Intended for healthcare professionals

Feature Essay

Shared decision making: why the slow progress? An essay by Neal Maskrey

BMJ 2019; 367 doi: (Published 05 December 2019) Cite this as: BMJ 2019;367:l6762
  1. Neal Maskrey, honorary professor, evidence informed decision making
  1. Keele University, Staffordshire
  1. nealmaskrey{at}

After more than 40 years of research and policy endorsement, adoption of shared decision making into routine practice has been remarkably slow. Neal Maskrey blames a lack of focus on doctors’ broader communication skills

Shared decision making occurs when clinicians and patients work together to select tests, treatments, management, or support packages, based on clinical evidence and patients’ informed preferences. It involves providing evidence based information about options, outcomes, and uncertainties, together with counselling and a system for recording and implementing patients’ informed preferences.1

Shared decision making has its roots in the ethical principles underpinning clinical practice. In the UK it is included in the NHS Constitution and is a requirement of the doctors’ regulator the General Medical Council (GMC), and it is recommended as usual practice by NHS England and in guidelines by the National Institute for Health and Care Excellence.

Every year from 2009, the healthcare services regulator the Care Quality Commission has asked hospital inpatients whether they were involved as much as they wanted to be in decisions about their care and treatment.2 In 2018 just over 50% of 75 000 respondents answered, “Yes, definitely,” and the pace of any progress seems glacial: about 25% of patients agreed when the CQC asked whether doctors had talked in front of them as if they weren’t there.

Recent Canadian observations of paediatric ear, nose, and throat consultations showed that only half of items that were listed as elements of shared decision making in a validated 12 item scale were present in over 90% of consultations.3 Clinicians were not observed assessing the patient’s preferred approach to receiving information to assist in decision making.

And, given the need identified by the late geriatrician Kate Granger for healthcare professionals to say, “Hello, my name is . . .” at the start of a consultation,4 the problems in doctor-patient communication run much deeper than just in shared decision making.

The rise of good consultation skills

The development of modern models of the consultation came with the arrival of the audio cassette and the video camera. In Manchester in 1976 Pat Byrne and Barrie Long analysed 2500 audiotaped consultations.5 They found that doctors made little effort to tailor their consultations to individual patients’ needs. Doctors varied very little in what they said once they had made a diagnosis or decided on a treatment.

Less than a decade later, David Pendleton’s videotaped consultations energised the teaching of, and research into, consultation skills.6

Many others contributed. A study of patients’ motivations about their treatment for tuberculosis in the early 1970s led to the Health Beliefs model of consultation,7 which introduced exploring the patient’s ideas, concerns, and expectations.

By the early 1990s most undergraduate medical students had some time in their crowded curriculum allocated to communication or consultation skills, beyond simply “taking a medical history.”

We now know that good doctor-patient communication helps to regulate patients’ emotions, facilitates comprehension of medical information, and allows better identification of patients’ needs, perceptions, and expectations.8 Patients’ agreement with the doctor about the nature of treatment and the need for follow-up is strongly associated with their recovery: patients are more likely to share pertinent information for accurate diagnosis of their problems, follow advice, and adhere to the prescribed treatment.8

Studies have shown correlations between a sense of control and the ability to tolerate pain, recovery from illness, decreased tumour growth, and daily functioning. Better mental health has also been reported. Some studies have found a decrease in the length of hospital stay, lower costs, and fewer referrals.8

A more patient centred encounter results in better patient and doctor satisfaction, and satisfied patients are less likely to lodge complaints or claim malpractice. Good communication is better for doctors in terms of greater job satisfaction, less work related stress, and reduced burnout.8

Potential reasons for slow progress

So, why hasn’t shared decision making become ubiquitous? In the UK, a large implementation programme funded by the Health Foundation found five key challenges: “We do it already,” “We don’t have the right tools,” “Patients don’t want shared decision making,” “How can we measure it?” and “We have too many other demands and priorities.”9 It robustly tackled each of these but acknowledged that progress in shared decision making was slow and difficult.

The reasons may be conflicts intimately related to medicine’s current core values. In the past century science has rightly come to dominate the curriculum. Medicine’s scientific basis is indisputable and has brought many benefits, not least a major contribution to an increase of more than 40 years in life expectancy. But also inescapable is the fact that healthcare, by definition, is delivered within relationships. And relationships are formed and sustained by communication, not by science, evidence, or data.

Evidence based medicine (EBM) has formalised the scientific approach, and a famous 1996 BMJ editorial defined it as the incorporation of the best available evidence into the traditional consultation, using clinical expertise and the patient’s views and preferences.10 Since then, EBM has defined the methodology for constructing systematic reviews and guidelines and has produced them in large numbers. But EBM has still not tackled the challenge of how best to incorporate the evidence into individual consultations.

This isn’t surprising. EBM processes identify factors that increase the probability of a good outcome on average in a population. Such data can rarely indicate the best approach for any individual patient. Even with large, well conducted, randomised controlled trials it’s impossible to predict whether a patient would have avoided a bad outcome without the intervention, been one of the population subgroup that benefited from the intervention, or been in the group that had a bad outcome despite having the intervention.

