Behavioural Science & Sociocultural Psychiatry · Paper A
Cognitive heuristics
Diagnostic mistakes often begin as ordinary shortcuts doing their job too quickly. A recent dramatic case makes one diagnosis feel common; a first impression keeps later evidence trapped around the original answer; a familiar-looking patient is forced into a prototype. The exam move is to name the pull, not just say 'clinical judgement was biased'.
Availability biasA systematic error that skews a study's results away from the truth in a particular direction.Read more → judges likelihood by ease of recall, so recent, vivid, or emotionally striking cases feel more probable than they are. Representativeness judges by resemblance to a prototype and can ignore base rates. Anchoring clings to the first value or diagnosis despite later information. Framing changes decisions when the same facts are presented as gains or losses. Confirmation bias searches for evidence that supports an existing belief, while premature closure stops the diagnostic search too early. A practical antidote is a diagnostic pause: ask what else could fit, check the base rate, and actively seek evidence that would disprove the favoured answer.
A HEURISTIC (a fast mental shortcut the brain uses to make a quick decision without weighing every possible option) is not automatically an error: it is usually a useful time-saver. What the exam actually tests is whether you can name the SPECIFIC heuristic at play in a scenario, because each one distorts judgement through a different, identifiable mechanism, and a vague answer of 'cognitive ' scores nothing.
Cognitive heuristics are shortcuts the brain uses instead of slow, effortful, fully rational analysis of every option. They usually work well, which is exactly why the brain relies on them by default, but each has a predictable blind spot where it produces a systematically wrong answer rather than a random one. Recognising that the error is systematic, and predictable, is what makes these worth studying rather than dismissing as random mistakes.
A common error is treating all heuristics as one interchangeable idea, '', rather than five distinct mechanisms with five distinct fixes. Another is assuming a heuristic is always harmful. In everyday low-stakes decisions a heuristic saves time with little cost; the exam specifically probes the clinical scenarios where that same shortcut produces a genuinely wrong judgement.
Rational, fully weighed decision-making is slow and effortful, so the brain defaults to fast pattern-matching whenever a decision feels routine or time-pressured. Imagine a clinician seeing patient after patient with the same common cold: pattern-matching serves them well, since most such presentations really are a common cold. The same shortcut becomes a trap the moment an unusual presentation is dressed up to resemble the routine one, because the brain's fast pathway does not pause to check the underlying probability, only the surface resemblance. That gap between surface resemblance and true probability is where nearly every named heuristic below does its damage.
Availability heuristic: frequency is judged by how easily an example comes to mind, so recent, vivid, or emotionally memorable cases get overweighted relative to their true frequency. A doctor who has just seen one dramatic case of a rare disease may overestimate how common that disease actually is in the weeks that follow.
Representativeness heuristic: likelihood is judged by how closely a case resembles a typical prototype, while the actual BASE RATE (how common a condition truly is in the relevant population) is neglected. A presentation that looks textbook-classic for a rare disease is still, on base rates alone, far more likely to be a common condition presenting atypically.
Anchoring: an initial value or first impression is given far more weight than later evidence deserves, so subsequent judgements stay pulled too close to that starting point. A first, tentative diagnosis mentioned early in a referral letter can anchor every clinician who reads it afterwards, even once contradicting evidence appears.
Framing effect: identical choices produce different decisions depending only on whether the outcome is presented as a gain or as a loss. A treatment described as '90% survival' is chosen more often than the identical treatment described as '10% mortality', despite the two statements being mathematically the same fact.
Confirmation and premature closure: confirmation bias is selectively seeking or interpreting evidence that supports an existing belief, and premature closure is stopping the search for alternative explanations too soon. A clinician who forms an initial diagnosis within the first minute may then unconsciously reinterpret every later symptom as confirming that first guess, a classic pairing of the two.
Availability versus representativeness is the pairing the exam loves most: availability is driven by ease of recall, so it is worsened by one recent, vivid case, whereas representativeness is driven by resemblance to a stereotype and is worsened by a textbook-perfect presentation, regardless of how recently anything similar was actually seen. Anchoring versus framing is told apart by what is being manipulated: anchoring is about which number or impression came first, while framing is about how the very same information is worded, as a gain or as a loss. Confirmation is told apart from premature closure by timing: confirmation bias can continue indefinitely, actively filtering new evidence, while premature closure is specifically the act of stopping the search itself.
There is no laboratory test for a heuristic; it is recognised through structured self-review, most often using a technique called a DIAGNOSTIC TIME-OUT (a deliberate pause where a clinician explicitly asks what else this could be, before committing to a working diagnosis). Reviewing a case retrospectively and asking which specific heuristic best explains a documented misjudgement is the exam's preferred format, and is also genuine clinical-audit practice.
Each heuristic has a matched countermeasure rather than one generic fix. Availability is countered by deliberately reviewing the true base rate rather than relying on memorable recent cases. Representativeness is countered the same way, by consciously checking base rates before pattern-matching to a stereotype. Anchoring is countered by deliberately reconsidering the case as if the initial value had never been mentioned. Framing effects are countered by restating the same choice in both gain and loss language before deciding. Confirmation bias and premature closure are both countered by actively asking what evidence would prove the current working diagnosis wrong, and deliberately seeking that disconfirming evidence out. Source-wise, this page is grounded in Kahneman-and-Tversky cognitive psychology and Croskerry-style diagnostic-error literature on cognitive forcing strategies; it is not a NICE, DSM-5, or ICD-11 treatment guideline topic. The management implication is educational and safety-system based: diagnostic time-outs, checklists, second opinions, morbidity-and-mortality review, and audit feedback reduce risk by making the clinician deliberately test the shortcut before acting on it.
When a vignette hands you a clinician who is clearly wrong, resist naming '' as the answer. Identify exactly which mechanism, ease of recall, resemblance to a prototype, an anchored first value, how the choice was worded, or a refusal to reconsider, produced that specific wrong answer.
Every heuristic is a fast shortcut with its own specific blind spot, and matching the blind spot to its name is what actually earns exam marks, not simply saying ''.