Supporting shared decision-making for older people with multiple health and social care needs: a realist synthesis

The vanguard sites

In England, the Five Year Forward View8 set out new models of integrated care as part of a wider system change in the NHS; these reinforce the aim for ‘a more engaged relationship with patients, carers and citizens’ in order to promote well-being and prevent ill health. 8 Fifty vanguard sites are piloting integrated care with the aim of achieving personalised, well co-ordinated care for patients and addressing traditional divides between primary care, community services and hospitals. 8,53 Their remit is to find new ways to deliver care in line with the five models of care (Box 1) identified in the Five Year Forward View. 8 The models are underpinned by preparatory and exploratory work carried out in collaboration with National Voices. 54 Integrated Primary and Acute Care Systems (PACS) and multispeciality community provider (MCP) vanguards currently cover about 8% of England but are similar to the Sustainability and Transformation Plans (STPs) now being introduced across England. 55 STPs seek to implement the NHS Five Year Forward View by creating new accountable models of care provision. Some are using a MCP model; larger STPs are using the PACS model. 56 The MCP model involves creating integrated community-based multidisciplinary teams (MDTs), which are supported by other specialist disciplines and by care co-ordinators who help patients and families with multiple LTCs to access services in other health and social care settings. 56

Shared decision-making in an integrated care setting

Integrated care sites, such as vanguard sites, offer a particular challenge for SDM as decision-making has to be negotiated between, and communicated to, multiple HSCPs, as well as patients and their families. Decision-making becomes more complex for people with increasing frailty as the capacity to self-manage is affected by the cumulative effects of LTCs. The nature of decisions is complicated by resource availability, polypharmacy, decline in decision-making abilities and concordance, availability of support networks, suitability of treatment, safeguarding and the increased likelihood of depression. 39,57–59 The skills for sharing and discussing personal information with vulnerable patients and their families can be hard to embed in services.

Shared decision-making for older people with complex health and care needs

Older people with complex health and social care needs often rely on family members to negotiate access to care or to advocate for them. 5,60,61 Person-centred approaches and models, like the triangle of care,32 recognise the crucial role of families for quality and safety and encourage their routine involvement in decision-making, with agreement from the patient. 2,4 If the decision-making ability of an older person is significantly compromised, family members are able to act as proxies62,63 if they have legal power of attorney for health and welfare decisions, and they should also be consulted if not. Recently, discussion about continuity of care favours a coconstruction approach, with patients, families and professionals as active partners. 58,64,65 Although service providers recognise the contribution of family carers, this recognition does not always translate into their routine engagement in decision-making with or for older people with multiple, often interlinked, health conditions. 58

To develop an understanding of the realities of working in and across complex overlapping systems of care, it is necessary to synthesise evidence from diverse strands of research. 66,67 Similarly, the evidence demonstrating the effectiveness of interventions designed to promote SDM is drawn from a wide variety of research and practice. 68,69 Realist methodology allows the deconstruction of component theories underpinning different interventions and enables us to consider relevant contextual data to test our understanding of the applicability of different approaches for older people with multimorbidity. It also helps us to examine how SDM might achieve desired outcomes such as improvements in patient safety, clinical effectiveness, quality of life and patient experience7 within the context of integration, austerity measures, growing demand for health and social care and personalisation of care.

Scoping the literature

The scoping review of the literature sought to identify existing theories on how and why the involvement of patients and carers is thought to be important, how the existing theories are defined in the literature, how interventions on SDM are meant to work and on which outcomes and how SDM might work in professional settings. The starting point was systematic reviews of SDM and related topics (such as PCC). To identify relevant reviews we searched PubMed and The Cochrane Library using the following Medical Subject Heading (MeSH) terms: shared decision-making, patient participation, patient decision-making, decision support, decision aid, expert patient, proxy decision-making, collaborative care, co-construction, coproduction and minimally disruptive medicine. These terms were combined with methodological search terms for systematic reviews. In addition, we undertook keyword searches on Google Scholar (Google Inc., Mountain View, CA, USA) for both reviews and primary studies and looked for relevant papers published by key authors in the area, such as Elwyn and Légaré. Full details of the search strategy are provided in Appendix 1.

We identified 39 relevant reviews20,40,66–69,93–124 and 35 primary studies or discussion papers. 17,40,67,83,93–102,125–145 Information extracted from the reviews included study aims, types of participants, definition of SDM, outcomes measured, relevance to our target group and any theoretical assumptions made. The reviews allowed us to understand the range and approach of research on SDM but provided limited detail with which to develop our programme theory. Time limitations meant that we were unable to carry out detailed data extraction on all 35 primary studies in phase 1, but they were used to generate ideas about possible contexts and mechanisms.

Scoping interviews

We conducted face-to-face or telephone scoping interviews with 13 stakeholders (Table 3). In phase 1 the purpose of the consultation with stakeholders was to explore key assumptions about what needs to be in place for effective SDM within integrated care initiatives, identify relevant outcomes and clarify the focus and scope of the searches in phase 2. Interview participants were purposively sampled to include a range of programme stakeholders. 81 They were recruited as a result of the networks of the research team. Participants were given a copy of the study information sheet – this provided contact details for the research team and a consent form – which they were asked to read and sign. Ethics approval was obtained from the University of Hertfordshire Health and Human Sciences Ethics Committee with delegated authority (reference number CSK/SF/UH/02387). Interviews were guided by a topic guide, which was used to explore understanding and experiences of SDM, including what works well and what could be improved. Interview transcripts were read by one author (BR) and discussed with a second (FB). Data from the interviews were regularly reviewed (BR and FB) and the topic guide was adapted as our programme theory developed. 146 Interviews were audio-recorded and transcribed.

Initial programme theory

From the literature and from stakeholder transcripts, a series of seven explanatory accounts were built up that contained ‘if–then’ statements, which helped to specify context and mechanism. These ‘if–then’ statements were illustrated with supporting evidence from the interviews and literature. ‘If–then’ statements identify an intervention/activity linked to outcome(s) and contain references to contexts and mechanisms (although these may not be very explicit at this stage) and/or barriers and enablers (which can be both mechanism and context). 147 The ‘if–then’ statements provided a useful way of structuring our thinking. They also helped to focus the process of taking ideas and assumptions about how interventions work and testing them against the evidence that we found.

To refine the theory, the seven ‘if–then’ statements were discussed at a half-day workshop that was attended by eight members of the research team. To ensure transparency of approach and an audit trail, we transcribed recordings of group discussions and maintained structured field notes on suggestions and decision-making processes. Following this meeting the statements were revised and then presented to the Project Advisory Group (PAG) for further discussion. The PAG included experts in the field of older people’s health, primary care, patient involvement and realist methods. It also included members of the University of Hertfordshire Public Involvement in Research Group (PIRG) (experts by experience). The PAG suggested that some of the statements were too complex and needed to be disaggregated. It also introduced the concept of ‘fake versus real’ SDM, to indicate when SDM appeared to be conducted as a tick box exercise rather than as a real process for decision-making. As a result, the seven ‘if–then’ statements were reduced to five (Table 4).

