1
|
Deiner MS, Honcharov V, Li J, Mackey TK, Porco TC, Sarkar U. Large Language Models Can Enable Inductive Thematic Analysis of a Social Media Corpus in a Single Prompt: Human Validation Study. JMIR INFODEMIOLOGY 2024; 4:e59641. [PMID: 39207842 PMCID: PMC11393503 DOI: 10.2196/59641] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 04/24/2024] [Revised: 06/04/2024] [Accepted: 07/01/2024] [Indexed: 09/04/2024]
Abstract
BACKGROUND Manually analyzing public health-related content from social media provides valuable insights into the beliefs, attitudes, and behaviors of individuals, shedding light on trends and patterns that can inform public understanding, policy decisions, targeted interventions, and communication strategies. Unfortunately, the time and effort needed from well-trained human subject matter experts makes extensive manual social media listening unfeasible. Generative large language models (LLMs) can potentially summarize and interpret large amounts of text, but it is unclear to what extent LLMs can glean subtle health-related meanings in large sets of social media posts and reasonably report health-related themes. OBJECTIVE We aimed to assess the feasibility of using LLMs for topic model selection or inductive thematic analysis of large contents of social media posts by attempting to answer the following question: Can LLMs conduct topic model selection and inductive thematic analysis as effectively as humans did in a prior manual study, or at least reasonably, as judged by subject matter experts? METHODS We asked the same research question and used the same set of social media content for both the LLM selection of relevant topics and the LLM analysis of themes as was conducted manually in a published study about vaccine rhetoric. We used the results from that study as background for this LLM experiment by comparing the results from the prior manual human analyses with the analyses from 3 LLMs: GPT4-32K, Claude-instant-100K, and Claude-2-100K. We also assessed if multiple LLMs had equivalent ability and assessed the consistency of repeated analysis from each LLM. RESULTS The LLMs generally gave high rankings to the topics chosen previously by humans as most relevant. We reject a null hypothesis (P<.001, overall comparison) and conclude that these LLMs are more likely to include the human-rated top 5 content areas in their top rankings than would occur by chance. Regarding theme identification, LLMs identified several themes similar to those identified by humans, with very low hallucination rates. Variability occurred between LLMs and between test runs of an individual LLM. Despite not consistently matching the human-generated themes, subject matter experts found themes generated by the LLMs were still reasonable and relevant. CONCLUSIONS LLMs can effectively and efficiently process large social media-based health-related data sets. LLMs can extract themes from such data that human subject matter experts deem reasonable. However, we were unable to show that the LLMs we tested can replicate the depth of analysis from human subject matter experts by consistently extracting the same themes from the same data. There is vast potential, once better validated, for automated LLM-based real-time social listening for common and rare health conditions, informing public health understanding of the public's interests and concerns and determining the public's ideas to address them.
Collapse
Affiliation(s)
- Michael S Deiner
- Department of Ophthalmology and Francis I Proctor Foundation, University of California San Francisco, San Francisco, CA, United States
| | - Vlad Honcharov
- Center for Vulnerable Populations, Zuckerberg San Francisco General Hospital, Department of Medicine, University of California San Francisco, San Francisco, CA, United States
- Division of General Internal Medicine, Zuckerberg San Francisco General Hospital, Department of Medicine, University of California San Francisco, San Francisco, CA, United States
| | - Jiawei Li
- S-3 Research, LLC, San Diego, CA, United States
| | - Tim K Mackey
- S-3 Research, LLC, San Diego, CA, United States
- Global Health Program, Department of Anthropology, University of California San Diego, La Jolla, CA, United States
| | - Travis C Porco
- Departments of Ophthalmology, Epidemiology and Biostatistics, Global Health Sciences, and Francis I Proctor Foundation, University of California San Francisco, San Francisco, CA, United States
| | - Urmimala Sarkar
- Center for Vulnerable Populations, Zuckerberg San Francisco General Hospital, Department of Medicine, University of California San Francisco, San Francisco, CA, United States
- Division of General Internal Medicine, Zuckerberg San Francisco General Hospital, Department of Medicine, University of California San Francisco, San Francisco, CA, United States
| |
Collapse
|
2
|
Nerella S, Bandyopadhyay S, Zhang J, Contreras M, Siegel S, Bumin A, Silva B, Sena J, Shickel B, Bihorac A, Khezeli K, Rashidi P. Transformers and large language models in healthcare: A review. Artif Intell Med 2024; 154:102900. [PMID: 38878555 DOI: 10.1016/j.artmed.2024.102900] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/01/2023] [Revised: 05/28/2024] [Accepted: 05/30/2024] [Indexed: 08/09/2024]
Abstract
With Artificial Intelligence (AI) increasingly permeating various aspects of society, including healthcare, the adoption of the Transformers neural network architecture is rapidly changing many applications. Transformer is a type of deep learning architecture initially developed to solve general-purpose Natural Language Processing (NLP) tasks and has subsequently been adapted in many fields, including healthcare. In this survey paper, we provide an overview of how this architecture has been adopted to analyze various forms of healthcare data, including clinical NLP, medical imaging, structured Electronic Health Records (EHR), social media, bio-physiological signals, biomolecular sequences. Furthermore, which have also include the articles that used the transformer architecture for generating surgical instructions and predicting adverse outcomes after surgeries under the umbrella of critical care. Under diverse settings, these models have been used for clinical diagnosis, report generation, data reconstruction, and drug/protein synthesis. Finally, we also discuss the benefits and limitations of using transformers in healthcare and examine issues such as computational cost, model interpretability, fairness, alignment with human values, ethical implications, and environmental impact.
