Shared decision‐making: how can we personalise risk?

We read with interest the PeriOperative Quality Initiative consensus statement by Leung et al., which has developed clear definitions and recommendations for shared decision-making in the peri-operative period [1]. The authors are to be commended on developing such practical guidance, particularly in stratifying resources and helping those patients who require additional decisional support. We would like to highlight some considerations in discussing patient-specific risk in this shared decision-making setting. First, many tools (e.g. the Surgical Outcome Risk Tool) that we use to aid risk discussions with patients have limitations [2]. The authors encourage clinicians to elicit non-mortality outcomes (e.g. complications, poor quality of life) that are unacceptable to patients, with the intention of counselling them accordingly, but many of these tools focus on mortality. Future shared decision-making conversations may be aided by outcome risk tools that give more meaningful and personalised outcome risks. Second, when we have patient-centred risk data to counsel patients, many clinicians have little training in how to understand, interpret or communicate this information [3]. Communicating the uncertainty in outcome risk models (e.g. 95% confidence intervals in risk score outcomes) to patients with a range of statistical literacy presents a challenge to clinicians and adds complexity to shared decision-making consultations. Guidance exists [4], but clinicians undertaking shared decision-making will need formalised training on how best to explain these nuances across a range of patients to avoid potential negative consequences [5] in addition to the experiential communication skill training advocated by the authors. Third, potential clinician preference for action over conservative management must be acknowledged. Communicating the risk of not proceeding with surgery is difficult, and clinicians in these shared decision-making conversations should be aware of this ‘action bias’. As highlighted by the authors, risks should be contextualised within the patient's health trajectory, but this is not simple. Often, peri-operative teams have limited means to refer patients promptly to teams who will lead non-operative treatment plans (e.g. palliative care). One way to alleviate this action bias could be to have established non-surgical pathways (e.g. to physiotherapy, chronic pain clinics, palliative care) available from these clinics so enabling fuller discussions of the options. Finally, a significant barrier to the operational delivery of shared decision-making is the persistence of traditional staffing models. The multidisciplinary model needed to support high-quality shared decision-making does not fit easily in conventional outpatient clinic structures. As a result, informal ‘corridor conversations’ between healthcare professionals become necessary workarounds, reflecting a mismatch between current working structures and the aspirations of delivering high-quality peri-operative care. Developing newer models of working will take time and will have to balance patient, staff needs and operational efficiency.