Causes for Retraction in the Biomedical Literature: A Systematic Review of Studies of Retraction Notices

The definition of research misconduct should be stated in the abstract when reporting research on research misconduct

Research integrity is the cornerstone for a reliable and trustworthy science. Research misconduct is classically defined as fabrication, falsification, or plagiarism. To be considered as such, the action must have been committed with the intent to mislead or deceive. There are many other research misbehaviors such as duplication, fake-peer review or lack of disclosure of conflicts of interest, that are often included in the definition of research misconduct in codes, policies, and professional documents. The definition of research misconduct varies among countries and institutions, the seriousness and intentionality of the action. This variability is also present in research articles on the prevalence of research misconduct because it is common for each author to use a different definition, creating confusion for readers. We argue that the definition of research misconduct used in a study should be stated already in the abstract, particularly because not all publications are in open access, so that readers can fully understand what the study found concerning research misconduct without needing to have access to the full article.

The politicization of retraction

The retraction of flawed scientific journal articles is one of the most important means by which science "self-corrects." The prevailing consensus is that retraction is appropriate only when the reported findings are unreliable due to research misconduct or honest errors, ethical violations have occurred, or there are legal concerns about the article. Recently, however, retractions seem to be occurring for political reasons. This trend is exemplified by recent editorial guidance from Nature and Human Behavior which advises the retraction of works that risk significant harm to members of certain social groups. This commentary argues that while "political" retractions may be appropriate in rare cases, retraction is typically not the best means to address potentially harmful research. The politicization of retraction risks harm to science in general as it may further undermine diminishing public trust in science and may encourage scientists to self-censor their work, leading to the under-exploration of some important scientific issues.

A systematic analysis of temporal trends, characteristics, and citations of retracted stem cell publications

BACKGROUND

The increasing prevalence of retracted publications in stem cell research presents significant challenges to scientific integrity. Although retraction notices are issued, retracted studies continue to be cited, facilitating the dissemination of unreliable findings. This study aimed to systematically explore the characteristics of retracted stem cell publications and evaluate the impact of retractions on subsequent citations.

METHODS

This study was conducted following the PRISMA guidelines. A comprehensive search of Web of Science, Retraction Watch Database, and PubMed was conducted from their inception through July 25, 2024, to identify retracted stem cell publications. Characteristics including publication details, retraction reasons, and citation counts were extracted. To assess the impact of retraction on subsequent citations, we compared citation patterns between a random sample of retracted papers and matched non-retracted controls from identical journals and issues. Further analysis was conducted to determine whether papers citing retracted articles had an elevated risk of subsequent retraction. Descriptive statistics, chi-squared tests, t-tests, and Mann-Kendall tests were used for data analysis.

RESULTS

The systematic search identified 1421 records, with 517 publications meeting inclusion criteria. Temporal analysis revealed two significant trends: an increasing retraction rate that peaked at 0.84% in 2023 and a declining time-to-retraction (median: 30 months, interquartile range: 13-60; Mann-Kendall, tau = - 0.29; P < 0.001). Hospital-affiliated researchers from China contributed to 244 (47.2%) of retractions. Data and image flaws were identified in 360 (69.6%) of retractions. Among 472 Web of Science-indexed retracted publications, 366 (77.5%) accumulated 4884 post-retraction citations, with 114 (24.2%) receiving more citations post-retraction than pre-retraction. Analysis of a random subset of retracted articles (n = 53) demonstrated that only 14 (4.2%) out of 334 post-retraction citations referenced the retraction notice. Compared with 639 non-retracted control publications, retracted articles showed significantly lower post-retraction citation rates (mean rank: 291.32 vs. 351.08; P = 0.01). Moreover, papers citing retracted articles exhibited an 11-fold higher risk of subsequent retraction (odds ratio (OR): 11.09; 95% confidence interval (CI): 7.06-17.43).

CONCLUSIONS

This analysis reveals substantial research integrity challenges within stem cell research. These findings suggest the necessity for enhanced surveillance mechanisms and standardized protocols to identify and curtail the dissemination of flawed research.

Inferential statistics for cardiothoracic surgeons: Part 3 - drawing valid conclusions from clinical data

Inferential statistics enable researchers to make predictions about a population based on sample data. This involves hypothesis testing where the null hypothesis assumes no effect, and the alternative hypothesis suggests a significant effect. Testing requires assumptions like normality and independence to be validated using tests like Shapiro-Wilk or Levene's for normality and variance. Significant findings are determined by p-values, with values under 0.05 typically indicating non-random effects. Choosing between parametric and non-parametric tests depends on data normality and variance homogeneity. Tools such as t-tests, analysis of variance (ANOVA), and their non-parametric counterparts like Mann-Whitney or Kruskal-Wallis are used based on these criteria, ensuring appropriate conclusions about clinical effects and interventions.

