Artificial intelligence-enabled social media analysis for pharmacovigilance of COVID-19 vaccinations in the United Kingdom: Observational Study
Hussain, Zain; Sheikh, Zakariya; Tahir, Ahsen; Dashtipour, Kia; Gogate, Mandar; Sheikh, Aziz; Hussain, Amir
Dr Kia Dashtipour K.Dashtipour@napier.ac.uk
Dr. Mandar Gogate M.Gogate@napier.ac.uk
Senior Research Fellow
Prof Amir Hussain A.Hussain@napier.ac.uk
The roll-out of vaccines for SARS-CoV-2 in the United Kingdom, started in December 2020. Uptake has been high, and there has been a subsequent reduction in infections, hospitalisations and deaths in vaccinated individuals. However, vaccine hesitancy remains a concern, in particular relating to adverse effects following immunisation (AEFI). Social media analysis has the potential to inform policymakers on AEFIs being discussed by the public, and on public attitudes towards the national immunisation campaign.
We sought to describe the frequency and nature of COVID-19 AEFI related mentions on social media, and provide insights on public sentiment towards COVID-19 vaccines, in the UK.
We extracted and analysed over 121,406 relevant Twitter and Facebook posts, from 8 December 2020 to 30 April 2021. These were thematically filtered using a two-step approach, initially using COVID-related keywords and then using vaccines and manufacturer related keywords. We identified AEFI related keywords and modelled their word frequency to monitor their trends over two-week periods. We also adapted and utilised our recently developed hybrid ensemble model, which combines a state-of-the-art lexicon rule based and deep learning based approaches, to analyse sentiment trends relating to the main vaccines available in the UK.
We identified an increasing trend in the number of mentions for each AEFI on social media over the period of study. The most frequent AEFI mentions were found to be: appetite (14%, n=79,132), allergy (9%, n=53,924), injection site (10%, n=56,152), and clots (8%, n=43,907) related symptoms. We also found more rarely reported AEFIs, such as Bell’s Palsy (2%, n=11,909) and Guillain Barre Syndrome (GBS) (2%, n=9,576), being discussed as frequently as more well-known side effects, such as headache (2%, n=10,641), fever (2%, n=12,707) and diarrhea (3%, n=16,559). Overall, we found public sentiment towards vaccines and their manufacturers to be largely positive (58%), with negative (22%) and neutral (19%) sentiment equally split. The sentiment trend was relatively steady over time and had minor variations, likely based on political and regulatory announcements and debates.
The most frequently discussed COVID-19 AEFIs on social media were found to be broadly consistent with those reported in the literature and by government pharmacovigilance. We also detected potential safety signals from our analysis, that have been detected elsewhere and are currently being investigated. As such, we believe our findings support the use of social media analysis to provide a complementary data source to conventional knowledge sources being used for pharmacovigilance purposes.
Hussain, Z., Sheikh, Z., Tahir, A., Dashtipour, K., Gogate, M., Sheikh, A., & Hussain, A. (2022). Artificial intelligence-enabled social media analysis for pharmacovigilance of COVID-19 vaccinations in the United Kingdom: Observational Study. JMIR Public Health and Surveillance, 8(5), Article e32543. https://doi.org/10.2196/32543
|Journal Article Type||Article|
|Acceptance Date||Feb 8, 2022|
|Online Publication Date||May 27, 2022|
|Deposit Date||Apr 26, 2022|
|Publicly Available Date||Jun 23, 2022|
|Journal||JMIR Public Health and Surveillance|
|Peer Reviewed||Peer Reviewed|
|Keywords||COVID-19, artificial intelligence, deep learning, Facebook, health informatics, natural language processing, public health, sentiment analysis, social media, Twitter, infodemiology, vaccination|
Artificial Intelligence-enabled Social Media Analysis For Pharmacovigilance Of COVID-19 Vaccinations In The United Kingdom: Observational Study
Publisher Licence URL
You might also like
Arabic sentiment analysis using dependency-based rules and deep neural networks
Intrusion Detection Framework for the Internet of Things Using a Dense Random Neural Network
Detecting Alzheimer’s Disease Using Machine Learning Methods