Advances in psychiatric treatment (2009), vol. 15, 72–79 doi: 10.1192/apt.bp.107.005298

ARTICLE

To err is human, and medicine is no exception(Horton 1999). In the USA, Kohn and colleagues(1999) reported that at least 44 000 deaths a year resulted rom medical error; this statistic generatedalarm not only among patients and the clinicalcommunity, but also in the Clinton White House(Pear 1999). As a result, subsequent years haveseen substantially increased interest in medicalerror in both scientic (Leape 2005) and popular literature (Gawande 2002). Indeed, the eld hasgrown to the point that sub-specialties in medicalerror research have opened up, including medica-tion error, diagnostic error and cognitive error.In

How Doctors Think,

Proessor Jerome Groop-man, a Harvard haematologist and writer with the

New Yorker

, has dened cognitive errors in medi-cine simply, as ‘errors in thinking that physicianscan make’ (Groopman 2007: p. 23). He argues that errors in thinking, rather than errors o technique,orm the majority o mistakes in modern medicine,i.e. there is a ‘cascade o cognitive errors’ that results in a clinical error (p. 260). Groopman cata-logues common cognitive errors in medical practiceand outlines practical strategies or acknowledgingand correcting them.

How Doctors Think

gener-ated many enthusiastic reviews (Crichton 2007), o which ew drew attention to the ootnote on page 7:‘I quickly realised’, wrote Groopman, ‘that tryingto assess how psychiatrists think was beyond myabilities’.The omission o psychiatry rom

How Doctors Think

, and or this reason, was arguably un-necessary: the cognitive style o psychiatrists issurely not so esoteric as to be un-understandable.We suspect that Proessor Groopman would haveound psychiatrists to be like any other doctors,had he applied the literature on cognitive error topsychiatry. In this article, we do just that.

Cgnitive errr and heristics

The study o cognitive error in medicine nds itsroots in the literature on cognitive psychology romthe past our decades (Redelmeier 2001). The keypoint o departure was the work o Amos Tverskyand Daniel Kahneman, two psychologists whosestudies o decision-making under conditions o un-certainty won the Nobel Prize or Economics in2002. In a seminal paper or the journal

Science

,they discussed reliance on heuristics in decision-making (Tversky 1974). Heuristics are cognitiveshortcuts that allow decisions to be reached inconditions o uncertainty. Many individual heur-istics are identiiable (Table 1), but what theyhave in common is that they reduce the time,resources and cognitive eort required to makea decision (Croskerry 2002). The use o heuristicscan be contrasted with the hypothetico-deductivemethod o decision-making, in which all necessaryevidence or and against any potential course o action is careully examined and weighed. Thelatter assumes no bias on the part o the decisionmaker, and optimal time and resources.Heuristics are useul, particularly when time andinormation are limited. Indeed, Groopman (2007:p. 36) argues that heuristics are ‘the oundationo all mature medical thinking’. However, theyare prone to bias. Decisions based on heuristicsare more likely to be wrong than decisions madeusing hypothetico-deductive methods (Croskerry2003). Tversky & Kahneman noted that relianceon heuristics leads to cognitive bias and ‘severeand systematic errors’ (Tversky 1974). Heuristicsthat result in error are called ‘ailed heuristics’(Croskerry 2002). In this article, we reer to error resulting rom ailed heuristics as cognitive error.

Wh shld edical practitiners be prnet cgnitive errr?

Heuristics are likely to be used in situations o highcomplexity or uncertainty (Tversky 1974), when

How psychiatrists think

Niall Crlish & Brendan D. Kell

Niall Crlish

is Lecturer in Psychiatryin Trinity College, Dublin. His primaryresearch interests are early psychosis,insight and transcultural psychiatry.

Brendan D. Kell

is Senior Lecturerin Psychiatry at University CollegeDublin. His research interests includethe epidemiology of psychosis andrelationships between mental illness andsocial factors.

