Shared Decision Making: A Model for Clinical Practice

The principles of shared decision making are well documented but there is a lack of guidance about how to accomplish the approach in routine clinical practice. Our aim here is to translate existing conceptual descriptions into a three-step model that is practical, easy to remember, and can act as a guide to skill development. Achieving shared decision making depends on building a good relationship in the clinical encounter so that information is shared and patients are supported to deliberate and express their preferences and views during the decision making process. To accomplish these tasks, we propose a model of how to do shared decision making that is based on choice, option and decision talk. The model has three steps: a) introducing choice, b) describing options, often by integrating the use of patient decision support, and c) helping patients explore preferences and make decisions. This model rests on supporting a process of deliberation, and on understanding that decisions should be influenced by exploring and respecting “what matters most” to patients as individuals, and that this exploration in turn depends on them developing informed preferences.


INTRODUCTION
Sharing decisions, as opposed to clinicians making decisions on behalf of patients, is gaining increasing prominence in health care policy. [1][2][3][4] Shared decision making (SDM) has been defined as: 'an approach where clinicians and patients share the best available evidence when faced with the task of making decisions, and where patients are supported to consider options, to achieve informed preferences". 2 The principles of SDM are well documented and the common elements have been summarized. 5 The earliest mention was in 1982, 6 but the idea draws on and deepens the principles of patient centered care. 7,8 Others 9,10 provided more detail and this led to a greater focus on the skills required. 11,12 Yet, despite attention to principles and competences, there remains a lack of clear guidance about how to accomplish SDM in routine practice. Our aim is to translate conceptual descriptions into a three-step model that is practical for clinicians. The purpose of this article is to advance a novel, yet pragmatic, proposal about how to do SDM in routine settings, in short to integrate good communication skills with the use of patient decision support tools.

GUIDING ETHICAL PRINCIPLES
The skills of SDM are unlikely to be developed, let alone exhibited, unless the clinician agrees with the guiding ethical principles. At its core, SDM rests on accepting that individual self-determination is a desirable goal and that clinicians need to support patients to achieve this goal, wherever feasible. Self-determination in the context of SDM does not mean that individuals are abandoned. SDM recognizes the need to support autonomy by building good relationships, respecting both individual competence and interdependence on others. These are the key tenets of both self-determination 13 and relational autonomy. 14 Self-determination theory is concerned with our intrinsic tendencies to protect and preserve our well-being. 13 Relational autonomy is the term used to describe the view that we are not entirely free, self-governing agents but that our decisions will always relate to interpersonal relationships and mutual dependencies. 15 As King and Moulton have noted, these principles extend the concept of informed consent beyond that of simple information transfer towards honoring informed preferences. 16 We acknowledge that good clinical practice balances these principles with those of beneficence and justice. 17 However, some healthcare professionals express doubts, saying that patients don't want to be involved in decisions, lack the capacity or ability, might make 'bad' decisions, or worry that SDM is just not practical, given constraints such as time pressure. Others claim they are 'already doing it', though data from patient experience surveys indicates that this is not generally the case. 18,19 It is therefore clear that the first step for those advocating the uptake of SDM is to ensure that clinicians and others support the underlying rationale.
Before doing so however, we need to note the challenges that clinicians will be navigating. Low health literacy or low numeracy will be barriers to SDM and some patients come from cultural backgrounds that lack a tradition of individuals making autonomous decisions.
We cannot therefore emphasize too strongly that SDM has to be built on the core skills of good clinical communication skills, as recognized in many seminal texts, [20][21][22][23] including building rapport and structuring the consultations. 24

WHY SHARE DECISIONS: BEYOND THE ETHICAL IMPERATIVE
SDM is supported by evidence from 86 randomized trials showing knowledge gain by patients, more confidence in decisions, more active patient involvement, and, in many situations, informed patients elect for more conservative treatment options. 25 We illustrate the arguments in favor of SDM by providing two hypothetical cases where more than one reasonable treatment option exist-see cases 1 and 2. 12 They illustrate that informed preferences are an optimal goal because the decisions made will be better understood, based on more accurate expectations about the negative and positive consequences 26 and more consistent with personal preferences.

