Action to protect the independence and integrity of global health research

Introduction

In a recent Viewpoint in the Lancet, some of us shared our experience of censorship in donor-funded evaluation research and warned about a potential trend in which donors and their implementing partners use ethical and methodological arguments to undermine research.¹

Reactions to the Viewpoint—and lively debate at the 2018 Global Symposium on Health Systems Research—suggest that similar experiences are common in implementation and policy research commissioned by international donors to study and evaluate large-scale, donor-funded health interventions and programmes, which are primarily implemented in low resource settings. ‘We all have the same stories’, was one of the first comments on the Viewpoint, followed by many private messages divulging instances of personal and institutional pressure, intimidation and censorship following attempts to disseminate unwanted findings. Such pressure comes from major donors and from international non-governmental organisations (NGOs) obliged to have an external assessment but who then maintain a high degree of confidentiality and control.

That such experiences are widespread reflects the deeply political nature of the field of ‘global health’ and the interconnections between priority setting, policy making and project implementation, which sit within a broader set of deeply entrenched power structures.^{2 3} Researchers in this field routinely find themselves working within—and studying—complex power relations and so experience challenges in negotiating their own position between interests of commissioning agencies and funders, implementers and country governments, as well as those of their own research institutions and their partnerships with other researchers spanning high-income, middle-income and low-income countries.^4–7 They often receive research funding from major donor agencies like the UK Department of International Development (DFID), the US Agency for International Development (USAID), the Agence Française de Développement (AFD), UNITAID and the Bill and Melinda Gates Foundation,⁸ who commission evaluations for their own funded projects, even though they have a stake in results that demonstrate the success of a multibillion-dollar investment.

Effects of interference in the research and evaluation process are compounded by more subtle acts of self-censorship and data embellishment that can arise as researchers become embroiled in what was recently called the global health ‘success cartel’.⁹ Their involvement in a collective drive to demonstrate success can unintentionally ‘instil a fear of failure, stifle risk-taking and innovation, and lead to the fabrication of achievement’.⁹ For example, research that threatens the position of powerful elites—such as research into high-level corruption—is lacking.¹⁰ Meanwhile, selective reporting of ‘unwelcome’ findings can be a way to avoid contractual terminations even though it undermines learning.^{1 11 12} Moreover, perverse incentives exist across the global health and development sectors to use simplistic indicators of success and bad or fudged data.^13–15 Donor agencies exacerbate the problem by distorting research findings to exaggerate their own successes.^16–19

Researchers are responsible for conducting research ethically and with integrity. Yet, without strong and reliable institutional support, they are often in a vulnerable position when faced with vested interests. What action is needed to avoid undermining independent and critical research findings? What kind of institutional structures and practices might support researchers in dealing with the ethical and political dilemmas associated with the dissemination of (potentially) contested research findings and evaluation results?

To start a discussion on ways forward, we invited input from an international network of global health, health systems and policy researchers from diverse disciplines. Below, we discuss suggestions, endorsed by more than 200 researchers based in 40 different countries (see the full list of signatories below), on how the organisations that commission, undertake and publish research and evaluations can safeguard independence and integrity.

Commissioning bodies

In the first instance, those commissioning external research must enable conditions for independence. Commissioning agencies should be transparent about the purpose and principles of external evaluation and research to their implementing partners and should commit to upholding the principles of good research: ethical, methodologically sound and responsive to population needs. They should specify in the grant contract to researchers that they can review and provide input but will not interfere in the design, data collection, analysis or dissemination of any findings and that they fully commit to making all findings publicly available, whatever their content, including through academic (peer-reviewed) publication. Contractual clauses that limit the dissemination of potentially critical findings—such as DFID’s new standard terms and conditions for service contracts (including evaluations), which prevent researchers from embarrassing DFID or bringing it into disrepute²⁰—should be deleted, since these terms jeopardise the independence of evaluation and research.

For each study, an independent research oversight committee should be established. The committee should include a broad range of stakeholders to avoid institutional bias and linkages with key funders, as well as fairly selected representatives from the communities that are being studied or civil society organisations who can assess the potential benefits and risks generated by the research. A key mandate of oversight committees would be to identify potential conflicts of interests and develop guidelines on rules of engagement between the commissioners and researchers. Such committees should be in a position to intervene or arbitrate if conflict arises, such as if the commissioner or implementing partners pressure, harass or threaten researchers, or if implementing partners feel that the researchers have misrepresented, traduced or misunderstood their work.

To prevent undue influence, donor agencies who commission research and evaluations should develop strong accountability measures between their operational departments and their research and evaluation departments. For example, it is well known in clinical medicine that pharmaceutical industry-funded trials are more likely to produce positive, flattering results than are independently funded trials.^21–25 It is time to debate this important issue in global health too and to ask the question as to whether donor agencies should issue tenders for, commission and oversee evaluation and research involving their own programmes or whether it would be better for an arm’s length body to do so. To increase transparency and reduce selective reporting of findings, we recommend establishing a global health evaluation registry, similar to existing clinical trial registries.^{12 26}

Researchers and research institutions

Today, universities and research organisations across the world depend heavily on external funding from government departments, private foundations and industry.². Therefore, they have an important responsibility to prevent conflicts of interest in research contracts. While better core funding would strengthen research institutions’ power over their own research priorities, they must also seek new ways to protect themselves from interference from external funders. Senior leadership in academia has a responsibility to discuss and develop terms of research with both funders and implementers. They should scrutinise all grants carefully and refuse those that have unfavourable contractual provisions (eg, those that limit researchers from disseminating potentially critical findings).

