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

Identifying Marijuana Use Behaviors Among Youth Experiencing Homelessness Using a Machine Learning-Based Framework: Development and Evaluation Study

BACKGROUND

Youth experiencing homelessness face substance use problems disproportionately compared to other youth. A study found that 69% of youth experiencing homelessness meet the criteria for dependence on at least 1 substance, compared to 1.8% for all US adolescents. In addition, they experience major structural and social inequalities, which further undermine their ability to receive the care they need.

OBJECTIVE

The goal of this study was to develop a machine learning-based framework that uses the social media content (posts and interactions) of youth experiencing homelessness to predict their substance use behaviors (ie, the probability of using marijuana). With this framework, social workers and care providers can identify and reach out to youth experiencing homelessness who are at a higher risk of substance use.

METHODS

We recruited 133 young people experiencing homelessness at a nonprofit organization located in a city in the western United States. After obtaining their consent, we collected the participants' social media conversations for the past year before they were recruited, and we asked the participants to complete a survey on their demographic information, health conditions, sexual behaviors, and substance use behaviors. Building on the social sharing of emotions theory and social support theory, we identified important features that can potentially predict substance use. Then, we used natural language processing techniques to extract such features from social media conversations and reactions and built a series of machine learning models to predict participants' marijuana use.

RESULTS

We evaluated our models based on their predictive performance as well as their conformity with measures of fairness. Without predictive features from survey information, which may introduce sex and racial biases, our machine learning models can reach an area under the curve of 0.72 and an accuracy of 0.81 using only social media data when predicting marijuana use. We also evaluated the false-positive rate for each sex and age segment.

CONCLUSIONS

We showed that textual interactions among youth experiencing homelessness and their friends on social media can serve as a powerful resource to predict their substance use. The framework we developed allows care providers to allocate resources efficiently to youth experiencing homelessness in the greatest need while costing minimal overhead. It can be extended to analyze and predict other health-related behaviors and conditions observed in this vulnerable community.

Citizens' digital footprints to support health promotion at the local level-PUHTI study, Finland

BACKGROUND

We aimed to explore to the possibilities of utilizing automatically accumulating data on health-owned for example by local companies and non-governmental organizations-to complement traditional health data sources in health promotion work at the local level.

METHODS

Data for the PUHTI study consisted of postal code level information on sport license holders, drug purchase and sales advertisements in a TOR online underground marketplace, and grocery sales in Tampere. Additionally, open population register data were utilized. An interactive reporting tool was prepared to show the well-being profile for each postal code area. Feedback from the tool's end-users was collected in interviews.

RESULTS

The study showed that buying unhealthy food and alcohol, selling or buying drugs, and participating in organized sport activities differed by postal code areas according to its socioeconomic profile in the city of Tampere. The health and well-being planners and managers of Tampere found that the new type of data brought added value for the health promotion work at the local level. They perceived the interactive reporting tool as a good tool for planning, managing, allocating resources and preparing forecasts.

CONCLUSIONS

Traditional health data collection methods-administrative registers and health surveys-are the cornerstone of local health promotion work. Digital footprints, including data accumulated about people's everyday lives outside the health service system, can provide additional information on health behaviour for various population groups. Combining new sources with traditional health data opens a new perspective for health promotion work at local and regional levels.

Potential of artificial intelligence in injury prevention research and practice

There is increasing interest and use of artificial Intelligence algorithms and methods in biomedical research and practice, particularly as the technology has made significant advances in the past decade and become more accessible to more disciplines. This editorial briefly reviews this technology and its potential for injury prevention research and practice, proposing ways that it can be used to advance the discipline, as well as the potential pitfalls, concerns and biases that accompany it.

Investigating Substance Use via Reddit: Systematic Scoping Review

BACKGROUND

Reddit's (Reddit Inc) large user base, diverse communities, and anonymity make it a useful platform for substance use research. Despite a growing body of literature on substance use on Reddit, challenges and limitations must be carefully considered. However, no systematic scoping review has been conducted on the use of Reddit as a data source for substance use research.

