It’s common ground that vulnerable populations were convenient and powerless subjects of scientific research, and some amount of harm to research participants is either unacceptable or must be the subject of conditions aimed at their protection.

Recap

1. The GSP acknowledges:

   1. no one wants a return to the days when researchers could experiment on people without their knowledge or consent, or inflict lasting harm on participants in order to obtain research data.

   2. the legacy of exploitation is most notably recognised by the principles set out in the judgment given in the Nuremberg doctors trials, which led the World Medical Assembly to adopt the Helsenki Declaration. In the US, the policy reaction to the atrocities of the Tuskagee study is founded in the adoption of the Belmont Report.

   3. a history of exploitative practices and neglect for the welfare of participants in research conducted by (US) researchers shows a trust deficit, and highlights the structural asymmetries between researcher and subject, researcher and research institution, research institution and funder.

   4. palpable frustration frames criticisms of scope creep and unchecked authority of a bureaucratic apparatus that can delay, distort, or even derail research.

2. GSP provides examples of what it assesses to be widespread petty tyranny,1 including:

   1. researchers being required to warn professors who participated in a study that using a laptop in a public place to complete a study survey might expose them to an intrusion on their privacy.

   2. after two years and 300 pages of forms were submitted in an engagement with an IRB, a study was abandoned, having been unable to identify risk mitigation protocols that address concerns arising in the context of a simple physiological study involving looking at the patient and monitor resting rate, saturation, exhaled CO2 temporal changes during initial treatment.

3. Such is the despondent state of researcher-IRB relations, that Alice Dreger2 asks whether both intellectual freedom and those humans involved in the research process can be protected.3 Alice’s complaints are partly about;

   1. the administrative burden of long forms that shed little light on how they relate to the nature of the proposed research;

   2. the delays that the process can impose;

   3. the legal liability mitigation mandate that dominates the institutions in which research is conducted; and

   4. what appears to Alice to be flexibility around ethics if money is connected to a proposal.

4. Referencing analysis by Laura Stark4 and Penny Mason,5 Alice vents her frustration with over-compliance and regulatory overreach (and the GSP echoes her views) when the rules appear to be more about shoring up a liability shield than protecting participants.

5. The GSP cites studies which demonstrate hinderances for reasons ranging from paternalism, ideological opposition to the subject of the proposed study, or a lack of knowledge in the discipline, including:

   1. a study in which participants role-played IRB members. It asserts scope-creep colouring ethical standards of judgment and asserting that the job of eliminating ethical problems in research might be pressing review boards to find problems where they do not exist.

6. an account (written by researchers who formerly served as IRB members) of:

   1. a study which proposed to study failed suicide bombers, on the basis that (despite claims that the prisoners were willing to consent):

      1. prisoners could not freely give informed consent; and

      2. a requirement that interviews be conducted in the presence of a lawyer who could intervene before prejudicial questions were answered.

   2. a study that restricted a white researcher to interviewing white subjects about career aspirations, on the basis that the topic could upset black participants.

   3. a study on digit ratio, which was rejected because the reviewer perceived (in the absence of substantiating evidence) that the study was being conducted to make approximations about penis size.

   4. a study proposing to observe HIV prevention practices in bathhouses of San Francisco on the basis that any request for consent from the bathhouse operator would compromise the ability to access data (because the operator would introduce new rules once they became aware that practices were being observed).

   5. practical costs of delay in conducting ethics reviews, particularly for graduate students who conduct and submit their research on funding and program timelines.

   6. studies indicating that IRBs prioritise liability management over their core service, which those researchers understand to be the evaluation of research proposals for the purpose of protecting against exploitation.6

7. To the GSP, the experience of dealing with ethics review is thought policing rather than a genuine and shared commitment to protecting participants and promising research. The Canceling of the American Mind uses the term Conformity Gauntlet to describe litmus tests used to thwart promising research where it questions the ideological orthodoxy.

8. The examples used to advance this argument bring up a range of factors, with many substantive criticisms of the process and gravity of the decisions made in ethics review. Some of these references assert that the primary objective should be discovery. Existing research should be accepted without renewed review (no acknowledgement of retractions, qualifications or reproducibility). These views are understandable backlash against perceived overreach, but they are insensitive to the other attributes of the system which call for integrity.

Research has value. People need it. Humans should be supported in safely participating in good science. Poor experiences with IRBs can adversely affect participants and researchers alike. This statement can be true, without being a complete encapsulation of the research ethics issue.

