Clinical Reasoning and Differential Diagnosis
Building the differential
Sound reasoning starts from a problem representation: age, sex, setting, tempo (acute versus chronic), and the pivotal finding. From that, generate a differential that spans the likely, the serious-but-uncommon, and the treatable. Use anatomical and pathophysiological frameworks (odontogenic versus non-odontogenic; pulpal versus periapical; local versus referred) so the list is organised rather than a memory dump. Vindicate each candidate with a discriminating test rather than collecting every test available.
Cognitive biases examiners probe
Anchoring (locking onto the first diagnosis), premature closure, availability bias (recently seen cases feel more likely), and confirmation bias (seeking only supporting evidence). Questions frequently present a case where the obvious finding is a distractor and a single historical detail reframes the diagnosis. Another trap is over-investigating a benign, self-limiting condition, or under-investigating a red flag because the patient looks well.
Rule of thumb
Apply the "worst-case first" rule: before settling on a benign diagnosis, ask what the most dangerous alternative is and whether you have excluded it. If a finding does not fit, revise the diagnosis rather than the finding.
Worked example: A 58-year-old with burning oral mucosa, fatigue, and angular cheilitis. The obvious label is candidiasis; the discriminating question is about diet, bowel change, and anaemia history, which points toward B12 or iron deficiency and prompts haematology referral rather than empirical antifungal therapy alone.