Selection criteria

In phase 2 we undertook systematic searches of the evidence to test and develop the theories identified in phase 1. The main inclusion criteria were as follows:

• Community-dwelling older people with complex health and care needs, such as those with frailty, multimorbidity and/or dementia. The focus was on those aged ≥ 65 years, although for certain groups (e.g. minority ethnic groups and homeless people) younger participants (aged ≥ 55 years) could be included if the issues were similar.

• Older people with complex health needs living in their own homes, in sheltered housing or in extra care housing (where people are very likely to still have the decision-making capacity to participate in SDM).

• Studies of any intervention or strategy designed to promote the ongoing engagement of older people with complex health needs, and/or their family carers, in decision-making relating to their health or social care needs [e.g. decision aids, physician or patient coaching, education or training, personalised care planning (PCP) or joint goal-setting]. The focus was on complex decision-making and personal goals rather than studies focused on single issues (such as whether or not to have a hip replacement).

• Studies of interprofessional SDM in which at least two health-care professionals (HCPs) collaborated to achieve SDM with the patient and/or family carer either concurrently or sequentially. 144

• Studies that provided evidence relating to the implementation and uptake of interventions designed to promote SDM for older people with complex health needs.

Identification of studies

We used a range of search techniques to identify relevant English-language publications:

• electronic databases such as PubMed, Scopus and The Cochrane Library [including Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Database of Systematic Reviews (CDSR), Database of Abstracts of Reviews of Effects (DARE) and Health Technology Assessment (HTA) database]

• checking of reference lists from primary studies and systematic reviews

• citation searches using the ‘cited by’ option in Scopus and Google Scholar and the ‘related articles’ option in PubMed (‘lateral searching’).

The searches were designed to reflect the five theory areas identified in phase 1 (see Table 4). For example, preparing for the SDM encounter (‘if–then’ statement 2) included searches around coaching, education and training and SDM for hard-to-engage groups. The latter (SDM for hard-to-engage groups) was built in because clinical members of the team (supported by stakeholder interviews) highlighted the difficulty that many people may have in engaging in SDM (e.g. those with chronic conditions and depression). Searches were conducted in December 2016. Two searches, interprofessional and reducing the workload, had date limits applied to the searches because literature in the area is relatively recent. Date limits and search terms used in PubMed are provided in Table 5. For full details of the searches see Appendix 1.

Screening and data extraction

Electronic search results were downloaded into bibliographic software and, when identified, duplicates were deleted. Documents from other sources were manually recorded in the same file. Two reviewers independently screened titles and abstracts for relevance. Full manuscripts of all potentially relevant citations were obtained and downloaded into Mendeley reference management and PDF (Portable Document Format) organisation software (Elsevier, Amsterdam, The Netherlands); all members of the research team had shared access to these manuscripts. Two reviewers screened full manuscripts for inclusion based on the relevance and rigour of the evidence, with disagreements resolved by discussion with a third team member.

Decisions on inclusion made at different points in time were recorded in a Microsoft Excel^® (Windows^® 10) spreadsheet (Microsoft Corporation, Redmond, WA, USA). Consistent with a realist synthesis approach, items were assessed for inclusion on the basis of whether or not they were considered ‘good enough and relevant enough’. 150,151 This was an iterative process that involved discussion between team members. ‘Good enough’ was based on the quality of evidence, for example was it of a sufficient standard for the type of research and were the claims made considered to be trustworthy. ‘Relevance’ related to whether or not the authors provided sufficient descriptive detail and/or theoretical discussion to contribute to the theories generated in phase 1. Poorly executed studies (low rigour) could still be included if the study was considered to contribute to understanding about how a programme was thought to work (high relevance).

A bespoke data extraction form was developed, piloted on five records and further refined as necessary. Once the final fields for data extraction were agreed, an electronic version was created in Microsoft Access^® (Microsoft Corporation, Redmond, WA, USA). The data extraction form included fields relating to study aims, design and methods; the types of participants (e.g. older people, people with LTCs, HSCPs); outcomes; information relevant to the theory areas; and emerging context–mechanism–outcome (CMO) configurations. Data were extracted by one reviewer, with 20% of the manuscripts checked by a second reviewer. PDFs in Mendeley were also annotated and relevant sections highlighted. Data in a realist sense are not just restricted to the study results or outcomes measured but also include author explanations and discussions, which can provide a rich source of ‘data’ that make explicit how an intervention was thought to work or not.

Synthesis

The Access database was used to identify prominent recurrent patterns of contexts and outcomes (demi-regularities) in the data and the possible means (mechanisms) by which they occurred. 152 This process enabled iteration from plausible explanations to the uncovering of potential CMO configurations. An example of the development of one of the original ‘if–then’ statements into a preliminary CMO is given in Table 6. In total, this CMO went through six iterations before the final programme theory was developed. Although all of the theories were developed and refined as the review progressed, some, such as that presented in Table 6, remained similar to the original theory developed in phase 1. In contrast, statement 5 (reducing the workload) was not taken forward as a CMO. This was largely because it was not supported by the emerging evidence but also because it was felt to be too conceptually different from SDM.

Data synthesis involved individual reflection and team discussion and was designed to (1) question the integrity of each theory, (2) adjudicate between competing theories, (3) consider the same theory in different settings and (4) compare the stated theory with practice experiences. Data from the studies or other evidence were then used to confirm, refute or refine the candidate theories. When theories failed to explain the data, alternative theories were sought.

Once the preliminary mapping of the evidence into tables was complete, the research team held a second workshop. This was attended by six members of the Research Management Group and one public and patient involvement (PPI) representative. This workshop was structured to include in-depth discussion of the findings and to develop and confirm the resultant hypotheses. As at the first workshop, we transcribed recordings of group discussions and took notes. This provided evidence of the reasoning processes and justification of inferential shifts. 74 Following the workshop the CMOs were revised (by FB, BR and CG) and circulated to all members of the research team for comments.

Understanding the needs and priorities of service users

Many studies recognise the importance of considering patients’ and, when appropriate, family carers’ preferences and values in the decision-making process,63,105,113,119,203 not least because ‘decision-making is grounded in the personal and social context of an individual’s life’. 105 Despite this, the evidence shows that individual needs and circumstances of patients and their family carers are frequently not understood or taken into account. 107,157,179 In a systematic review observing the extent to which HCPs involve patients in decision-making across a range of clinical contexts, the least observed behaviours were those that involved enquiring about the patients’ preferences. 107 This implies that interventions aimed at improving the tailoring of care to patient preferences are needed. This is supported by the Making Good Decisions In Collaboration (MAGIC) study,177 a 3-year project to test and identify the best ways to embed SDM into routine primary and secondary care using quality improvement methods. In their training sessions, the researchers found that ‘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’177 (risk is discussed further in CMO 2). An observational study of people with cancer and multimorbidities suggests that it is more important for HCPs to consider values that are stable over time rather than goals and preferences that are more context or circumstance specific. 63 This of course assumes an ongoing and sustained relationship between service user/patient and provider.