OBJECTIVE

This review aims to investigate the use of Reddit for studying substance use by examining previous studies' objectives, reasons, limitations, and methods for using Reddit. In addition, we discuss the implications and contributions of previous studies and identify gaps in the literature that require further attention.

METHODS

A total of 7 databases were searched using keyword combinations including Reddit and substance-related keywords in April 2022. The initial search resulted in 456 articles, and 227 articles remained after removing duplicates. All included studies were peer reviewed, empirical, available in full text, and pertinent to Reddit and substance use, and they were all written in English. After screening, 60 articles met the eligibility criteria for the review, with 57 articles identified from the initial database search and 3 from the ancestry search. A codebook was developed, and qualitative content analysis was performed to extract relevant evidence related to the research questions.

RESULTS

The use of Reddit for studying substance use has grown steadily since 2015, with a sharp increase in 2021. The primary objective was to identify tendencies and patterns in various types of substance use discussions (52/60, 87%). Reddit was also used to explore unique user experiences, propose methodologies, investigate user interactions, and develop interventions. A total of 9 reasons for using Reddit to study substance use were identified, such as the platform's anonymity, its widespread popularity, and the explicit topics of subreddits. However, 7 limitations were noted, including the platform's low representativeness of the general population with substance use and the lack of demographic information. Most studies use application programming interfaces for data collection and quantitative approaches for analysis, with few using qualitative approaches. Machine learning algorithms are commonly used for natural language processing tasks. The theoretical, methodological, and practical implications and contributions of the included articles are summarized and discussed. The most prevalent practical implications are investigating prevailing topics in Reddit discussions, providing recommendations for clinical practices and policies, and comparing Reddit discussions on substance use across various sources.

CONCLUSIONS

This systematic scoping review provides an overview of Reddit's use as a data source for substance use research. Although the limitations of Reddit data must be considered, analyzing them can be useful for understanding patterns and user experiences related to substance use. Our review also highlights gaps in the literature and suggests avenues for future research.

WLCD: a dataset of lifestyle in relation with women's cancer

OBJECTIVES

Social media text mining has been widely used to extract information about the experiences and needs of patients regarding various diseases, especially cancer. Understanding these issues is necessary for further management in primary care. Researchers have identified that lifestyle factors such as diet, exercise, alcohol, and Smoking are associated with cancer risks, particularly women's cancer. Considering the growing trend in the global burden of women's cancer, it is essential to monitor up-to-date data sources using text mining.

DATA DESCRIPTION

We have prepared six independent datasets regarding lifestyle components and women's cancer: (1) a dataset of nutrition containing 10,161 tweets; (2) a dataset of exercise containing 9412 tweets; (3) a dataset of alcohol containing 2132 tweets; (4) a dataset of Smoking containing 4316 tweets; and (5) a dataset of lifestyle (term) containing 1861 tweets. We also construct an additional dataset: (6) a dataset by summing other components containing 27,882 tweets. These data are provided to discover people's perspectives, knowledge, and experiences regarding lifestyle and women's cancer. Hence, it should be valuable for healthcare providers to develop more efficient patient management approaches.

Digital health technologies for peripartum depression management among low-socioeconomic populations: perspectives from patients, providers, and social media channels

BACKGROUND

Peripartum Depression (PPD) affects approximately 10-15% of perinatal women in the U.S., with those of low socioeconomic status (low-SES) more likely to develop symptoms. Multilevel treatment barriers including social stigma and not having appropriate access to mental health resources have played a major role in PPD-related disparities. Emerging advances in digital technologies and analytics provide opportunities to identify and address access barriers, knowledge gaps, and engagement issues. However, most market solutions for PPD prevention and management are produced generically without considering the specialized needs of low-SES populations. In this study, we examine and portray the information and technology needs of low-SES women by considering their unique perspectives and providers' current experiences. We supplement our understanding of women's needs by harvesting online social discourse in PPD-related forums, which we identify as valuable information resources among these populations.

METHODS

We conducted (a) 2 focus groups (n = 9), (b) semi-structured interviews with care providers (n = 9) and low SES women (n = 10), and (c) secondary analysis of online messages (n = 1,424). Qualitative data were inductively analyzed using a grounded theory approach.

