Discourse on COVID-19 and illicit drug toxicity on social media

This Health System Impact Fellowship is co-funded by ­CIHR Institute of Population and Public Health (CIHR-IPPH), Michael Smith Health Research BC, and the BC Centre for Disease Control (health system partner), to help build BC’s health policy research capacity for the integration of policy research into decision-making.

The illicit drug overdose crisis in North America has had a profound impact on individuals, families and communities, often leading to premature loss of life and lowering of life expectancy. Since 2016, British Columbia (BC) has been experiencing an epidemic of toxic drug supply leading to a large increase in the number of drug overdose events and related deaths. Coronavirus disease 2019 (COVID-19) and measures taken to limit the transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus that causes COVID-19, have directly and indirectly disrupted access to healthcare and social services worldwide, including harm reduction and social support services. Since the beginning of the COVID-19 pandemic, BC witnessed large increases in drug overdose related deaths, with 2021 being the deadliest year. This requires scaling up of existing interventions and introducing new, targeted interventions to address the overdose crisis. The level and type of response is also affected by the perception of the population towards health issues and debate on intervention options. The information available on social media could help decision makers understand the public discourse about opioid use and intervention options. In this project, the overall aim is to understand public perceptions and discourse related to overdose in social media using Artificial Intelligence (AI) methods and techniques to inform the overdose response.

Source: CIHR Funding Decisions Database

End of Award Update – December 2024

Results

1. We identified five predominant themes in Long COVID conversations:

• Long COVID in People, Including Children, in the Context of Vaccination – Focused on objective perspectives such as research, monitoring, experiences, and news items highlighting the significant impact of Long COVID symptoms in adults and children, including the potential benefits of vaccination in reducing Long COVID risk.
• Duration and Suffering Associated with Long COVID – Captures the subjective experiences of suffering and frustration associated with Long COVID, emphasizing the prolonged duration of symptoms.
• Persistent Symptoms of Long COVID – Focuses on the persistent symptoms of Long COVID, highlighting both formal and informal advocacy and awareness-raising efforts.
• Need for Research on Long COVID Treatment – Represents calls to address the plight of individuals with Long COVID, either by researchers or governments, expressing frustration over neglect and ongoing symptoms.
• Measuring Long COVID Symptoms –Discusses the large number of people affected by or projected to be affected by Long COVID symptoms, along with associated societal impacts.

2. Sentiment Trends:

• Dynamic Emotional Responses: Our sentiment analysis revealed fluctuations in positive, negative, and neutral sentiments over the months, corresponding to significant events and announcements related to Long COVID.
• Regional Variations: Different regions exhibited unique sentiment trends, reflecting varying public concerns and levels of awareness.

3. Regional Comparisons:

• Canada and Europe: Higher prevalence of discussions about personal experiences and research on Long COVID.
• United States: More diverse and sometimes less clear topics, with notable discussions around government policies and advocacy efforts.

4. Publications:

• Journal Publication Available: Monthly Variations in Long COVID Discourse on Twitter – Full Paper
• Upcoming Publication: A second paper detailing the comparative analysis of topic evolution and sentiment trends across the regions is currently being prepared for submission and will be available soon.

Impact

Our project has made significant contributions to understanding public perceptions and sentiments surrounding Long COVID, with several immediate impacts:

• Enhanced Public Health Monitoring: By mapping Long COVID discourse on Twitter, our research provides valuable insights into public concerns, misinformation, and areas requiring attention. This enables health authorities to tailor their communication strategies more effectively.
• Informed Policy Development: Insights into public sentiment regarding government policies and advocacy efforts have informed policymakers about the effectiveness and reception of their initiatives, guiding future policy adjustments to better address Long COVID challenges.
• Awareness and Advocacy: The identification of themes related to research needs and advocacy highlights the growing demand for increased support and resources for individuals affected by Long COVID, prompting organizations to prioritize Long COVID in their agendas.
• Academic Contributions: Our study advances the methodology of analyzing health-related social media discourse by integrating Contextualized Topic Modeling and Large Language Models, offering a robust framework for future research in similar domains.

Potential Influence

Looking ahead, our research holds substantial potential to shape future health strategies and interventions both locally and globally:

• Enhanced Surveillance Systems: Integrating social media analysis into public health surveillance can provide real-time insights, enabling quicker responses to emerging public concerns about Long COVID.
• Targeted Public Health Campaigns: Understanding regional sentiment and topic prevalence allows for the creation of personalized and effective public health campaigns that address specific local needs and misconceptions.
• Global Health Strategies: By comparing discourse across multiple regions, our findings can contribute to international collaborations aimed at managing and mitigating Long COVID on a global scale.
• Improved Patient Support: Insights into the experiences and persistent symptoms of Long COVID can inform the development of targeted support programs and resources for affected individuals.