Information-sharing difficulties and goal divergence may arise for any number of reasons. These include HSCPs having difficulty identifying and explaining patient preferences,68,104 differences in the way patients and clinicians interpret and frame the patient’s health problems,189 clinicians being reluctant to engage in SDM when the patient’s preferences are not in line with clinical guidelines185 or when there are concerns about safety or cognitive function. 179

There are also system-level barriers. A qualitative study in primary care in the UK investigated gaps between lay and medical discourse in care planning. 189 It suggests that electronic records ‘can marginalise aspects of care which lie beyond a biomedical focus or contractual requirements’. 189 Patients’ narratives could be misunderstood or reframed when they were formally recorded. A focus on a biomedical agenda was also identified in a qualitative study exploring the extent of PCC and SDM with adolescents with type 1 diabetes mellitus. 175 It found that when patients tried to raise themes or concerns of their own, they were often interrupted by professionals who were trying to restore a biomedical agenda. 175 The authors argue that health professionals concentrated on educating patients about biomedically optimal self-care rather than considering the patient’s preferences and how these might affect adherence.

Several studies explored interventions aimed at helping to identify patient and carer preferences. A qualitative study from the UK reported on the development of a primary care-based complex intervention for patients with multimorbidity. 181 The intervention involved establishing and maintaining therapeutic relationships with patients and a person-centred approach. Patients appreciated the time, being able to ask questions, being listened to and having their goals acknowledged. The intervention is due to be tested in a cluster RCT. A RCT conducted in the Netherlands155 evaluated the impact of proactive care planning and goal-orientated integrated care. Although GPs reported some improvements (e.g. a perception that there were fewer unexpected demands on care) there was no significant change in patient quality of life or functional status and only a small improvement in satisfaction levels. A pilot study tested the feasibility of a conversation tool for outcome prioritisation in medication reviews with older people. 192 The tool increased satisfaction with medication use and the authors contend that knowing the individual’s preferences appears to deepen the patient–doctor relationship. However, the mean consultation duration was 31 minutes, which, the authors caution, does not fit with routine medication reviews in family practice.

Developing relationships

Achieving SDM depends on building a good relationship in the clinical encounter. 17,113,182,194 The ability to build and maintain trusting relationships between HSCPs and patients/carers was seen as beneficial in the implementation of collaborative approaches by staff157 and patients/service users165,171,178,181,183,188 and was found to have an impact on patient and carer perceptions of the quality of care. 109,162,164 A systematic review109 found evidence of a positive association between the quality of clinician–patient communications and adherence to medical treatments. For example, the odds of patient adherence were higher when physicians had received communication training. 109 Longer consultations and physician verbal behaviour (such as exploring the impact of the condition or illness on the patient) have been shown to be associated with increased trust. 210

Trust is also facilitated by continuity of care. 68,113,174 However, a long-standing relationship with a clinician may lead a patient to assume that the clinician is aware of his or her values,159 whereas evidence suggests that there is often a mismatch between patients’ and clinicians’ views. For example, a qualitative study179 investigated alignment of goals between people with multimorbidity, their family carers and HCPs, and found that, although there were some common goals (such as maintaining independence), there was also a significant amount of goal divergence. This was particularly the case when patients had significant illness complexity, such as unstable/fluctuating health problems or cognitive decline, which posed immediate or anticipated threats to the safety of the patient or carer/caregiver. 179

The importance of ongoing relationships and the ability to reassess changing priorities were highlighted in several studies. They were particularly important for people with complex needs or dementia as ‘the dominant chronic illness shifts over time as conditions and treatments change, and re-prioritisation occurs’;105 decision-making responsibility may shift over time, from the person with dementia to the informal or family carer/caregiver. 172 A review to develop a communication model to enable greater understanding of patient-centred communication and decision-making says that communication about complex medical issues often occurs as ‘a series of conversations over time, with multiple clinicians involved.’182 In one review the authors suggested that a failure to identify an association between empirical measures of SDM and health outcomes is because most studies focus on a one-off encounter. 68 They added:

. . . one discussion between a clinician and patient may not lead to improved health outcomes. Instead, a long-standing relationship between a clinician and patient marked by patient-centred care and SDM may affect outcomes over time.

Shay et al. 68

Interprofessional working

Trust and partnership working between different HCPs were described as key to decision-making for older adults with complex needs. 156,161,172,178 Facilitators of relational coproduction and interprofessional working (see Table 1 for definitions) include familiarity with one another and a history of working together, mutual knowledge and understanding of disciplinary roles, trust and respect, a shared understanding of SDM and effective communication between individuals (including different HCPs and patients and carers). 84,162,199 However, few studies addressed an interprofessional approach to SDM, with most studies targeting a single professional group. 144 A systematic review118 of interventions to improve HCPs’ adoption of SDM found that only three studies targeted more than one HCP.

Légaré et al. 84,93,144 have published several qualitative or descriptive studies exploring interprofessionalism and SDM. They created a model to help health professionals envisage a common goal and enhance the contribution of different health professionals to SDM. 84,144 The model ‘factors in’ determinants at the individual or micro level such as patient priorities, and at the meso and macro health-care system levels. Meso-level considerations include the influence of individual team members’ professional roles and macro-level factors encompass health policies and social context. A decision coach who is trained to support the patient’s involvement in decision-making is seen as key to the process. They argued that by assigning particular tasks to specific members of the team the process can become more efficient. A UK initiative ‘The Year of Care’ also involved team members working together to facilitate decision-making. 25 The initiative (which was piloted in over 50 practices in England) first focused on people with diabetes mellitus. Care planning involved an initial consultation with a health-care assistant (HCA) to work out what the patient wanted to do and undertake weights and measurements; there was then a second consultation with a GP or specialist nurse to make a care plan. They reported improved patient experience and biomedical outcomes, although they suggested that it may take 3–5 years for improvements to biomedical outcomes to become apparent. 25

Several reports describe UK initiatives to promote interprofessional working and integrated care. For example, a case study of structured medication reviews in care homes, involving residents, family members, care home staff and HCPs provided qualitative evidence of the benefits of an interprofessional approach to decision-making:198

It used to be just the doctor and the pharmacist talking to each other. But actually having everybody round the table, it’s just wonderful. It’s a game changer in terms of the decisions you make.

