Background: COVID-19 is a scientifically and medically novel disease that is not fully understood because it has yet to be consistently and deeply studied. Among the gaps in research on the COVID-19 outbreak, there is a lack of sufficient infoveillance data.
Objective: The aim of this study was to increase understanding of public awareness of COVID-19 pandemic trends and uncover meaningful themes of concern posted by Twitter users in the English language during the pandemic.
Methods: Data mining was conducted on Twitter to collect a total of 107,990 tweets related to COVID-19 between December 13 and March 9, 2020. The analyses included frequency of keywords, sentiment analysis, and topic modeling to identify and explore discussion topics over time. A natural language processing approach and the latent Dirichlet allocation algorithm were used to identify the most common tweet topics as well as to categorize clusters and identify themes based on the keyword analysis.
Results: The results indicate three main aspects of public awareness and concern regarding the COVID-19 pandemic. First, the trend of the spread and symptoms of COVID-19 can be divided into three stages. Second, the results of the sentiment analysis showed that people have a negative outlook toward COVID-19. Third, based on topic modeling, the themes relating to COVID-19 and the outbreak were divided into three categories: the COVID-19 pandemic emergency, how to control COVID-19, and reports on COVID-19.
Conclusions: Sentiment analysis and topic modeling can produce useful information about the trends in the discussion of the COVID-19 pandemic on social media as well as alternative perspectives to investigate the COVID-19 crisis, which has created considerable public awareness. This study shows that Twitter is a good communication channel for understanding both public concern and public awareness about COVID-19. These findings can help health departments communicate information to alleviate specific public concerns about the disease.

/pdf/public-perception-of-the-covid-19-pandemic-on-twitter-4dkjnsodb0.pdf

Public Perception of the COVID-19 Pandemic on Twitter: Sentiment Analysis and Topic Modeling Study.

In December 2019, a novel virus named COVID-19 emerged in the city of Wuhan, China. In early 2020, the COVID-19 virus spread in all continents of the world except Antarctica, causing widespread infections and deaths due to its contagious characteristics and no medically proven treatment. The COVID-19 pandemic has been termed as the most consequential global crisis since the World Wars. The first line of defense against the COVID-19 spread are the non-pharmaceutical measures like social distancing and personal hygiene. The great pandemic affecting billions of lives economically and socially has motivated the scientific community to come up with solutions based on computer-aided digital technologies for diagnosis, prevention, and estimation of COVID-19. Some of these efforts focus on statistical and Artificial Intelligence-based analysis of the available data concerning COVID-19. All of these scientific efforts necessitate that the data brought to service for the analysis should be open source to promote the extension, validation, and collaboration of the work in the fight against the global pandemic. Our survey is motivated by the open source efforts that can be mainly categorized as (a) COVID-19 diagnosis from CT scans, X-ray images, and cough sounds, (b) COVID-19 case reporting, transmission estimation, and prognosis from epidemiological, demographic, and mobility data, (c) COVID-19 emotional and sentiment analysis from social media, and (d) knowledge-based discovery and semantic analysis from the collection of scholarly articles covering COVID-19. We survey and compare research works in these directions that are accompanied by open source data and code. Future research directions for data-driven COVID-19 research are also debated. We hope that the article will provide the scientific community with an initiative to start open source extensible and transparent research in the collective fight against the COVID-19 pandemic.

/pdf/covid-19-open-source-data-sets-a-comprehensive-survey-5ft1knk163.pdf

COVID-19 open source data sets: a comprehensive survey

https://onlinelibrary.wiley.com/doi/pdf/10.1002/wps.20883

The growing field of digital psychiatry: current evidence and the future of apps, social media, chatbots, and virtual reality.

