Shared decision making in realistic medicine: what works

3. Barriers

The literature around SDM (Carter & Martin, 2018; Coulter, 2006; Elwyn, et al., 2016; Legare & Thompson-Leduc, 2014; Legare, et al., 2008) highlights a number of barriers preventing greater engagement with, and uptake of SDM within clinical consultations, which may explain why SDM is not part of everyday medical practice (Gulbrandsen, et al., 2016).

The barriers may be actual or perceived, and have been grouped within three overarching categories: barriers surrounding knowledge; around expectations; and around the meaningful conversation. Solutions require a mix of individual, board and system level changes (see Section 4 for examples). Each barrier will now be discussed in more detail.

3.1 Knowledge

Both clinicians and patients reported a number of barriers concerning knowledge. For clinicians current knowledge barriers include a lack of knowledge of how to carry out an SDM process (Legare & Thompson-Leduc, 2014), often manifested in the (often incorrect) belief of clinicians that they are 'already doing SDM'. In reality, there may be a focus only on particular aspects (i.e. providing information to patients).

Another clinician-cited knowledge barrier was highlighted by Joseph-Williams et al (2014), whereby clinicians (particularly less experienced clinicians) can worry that asking patients for their views and preferences may reduce a patient's trust in their skills as a clinician.

A common barrier for clinicians was difficulties responding to patients who disagree with a recommendation (Zeuner, et al., 2014). However, evidence suggests that generally patients would prefer to be asked for their views and any lack of trust would more likely arise from poorly communicated information, or a poor evidence base to draw from (i.e. inconsistent access to information of the risks and benefits of certain treatments). This can be improved through additional training and creation of a better evidence base (Joseph-Williams, et al., 2017).

Furthermore, increased uncertainty may occur by raising patients' awareness of additional options with uncertain outcomes in terms of risks and benefits (Gulbrandsen, et al., 2016). However this again can stem from poorly communicated information, or poorly executed SDM processes, but may be a particular concern for patients with lower health literacy.

Low health literacy is the main knowledge barrier for some patients. Generally, patients with lower levels of health literacy:

• have poorer health status and self-reported health;
• are less able to communicate with health professionals;
• wait until they are sicker before going to the doctor;
• have an increased risk of developing multiple health problems; and,
• are less likely to engage with health promotions such as flu vaccines and breast screenings. (Making it easy, 2014)

Low health literacy will remain a barrier towards greater SDM unless changes can be made to make active participation in healthcare easier. Calls for health and social care services to do more to support better health literacy and numeracy are often promoted as solutions to these barriers through the use of simple language and providing information in various forms (Coulter & Collins, 2011).

3.2 Expectations

There appears to be a mismatch of expectations of healthcare services between patients, clinicians and wider healthcare systems which has created a number of barriers (Joseph-Williams, et al., 2014; Zeuner, et al., 2014; Fisher, et al., 2018). Some patients have the expectation that instigating a process of SDM (i.e. more active involvement in decisions) may lead to poorer treatment from their clinician (i.e. because the clinician may feel the patient has disrespected their knowledge or role) (Zeuner, et al., 2014). This expectation may stem from the paternalistic healthcare model in which 'doctors knows best'.

Meanwhile some patients and clinicians both reported they do not pursue SDM because they do not believe it will actually happen. This may be connected to self-efficacy (one's belief in one's ability to succeed at a task) (Legare & Thompson-Leduc, 2014).

Furthermore, there can be inconsistencies between how patients and clinicians conceive of and understand 'a shared decision', a decision in general (Leppin, et al., 2014), or the most appropriate style of SDM for each situation (Fisher, et al., 2018).

Perhaps the most frequently cited reason for the lack of SDM uptake by clinicians remains the lack of time in consultations (Joseph-Williams, et al., 2014), with the expectation that SDM discussions will add to an already pressured 10 minute consultation (Legare & Thompson-Leduc, 2014). However much literature exists to show that often SDM can reduce the length of time per patient overall in terms of more targeted follow up appointments and discussions (Legare & Thompson-Leduc, 2014), especially if the patients are given information beforehand and come to their appointment prepared. For example, Butow et al (1995) found that on average patients spoke for only 24% of a consultation and asked 5.6 questions on average, which took up less than 0.1% of the total consultation time. In comparison clinicians spoke for 44% of the consultation time on average, and only 5% of this time was spent answering patient questions.

In addition, SDM should not be expected to always occur within a single consultation (Elwyn, et al., 2017). SDM is about exchanging information in a more useful and accessible way, e.g. personalised to the patient in terms of their preferences. It is argued that 'any change that is proposed for the busy clinical context is always assumed to add more time to the consultation' (Legare & Thompson-Leduc, 2014, p. 283), when in actual fact no evidence has yet been produced to support this claim. A Cochrane systematic review highlighted research on ten decision aids specifically embedded in interventions that measured consultation lengths and the majority of studies found that there was no statistically significant change in consultation length (Stacey, et al., 2014). Furthermore some of the decision aids reduced indecision in participants overall, had a positive effect on patient-clinician communication, and reduced the number of decisions for major invasive surgery, in favour of more conservative decision options (Stacey, et al., 2014).