Pharmacist

However, it was not always easy for GPs to attend these meetings. Evaluation of holistic primary care for people with multimorbidities reported an increase in patient satisfaction and engagement and fewer unplanned attendances at the practice (although how this was measured is not clear). The initiative involved cross-discipline training and the involvement of all clinical staff and the patient and their family in producing, monitoring and updating a care plan to focus on quality of life for the patient. 200

Patient/service user outcomes

In CMO 1 we suggested that resources such as a person-centred approach, continuity, good relationships between service users and providers and interprofessional collaboration in SDM lead to benefits such as improved adherence and greater patient satisfaction. A systematic review110 including 43 RCTs found that interventions to promote a patient-centred approach in clinical consultations generally had a positive effect on a range of measures, although the impact on satisfaction, behaviour and health status was mixed. Another review focused on PCP. 66 It reported improvements in some indicators of physical and psychological health status, and people’s capacity to self-manage their condition when in comparison with usual care. The impact was greater when the intervention was more comprehensive, intensive and better integrated into routine care. 66

Systematic reviews of SDM have identified positive benefits to patients and carers such as feeling more involved,117,119 an improved quality of life and reduced depression in carers,119 and improved affective cognitive outcomes for patients, such as enhanced satisfaction and reduced decisional conflict. 68 These impacts (particularly on patient/service user satisfaction) are echoed in many of the other studies we accessed. 109 There is also some evidence that SDM leads to better treatment adherence. 197 There is little evidence, however, to suggest that there is an association between empirical measures of SDM and health outcomes. 68

Context–mechanism–outcome 1: summary

The evidence shows how systems that enable HCPs to develop relationships with patients/service users and their family carers trigger feelings of trust, engagement and respect that can lead to improved outcomes such as patient and carer satisfaction with services and decisions. The quality of individual clinicians’ communication skills, and their ability to foster trusting relationships with older people and their families, is fundamental to SDM. In addition, there is also a need for systems that promote continuity of care both through ongoing relationships with one clinician (relationship continuity) and through system-based approaches that develops ways of working whereby the patient is linked to multiple professionals (management and informational continuity). SDM with older people with complex needs is likely to increase appointment length. Although it is thought to improve adherence to treatment regimens, there is currently little evidence to suggest a link to health outcomes or service use.

Organisation support

The MAGIC study177 supported the link between organisational buy-in (e.g. identifying SDM as an organisational priority) and an increase in HCP engagement with, and prioritisation of, SDM. It also affects patients’ perceptions that the health-care organisation and clinicians want them to be involved. 177 However, although SDM is a core part of policy in many countries, including the UK,211 at a service level, systems are not in place to incentivise or appropriately reward patient-centred practices and SDM. 17,177

Systems to support shared decision-making

In the UK, the Year of Care initiative aimed to improve care for people with LTCs. A report from one practice in North East England31 described the implementation of this initiative for people with multiple LTCs (the ages of participants were not specified). Although the report does not provide data on patient outcomes, the staff felt that it improved the quality of conversations with patients. The report documented key stages in the SDM process, such as preparatory work before the actual consultation with the GP, when relevant information about the patient is gathered by other members of the primary care team. Appointments are longer than normal, with the length of the appointment being adapted to reflect patient needs. In addition, all chronic disease monitoring is combined into one annual review, which involves a collaborative consultation, based on SDM and self-management support, via care planning. 31,200 A qualitative study on information sharing also supports the idea of looking at health as a whole rather than in a series of reviews focusing on specific chronic diseases. 189

In the Year of Care approach adopted by one practice in the UK,200 patients are sent relevant test results before the consultation. This is an increasingly common practice in the UK. We found no formal evaluations of this approach. Although a report from a GP practice in the UK said that despite concerns that such an approach would increase patient anxiety and generate more work for clinic staff, this has not been the case. It stated that ‘people are keen to have their own information and value the recognition that gives about their involvement’. 200

Managing risk and uncertainty

Several studies highlighted the way that perceptions of risk or uncertainty affect decision-making. A paper on collaborative decision-making concluded that the uncertainty that complicates medical decisions is rarely explicitly addressed in decision support tools or medical consultations. 182 One paper discussed the importance of addressing uncertainty in SDM and proposed an uncertainty toolbox that includes the principles of honesty, recognition of emotion, hope, support/co-ordination of care, willingness to readdress issues, respect for personal decisions and the acceptance that a lack of decision is possible. 202 A cluster RCT tested a SDM training programme for newly qualified GPs in the UK. The training involved two aspects: (1) SDM training and (2) training in risk communication. The authors stated that doctors’ confidence and satisfaction with the process were greater if they received the risk communication training before the SDM training. 167 However, the risk communication training was based on four discrete conditions and the study excluded people aged ≥ 75 years.

Training

The need for enhanced communication skills for clinicians was a common theme across the papers. 66,107,110,113,168 For example, a review of qualitative studies suggested that GPs need better communication skills, particularly when working with patients with complex needs. 68 A number of the included studies evaluated the impact of training for HCPs on SDM. A review of interventions for providers to promote a patient-centred approach (of which SDM was seen as an important component) in clinical consultations found ‘generally positive effects on a range of measures relating to clarifying patients’ concerns and beliefs; communicating about treatment options; levels of empathy; and patients’ perception of providers’ attentiveness to them and their concerns as well as their diseases’. 110 Short training (< 10 hours) was as successful as longer training. 110 Another review evaluating interventions to promote SDM found that interventions targeted at HCPs did lead to an improvement in SDM. Interventions that targeted professionals as well as patients were considered more effective than interventions that targeted only one of these groups. 118

A cluster RCT of a SDM training programme for newly qualified GPs in the UK found that, although training increased GPs’ SDM skills (measured on the option scale), there were no statistically significant changes in patient-focused outcomes. 167 The length of the consultation and being seen in a protected environment were linked to greater patient confidence in the decision and expectation to adhere to their chosen treatments. A more recent RCT of SDM training for GPs found that, although trained GPs exhibited less paternalistic decision-making, they were not judged (according to the control preference scale) to have engaged in SDM. 185 The study did not consider the impact on particular subgroups of patients such as older people or those with complex health needs.

In a discussion piece, Elwyn et al. 17 argued that the best way for clinicians to learn SDM skills is using simulations. This was supported by evidence from the MAGIC study in the UK. They reported that interactive skills training workshops based on a SDM model helped build coherence, improving skills and promoting positive attitudes. They added that role play-based training, which emphasised practical skills, worked better than theory-heavy presentations. It was also considered important that clinical teams were able to develop a shared understanding of how SDM might differ from their current practice. 177 The impact on patients was unclear as the authors reported that ‘we found it difficult to identify or develop suitable patient reported measures to capture experience of shared decision making’. 177

Context–mechanism–outcome 2: summary

There is evidence that organisational support, appropriate training for HCPs and system-based aspects, such as longer appointments, lead to HCPs feeling more supported and having the confidence to engage with SDM. It can also result in increased patient satisfaction with decision-making because they feel that service providers are attentive to them and their concerns. There is a lack of studies addressing interprofessional approaches to SDM or the training needs of providers other than doctors.

Decision aids

Much of the literature we found on preparing patients and carers relates to the use of PDAs (see Appendix 2 for more details of these studies). PDAs are tools designed to help people participate in decision-making about health-care options. They provide information on reasonable health-care options and are designed to help patients clarify and communicate the personal value they associate with different features of the options. 212 They differ from health education information because they focus specifically on the decision to be made. 20 Systematic reviews provide good evidence that PDAs can have a positive impact on patient knowledge, decisional conflict, informed choice, participation in SDM and decision self-efficacy,20,69,103,108,116,121 including for those who are socially disadvantaged. 69 Potential mechanisms relating to the likely benefits of decision aids include patients becoming more engaged,69 greater decisional self-efficacy,69 a greater feeling for patients that they are involved in decisions,167 and that ‘being offered the opportunity to prepare for a meeting is seen as an indicator of mutuality’. 171 However, those same systematic reviews provide little evidence that decision aids improve health outcomes or patient adherence.

