In shared decision-making (SDM), patients and clinicians work together to determine the best next steps in health and care. Elwyn et al’s recent paper1 makes an important contribution to the literature about the applicability and limits of SDM.2 Their argument requires adherence to the traditional definition of SDM.

The definition of SDM the authors use—that is, the preference-based selection of an alternative based on its relative merits3—has enabled progress. It has fuelled work on standards for patient decision aids, measurement instruments, efficacy and implementation studies, and policy initiatives. Experts and policy makers believe implementing SDM, considered as something new or as an extrinsic addition to clinical practice, can promote patient engagement and empowerment, reduce the uptake of invasive and expensive care, improve outcomes (particularly adherence to therapy), and correct distortions in healthcare such as unwarranted variations in healthcare.4–6 Cottage industries of decision aid production and decision coaching have emerged as well as a policy initiative, far and few apart in the USA, for example, mandating SDM as a condition for payer reimbursement.7 The evidence to date would support the statement that despite four decades of work by our SDM community, the uptake of SDM is very scant in practice and that, therefore, the improvements in healthcare that SDM should have produced remain largely unrealised. Defined in this way, SDM is not an appropriate method to address some issues common in care. Elwyn et al identify some situations in which the options that can be subjected to deliberation are restricted and for which SDM, as defined, is inappropriate.

One could see this situation as a turning point. Hargraves has proposed that, indeed, there are problems in clinical care that are not appropriate for SDM when defined as a task of matching preferences to the features of available options; selecting care from a ‘menu’.8 Yet, rather than describing these instances as limits, Hargraves has uncovered that patients and clinicians use different joint, deliberative conversations to decide how to address these kinds of problems. The canonical definition of SDM does not include these other methods. One would not bring a flamethrower, an otherwise useful tool, but one that is rarely useful in clinical encounters because of the kind of problems that emerge there. Similarly, patients and clinicians make use of certain forms of SDM and fluidly change across these forms in their effort to uncover the best method to address the problematic human situation of the patient as it becomes clearer in conversation. He has proposed that there are (at least) three other forms of SDM—negotiating conflicts, solving problems or developing insight.9

There is no doubt that a problem can be addressed with one or another form of SDM or that several forms of SDM may need to be used. In some situations, therefore, patients and clinicians may find themselves selecting from existing options after considering their relative merits. In many other situations, however, preference-driven option selection may not be the main form of SDM used.

In a secondary analysis of encounters in our conversation bank, drawn from studies testing the incremental value of SDM tools used during the encounter, we found evidence of SDM in close to 90% of encounters, in which about one-third involve the use of the canonical form of SDM.10 This fundamentally changes the situation. SDM, defined using the purposeful SDM framework, may very well be a commonly used method of care.11 Indeed, our observation may confirm what clinicians have been telling us for years: that they use SDM frequently as a method of collaborative deliberation to solve patient problems.12 13 Like Elwyn et al, we understand that there are situations (interrupted, noisy and accelerated encounters, clinical emergencies, strong preference from participants to not engage in collaborative deliberation) in which caring with SDM is a more arduous or even impossible undertaking.

A criticism of the broader purposeful SDM framework is that it is so general that perhaps it simply describes patient-centred care. Purposeful SDM may be so encompassing and commonplace that clinicians may not find it worth spending time to learn SDM and improve its practice, or for researchers to take part in separate academic societies, or for agencies to issue targeted funding opportunities for research or policy to support SDM practice.

We agree that purposeful SDM contributes to patient-centred care, yet it does so intrinsically, not as an add-on, but as a method of care.11 A common occurrence in practice, SDM can be more easily studied across a broad range of problematic situations, and its improvement now calls for innovation. Enabling unhurried conversations and continuity of care become, for example, important tactics, beyond decision aids, to enable SDM. Thus, purposeful SDM shifts the target of these tactics from promoting SDM as a goal to enabling SDM as a method of care that helps patients and clinicians in the work of making care fit.14

We agree with Elwyn et al that there are situations in which the clinician is tasked with enforcing public health mandates (vaccinations), adhering to legal requirements (reporting of abuse), gatekeeping to enable health system goals (restricting access to specialty services) or upholding the clinician’s duty to act in the patient’s best interest (use evidence-based treatments over untested or ineffective alternatives). SDM may very well be inappropriate in these situations, and we agree with Elwyn et al that these are not situations in which canonical SDM may be helpful. And yet, for example, a form of SDM focused on clarifying positions (not preferences) and negotiating conflict (not matching preferences with the features of options) may prove useful in some instances. In fact, Elwyn et al propose patients and clinicians negotiate agreements for unscheduled opiate prescriptions or antibiotics for self-limiting illness.

Although focused on its limits, Elwyn et al state that SDM may be applied more widely than many clinicians think possible. We would add that perhaps SDM, as a method of care, may be applied even more widely than the field of SDM currently assumes.