Introduction
In this article I want to pick out some features of a talk that I gave recently about ethical decision-making in a medical context and apply these to another context; namely, to organisational data governance. In doing this, I’ll also make some general reflections on what it is that we’re doing when we make moral judgements and why close engagement with the ethical commitments that are in tension in any particular case matters for optimising decision-making.
A key aim of this article is to articulate a central principle of IGS’ approach to data governance; namely, that data ethics isn’t a desirable and optional feature of optimal data governance, but a necessary and integral feature, in all sectors, industries, and professions.
So, first, I’ll sketch out the medical context-specific talk that I gave; second, I’ll pick out an aspect of general relevance for ethical decision-making and explain why it matters; third, I’ll apply this to the data governance context in particular to show how applying the approach can help to ensure that the necessary processes are conducted in a way that is not only legally compliant, but also meets the necessary ethical standards.
Fourth, and finally, I’ll offer some reflections on how IGS’ consultancy services in both data ethics and compliance can help you and your organisation ensure that you are equipped with the skills required to make sound, well-balanced, judicious and robustly defensible decisions about the data governance challenges that you face.
Forensic Psychiatry Research Nottingham Summer Symposium 2024
The talk that I gave was in response to an invitation from Dr. John Tully and fellow clinicians in the forensic psychiatry group at the University of Nottingham’s Institute for Mental Health, to speak at their annual symposium. Forensic psychiatry is that bit of psychiatry which deals with the diagnosis, treatment, management, clinical decision-making and so on of patients who are either being held in secure psychiatric units or in prison, or who have been deemed sufficiently low risk to live in the community, subject to monitoring.
Forensic psychiatry is highly specialised as it deals with people who not only are affected by serious mental health problems, psychological disorders, learning disabilities, or capacity issues, but also people who have represented or might represent a risk of harm to themselves or others, often through criminal behaviour. Given the difficult balance that must be struck between restricting the liberty and privacy of these people as minimally as necessary but also ensuring that the public are protected and kept safe, forensic psychiatry is a profoundly ethically complex and serious domain of practice, since sometimes ensuring the latter can mean having to impose restrictions on the liberty and privacy of these individuals that are severe.
Given the moral dilemmas that this domain of practice often raises, the theme for the 2024 symposium was ethical and philosophical issues in forensic psychiatry. Within this I was asked to speak about ethical challenges involved in imposing restrictions on liberty and privacy, or what is known generally as ‘restrictive practice’
Ethics of restrictive practice in forensic psychiatry
Restrictive practice is, in short, the forced curtailment of the freedoms of psychiatric patients or prisoners deemed a danger to themselves or others. Methods used can include physical restraints, chemical restraints, seclusion, segregation, surveillance, and so on. Given my role at IGS in data ethics, in several places my talk touched on – but was not limited to -an intervention especially pertinent to the data ethics context; namely, electronic monitoring, more colloquially known as ‘tagging’.
Having introduced the scope of the talk, next I’m going to pick out and unpack one of its central features. This is a part of the argument that is relevant to practical ethical decision-making in general, irrespective of the particular domain of personal or professional life.
Ethical dilemmas and ethical judgements
When faced with a hard ethical choice, or what we might call an ethical dilemma, we need to figure out what we ought to do, given the options in front of us. If the decision is a true dilemma, it will be so in virtue of it being a difficult decision to make, in the sense that both outcomes produce a trade-off, which is to say, where both outcomes have some kind of morally serious negative consequence. This is quite straightforward to grasp, in the sense that if it were obvious or incontestable what we ought to do, because one option had no negative consequence of any kind, the dilemma would vanish.
What this means in practice is that an ethical judgement – arriving at a conclusion about what we ought to do – made in response to a dilemma is an all-things-considered matter, rather than something which is ‘perfectly’ morally correct, in the sense that deciding to take option A rather than option B would mean that there are no remaining ethical concerns whatsoever about option A.
As such, while we might eventually take a view about what, on balance, should be done to get off the horns of the dilemma and act, it doesn’t follow that the reasoning which led to that view completely dissolves the underlying moral tension which gave rise to the dilemma in the first place. The ethical tension remains, which is why ethical judgements are just that; namely, judgements, arrived at by rational deliberation and an attempt to arrive at a conclusion that is balanced.
Two examples: animal research and physician-assisted suicide
This all might seem a bit abstract so far, but there are numerous familiar examples which, once we think about them even briefly, underline the point.
