Trajectories of and spatial variations in HPV vaccine discussions on Weibo, 2018-2023: a deep learning analysis DOI Creative Commons
Wang You,

Haoyun Yang,

Zhijun Ding

и другие.

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2023, Номер unknown

Опубликована: Дек. 9, 2023

Summary Research in context Evidence before this study We first searched PubMed for articles published until November 2023 with the keywords “(“HPV”) AND (“Vaccine” or “Vaccination”) (“Social Media”)”. identified about 390 studies, most of which were discussions on potentials feasibility social media HPV vaccination advocacy research, manual coding-driven analyses text (eg., tweets) vaccines emerged platforms. When we added keyword “Machine Learning”, only 12 several them using AI-driven approach, such as deep learning, machine and natural language process, to analyze extensive data public perceptions perform monitor platforms, X (Twitter) Reddit. All these studies are from English-language platforms developed countries. No date has monitored developing countries including China. Added value This is deep-learning monitoring expressed Chinese (Weibo our case), revealing key temporal geographic variations. found a sustained high level positive attitude towards exposure norms facilitating among Weibo users, lower national prevalence negative attitude, perceived barriers accepting vaccination, misinformation indicating achievement relevant health communication. High practical was associated relatively insufficient vaccine accessibility China, suggesting systems should prioritize addressing issues supply. Lower perception male higher hesitancy 2-valent vaccine, provincial-level spatial cluster indicate that tailored strategies need be formed targeting specific population, areas, type. Our practice shows realizing surveillance potential listening context. Leveraging recent advances approach could cost-effective supplement existing techniques. Implications all available evidence highlights learning-driven convenient effective identifying emerging trends inform interventions. As techniques, it particularly helpful timely communication resource allocation at multiple levels. Key stakeholders officials maintain focus education highlighting risks consequences infections, benefits safety types vaccines; aim resolve accessibility. A proposed research area further development learning models analyzing Background rate low Understanding multidimensional impetuses by individuals essential. assess perceptions, barriers, facilitators platform Weibo. Methods collected posts regarding between 2018 2023. annotated 6,600 manually according behavior change theories, subsequently fine-tuned annotate collected. Based results models, conducted attitudes its determinants. Findings Totally 1,972,495 vaccines. Deep reached predictive accuracy 0.78 0.96 classifying posts. During 2023, 1,314,510 (66.6%) classified attitudes. And 224,130 (11.4%) misinformation, 328,442 (16.7%) vaccines, 580,590 (29.4%) vaccination. The increased 15.8% March 79.1% mid-2023 (p < 0.001), declined 36.6% mid-2018 10.7% (P .001). Central regions exhibited norms, whereas Shanghai, Beijing megacities northeastern showed misinformation. Positive significantly (65.7%), than 4-valent 9-valent (79.6% 74.1%). Interpretation Social represents promising can enable strategies.

Язык: Английский

A Comparison of ChatGPT and Fine-Tuned Open Pre-Trained Transformers (OPT) Against Widely Used Sentiment Analysis Tools: Sentiment Analysis of COVID-19 Survey Data DOI Creative Commons
Juan Antonio Lossio-Ventura, Rachel Weger, Angela Y. Lee

и другие.

JMIR Mental Health, Год журнала: 2023, Номер 11, С. e50150 - e50150

Опубликована: Ноя. 17, 2023

Background Health care providers and health-related researchers face significant challenges when applying sentiment analysis tools to free-text survey data. Most state-of-the-art applications were developed in domains such as social media, their performance the health context remains relatively unknown. Moreover, existing studies indicate that these often lack accuracy produce inconsistent results. Objective This study aims address of comparative on applied data COVID-19. The objective was automatically predict sentence for 2 independent COVID-19 sets from National Institutes Stanford University. Methods Gold standard labels created a subset each set using panel human raters. We compared 8 both evaluate variability disagreement across tools. In addition, few-shot learning explored by fine-tuning Open Pre-Trained Transformers (OPT; large language model [LLM] with publicly available weights) small annotated zero-shot ChatGPT (an LLM without weights). Results comparison revealed high evaluated OPT demonstrated superior performance, outperforming all other outperformed OPT, exhibited higher 6% F-measure 4% 7%. Conclusions demonstrates effectiveness LLMs, particularly approaches, These results have implications saving labor improving efficiency tasks, contributing advancements field automated analysis.

