Written by Ben Davies
Personal autonomy is the guiding light of contemporary clinical and research practice, at least in the UK. Whether someone is a potential participant in a research trial, or a patient being treated by a medical professional, the gold standard, violated only in extremis, is that they should decide for themselves whether to go ahead with a particular intervention, on the basis of as much relevant information as possible.
Roger Crisp recently discussed Professor Gopal Sreenivasan’s New Cross seminar, which argued against a requirement for informational disclosure in consenting to research participation. Sreenivasan’s argument was, at least in its first part, based on a straightforward appeal to autonomy: if autonomy is what matters most, I should have the right to autonomously refuse information.
I have previously outlined a related argument in a clinical context, in which I sought to undermine arguments against a putative ‘Right Not to Know’ that are themselves based in autonomy. In brief, my argument is, firstly, that a decision can itself be autonomous without promoting the agent’s future or overall autonomy and, second, that even if there is an autonomy-based moral duty to hear relevant information (as scholars such as Rosamond Rhodes argue), we can still have a right that people not force us to hear such information.
In a recent paper, Julian Savulescu and I go further into the details of the Right Not to Know, setting out the scope for a degree of compromise between the two central camps.
These camps are grounded in two opposing views of autonomy, namely:
the ‘liberty’ approach, based on the work of John Stuart Mill, which sees autonomy as a matter of doing what one prefers; and a ‘duty’ approach, thought by some proponents to derive from Immanuel Kant’s work, which frames autonomy as a duty of self-government.
The liberty view is often supported by appeal to the idea (which was central to Gopal Sreenivasan’s understanding) that respecting autonomy involves respecting individual decisions that people make. The central example of Mill’s that is relevant here is his case of a traveller crossing a bridge that you know to be unsafe. In Mill’s view, you may stop the traveller to inform them of the danger but, if they decide to press on regardless, you may not interfere further.
Some see even in Mill the grounds for opposing a right not to know. For the unsafe bridge case rules out preventing an informed traveller from acting, but not from informing them in the first place. As we suggest, though, things are not so simple. We distinguish between two levels of decision, arguing that even if lacking information about their health means that a range of ‘practical’ decisions patients will make will lack autonomy, the ‘informational’ decision to refuse information can itself still be autonomous. However, we note that there are limits here. If a patient has no inkling that there is any information even to refuse, then their informational decision also cannot be autonomous. We imagine:
a patient who is informed by her doctor that ‘something has come up in your test results’ that the doctor would like to investigate further. Of course, the patient could theoretically imagine each possible outcome, consider the potential consequences, and make a somewhat informed decision not to receive further information. But this is practically beyond even most medical professionals, let alone those who lack relevant knowledge and expertise. A patient who refused information at this stage could not, in our view, be said to be making even a minimally informed decision.
Turn now to what we call the ‘duty’ view. Rather than seeing autonomy as requiring that one can do whatever one wants, defenders of this view think that autonomy requires acting rationally. And, as they sometimes point out, it is hard to act rationally if one lacks crucial information.
However, advocates of this view need to, but often don’t, take a stand on precisely how ‘being informed’ relates to ‘being autonomous’. One option is that ‘being autonomous’ is a binary state: you either are or you aren’t. But it is far from clear where such a line would lie: what level of information would one need to act autonomously? Furthermore, we suggest that such a view involves proponents of the duty approach into a paradox. As we suggest:
Consider Elizabeth, who lived in the late nineteenth century, long before the possibility of discovering whether one had the genetic basis for Huntington’s. Elizabeth lacks the very same information as Joe, who lives today and refuses a Huntington’s-related diagnosis. If duty theorists insist that Joe lacks autonomy (in, recall, a binary sense) because he lacks this information, they must also decide whether Elizabeth lacks autonomy.
But neither answer is satisfactory. If Elizabeth lacks autonomy, so did everyone living in her time. Moreover, it is unclear why we should set the bar here; there are plausibly many tests which are unavailable now, but will become possible in the future. Why not say that the information they will provide is also crucial for autonomy? But this would imply that nobody living today is autonomous either.
More plausible, of course, is a degree-based approach. Both Elizabeth and Joe are less autonomous than they could be with the knowledge they lack, but they do not lack autonomy entirely. Yet this makes the duty view, at the very least, more complex. Even if we have a duty to be autonomous rather than non-autonomous, it is much harder to argue that we have a continuous duty always to be more autonomous.
As we argue, the more plausible view is a satisficing one: as agents, we have a duty to receive sufficient information. Sufficient for what? Well, that is a task for individual duty theorists to decide. But as we suggest, the admission that the duty is not maximising at least brings the duty view closer to the revised version of the liberty view that we outlined previously.
We do not suggest that our arguments show that there is no daylight between the two approaches. Rather, at a practical level we hope to have shown that what is at stake between reasonable versions of these views is the degree of information required, not the basic question of whether patients can ever, or always, legitimately refuse information and remain autonomous.