CCBYNC Open access

Implementing shared decision making in the NHS: lessons from the MAGIC programme

BMJ 2017; 357 doi: (Published 18 April 2017) Cite this as: BMJ 2017;357:j1744
  1. Natalie Joseph-Williams, lecturer1,
  2. Amy Lloyd, research fellow2,
  3. Adrian Edwards, professor1,
  4. Lynne Stobbart, senior research associate3,
  5. David Tomson, executive partner and freelance consultant in patient centred care44,
  6. Sheila Macphail, consultant in obstetrics and fetal medicine and assistant medical director5,
  7. Carole Dodd, director6,
  8. Kate Brain, reader1,
  9. Glyn Elwyn, professor6,
  10. Richard Thomson, professor3
  1. 1Division of Population Medicine, Cardiff University, Cardiff, UK
  2. 2Centre for Trials Research, Cardiff University, Cardiff, UK
  3. 3Institute of Health and Society, Newcastle University, Newcastle upon Tyne, UK
  4. 4Collingwood Surgery, North Shields, UK
  5. 5Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle, UK
  6. 6CK Health Consultancy, Newcastle upon Tyne, UK
  7. 7Dartmouth Institute for Health Policy and Clinical Practice, New Hampshire, USA
  1. Correspondence to: Natalie Joseph-Williams josephnj1{at}

Shared decision making requires a shift in attitudes at all levels but can become part of routine practice with the right support, say Natalie Joseph-Williams and colleagues

Adoption of shared decision making into routine practice has been remarkably slow, despite 40 years of research and considerable policy support.12345678 In 2010, the Health Foundation in the UK commissioned the MAGIC (Making Good Decisions in Collaboration) programme to design, test, and identify the best ways to embed shared decision making into routine primary and secondary care using quality improvement methods (box 1).910

Box 1: Summary of MAGIC programme

Phase 1 (August 2010–February 2012): Designing and testing the best ways to implement shared decision making in routine clinical settings
  • Sites

  • Cardiff University (lead), Cardiff and Vale University Health Board

  • Newcastle University (lead), Newcastle upon Tyne Hospitals NHS Foundation Trust, Northumbria Healthcare NHS Foundation Trust

Clinical areas

  • General practices (four each in Cardiff and Newcastle)

  • Breast cancer (surgical options for early stage breast cancer)

  • Chronic kidney disease (treatment for end stage renal failure), Cardiff

  • Head and neck cancer (treatment options for head and neck cancer), Cardiff

  • Paediatric ears, nose and throat (surgery for recurrent tonsillitis), Cardiff

  • Maternity (options relating to mode of delivery after caesarean, place of birth, and screening for Down’s syndrome), Newcastle

  • Urology (treatment options for prostate cancer), Newcastle

Phase 2 (February 2012–October 2013): Demonstrate that shared decision making can become part of routine clinical care (wider dissemination, implementation and sustainability) and build practical and transferable knowledge about the conditions for success


  • Work with up to 10 clinical teams across the Cardiff and Vale University Health Board to involve more patients in shared decision making (spread from phase 1)

  • Teams had access to support and interventions/approaches developed during phase 1: shared decision making skills, decision support tools, patient participation, quality improvement support


  • Provide tools, materials, and resources to enable clinical implementation teams from phase 1 to continue their plans for sustainable change

  • Engage with four new clinical teams (two each from primary and secondary care)

  • Engage with third sector organisations, providing versions of skills training workshops to service users and advocates

  • Further develop tools, materials, and processes to facilitate spread of shared decision making and to demonstrate sustainable change

The learning from MAGIC derives from a variety of sources, including facilitated shared learning events, clinic and consultation observations, interviews with clinicians and patients, patient and public involvement panels, focus groups, and questionnaires. We assessed progress using “plan do study act” data collection tools,11 monthly project team meetings (including researchers, clinical teams, healthcare organisations, and patient representatives), and an independent evaluation report of phase 1.10 Here, we draw on our learning from the three year programme and subsequent experience to summarise the key challenges of implementing shared decision making and to offer some practical solutions (table 1).

Table 1

Recommended solutions to challenges to implementing shared decision making*

View this table:

Challenge 1: “We do it already”

Changing attitudes is a key challenge for any change programme. Both structural change, in terms of healthcare pathways and delivery, and culture and attitudinal change among clinicians and patients are required for shared decision making and patient choice to become routine. Many clinicians feel that they already involve patients in decisions about their care, so often do not see how shared decision making differs from their usual practice. A minority of clinicians view their role as “decision maker” to act in the best interests of their patients. Clinicians’ long held commitment to doing what they perceive to be the best for their patients is a key barrier to attitudinal change. This is well intended, but fails to recognise that patients’ values, opinions, or preferences are important and might differ from their own.

