Online Conversation Monitoring to Understand the Opioid Epidemic: Epidemiological Surveillance Study
Bibliographic record
Abstract
BACKGROUND: Between 2016 and 2017, the national mortality rate involving opioids continued its escalation; opioid deaths rose from 42,249 to 47,600, bringing the public health crisis to a new height. Considering that 69% of adults in the United States use online social media sites, a resource that builds a more complete understanding of prescription drug misuse and abuse could supplement traditional surveillance instruments. The Food and Drug Administration has identified 5 key risks and consequences of opioid drugs-misuse, abuse, addiction, overdose, and death. Identifying posts that discuss these key risks could lead to novel information that is not typically captured by traditional surveillance systems. OBJECTIVE: The goal of this study was to describe the trends of online posts (frequency over time) involving abuse, misuse, addiction, overdose, and death in the United States and to describe the types of websites that host these discussions. Internet posts that mentioned fentanyl, hydrocodone, oxycodone, or oxymorphone were examined. METHODS: Posts that did not refer to personal experiences were removed, after which 3.1 million posts remained. A stratified sample of 61,000 was selected. Unstructured data were classified into 5 key risks by manually coding for key outcomes of misuse, abuse, addiction, overdose, and death. Sampling probabilities of the coded posts were used to estimate the total post volume for each key risk. RESULTS: Addiction and misuse were the two most commonly discussed key risks for hydrocodone, oxycodone, and oxymorphone. For fentanyl, overdose and death were the most discussed key risks. Fentanyl had the highest estimated number of misuse-, overdose-, and death-related mentions (41,808, 42,659, and 94,169, respectively). Oxycodone had the highest estimated number of abuse- and addiction-related mentions (3548 and 12,679, respectively). The estimated volume of online posts for fentanyl increased by more than 10-fold in late 2017 and 2018. The odds of discussing fentanyl overdose (odds ratios [OR] 4.32, 95% CI 2.43-7.66) and death (OR 5.05, 95% CI 3.10-8.21) were higher for social media, while the odds of discussing fentanyl abuse (OR 0.10, 95% CI 0.04-0.22) and addiction (OR 0.24, 95% CI 0.15-0.38) were higher for blogs and forums. CONCLUSIONS: Of the 5 FDA-defined key risks, fentanyl overdose and death has dominated discussion in recent years, while discussion of oxycodone, hydrocodone, and oxymorphone has decreased. As drug-related deaths continue to increase, an understanding of the motivations, circumstances, and consequences of drug abuse would assist in developing policy responses. Furthermore, content was notably different based on media origin, and studies that exclusively use either social media sites (such as Twitter) or blogs and forums could miss important content. This study sets out sustainable, ongoing methodology for surveilling internet postings regarding these drugs.
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 | 0.004 | 0.012 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".