DOING SHARED DECISION MAKING
We propose that achieving SDM depends on tasks that help confer agency, where agency refers to the capacity of individuals to act independently and to make their own free choices. 27,28 SDM aims to confer agency by 1) providing information and 2) supporting the decision making process.
Providing Information. We help patients participate by providing high quality information. We also need to elicit

Case 1 Katherine: early stage breast cancer
Katherine (aged 67) had recently been diagnosed with breast cancer. She was widowed, living alone in a rural location and did not drive. She was offered a choice between lumpectomy with radiotherapy (breast conservation surgery) or mastectomy, and was told of the equal survival rates for the two procedures. She was surprised by this choice and became anxious. She listened to the advice and, although she was given good information, felt steered towards having a lumpectomy and radiotherapy as the "less invasive" option. She became very tired during the radiotherapy, and her breast became tender and much smaller, an effect that she did not anticipate. Two years later, an ipsilateral local recurrence of the breast cancer necessitated a mastectomy. At this point, she became aware that there was a higher (double) rate of local recurrence after lumpectomy. She felt regret and considered that her decision might have been different if she had been given more information and a chance to express her strong wish to avoid recurrence.

Case 2 Edward: symptoms due to an enlarged prostate
Edward (aged 75) had recently been diagnosed as having an enlarged prostate gland causing him bothersome urinary symptoms. He was offered surgery as the most effective treatment and accepted the recommendation. Before surgery he enjoyed an active sex life which was important to him and his wife but this was seriously affected by the surgery. He had been made aware that some men have sexual problems after surgery but he did not feel as if he'd had a chance to consider the extent of this risk or to consider whether this was a concern to him personally. Looking back, he feels that if he had been given more of a chance to discuss his preferences, he may have postponed surgery in favor of "watchful waiting".
what patients already know, and whether it is correct. People place different importance on the outcomes associated with different options and have different preferences about the processes and paths that lead to these outcomes. If patients are not informed, they will be unable to assess 'what it is important to them', and so establish informed preferences. The first task of SDM is to ensure that individuals are not making decisions when insufficiently informed about key issues, not 'making decisions in the face of avoidable ignorance' (Al Mulley, personal communication). Many tools have been designed to help achieve this goal. 29 Detail about these tools and their effects can be found elsewhere: 11 in this article we will describe how to deploy them as part of doing SDM.
Supporting Deliberation. The second task is to support patients to deliberate about their options (see Fig. 1), by exploring their reactions to information. When offered a role in decisions, some patients feel surprised, unsettled by the offer of options and uncertainty about what might be best. 30 If all responsibility for decision making is transferred to patients they may feel 'abandoned'. 31 Some patients initially decline decisional responsibility role, and are wary about participating. 32

A MODEL FOR CLINICAL PRACTICE
To accomplish SDM, we propose a three-step model for clinical practice (see Fig. 1). We want to emphasize that this is a simplified model that illustrates the process of moving from initial to informed preferences. We acknowledge that this process also has psychological, social and emotional factors that will influence this deliberation space and that will need to be managed by an effective clinician-patient dialogue, seeking what Epstein has termed a 'shared mind'. 33 However, accepting these requirements, we aim for parsimony. We describe three key steps of SDM for clinical practice, namely: choice talk, option talk and decision talk, where the clinician supports deliberation throughout the process ( Fig. 1 and Boxes 1, 2 and 3). Choice talk refers to the step of making sure that patients know that reasonable options are available. Option talk refers to providing more detailed information about options and decision talk refers to supporting the work of considering preferences and deciding what is best. The model outlines a step-wise process, although it is important to recognize that the model is not prescriptive-clinical interactions are by necessity fluid. Decision support tools provide crucial inputs into this process.
Choice talk is about making patients that reasonable options exist. This step does not necessarily have to be done face-to-face -an email, letter or a telephone call can also be effective: e.g. asking a patient whose tests come back showing a herniated intervertebral disc to use a decision support website.
'Choice talk' is a planning step (45). Components of the choice talk include: a) Step back. Summarise and say: "Now that we have identified the problem, it's time to think what to do next" b) Offer choice. Beware that patients often misconstrue the presentation of choice and think that the clinician is either incompetent or uninformed, or both. Reduce this risk by saying: "There is good information about how these treatments differ that I'd like to discuss with you." c) Justify choice. Emphasise: 1) the importance of respecting individual preferences and, 2) the role of uncertainty.
Personalizing preferences: Explaining that different issues matter more to some people than to others should be easily grasped. Say: "Treatments have different consequences … some will matter more to you than to other people…" Uncertainty: Patients are often unaware about the extent of uncertainty in medicine: that evidence may be lacking and that, individual outcomes are unpredictable at the individual level. Say: "Treatments are not always effective and the chances of experiencing side effects vary…" d) Check reaction. Choice of options may be disconcerting: some patients may express concern. Suggested phrases: "Shall we go on" or 'Shall I tell you about the options?" e) Defer closure. Some patients react by asking clinicians to "tell me what to do …" We suggest that deferring closure if this occurs, reassuring that you are willing to support the process. Say: "I'm happy to share my views and help you get to a good decision. But before I do so, may I describe the options in more detail so that you understand what is at stake?"