Collapse
Affiliation(s)
- Subhash Nerella
- Department of Biomedical Engineering, University of Florida, Gainesville, United States
| | | | - Jiaqing Zhang
- Department of Electrical and Computer Engineering, University of Florida, Gainesville, United States
| | - Miguel Contreras
- Department of Biomedical Engineering, University of Florida, Gainesville, United States
| | - Scott Siegel
- Department of Biomedical Engineering, University of Florida, Gainesville, United States
| | - Aysegul Bumin
- Department of Computer and Information Science and Engineering, University of Florida, Gainesville, United States
| | - Brandon Silva
- Department of Computer and Information Science and Engineering, University of Florida, Gainesville, United States
| | - Jessica Sena
- Department Of Computer Science, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil
| | - Benjamin Shickel
- Department of Medicine, University of Florida, Gainesville, United States
| | - Azra Bihorac
- Department of Medicine, University of Florida, Gainesville, United States
| | - Kia Khezeli
- Department of Biomedical Engineering, University of Florida, Gainesville, United States
| | - Parisa Rashidi
- Department of Biomedical Engineering, University of Florida, Gainesville, United States.
| |
Collapse
|
3
|
Chepo M, Martin S, Déom N, Khalid AF, Vindrola-Padros C. Twitter Analysis of Health Care Workers' Sentiment and Discourse Regarding Post-COVID-19 Condition in Children and Young People: Mixed Methods Study. J Med Internet Res 2024; 26:e50139. [PMID: 38630514 PMCID: PMC11063881 DOI: 10.2196/50139] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/20/2023] [Revised: 02/14/2024] [Accepted: 03/08/2024] [Indexed: 04/19/2024] Open
Abstract
BACKGROUND The COVID-19 pandemic has had a significant global impact, with millions of cases and deaths. Research highlights the persistence of symptoms over time (post-COVID-19 condition), a situation of particular concern in children and young people with symptoms. Social media such as Twitter (subsequently rebranded as X) could provide valuable information on the impact of the post-COVID-19 condition on this demographic. OBJECTIVE With a social media analysis of the discourse surrounding the prevalence of post-COVID-19 condition in children and young people, we aimed to explore the perceptions of health care workers (HCWs) concerning post-COVID-19 condition in children and young people in the United Kingdom between January 2021 and January 2022. This will allow us to contribute to the emerging knowledge on post-COVID-19 condition and identify critical areas and future directions for researchers and policy makers. METHODS From a pragmatic paradigm, we used a mixed methods approach. Through discourse, keyword, sentiment, and image analyses, using Pulsar and InfraNodus, we analyzed the discourse about the experience of post-COVID-19 condition in children and young people in the United Kingdom shared on Twitter between January 1, 2021, and January 31, 2022, from a sample of HCWs with Twitter accounts whose biography identifies them as HCWs. RESULTS We obtained 300,000 tweets, out of which (after filtering for relevant tweets) we performed an in-depth qualitative sample analysis of 2588 tweets. The HCWs were responsive to announcements issued by the authorities regarding the management of the COVID-19 pandemic in the United Kingdom. The most frequent sentiment expressed was negative. The main themes were uncertainty about the future, policies and regulations, managing and addressing the COVID-19 pandemic and post-COVID-19 condition in children and young people, vaccination, using Twitter to share scientific literature and management strategies, and clinical and personal experiences. CONCLUSIONS The perceptions described on Twitter by HCWs concerning the presence of the post-COVID-19 condition in children and young people appear to be a relevant and timely issue and responsive to the declarations and guidelines issued by health authorities over time. We recommend further support and training strategies for health workers and school staff regarding the manifestations and treatment of children and young people with post-COVID-19 condition.
Collapse
Affiliation(s)
- Macarena Chepo
- School of Nursing, Universidad Andrés Bello, Santiago, Chile
| | - Sam Martin
- Department of Targeted Intervention, University College London, London, United Kingdom
- Oxford Vaccine Group, Churchill Hospital, University of Oxford, Oxford, United Kingdom
| | - Noémie Déom
- Department of Targeted Intervention, University College London, London, United Kingdom
| | - Ahmad Firas Khalid
- Canadian Institutes of Health Research Health System Impact Fellowship, Centre for Implementation Research, Ottawa Hospital Research Institute, Otawa, ON, Canada
| | | |
Collapse
|
4
|
Ullah N, Martin S, Poduval S. A Snapshot of COVID-19 Vaccine Discourse Related to Ethnic Minority Communities in the United Kingdom Between January and April 2022: Mixed Methods Analysis. JMIR Form Res 2024; 8:e51152. [PMID: 38530334 DOI: 10.2196/51152] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/22/2023] [Revised: 02/27/2024] [Accepted: 02/29/2024] [Indexed: 03/27/2024] Open
Abstract
BACKGROUND Existing literature highlights the role of social media as a key source of information for the public during the COVID-19 pandemic and its influence on vaccination attempts. Yet there is little research exploring its role in the public discourse specifically among ethnic minority communities, who have the highest rates of vaccine hesitancy (delay or refusal of vaccination despite availability of services). OBJECTIVE This study aims to understand the discourse related to minority communities on social media platforms Twitter and YouTube. METHODS Social media data from the United Kingdom was extracted from Twitter and YouTube using the software Netlytics and YouTube Data Tools to provide a "snapshot" of the discourse between January and April 2022. A mixed method approach was used where qualitative data were contextualized into codes. Network analysis was applied to provide insight into the most frequent and weighted keywords and topics of conversations. RESULTS A total of 260 tweets and 156 comments from 4 YouTube videos were included in our analysis. Our data suggests that the most popular topics of conversation during the period sampled were related to communication strategies adopted during the booster vaccine rollout. These were noted to be divisive in nature and linked to wider conversations around racism and historical mistrust toward institutions. CONCLUSIONS Our study suggests a shift in narrative from concerns about the COVID-19 vaccine itself, toward the strategies used in vaccination implementation, in particular the targeting of ethnic minority groups through vaccination campaigns. The implications for public health communication during crisis management in a pandemic context include acknowledging wider experiences of discrimination when addressing ethnic minority communities.