Crrespndence

Dr Niall Crumlish,Jonathan Swift Clinic, St James’sHospital, James’s Street, Dublin 8,Ireland. Email: niall.crumlish

tcd.ie

SummARy

Over the past decade, the study of error in medicine hasexpanded to incorporate new insights from cognitive psy-chology, generating increased research and clinical interestin cognitive errors and clinical decision-making. The study ofcognitive error focuses on predictable errors in thinking thatresult from the use of cognitive shortcuts or ‘heuristics’.Heuristics reduce the time, resources and cognitive effortrequired for clinical decision-making and are a feature ofmature clinical thinking. Heuristics can also lead to bias andmust be used with an awareness of their weaknesses. Inthis article, we describe heuristics commonly used in clinicaldecision-making and discuss how failure of heuristics resultsin cognitive error. We apply research findings on decision-making in medicine to decision-making in psychiatry andsuggest directions for training and future research intocognitive error in psychiatry.

DECLARATIoN of INTEREST

None.

How psychiatrists thinkAdvances in psychiatric treatment (2009), vol. 15, 72–79 doi: 10.1192/apt.bp.107.005298

there is a high cognitive load or a high densityo decision-making (Croskerry 2002) and whentime or individual decisions is short (Groopman2007). These conditions are most obviously met inemer gency medicine (Croskerry 2002), but in any branch o medicine, time is inadequate (Davido 1997) and cognitive eort is high (Schwarz 2005),while decisions are complex and must be madedespite inherent uncertainty (uncertainty that israrely acknowledged; Coles 2006).

Examples of cognitive error in medicine

The list o potential cognitive errors is long, with30 ailed heuristics described in an infuentialpaper on error in emergency medicine (Croskerry2002). Here we discuss cognitive errors in medicinethat may arise rom the ten heuristics listed inTable 1. They include those discussed by Groop-man (2007), with others that recur in the literature(Tversky 1974; Redelmeier 2001; Croskerry 2002,2003).

Representativeness

Representativeness occurs when thinking is guided by a prototype, so that an event is not consideredprobable unless the presentation is prototypical o it. (In medicine, the event is oten a diagnosis.) Therepresentativeness heuristic may be useul whenthe doctor is conronted with a prototypical pres-entation: pulmonary embolism can be diagnosedalmost without cognitive eort in a patient whopresents with pleuritic chest pain o acute onset with dyspnoea ollowing a deep venous thrombo-sis. A representativeness error may occur when theabsence o prototypical eatures leads to atypicalvariants being missed: or example, i pulmonaryembolism is not considered in the absence o severepleuritic chest pain. In act, only 60% o patientsover 65 years old who have a pulmonary embolismpresent with chest pain (Timmons 2003).

Availability

The availability heuristic is seen when a doctor’sassessment o the probability o an event is deter-mined by the ease with which an example comesto mind; a doctor reviewing a patient with headachemay over estimate the probability o subarachnoidhaemorrhage i they have recently seen such a case.Oten, availability is a useul heuristic, as eventscome easily to mind either because they are com-mon or, i occurring more rarely, serious enoughalways to be considered as a possibility (e.g.meningitis). An availability error occurs when theprobability o an event is overestimated because it comes easily to mind, or underestimated becauseit does not. In the above example, the doctor’srecent encounter with subarachnoid haemorrhagehas no bearing on the likelihood that the current presentation is that o tension headache, migraineor a rarer, potentially serious condition such astemporal arteritis.

Anchoring

Anchoring is the tendency to ocus on prominent eatures o a presentation too early in the decision-making process, to arrive at an early hypothesisand to ail to adjust it in the light o later inormation.First impressions are oten accurate, particularlyamong clinicians with highly developed patternrecognition skills, but they may be wrong. Tversky& Kahneman demonstrated that adjustments romrst impressions are ‘typically insucient … payosor accuracy did not reduce the anchoring eect’(Tversky 1974); that is, irst impressions have

TABLE 1Ten heristics, with strengths and weaknesses  eachHeristicStrengthWeakness

RepresentativenessQuick diagnosis, action through pattern recognitionNon-prototypical variants may be missedAvailabilityEvents that come to mind easily are common and should therefore beconsideredEvents that do not come quickly to mind are not consideredAnchoringFirst impressions often give valuable information

It is difcult to move from incorrect rst impressionsConrmation bias

NoneCan compound the failure to adjust from initial impressions (anchoring)Search satisfyingSaves the time and effort of a search for comorbidity, as often noneexistsComorbidity, which is particularly common in psychiatry, is missedDiagnosis momentumNoneInaccurate diagnostic labels persist, potentially resulting in incorrecttreatment and stigmaCommission biasAvoids omission bias; optimal information is not always available in thereal world

