The use of artificial intelligence methods in Reddit to investigate opioid use: a scoping review
Bibliographic record
Abstract
Opioid use disorder (OUD) is a chronic condition that affects more than 40 million people worldwide1. In 2019, the Global Burden of Disease study estimated 494,000 deaths and 30.9 million years of "healthy" life lost because of premature death and disability attributable to the use of drugs. Twenty-six percent (128,000) of these deaths were attributed to drug use disorders, of which OUD contributed to 69%. Most of the disability-adjusted life years (DALYs) (71%) attributed to drug use disorders were caused by OUD, corresponding to 12.9 million DALYs2. The United States (U.S.) accounts for a large share of these deaths, with over 107,000 deaths directly attributable to overdose in 2021, of which 75% were related to opioid use3. Among the 2.5 million people aged 12 or older with a past year of OUD in the U.S., only 11.2% (or 278,000 people) received medications for OUD (MOUD) in the past year4 and among people with OUD (PWOUD) receiving MOUD, drug use and relapse are the leading cause of death5–8. Understanding experiences, concerns, challenges, and sources of support among people who use opioids is the first step to designing interventions tailored to their needs. Social media platforms represent an important and accessible source of community support for people who use opioids, given they facilitate getting technical "know-"ow" and support from peers. These platforms often allow for anonymous participation, leading to authentic accounts of both positive and negative experiences with drugs (including in the context of treatment) and daily life situations that may impact their physical and mental health and that may be addressed through appropriate interventions. Reddit, a community-based social media platform where people participate in discussions anonymously, has been studied as a rich virtual place gathering people willing to share their experiences with opioids and related issues. This platform is one of the most popular online platforms of interaction and has provided a space for exchange and discussion since 2005, mainly used by English speakers9. Given the high rate of OUD in the U.S. and Canada, Australia, and the United Kingdom, it represents a valuable source of data to better understand the daily experiences of people who use opioids in these settings. Analyzing such large amounts of data is challenging. Beyond the traditional qualitative approach to analyzing textual data, artificial intelligence methods such as natural language processing, sentiment analysis, and supervised and unsupervised machine learning techniques have been used to facilitate content extraction, unveil discussion topics, and learn and predict "behaviors". These techniques enable the use of a massive textual corpus and the exploration of a multitude of research questions based on the perspectives of people sharing their experiences on social media platforms such as Reddit. This study aims to rigorously document how artificial intelligence methods have been applied to Reddit forums to study opioid use. More specifically, we are interested in mapping the main questions asked in these studies, their overarching goals and key objectives, the methodologies and software used, and their main limitations. We also aim to systematically collect information on a dictionary or key terms used in these studies to synthesize a comprehensive dictionary of embedding words allowing for the broader research community to rigorously investigate opioid use in Reddit using textual pre-processing algorithms.
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 | 0.019 | 0.075 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.006 | 0.007 |
| Bibliometrics | 0.025 | 0.020 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.005 | 0.004 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.009 | 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".