Interpreting data in consultations is part of a doctor’s role. Yet many doctors don’t possess the skills needed to understand and then communicate the risks and benefits of diagnostic and management options.11 The patient decision aids developed to help with this are underused, and their relevance in consultations, where emotions dominate over logic and rationalism, is still not clear.

A lack of research and education

A huge imbalance exists between research on communication skills and research devoted to EBM. A PubMed search for “communication skills” yields 1672 academic papers published in the past 10 years. But the term “guideline” uncovers 227 123 papers, and “systematic review” finds 127 351 papers.

Current data on undergraduate medical curriculums are elusive, but a 2001 survey of 24 UK medical schools found that the number of teaching hours dedicated to communication skills ranged from 18 to 91 hours.12 Assuming that the entire taught curriculum occupies 30 hours a week, 40 weeks a year, for five years, only 0.3%-1.5% is allocated to teaching such skills. Undergraduate communication curriculums are extensive and almost certainly difficult to cover in such limited time.

Data on how postgraduate specialist training programmes cover communication skills and shared decision making are also lacking, but some specialties seem to focus much more on training and assessment than others. Such variation is unwarranted.

Shared decision making is a part of high level, complex consultation skills and not separate from them. It’s no wonder that shared decision making, usually mentioned as a standalone entity, is far from routine.

Different approaches worth trying

In general, when implementation approaches have been unsuccessful it can be helpful to look to complexity theory and social science approaches to try to understand why people act the way they do.13 If shared decision making were easily assimilated into routine practice the evidence on how to do it would be clear—but it isn’t.14 It now seems obvious: no simple solutions exist. Each consultation is unique and incredibly complex, taking place under pressure from emotions, time, and limited health system resources.

We need new approaches designed to improve communication and shared decision making that incorporate routine evaluation and timely reporting of findings. Here I suggest three to start with.

Assessing doctors’ communication

A decade ago, in response to concerns about prescribing errors, the GMC commissioned research to describe and quantify the problems and then supported the development of the prescribing safety assessment (PSA). All UK medical graduates now must pass this summative test of drug dosage calculations and safe prescription writing. The 2009 problem of prescribing errors seems analogous to the 2019 problem of shared decision making and consultation skills. The GMC, given its commitment to shared decision making, should consider developing and introducing, alongside the PSA, a national assessment of consultation skills for all UK medical graduates.

New models of consultation

The Calgary-Cambridge consultation model used most in undergraduate teaching has been honed and tested by experts over many years. But it comprises 78 items. Learning, understanding, and using detailed theoretical frameworks is necessary to develop expertise, but some truncation is needed for practical use. This has not explicitly occurred until recently, when I supported Richard Byng and Edmund Jack in developing their SHERPA (Sharing Evidence Routine for a Patient Centred Plan of Action) model.15 If shared decision making is synonymous with great consultation skills it’s preceded by shared agenda setting and shared understanding of the problems. The SHERPA model has just three items: share, link, and plan. Evaluation is under way.

Recent changes have been made to the Quality and Outcomes Framework, the pay-for-performance system for GPs in England. These encourage more individualised decision making: for example, by including the ability to record a preference, “The patient has chosen not to receive the intervention described in the indicator.” These changes are welcome but late in coming, and their impact is yet to be assessed.

Greater local responsibility

I was recently taken aback by a secondary care doctor who spoke to me after I’d given a teaching session covering cognitive aspects of decision making. He described some colleagues whose consultations he was desperate to observe and emulate, not only for their grasp of the evidence base and its relevance to each colleague but also for the emotional intelligence and consultation skills needed to interact successfully with the patient and the family.

But he described other colleagues whose consultations were so poorly done that he would often try to exclude himself from them, knowing that he would personally undertake a follow-up consultation to deal with the fallout. The CQC data indicate that such variability isn’t unusual.

So, who takes responsibility for improving shared decision making locally? Local quality improvement schemes seem to focus on the more easily measured technical outcomes of care rather than the complexity and uniqueness of patient centred consultation skills.

As Don Berwick, president emeritus and senior fellow of the US Institute for Healthcare Improvement and former NHS adviser, says, “Everyone in the NHS has two jobs: doing the work and improving it.” We work hard to improve the delivery of technically excellent care. But it would sometimes be good for us, as well as for our patients, if we took more time to consider which words (and silences) work best when we’re making decisions with, and for, our patients.


Neal Maskrey spent 20 years as a GP and another 20 years working out why decision making for individual patients is very different from decision making based on population averages. He was medical director of the National Prescribing Centre and programme director of the Medicines and Prescribing Centre at the National Institute for Health and Care Excellence. He now works as a freelance academic based at Keele University. He continues to teach and learn from many colleagues, including GPs and their GP registrars, having taught on MRCGP courses continuously since 1991. This essay is based on an invited presentation to the NICE Shared Decision Making Collaborative meeting held in London on 6 June 2019.


  • Competing interests: I have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; not externally peer reviewed.