For instance, we could think about a) the ethical permissibility or impermissibility of using animals in medical research, and b) the question of whether physician-assisted suicide should be decriminalised. Both are deeply contentious moral issues, and there are strong, principled arguments both for and against in each case.
For instance, on one hand animal research enables the development of therapies which can be used to treat disease and alleviate suffering in humans; but on other, this is achieved at the cost of often causing considerable harm and suffering to huge numbers of animals.
Likewise, on one hand, the decriminalisation of physician-assisted suicide would allow terminally patients to relieve themselves of suffering if they wished and uphold their rights to autonomy and self-determination in doing so; but on the other, it might arguably undermine the foundational clinical ethical commitment to the preservation of life and / or could increase the risk of people with very little time left feeling under pressure to shorten it even further because doing so would free up scarce medical resources that could be used for people who are less seriously ill.
Debate about the fundamental ethical permissibility of practices such as animal research or physician-assisted suicide goes back and forth and engaging in it productively requires great sensitivity to the various ethical commitments in tension to protect individuals, society, animals, and so on. Whichever way one happens to fall on either of the issues, and however strongly one disagrees with the opposing view, nevertheless, there are strong, persuasive, serious arguments to be made for and against both. It is not clear that there is an absolute right or wrong view to take about either, even though one probably finds one view more compelling than the other in each case. Rather, the best that we can hope for is as well-justified a reason, or argument, as possible for why one adopts one view rather than another. Why is this?
Well – and to return to a point I raised earlier – whether one is for or against either of the practices in question, each view produces a trade-off. Neither is completely without a moral cost of some kind.
For instance, if we think that animal research is, on balance, ethically permissible because it helps to reduce suffering in humans, we must accept that we are prepared to allow deliberately causing suffering to thousands of other animals which, although less cognitively sophisticated and self-determining than humans, can nevertheless experience pain and distress.
Likewise, if, all-things-considered, we endorse the decriminalisation of physician-assisted suicide, then we promote the autonomy of terminally ill people who wish to end their suffering and have control over when they die. But in doing so we have to accept that we are prepared to risk increasing the vulnerability of and potential perception of coercion among other terminally ill people who wish to be sure that their right to life is protected as far as medically possible and their death is not hastened in any way.
Note that in neither case is there an outcome where suffering or the risk thereof is avoided completely. As such, the decision about what, on balance, ought to be done is based on an evaluation of which harms should be prevented and which can be accepted as a price worth paying to produce the outcome that we endorse, and what our reasons are for evaluating them in the way that we do.
Ethical decision-making and the adequacy of reasons
Assuming that we think it’s important to try and justify difficult decisions like these, rather than thinking either that it doesn’t matter what we do either way, or that the best way to decide would be to flip a coin and leave the decision to chance, when it comes to ethical dilemmas, the justification for doing X or Y – where neither can prevent some harm or suffering from occurring – is and must be grounded in the adequacy of the reasons (the importance of which I touched on briefly in an earlier article) that one has for prioritising one outcome over another. And the adequacy of the reasons is key to the justifiability of the eventual decision, because all the competing factors have moral weight. This means that the decision-making process, the weighing of priorities, the attention to what constitutes an adequate reason must be carried out with a seriousness that is commensurate with the moral weight of what is stake in each course of action.
So, how can we summarise where we are at this point, before we consider what it means for ethical decision-making in data governance in particular? I think there are three things to say.
First, I hope it is clear from the unpacking done so far that arriving at a decision about what we ought to do when faced with an ethical dilemma does not neutralise the moral gravity of the option we decide not to take, nor the underlying values-based tension which gave rise to the dilemma.
Second, because of this, we should not assume that arriving at a judgement about what, on balance, we ought to do, is equivalent to eliminating further legitimate ethical disagreement or challenge.
Third, drawing the first two points together, it follows that we must be equipped, based on good reasons, with the skills needed to defend our decision against challenges that might be legitimate even if we disagree with them.
Having derived this general point about ethical decision-making, next, let’s think about what it means in the particular context of data governance.
Respect for human rights AKA the roots of ethical data governance
The first context-specific point to be made on the basis of the analysis so far is a reminder that wherever data pertains to people – which is to say autonomous, self-determining, moral agents who possess rights commensurate with such a status, and so on – the data is worthy of the same respect as those people to whom it pertains, because those people could come to harm if their data is handled irresponsibly.