Background: With restrictions on movement and stay-at-home orders in place due to the COVID-19 pandemic, social media platforms such as Twitter have become an outlet for users to express their concerns, opinions, and feelings about the pandemic. Individuals, health agencies, and governments are using Twitter to communicate about COVID-19.
Objective: The aims of this study were to examine key themes and topics of English-language COVID-19–related tweets posted by individuals and to explore the trends and variations in how the COVID-19–related tweets, key topics, and associated sentiments changed over a period of time from before to after the disease was declared a pandemic.
Methods: Building on the emergent stream of studies examining COVID-19–related tweets in English, we performed a temporal assessment covering the time period from January 1 to May 9, 2020, and examined variations in tweet topics and sentiment scores to uncover key trends. Combining data from two publicly available COVID-19 tweet data sets with those obtained in our own search, we compiled a data set of 13.9 million English-language COVID-19–related tweets posted by individuals. We use guided latent Dirichlet allocation (LDA) to infer themes and topics underlying the tweets, and we used VADER (Valence Aware Dictionary and sEntiment Reasoner) sentiment analysis to compute sentiment scores and examine weekly trends for 17 weeks.
Results: Topic modeling yielded 26 topics, which were grouped into 10 broader themes underlying the COVID-19–related tweets. Of the 13,937,906 examined tweets, 2,858,316 (20.51%) were about the impact of COVID-19 on the economy and markets, followed by spread and growth in cases (2,154,065, 15.45%), treatment and recovery (1,831,339, 13.14%), impact on the health care sector (1,588,499, 11.40%), and governments response (1,559,591, 11.19%). Average compound sentiment scores were found to be negative throughout the examined time period for the topics of spread and growth of cases, symptoms, racism, source of the outbreak, and political impact of COVID-19. In contrast, we saw a reversal of sentiments from negative to positive for prevention, impact on the economy and markets, government response, impact on the health care industry, and treatment and recovery.
Conclusions: Identification of dominant themes, topics, sentiments, and changing trends in tweets about the COVID-19 pandemic can help governments, health care agencies, and policy makers frame appropriate responses to prevent and control the spread of the pandemic.

/pdf/topics-trends-and-sentiments-of-tweets-about-the-covid-19-5ofyw5t71f.pdf

Topics, Trends, and Sentiments of Tweets About the COVID-19 Pandemic: Temporal Infoveillance Study.

During the unprecedented situation that all countries around the globe are facing due to the Coronavirus disease 2019 (COVID-19) pandemic, which has also had severe socioeconomic consequences, it is imperative to explore novel approaches to monitoring and forecasting regional outbreaks as they happen or even before they do so. To that end, in this paper, the role of Google query data in the predictability of COVID-19 in the United States at both national and state level is presented. As a preliminary investigation, Pearson and Kendall rank correlations are examined to explore the relationship between Google Trends data and COVID-19 data on cases and deaths. Next, a COVID-19 predictability analysis is performed, with the employed model being a quantile regression that is bias corrected via bootstrap simulation, i.e., a robust regression analysis that is the appropriate statistical approach to taking against the presence of outliers in the sample while also mitigating small sample estimation bias. The results indicate that there are statistically significant correlations between Google Trends and COVID-19 data, while the estimated models exhibit strong COVID-19 predictability. In line with previous work that has suggested that online real-time data are valuable in the monitoring and forecasting of epidemics and outbreaks, it is evident that such infodemiology approaches can assist public health policy makers in addressing the most crucial issues: flattening the curve, allocating health resources, and increasing the effectiveness and preparedness of their respective health care systems.

/pdf/covid-19-predictability-in-the-united-states-using-google-190wker7vs.pdf

COVID-19 predictability in the United States using Google Trends time series

Self-reported COVID-19 symptoms on Twitter: an analysis and a research resource.

Objective To mine Twitter to quantitatively analyze COVID-19 symptoms self-reported by users, compare symptom distributions against clinical studies, and create a symptom lexicon for the research community. Materials and methods We retrieved tweets using COVID-19-related keywords, and performed several layers of semi-automatic filtering to curate self-reports of positive-tested users. We extracted COVID-19-related symptoms mentioned by the users, mapped them to standard IDs, and compared the distributions with multiple studies conducted in clinical settings. Results We identified 203 positive-tested users who reported 932 symptoms using 598 unique expressions. The most frequently-reported symptoms were fever/pyrexia (65%), cough (56%), body aches/pain (40%), headache (35%), fatigue (35%), and dyspnea (34%) amongst users who reported at least 1 symptom. Mild symptoms, such as anosmia (26%) and ageusia (24%) were frequently reported on Twitter, but not in clinical studies. Conclusion The spectrum of COVID-19 symptoms identified from Twitter may complement those identified in clinical settings.

