Utilizing Machine Learning for Early Intervention and Risk Management in the Opioid Overdose Crisis
Bibliographic record
Abstract
ABSTRACT This systematic review and meta‐analysis seek to identify prevalent machine learning (ML) models applied to outcomes related to illicit opioid use. Following PRISMA guidelines, we reviewed databases including MEDLINE, Embase, CINAHL, PsycINFO, and Web of Science, yielding 10,666 records. Of these, 6029 were unique, leading to 155 full‐text publications, with 69 studies meeting inclusion criteria. The inclusion criteria focused on two primary themes: the application of artificial intelligence and machine learning techniques, and opioid related substance use outcomes. The meta‐analysis focused on Area Under the Receiver Operating Characteristic curve (AUC/AUROC). Most of the studies used classification models and evaluated them using the AUC metric. Cohen's d effect sizes were 1.22 for logistic regression (AUC = 0.806), 1.26 for decision trees/random forests (AUC = 0.814), 1.54 for deep learning (AUC = 0.862), and 1.27 for boosting algorithms (AUC = 0.815). Regarding outcomes, effect sizes were 1.42 for opioid use disorder (OUD) (AUC = 0.842), 1.37 for opioid overdoses (AUC = 0.842), and 1.25 for risk of drug use (AUC = 0.812). The study reveals the efficacy of ML in illicit opioid use, with a notable predominance of supervised ML models, particularly Logistic Regression. The underutilization of regression models, despite their potential in outcome quantification, was surprising. Deep learning emerged as the most effective model, demonstrating the complexity of data in addiction psychiatry. ML algorithms provide a powerful framework for informed decision‐making in addiction care, leading toward personalized medicine and reducing unregulated drug use and related harms.
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 | 0.014 | 0.040 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.007 | 0.013 |
| Bibliometrics | 0.004 | 0.003 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 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".