Some patients (e.g. often older generations) can feel pressured to cut appointments short if the waiting room is busy (as they don't want to waste valuable clinician time) (Bastiaens, et al., 2007). Giving clinicians and patients the right tools, skills and conditions for undertaking SDM, away from pressurised targets and waiting lists, may alleviate this barrier.

Finally misaligned expectations around patient preferences can also be a barrier. Lack of easy access to robust evidence of certain treatments, as well as a perceived lack of time to assess patient preferences can lead clinicians to have incorrect assumptions over what matter most to their patients (Mulley, et al., 2012; Zeuner, et al., 2014). For instance Lee et al (2010) highlighted research in which clinicians believed that 71% of patients with breast cancer consider keeping their breast as a top priority, when in reality the figure was only 7%. This highlights the importance of understanding patient preferences, and also reducing waste through lowering the incidence of unwarranted treatments and having more meaningful conversations.

3.3 Meaningful conversation

An effective SDM process requires meaningful conversations between patient and healthcare professional to determine the preferences of the patient (Elwyn, et al., 2017). A number of factors can act as barriers to these conversations, namely: unrecognised power imbalances, lack of trust and familiarity; and lack of self-belief.

SDM requires a change, at least to some degree, in the power dynamics between patient and clinician from the traditional paternalistic healthcare model (Fisher, et al., 2018). This transfer of power needs to be recognised by both clinician and patient to enable an SDM process to take place. Nevertheless the balance of power is still held by the healthcare professional, given that, along with the diagnosis and prognosis information they hold, these SDM processes often take place in environments that are more familiar to the healthcare professional (i.e. GP surgery/ hospital) (Joseph-Williams, et al., 2014). These environments are often not attune to meaningful conservations. Physical barriers such as desks and computers (Joseph-Williams, et al., 2014) may be a hindrance to trust building and may reinforce the power imbalance between patient and clinician.

However some patients (e.g. some older people) will not be willing to participate in SDM to the same degree due to long-term acceptance of and familiarity with, the 'doctor knows best' outlook for their healthcare (Bastiaens, et al., 2007). As such clinicians need to be more aware of the preferences of the patient in relation to their desire for involvement in SDM.

SDM also necessitates that the patient trusts their clinician to engage in a discussion about their preferences. The issues of power and trust are dependent to an extent on the way in which information is offered and retained (e.g. clinicians need to carefully consider when to provide certain information and through which means) (Gulbrandsen, et al., 2016; Joseph-Williams, et al., 2014). The SDM literature (e.g. (Legare, et al., 2008)) highlights that concern exists for some clinicians around how to effectively begin an SDM discussion, particularly in times when the discussion may concern end-of-life options. Difficulties may arise when trying to elicit preferences from patients as 'what matters to you' is not easy to answer (Jacobsen, et al., 2018). This barrier can lead some clinicians being more likely to suggest a treatment that they are more comfortable performing (even if it may not match with the patient's preferences) (Mulley, et al., 2012). Furthermore clinicians may find it difficult to 'hide' biases they have toward particular treatments, which may almost subconsciously appear through use of a particular tone, or time given to discussing each option (Joseph-Williams, et al., 2014). This may damage trust and deter patients from further SDM.

A final barrier concerning difficulties around enacting meaningful conversation relates to a lack of familiarity between the healthcare professional and patient. Whilst SDM remains a relatively new means of healthcare delivery greater importance is placed on trust and familiarity between clinician and patient. Thus having appointments with different clinicians, can be a perceived barrier to SDM as then extra, already pressurised, consultation time is spent retelling medical/ treatment history (Joseph-Williams, et al., 2014). This is a particular issue for patients with longer term or chronic conditions who have more contact with healthcare professionals. Chronic conditions affect over 15 million people in the UK, particularly as we age, and recent research highlighted that 42% of the population have at least one chronic condition (Coulter, et al., 2013).

In terms of specific examples, Beaver et al (2005) conducted a qualitative study into the preferences for SDM of patients undergoing treatment for colorectal cancer, concluding that trust in their clinician and continuity of care were important factors in their likelihood to participate in SDM. Furthermore Belcher et al. (2006) found that older people were more likely to participate in their own medical decision making if they trusted their clinician.

3.4 Summary

Overall, the barriers around knowledge, expectations and the meaningful conversation require a mix of individual, board and system level solutions. There is a lack of understanding (patient and clinician) of what the SDM process could look like in practice, and some barriers related to the legacy of paternalistic healthcare (e.g. 'I'm already doing it', or 'it's disrespectful to my clinician'). Raising public and clinician awareness of 'SDM in practice' may subsequently help with other barriers such as lower health literacy, the importance of key conditions (e.g. trust and power imbalances) and misaligned expectations of preferences between patient and clinician. Some examples of solutions to these barriers are presented in Section 4.