outbreak era more frequently warned about the dangers of vaping and whether internet searches for vaping cessation increased.

METHODS

Using Tobacco Watcher, a media monitoring platform that automatically identifies and categorises news articles from sources across the globe, we obtained all articles that (a) discussed the outbreak and (b) primarily warned about the dangers of vaping. We obtained internet search trends originating from the USA that mentioned 'quit' or 'stop' and 'e cig(s),' 'ecig(s),' 'e-cig(s),' 'e cigarette(s),' 'e-cigarette(s),' 'electronic cigarette(s),' 'vape(s),' 'vaping' or 'vaper(s)' from Google Trends (eg, 'how do I quit vaping?'). All data were obtained from 1 January 2014 to 18 February 2020 and ARIMA models were used with historical trends to forecast the ratio of observed to expected search volumes during the outbreak era.

RESULTS

News of the vaping-induced pulmonary disease outbreak was first reported on 25 July 2019 with 195 articles, culminating in 44 512 articles by 18 February 2020. On average, news articles warning about the dangers of vaping were 130% (95% prediction interval (PI): -15 to 417) and searches for vaping cessation were 76% (95% PI: 28 to 182) higher than expected levels for the days during the period when the sources of the outbreak were unknown (25 July to 27 September 2019). News and searches stabilised just after the US Centers for Disease Control and Prevention reported that a primary source of the outbreak was an additive used in marijuana vapes on 27 September 2019. In sum, there were 12 286 articles archived in Tobacco Watcher primarily warning about the dangers of vaping and 1 025 000 cessation searches following the outbreak.

CONCLUSION

The vaping-induced pulmonary disease outbreak spawned increased coverage about the dangers of vaping and internet searches for vaping cessation. Resources and strategies that respond to this elevated interest should become a priority among public health leaders.

Recent Advances in Using Natural Language Processing to Address Public Health Research Questions Using Social Media and ConsumerGenerated Data

OBJECTIVE

We present a narrative review of recent work on the utilisation of Natural Language Processing (NLP) for the analysis of social media (including online health communities) specifically for public health applications.

METHODS

We conducted a literature review of NLP research that utilised social media or online consumer-generated text for public health applications, focussing on the years 2016 to 2018. Papers were identified in several ways, including PubMed searches and the inspection of recent conference proceedings from the Association of Computational Linguistics (ACL), the Conference on Human Factors in Computing Systems (CHI), and the International AAAI (Association for the Advancement of Artificial Intelligence) Conference on Web and Social Media (ICWSM). Popular data sources included Twitter, Reddit, various online health communities, and Facebook.

RESULTS

In the recent past, communicable diseases (e.g., influenza, dengue) have been the focus of much social media-based NLP health research. However, mental health and substance use and abuse (including the use of tobacco, alcohol, marijuana, and opioids) have been the subject of an increasing volume of research in the 2016 - 2018 period. Associated with this trend, the use of lexicon-based methods remains popular given the availability of psychologically validated lexical resources suitable for mental health and substance abuse research. Finally, we found that in the period under review "modern" machine learning methods (i.e. deep neural-network-based methods), while increasing in popularity, remain less widely used than "classical" machine learning methods.

Artificial intelligence (AI) and cancer prevention: the potential application of AI in cancer control programming needs to be explored in population laboratories such as COMPASS

Understanding the risk factors that initiate cancer is essential for reducing the future cancer burden. Much of our current cancer control insight is from cohort studies and newer large-scale population laboratories designed to advance the science around precision oncology. Despite their promise for improving diagnosis and treatment outcomes, their current reductionist focus will likely have little impact shifting the cancer burden. However, it is possible that these big data assets can be adapted to have more impact on the future cancer burden through more focus on primary prevention efforts that incorporate artificial intelligence (AI) and machine learning (ML). ML automatically learns patterns and can devise complex models and algorithms that lend themselves to prediction in big data, revealing new unexpected relationships and pathways in a reliable and replicable fashion that otherwise would remain hidden given the complexities of big data. While AI has made big strides in several domains, the potential application in cancer prevention is lacking. As such, this commentary suggests that it may be time to consider the potential of AI within our existing cancer control population laboratories, and provides justification for why some small targeted investments to explore their impact on modelling existing real-time cancer prevention data may be a strategic cancer control opportunity.