Misdiagnosis: analysis based on case record review with proposals aimed to improve diagnostic processes

Editor–Neale and colleagues provide a valuable reminder that misdiagnosis is commonly related to cognitive errors (Clin Med August 2011 pp 317–21). They encourage a shift in emphasis from intuitive (system 1) to analytical (system 2) thinking because ‘over-reliance on a simple perceptive approach to diagnosis may forestall analysis’. However, general strategies for correcting cognitive biases can be problematic. Croskerry¹ advises against viewing the two systems separately and has promoted a combined approach, emphasising the complexity of decision making in practice. Norman and Eva² have drawn attention to biases inherent in the analysis of errors. They note that similar errors are attributable to several mechanisms but cite some gains from combining the analytic and intuitive. They suggest that simple prompting strategies may be better than formal teaching about cognitive biases. Thus, to avoid the common bias of premature closure, ‘think of the first thing that comes to mind but think of other possibilities’. Some studies provide evidence that teaching more analytical reasoning may sometimes, paradoxically, worsen results.³ Norman and Eva reference a similar point.² None of this, of course, negates the need for analytic thought in context.

We have recently proposed that memorable aphorisms can still be valuable aids to judgement.⁴ While such heuristics (short cuts) have come in for criticism they are not inherently bad (or good) but must be applied in context and reviewed critically. Neale and colleagues seem to be thinking along these lines where they say ‘To remind clinicians not to ignore the pelvis perhaps the term ‘PR (per rectum)’ might be replaced by ‘RPE’ (rectal and pelvic examination)’. Could this be ripe for an aphorism? This important point about terminology influencing behaviour deserves more study. One of us has observed that the ubiquitous use of abbreviations such as ACS or TIA can cause diagnostic error by turning a verbal short cut into thought cut short.⁵ Few would argue with their suggestion for structural prompts in records. Too often ‘clerking’ is seen as an end in itself. Physicians will support their call for reflection, resisting speed of throughput at the expense of time for thought.

The authors treat the unfolding case as a series of links in a chain, noting error at various points. However, the linear chain is not always a good model for healthcare. Working with patients is a complex system involving uncertainty and unfolding over time. Within such a system, decisions require a wider process of sensemaking and situation awareness that must include networks of persons, as indeed the authors suggest in their recommendation for more consultation. Encouraging doctors (and others) to challenge diagnoses and voice uncertainty is essential, and inculcating these behaviours and collaborative work habits should start in medical school.