Analyzing Reddit Social Media Content in the United States Related to H5N1: Sentiment and Topic Modeling Study
Bibliographic record
Abstract
BACKGROUND: The H5N1 avian influenza A virus represents a serious threat to both animal and human health, with the potential to escalate into a global pandemic. Effective monitoring of social media during H5N1 avian influenza outbreaks could potentially offer critical insights to guide public health strategies. Social media platforms like Reddit, with their diverse and region-specific communities, provide a rich source of data that can reveal collective attitudes, concerns, and behavioral trends in real time. OBJECTIVE: This study aims to analyze Reddit comments from state-specific subreddits in the United States from the most recent outbreak period of 2022 to 2024 to (1) assess the sentiments expressed as the H5N1 outbreak progresses; (2) identify predominant topics discussed, particularly those corresponding to negative sentiments; and (3) explore correlations between these sentiments or topics and the severity and spread of the outbreak in respective regions. METHODS: We collected 2152 Reddit comments from 160 subreddits across 11 highly impacted states from February 2022 to July 2024. Outbreak data comprising almost 600 entries were obtained from the US Department of Agriculture database. Sentiment classification was performed using a fine-tuned Bidirectional Encoder Representations From Transformers (BERT) base model, and comments were categorized into 6 emotions: anger, fear, joy, love, sadness, and surprise, with a seventh "neutral" category added for low-confidence classifications. Topic modeling was conducted using BERTopic and latent Dirichlet allocation models. Statistical analyses included calculating correlations between sentiment intensity and outbreak severity levels and applying the Mann-Whitney U test to assess differences between sentiment categories. RESULTS: The findings illustrate that H5N1 unfolded in mostly discrete national waves and that only a subset of states-Minnesota and Iowa-experienced chronic, multiwave exposure, a pattern obscured in national aggregates. Sentiment intensity scoring revealed that although 90% (n=1931) of discourse was negative, emotions differed in how they tracked the epidemic: fear aligned weekly with real-time case counts (r=0.11), whereas anger, sadness, and even joy surged 3 weeks after the outbreak (r=0.20-0.24 after the lag was considered). When both the 3-week lag and an outlier month in terms of outbreak cases were adjusted for simultaneously, those associations strengthened further (overall r=0.223), showing how delayed reactions and anomalous surges can mask true sentiment-epidemiology links if left uncorrected. This defines the window in which risk communicators can pre-empt misinformation and economic anxiety. Topic modeling uncovered recurring themes of concern: avian flu culling, sharp egg-price hikes, and frustration over prolonged biosecurity measures. BERTopic provided more coherent and locally specific topics than latent Dirichlet allocation. CONCLUSIONS: Overall, these results underscore the critical role of social media analysis in understanding public reactions, including prevalent themes and sentiments, and guiding timely, targeted public health interventions during the H5N1 outbreak.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".