Collapse
Affiliation(s)
- Nazifa Ullah
- Research Department of Primary Care & Population Health, University College London, London, United Kingdom
| | - Sam Martin
- Vaccines and Society Unit, Oxford Vaccine Group, University of Oxford, Oxford, United Kingdom
| | - Shoba Poduval
- Institute of Health Informatics, University College London, London, United Kingdom
| |
Collapse
|
5
|
Vera San Juan N, Martin S, Badley A, Maio L, Gronholm PC, Buck C, Flores EC, Vanderslott S, Syversen A, Symmons SM, Uddin I, Karia A, Iqbal S, Vindrola-Padros C. Frontline Health Care Workers' Mental Health and Well-Being During the First Year of the COVID-19 Pandemic: Analysis of Interviews and Social Media Data. J Med Internet Res 2023; 25:e43000. [PMID: 37402283 PMCID: PMC10426381 DOI: 10.2196/43000] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/27/2022] [Revised: 06/29/2023] [Accepted: 07/04/2023] [Indexed: 07/06/2023] Open
Abstract
BACKGROUND The COVID-19 pandemic has shed light on fractures in health care systems worldwide and continues to have a significant impact, particularly in relation to the health care workforce. Frontline staff have been exposed to unprecedented strain, and delivering care during the pandemic has affected their safety, mental health, and well-being. OBJECTIVE This study aimed to explore the experiences of health care workers (HCWs) delivering care in the United Kingdom during the COVID-19 pandemic to understand their well-being needs, experiences, and strategies used to maintain well-being (at individual and organizational levels). METHODS We analyzed 94 telephone interviews with HCWs and 2000 tweets about HCWs' mental health during the first year of the COVID-19 pandemic. RESULTS The results were grouped under 6 themes: redeployment, clinical work, and sense of duty; well-being support and HCW's coping strategies; negative mental health effects; organizational support; social network and support; and public and government support. CONCLUSIONS These findings demonstrate the need for open conversations, where staff's well-being needs and the strategies they adopted can be shared and encouraged, rather than implementing top-down psychological interventions alone. At the macro level, the findings also highlighted the impact on HCW's well-being of public and government support as well as the need to ensure protection through personal protective equipment, testing, and vaccines for frontline workers.
Collapse
Affiliation(s)
- Norha Vera San Juan
- Rapid Research Evaluation and Appraisal Lab (RREAL), Department of Targeted Intervention, University College London, London, United Kingdom
- Centre for Global Mental Health and Centre for Implementation Science, Health Services and Population Research Department, Institute of Psychiatry, Psychology and Neuroscience, King's College London, London, United Kingdom
| | - Sam Martin
- Rapid Research Evaluation and Appraisal Lab (RREAL), Department of Targeted Intervention, University College London, London, United Kingdom
- Ethox Centre, Big Data Institute, University of Oxford, Oxford, United Kingdom
| | - Anna Badley
- Academy Research and Improvement, Solent Trust, Southampton, United Kingdom
- School of Health Sciences, University of Southampton, Southampton, United Kingdom
| | - Laura Maio
- Rapid Research Evaluation and Appraisal Lab (RREAL), Department of Targeted Intervention, University College London, London, United Kingdom
| | - Petra C Gronholm
- Centre for Global Mental Health and Centre for Implementation Science, Health Services and Population Research Department, Institute of Psychiatry, Psychology and Neuroscience, King's College London, London, United Kingdom
| | - Caroline Buck
- Department of Behavioural Science and Health, University College London, London, United Kingdom
| | - Elaine C Flores
- Centre on Climate Change & Planetary Health, London School of Hygiene and Tropical Medicine, London, United Kingdom
- Stanford Center for Innovation in Global Health, Stanford Woods Institute for the Environment,, Stanford University, Stanford, CA, United States
| | - Samantha Vanderslott
- Oxford Vaccine Group, Churchill Hospital, University of Oxford, Oxford, United Kingdom
- NIHR Oxford Biomedical Research Centre, University of Oxford, Oxford, United Kingdom
| | - Aron Syversen
- Institute of Epidemiology and Healthcare, University College London, London, United Kingdom
| | - Sophie Mulcahy Symmons
- Centre for Interdisciplinary Research, Education and Innovation in Health Systems, School of Nursing, Midwifery and Health Systems, University College Dublin, Dublin, Ireland
| | - Inayah Uddin
- Division of Psychiatry, Marie Curie Palliative Care Research Department, University College London, London, United Kingdom
| | - Amelia Karia
- Rapid Research Evaluation and Appraisal Lab (RREAL), Department of Targeted Intervention, University College London, London, United Kingdom
| | - Syka Iqbal
- Rapid Research Evaluation and Appraisal Lab (RREAL), Department of Targeted Intervention, University College London, London, United Kingdom
- Department of Psychology, University of Bradford, Bradford, United Kingdom
| | - Cecilia Vindrola-Padros
- Rapid Research Evaluation and Appraisal Lab (RREAL), Department of Targeted Intervention, University College London, London, United Kingdom
| |
Collapse
|
6
|
Dupuy-Zini A, Audeh B, Gérardin C, Duclos C, Gagneux-Brunon A, Bousquet C. Users' Reactions to Announced Vaccines Against COVID-19 Before Marketing in France: Analysis of Twitter Posts. J Med Internet Res 2023; 25:e37237. [PMID: 36596215 PMCID: PMC10132828 DOI: 10.2196/37237] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/11/2022] [Revised: 07/17/2022] [Accepted: 08/09/2022] [Indexed: 01/04/2023] Open
Abstract
BACKGROUND Within a few months, the COVID-19 pandemic had spread to many countries and had been a real challenge for health systems all around the world. This unprecedented crisis has led to a surge of online discussions about potential cures for the disease. Among them, vaccines have been at the heart of the debates and have faced lack of confidence before marketing in France. OBJECTIVE This study aims to identify and investigate the opinions of French Twitter users on the announced vaccines against COVID-19 through sentiment analysis. METHODS This study was conducted in 2 phases. First, we filtered a collection of tweets related to COVID-19 available on Twitter from February 2020 to August 2020 with a set of keywords associated with vaccine mistrust using word embeddings. Second, we performed sentiment analysis using deep learning to identify the characteristics of vaccine mistrust. The model was trained on a hand-labeled subset of 4548 tweets. RESULTS A set of 69 relevant keywords were identified as the semantic concept of the word "vaccin" (vaccine in French) and focused mainly on conspiracies, pharmaceutical companies, and alternative treatments. Those keywords enabled us to extract nearly 350,000 tweets in French. The sentiment analysis model achieved 0.75 accuracy. The model then predicted 16% of positive tweets, 41% of negative tweets, and 43% of neutral tweets. This allowed us to explore the semantic concepts of positive and negative tweets and to plot the trends of each sentiment. The main negative rhetoric identified from users' tweets was that vaccines are perceived as having a political purpose and that COVID-19 is a commercial argument for the pharmaceutical companies. CONCLUSIONS Twitter might be a useful tool to investigate the arguments for vaccine mistrust because it unveils political criticism contrasting with the usual concerns on adverse drug reactions. As the opposition rhetoric is more consistent and more widely spread than the positive rhetoric, we believe that this research provides effective tools to help health authorities better characterize the risk of vaccine mistrust.
