Trial 1

Improving medical diagnosis

Trial 1

Can a mini review of a working diagnosis with another doctor reduce cognitive bias?

Hospital Emergency Departments (EDs) are busy environments where clinicians must quickly and accurately diagnose cases that are often complex and where they are not familiar with the patient.

In this environment there is a risk of misdiagnosis, which can affect patients receiving timely and appropriate care. While Australia faces similar issues to other Western counties, estimates suggest that diagnostic errors happen in 10-15% of cases.

One of the main contributing factors to misdiagnosis is cognitive bias (the mental shortcuts we all use), which can influence clinician decision-making, leading to diagnostic errors, poor patient outcomes and possibly even deaths.

Cognitive bias is thought to be at play in more than three-quarters of misdiagnosis cases, with two biases being:

  • anchoring bias, where clinicians lock onto salient information early in the patient’s presentation and fail to adjust this impression in light of further information and,
  • premature closure, where the investigative process ends too early and disconfirming information is not sought.

To address the issue, we developed and trialled a Rapid Diagnostic Discussion (RaDD) tool to encourage clinicians to have a structured discussion about the possible causes of abdominal pain – the most frequent presentation to Victorian EDs.

What did the literature tell us?

The first phase of the trial involved a Facilitated Dialogue with key stakeholders.

To ensure that participants had a good understanding of the issue – and what others around the world had done to address it – they were briefed with the results of a Rapid Evidence and Practice review (below).

Briefing document PDF

What's likely to fix the problem?

From the Facilitated Dialogue, participants drew the following conclusions on how to best approach the issue of misdiagnosis.

Dialogue summary PDF

What did we do?

We developed and trialled a Rapid Diagnostic Discussion (RaDD) tool to help reduce cognitive bias in the diagnosis of abdominal pain – the most common presentation to Victorian hospital EDs and one that can be complex to diagnose.

Doctors at Eastern Health’s Box Hill Emergency Department used the tool to query diagnoses on 155 patient cases. Patients were first examined by one doctor, then independently reviewed by a second doctor.

Both doctors then engaged in a prompted discussion about their working diagnoses before proceeding with patient management.

Question prompts were:

  • “What else could this be?”
  • “Are there any symptoms or signs that don’t fit? Any red flags?”
  • “Are additional tests/investigations needed, or do you already have sufficient confidence in the diagnosis to proceed without them?”
  • “Do we have all the relevant information about the patient?”

The trial involved participants at three Victorian hospitals. Box Hill represented the ‘test’ hospital. Two other ‘control’ hospitals were involved (they did not use the tool).

For further information, visit the Trial Registration page on the Australian New Zealand Clinical Trials Registry.

What did we find?

RaDD was shown to be an effective tool for reducing cognitive bias and improving communication between clinicians. The trial showed that RaDD can result in doctors rethinking their working diagnoses.

Of the 155 patients enrolled in RaDD, the original working diagnosis of the first doctor was changed in almost a quarter (24.7%) of cases.

Clinician confidence in their diagnoses also increased and the use of RaDD led to fewer patients being discharged home.

Media and other information

Full trial infographic PDF