Older age, depressive symptoms and difficulties with the activities of daily living are associated with decreased patient activation;170 interventions may therefore need to be tailored to older adults’ level of patient activation and address issues such as depression. 170 A review of decision aids for older people reported that they improve older adults’ knowledge, increase their risk perception, decrease decisional conflict and seem to enhance participation in SDM. 121 These findings were supported by a patient-level meta-analysis that looked at the impact of characteristics such as age on the efficacy of decision aids. 108 They found no difference, however, in concordance with chosen treatment. The authors highlighted several limitations of the current literature in this area; they found only one study108 in which the tool had specifically been designed for older people, most participants in the study were younger older people (≤ 70) rather than the oldest old (≥ 80) and tools were not tailored to the needs of people with multimorbidity. The authors argue that:

. . . it is doubtful whether it will ever be possible to provide relevant and evidence based information tailored to all possible combinations of comorbidities in a single decision aid. This indicates that decision aids for older adults might be most effective when combined with high quality patient-provider interaction during consultation, with personal tailoring of the decision aid to the individual persons’ context.

van Weert et al. 121

In contrast, most of the tools included in the review were designed to be used by the patient at home or in a research interview rather than as part of a routine consultation with a clinician.

Although decision aids are seen to have advantages, a number of studies also outlined potential problems with their use. In the MAGIC study177 the authors noted that a key finding of their programme was that ‘skills trump tools, and attitudes trump skills’. They argued that:

. . . 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.

Joseph-Williams et al. 177

They concluded that short tools to use in the consultation are better than information sources used outside the consultation because they are better at facilitating discussion between the patient and clinician and are cheaper and easier to keep updated than an information source, such as a website. They cautioned, however, that there is a risk that brief decision aids may be used purely to enhance information transfer, rather than improving the way clinicians work with patients.

Another criticism of decision aids is that they do not address the essential first step of SDM, which is preparing for the encounter. 116 This includes perceiving the opportunity and personal ability to be involved, both of which were very important in our programme theory. A qualitative study with people with mental health problems suggested that ‘concrete aids for considering and contributing to the preparation of the decision-making occasion might reduce power differentials’. 171 However, in a systematic review of barriers and facilitators of SDM, Joseph-Williams et al. 116 argued that there is a need ‘to address the entry level factors to SDM, such as subjective norms and patients’ roles, before secondary process factors such as information provision and value clarification’. 116 In addition, several studies proposed that the provision of decision aids will not necessarily mean that they are used, or, if they are used, that they will lead to SDM. 114,166

Coaching

Another proposed way to help people in SDM is via coaching or facilitation. Studies on patients with cancer have found that when trained facilitators or navigators met with patients prior to their consultation with a clinician, there was improved self-efficacy and reduced decisional conflict. 153,173 However, one was a feasibility study173 and in the other rigour was low. 153 A systematic review of coaching or guidance as part of SDM concluded that the evidence supports the use of coaching or guidance to better guide patients in the process of thinking about a decision and in communicating their values/preferences with others. The mechanisms inferred from the paper are that if you improve patients’ deliberation and communication skills, this will lead to empowerment and thus patients will feel better supported. But the impact on other outcomes, such as participation in decision-making or satisfaction with the option chosen is mixed. 120 Moreover, the relevance of this review to our synthesis is only moderate as the studies mostly focused on single issues, such as screening.

Family-centred approaches

Research has underscored the importance of family-centred approaches for older people with complex needs. 58,213 However, similar to a previous realist review on engaging older adults in health-care decision-making,113 we found few studies that considered the involvement of family members and friends in SDM. A German pilot study reported on the development of a tailored e-health service for caregivers/carers of people living with dementia and HSCPs. 186,187 The aim of the tool was to facilitate communication between HCPs and informal carers and to empower the carers. Although not aimed at SDM, the intervention included an element related to preparation for doctors’ visits. The study suggested that such a tool might help informal carers to feel more prepared for doctors’ visits (e.g. identifying questions).

One initiative to make it easier for patients and their families to engage in evidence-based decision-making is ‘the three questions’ approach (AskShareKnow). 214 Patients are encouraged to ask the following: what are my options? What are the benefits and harms? And how likely are these?214 These questions have been shown to improve the information given by family physicians and increase physician facilitation of patient involvement. The authors considered the approach to be a potentially powerful intervention for affecting physician behaviour without affecting the length of the consultation. 214 It is possible that the approach may be less successful with adults with lower-than-general levels of literacy. 215

Permission/space to discuss options

Key to CMO 3 is that SDM is undertaken in a context in which HSCPs, patients and their families are able to discuss the value and effectiveness of proposed treatments and to set the agenda. A positive example of this comes from the Year of Care initiative, which described patients feeling free to ask questions, meaning that they feel that HCPs are interested in them as people and not just in the condition. The report added that the agenda-setting prompt had given patients permission to talk about things. 31 This ties in with a key part of our CMO: that interventions or programmes need to create a culture that enables people to ask questions without feeling judged. Longer consultations are linked to greater patient satisfaction and improved SDM,31,107,157,165,181,183,192 which is likely to be related to the opportunity for patients to ask questions and ‘feel listened to’. 165,181 Patients also want to feel respected. 178,188

However, there is evidence to suggest that clinicians’ attitudes act as a key facilitator or barrier to SDM. 116,177 A study of older people’s experiences of health-care decision-making found that participants often felt unable to make their needs heard because clinicians rushed them or ‘closed them down’. 159 A qualitative study of information exchange in consultations found that patients and HSCPs may not be able to reach a shared understanding because HCPs omitted relevant information or because patients often omitted relevant context or preferences during a consultation. 158 Older people may be reluctant to voice their concerns or openly disagree with a doctor but instead avoid conflict by choosing not to adhere to prescribed treatment. 159

Understanding that there is a choice

Studies have illustrated that the presentation of choices can substantially influence the decisions people make. For example, framing risk in survival rather than mortality terms increases the probability that patients will consent to an intervention. 207 Moreover, older people may not always be aware that there is a choice to be made. 159 A qualitative study with older people receiving dialysis found that patients were often unaware that dialysis initiation was voluntary and held mistaken beliefs about their prognosis. 180 Clinicians may find it difficult to present choices to patients because they are trained to display confidence to patients. Their own discomfort with uncertainty might lead them to engage in more paternalistic styles of decision communication. 182

In some instances, the choice being made may be constrained by resource limitations, health-care policies or evidence-based recommendations (such as guidelines). A qualitative systematic review of decision-making communication practices found that even when there is no alternative, exploring patients’ choices and reasons can validate the patients’ participation and mean that they have still ‘participated in the decision making process’. 117 They added that pursuing agreement without engaging with patients’ reasoning may be treated as coercive. This was echoed in a knowledge synthesis182 of medical decision making which concluded that:

Achieving shared mind will be more difficult when patient preferences are counter to the strongest clinical evidence. Under these circumstances, quality decision making will depend not only on clinician(s) and patient becoming aware of and understanding the other’s position, but also engaging in a dialogue to find common ground, identify and reconcile differences, and negotiate or compromise to reach a collaborative, agreed-upon decision.

