Ethical review should apply to non-academic research, too

When I served on NHS research and development committees a few years ago, we were often tasked with deciding whether a proposal constituted fully fledged research or “merely” an audit.

The latter involves the evaluation of services supplementary to treatments, such as changes to a patient waiting area or feedback on the experience of care. It is deemed not to be research because it might not produce new and effective knowledge. That is despite the fact that, to be useful, an audit should be as robust as standard research in terms of accurate baseline and outcome measures, as well as awareness (if not control) of potentially confounding variables.

The distinction between research and audit matters because audits are not considered by the NHS to require the same ethics appraisal as research. Yet such interventions could raise just as many moral questions as research does.

It largely comes down to administrative convenience. Given the amount of service evaluations conducted in the health context, the ethics review system could be overwhelmed if all research-like activity required its approval. Yet administrative convenience is obviously not a good reason to determine where ethical oversight is appropriate.

More recently, I have been challenged to define research in my capacity as leader of a European Commission project, known as PRO-RES, that aims to promote ethics and integrity in non-medical research – in certain fields of which, the requirement to submit to ethics review is sometimes actively opposed.

My colleague Fabian Zuleeg, CEO and chief economist of the European Policy Centre, points out that not all research agencies are in a position to adhere to PRO-RES’ foundational statement on the values, principles and standards to be sought in research. For instance, it could clash with some thinktanks’ business models by offering an advantage to less reputable organisations that would be unlikely to follow such principles.

But while analysts in the public and private sectors might not carry the label “researchers”, they certainly collect data, as well as synthesise research findings from other sources. They use their judgement and experience to make recommendations based on the research. So why, unlike academic researchers, are they currently unlikely to be subject to any form of ethics review?

Nor is it just public policy researchers and advocacy organisations that evade ethical oversight. Data gathering is also part of the modus vivendi of most large internet-based corporations – and the scale of profit realised by companies such as Google, Amazon and Facebook is testimony to the value, accuracy and effectiveness of such data. Yet there are currently no adequate legal limits on their actions. They allow themselves to be subject only to their own internal mission statements.

How can it be allowed that standards in research apply to some researchers but not to others? This only enhances the unfortunate incentives for skilled researchers to migrate from ethics-appraised academic research in universities to better-paid commercial research organisations, in which innovation might even override integrity and which are highly unlikely to subscribe to those supposed fundamental principles of science: openness, transparency and the sharing of knowledge.

The research that elected Donald Trump and “caused” the Brexit referendum result was not subject to the kinds of checks on ethics and integrity required of academic researchers. Cambridge Analytica was a research mercenary – employed to gather data that could influence opinion on a massive scale and immune to the need for transparency, balance and justice in its exploitation of data.

The mathematician Hannah Fry recently argued for a Hippocratic Oath for scientists – especially mathematicians. She hopes that this would encourage them to pay closer attention to the consequences of hidden algorithms and invisible processes that gather and manipulate personal data with varying degrees of permission and perform functions that have real-life consequences for individuals and their communities.

But we also need to find something equivalent to the conventional models for monitoring ethical research that can be applied to corporations, thinktanks, advocacy agencies, lobbying groups and the more murky “advisory bodies” with mass behaviour change as their raison d’être. This is particularly true in areas such as personal data, artificial intelligence, robotics, food and agricultural science.

The reason that this has not happened already is primarily that no one has thought through how it could be done. But the authorities that should address these issues must not be dissuaded merely because the task is daunting. It may be easier to establish regulatory mechanisms for “pure” academic research – but such research may be considerably less dangerous than the research that is currently unregulated.

If some research is worth monitoring, then all should be – in ways suited to their forms and function. It is not fair that non-academic researchers are allowed to get away with things that academics cannot. And history has already shown that relying on the vagaries of self-regulation outside the academic field is an invitation to abuse.

Ron Iphofen is an independent research consultant. He draws these observations from The Handbook of Research Ethics and Scientific Integrity, which he is editing for Springer Nature. With some chapters already published online, the full version will be published early 2020.