Why research integrity matters

Clinical practice and health policies require high-quality scientific evidence. Such evidence is produced when research is adequately funded and conforms to guidance that sets out ‘standards’ for research integrity and publication ethics.^1,2 Psychiatric research and practice have historically been undermined by a lack of the investment seen in other areas, and may be affected by competing interests of individuals or groups. Evidence is often contested; we are dealing with complex and multiple interacting causal influences and complex interventions. For example, environmental, social, cultural, psychological and biological factors are relevant, as are specific forms of adversity such as poverty, racism and trauma. There may be much disagreement or challenge. These complexities make the task of journal editors and editorial boards challenging and even more important for advancing science. We need to respond by promoting high-quality scientific evidence to inform mental health practice, policy and legislation. This includes following ethical standards, which mandate that editorial decisions are made independent of external vested interests.

What is editorial independence?

Editorial independence can be defined as the freedom of editors to make decisions about the scientific publication record. To promote good science, editorial independence must be a non-negotiable cornerstone of all scientific journals seeking to prevent influence from journal owners or from interest groups.^{Reference Davies and Rennie3} However, editorial independence is a multidimensional and complex concept, as laid out in guidelines from the World Association of Medical Editors (WAME), the Committee on Publication Ethics (COPE) and the International Committee of Medical Journal Editors (ICMJE) (see Table 1 for a summary). This places full responsibility for advancing scientific debate on the Editor in Chief, working with editorial boards, authors and reviewers, in order to uphold standards. Editors’ responsibilities are to the public and patients (for health journals), as well as to professional, scientific and commissioning organisations.

Table 1

Editorial independence: guidance and protections

The guidelines make clear that journal owners should provide the conditions for editorial independence to be realised and protected. Owners may hire and fire editors and, from a distance, be satisfied that appropriate editorial policies are in place. They should not influence editorial decisions, nor prioritise their financial concerns over good ethics and science. These standards are the oxygen of editorial practice, and editors work hard to uphold them, working with reviewers, boards, authors and public commentators under an implicit contract of trust in editorial equipoise, judiciousness and authority. The editorial role is to find and present the best research for peer review, scientific critique and public scrutiny, and not take specific political or policy positions.

Correct or retract?

Too much published research is unsound: 25% of randomised trials and about 10% of large-scale randomised trials suffer from major flaws.^{Reference Ioannidis4} Research cultures are part of the problem. For example, in a Nuffield Council on Bioethics study, 26% of researchers were tempted to or felt under pressure to falsify data.^{Reference Joynson and Leyser5} Clarivate delisted 82 journals in March 2023, citing concerns about bypassing peer review and publishing articles that lacked scientific rigour in exchange for publication fees, mostly funded by taxpayers. Corrections and retractions may be perceived to challenge trust in science, but eventually such honest and explicit actions should improve it and must be carefully considered by Editors in Chief, who in accord with guidance have the final say. Corrections and retractions should not be stigmatised but are an accepted part of the contract between authors, editors, readers and the wider group of stakeholders. All work in trust to establish and correct the scientific record.

Famous retractions or expressions of concerns about Hans Eysenck's research, Rosenhan's discredited study of feigning (which remains unretracted), as well as the retraction of a paper on measles, mumps and rubella (MMR) vaccinations, suggest that reliability, validity and dependability are the drivers of retractions. In our experience, flawed research designs, unclear search strategies, selective reporting, outcome switching, p-hacking and failure to replicate are commonly seen. Blatant data fabrication is less common. Many errors are picked up not at the time of publication but much later. Dealing with older papers is especially complex, given the amount of time that has passed since publication. Scientific knowledge is contextual. Dismissing ‘old’ research entirely based on modern standards may overlook the incremental contributions that were the foundation for subsequent advancements. Retracting seminal papers because they fail to meet current academic criteria could entirely negate the importance of their advances in their respective field at the time. However, older papers can and should be legitimately appraised and judged against best practice at the time of publication; if found to be flawed in some fundamental scientific way, removal from the scientific record is appropriate. Older publications undertaken with best research practice of the time should be retained.

Threats to editorial independence and recommendations

COPE, WAME and ICMJE provide some fundamental standards. Despite the plethora of guidance, the reality is not always so rosy. Why do poor studies remain in the scientific record? There are no or few requests for correction or retraction of uninfluential or uncited papers. There is often a pressure to not retract if a paper is contributing to debate, is controversial or is topical, as it raises the profile of a journal. For example, the recent COVID-19 pandemic saw large numbers of submissions being processed rapidly, with many retracted by May 2023. However, the retracted papers are still highly cited.

Public scrutiny, for example by Retraction Watch, and peer review by scientists provide a form of self-regulation. However, membership of COPE is not enough of a safeguard if its members are violating principles, and there are no powers to expel or to enforce compliance. Some larger publishers or owners take charge of complaints, calls for retractions and any legal threats to the publisher or owner, rather than following the guidance on editorial independence from COPE, WAME and ICMJE. This risks non-scientific interests becoming the basis for decisions on how to respond to flawed science.

There are examples of the guidance being ignored. For example, an owner and membership society tried to influence editorial decisions and the wording of the retraction and of subsequent correspondence which criticised the owner's position.^{Reference Gunn6} WAME guidance encourages editors to make public any violations of publication ethics:

This can be challenging if an owner removes editors, does not support the guidance and disrupts the execution of decisions.

A recent but alarming trend is the strategic use of legal threat to control the dissemination of public opinion, journalism and scientific findings. There are examples of companies and individuals threatening litigation on grounds of defamation in response to scientific critique.^{Reference Rennie7} Legal challenges to editorial independence are especially difficult as settlements can be costly. To avoid this risk, owners may choose to not follow editorial recommendations. Thus editorial independence is a dynamic concept, only realised when tested in specific contexts. There are pilots in Sweden where the courts, rather than journal owners or editors, decide what is or is not retracted. This expensive legal proposal cannot replace scientific scrutiny and may weaken scientific judgement as the basis of what is and is not retracted. Such processes are potentially as vulnerable to political and financial pressures. Only those with the most resources are likely to resort to the courts to defend editorial independence and the scientific record. Indeed, the industry is not designed to operate in such an environment. We need better guidance to deter legal threats designed to intimidate and influence editorial decisions; for example, one proposal is that legal threats be posted on journal websites to ensure that such threats are transparent.^{Reference Persaud, Ringer and Lemmens8}

When there are complaints or allegations of error, it is imperative that the editor and author can discuss potential remedies, for example corrections, data re-analysis or the reporting and interpretation of the findings, long before retraction is considered necessary. Even when retraction is necessary, this should ideally be by mutual agreement. Sometimes authors themselves request (or agree to) retraction after discovering errors in data gathering or processing or reporting.^{Reference Kaufman, Malhi and Bhui9} A mutually satisfactory decision to retract may not be possible if authors object, or worse, if they resort to legal threat, thereby blocking any meaningful dialogue about the validity of their research.

The most important quality-control mechanism for research integrity is editorial independence guided by publication ethics to ensure that there is an uncompromising insistence on meeting the highest standards of scientific research, reporting and publication. A number of issues require further discussion. For example, regulation could include a register of breaches and scrutiny to learn lessons, as well as legal advice on issues of public interest. How might we assess journal owners’ compliance, competence and capability? Should insurance for clinical roles be extended to editorial roles? Continuing professional development should include editorial independence. We look forward to debate and discussion on how to protect and promote editorial independence.