Язык: Английский

Процитировано

28
Vaccination, Public Health and Health Communication: A Network of Connections to Tackle Global Challenges DOI Creative Commons
Antonella Arghittu, Giovanna Deiana, Marco Dettori

и другие.

Vaccines, Год журнала: 2025, Номер 13(3), С. 245 - 245

Опубликована: Фев. 27, 2025

Vaccination constitutes one of the most significant milestones in history Public Health [...]

Язык: Английский

Процитировано

0
A thematic analysis of Chinese university students’ digital art therapy experience: an embodied perspective DOI Creative Commons
Chenmu Xie

Cogent Arts and Humanities, Год журнала: 2025, Номер 12(1)

Опубликована: Апрель 19, 2025

Язык: Английский

Процитировано

0
COVID-19 information seeking and individuals’ protective behaviors: examining the role of information sources and information content DOI Creative Commons
Xuefeng Zhang, Lin Du, Yelin Huang

и другие.

BMC Public Health, Год журнала: 2024, Номер 24(1)

Опубликована: Янв. 29, 2024

Abstract Background Seeking COVID-19 information promotes individuals to adopt preventive behaviors, including wearing a mask, social distancing, staying away from risky places, and washing hands. This study aims investigate which sources relied on in seeking further examine their roles individuals’ adoption of behaviors. Methods Through statistical analysis 1027 valid responses citizens different Chinese cities 2022 the self-designed items an online survey, this identified preferred content COVID-19. Regarding content, used multiple regression associations with applied fuzzy-set qualitative comparative (fsQCA) explore configurations that increase likelihood adopting Results Individuals about newest prevention control policies, precautions treatment, symptoms workplace community, media, live streaming services. Additionally, behaviors were positively related community ( β = 0.202, p <.001), services 0.089, <.01), government department websites 0.079, <.05), television 0.073, news media 0.069, but negatively associated newspapers =-0.087, <.05). treatments 0.211, policies 0.173, 0.152, official rumor-dispelling 0.082, <.05) had positive relationship In addition, fsQCA results presented eight promote The total coverage solution consistency values 0.869 0.987, respectively. Furthermore, interpersonal sources, played essential role increasing Conclusions Our findings demonstrated seek various sources. direct degree association vary source content. Information could combinatorially through several configurations.

Язык: Английский

Процитировано

3
Internet of Things and Bigdata Analytics in Preventive Healthcare: A Synthetic Review DOI Open Access

Urška Šajnovič,

Helena Blažun Vošner, Jernej Završnik

и другие.

Опубликована: Авг. 7, 2024

Background: IoT and Big Data are newer technologies that can provide substantial support for healthcare systems help overcome their shortcomings. The aim of this paper was to analyze literature production descriptively, thematically, chronologically from an interdisciplinary perspective in a holistic way identify the most prolific research entities themes. Methods: synthetic knowledge synthesis qualitatively quantitatively analyzes through combination descriptive bibliometrics, bibliometric mapping content analysis. For analysis Scopus database used. Results: In database, 2272 publications were found, which published between 1985 June 10, 2024. first article field 1985. Until 2012 steady increasing, after exponential growth began, reached its peak 2023. productive countries USA, India, China, United Kingdom, South Korea, Germany Italy. resulted 8 themes (four computer science four medicine) 21 thematic concepts (eight 13 medicine). Conclusions: results show have become key preventive health. study outcomes might represent starting point further development combines multidisciplinary aspects healthcare.

Язык: Английский

Процитировано

3
Deep Learning Analysis of COVID-19 Vaccine Hesitancy and Confidence Expressed on Twitter in 6 High-Income Countries: Longitudinal Observational Study DOI Creative Commons
Xinyu Zhou, Suhang Song, Ying Zhang

и другие.

Journal of Medical Internet Research, Год журнала: 2023, Номер 25, С. e49753 - e49753

Опубликована: Окт. 3, 2023

An ongoing monitoring of national and subnational trajectory COVID-19 vaccine hesitancy could offer support in designing tailored policies on improving uptake.We aim to track the temporal spatial distribution confidence expressed Twitter during entire pandemic period major English-speaking countries.We collected 5,257,385 English-language tweets regarding vaccination between January 1, 2020, June 30, 2022, 6 countries-the United States, Kingdom, Australia, New Zealand, Canada, Ireland. Transformer-based deep learning models were developed classify each tweet as intent accept or reject belief that is effective unsafe. Sociodemographic factors associated with States analyzed using bivariate multivariable linear regressions.The countries experienced similar evolving trends confidence. On average, prevalence decreased from 71.38% 44,944 March 2020 34.85% 48,167 2022 fluctuations. The believing vaccines be unsafe continuously rose by 7.49 times (2.84% tweets) (21.27% tweets). varied country, manufacturer, states within a country. democrat party higher significantly lower across US states.COVID-19 evolved influenced development viruses pandemic. Large-scale self-generated discourses social media provide cost-efficient approach routine hesitancy.