Linking citation and retraction data reveals the demographics of scientific retractions among highly cited authors

Retractions are becoming increasingly common but still account for a small minority of published papers. It would be useful to generate databases where the presence of retractions can be linked to impact metrics of each scientist. We have thus incorporated retraction data in an updated Scopus-based database of highly cited scientists (top 2% in each scientific subfield according to a composite citation indicator). Using data from the Retraction Watch database (RWDB), retraction records were linked to Scopus citation data. Of 55,237 items in RWDB as of August 15, 2024, we excluded non-retractions, retractions clearly not due to any author error, retractions where the paper had been republished, and items not linkable to Scopus records. Eventually, 39,468 eligible retractions were linked to Scopus. Among 217,097 top-cited scientists in career-long impact and 223,152 in single recent year (2023) impact, 7,083 (3.3%) and 8,747 (4.0%), respectively, had at least 1 retraction. Scientists with retracted publications had younger publication age, higher self-citation rates, and larger publication volume than those without any retracted publications. Retractions were more common in the life sciences and rare or nonexistent in several other disciplines. In several developing countries, very high proportions of top-cited scientists had retractions (highest in Senegal (66.7%), Ecuador (28.6%), and Pakistan (27.8%) in career-long citation impact lists). Variability in retraction rates across fields and countries suggests differences in research practices, scrutiny, and ease of retraction. Addition of retraction data enhances the granularity of top-cited scientists' profiles, aiding in responsible research evaluation. However, caution is needed when interpreting retractions, as they do not always signify misconduct; further analysis on a case-by-case basis is essential. The database should hopefully provide a resource for meta-research and deeper insights into scientific practices.

Evaluation of Incurred Subject Period Re-analysis (ISPR) as a Tool to Distinguish Fraudulent Pharmacokinetic Profile Pairs from Non-fraudulent Pairs

Duplicate pharmacokinetic profiles in bioequivalence trials is an issue which has caused hundreds of retracted marketing authorizations. No formal test for profile duplication exists in spite of the existence of profile comparison algorithms, so defining a threshold that distinguishes a naturally occurring pair from a duplication remains difficult. An idea called ISPR (incurred subject period analysis) was aired in 2023 and is evaluated in this paper along with three new profile comparison methods. ISPR involves analysis of entire PK-profiles within a study. It is shown that when ISPR is combined with appropriate PK-profile comparison methods, the duplicate pairs display a lower score (better similarity) than pair that do not arise out of duplication. Therefore, ISPR may help establish a threshold that distinguishes fraudulent profile pairs from non-fraudulent profile pairs. ISPR therefore may be used as QA tool, serves as a method by which a CRO can -to some extent- show that their studies do not contain duplicates in the primary analysis, and thus also may be a means by which sponsor can argue that their studies are trustworthy, in case the suspicion about duplication arises. This paper does not introduce a formal test for this type of fraud; rather the authors see it as a first moderate step in that direction. Hopefully, if or when ISPR data is submitted to authorities as part of general dossier submission, data will accumulate to the extent that they may be able to develop models that allow formal testing for profile duplication.

Opening the black box of article retractions: exploring the causes and consequences of data management errors

The retraction of an article is probably the most severe outcome of a scientific project. While great emphasis has been placed on articles retracted due to scientific misconduct, studies show many retractions are due to honest errors. Unfortunately, in most cases, retraction notices do not provide sufficient information to determine the specific types and causes of these errors. In our study, we explored the research data management (RDM) errors that led to retractions from the authors' perspectives. We collected responses from 97 researchers from a broad range of disciplines using a survey design. Our exploratory results suggest that just about any type of RDM error can lead to the retraction of a paper, and these errors can occur at any stage of the data management workflow. The most frequently occurring cause of an error was inattention. The retraction was an extremely stressful experience for the majority of our sample, and most surveyed researchers introduced changes to their data management workflow as a result. Based on our findings, we propose that researchers revise their data management workflows as a whole instead of focusing on certain aspects of the process, with particular emphasis on tasks vulnerable to human fallibility.

Scientific misconduct in infectious diseases-European Society of Clinical Microbiology and Infectious Diseases survey

OBJECTIVES

We aimed to evaluate the prevalence and perception of scientific misconduct in infectious diseases (ID) and clinical microbiology (CM), as reported by the ID/CM community.

METHODS

An anonymous online European Society of Clinical Microbiology and Infectious Diseases survey circulated among society members from October 2023 to June 2024; the questionnaire included data on participants' views on their own and their colleagues' scientific misconduct in the last 5 years.