https://medrxiv.org/content/10.1101/2020.04.16.20067421v1.full.pdf

Self-reported COVID-19 symptoms on Twitter: An analysis and a research resource

Breast cancer patients often discontinue their long-term treatments, such as hormone therapy, increasing the risk of cancer recurrence. These discontinuations may be caused by adverse patient-centered outcomes (PCOs) due to hormonal drug side effects or other factors. PCOs are not detectable through laboratory tests, and are sparsely documented in electronic health records. Thus, there is a need to explore complementary sources of information for PCOs associated with breast cancer treatments. Social media is a promising resource, but extracting true PCOs from it first requires the accurate detection of real breast cancer patients. We describe a natural language processing (NLP) pipeline for automatically detecting breast cancer patients from Twitter based on their self-reports. The pipeline uses breast cancer-related keywords to collect streaming data from Twitter, applies NLP patterns to filter out noisy posts, and then employs a machine learning classifier trained using manually-annotated data (n = 5,019) for distinguishing firsthand self-reports of breast cancer from other tweets. A classifier based on bidirectional encoder representations from transformers (BERT) showed human-like performance and achieved F\(_1\)-score of 0.857 (inter-annotator agreement: 0.845; Cohen’s kappa) for the positive class, considerably outperforming the next best classifier—a recurrent neural network with bidirectional long short-term memory (F\(_1\)-score: 0.670). Qualitative analyses of posts from automatically-detected users revealed discussions about side effects, non-adherence and mental health conditions, illustrating the feasibility of our social media-based approach for studying breast cancer related PCOs from a large population.

Automatic Breast Cancer Cohort Detection from Social Media for Studying Factors Affecting Patient-Centered Outcomes

Americans bear a high chronic stress burden, particularly during the COVID-19 pandemic. Although social media have many strengths to complement the weaknesses of conventional stress measures, including surveys, they have been rarely utilized to detect individuals self-reporting chronic stress. Thus, this study aimed to develop and evaluate an automatic system on Twitter to identify users who have self-reported chronic stress experiences. Using the Twitter public streaming application programming interface, we collected tweets containing certain stress-related keywords (eg, "chronic," "constant," "stress") and then filtered the data using pre-defined text patterns. We manually annotated tweets with (without) self-report of chronic stress as positive (negative). We trained multiple classifiers and tested them via accuracy and F1 score. We annotated 4195 tweets (1560 positives, 2635 negatives), achieving an inter-annotator agreement of 0.83 (Cohen's kappa). The classifier based on Bidirectional Encoder Representation from Transformers performed the best (accuracy of 83.6% [81.0-86.1]), outperforming the second best-performing classifier (support vector machines: 76.4% [73.5-79.3]). The past tweets from the authors of positive tweets contained useful information, including sources and health impacts of chronic stress. Our study demonstrates that users' self-reported chronic stress experiences can be automatically identified on Twitter, which has a high potential for surveillance and large-scale intervention.

Automatic Detection of Twitter Users Who Express Chronic Stress Experiences via Supervised Machine Learning and Natural Language Processing.

Breast cancer patients often discontinue their long-term treatments, such as hormone therapy, increasing the risk of cancer recurrence. These discontinuations are often caused by adverse patient-centered outcomes (PCOs) due to hormonal drug side effects or other factors. PCOs are not detectable through laboratory tests and are sparsely documented in electronic health records. Thus, there is a need to explore other sources of information for PCOs associated with breast cancer treatments. Social media is a promising resource, but extracting true PCOs from it first requires the accurate detection of breast cancer patients. We describe a natural language processing (NLP) architecture for automatically detecting breast cancer patients from Twitter based on their self-reports. The architecture employs breast cancer-related keywords to collect streaming data from Twitter, applies NLP patterns to pre-filter noisy posts, and then employs a machine learning classifier trained using manually-annotated data (n=5019) for distinguishing firsthand self-reports of breast cancer from other tweets. A classifier based on bidirectional encoder representations from transformers (BERT) showed human-like performance and achieved F1-score of 0.857 (inter-annotator agreement: 0.845; Cohen9s kappa) for the positive class, considerably outperforming the next best classifier--a deep neural network (F1-score: 0.665). Qualitative analyses of posts from automatically-detected users revealed discussions about side effects, non-adherence, and mental health conditions, illustrating the feasibility of our social media-based approach for studying breast cancer-related PCOs from a large population.

/pdf/automatic-breast-cancer-survivor-detection-from-social-media-1eiq061fd2.pdf

Angel Xie

Papers

Self-reported COVID-19 symptoms on Twitter: an analysis and a research resource.

Self-reported COVID-19 symptoms on Twitter: An analysis and a research resource

Automatic Breast Cancer Cohort Detection from Social Media for Studying Factors Affecting Patient-Centered Outcomes

Automatic Detection of Twitter Users Who Express Chronic Stress Experiences via Supervised Machine Learning and Natural Language Processing.

Automatic Breast Cancer Survivor Detection from Social Media for Studying Latent Factors Affecting Treatment Success