Cognitive biases in medical decision-making relevant to paediatric RTI
| Cognitive bias | Description | Example/consequence of relevance to paediatric RTI |
| (1) Anticipated regret | The probability of a diagnosis with a severe outcome is overestimated due to a heightened sense of future regret in the event of missing the diagnosis. | Clinicians’ fear of ‘missing the sick child’ leading to prescribing ‘just in case’,41 due to perceptions that not prescribing carries greater potential threat. |
| (2) Anchoring and adjustment | Assessing new cases in relation to a previous case, rather than a population baseline. | Assessing a child’s RTI as severe/not in comparison to the last sick child/ren seen, rather than as a new case against a broad population baseline. |
| (3) Confirmation bias | Selectively gathering and interpreting evidence to confirm a diagnosis, and ignoring evidence that may disconfirm it. | Deciding a child needs antibiotics based on a ‘gut’ feeling and looking for reasons to prescribe. |
| (4) The availability bias | Information that is easily recalled is given high importance. That is, salience correlates with decision-making, regardless of the quality of the evidence. Information salience is increased by being: frequent, recent, unusual, emotive or high profile. Research shows that simply imagining a diagnostic outcome (therefore making it salient) will raise a clinician’s subjective probability of its likelihood.42 | Remembering a child with RTI symptoms who deteriorated when not offered antibiotics; media reporting of a child deteriorating after seeing their GP. |
| (5) Representativeness | Assuming that what presents in clinic represents a ‘real’ state of events, includes: (A) not accounting for regression to the mean by assuming acute symptoms are representative of the illness, rather than an anomalous peak; (B) assessing only by the similarity of symptoms with possible diagnoses, and ignoring relevant base rate probabilities of diagnostic options; (C) the gambler’s fallacy of reasoning that sequential cases represent the spectrum of probabilities, for example, after four similar successive cases given diagnosis A (80% probability), similar case number 5 is given diagnosis B (20% probability), rather than being assessed independently as having 80% probability of diagnosis A. | Prescribing antibiotics to a proportion of children presenting with RTI, based on symptoms on the day. |
GP, general practitioner; RTI, respiratory tract infection.