Язык: Английский

Процитировано

7
Emoji and visual complexity in health information design: A moderated serial mediation model DOI
Tingyi S. Lin, Yue Luo

Telematics and Informatics, Год журнала: 2023, Номер 85, С. 102065 - 102065

Опубликована: Окт. 16, 2023

Язык: Английский

Процитировано

5
The Collaboverse: A Collaborative Data-Sharing and Speech Analysis Platform DOI

Justin Dvorak,

Frank Boutsen

Journal of Speech Language and Hearing Research, Год журнала: 2024, Номер 67(10S), С. 4137 - 4156

Опубликована: Июль 12, 2024

Purpose: Collaboration in the field of speech-language pathology occurs across a variety digital devices and can entail usage multiple software tools, systems, file formats, even programming languages. Unfortunately, gaps between laboratory, clinic, classroom emerge part because siloing data workflows, as well divide users. The purpose this tutorial is to present Collaboverse, web-based collaborative system that unifies these domains, describe application tool common tasks pathology. In addition, we demonstrate its utility machine learning (ML) applications. Method: This outlines key concepts divide, management, distributed computing, ML. It introduces Collaboverse workspace for researchers, clinicians, educators who wish improve their network leverage advanced computation abilities. also details an ML approach prosodic analysis. Conclusions: shows promise narrowing capable generating clinically relevant data, specifically area prosody, whose computational complexity has limited widespread analysis research clinic alike. it includes augmentative alternative communication app allowing visual, nontextual communication.

Язык: Английский

Процитировано

1
Internet of Things and Big Data Analytics in Preventive Healthcare: A Synthetic Review DOI Open Access

Urška Šajnovič,

Helena Blažun Vošner, Jernej Završnik

и другие.

Electronics, Год журнала: 2024, Номер 13(18), С. 3642 - 3642

Опубликована: Сен. 12, 2024

Background: The IoT and big data are newer technologies that can provide substantial support for healthcare systems, helping them overcome their shortcomings. aim of this paper was to analyze the relevant literature descriptively, thematically, chronologically from an interdisciplinary perspective in a holistic way identify most prolific research entities themes. Methods: Synthetic knowledge synthesis qualitatively quantitatively analyzes production through combination descriptive bibliometrics, bibliometric mapping, content analysis. For analysis, Scopus database used. Results: In database, 2272 publications were found; these published between 1985 10 June 2024. first article field 1985. Until 2012, such steadily increasing; after that, exponential growth began, peaking 2023. productive countries United States, India, China, Kingdom, South Korea, Germany, Italy. analysis resulted eight themes (four computer science four medicine) 21 thematic concepts (8 13 medicine). Conclusions: results show have become key employed preventive healthcare. study outcomes might represent starting point further development combines multidisciplinary aspects

Язык: Английский

Процитировано

1
Prioritization of Vaccines for Introduction in the National Immunization Program in the Republic of Korea DOI Creative Commons
Won Suk Choi,

Yeonhee Sung,

Jimin Kim

и другие.

Vaccines, Год журнала: 2024, Номер 12(8), С. 886 - 886

Опубликована: Авг. 4, 2024

This study presents a framework for determining the prioritization of vaccine introduction in National Immunization Program (NIP) Republic Korea, with focus on case examples assessed 2021 and 2023. We describe predefined criteria evaluating vaccines NIP established process Korea. These included disease characteristics, rationality efficiency resource allocation, acceptance immunization. The prioritizing involved several sequential steps: demand survey, evidence collection, preliminary evaluation, priority decision making. In 2023, 14 25 committee members participated NIP, respectively. Overall, 13 19 candidates were 2023 evaluations, Through Delphi survey consensus processes, order was determined: vaccination against Rotavirus infection top 2021, while Influenza 4v (for chronic patients) took precedence demonstrates an evidence-based decision-making within healthcare field. outlined approach may provide valuable guidance policymakers other countries seeking to prioritize inclusion new their NIP.

Язык: Английский

Процитировано

0