RESULTS

The survey received 220 responses. Responders were 73% ID physicians, 52% men, 56% aged 35-54 years, and represented 48 countries, mainly European (126 participants). The vast majority of participants (78%) reported that they did not personally commit scientific misconduct, whereas 54% reported witnessing misconduct by colleagues in their field. The most commonly committed misconduct by both responders and their colleagues was misconduct of authorship rules, 14% and 41%, respectively. Overall, 18% reported witnessing misleading reporting and 14% reported witnessing nonaccurate reporting of conflict of interest. Nevertheless, the majority (>60%) of responders reported high confidence in the integrity of published work in the field of ID/CM. Approximately one-third of responders were not aware of the European Society of Clinical Microbiology and Infectious Diseases ethics advisory committee as an authority to which members can report misconduct.

DISCUSSION

Scientific misconduct, mostly related to violation of authorship rules, seems to be common in ID/CM. Efforts to improve scientific integrity should be made to keep trust in the scientific process.

Retracted articles in scientific literature: A bibliometric analysis from 2003 to 2022 using the Web of Science

Retractions serve a crucial role in maintaining the integrity and accuracy of scientific literature. There has been growing interest in understanding the patterns behind retractions. This bibliometric study analyzed retracted articles published between 2003 and 2022, indexed by the Science Citation Index Expanded of the Web of Science Core Collection database. A total of 8466 retracted articles were identified, revealing an overall increase up to 2019, followed by a decline. A total of 109 countries contributed to the retracted articles, with China and the United States having the highest absolute numbers. In addition, the articles were published in 2347 different journals, with Tumor Biology recording the largest number of retracted articles. The top 10 most cited retracted articles indicated that data and image integrity issues were the main reasons for retraction. The primary reasons for retractions, identified by linking the retracted articles to the Retraction Watch Database, were data and results issues followed by plagiarism and duplication. In conclusion, the present bibliometric study offered an overview of the status of retracted articles indexed by the Web of Science Core Collection over the past two decades. These findings provide insight into areas where scientific integrity may be compromised and serve as a guide to foster a responsible research environment.

Publication Ethics in the Era of Artificial Intelligence

The application of new technologies, such as artificial intelligence (AI), to science affects the way and methodology in which research is conducted. While the responsible use of AI brings many innovations and benefits to science and humanity, its unethical use poses a serious threat to scientific integrity and literature. Even in the absence of malicious use, the Chatbot output itself, as a software application based on AI, carries the risk of containing biases, distortions, irrelevancies, misrepresentations and plagiarism. Therefore, the use of complex AI algorithms raises concerns about bias, transparency and accountability, requiring the development of new ethical rules to protect scientific integrity. Unfortunately, the development and writing of ethical codes cannot keep up with the pace of development and implementation of technology. The main purpose of this narrative review is to inform readers, authors, reviewers and editors about new approaches to publication ethics in the era of AI. It specifically focuses on tips on how to disclose the use of AI in your manuscript, how to avoid publishing entirely AI-generated text, and current standards for retraction.

Taking it back: A pilot study of a rubric measuring retraction notice quality

The frequency of scientific retractions has grown substantially in recent years. However, thus far there is no standardized retraction notice format to which journals and their publishers adhere voluntarily, let alone compulsorily. We developed a rubric specifying seven criteria in order to judge whether retraction notices are easily and freely accessible, informative, and transparent. We mined the Retraction Watch database and evaluated a total of 768 retraction notices from two publishers (Springer and Wiley) over three years (2010, 2015, and 2020). Per our rubric, both publishers tended to score higher on measures of openness/availability, accessibility, and clarity as to why a paper was retracted than they did in: acknowledging institutional investigations; confirming whether there was consensus among authors; and specifying which parts of any given paper warranted retraction. Springer retraction notices appeared to improve over time with respect to the rubric's seven criteria. We observed some discrepancies among raters, indicating the difficulty in developing a robust objective rubric for evaluating retraction notices.

Quantitative research assessment: using metrics against gamed metrics

Quantitative bibliometric indicators are widely used and widely misused for research assessments. Some metrics have acquired major importance in shaping and rewarding the careers of millions of scientists. Given their perceived prestige, they may be widely gamed in the current "publish or perish" or "get cited or perish" environment. This review examines several gaming practices, including authorship-based, citation-based, editorial-based, and journal-based gaming as well as gaming with outright fabrication. Different patterns are discussed, including massive authorship of papers without meriting credit (gift authorship), team work with over-attribution of authorship to too many people (salami slicing of credit), massive self-citations, citation farms, H-index gaming, journalistic (editorial) nepotism, journal impact factor gaming, paper mills and spurious content papers, and spurious massive publications for studies with demanding designs. For all of those gaming practices, quantitative metrics and analyses may be able to help in their detection and in placing them into perspective. A portfolio of quantitative metrics may also include indicators of best research practices (e.g., data sharing, code sharing, protocol registration, and replications) and poor research practices (e.g., signs of image manipulation). Rigorous, reproducible, transparent quantitative metrics that also inform about gaming may strengthen the legacy and practices of quantitative appraisals of scientific work.