Advertisement

Protecting data integrity in reproductive medicine

Published:December 08, 2021DOI:https://doi.org/10.1016/j.rbmo.2021.12.005
      Historically, medical research aiming to improve the health of mankind has been based on the principle of trust and the open global exchange of ideas, researchers, and data. Intense global interaction during the early days of IVF represents a fascinating recent example of how collaboration can stimulate the rapid advance and clinical introduction of a newly developed technology. Unfortunately, the countercurrent contribution published in this issue of RBMO (
      • Li W.
      • Gurrin L.
      • Mol B.W.
      Violation of research integrity principles occur more often than we think.
      ) provides us with the very disturbing message that trust may no longer be good enough in publishing newly generated knowledge in scientific journals.
      Reported research observations should be trustworthy, methodologically robust and reproducible. We are becoming increasingly aware, however, that this may regularly not be the case. Multiple factors could underlie the currently much-debated phenomenon of ‘irreproducible science’, such as inconsistencies in populations studied or patient inclusion criteria used, modifications to tools or regimens investigated, along with selective practices such as P-hacking (inappropriate manipulation of data analysis to enable a favoured result to be presented as statistically significant) or HARKing (proposing a Hypothesis After the Results are Known). It is, however, becoming increasingly evident that the incorrect analysis or interpretation of data, or even the use of falsified or fabricated data, should now be added to the many factors contributing to ‘science waste’ (
      • Smith R.
      Time to assume that health research is fraudulent until proven otherwise?.
      ).
      It is suggested that in the current scientific literature up to 20% of reported clinical trials are false. PubMed – under the heading ‘retraction of publication’ – reports over 11,000 retractions (representing less than 1% of the literature) of which over 250 are in the research fields obstetrics, gynecology and reproduction. A very informative overview of all retracted publications can also be found on the ‘Retraction Watch’ website (https://retractionwatch.com/). Without a doubt, these numbers represent only the tip of the iceberg of questionable science. It is noteworthy that when papers are retracted the reason is rarely given. Moreover, it is not known how often journals elect not to withdraw questionable publications. A more detailed analysis of retracted publications revealed plagiarism, duplication or errors in methods, and concerns regarding misconduct as the most common reasons for retraction. Articles published in the field of Obstetrics and Gynecology represent no exception (
      • Chambers L.M.
      • Michener C.M.
      • Falcone T.
      Plagiarism and data falsification are the most common reasons for retracted publications in obstetrics and gynaecology.
      ;
      • Bennett C.
      • Chambers L.M.
      • Al-Hafez L.
      • Michener C.M.
      • Falcone T.
      • Yao M.
      • Berghella V.
      Retracted articles in the obstetrics literature: lessons from the past to change the future.
      ;
      • Bordewijk E.M.
      • Wang R.
      • Askie L.M.
      • Gurrin L.C.
      • Thornton J.G.
      • van Wely M.
      • Li W.
      • Mol B.W.
      Data integrity of 35 randomised controlled trials in women' health.
      ).
      Scientific journals are now faced with the challenge of balancing the basic principle of trust with a more proactive approach assessing the robustness and integrity of the submitted work at an early stage in the review process (
      • Li W.
      • Bordewijk E.M.
      • Mol B.W.
      Assessing Research Misconduct in Randomized Controlled Trials.
      ). There is a need to work jointly on a transparent system by which a scientific contribution can be rejected during the review process, or withdrawn following publication, in an objective and timely manner, preventing contamination of the scientific literature with studies of questionable quality which may harm the interest of patients. The increasing adoption of preprints in medical research, to share studies prior to peer review, increases the difficulty of preventing findings based on scientific misconduct or falsified data from entering the literature, and subsequently dealing with the repercussions.
      For sure, scientific journals should also screen submitted work in a more rigorous manner during the review process and should develop more proactive guidance to authors concerning proper study conduct and handling. Checklists assessing the integrity of trials are proposed, including parameters such as the plausibility of patient inclusion and characteristics, study interventions and study outcomes (Li et al., 2011). The REAPPRIASED checklist for evaluating article integrity (
      • Grey A.
      • Bolland M.J.
      • Avenell A.
      • Klein A.A.
      • Gunsalus C.K.
      Check for publication integrity before misconduct.
      ) provides a framework of questions for journals and authors to answer about individual cases, to better inform editors of the information needed to more effectively handle ethical issues raised over papers before and after publication. The increasing need for more extensive scrutiny of submitted work puts added burden on the editorial process of scientific journals. Historically, taking part in the peer review process has been considered as noblesse oblige. If you as researcher benefit from the peer review process by publishing in the international literature, you feel obliged to contribute to that system yourself. However, there are limits to the added burden that can be put on the shoulders of unpaid referees to assess more intensely the veracity of the data and the methodologies applied in every single manuscript being reviewed.
      The most difficult part remains proving beyond reasonable doubt that the work at hand is indeed incorrect or outright fraudulent. This is likely to account – at least in part – for the time-consuming and frustrating nature of the withdrawal procedures for all parties concerned (
      • Grey A.
      • Bolland M.J.
      • Avenell A.
      • Klein A.A.
      • Gunsalus C.K.
      Check for publication integrity before misconduct.
      ). Studies into retraction timescales have shown that it may take anything from 2 to 11 years (
      • Dal-Ré R.
      • Ayuso C.
      Reasons for and time to retraction of genetics articles published between 1970 and 2018.
      ;
      • Li W.
      • Bordewijk E.M.
      • Mol B.W.
      Assessing Research Misconduct in Randomized Controlled Trials.
      ) between publication and retraction, a figure that may continue to grow as older papers are re-examined. The Committee on Publication Ethics (COPE) provides editors, journals, publishers, and institutions with a series of guidance documents, flowcharts, case studies and core practices (https://publicationethics.org/core-practices), to support the resolution of a wide range of ethical issues and suspected misconduct in published articles, and those under review. However, COPE guidelines deal predominantly with procedures regarding communication between journals, authors and institutions involved in a case. Relevant additional information concerning research integrity and its many aspects can be found on the following websites; https://wcrif.org/, https://embassy.science/wiki/Main_Page, and http://www.icmje.org/.
      The only way forward to deal effectively with suspicion of misconduct seems to be to join forces. Science fraud is not unique to reproductive medicine, and comparable tools may be developed to deal with suspected misconduct in all areas of medical research. Establishing an independent committee with the formal task of assessing the validity of the work in question, along with the criticism put forward concerning the work, would represent an important step forward. Such an authoritative committee should incorporate multiple disciplines, especially ethics, research methodology, statistics and data interpretation.
      As important and urgent as it is to address unethical and fraudulent research, we are also conscious of the effect on all parties concerned when action is taken. Retraction decisions have potentially profound implications, including the future academic careers of authors of retracted papers, the reputation of institutions and for editors who may open themselves up to legal action by the authors (or their institutions) arguing that the retraction was not valid (see also https://publicationethics.org/node/19896). We have to acknowledge, however, that increasing distrust is a common phenomenon in current societies, and it may not be realistic to assume that science will remain immune from such developments. The ever-increasing pressure imposed by institutions on young scientists to publish high impact research in prestigious scientific journals – preferably coinciding with vast media coverage highlighting these novel and important findings – is, for sure, also not helpful.
      At the end of the day, everyone in the scientific community has to take responsibility for the proper conduct and reporting of science. Such responsible behaviour starts with all investigators taking full accountability for the proper design and conduct of the trial, data generation, storage, analysis and interpretation. Co-authors should also take full responsibility for the content of the publication, which means that they should ascertain that the generation of data, its handling and interpretation have been performed in a correct manner. Hence, data generated should at least be shared with co-authors and new systems are currently being developed guaranteeing that everyone has access to the data underlying any given publication. Some journals have already made the public sharing of data a prerequisite of publication. In addition, research institutions should take responsibility for the proper governance of research taking place in their institutions, since there is a perception that the research integrity climate varies greatly between academic ranks and disciplines (
      • Haven T.L.
      • Tijdink J.K.
      • Martinson B.C.
      • Bouter L.M.
      Perceptions of research integrity climate differ between academic ranks and disciplinary fields: Results from a survey among academic researchers in Amsterdam.
      ).
      In conclusion, the urgent need to better deal with scientific wrongdoing remains; it is destructive and bad for the advancement of healthcare. It is time to join forces and create together an effective, transparent research and publications system, where – if needed – misconduct can be detected or corrected as soon as possible.

