Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 52 of the 52 references it cites were found in a registry or an index; 12 of the 37 sources its claims cite were checked for support.

Fake publications in biomedical science: red-flagging method indicates mass production

Open access
Sabel, Bernhard A., Knaack, Emely, Gigerenzer, Gerd, Bilc, Mirela-Ioana · 2026 · Naunyn-Schmiedeberg's Archives of Pharmacology

Abstract

Integrity of academic publishing is increasingly undermined by fake publications massively produced by commercial “editing services” (so-called “paper mills”). These services use AI-supported production techniques at scale and sell fake publications to students, scientists, and physicians under pressure to advance their careers. Because the scale of fake publications in biomedicine is unknown, we developed an easy-to-apply rule to red-flag potentially fake publications and estimate their number. After analyzing questionnaires sent to authors of published papers, we developed simple classification rules and tested them in a 9-step bibliometric analysis in a sample of 17,120 publications listed in PubMed®. We first validated various simple rules and finally applied a multifactorial tallying rule comparing 400 known fakes with 400 random (presumed) non-fakes. This rule was then applied to 1,000 random publications each from 2020 and 2023. The multifactorial tallying rule was the best red-flagging method, with a 94% sensitivity and only a 11.5% false-alarm rate. The rate of red-flagged articles increased during the last decade, reaching an estimated 14.9% in 2020 and 16.3% in 2023. Countries with the highest proportion of read-flagged publications were China, India, Iran, Russia, and Turkey, with China and India the largest absolute contributors globally. Applying Bayes’ rule resulted in an estimate of 5.8% actual fakes in the biomedical literature. Given 1.86 million Scimago-listed biomedical publications in 2023, we estimate the actual number of true fakes at 107.800 articles per year, growing steadily. Scientific publications in biomedicine can be red-flagged as potentially fake using fast-and-frugal classification rules to earmark them for subsequent scrutiny. When applying Bayes´rule, the annual true scale of fake publishing in biomedicine is about 19 times that of the 5.671 biomedicine retractions in 2023. This scale of fraudulent publishing is concerning as it can damage trust in science, endanger public health, and impact economic spending. But fake detection tools can enable retractions of fake publications at scale and help prevent further damage to the permanent scientific record.

Overview

CoChat Summary

TL;DR

The paper develops and validates a fast, multifactorial red-flagging rule for detecting potentially fake biomedical publications, reporting 94% sensitivity and an 11.5% false-alarm rate against 400 known fakes and 400 presumed non-fakes. Applying it to 1,000 publications each from 2020 and 2023 yields red-flag rates of 14.9% and 16.3%, and Bayes’ rule estimates 5.8% true fakes, or about 107,800 articles per year in 2023.

Key findings

  • The multifactorial tallying rule was the best red-flagging method, with 94% sensitivity and an 11.5% false-alarm rate; the simpler email-hospital rule had 85% sensitivity and false-alarm rates of 1% (2000) or 8% (2010).
  • Red-flagged publication rates increased over the last decade, reaching 14.9% in 2020 and 16.3% in 2023. Countries with the highest within-country proportions were Russia (66.7%), India (54.2%), China (38.7%), Iran (29.6%), and Turkey (20.8%); China (42.3%) and India (33.0%) were the largest between-country contributors.
  • Using Bayes’ rule, the estimated rate of true fakes was 5.8%, with a positive predictive value of a red flag of 33.5%. Given 1.86 million Scimago-listed biomedical publications in 2023, the estimated annual number of true fakes was 107,800, about 19 times the 5,671 biomedical retractions in 2023.
  • A paper mill interview provided 98 publications, 96 of which were red-flagged by the indicators; paper mills advertised fees of $1,000–$25,000 per publication, and the estimated annual paper mill revenue in biomedicine was about $1 billion in 2023.