Senior leadership should also create a supportive, collegial environment for all research staff facing attempts at censorship, including providing legal support when necessary and, ideally, referral to a cross-institutional or national ombudsperson who can serve as a reference point for particular research areas or disciplines. They should extend support to individuals subcontracted to conduct research on behalf of institutions, who may be in especially vulnerable positions.

In addition, senior leadership should encourage methodological and disciplinary diversity to capture complexity and value the dissemination of both positive and negative research findings. Senior research staff being prepared to disseminate controversial and politically contentious analyses can pave the way for more junior researchers to do the same. Research ethics and integrity issues should be part of research training programmes. Research institutions can also provide researchers with access to mentors external to their research group, particularly for junior staff with soft funding. Unions can play an important role if institutional leadership fails.

Ethics and research governance committees

Ethics committees play a crucial role in ensuring the independence and integrity of research. Researchers seek approval from ethics committees, usually both at their research institutions and in the countries in which research is undertaken. Such committees have a remit to safeguard ethical conduct of research and protect the rights and welfare of research subjects, and primarily draw on biomedical research paradigms to do so. Although research ethics committees do often consider the safety risks posed to individual researchers (injury and incarceration), they do not typically consider concerns about protecting the researchers from interference and threats to their credibility. Therefore, they could play a fuller role in helping researchers navigate related unforeseen ethical dilemmas that arise in the course of research. For example, they can provide guidance on whether to extend protections intended for individual research subjects (eg, to ‘do no harm’) to organisations. They could also advice researchers on how to balance their ethical obligations to research participants and their obligations to wider society in the face of pressure from vested interests.²⁷

Ethics committees should have representation from different research fields, with members who are trained in the epistemological and methodological bases of different disciplines.²⁸ In addition, procedures for ensuring the independence of ethics committees are vital to prevent a situation in which members’ close ties to senior management, funders, ruling parties, governments or commercial interests lead them to use regulatory frameworks to shut down ‘unfavourable’ or disruptive research.

University research governance offices, where they exist, can complement ethics committees by protecting the rights and welfare of researchers especially where research challenges powerful agendas. They can offer advice and arbitration assistance to researchers on conflicts of interest arising in relation to external research funding. They can monitor for instances of unethical practice to enable research institutions to push back if powerful external actors manipulate research ethics regulations to constrain the research process, as some British universities have done in the past.²⁹ Research institutions should develop clear value statements and commit to implementing them through their ethics and governance protocols.

Academic journals and editors

The current practice is that academic journals ask or expect authors to declare any conflicts of interest relating to a publication. Journals ought to be challenging of these statements and refrain from publishing what is stated by authors in cases when it is obvious there is a gross conflict. Additional conflict of interest statements should be required from any co-authors that are part of the funding organisation. This can prevent funders from putting pressure on researchers to be included as co-authors of papers emanating from the research and use this role to influence how the results are reported.

Academic journal editors have considerable potential to contribute towards dismantling the ‘success cartel’ within global health, for example, by publishing negative findings, and encouraging papers that explain the ‘hows and whys’ of both positive and negative findings.³⁰ This includes process evaluations and in-depth political and social analyses of global health policy and practice, especially when these challenge the status quo. Editors of academic journals that publish global health research and evaluations should create procedures to select diverse peer reviewers without a vested interest and support them to rigorously question manuscripts that present uncritical and unexplained success stories. Editors should ensure diversity among peer reviewers and moderate dialogue between authors and peer reviewers where, for example, junior authors can challenge unduly hostile or politically motivated reviews by senior academics. They should ideally invite commentaries and responses from donor agencies, NGOs, civil society members, policy makers and researchers from the countries in which research and evaluations have been commissioned.

Conclusion

The tensions between research ethics and the wider politics of the global health field are increasingly recognised. However, the repercussions of these tensions for individuals and research institutions need careful consideration. While ‘rocking the boat’ is uncomfortable and may threaten individual career progression and research institutions’ external income, biased evidence can harm health programme beneficiaries and public trust in research. There are certainly no simple, fail-safe, technocratic quick fixes to resolving issues of power and politics, but the ideas proposed here should at least create better relationships between the institutions involved in commissioning, undertaking and publishing research, and feed into more sophisticated and thoughtful mechanisms of accountability, which do not simply re-enforce existing frameworks that favour accountability towards donors. The ideas we propose should be considered within broader discussions on how to address north–south power imbalances within the research community, and will hopefully catalyse wider action on protecting the independence of public universities and other research institutions globally. We believe this is necessary to enable researchers to hold power to account and advance informed and healthy debate on issues of public interest.