What to do about it

1. The GSP:

   1. acknowledges the Foundation for Individual Rights and Expression’s (FIRE) proposal that:

      1. the gold-standards in the Belmont Report be declared as the boundaries of a scope of any IRB review; and

      2. a reporting system which logs poor experiences with IRBs be established.

   2. applauds the Academic Freedom Alliance (AFA) for shedding light on overreach and commends the Mudd Code as a concrete and principled reform proposal. It proposes 10 (actually 11) principles to be adopted as a counterweight to (what the AFA perceives as) IRB’s adopting a mandate to totally eliminate actual or perceived harm without regard to its effect on proposed science.

2. Our specific comments on the Mudd Code principles are set out below. They’re offered as a contribution to developing principles that serve to address the frustrations ventilated by the GSP and the AFA in its criticisms of their experiences with ethics review protocols. The general observation is that they seek specificity (anchored in the perspective of a niche field of research interest) in order to exclude the room for overreach. That specificity does not neatly reconcile with the unknown elements and vast diversity across fields (much less new discoveries that unlock new fields). A practical starting point is to separate the legitimate frustrations with process from the proposed changes (or clarifications) to purpose. The first deserves reform. The second deserves more careful argument than it has received.

3. The examples linked in the GSP essay present a variety of ethics review concerns. Some are plainly absurd: the laptop-in-a-public-place warning is its own punchline (but also capable of satisfaction in less time than it takes to write up the complaint). Others involve genuine ethics questions where the IRB came down on one side and the GSP (or the essays is cites) frames the disagreement as obvious overreach. For example, the bathhouse study is nuanced. Limited disclosure (and to take the point further, active deception) has a recognised place in research ethics. It is neither categorically forbidden nor permitted. The public-private character of the setting, whether the data could be obtained by other means, the level of risk to those observed (acknowledging that patrons have some expectation of privacy - even in the setting described), the strength of the public benefit (along with any risk to the community’s trust in the research and the researchers), the adequacy of the debrief, the protections around identity are considerations where covert observation is proposed.

4. The complaint, as presented, is not that the IRB's analysis was wrong but that there was an analysis at all. The answer was not obvious (given the facts as recounted). The researchers' view of what data they need and how to obtain it is one input among several. An IRB that engages the question is doing its job. I would be interested to understand whether this complaint would have been made had the ethics review process involved less form filling, a faster pace and a more transparent focus on balancing the interests.

Oceans divide us. I’m Australian and primarily influenced by the National Statement on Ethical Conduct in Human Research. Having come through local and British institutions, I have had opportunities to work and collaborate with researchers located elsewhere, US included. Each institution involved can bring its own interpretation of the core guiding principles, supplement them with institutional constraints or preferences, and layer in the sources of external influence, such as funding requirements or niche regulatory compliance requirements (for example, foreign interest reviews). My experience with cross-border or cross-institutional collaboration is that the highest standard applies for everyone involved. One program did not proceed with US collaborators given tensions arising in connection with publication controls.

Nuance

The scope of the IRB remit

1. When the GSP frames the IRB’s job as ensuring that nobody is harmed and that participants understand the risks they’re accepting, it describes two of the IRB’s responsibilities. There are more.

2. An IRB acts as a steward for multiple parties, each with legitimate interests that generate requirements for how research is conducted. Participants need protection from exploitation, but also fair treatment in the use of their data, equitable access to research that could benefit them, and the right to walk away. Researchers need fair process, timely review, and an institution in which they can do their work. The IRB serves researchers too. The examples raised in the GSP's essay show where that relationship has become adversarial, and the frustration is fair, but adversarial moments do not redefine the function.

3. The institution carries reputational and fiduciary interests, and public trust is part of its capital. Those who participate in research need to trust researchers that they may not know. Funders and regulators set conditions that form part of the framework the IRB should consider. The broader community served by the research should be considered, particularly where the institution is a public-benefit body. The public has an interest in benefit, in equitable inclusion, and in not having research conducted on its members without producing anything that helps them or which unfairly burdens their cohort of the population. Future participants and future researchers depend on the trust capital being built or burned. The scope is wider than the researcher’s important quest for discovery.