Collapse
Affiliation(s)
- Alexandre Dupuy-Zini
- Laboratoire d'Informatique Médicale et d'Ingénierie des connaissances en e-Santé, LIMICS, Sorbonne Université, Université Sorbonne Paris Nord, Institut national de la santé et de la recherche médicale, INSERM, Paris, France
| | - Bissan Audeh
- Laboratoire d'Informatique Médicale et d'Ingénierie des connaissances en e-Santé, LIMICS, Sorbonne Université, Université Sorbonne Paris Nord, Institut national de la santé et de la recherche médicale, INSERM, Paris, France
| | - Christel Gérardin
- Institut Pierre Louis d'Epidémiologie et de Santé Publique, Département de médecine interne, Sorbonne Université, Paris, France
| | - Catherine Duclos
- Laboratoire d'Informatique Médicale et d'Ingénierie des connaissances en e-Santé, LIMICS, Sorbonne Université, Université Sorbonne Paris Nord, Institut national de la santé et de la recherche médicale, INSERM, Paris, France
| | - Amandine Gagneux-Brunon
- Groupe sur l'Immunité des Muqueuses et Agents Pathogènes, Centre International de Recherche en Infectiologie, University of Lyon, Saint Etienne, France
- Vaccinologie, Centre Hospitalier Universitaire de Saint-Etienne, Saint Etienne, France
| | - Cedric Bousquet
- Laboratoire d'Informatique Médicale et d'Ingénierie des connaissances en e-Santé, LIMICS, Sorbonne Université, Université Sorbonne Paris Nord, Institut national de la santé et de la recherche médicale, INSERM, Paris, France
- Service de santé publique et information médicale, Centre Hospitalier Universitaire de Saint Etienne, Saint Etienne, France
| |
Collapse
|
7
|
Müller M, Salathé M, Kummervold PE. COVID-Twitter-BERT: A natural language processing model to analyse COVID-19 content on Twitter. Front Artif Intell 2023; 6:1023281. [PMID: 36998290 PMCID: PMC10043293 DOI: 10.3389/frai.2023.1023281] [Citation(s) in RCA: 14] [Impact Index Per Article: 14.0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/19/2022] [Accepted: 01/31/2023] [Indexed: 03/15/2023] Open
Abstract
IntroductionThis study presents COVID-Twitter-BERT (CT-BERT), a transformer-based model that is pre-trained on a large corpus of COVID-19 related Twitter messages. CT-BERT is specifically designed to be used on COVID-19 content, particularly from social media, and can be utilized for various natural language processing tasks such as classification, question-answering, and chatbots. This paper aims to evaluate the performance of CT-BERT on different classification datasets and compare it with BERT-LARGE, its base model.MethodsThe study utilizes CT-BERT, which is pre-trained on a large corpus of COVID-19 related Twitter messages. The authors evaluated the performance of CT-BERT on five different classification datasets, including one in the target domain. The model's performance is compared to its base model, BERT-LARGE, to measure the marginal improvement. The authors also provide detailed information on the training process and the technical specifications of the model.ResultsThe results indicate that CT-BERT outperforms BERT-LARGE with a marginal improvement of 10-30% on all five classification datasets. The largest improvements are observed in the target domain. The authors provide detailed performance metrics and discuss the significance of these results.DiscussionThe study demonstrates the potential of pre-trained transformer models, such as CT-BERT, for COVID-19 related natural language processing tasks. The results indicate that CT-BERT can improve the classification performance on COVID-19 related content, especially on social media. These findings have important implications for various applications, such as monitoring public sentiment and developing chatbots to provide COVID-19 related information. The study also highlights the importance of using domain-specific pre-trained models for specific natural language processing tasks. Overall, this work provides a valuable contribution to the development of COVID-19 related NLP models.
Collapse
Affiliation(s)
| | | | - Per E. Kummervold
- FISABIO-Public Health, Vaccine Research Department, Valencia, Spain
- *Correspondence: Per E. Kummervold
| |
Collapse
|
8
|
"Any idea how fast 'It's just a mask!' can turn into 'It's just a vaccine!'": From mask mandates to vaccine mandates during the COVID-19 pandemic. Vaccine 2022; 40:7488-7499. [PMID: 34823912 PMCID: PMC8552554 DOI: 10.1016/j.vaccine.2021.10.031] [Citation(s) in RCA: 13] [Impact Index Per Article: 6.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/05/2021] [Revised: 09/23/2021] [Accepted: 10/12/2021] [Indexed: 01/28/2023]
Abstract
Protests starting in the summer of 2020, notedly in the US and UK, have brought together two constituencies: pre-existing anti-vaccine groups and newly formed oppositional COVID-19 groups. The oppositional COVID-19 groups vary in composition and nature, but the central focus is a disagreement about the seriousness and threat of COVID-19 and with the public health measures to control COVID-19. What unites many disparate interests is an aversion to mandates. The compulsion to undertake particular public health activities such as mask-wearing and vaccination is a complex topic of public attitudes and beliefs alongside public health goals and messaging. We aim to analyse social media discussions about facemask wearing and the adoption of potential vaccines for COVID-19. Using media monitoring software MeltwaterTM, we analyse English-language tweets for one year from 1st June 2020 until 1st June 2021. We pay particular attention to connections in conversations between key topics of concern regarding masks and vaccines across social media networks. We track where ideas and activist behaviours towards both health interventions have originated, have similarities, and how they have changed over time. Our aim is to provide an overview of the key trends and themes of discussion concerning attitudes to and adoption of health measures in the control of COVID-19 and how publics react when confronted with mandatory policies. We draw on an already extensive literature about mandatory vaccination policies to inform our assessment, from psychology and behavioural science to ethics, political theory, sociology, and public policy.