RESULTS

A total of 134 open concepts resulted from patient interviews, 185 from provider interviews, and 106 from focus groups. These revealed six core themes for PPD management, including "Use of Technology/Features", "Access to Care", and "Pregnancy Education". Our social media analysis revealed six PPD topics of importance in online messages, including "Physical and Mental Health" (n = 725 messages), and "Social Support" (n = 674).

CONCLUSION

Our data triangulation allowed us to analyze PPD information and technology needs at different levels of granularity. Differences between patients and providers included a focus from providers on needing better support from administrative staff, as well as better PPD clinical decision support. Our results can inform future research and development efforts to address PPD health disparities.

Discerning conversational context in online health communities for personalized digital behavior change solutions using Pragmatics to Reveal Intent in Social Media (PRISM) framework

BACKGROUND

Online health communities (OHCs) have emerged as prominent platforms for behavior modification, and the digitization of online peer interactions has afforded researchers with unique opportunities to model multilevel mechanisms that drive behavior change. Existing studies, however, have been limited by a lack of methods that allow the capture of conversational context and socio-behavioral dynamics at scale, as manifested in these digital platforms.

OBJECTIVE

We develop, evaluate, and apply a novel methodological framework, Pragmatics to Reveal Intent in Social Media (PRISM), to facilitate granular characterization of peer interactions by combining multidimensional facets of human communication.

METHODS

We developed and applied PRISM to analyze peer interactions (N = 2.23 million) in QuitNet, an OHC for tobacco cessation. First, we generated a labeled set of peer interactions (n = 2,005) through manual annotation along three dimensions: communication themes (CTs), behavior change techniques (BCTs), and speech acts (SAs). Second, we used deep learning models to apply our qualitative codes at scale. Third, we applied our validated model to perform a retrospective analysis. Finally, using social network analysis (SNA), we portrayed large-scale patterns and relationships among the aforementioned communication dimensions embedded in peer interactions in QuitNet.

RESULTS

Qualitative analysis showed that the themes of social support and behavioral progress were common. The most used BCTs were feedback and monitoring and comparison of behavior, and users most commonly expressed their intentions using SAs-expressive and emotion. With additional in-domain pre-training, bidirectional encoder representations from Transformers (BERT) outperformed other deep learning models on the classification tasks. Content-specific SNA revealed that users' engagement or abstinence status is associated with the prevalence of various categories of BCTs and SAs, which also was evident from the visualization of network structures.

CONCLUSIONS

Our study describes the interplay of multilevel characteristics of online communication and their association with individual health behaviors.

Assessing vulnerability to psychological distress during the COVID-19 pandemic through the analysis of microblogging content

In recent years we have witnessed a growing interest in the analysis of social media data under different perspectives, since these online platforms have become the preferred tool for generating and sharing content across different users organized into virtual communities, based on their common interests, needs, and perceptions. In the current study, by considering a collection of social textual contents related to COVID-19 gathered on the Twitter microblogging platform in the period between August and December 2020, we aimed at evaluating the possible effects of some critical factors related to the pandemic on the mental well-being of the population. In particular, we aimed at investigating potential lexicon identifiers of vulnerability to psychological distress in digital social interactions with respect to distinct COVID-related scenarios, which could be "at risk" from a psychological discomfort point of view. Such scenarios have been associated with peculiar topics discussed on Twitter. For this purpose, two approaches based on a "top-down" and a "bottom-up" strategy were adopted. In the top-down approach, three potential scenarios were initially selected by medical experts, and associated with topics extracted from the Twitter dataset in a hybrid unsupervised-supervised way. On the other hand, in the bottom-up approach, three topics were extracted in a totally unsupervised way capitalizing on a Twitter dataset filtered according to the presence of keywords related to vulnerability to psychological distress, and associated with at-risk scenarios. The identification of such scenarios with both approaches made it possible to capture and analyze the potential psychological vulnerability in critical situations.