Politi and Street182

The subject of coercion was also explored in a discussion paper that considered whether or not ‘clinical nudges’ are ever compatible with SDM. The authors argued that in instances in which patients do not have authentic preferences, clinicians are justified in using nudges in accordance with the best interest standard (e.g. the ethical requirement that people who care for others will do so in good faith, placing their assessment of that person’s best interests above their own). 207

Context–mechanism–outcome 3: summary

Older people with complex health needs are likely to need support to participate in SDM. Although the evidence suggested that interventions, such as decision aids and coaching, can improve involvement in SDM, the impact on adherence or health outcomes is not proven. Moreover, most tools were not designed (and have not been tested) for the oldest old. Evidence suggests that if tools are used they need to be brief, designed for use within a consultation and focused on facilitating discussion between the patient, family carer and HSCP. The right culture, that allows people time to ask questions and to discuss options, and staff with positive attitudes towards SDM are likely to be more important than tools for older people with complex health and care needs.

Time and resources

The programme theory outlined in CMOs 1–4 puts forward many barriers to SDM and it is clear that relying on individual clinicians or patients to implement SDM without system-based support is unlikely to be successful or sustainable. 114,116,118 Several of the papers we included have described system-based changes that involve person-centred, integrated approaches to people with LTCs,25,31,164,200 of which SDM is an integral part. A report describing the implementation of the Year of Care initiative in the UK suggested that valuing the development of staff is as important as valuing the expertise and lived experience of the patient. 31 This initiative reported increased staff and patient satisfaction. 25,31,200 One site reported that the quality outcomes framework (QOF) figures had not been negatively affected by the new ways of working. 200 Overall, it is suggested that similar initiatives (data not given) will take 2 or 3 years to make a difference to clinical outcomes as ‘habits of both patients and professionals die hard and engagement increases over a number of care planning cycles’. 200 This is reflected in our programme theory, which argues that familiarity and a shared expectation of new ways of working (which include SDM) are crucial, but these are likely to take time to develop.

Patient activation or engagement

The willingness or ability of patients to participate in SDM is a key contextual factor in our programme theory (see CMO 3). This is supported by the literature107 and also underscored by our interviews with stakeholders. In general, the consensus is that the majority of older people would wish to be involved in decision-making104 but in practice they are often not encouraged to participate in SDM. 104 Reasons for this include limited time, poor continuity of care, environmental conditions, organisational inertia, a biomedical focus, concern about disruption to routines, clinicians’ belief that they are already practising SDM and power imbalances. 114,116,168,202,216 Patients may feel unable rather than unwilling to share in decision-making. 116,177 Joseph-Williams et al. 177 suggested that, for older patients, their previous experiences are likely to give them an expectation of a paternalistic approach from clinicians. Although many SDM initiatives involve giving patients more information, this alone is not enough; patients need knowledge and power to participate in SDM. 102,116 A systematic review of patient-reported barriers to, and facilitators of, SDM suggested that power may be linked to perceptions of permission to participate in decision-making, perceived influence on decision-making, confidence in own knowledge and self-efficacy in SDM. 116

Outcomes

A systematic review of tools to promote SDM in serious illness suggested that decision aids have the potential to reduce health-care intensity and costs by decreasing unwanted major high-cost interventions or hospitalisations but that these outcomes have not been studied. 103 Another review suggested that as a result of SDM, patients tend to choose more conservative options. 20 One particular area in which there has been overlap between SDM and care for older adults with complex needs (which has a potential impact on resource use and patient outcomes) is around medication and, in particular, deprescribing. Deprescribing has been defined as ‘the process of withdrawal of an inappropriate medication, supervised by a health care professional with the goal of managing polypharmacy and improving outcomes’. 217 Inappropriate medication would include that which does not fit with the goals of treatment, does not align with patient values or is overly burdensome. 218 Although this was not the focus of our review, we did consider several studies that focused on deprescribing in a context of SDM. A project in care homes in the UK investigated structured pharmacy reviews for older adults. 198 The reviews, which involved the pharmacist, patient, family carer, GP and care home staff, significantly reduced unnecessary prescribing across care homes and reduced the risks of harm to residents caused by medication.

Context–mechanism–outcome 4: summary

We suggest that SDM is likely to be most effective when it involves service providers who have the right skills, attitudes and tools, working in systems that are structured to support service providers and users to engage in SDM. Key to this is a culture that involves person- (and family-) centred approaches. Although a shared expectation of, and familiarity with, the process is essential to its success, this is likely to take time to develop. The evidence indicates that such approaches may lead to improved service user and provider satisfaction with services and with the quality of decisions, but there is little research on the impact on health-related outcomes.

Context

The CMOs highlight several contextual factors that are likely to affect the success or otherwise of interventions or programmes to facilitate SDM with older people with complex health and care needs. These operate at individual, team and broader organisational levels (Figure 8):

• At the macro or broader organisational level, a key context is to what extent systems legitimise and prioritise SDM for people with complex needs. Evidence from this review suggests that, for SDM to become embedded, there needs to be a level of organisational support that will trigger confidence in HSCPs that SDM is an organisational priority. 114,219

• Meso-level factors relate to the way the team is organised and the extent to which services support and resource professionals, patients and family carers to have the time and skills necessary to be able to engage in SDM.

• At the individual level, there are specific requirements needed of the workforce so that they can work in partnership with patients and family carers. For example, developmental needs around communication skills and how they define their role as a health or care professional working with and for older people over time. Maskrey and Gordon220 talk about the need for ‘a lifelong curriculum of learning’ in order to develop a shared understanding with patients.

FIGURE 8.

Summary of contextual factors at individual, team and system levels. IP, interprofessional working.

Mechanisms

A realist approach argues that exposing the resources and reasoning within mechanisms and their relationship to the context of their implementation is key to the evaluation of a complex programme of change. 221 Mechanisms are dynamic and may be interacting with each other;222 for example, in our programme theory, trust in HSCPs is likely to operate in parallel to patients and carers feeling heard and respected. In addition, what is seen as a mechanism or outcome, in one CMO, such as SDM becoming embedded, may, in turn, become a context in a subsequent CMO. 223

Our review suggests that interventions or programmes to promote SDM between HSCPs and older people with complex needs will have a greater chance of ‘working’, that is, engaging people in SDM and potentially improving satisfaction and health outcomes by paying attention to those activities or contexts that engender the following mechanisms:

• Trust between patients, family carers and HSCPs.

  The evidence60 suggests that relationship continuity (e.g. a continuous therapeutic relationship with one or more health professionals over time), is an important resource that engenders trust and the related mechanism of confidence in patients and family carers. This requires face-to-face contact to encourage the person to participate in SDM.

• Trust between HSCPs.

  Evidence84 shows that in order for HSCPs to take an interprofessional approach to SDM there needs to be trust and an understanding of each other’s roles.

• Confidence.

  Confidence is an important mechanism for HSCPs, patients and family carers. When HSCPs know that SDM is an organisational priority and that services are organised to facilitate person-centred approaches to care they are more likely to engage in SDM. For patients and carers, confidence might be engendered by interventions that increase their communication and deliberation skills (e.g. coaching, PDAs, longer appointments) and environments that mean that they are not afraid of ‘doing the wrong thing’ or being labelled difficult. 224 Confidence may then in turn become a context that enables people to persevere with SDM because they know that it is valued by those around them.