Collapse
|
9
|
Cheatham S, Kummervold PE, Parisi L, Lanfranchi B, Croci I, Comunello F, Rota MC, Filia A, Tozzi AE, Rizzo C, Gesualdo F. Understanding the vaccine stance of Italian tweets and addressing language changes through the COVID-19 pandemic: Development and validation of a machine learning model. Front Public Health 2022; 10:948880. [PMID: 35968436 PMCID: PMC9372360 DOI: 10.3389/fpubh.2022.948880] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/20/2022] [Accepted: 07/11/2022] [Indexed: 11/13/2022] Open
Abstract
Social media is increasingly being used to express opinions and attitudes toward vaccines. The vaccine stance of social media posts can be classified in almost real-time using machine learning. We describe the use of a Transformer-based machine learning model for analyzing vaccine stance of Italian tweets, and demonstrate the need to address changes over time in vaccine-related language, through periodic model retraining. Vaccine-related tweets were collected through a platform developed for the European Joint Action on Vaccination. Two datasets were collected, the first between November 2019 and June 2020, the second from April to September 2021. The tweets were manually categorized by three independent annotators. After cleaning, the total dataset consisted of 1,736 tweets with 3 categories (promotional, neutral, and discouraging). The manually classified tweets were used to train and test various machine learning models. The model that classified the data most similarly to humans was XLM-Roberta-large, a multilingual version of the Transformer-based model RoBERTa. The model hyper-parameters were tuned and then the model ran five times. The fine-tuned model with the best F-score over the validation dataset was selected. Running the selected fine-tuned model on just the first test dataset resulted in an accuracy of 72.8% (F-score 0.713). Using this model on the second test dataset resulted in a 10% drop in accuracy to 62.1% (F-score 0.617), indicating that the model recognized a difference in language between the datasets. On the combined test datasets the accuracy was 70.1% (F-score 0.689). Retraining the model using data from the first and second datasets increased the accuracy over the second test dataset to 71.3% (F-score 0.713), a 9% improvement from when using just the first dataset for training. The accuracy over the first test dataset remained the same at 72.8% (F-score 0.721). The accuracy over the combined test datasets was then 72.4% (F-score 0.720), a 2% improvement. Through fine-tuning a machine-learning model on task-specific data, the accuracy achieved in categorizing tweets was close to that expected by a single human annotator. Regular training of machine-learning models with recent data is advisable to maximize accuracy.
Collapse
Affiliation(s)
- Susan Cheatham
- Multifactorial and Complex Diseases Research Area, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy
| | | | - Lorenza Parisi
- Department of Human Sciences, Link Campus University, Rome, Italy
| | - Barbara Lanfranchi
- Multifactorial and Complex Diseases Research Area, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy
| | - Ileana Croci
- Multifactorial and Complex Diseases Research Area, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy
| | - Francesca Comunello
- Department of Communication and Social Research, Sapienza University, Rome, Italy
| | - Maria Cristina Rota
- Department of Infectious Diseases, Istituto Superiore di Sanità, Rome, Italy
| | - Antonietta Filia
- Department of Infectious Diseases, Istituto Superiore di Sanità, Rome, Italy
| | - Alberto Eugenio Tozzi
- Multifactorial and Complex Diseases Research Area, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy
| | - Caterina Rizzo
- Multifactorial and Complex Diseases Research Area, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy
- Department of Translational Research and New Technologies in Medicine and Surgery, Pisa University, Pisa, Italy
- *Correspondence: Caterina Rizzo
| | - Francesco Gesualdo
- Multifactorial and Complex Diseases Research Area, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy
| |
Collapse
|
10
|
Yin JDC. Media Data and Vaccine Hesitancy: Scoping Review. JMIR INFODEMIOLOGY 2022; 2:e37300. [PMID: 37113443 PMCID: PMC9987198 DOI: 10.2196/37300] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 02/15/2022] [Revised: 06/16/2022] [Accepted: 07/14/2022] [Indexed: 04/29/2023]
Abstract
Background Media studies are important for vaccine hesitancy research, as they analyze how the media shapes risk perceptions and vaccine uptake. Despite the growth in studies in this field owing to advances in computing and language processing and an expanding social media landscape, no study has consolidated the methodological approaches used to study vaccine hesitancy. Synthesizing this information can better structure and set a precedent for this growing subfield of digital epidemiology. Objective This review aimed to identify and illustrate the media platforms and methods used to study vaccine hesitancy and how they build or contribute to the study of the media's influence on vaccine hesitancy and public health. Methods This study followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. A search was conducted on PubMed and Scopus for any studies that used media data (social media or traditional media), had an outcome related to vaccine sentiment (opinion, uptake, hesitancy, acceptance, or stance), were written in English, and were published after 2010. Studies were screened by only 1 reviewer and extracted for media platform, analysis method, the theoretical models used, and outcomes. Results In total, 125 studies were included, of which 71 (56.8%) used traditional research methods and 54 (43.2%) used computational methods. Of the traditional methods, most used content analysis (43/71, 61%) and sentiment analysis (21/71, 30%) to analyze the texts. The most common platforms were newspapers, print media, and web-based news. The computational methods mostly used sentiment analysis (31/54, 57%), topic modeling (18/54, 33%), and network analysis (17/54, 31%). Fewer studies used projections (2/54, 4%) and feature extraction (1/54, 2%). The most common platforms were Twitter and Facebook. Theoretically, most studies were weak. The following five major categories of studies arose: antivaccination themes centered on the distrust of institutions, civil liberties, misinformation, conspiracy theories, and vaccine-specific concerns; provaccination themes centered on ensuring vaccine safety using scientific literature; framing being important and health professionals and personal stories having the largest impact on shaping vaccine opinion; the coverage of vaccination-related data mostly identifying negative vaccine content and revealing deeply fractured vaccine communities and echo chambers; and the public reacting to and focusing on certain signals-in particular cases, deaths, and scandals-which suggests a more volatile period for the spread of information. Conclusions The heterogeneity in the use of media to study vaccines can be better consolidated through theoretical grounding. Areas of suggested research include understanding how trust in institutions is associated with vaccine uptake, how misinformation and information signaling influence vaccine uptake, and the evaluation of government communications on vaccine rollouts and vaccine-related events. The review ends with a statement that media data analyses, though groundbreaking in approach, should supplement-not supplant-current practices in public health research.