Box 1. Choice talk
Patients will want time to study new information and to consider their personal preferences, particularly for futures that are unknown to them, to think about outcome states that they have never experienced. 34,35 Deliberation may, in part, be done outside the clinical encounter, although often patients wish to consolidate their views with a trusted clinician. Individuals often want to discuss options with others and it would be best if those involved could potentially use the same information resources (see below). Rapley has referred to this need to talk to others, at different times and places, as a 'distributed' deliberation process. 36 Recognizing this need, and allowing time for it, is a cornerstone for effective SDM . 33,37 The model also includes the use of decision support interventions, 38 which summarize information in formats that are accessible to patients, using the most up to date evidence about the harms and the benefits. 39 Some tools also include preference clarification exercises. 39 Decision support for patients can be in concise formats, such as in brief text or  (46)(47), about framing effects and the importance of providing risk data in absolute as well as relative terms. Try giving information in 'chunks' (chunking and checking) (48).
d) Provide patient decision support. These tools make options visible and may save time. Some are sufficiently concise to use in clinical encounters (38). Examples of these short tools are Issues Cards (49), Decision Boards (50), and Option Grids (http://www.optiongrid.co.uk/) (42). SDM may need more than one encounter. More extensive patient decision support tools may play a crucial role (51). Say: "These tools have been designed to help you understand options in more detail. Use them and come back so that I can answer your questions ". e) Summarize. List the options again and assess understanding by asking for reformulations. This is called a 'teach-back' method and is a good check for misconceptions.
a) Focus on preferences. Guide the patient to form preferences. Suggested phrases: "What, from your point of view, matters most to you?" b) Elicit a preference. Be ready with a back-up plan by offering more time or being willing to guide the patient, if they indicate that this is their wish.
c) Moving to a decision. Try checking for the need to either defer a decision or make a decision. Suggested phrases: "Are you ready to decide?" or "Do you want more time? Do you have more questions?" "Are there more things we should discuss?" d) Offer review. Reminding the patient, where feasible, that decisions may be reviewed is a good way to arrive at closure.

Box 2. Option talk
Box 3. Decision talk diagrams, and used during encounters to initiate SDM. They can also be extensive: typical of the many tools already developed-booklets, websites, videos, DVDs-to used by patients, their family and friends, outside the encounter, and at different time points 38 -see Box 1, 2 and 3, and synopsis in Box 4.

DELIBERATION
We use the term "deliberation" (see Fig. 1) to describe a process of considering information about the pros and cons of their options, to assess their implications, and to consider a range of possible futures, practical as well as emotional. This 'deliberation' space, colored grey in the figure, encompasses the need to work collaboratively with professionals as well as with the wider networks that patients will use. 36 Deliberation begins as soon as awareness about options develops. The process is iterative and recursive, and the intensity increases after options have been described and understood.

DISCUSSION
We have proposed a model of how to do SDM in clinical practice ( Fig. 1), based on three key steps, namely choice talk (Box 1), option talk (Box 2) and decision talk (Box 3), whilst also being aware that many other people may be contributing. There are implications for training: in our experience the best way to learn these skills is to use simulations, either with colleagues or with trained actors 11,40,41 Figure 1. A shared decision making model.

Choice talk
Step and use brief patient decision support tools. 38,42 There are measurement scales to assess skillfulness, 43,44 although we lack a measure to assess proficiency in risk communication. The use of brief patient decision support tools can catalyze SDM. 38,42 This model builds on the previous work in this field by integrating a number of contributions. It acknowledges the foundations in ethics 9,10 as well as the work that describes the stages and skills required. 5 However, it was our experience in implementation studies that gave rise to this three-step model that aims to help clinicians integrate SDM and patient decision support into their work. 42 Many clinicians will push back at the suggestion that yet more has to be accomplished in clinical encounters. We acknowledge this concern and argue that new systems will be required to appropriately reward truly patient centered practice. The introduction of brief decision support interventions can act as a catalyst for a new discourse and help make SDM a practical reality in busy clinics, albeit one that may lead to some patients needing more than one encounter where they can discuss important decisions. We realize that this model is a simplification of a complex, dynamic process, yet its simplicity may help others accomplish and teach shared decision making. That was our goal.