An essential step in implementing shared decision making is to increase understanding of what it entails.1113 Clinical teams need support to review current practice, to build a shared understanding of how shared decision making differs from their current practice, and to decide how they want to make decisions with patients. We found interactive skills training workshops based on a shared decision making model (fig 1) 12 helped build coherence, improving skills, and promoting positive attitudes.14


Fig 1 Shared decision making model for clinical practice12

Workshop feedback indicated that role play based training, which emphasised practical skills, worked better than theory heavy presentations. We increasingly included exercises that challenged embedded attitudes and promoted discussion around them. The teams were generally already good at recognising options and discussing them with patients, but there was potential to improve their communication of risk and the task of exploring what matters to patients. The training helped clinicians understand how shared decision making differed from their current ways of working. Some clinicians reported changing their view from, “We do this already” to, “We could do this better.”

Challenge 2: “We don’t have the right tools”

Many clinicians believe that a decision aid will itself enable shared decision making and that decisions cannot be shared without a tool to give their patients. A key learning point from the MAGIC programme was that “skills trump tools, and attitudes trump skills.” Developing attitudes and understanding is essential, but then clinicians need to consider their communication skills to engage patients in decision making, drawing on evidence based tools when appropriate. There will never be decision support tools for every decision; nor will every patient find them acceptable or helpful. The skills to have different types of conversations with patients are paramount, with or without an available tool.

In the skills training workshops, role play was particularly effective for showing that tools may support the process but do not replace communication skills. Patient information sources designed for use outside the consultation, such as websites, are costly and time consuming to develop and keep updated. Developing brief decision support tools helped overcome this challenge.1516171819 Clinicians willingly designed brief evidence based tools to use inside the consultation, such as Option Grids in the Cardiff sites15 and brief decision aids in the Newcastle sites.19 These provide short (one to three pages) summaries of the treatment choices, the likely outcomes, and the factors that patients might consider when making their decision, including risk and benefit data.

Experience from MAGIC suggests that in-consultation tools are often better at facilitating discussion between patient and clinician than those used outside the consultation. Some patients used the tools to guide their questions for the clinician and it prompted them to discuss what mattered to them—the nub of shared decision making. However, the risk is that clinicians use brief decision aids to enhance information transfer and talk at patients, rather than improving how they work with patients.

Challenge 3: “Patients don’t want shared decision making”

Clinicians often report that their patients do not want to be involved in making healthcare decisions. This might be the case for some patients, and some patients want different levels of involvement, so the shared decision making process should respect the patient’s preference. This preference should itself be informed, rather than based on a clinician’s presumption about what the patient wants. Many patients feel unable rather than unwilling to share in decision making.202122 Some patients think they will annoy clinicians by trying to be more involved,23 and their desire to be a “good” patient over-rides their desire for sharing decisions. Often this comes from longstanding experience and expectations of a paternalistic approach and seems to be more common in older people.2425 This can be mistaken for lack of interest in engaging in decision making.

Clinicians’ misconceptions about what patients want must be addressed. It is neither possible nor desirable to make every patient be more involved, but more can be done to make patients feel included and respected, and clinicians can often make more effort to understand what is important to the patient. However, patients may also need support and preparation to take part in a different type of consultation.

Patient activation and preparation can increase the likelihood of mutually useful conversations between patients and clinicians. Patient activation campaigns focus on changing patients’ attitudes about involvement in healthcare decisions, explaining what shared decision making involves, why it might help, and provide interventions such as question prompt lists. Interventions such as the Ask 3 Questions campaign can help patients know what to expect and give “permission” and encouragement to be involved.262728

Successful implementation also needs wider patient and public involvement. In the MAGIC programme we included patients with the relevant conditions in designing local interventions as well as having a wider user panel to guide the broader implementation process. The lay panel was crucial in identifying areas for improvements in the implementation work, identifying users’ needs, and aiding intervention development and testing.

Challenge 4: “How can we measure it?”

Clinicians and managers implementing shared decision making want to know what difference it makes to their patients and to clinical practice. In the MAGIC programme we found it difficult to identify or develop suitable patient reported measures to capture experience of shared decision making. Patient reported measures are hampered by social desirability bias (wanting to give high satisfaction ratings), and patients may also not fully understand and identify shared decision making if they have not experienced it previously.29 However, the three item CollaboRATE measure (used in over 40 studies worldwide) shows promise in overcoming these problems.30

Tensions exist between the need for validated and reliable measures for research and measuring for quality improvement. Focusing on quality improvement helped embed shared decision making more readily with some clinical teams. When measures directly informed practice and improved patient care we witnessed greater motivation to improve and to sustain the improvement (eg, the decision quality measure in breast cancer, table 1).31 Linking with local quality improvement expertise and resources, where available, was valuable. For example, the 1000 Lives Plus national improvement service for NHS Wales had supported development of quality improvement skills across the Cardiff site, which was used to implement shared decision making.32

Challenge 5: “We have too many other demands and priorities”

Changing attitudes and behaviours takes effort at all organisational levels. Clinical teams face many competing demands and priorities, some of which are compulsory, some even incentivised (whether financially or by targets). For example, the Quality and Outcomes Framework rewards general practitioners for behaviours that are evidence based but not necessarily about what matters most to patients. Tensions are also increasing between shared decision making and clinical threshold guidelines or referral management schemes, which are widely applied within commissioning in England. Similarly, for cancer treatment time targets, the current emphasis is on time to treatment but patients may prioritise time to make the decision.