Collapse
Affiliation(s)
- Jason Dean-Chen Yin
- School of Public Health Li Ka Shing Faculty of Medicine The University of Hong Kong Hong Kong China (Hong Kong)
| |
Collapse
|
11
|
Sauvayre R, Vernier J, Chauvière C. Using supervised learning to analyze the French vaccine debate on Twitter. JMIR Med Inform 2022; 10:e37831. [PMID: 35512274 PMCID: PMC9116457 DOI: 10.2196/37831] [Citation(s) in RCA: 4] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/08/2022] [Revised: 05/01/2022] [Accepted: 05/04/2022] [Indexed: 11/13/2022] Open
Abstract
Background As the COVID-19 pandemic progressed, disinformation, fake news, and conspiracy theories spread through many parts of society. However, the disinformation spreading through social media is, according to the literature, one of the causes of increased COVID-19 vaccine hesitancy. In this context, the analysis of social media posts is particularly important, but the large amount of data exchanged on social media platforms requires specific methods. This is why machine learning and natural language processing models are increasingly applied to social media data. Objective The aim of this study is to examine the capability of the CamemBERT French-language model to faithfully predict the elaborated categories, with the knowledge that tweets about vaccination are often ambiguous, sarcastic, or irrelevant to the studied topic. Methods A total of 901,908 unique French-language tweets related to vaccination published between July 12, 2021, and August 11, 2021, were extracted using Twitter’s application programming interface (version 2; Twitter Inc). Approximately 2000 randomly selected tweets were labeled with 2 types of categorizations: (1) arguments for (pros) or against (cons) vaccination (health measures included) and (2) type of content (scientific, political, social, or vaccination status). The CamemBERT model was fine-tuned and tested for the classification of French-language tweets. The model’s performance was assessed by computing the F1-score, and confusion matrices were obtained. Results The accuracy of the applied machine learning reached up to 70.6% for the first classification (pro and con tweets) and up to 90% for the second classification (scientific and political tweets). Furthermore, a tweet was 1.86 times more likely to be incorrectly classified by the model if it contained fewer than 170 characters (odds ratio 1.86; 95% CI 1.20-2.86). Conclusions The accuracy of the model is affected by the classification chosen and the topic of the message examined. When the vaccine debate is jostled by contested political decisions, tweet content becomes so heterogeneous that the accuracy of the model drops for less differentiated classes. However, our tests showed that it is possible to improve the accuracy by selecting tweets using a new method based on tweet length.
Collapse
Affiliation(s)
- Romy Sauvayre
- Laboratoire de Psychologie Sociale et Cognitive, Université Clermont Auvergne, CNRS, Clermont-Ferrand, FR.,Polytech Clermont, 2 avenue Blaise Pascal, Aubiere, FR
| | - Jessica Vernier
- Laboratoire de Psychologie Sociale et Cognitive, Université Clermont Auvergne, CNRS, Clermont-Ferrand, FR
| | - Cédric Chauvière
- Laboratoire de Mathématiques Blaise Pascal, Université Clermont Auvergne, CNRS, Clermont-Ferrand, FR.,Polytech Clermont, 2 avenue Blaise Pascal, Aubiere, FR
| |
Collapse
|
12
|
Sun Y, Gao D, Shen X, Li M, Nan J, Zhang W. Multi-Label Classification in Patient-Doctor Dialogues With the RoBERTa-WWM-ext + CNN (Robustly Optimized Bidirectional Encoder Representations From Transformers Pretraining Approach With Whole Word Masking Extended Combining a Convolutional Neural Network) Model: Named Entity Study. JMIR Med Inform 2022; 10:e35606. [PMID: 35451969 PMCID: PMC9073616 DOI: 10.2196/35606] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/10/2021] [Revised: 02/24/2022] [Accepted: 02/25/2022] [Indexed: 12/03/2022] Open
Abstract
Background With the prevalence of online consultation, many patient-doctor dialogues have accumulated, which, in an authentic language environment, are of significant value to the research and development of intelligent question answering and automated triage in recent natural language processing studies. Objective The purpose of this study was to design a front-end task module for the network inquiry of intelligent medical services. Through the study of automatic labeling of real doctor-patient dialogue text on the internet, a method of identifying the negative and positive entities of dialogues with higher accuracy has been explored. Methods The data set used for this study was from the Spring Rain Doctor internet online consultation, which was downloaded from the official data set of Alibaba Tianchi Lab. We proposed a composite abutting joint model, which was able to automatically classify the types of clinical finding entities into the following 4 attributes: positive, negative, other, and empty. We adapted a downstream architecture in Chinese Robustly Optimized Bidirectional Encoder Representations from Transformers Pretraining Approach (RoBERTa) with whole word masking (WWM) extended (RoBERTa-WWM-ext) combining a text convolutional neural network (CNN). We used RoBERTa-WWM-ext to express sentence semantics as a text vector and then extracted the local features of the sentence through the CNN, which was our new fusion model. To verify its knowledge learning ability, we chose Enhanced Representation through Knowledge Integration (ERNIE), original Bidirectional Encoder Representations from Transformers (BERT), and Chinese BERT with WWM to perform the same task, and then compared the results. Precision, recall, and macro-F1 were used to evaluate the performance of the methods. Results We found that the ERNIE model, which was trained with a large Chinese corpus, had a total score (macro-F1) of 65.78290014, while BERT and BERT-WWM had scores of 53.18247117 and 69.2795315, respectively. Our composite abutting joint model (RoBERTa-WWM-ext + CNN) had a macro-F1 value of 70.55936311, showing that our model outperformed the other models in the task. Conclusions The accuracy of the original model can be greatly improved by giving priority to WWM and replacing the word-based mask with unit to classify and label medical entities. Better results can be obtained by effectively optimizing the downstream tasks of the model and the integration of multiple models later on. The study findings contribute to the translation of online consultation information into machine-readable information.
