D1.276 - Social #anaphylaxis: how social media reveals shifting trends in public health concerns

Social media platforms are invaluable tools for gathering real-time data on public health events, including life-threatening reactions such as anaphylaxis. Leveraging artificial intelligence, this study aimed to analyze trends in online conversations about anaphylaxis, particularly during the COVID-19 pandemic and global vaccination rollout.

We used the Symplur artificial intelligence engine to analyze all tweets posts on X with the hashtag #anaphylaxis from January 2019 to December 2020. This analysis included quantitative data, message characteristics, link sources, and user profiles. Additionally, we assessed the hashtag score, a metric for evaluating hashtag popularity and engagement.

A total of 22,696 posts with the hashtag #anaphylaxis were identified (2019: 13,812; 2020: 11,095). Until late March 2020, tweet volumes remained stable. After the WHO declared the pandemic, tweets decreased notably: April 2019:1,322 tweets, compared to 889 in April 2020. A rebound occurred in December 2020 (1,882 tweets), coinciding with the global vaccination rollout and reports of vaccine-related anaphylaxis (Figure 1).

Tweets originated from diverse global sources, including professional organizations physicians, researchers, governmental bodies (e.g., NIAID News), and anonymous users (6,328 posts in 2019, 6,121 in 2020). This participation highlighted an international communication network.

A subset of tweets linked to high-impact journal articles, such as those from the European Journal of Allergy and Clinical Immunology and Science Immunology, showcasing Twitter's role in disseminating evidence-based information. The highest hashtag scores were from accounts with large followings, including FDA (415,535 followers), NIAID News (65,088 followers), and physician Kevin Pho (161,675 followers).

This study demonstrates how Twitter (now X) provides insights into public health trends, such as the decline in #anaphylaxis activity during 2020 (reflecting a shift in attention to COVID-19) and the rebound during the vaccination campaign. Tools like Symplur are useful for efficiently analyzing large datasets, identifying meaningful patterns, and fostering a deeper understanding of public health dynamics.