Medical Education e-Professionalism (MEeP) framework; from conception to development

BACKGROUND

Medical professionalism education intends to produce virtuous and humanistic healthcare professionals who demonstrate perseverance and professional integrity. However, today's medicine has embodied a mammoth transformation of medical practice towards sns and the digital realm. Such paradigm shift has challenged the medical professional's values, behaviors, and identities, and the distinct boundaries between personal and professional lives are blurred. This study aims to develop a framework for healthcare professionals coping with the challenges of medical professionalism in the digital realm.

METHODS

We followed a systematic approach for the development of a framework about e-professionalism. Qualitative data was collected from a systematic review and a delphi study, while quantitative data was collected by administering a validated questionnaire social networking sites for medical education (snsme). Subsequently, categorization of the selected data and identifying concepts, deconstruction and further categorizing concepts (philosophical triangulation), integration of concepts (theoretical triangulation), and synthesis and resynthesis of concepts were performed.

RESULTS

The initial process yielded six overlapping concepts from personal, professional, character (implicit) and characteristic (explicit) domains: environment, behavior, competence, virtues, identity, and mission. Further integration of data was done for the development of the medical education e-professionalism (meep) framework with a central concept of a commitment to mission. The mission showed deep connections with values (conformity, beneficence, universalism, and integrity), behaviours (communication, self-awareness, tolerance, power), and identity (reflection, conscientiousness, self-directed, self-actualization). The data demonstrated that all medical professionals require updated expertise in sns participation.

CONCLUSION

The meep framework recognises a mission-based social contract by the medical community. This mission is largely driven by professional values, behaviors and identity. Adherence to digital standards, accountability, empathy, sensitivity, and commitment to society are essential elements of the meep framework.

Obesity-Related Communication in Digital Chinese News From Mainland China, Hong Kong, and Taiwan: Automated Content Analysis

BACKGROUND

The fact that the number of individuals with obesity has increased worldwide calls into question media efforts for informing the public. This study attempts to determine the ways in which the mainstream digital news covers the etiology of obesity and diseases associated with the burden of obesity.

OBJECTIVE

The dual objectives of this study are to obtain an understanding of what the news reports on obesity and to explore meaning in data by extending the preconceived grounded theory.

METHODS

The 10 years of news text from 2010 to 2019 compared the development of obesity-related coverage and its potential impact on its perception in Mainland China, Hong Kong, and Taiwan. Digital news stories on obesity along with affliction and inferences in 9 Chinese mainstream newspapers were sampled. An automatic content analysis tool, DiVoMiner was proposed. This computer-aided platform is designed to organize and filter large sets of data on the basis of the patterns of word occurrence and term discovery. Another programming language, Python 3, was used to explore connections and patterns created by the aggregated interactions.

RESULTS

A total of 30,968 news stories were identified with increasing attention since 2016. The highest intensity of newspaper coverage of obesity communication was observed in Taiwan. Overall, a stronger focus on 2 shared causative attributes of obesity is on stress (n=4483, 33.0%) and tobacco use (n=3148, 23.2%). The burdens of obesity and cardiovascular diseases are implied to be the most, despite the aggregated interaction of edge centrality showing the highest link between the "cancer" and obesity. This study goes beyond traditional journalism studies by extending the framework of computational and customizable web-based text analysis. This could set a norm for researchers and practitioners who work on data projects largely for an innovative attempt.

CONCLUSIONS

Similar to previous studies, the discourse between the obesity epidemic and personal afflictions is the most emphasized approach. Our study also indicates that the inclination of blaming personal attributes for health afflictions potentially limits social and governmental responsibility for addressing this issue.