Collapse
Affiliation(s)
- Yuanyuan Sun
- Institute of Medical Information, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China.,Department of Internal Medicine, Chinese Academy of Medical Sciences, Peking Union Medical College Hospital, Beijing, China
| | - Dongping Gao
- Institute of Medical Information, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China
| | - Xifeng Shen
- Institute of Medical Information, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China
| | - Meiting Li
- Institute of Medical Information, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China
| | - Jiale Nan
- Institute of Medical Information, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China
| | - Weining Zhang
- Institute of Medical Information, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China
| |
Collapse
|
13
|
Portelli B, Scaboro S, Tonino R, Chersoni E, Santus E, Serra G. Monitoring user opinions and side effects on COVID-19 vaccines in the Twittersphere: Infodemiology Study of Tweets. J Med Internet Res 2022; 24:e35115. [PMID: 35446781 PMCID: PMC9132143 DOI: 10.2196/35115] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/22/2021] [Revised: 01/29/2022] [Accepted: 03/09/2022] [Indexed: 11/18/2022] Open
Abstract
Background In the current phase of the COVID-19 pandemic, we are witnessing the most massive vaccine rollout in human history. Like any other drug, vaccines may cause unexpected side effects, which need to be investigated in a timely manner to minimize harm in the population. If not properly dealt with, side effects may also impact public trust in the vaccination campaigns carried out by national governments. Objective Monitoring social media for the early identification of side effects, and understanding the public opinion on the vaccines are of paramount importance to ensure a successful and harmless rollout. The objective of this study was to create a web portal to monitor the opinion of social media users on COVID-19 vaccines, which can offer a tool for journalists, scientists, and users alike to visualize how the general public is reacting to the vaccination campaign. Methods We developed a tool to analyze the public opinion on COVID-19 vaccines from Twitter, exploiting, among other techniques, a state-of-the-art system for the identification of adverse drug events on social media; natural language processing models for sentiment analysis; statistical tools; and open-source databases to visualize the trending hashtags, news articles, and their factuality. All modules of the system are displayed through an open web portal. Results A set of 650,000 tweets was collected and analyzed in an ongoing process that was initiated in December 2020. The results of the analysis are made public on a web portal (updated daily), together with the processing tools and data. The data provide insights on public opinion about the vaccines and its change over time. For example, users show a high tendency to only share news from reliable sources when discussing COVID-19 vaccines (98% of the shared URLs). The general sentiment of Twitter users toward the vaccines is negative/neutral; however, the system is able to record fluctuations in the attitude toward specific vaccines in correspondence with specific events (eg, news about new outbreaks). The data also show how news coverage had a high impact on the set of discussed topics. To further investigate this point, we performed a more in-depth analysis of the data regarding the AstraZeneca vaccine. We observed how media coverage of blood clot–related side effects suddenly shifted the topic of public discussions regarding both the AstraZeneca and other vaccines. This became particularly evident when visualizing the most frequently discussed symptoms for the vaccines and comparing them month by month. Conclusions We present a tool connected with a web portal to monitor and display some key aspects of the public’s reaction to COVID-19 vaccines. The system also provides an overview of the opinions of the Twittersphere through graphic representations, offering a tool for the extraction of suspected adverse events from tweets with a deep learning model.
Collapse
Affiliation(s)
- Beatrice Portelli
- Department of Mathematics, Computer Science and Physics, University of Udine, via delle Scienze 206, Udine, IT
| | - Simone Scaboro
- Department of Mathematics, Computer Science and Physics, University of Udine, via delle Scienze 206, Udine, IT
| | - Roberto Tonino
- Department of Mathematics, Computer Science and Physics, University of Udine, via delle Scienze 206, Udine, IT
| | | | - Enrico Santus
- Decision Science and Advanced Analytics for MAPV & RA, Bayer, Bayer Pharmaceuticals, Whippany, US
| | - Giuseppe Serra
- Department of Mathematics, Computer Science and Physics, University of Udine, via delle Scienze 206, Udine, IT
| |
Collapse
|
14
|
Blane J, Bellutta D, Carley KM. Social-Cyber Maneuvers Analysis During the COVID-19 Vaccine Initial Rollout. J Med Internet Res 2022; 24:e34040. [PMID: 35044302 PMCID: PMC8903203 DOI: 10.2196/34040] [Citation(s) in RCA: 5] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/04/2021] [Revised: 12/18/2021] [Accepted: 01/08/2022] [Indexed: 01/25/2023] Open
Abstract
BACKGROUND During the time surrounding the approval and initial distribution of Pfizer-BioNTech's COVID-19 vaccine, large numbers of social media users took to used their platforms to voice opinions on the vaccine. They formed pro- and anti-vaccination groups towards the purpose of influencing behaviors to vaccinate or not to vaccinate. The methods of persuasion and manipulation for convincing audiences online can be characterized under a framework for social-cyber maneuvers known as the BEND maneuvers. Previous studies have been conducted on the spread of COVID-19 vaccine disinformation. However, these previous studies lacked comparative analyses over time on both community stances and the competing techniques of manipulating both the narrative and network structure to persuade target audiences. OBJECTIVE This study aimed to understand community response to vaccination by dividing Twitter data from the initial Pfizer-BioNTech COVID-19 vaccine rollout into pro-vaccine and anti-vaccine stances, identifying key actors and groups, and evaluating how the different communities use social-cyber maneuvers, or BEND maneuvers, to influence their target audiences and the network as a whole. METHODS COVID-19 Twitter vaccine data was collected using the Twitter API for one-week periods before, during, and six weeks after the initial Pfizer-BioNTech rollout (December 2020-January 2021). Bot identifications and linguistic cues were derived for users and tweets, respectively, to use as metrics for evaluating social-cyber maneuvers. ORA-PRO software was then used to separate the vaccine data into pro-vaccine and anti-vaccine communities and to facilitate identification of key actors, groups, and BEND maneuvers for a comparative analysis between each community and the entire network. RESULTS Both the pro-vaccine and anti-vaccine communities used combinations of the 16 BEND maneuvers to persuade their target audiences of their particular stances. Our analysis showed how each side attempted to build its own community while simultaneously narrowing and neglecting the opposing community. Pro-vaccine users primarily used positive maneuvers such as excite and explain messages to encourage vaccination and backed leaders within their group. In contrast, anti-vaccine users relied on negative maneuvers to dismay and distort messages with narratives on side effects and death and attempted to neutralize the effectiveness of the leaders within the pro-vaccine community. Furthermore, nuking through platform policies showed to be effective in reducing the size of the anti-vaccine online community and the quantity of anti-vaccine messages. CONCLUSIONS Social media continues to be a domain for manipulating beliefs and ideas. These conversations can ultimately lead to real-world actions such as to vaccinate or not to vaccinate against COVID-19. Moreover, social media policies should be further explored as an effective means for curbing disinformation and misinformation online. CLINICALTRIAL Not applicable.
