An article published in Nature piqued my interest this month. Entitled ‘Guidelines for academics aim to lessen ethical pitfalls in generative-AI use’, the byline warned about normalising the use of artificial intelligence (AI) in research without safeguarding against the risks. Rather than scaremongering, the bigger aim of the piece was to signpost to a project currently underway to develop guidelines for researchers and university ethics committees in the UK on the use of generative AI with participants’ data to protect participant privacy. The article also signposts to similar guidelines being developed by the European Research Area Forum with the European Commission. The impetus for the UK project is described as a sudden realisation by the project lead that feeding information they had collected via research interviews into a generative AI programme ran the risk of making participants accidentally ‘re-identifiable’ if the tool itself pieced together the anonymised data it was being given with data publicly available elsewhere. As the British Psychological Society’s ‘Ethics guidelines for internet mediated research’ point out: internet-mediated research (IMR) can raise particular, sometimes non-obvious, challenges in adhering to existing ethics principles. These concerns are couched against the existing four pillars of: respect for the autonomy, privacy and dignity of individuals and communities; scientific integrity; social responsibility; and maximising benefits and minimising harm.
Of course, IMR can encompass a whole range of methods and approaches to data collection, but one of these is the collection of online information as potential research data. Consider, for example, that what people disclose in discussion forum posts, social media sites and app-enabled fields can all be considered ‘data’ and analysed in research. Should such information really be considered in the public domain and/or readily available for use as research data, for example, without gaining informed consent? Is it ethically acceptable to use such data freely for research purposes? And if we do, then what happens to the visibility, traceability and permanence of such data when it is not only collected from ‘publicly available’ domains, but then fed into generative AI tools for ease of analysis? These are the important questions of our time that, ideally, I would like us to contemplate as pharmacists and scientists. To consider them carefully, is dependent on what I call ‘research ethics literacy’. I would of course be very interested to hear from readers their thoughts on where we are with this form of literacy in pharmacy and how we might improve it. So, this month, I review a paper published in the suite of Pharmaceutical Press publications against the notion of research ethics, to consider specifically the authors’ description of good research ethics practices.
The paper is published in the International Journal of Pharmacy Practice. Entitled ‘Understanding medication recycling practices in Canadian hospitals’, the authors used an online survey to determine if Canadian hospitals recycle medications, gain an estimate of what amount (percentage) is recycled based on the dosage form, and examine the barriers to such practices. In terms of the researchers’ adherence to ethical standards, I was delighted to see an extensive section of the paper illustrating the principle of respect for the autonomy, privacy and dignity of the participants, via the collection of valid consent. It is not often that I see research papers in pharmacy devote substantial text in their methods to highlight the pursuit of this principle. To address the issue of ‘autonomy’, the authors first made sure that they only sent their survey to consenting sites; therefore, while freely available resources were used to identify a list of 896 potential hospital pharmacies and contact information, the survey was only sent to sites that actively agreed to receive it during a follow-up call that followed the distribution of a letter of intent. These elements indicate not only that the participating sites consented to take part in the research, but that they did so based on valid information about the study (i.e. informed consent).
Additionally, while two survey questions collected information about the name and nature of the hospital site, the methods section also explained that the responses to these questions were used only to remove duplicate responses from the same site and were then eliminated from the dataset used for the analysis (data were de-identified). In this sense, the researchers are explaining how they adhered to the same principle named above by protecting the privacy of the participants and ensuring that the sites/respondents are not identifiable. Participants were also given the option to take part in further research to explore their medication recycling practices, to which they could actively consent by providing their email address in the final two questions of the survey, another illustration of how the researchers respected the ‘autonomy’ of their participants.
On the principle of scientific integrity, the researchers explain the quality and design of their survey by detailing the involvement of experienced pharmacists in the study design, as well as efforts made to ensure the face validity of the questionnaire before its distribution. After all, research that is poorly designed could devalue the contribution of the participants and waste their time, as well as the researchers’ time.
As with most research, the principle of social responsibility is addressed in the introduction to the paper, which reviews the existing literature and sets out a rationale for the work itself. This is nicely summarised in the abstract as the fact that although medication recycling within hospitals in Canada has proven financial and potentially beneficial environmental effects, the extent of such practice remains unclear.
The final principle of maximising benefit and minimising harm is, in my view, implicit in the work; the risk of harm to participants for completion of a short survey being very low against the potential to unearth medication recycling practices that could help increase the responsible use of medicines and reduce their negative environmental impact. I would be interested to know what other ethics-related considerations readers might identify.