Five steps for managing uncertainty and diagnostic error
This is a summary of the slides presneted on 12 November 2015 at the Recently Appointed Consultants Evening Meeting at RCPE.
1. Understand it
We all make mistakes - to err is human! Even conservative estimates of diagnostic error rates in published literature suggest around a 20% misdiagnosis rate. We will therefore all make diagnostic errors, irrespective of expertise, because of our innate and evolutionarily ancient tendency to rely on heuristic-based thinking. Heuristics are cognitive shortcuts our brains intuitively, automatically use to solve particular problems; they are quick and reflexive, and used to generate an approximate answer to a reasoning question, but are prone to error, often introducing or relying upon bias and prejudice. Decision making ability is particularly impaired when tired, unwell or over-worked - familiar circumstances!
2. Accept it
The Physician's Reaction to Uncertainty (PRU) scale shows that urologists and orthopaedic surgeons typically encounter the least uncertainty, and general physicians, GPs and psychiatrists the most - perhaps no surprises there! Higher scores are also found in younger physicians. Uncertainty as a recently appointed consultant physician is therefore inevitable.
3. Share it
Your colleagues all share the same insecurities about uncertainty/diagnostic error, and are there to help you! There are also a wealth of resources available to support decision making, usually available to your Board/Trust for free, e.g. via the Knowledge Network (Scotland), including UpToDate, Toxbase, SIGN/NICE guidelines etc.
Learning from our own and others’ mistakes is hard but useful - e.g. shared as an organisation/department via Datix or M&M meetings. Involving trainees in these during early stages of training will help embed this as normal practice and hopefully make these less intimidating for generations to come.
Patients will generally accept uncertainty if it is skilfully handled. For example, there is good evidence that patients prefer behavioural expressions of uncertainty (looking something up; calling a specialty colleague) to verbal expressions of uncertainty ('I don't know', 'we'll have to see'). A widely shared format for handling medical uncertainty is included in section 4 below.
4. Safety net it
One recognised strategy for handling medical uncertainty and complexity is as follows:
- Definite the context of the diagnosis and explain the patient's symptoms and signs as part of the expected spectrum of the disease;
- Eliminate alternative diagnoses by dealing with patient fears, giving reasons in the context of the patient's belief system;
- Describe the prognosis in terms of the likely course of the disease and expectations of treatment;
- Negotiate key problems or issues that are important to both patient and physician;
- Negotiate the plan and ensure the patient understands, and is willing and able to comply, given their particular context;
- Keep diagnostic options open by making provisional diagnoses while keeping alternatives in mind;
- Remain circumspect and take action to minimise the possibility of missing other critical diagnoses;
- Play for time if necessary by allowing signs and symptoms to develop to help clarify the diagnosis;
- Plan for contingencies by providing appropriate if/then statements concerning situations requiring further action.
We are also surrounded by a wealth of validated tools, bundles and checklists to support this process.
Diagnostic support apps, such as Isabel and Dxplain, are more commonly used in the USA, but are likely to become more commonplace as digital solutions are integrated and accepted in the UK.
5. Teach it!
This is part of the unwritten curriculum that we can help embed in undergraduate teaching and postgraduate training. Consider including aspects of the following:
- Learning about common heuristics and cognitive bias.
- Encouragement of systematic critical thinking.
- Embracing zebras ! 'Zebra' is common medical slang for arriving at an esoteric or exotic diagnosis where a more commonplace explanation is significantly more likely, as in 'if you hear hoof beats outside, think horse, not zebra'.
- Admitting one's own mistakes and lead by example, e.g. submit a Datix report with your trainee; invite them to co-present at an M&M meeting.
- Encouraging trainees to seek advice more readily (a key recommendation in consecutive NCEPOD reports).
- Empowering trainees not to be afraid of 'covering your bases', and promote 'worst case scenario medicine'.
- Seeking evidence for things that fit AND things that don't, and have a shortlist of things to exclude (for an acute physician, sepsis and pulmonary embolism are always on that list, for example!)
References
Hewson MG et al. Strategies for managing uncertainty and complexity. J Gen Intern Med. 1996;11(8):481-5.
Trowbridge RL. Twelve tips for teaching avoidance of diagnostic errors. Medical Teacher 2008;30:496-500.