Collapse
Affiliation(s)
- Janice Blane
- School of Computer Science, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, US
| | - Daniele Bellutta
- School of Computer Science, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, US
| | - Kathleen M Carley
- School of Computer Science, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, US
| |
Collapse
|
15
|
Dowrick A, Mitchinson L, Hoernke K, Mulcahy Symmons S, Cooper S, Martin S, Vanderslott S, Vera San Juan N, Vindrola‐Padros C. Re-ordering connections: UK healthcare workers' experiences of emotion management during the COVID-19 pandemic. SOCIOLOGY OF HEALTH & ILLNESS 2021; 43:2156-2177. [PMID: 34706107 PMCID: PMC8652548 DOI: 10.1111/1467-9566.13390] [Citation(s) in RCA: 13] [Impact Index Per Article: 4.3] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Received: 02/11/2021] [Revised: 09/10/2021] [Accepted: 09/26/2021] [Indexed: 05/15/2023]
Abstract
This paper examines the impact of disruptions to the organisation and delivery of healthcare services and efforts to re-order care through emotion management during the COVID-19 pandemic in the UK. Framing care as an affective practice, studying healthcare workers' (HCWs) experiences enables better understanding of how interactions between staff, patients and families changed as a result of the pandemic. Using a rapid qualitative research methodology, we conducted interviews with frontline HCWs in two London hospitals during the peak of the first wave of the pandemic and sourced public accounts of HCWs' experiences of the pandemic from social media (YouTube and Twitter). We conducted framework analysis to identify key factors disrupting caring interactions. Fear of infection and the barriers of physical distancing acted to separate staff from patients and families, requiring new affective practices to repair connections. Witnessing suffering was distressing for staff, and providing a 'good death' for patients and communicating care to families was harder. In addition to caring for patients and families, HCWs cared for each other. Infection control measures were important for limiting the spread of COVID-19 but disrupted connections that were integral to care, generating new work to re-order interactions.
Collapse
Affiliation(s)
- Anna Dowrick
- Nuffield Department of Primary Care Health SciencesUniversity of OxfordOxfordUK
| | - Lucy Mitchinson
- Marie Curie Palliative Care Research DepartmentUniversity College LondonLondonUK
| | - Katarina Hoernke
- Institute of Epidemiology and HealthcareUniversity College LondonLondonUK
| | | | - Silvie Cooper
- Institute of Epidemiology and HealthcareUniversity College LondonLondonUK
| | - Sam Martin
- Oxford Vaccine GroupChurchill HospitalUniversity of OxfordOxfordUK
| | | | - Norha Vera San Juan
- Health Service and Population Research DepartmentKing's College LondonLondonUK
| | - Cecilia Vindrola‐Padros
- Department of Targeted Intervention and Rapid Research Evaluation and Appraisal Lab (RREAL)University College LondonLondonUK
| |
Collapse
|
16
|
Usher K, Durkin J, Martin S, Vanderslott S, Vindrola-Padros C, Usher L, Jackson D. Public Sentiment and Discourse on Domestic Violence During the COVID-19 Pandemic in Australia: Analysis of Social Media Posts. J Med Internet Res 2021; 23:e29025. [PMID: 34519659 PMCID: PMC8489563 DOI: 10.2196/29025] [Citation(s) in RCA: 6] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/23/2021] [Revised: 08/02/2021] [Accepted: 09/04/2021] [Indexed: 12/25/2022] Open
Abstract
BACKGROUND Measuring public response during COVID-19 is an important way of ensuring the suitability and effectiveness of epidemic response efforts. An analysis of social media provides an approximation of public sentiment during an emergency like the current pandemic. The measures introduced across the globe to help curtail the spread of the coronavirus have led to the development of a situation labeled as a "perfect storm," triggering a wave of domestic violence. As people use social media to communicate their experiences, analyzing public discourse and sentiment on social platforms offers a way to understand concerns and issues related to domestic violence during the COVID-19 pandemic. OBJECTIVE This study was based on an analysis of public discourse and sentiment related to domestic violence during the stay-at-home periods of the COVID-19 pandemic in Australia in 2020. It aimed to understand the more personal self-reported experiences, emotions, and reactions toward domestic violence that were not always classified or collected by official public bodies during the pandemic. METHODS We searched social media and news posts in Australia using key terms related to domestic violence and COVID-19 during 2020 via digital analytics tools to determine sentiments related to domestic violence during this period. RESULTS The study showed that the use of sentiment and discourse analysis to assess social media data is useful in measuring the public expression of feelings and sharing of resources in relation to the otherwise personal experience of domestic violence. There were a total of 63,800 posts across social media and news media. Within these posts, our analysis found that domestic violence was mentioned an average of 179 times a day. There were 30,100 tweets, 31,700 news reports, 1500 blog posts, 548 forum posts, and 7 comments (posted on news and blog websites). Negative or neutral sentiment centered on the sharp rise in domestic violence during different lockdown periods of the 2020 pandemic, and neutral and positive sentiments centered on praise for efforts that raised awareness of domestic violence as well as the positive actions of domestic violence charities and support groups in their campaigns. There were calls for a positive and proactive handling (rather than a mishandling) of the pandemic, and results indicated a high level of public discontent related to the rising rates of domestic violence and the lack of services during the pandemic. CONCLUSIONS This study provided a timely understanding of public sentiment related to domestic violence during the COVID-19 lockdown periods in Australia using social media analysis. Social media represents an important avenue for the dissemination of information; posts can be widely dispersed and easily accessed by a range of different communities who are often difficult to reach. An improved understanding of these issues is important for future policy direction. Heightened awareness of this could help agencies tailor and target messaging to maximize impact.
Collapse
Affiliation(s)
- Kim Usher
- University of New England, Armidale, Australia
| | | | - Sam Martin
- Oxford Vaccine Group, University of Oxford, Oxford, United Kingdom
| | | | | | - Luke Usher
- Griffith University, Goldcoast, Australia
| | | |
Collapse
|