• Respect.

  Strategies to foster SDM are likely to be most effective when they are based on person-centred approaches that trigger a feeling in patients and family carers that they are respected, listened to and understood. 29,225 SDM is often conceptualised as an integral part of person-centred approaches to care;110 indeed, it has been described as the ‘pinnacle of person centred care’. 23 Respect among different HSCPs as well as between team members and the patient is also vital for interprofessional SDM. This includes a sharing of power among the team members so that all members can share knowledge and communicate effectively. 84

• Familiarity/shared expectation.

  We suggest that confidence is also related to all parties (HSCPs, patients, family carers) being familiar with the process of SDM and having a shared expectation of what this involves. Evidence (primarily from stakeholders) indicates that this may take time as both sets of actors have to assimilate the core assumptions of SDM and learn how to work together. 200

• Permission to participate.

  Programmes and strategies to promote SDM should also look to trigger a sense of permission so that all parties (patients, family carers and HSCPs) feel that they are able to participate in SDM. This may arise, for example, through knowing that SDM is an organisational priority.

Shared decision-making for older people with complex health and care needs

For those most reliant on health and social care support, such as older people living with frailty and/or severe disabilities arising from LTCs, decision-making may be particularly complex. It requires consideration of resource availability, the potential impacts of polypharmacy, the ability to obtain informed consent, the capacity of patients to attend to health-care demands, the support networks available, potential safeguarding issues and the appropriateness of individual treatments for people with multimorbidity. Initiatives to address these have focused on aspects of care such as interprofessional working,226 case management227,228 and the use of shared methods of assessment and planning (e.g. Comprehensive Geriatric Assessment). 229 All of these tend to be professionally initiated and directed. Although there is evidence230,231 that many older people and their family carers would like to be involved in decision-making, there is little evidence that relates specifically to SDM with older people with complex health needs. The overwhelming majority of the evidence on SDM relates to decisions about single conditions (e.g. prostate cancer or diabetes mellitus). It is more difficult to extrapolate this single condition-focused research to older people with multimorbidity. In addition, when older people are involved they are generally not the ‘oldest old’. 121 There is no consensus on how SDM should be structured or who should be involved, and few studies considered the involvement of family members in SDM when older people rely on them for ongoing support and care.

Person-centred approaches to health and care and consideration of the patient’s preferences and values is central to the SDM process. 29 For older people with complex needs, eliciting preferences is likely to involve regularly revisiting decisions as the dominant illness, and priorities, may shift over time. 105,172 However, the evidence suggests that doctors are better at recognising and discussing options than eliciting patient preferences (see CMO 1). This may reflect the fact that different HSCPs conceptualise PCC in different ways. 18,232 A review of literature on PCC suggests that, whereas the nursing literature tends to focus more on respecting patients’ values and beliefs in promoting PCC, the medical literature has devoted more attention to understanding the nature of the informed decision-making process between the doctor and the patient. 232 What is not explored in the literature is whether or not integrated care and interprofessional working might enable different members of the MDT to draw on the skills of others in order to promote effective person-centred approaches to SDM.

The quality of individual clinicians’ communication skills, and their ability to foster trusting relationships with older people and their families, is fundamental to SDM. In addition, there is also a need for systems that promote continuity of care. Continuity can be achieved through an ongoing relationship with one clinician (relationship continuity) or a system-based approach that develops ways of working whereby the patient is linked to multiple professionals (management and informational continuity). 60,233,234 The evidence would suggest that both need to be in place. Informational continuity is, however, often hindered by electronic systems not set up to record information relating to patient preferences and goals. For example, a study of record keeping in UK primary care found a gap between patients’ oral narratives and the way in which these are understood and recorded by clinicians. This suggests that universal adoption of electronic patient records in their present state will not alone achieve informational continuity between organisations. 189

Enabling older people to participate in shared decision-making

Context–mechanism–outcome 3 relates to the need to help prepare and support patients and their family carers to engage in SDM. This is likely to be particularly important for older people with complex needs, as older age, depression (which is often linked to LTCs) and difficulties with activities of daily living are associated with decreased patient activation. 170 The evidence highlights key contextual factors such as longer consultations, clinicians’ communication skills and a culture that allows people to ask questions without feeling judged. These are likely to lead to an increased ability and willingness to engage in decision-making through mechanisms such as feeling respected and understood.

Much of the literature on preparing patients and carers relates to the use of PDAs. Although there is some evidence that PDAs may enhance participation in SDM and decrease decisional conflict, very few studies include the oldest old, and tools are not designed for use with older people or those with multimorbidities. 121 Moreover, PDAs on their own are unlikely to address what Joseph-Williams et al. 116 calls the entry-level factors to SDM, such as subjective norms and patients’ roles. For older people, subjective norms235 may include the perception that there is a right and wrong decision, or that decisions should be made by doctors. HSCPs should actively consider the possibility that older people may be more hesitant than other age groups to express a wish that they feel is counter to that of the clinician,159 and that their attitude towards SDM may be affected by past experiences of a paternalistic approach from clinicians. 116

Patient decision aids for older people are likely to be most effective when used as part of face-to-face interactions, ongoing patient–clinician relationships and patient-centred approaches. 236 Hargraves and Montori237 argue that the real work of SDM is ‘collaborative deliberation in the face of uncertainty’237 which happens within the encounter between patients and doctors. Although this is a narrow conceptualisation of SDM, excluding family carers and other types of HSCPs, it does emphasise the importance of the conversation and the need to understand how factors such as decision aids, clinician skills, patient power and uncertain scientific evidence interact in the SDM encounter. 237

Shared decision-making and the new models of care

Many health-care systems, including the English NHS, are moving towards a more interprofessional health-care team-based approach. 8 In integrated care sites, decision-making and communication may need to be negotiated between, and communicated to, multiple HSCPs and personal assistants directly employed by care users, as well as patients and their family carers. Despite this, the vast majority of the SDM literature relates to encounters between a single clinician (usually a doctor) and a patient, and there is a lack of studies addressing interprofessional approaches to SDM. 118 In addition, when studies address the skills and attitudes needed for SDM, these focus almost exclusively on doctors. For interprofessional SDM to work, the development, and involvement, of all staff is important and power imbalances need to be addressed. 25 Doctors’ real and symbolic power is still strong and may be a barrier to interprofessional decision-making. 84

The literature does provide some examples of how decision-making might be negotiated between a patient and several clinicians. For example, a case study of medication reviews in care homes involved meetings between residents, family members, care home staff, a pharmacist and (when possible) a GP. 191 Another example of an interprofessional approach to care planning (which involves SDM) comes from the Year of Care initiative when care planning involved a first consultation with a HCA who helped the patient work out ‘what they want to do’ followed by a second consultation with a GP or specialist nurse to make a care plan. 25 Both reported positive outcomes.