Adverse effects of unjustied treatment may violate the ethic of

primo non nocere

Affective heuristic Clinicians should be sympathetic towards patientsUnpleasant diagnoses or interventions may not be adequately consideredPlaying the oddsAssumption of benign diagnosis or positive outcome is usually correctNegative diagnoses or outcome may not be adequately consideredFundamental attribution errorNot applicablePatients may be inappropriately blamed and judged, to the detriment oftheir care

Crumlish & Kelly

Advances in psychiatric treatment (2009), vol. 15, 72–79 doi: 10.1192/apt.bp.107.005298

lasting power, even when they are wrong and whencorrecting them in the light o contradictoryinormation is rewarded.

Confirmation bias

Conrmation bias is the tendency to seek onlyinormation that will support rather than reutean initial hypothesis, or to selectively interpret inormation acquired ater the hypothesis is ormedin a way that supports it. The bias here is evident:a hypothesis that is true can withstand attemptsto disprove it and should be subjected to suchattempts. Conrmation bias is always an error,as it aims simply to avoid the cognitive eort that would be required to revise an initial impression,regardless o whether or not the hypothesis iscorrect.

Search satisfying

Search satisying may ollow on rom anchoring andconrmation bias. Search satisying is the tendencyto stop the diagnostic process once one diagnosishas been made. Even in the event that the rst diagnosis is correct, search satisying may be anerror, as comorbid conditions are not considered.Examples are the second racture in an X-ray or co-ingestants in poisoning (Croskerry 2003).

Diagnosis momentum

Diagnosis momentum occurs when a diagnosticlabel applied to a patient sticks, whether or not subsequent events conirm the diagnosis. Aworking diagnosis may become a nal diagnosiswithout any new diagnostic inormation having been acquired.

Commission bias

Commission bias is the tendency to action rather than inaction, even when the correct courseo action is unclear and inaction may be moreappropriate. A doctor exhibiting commission biasmay decide to institute treatment without adequateinormation to guide it, believing that it is better todo something than nothing.

The affective heuristic

Aective error occurs when the clinician’s judge-ments are biased by their emotions or hopes: judgements o likelihood may be based on what the clinician would like to be the case rather thanwhat actually is. A doctor may allow positiveeelings towards a patient to infuence their clinical judgement: because the doctor wishes the patient well, a symptom may be interpreted benignly whena more ominous interpretation is valid.

Playing the odds

Aective error may combine with the heuristico playing the odds. The latter is the tendency inambiguous situations to opt or a benign inter-pretation, on the basis that benign causes andoutcomes are more common than more ominousones (tension headaches are more common thantemporal arteritis). Playing the odds ails when arare and serious disease similar in presentation toa common benign disease is missed.

Fundamental attribution error

The undamental attribution error is the tendencyto attribute someone’s behaviour to their dis-positional qualities rather than to environmentalor situational actors (Ross 1977). However,people systematically under estimate the extent to which other people’s behaviour is infuenced by external actors (Fiske 1991). In medicine, theundamental attribution error is the tendency to be judgemental and blame patients inappropriatelyor their illnesses. Classically, it occurs whenpatients present with symptoms that are in someway precipitated or perpetuated by their own behaviour, or example smokers who present withexacerbations o pulmonary disease or intravenousdrug users who present with skin abscesses ater injecting. This may have implications or the levelo care received, as it may be elt that patients withillnesses that are not o their own making are moredeserving o care.

Errr in pschiatr

Mistakes in psychiatry can have serious conse-quences or patients, clinical teams and the wider community (Kapur 2000). However, the literatureon error in psychiatry is small (Grasso 2003) andnarrow, with most studies ocusing on medication.Little has been written on diagnostic error, whichwas just briefy touched on in the most thoroughreview o error in psychiatric practice (Nath 2006).Some work has been done on error in predictingorensic risk (Freedman 2001). Other than a noveltechnical paper on cognition in emergency psy-chiatry (Cohen 2006), there has been no systematicstudy o cognitive error in psychiatry. There are,however, reasons why the practice o psychiatrymight be prone to error o this type.

Wh shld pschiatrists be prnet cgnitive errr?

As noted above, heuristics are likely to be used,with their attendant risk o cognitive error, whenthere is a high cognitive load and limited time tomake decisions, and in situations o complexity