This basic principle matters for ethical robustness and integrity in data governance, because it helps us not to lose sight of the fact that people are affected by our decisions. As we saw, true dilemmas always, by definition, entail negative consequences of some kind for some party whatever course of action is taken. Given, then, that if we are faced with a data-related dilemma, some non-trivial risk of harm some people or person will result from whatever we do, and the aim of data governance must be the prevention of that harm from occurring. Since we will therefore have to bear responsibility for any consequences of our decision, we need to be sure that the justification for taking that risk or, if the worst happens, causing that harm is grounded in the strongest possible reasons.
Another way to look at this which reinforces the central point is from a human rights perspective. Legislation such as the Human Rights Act and European Convention on Human Rights exists to protect the most foundational and incontrovertible freedoms – such as the right to liberty and the right to a private life, for example – that humans require as basic conditions necessary for them to be able to live good lives, where ‘good’ is understood as something like ‘flourishing’, or ‘in accordance with their own values, wishes, plans, desires, beliefs’ and so on. And, needless to say, we see human rights such as the right to privacy protected in relevant data protection legislation such as the Data Protection Act and UK General Data Protection Regulation.
It’s alarmingly easy to take the fundamental importance of these rights for granted, given that in a country like the UK we are, on the whole, fortunate enough to be able to enjoy them relatively freely, relative to many other countries around the world. That freedom can mean that active awareness of our basic rights being – for the most part – protected effectively can slip into the background.
Given the profound importance of these rights, then, it’s important to take seriously instances where we think it’s justifiable to override them – as we do, for example, in certain cases of incarceration in response to various kinds of criminal behaviour. In instances like these, we might also be so used to the idea that some people are imprisoned and have their liberty removed because of crimes that they’ve committed, that we don’t give it as much of a second thought as perhaps we should.
This is important, because if the right to something like liberty is so fundamental for every human, then it should be a reminder that the removal of such a right in response to criminal behaviour is a weighty decision that must be taken with a seriousness proportionate to the ethical significance of what is being done, even if, on balance, we deem it permissible or necessary, given harms that were done and / or still might be done, were the individual’s liberty not removed. This is to say, even if we think the removal of liberty is, all-things-considered, justified, this does not mean that the decision should be taken lightly, or that arriving at a judgement about what to do strips the factors being weighed up of their moral significance.
Data ethics in restrictive practice: electronic monitoring
Let’s think about an example in data ethics which I covered in the Nottingham talk to illustrate this point further.
I mentioned earlier the use in forensic psychiatry of electronic monitoring, more commonly known as ‘tagging’. When this is done, an electronic tag is attached to the patient for keeping track of their movements when, for example, on daytime release from hospital, which is a stage in the treatment and rehabilitation of patients where gradual reintroduction to society is deemed safe and appropriate.
Location data transmitted by the tag is collected continuously, for several reasons: to ensure that the patient observes the conditions of their temporary release, for example if there are places that they are prohibited from visiting or if they have an evening curfew by which time they are obliged to return to hospital; and so that if there is a risk of criminal activity, including harm to others such that they need to be involuntarily returned to hospital, their whereabouts is known.
The rationale for surveillance, then, is that the denial of the patient’s otherwise fundamental rights to liberty and privacy, such that their whereabouts are subject to restrictions and known at all times, is an acceptable and necessary price to pay for protecting the individual if they pose a risk of harm to themselves and / or others.
The ethical justification for tagging is complex and involves a tension between other factors as well. For example, we might wonder how the goal of rehabilitation towards independence and safe reintroduction to society without scrutiny can best be achieved if, when some degree of freedom is returned to them by being released on a tag, the patient is necessarily always aware that they are under surveillance. There is insufficient space to consider matters such as these here. For our purposes, it suffices just to sketch out the deliberative process according to which the denial of rights as fundamental as liberty and privacy might be justifiably overridden.
The essential point to be drawn from this case is that even if we conclude that it is justifiable to override the patient’s right to liberty and privacy for reasons X, Y, and Z, it is not as if we infer from this that the rights themselves are in some way less important than we thought they were before. Our assessment of the value of those rights remains the same despite our judgement about what ought to be done.