Légaré et al. 84,144 have developed a model of interprofessional SDM, which incorporates the use of a decision coach who is trained to support the patient’s involvement in decision-making. They suggest that the decision coach could be a member of the team, such as a nurse or a therapist. A lack of time is a commonly cited barrier to SDM114,116 and the employment of members of the MDT as coaches could create more space and time for SDM. Although there is evidence to support the use of coaching to aid patients in the decision-making process, we found little evidence specific to older people with complex needs. In addition, the model proposed by Légaré et al. 84,144 does not appear to have been tested in practice; forward citation tracking from these papers suggests that the literature continues to be descriptive and explorative rather than describing the development and evaluation of interventions to promote interprofessional SDM. Although it is suggested that interprofessional approaches to SDM could improve quality of care by fostering continuity in the decision-making process, this has not yet been tested. 144

Shared decision-making and the reduction of unnecessary health services

There are several different paradigms underpinning SDM. 13 Imperatives behind SDM include policy drivers, moral arguments, the need to address uncertainty, the wish to deliver PCC and the potential for SDM to increase patient activation leading to improved self-management of LTCs and, in turn, improved health outcomes. There is also a suggestion that SDM can be used as a tool to reduce resource use or unnecessary treatment. 143 All of these assumptions are apparent in the literature although perhaps the strongest is that SDM (either on its own or as part of PCC) is morally right.

There was some evidence from the literature of a crossover between SDM and campaigns to reduce the use of harmful or unnecessary health services. For example, the Canadian initiative ‘choosing wisely’ includes in its framework the need for supporting SDM at the clinical level,238 and the promotion of better conversations between patients and physicians to facilitate informed decisions. 239 However, an evaluation of the ‘choosing wisely’ literature suggests that this does not meet criteria for SDM documentation because the patient materials presented imbalanced information and most did not provide an opportunity to clarify patient values and preferences. 145 It is also possible that if such approaches are seen by patients and family members as a way of rationing care or denying access to treatment, then it would be difficult to trigger the important mechanisms of trust and confidence needed for SDM.

Changing the culture

In CMO 4 we suggest that SDM needs to be situated in a context of wider changes in the culture and organisation of health-care systems. This CMO incorporates components from the previous CMOs, such as organisational resources (time and space), systems to support SDM and skills development through continuous practice development. These wider changes are important to trigger mechanisms such as familiarity and confidence. This familiarity, which develops over time, is necessary for both HCPs and patients and their families. It includes an understanding that patients and (when appropriate) their family carers have responsibility for their health and the decisions that affect them. Theories of organisational readiness for change in health-care settings suggest that, in order for change to occur, members of an organisation need to feel committed to implementing an organisational change and confident in their collective abilities to do so. 240,241 Such confidence is linked to an understanding of how to implement the change, having the available resources and current situational factors.

Clearly important to CMO 4 are person-centred approaches, and we suggest that desirable outcomes of a SDM approach would be that care is organised around the needs of the patient and family carer and that the patient’s priorities are incorporated into actions, notes and care plans. McCormack and McCane86 argue that PCC is about having a specific person-centred culture rather than PCC by itself, and that it should be practised by everyone in an organisation. He writes that ‘person-centredness can only happen if there is a person-centred culture in place in care settings that enables staff to experience person-centredness and work in a person-centred way’. 86 Person-centred SDM includes supporting patients so that ‘they feel capable of acquiring and understanding knowledge about the available options, and so that they value their personal knowledge contribution to SDM’. 116

Systems and culture

• The evidence suggests that SDM is only likely to become embedded if it is regarded as an essential component of ‘good’ healthcare and is linked to a culture of person-centred approaches throughout an organisation.

• The evidence suggests that a culture that allows people time to ask questions and to discuss options, and staff with positive attitudes towards SDM, are likely to be more important than decision support tools for older people with complex health and care needs.

• The evidence suggests that there is a need for visible organisational support in order for SDM to become embedded. This includes visibility in internal policy documents, through financial and organisational support (e.g. enabling longer appointments when necessary, providing appropriate administrative support) and through the inclusion of SDM in continuing professional development.

• The evidence suggests that systems that foster continuity of care both through ongoing relationships with one clinician (relationship continuity) and through system-based approaches that develop ways of working whereby the patient is linked to multiple professionals (management and informational continuity) are important for SDM.

• It appears likely that in instances in which choices are constrained by resource limitations, health-care policies or evidence-based recommendations, it is still valuable to explore patients choices and reasons.

• Although properly conducted SDM may increase the length of consultations (such as those in primary care), there is evidence to suggest that they may be ameliorated by involving other members of the MDT in the SDM process.

• It is important that service providers and service users have shared expectations of, and familiarity with, SDM for it to become properly embedded. This is likely to take time to develop.

Education and training

• Our findings suggest that SDM education and training should be focused on all members of the MDT and not just doctors or lead clinicians. It should be part of undergraduate training programmes but also part of ongoing professional development.

• Evidence points to a need for SDM and communication skills training to include the task of exploring what matters to patients and how to elicit their goals and priorities.

• SDM training should include information on risk communication.

Preparation for shared decision-making

• Evidence suggests that many older people with complex health and care needs, particularly those with lower health literacy or conditions, such as depression, are likely to need support to take part in consultations involving SDM.

• The evidence indicates that PDAs for older people with complex needs are likely to be most effective when used as part of a face-to-face interaction with a HCP, for example for facilitating discussion between the patient, family carer and HCP.

The Cochrane Library – searched on 25 July 2016 (no date restrictions)

“shared decision making”:ti,ab,kw (Word variations have been searched) OR “patient participation” or “patient decision making”:ti,ab,kw (Word variations have been searched) OR “decision support” or “decision aid”:ti,ab,kw (Word variations have been searched)

Used just these terms for Cochrane Database of Systematic Reviews, but added terms below for Cochrane Central Register of Controlled Trials (CENTRAL):

AND

old or older or frail* or dement* or alzheimer* or complex or aged or elderly or geriatric:ti (Word variations have been searched)

Google Scholar (July/August 2016)

• Search Google Scholar and Google (searches from 2000 onwards):

  • Shared decision making AND review

  • Patient participation AND review

  • Patient decision making AND review

  • Decision support AND review

  • Decision AID AND review

  • Health communication AND review

  • Informed decision making AND review.

Keyword searches for papers by key authors, such as Elwyn and Légaré.

Keyword searches for SDM and MDTs.

Reflecting patient and carer values

Preparing patients for the shared decision-making encounter

Sharing the communication of a decision

Fake versus real shared decision-making

Reducing the workload (no date restrictions)

Additional searches: in response to discussion with project team about areas where there appear to be gaps

Context–mechanism–outcome 1: reflecting patient and carer values

Question: the evidence suggests that for SDM to happen it is very important that systems enable health and care professionals to develop relationships with patients and carers that allow them to understand patients’ and carers’ individual emotional, cultural and cognitive needs (e.g. person-centred approaches to care). Our hypothesis is that this will help patients and carers to feel valued, listened to and involved in the decision.

Do you have any examples where this has been done well?

If yes – why do you think it worked well? What needed to be in place? What were the outcomes?

Could follow up with prompts around:

• continuity of care

• effective communication skills (what would these look like)

• less focus on biomedical targets.