4. Each of these interests generates requirements, some explicit (there are laws regulating some issues), some are founded in standards such as the Belmont Report and some judgment calls that an IRB has to navigate. That is why reviews are not checklist exercises and why the Belmont Report frames its principles at a level of generality. The GSP defines the function of the IRB as performed by enforcing the principles set out in the Belmont Report at the institutional level. Neither the GSP authors nor the AFA team has indicated how they have determined that the Belmont Report could or should be understood to be the gold standard statement of (exhaustive) review criteria. The document itself excludes such characterisation:

   Three principles, or general prescriptive judgments, that are relevant to research involving human subjects are identified in this statement. Other principles may also be relevant. These three are comprehensive, however, and are stated at a level of generalization that should assist scientists, subjects, reviewers and interested citizens to understand the ethical issues inherent in research involving human subjects. These principles cannot always be applied so as to resolve beyond dispute particular ethical problems. The objective is to provide an analytical framework that will guide the resolution of ethical problems arising from research involving human subjects.

5. The document guides with generality rather than prescribes with specificity. It establishes values that leave the direct implementation to those working within the context of a research program. That judgment may involve the introduction of tailored supplements or nuanced interpretation. The GSP critique reads the review board as a single-function gatekeeper and then complains that the gate is too tight. We do not start from the common view of the function as singular.

6. Responsibility for the quality, safety and ethical practices involved in research is shared. IRBs should lead the process. It is a steward representing the various interests of a broad set of stakeholders. Unlocking the knowledge frontier is an important goal. The ethical questions around how that is done are not second order functions.

The Mudd Code

The Mudd Code's 11 principles fall into four groups. Most of the transparency proposals address the process complaints raised by the GSP and AFA. Several of the risk-evaluation and risk-reduction principles reflect scope complaints: they ask the IRB to treat as settled questions that remain contested. Some thoughts and suggestions on the proposal:

Sources of friction

Stiff laws across the books often reflect a reaction to toothless sympathetic platitudes proffered in response to abuses. Where good judgment is shown to fail, fixed rails tend to show up in regulations. They’re sticky. Research is not. There is valid criticism for the friction this creates.

In the research context, while the worst abuses are mostly7 memories in the US, historically grounded fears cast long shadows over the present day. There may be different points of view as to its extent. In Medical Apartheid: The Dark History of Medical Experimentation on Black Americans from Colonial Times to Present, Harriet Washington identifies how modern research leaders can dismiss doubts as anti-science in order to deny present research dangers.8 Her work also identifies how exploitative practices have been exported across the developing world.9

To get around consent forms and a skeptical public, many researchers are turning their attention to the African and other developing countries. — Robert F Murray, MD, Chief of the Division of Medical Genetics at Howard University

Harriet’s point is that whilst research is needed, research subjects need trust. A trust deficit lingers in the wake of past abuses. It’s more than a mass torts premise for a John Grisham novel. Trust in our research institutions is under strain. Curiosity and profit are pitted against participant welfare.

The potential for exploitation still exists and past abuses offer lessons in why independent oversight and clear ethical frameworks could serve as research infrastructure to guide the implementation of ethical research to the point where participants could have confidence in its practices.

The frustrations the GSP describes are real. The reforms it endorses for IRB process transparency, timeliness,10 accountability for reviews, insights around ideology could improve trust and deserve support, and the table above sets out where I agree with parts of the sentiment (and propose adjustments to show how I would adjust it). The reforms it endorses for IRB scope: confining review to a narrow remit (protecting people from genuine physical harm and treating the Belmont Report’s general principles as exhaustive boundaries) distorts the frame. We don’t share a common starting point in our respective appreciation of the objectives for an ethics review.11

Friction in research ethics is not the failure. It is the cost of operating in a context where the consequences of getting it wrong have historically been borne by those least able to refuse. Unnecessary friction should be addressed. Ceaseless and heedless growth can be criticised across the broader legal and regulatory systems.12 Trust is the asset the review board exists to protect, and trust is harder to rebuild than to lose. I favour reasonable friction where it serves to strengthen trust.

The GSP commends the AFA for creating the Mudd Code. To the extent that both teams promote efforts to address what isn’t working in the system, we are in agreement. This is an invitation to share your views on:

• how you rationalise the various stakeholders’ interests in research (if you would prioritise discovery and autonomy, how do you set precedent which defines acceptable risk for other stakeholders?);

• if you support the Mudd Code proposal:

  • how would you transform the generalisations of the Belmont Report to the particular promulgations sought by the AFA for each discipline of research?

  • how do you see the review boards role in the context of maintaining discipline specific exhaustive checklists for a vast and expanding range of research interests?