So, although there are circumstantial, harm-related, factors which mean that, on balance, regrettably it might be justifiable to deny those rights to someone, the basic value of the rights is unaffected. Since the basic value of the rights is unaffected, the underlying tension persists between obligations to, for example, a particular individual on one hand and to society on the other. So, given that the basic importance of the rights in tension is unaffected, we still need to consider what it means, morally speaking, to deny someone of them, even if we conclude that to do so is on balance morally permissible in some situations.
In pointing this out, I’m trying to show that the ethical importance of what is at stake when faced with a dilemma about what we ought to do persists, even when we arrive at a judgement about what that course of action is, all things considered.
Practical ethical implications for organisational data governance
So, next to bring the analysis closer to the concrete day-to-day realities of data governance. Assuming that the arguments I’ve put forward here are plausible, what can we say about their organisational implications in practice?
The unreliable relation of ethics to the law
Well, first, I suggest it’s a reminder that optimal governance is not satisfied by legal compliance alone. For sure, if we’re faced with a dilemma and need to decide, we can – and indeed should – use what is legally permitted and prohibited as a guide to action. It would make no sense to recommend deliberately flouting data protection regulations, given not only the betrayal of trust of those whose privacyis being breached but also the legal consequences and reputational damage that would follow for any organisation that chooses to do so.
Rather, the point I’m making here is somewhat different, which is that compliance with the law does not extinguish the need for reflection on what is at stake, ethically speaking. To return to one of the examples I used earlier by way of illustration – does it follow that the use of animals in medical research is devoid of ethical concerns just because it is legally permitted, and even if one thinks, on balance, that until there are better alternatives (and no longer than that) it should be legally permitted? I would suggest that even if this is what we conclude on balance, the fundamental ethical problem remains that we are allowing harm to be caused to creatures capable of experiencing suffering, which is something that we should generally commit to being a moral wrong.
The principle here is the same in data-related cases. Just because it is legal to electronically monitor psychiatric patients on day release, given the various interests in tension that must be managed, the underlying fact remains that even in complying with the law we must override an otherwise fundamental human right, and this is something that we should take seriously. And finally, to reinforce the point about the uncertain relation between ethics and the law, it is worth remembering historical examples of the law being used to protect very great moral wrongs. For a very long time, slavery was legal and homosexuality was illegal; but it does not follow from acting in compliance with the law either by keeping slaves or imprisoning people because of their sexual orientation that to do so was ethical. Likewise, women were long denied the right to vote, but in this case too it does not follow that it was ethical to deny women the right to vote just because the denial was protected in law.
Examples such as these and many others should remind us that although compliance with the law will be sufficient for an organisation to ensure that it can continue to operate, it cannot be assumed that compliance extinguishes every factor relevant to whether that organisation is or is not doing what it ought to do, in the more fundamental sense of a sufficiently close engagement with why it should and should not make particular data handling decisions.
The contestability of ethical judgements: avoiding decisional paralysis
The second point to make follows on here, and relates to this need for ongoing, close engagement with the ethically-relevant features of data handling decisions, as a standard component of good governance. Even though any organisation must, of course, follow data compliance regulations, the residual uncertainty about whether lawfulness necessarily guarantees ethical practice obviates the need to engage in the kind of critical reflection that I’ve encouraged here.
Now, it is fair to note a potential challenge here. One could object that if ethical judgements are, for the most part, always contestable, then there is no end to how long we could continue reasoning about what we ought or ought not to do, for example with respect to data handling. And this in turn might be self-defeating because, eventually, a decision must be made about how to handle data or not, and there is a risk that recognition of the fundamental uncertainty of ethical judgements could cause paralysis in decision-making. After all, if an absolutely watertight and incontestably ethically correct answer never emerges, what are we to do and on what basis?
A response to this challenge would be to say that even if close engagement never fully extinguishes ethical contestability – in the sense of never neutralising what is at stake; namely, how well people’s lives go – then a procedural commitment to as close engagement as possible can at least improve the decision-making process, which is a highly valuable exercise. I say this for a couple of reasons.
Ethical judgements can’t be perfect, but they can be better or worse
One reason is that the more closely engaged you are with the ethically relevant factors that need to be weighed when faced with an ethical dilemma – in which no outcome is risk-free – the more likely you are to be able to give a well-reasoned justification for what you eventually decide to do. By this I mean that engaging as closely as time and other practical considerations allow is likely to improve the adequacy of the reasons that you have for whatever conclusion you arrive at, as the depth of understanding of the issues will be reflected in the weighing that needs to be done. And this in turn helps to promote the integrity of your decision-making if it is challenged, which is always possible when trying to make ethical decisions, for reasons I’ve outlined.