Methods

  • The study used a 9-step bibliometric analysis in a sample of 17,120 publications listed in PubMed and Web of Science in neuroscience and medicine, beginning with questionnaires sent to corresponding authors of suspected and control articles.
  • Validation compared 400 known fakes (100 paper-mill retractions, 100 tadpole-like image plagiarisms, 100 fake gene sequences, and 100 retractions from the Journal of Cellular Biochemistry) with 400 presumed non-fakes from 2000 and 2010. The multifactorial tallying rule red-flagged publications scoring >2 points, with points for private email, hospital affiliation, suspicious country, and low institutional output, and points subtracted for co-author number and high institutional output.
  • The validated rule was applied to 1,000 publications from 2020 (10 journals) and 1,000 from 2023 (random keyword search selecting the first PubMed item); Bayes’ rule estimated the true fake rate. A 2D landscape map of 21 million biomedical abstracts using PubMedBERT was also used, along with analysis of >1,000 editing-service websites and a paper mill interview.

Limitations

  • The paper states that journal and topic selection limits the study to biomedicine, and that the sample of 17,120 screened publications is several magnitudes smaller than the 1.86 million biomedical publications in 2023 alone.
  • The paper states that the tallying method has an 11.5% false-alarm rate and 6% false negatives but does not allow an unambiguous judgment for a given paper; the probability of a true fake given a red flag is 33.5%, and the red-flagging method does not directly translate to other science disciplines.
  • The paper states that the control publications were presumed non-fake and one cannot be certain each control was indeed not fake, though the probability was deemed low for 2000 and 2010; it also notes that the true fake estimate depends on the specific indicators, rules, and databases surveyed.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s00210-025-04275-9

Cites

52 of 52 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Screening plan for journals and publications. Studies 1–6 identified potential indicators of fake publications on the basis of hypotheses 1 to 3 (see text), using the indicators “private email” and “no international author”, Studies 7 used the indicator combination “email” and “hospital” as an alternative classification rule. To estimate the incidence of potential fake publications in Studies 8 and 9, we used a more complex multifactorial tallying rule. Note that some publications were used in several screening studies. Total number of unique publications analyzed: n = 17,120
Fig. 1 Screening plan for journals and publications. Studies 1–6 identified potential indicators of fake publications on the basis of hypotheses 1 to 3 (see text), using the indicators “private email” and “no international author”, Studies 7 used the indicator combination “email” and “hospital” as an alternative classification rule. To estimate the incidence of potential fake publications in Studies 8 and 9, we used a more complex multifactorial tallying rule. Note that some publications were used in several screening studies. Total number of unique publications analyzed: n = 17,120CC BY · PubMed Central
Fig. 2 2D landscape map of biomedical topics of 21 million biomedical publications. A Color patches represent different disciplines (topics) in basic science to the left of the white line (e.g., immunology, physiology, and chemistry) and in clinical science to the right (e.g., pharmacology, neurology, and radiology) (for details, see González-Márquez et al. 2023). The position of actual paper mill publications obtained by one of us (B.S.) (98 red triangles in A) shows that the large majority are from clinical and not basic science disciplines. The paper mill publications in the 12 o’clock cluster are from cancer research and the cluster in the center from microbiology/virology. B Distribution of publications with hospital affiliation only (black) or hospital affiliation and private email address (red). The distribution shows that private email addresses are widespread across fields, forming clusters in some topics but not in others. The private/institutional email address ratio is 894,124/5,800,617 = 15.4%
Fig. 2 2D landscape map of biomedical topics of 21 million biomedical publications. A Color patches represent different disciplines (topics) in basic science to the left of the white line (e.g., immunology, physiology, and chemistry) and in clinical science to the right (e.g., pharmacology, neurology, and radiology) (for details, see González-Márquez et al. 2023). The position of actual paper mill publications obtained by one of us (B.S.) (98 red triangles in A) shows that the large majority are from clinical and not basic science disciplines. The paper mill publications in the 12 o’clock cluster are from cancer research and the cluster in the center from microbiology/virology. B Distribution of publications with hospital affiliation only (black) or hospital affiliation and private email address (red). The distribution shows that private email addresses are widespread across fields, forming clusters in some topics but not in others. The private/institutional email address ratio is 894,124/5,800,617 = 15.4%CC BY · PubMed Central
Fig. 3 The rise of red-flagged fake publications (RFPs). Number of publications and percentage of red-flagged fake publications in medicine (top) and neuroscience (bottom) (Study 3). The red line (y1-axis) shows the percentage of red-flagged publications per year; the grey bars (y2-axis) show the estimates of non-fakes, and the red bars show those of potentially fake publications, extrapolated to overall articles published in journals of the Neuroscience Peer Review Consortium and in medicine (selected from scimagojr.com). Note the relatively slow increase in potential fakes in medicine and the rapid rise in neuroscience
Fig. 3 The rise of red-flagged fake publications (RFPs). Number of publications and percentage of red-flagged fake publications in medicine (top) and neuroscience (bottom) (Study 3). The red line (y1-axis) shows the percentage of red-flagged publications per year; the grey bars (y2-axis) show the estimates of non-fakes, and the red bars show those of potentially fake publications, extrapolated to overall articles published in journals of the Neuroscience Peer Review Consortium and in medicine (selected from scimagojr.com). Note the relatively slow increase in potential fakes in medicine and the rapid rise in neuroscienceCC BY · PubMed Central

