Citation Contamination of Systematic Review Literature in the Life Sciences by Paper Mills
Bibliographic record
Abstract
Abstract Importance Paper mills are systematic fraud organizations that mass-produce fabricated papers, submit them under researchers’ names, and profit by charging fees or selling authorship. Their products have infiltrated scientific literature databases, yet the academic community remains uncertain about their potential impact on scholarly research. Objective Our study investigates the citation contamination of systematic reviews in the life sciences by paper mill articles, aiming to determine whether these fabricated papers undermine the status of systematic reviews as the “gold standard” of evidence synthesis. Evidence Review We conducted a cross-sectional study of 100,000 systematic reviews published in the life sciences between 2013 and 2023, as indexed in the Web of Science. We extracted their references and matched them against retracted articles in the Retraction Watch Database, specifically those retracted due to paper mills. Descriptive statistical analyses were used to characterize the features of contaminated systematic reviews, including their subject areas, journals of publication, and citation patterns. Findings A total of 179 systematic reviews were contaminated by paper mill articles, representing a contamination rate of 0.179%. Although the overall extent of contamination was small, an increasing trend was observed. Notably, 61 citations occurred after the articles had been retracted. Oncology was the most severely affected field. Four systematic reviews each cited five or more paper mill articles; all were published by journals under the same academic publisher. Among those citing three or more paper mill articles, 23 (76.67%) were published in journals ranked in the top 50% of their respective fields by impact factor (i.e., those classified as Q1 or Q2). Conclusions and Relevance Systematic reviews, as comprehensive syntheses of high-quality evidence, must not incorporate systematically fabricated articles. The scientific community should remain vigilant about the growing trend of paper mill contamination in life sciences systematic reviews. Correcting and retracting previously contaminated reviews, and developing new tools to help researchers identify potential paper mill articles in literature databases, will be essential steps to ensure the integrity of evidence synthesis. Key Points Question Do paper mill articles contaminate systematic reviews in the life sciences, and what is their potential impact on the integrity of evidence synthesis? Findings In a cross-sectional study of 100,000 systematic reviews published between 2013 and 2023, 0.179% were contaminated by paper mill articles, with an increasing trend observed over time. Oncology was the most affected field, and 61 post-retraction citations were identified. Meaning These findings underscore the need for heightened vigilance, systematic correction of contaminated reviews, and development of tools to identify paper mill articles to safeguard the integrity of evidence synthesis.
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
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gemma | MetaresearchResearch integrityBibliometrics Domain: Evaluation · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | low |
| gpt | MetaresearchBibliometricsResearch integrity Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | low |
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 | 0.251 | 0.715 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.004 | 0.005 |
| Bibliometrics | 0.043 | 0.064 |
| Science and technology studies | 0.002 | 0.005 |
| Scholarly communication | 0.008 | 0.006 |
| Open science | 0.003 | 0.006 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 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, unvalidatedLabeled directly by 2 models reading the full record.
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".