A second reason is that, in general, the more thoroughly and deeply you engage with a complex topic, the more you tend to become aware of what you don’t know. Indeed, the deeper and more extensive your knowledge and understanding of something, the more specific you can be about what the gaps in your knowledge and understanding are. Moreover, it’s always possible that you could learn something which would confound or compel you to change a view that you currently hold. Given the contestability of ethical judgements, approaching ethical decision-making about data or anything else in an open-minded way, rather than being dogmatic about what the right answer definitely must be, is good practice for ensuring the level of care and attention that is required for making the reasoning process as robust as it can be. And given what is at stake in practical ethical decision-making is of the greatest possible significance, namely people’s lives and their rights, whether directly or via what happens to their data, then giving sufficient care and attention to optimise the robustness of the reasoning process is appropriate and required.
Intrinsic and instrumental reasons
So, before concluding, what can we draw from all this analysis? Perhaps we could summarise it in the following way.
There are what we might call both intrinsic reasons and instrumental reasons for integrating ethical reasoning into data governance processes alongside compliance with law and regulations.
We might understand intrinsic reasons as reasons that are valuable for their own sake, which speaks to the point I made earlier about the fundamental rights and values that are at stake in ethical decision-making and dilemmas. It matters what we decide to do because all people have moral worth and to the same degree. In virtue of this, we are obliged to make sure that our decisions about, for example, how to handle people’s data, are based on a respect for the moral worth of those people. And since ethical dilemmas produce trade-offs, then even if our decision is lawful, we need to make sure that what we decide is grounded in a deliberation of what it is ethically permissible to trade off and why.
Instrumental reasons, by contrast, are reasons that we might understand as valuable not by themselves but in the service of achieving a further goal. An instrumental justification would be to say that we should include ethical reasoning into data governance processes because doing so will help to make those processes more comprehensive, more attuned to the full range of relevant factors which come to bear on decisions about data handling, and so on. This, in turn, will have the effect of improving the decision-making process, ensuring that it has integrity and is trustworthy, by virtue of creating the conditions in which decisions and the trade-offs that they entail can be grounded in an optimally rigorous deliberation.
Finally, here, these intrinsic and instrumental reasons come together when we consider something I mentioned earlier; namely, that even when we come to a rigorously justified conclusion, firmly grounded by good, defensible reasons, according to which we can justify the trade-off entailed by whatever we decide to do, nevertheless the contestability of the ethical judgement does not completely go away.
Since rights, obligations, and values will always come into tension when we’re trying to determine whether we should do X or Y about a situation in which humans and their interests are involved, then if the decision is ethically challenging at all, it is so by virtue of it being about something that is not trivial and should be treated with a seriousness proportionate to that. And if, for intrinsic and instrumental reasons, we incorporate ethical deliberations into our data governance processes as an integral feature of those processes, we are more likely to identify what is, all-things-considered and on balance, the right course of action.
Conclusions
In conclusion it might help to use this intrinsic / instrumental reasons distinction to make some final remarks.
If your organisation handles people’s data, there are intrinsic reasons to consider what you ought and ought not to do with it and why. The data pertains to moral agents with rights, and as such there is a basic obligation to ensure that this is being respected in data governance decisions alongside compliance with data protection regulation. And if your organisation handles people’s data, there are instrumental reasons why it is good practice to do so as well. Reputational damage caused by untrustworthy, unreliable, insufficiently robust data governance processes can threaten the bottom line of any organisation and it is therefore in the interests of any organisation that wishes to hold people’s data that it can demonstrate trustworthiness. And since trust is a moral concept, engagement with what this requires in practice is rational with respect to the protection of that bottom line.
I hope the analysis laid out here helps to demonstrate IGS’ approach to data governance and the advice thereon that we give. We believe that ethical deliberation is not just a nice but basically unnecessary requirement of data governance above and beyond legal compliance. For the reasons I’ve outlined here, our view is that ethical deliberation is an integral part of data governance, which interlocks with compliance processes, and helps to further minimise risks and optimise organisational data governance across the board. If your organisation needs advice and support with any aspect of ethical and legal aspects of data governance, we are here to help.