      References

        • Bennett C.
        • Chambers L.M.
        • Al-Hafez L.
        • Michener C.M.
        • Falcone T.
        • Yao M.
        • Berghella V.
        Retracted articles in the obstetrics literature: lessons from the past to change the future.
        Am. J. Obstet. Gynecol. 2020; 2 (Nov)100201
        • Bordewijk E.M.
        • Wang R.
        • Askie L.M.
        • Gurrin L.C.
        • Thornton J.G.
        • van Wely M.
        • Li W.
        • Mol B.W.
        Data integrity of 35 randomised controlled trials in women' health.
        Eur. J. Obstet. Gynecol. Reprod. Biol. 2020; 249: 72-83
        • Chambers L.M.
        • Michener C.M.
        • Falcone T.
        Plagiarism and data falsification are the most common reasons for retracted publications in obstetrics and gynaecology.
        BJOG. 2019; 126: 1134-1140
        • Dal-Ré R.
        • Ayuso C.
        Reasons for and time to retraction of genetics articles published between 1970 and 2018.
        Journal of Medical Genetics. 2019; 56: 734-740
        • Grey A.
        • Bolland M.J.
        • Avenell A.
        • Klein A.A.
        • Gunsalus C.K.
        Check for publication integrity before misconduct.
        Nature. 2020; 577: 167-169
        • Haven T.L.
        • Tijdink J.K.
        • Martinson B.C.
        • Bouter L.M.
        Perceptions of research integrity climate differ between academic ranks and disciplinary fields: Results from a survey among academic researchers in Amsterdam.
        PLoS One. 2019; 18 (Jan): 1-16
        • Li W.
        • Bordewijk E.M.
        • Mol B.W.
        Assessing Research Misconduct in Randomized Controlled Trials.
        Obstet. Gynecol. 2021; 138: 338-347
        • Li W.
        • Gurrin L.
        • Mol B.W.
        Violation of research integrity principles occur more often than we think.
        Reprod. BioMed. Online. 2022; (in this issue)
        • Smith R.
        Time to assume that health research is fraudulent until proven otherwise?.
        The BMJ Opinion. July 5 2021;