Verification

Claims the paper makes

11 Valid · 3 Not valid · 13 Not checked · 70 Own results and interpretations

25 more claims

Create a free account to see all 97 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 55
  • Probably not a substantial claim. 25

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

2 of 2

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

3 of 3

Citations

Cites

3 retracted

Show all 47 references

Research works printed but not linked: 1

  1. Hu Z-w, Wu Y (2013) An empirical analysis on number and monetary value of ghostwritten papers in China. Curr Sci 105:1230–1234

Cited by

See all 51

Metadata

Authors
Sabel, B. A., Knaack, E., Gigerenzer, G., & Bilc, M.-I.
Affiliations
  • Center for Behavioral Brain Sciences (DE)
  • Otto-von-Guericke-Universität Magdeburg (DE)
  • Max Planck Institute for Human Development (DE)
Corresponding author
Sabel, Bernhard A.
Venue
Naunyn-Schmiedeberg's Archives of Pharmacology
Published
2025-09-24
Volume
399
Issue
2
Pages
2943-2955
ISSN
  • 0028-1298 (ISSN-L)
  • 1432-1912
OpenAlex
W4414490687
Cited by
51
References
47
Citations by year
2026: 12 · 2025: 15
FWCI
27.79
Type
journal article
Language
en
Open access
Licence
Locations
  • Naunyn-Schmiedeberg s Archives of Pharmacology (publishedVersion, cc-by)
  • medRxiv (acceptedVersion, cc-by-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • MPG.PuRe (Max Planck Society) (submittedVersion, other-oa)
Funders
  • State of Sachsen-Anhalt Ministry of Science, Energy, Climate Protection and Environment (FKZ I 167)
  • Otto-von-Guericke-Universität Magdeburg
Topics
  • Social Sciences › Education › Academic integrity and plagiarism
  • Social Sciences › Library and Information Sciences › scientometrics and bibliometrics research
  • Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
Keywords
paper mills, research integrity, bibliometric analysis, Bayes' theorem, publication fraud, biomedical literature, retraction, AI-generated content
MeSH terms
Humans, Periodicals as Topic, Publishing, Surveys and Questionnaires, Bibliometrics, Biomedical Research

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Fake publications in biomedical science: red-flagging method indicates mass production. https://app.cochat.ai/papers/10.1007/s00210-025-04275-9
The paper
Sabel, B. A., Knaack, E., Gigerenzer, G., & Bilc, M.-I. (2026). Fake publications in biomedical science: red-flagging method indicates mass production. Naunyn-Schmiedeberg's Archives of Pharmacology. https://doi.org/10.1007/s00210-025-04275-9

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.