Analysis of Online Peripartum Depression Communities: Application of Multilabel Text Classification Techniques to Inform Digitally-Mediated Prevention and Management

Peripartum depression (PPD) is a significant public health problem, yet many women who experience PPD do not receive adequate treatment. In many cases, this is due to social stigmas surrounding PPD that prevent women from disclosing their symptoms to their providers. Examples of these are fear of being labeled a “bad mother,” or having misinformed expectations regarding motherhood. Online forums dedicated to PPD can provide a practical setting where women can better manage their mental health in the peripartum period. Data from such forums can be systematically analyzed to understand the technology and information needs of women experiencing PPD. However, deeper insights are needed on how best to translate information derived from online forum data into digital health features. In this study, we aim to adapt a digital health development framework, Digilego, toward translation of our results from social media analysis to inform digital features of a mobile intervention that promotes PPD prevention and self-management. The first step in our adaption was to conduct a user need analysis through semi-automated analysis of peer interactions in two highly popular PPD online forums: What to Expect and BabyCenter. This included the development of a machine learning pipeline that allowed us to automatically classify user post content according to major communication themes that manifested in the forums. This was followed by mapping the results of our user needs analysis to existing behavior change and engagement optimization models. Our analysis has revealed major themes being discussed by users of these online forums- family and friends, medications, symptom disclosure, breastfeeding, and social support in the peripartum period. Our results indicate that Random Forest was the best performing model in automatic text classification of user posts, when compared to Support Vector Machine, and Logistic Regression models. Computerized text analysis revealed that posts had an average length of 94 words, and had a balance between positive and negative emotions. Our Digilego-powered theory mapping also indicated that digital platforms dedicated to PPD prevention and management should contain features ranging from educational content on practical aspects of the peripartum period to inclusion of collaborative care processes that support shared decision making, as well as forum moderation strategies to address issues with cyberbullying.

Deep Learning for Identification of Alcohol-Related Content on Social Media (Reddit and Twitter): Exploratory Analysis of Alcohol-Related Outcomes

BACKGROUND

Many social media studies have explored the ability of thematic structures, such as hashtags and subreddits, to identify information related to a wide variety of mental health disorders. However, studies and models trained on specific themed communities are often difficult to apply to different social media platforms and related outcomes. A deep learning framework using thematic structures from Reddit and Twitter can have distinct advantages for studying alcohol abuse, particularly among the youth in the United States.

OBJECTIVE

This study proposes a new deep learning pipeline that uses thematic structures to identify alcohol-related content across different platforms. We apply our method on Twitter to determine the association of the prevalence of alcohol-related tweets with alcohol-related outcomes reported from the National Institute of Alcoholism and Alcohol Abuse, Centers for Disease Control Behavioral Risk Factor Surveillance System, county health rankings, and the National Industry Classification System.

METHODS

The Bidirectional Encoder Representations From Transformers neural network learned to classify 1,302,524 Reddit posts as either alcohol-related or control subreddits. The trained model identified 24 alcohol-related hashtags from an unlabeled data set of 843,769 random tweets. Querying alcohol-related hashtags identified 25,558,846 alcohol-related tweets, including 790,544 location-specific (geotagged) tweets. We calculated the correlation between the prevalence of alcohol-related tweets and alcohol-related outcomes, controlling for confounding effects of age, sex, income, education, and self-reported race, as recorded by the 2013-2018 American Community Survey.

RESULTS

Significant associations were observed: between alcohol-hashtagged tweets and alcohol consumption (P=.01) and heavy drinking (P=.005) but not binge drinking (P=.37), self-reported at the metropolitan-micropolitan statistical area level; between alcohol-hashtagged tweets and self-reported excessive drinking behavior (P=.03) but not motor vehicle fatalities involving alcohol (P=.21); between alcohol-hashtagged tweets and the number of breweries (P<.001), wineries (P<.001), and beer, wine, and liquor stores (P<.001) but not drinking places (P=.23), per capita at the US county and county-equivalent level; and between alcohol-hashtagged tweets and all gallons of ethanol consumed (P<.001), as well as ethanol consumed from wine (P<.001) and liquor (P=.01) sources but not beer (P=.63), at the US state level.

CONCLUSIONS

Here, we present a novel natural language processing pipeline developed using Reddit's alcohol-related subreddits that identify highly specific alcohol-related Twitter hashtags. The prevalence of identified hashtags contains interpretable information about alcohol consumption at both coarse (eg, US state) and fine-grained (eg, metropolitan-micropolitan statistical area level and county) geographical designations. This approach can expand research and deep learning interventions on alcohol abuse and other behavioral health outcomes.