Visible organisational buy-in and support are essential. During the MAGIC programme, key organisational leaders showed clinicians that shared decision making was an important organisational priority to drive improvement (table 1) and clinical leadership was critical to implementation. This led to greater engagement because clinicians then saw shared decision making as something the organisation does, rather than as another initiative being imposed on them and competing with other demands. Teams sometimes needed support from the organisation to adapt clinical pathways to support effective shared decision making. Shared decision making is not the sole responsibility of doctors; it should be supported by all members of the clinical team. Responsibility can be distributed—for example, a breast care surgeon can explain the choices, but the specialist nurse can elicit the patient’s preferences in more detail. It is important to share learning about what works both within and between teams to avoid “reinventing the wheel.”

Recommendations for implementation

We have summarised the key challenges we faced during a programme to implement shared decision making in organisations across primary and secondary care, but many other factors also affect implementation (fig 2). More detailed discussion of the barriers and solutions is available elsewhere.42233 Similar findings have been reported in the US.343536


Fig 2 Summary of key factors influencing implementation of shared decision making (SDM). (QOF=Quality and Outcomes Framework)

In the real world, finance, resources, and time are all scarce. Although shared decision making is mentioned in key policy documents, such as the NHS Constitution 2015,2 it has no incentives and is not promoted systematically at national, regional, or organisational levels. We still need to foster cultural change among clinicians and patients, and this is a momentous challenge. A receptive culture will truly exist only if clinicians view shared decision making as usual practice and as a fundamental component of safe, effective, and compassionate healthcare for patients. It needs to be embedded in the medical and nursing curriculums and interprofessional training programmes. Similarly, much health policy and professional practice has led to feelings of powerlessness and passivity for many patients. Increasing patient agency, activation, and health literacy are equally important.37

Although these actions may help embed shared decision making, they do not standardise the process: every discussion varies, depending on the patient, clinician, their preferences, and the type of decision being made. Shared decision making may not necessarily result in, or depend on, complete agreement between a clinician and a patient. Instead, it is about bringing both types of expertise together, and weighing up the available options in light of both of these perspectives; it makes it more likely that the final decision is informed by what the clinician knows (medical evidence, clinical experience) and by what the patient knows (what matters to them, the outcomes they are prepared to accept).

We need to change attitudes to reflect this, so that patients are not seen as “non-compliant” if they have a different opinion and clinicians are not seen as overly paternalistic if they are listening to the patient’s preferences and considering this in a recommended treatment plan. Furthermore, we should emphasise that shared decision making is not confined to one patient and one clinician, during one consultation. The process is distributed across healthcare teams, and between patients and their families, all of whom will influence the process, especially for patients with long term conditions.38

Implementing shared decision making is challenging but possible. No one intervention will succeed in isolation. It requires interventions to support organisations, clinicians, and patients: a bundle of interventions working together holistically across all healthcare settings.

Key messages

  • Shared decision making is about more than tools: skills trump tools, but attitudes trump skills

  • Successful implementation relies on a combination of interventions supporting the organisation, clinicians, and patients

  • Organisational support and local ownership are vital for engagement


  • This paper originated from the MAGIC programme, funded by the Health Foundation. This programme involved collaboration between the following organisations: Cardiff and Vale University Health Board, Cardiff University, the Newcastle upon Tyne Hospitals NHS Foundation Trust, and Newcastle University. The learning derived is based on the involvement and commitment of various clinical teams and our patient and public involvement panels. We are grateful to all of the MAGIC contributors, including Andrew Rix (independent chair of the Delivery Board of Health and Care research Wales, responsible for promoting public involvement in health and social care in Wales), the organisational representatives, clinical teams, and patients for their contribution to the learning.

  • Contributors and sources: This paper originated from the learning from the Making Good Decisions in Collaboration (MAGIC) Programme, funded by the Health Foundation. NJW drafted the manuscript, and all authors worked collaboratively to contribute to the content, edit, and agree the final version. All authors were involved in the MAGIC programme and contributed to the learning that is described throughout this manuscript.

  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following interests. AE reports personal fees outside the submitted work, for the delivery of shared decision making training to clinicians in primary care organisations in England and Wales. He has also received royalties for editing Shared Decision Making in Healthcare: Achieving Evidence Based Patient Choice (Oxford University Press). GE reports personal fees outside the submitted work from Emmi Solutions, National Quality Forum, and Washington State Health Department. He has also received royalties for editing Shared Decision Making in Healthcare: Achieving Evidence Based Patient Choice (Oxford University Press) and Groups (Radcliffe Press). GE initiated and led the Option Grid Collaborative, which produces and publishes patient knowledge tools ( He has been a member of teams that have developed measures of shared decision making and care integration. These tools and measures are published and are available for use. For further information see DT reports personal fees outside the submitted work for the delivery of lectures on shared decision making to professionals, funded by TEVA, and from workshops and seminars on shared decision making to various organisations in the north east of England.

  • Provenance and peer review: Not commissioned; externally peer reviewed.

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See:


View Abstract