Dual Process Theory (System 1 and System 2 Clinical Reasoning)

Reza Manesh, Denise M. Connor, Jeff Kohlwes

The Dual Process Theoryhas been adaptedfrom the psychology literature todescribe how clinicians think when reasoning through a patient’s case.1The dual processes, or System 1 and System 2, work together by enabling a clinician to think both fast and slow when reasoning through a patient's presentation.

System 1 is intuitive, efficient, and based on pattern recognition.2Reasoning using System 1 often occurs so quickly that we do not explicitly recognize it as a distinct cognitive process. For example, a post-operative patient with sinus tachycardia, asymmetric lower extremity edema, and hypoxia is recognized immediately as having a pulmonary embolus by an experienced clinician. This rapidthinking draws on prior clinical experience, and is invaluable in helping busy clinicians accurately assess and treat patients with straightforward presentations.

In contrast,System 2 is an analytical cognitive process that is time intensive and deliberate.3 It involves the conscious, explicit application of an analytical approach to arrive at the correct diagnosis.An HIV positive patient with a CD4 count of 50 with fevers, weight loss, headaches, diarrhea, andrecent travel to South Africa would likely activate System 2 reasoninggiven the myriad diagnostic possibilities. Complicated or atypical patient presentations that do not closely match known patterns require clinicians to slow down and systematically consider multiple potential etiologies to avoid making diagnostic errors.

Which system is activated at a given time depends on two factors:the providers’ priorexperience with a particularclinical presentation, and their ability to activate the appropriate illness scriptthat sufficientlyexplains the patient’s clinical syndrome.3,4Medical students tend to utilizeSystem 2thinking more often thanSystem 1thinking since they have insufficient clinical experience to accurately reason through a case using pattern recognition alone. In contrast, seasoned clinicians practicing in a familiar setting spend more time utilizingSystem 1 reasoning unless triggered to switch to System 2 when a patient does not neatly matchone of their storedillness scripts.

Like most models, the Dual Process Theoryoversimplifies reality.5 In real-world practice, a clinician’s reasoning process is unlikely to fall exclusively into either category, but rather oscillates between the two, even within a single case.6 Choosing the right analytic frameworks to use, and selecting the appropriate clinicalfeatures to consider are difficult tasks and require practice. To avoid mistakes, experts often check a diagnosis they arrived at quickly through System 1reasoning by applying System 2reasoning to the case.

Diagnostic errors can occur withSystem 1 or System 2 thinking. However, the types of errors clinicians are most at risk for making differ depending on which end of the spectrum they are operating in. For example, exclusively utilizing System 1,and being too reliant on fitting a patient into a previously stored pattern,may lead clinicians to unconsciously ignore key aspects (history, exam, labs, or imaging) of a patient’s presentation that do not fit with their initial diagnosis,leading them to anchoron an incorrect diagnosis. An elderly man with jaundice, weight loss, and a pancreatic mass should make a clinician reflexively consider pancreatic adenocarcinoma. However, if the endoscopic biopsy results are benign, then presuming the biopsy results are false and proceeding with a Whipple procedure without firstpausing to consider other causes of this clinical picture (e.g., IgG4-related disease) would represent a failure to slow down and switch to System 2 thinking.In contrast, System 2may place undue emphasis on a particular finding from history, exam, or diagnostic testing. For example, a patient with coronary artery disease, dyspnea on exertion, and worsening bilateral lower extremity edema—but normal pro-BNP—still likely has heart failure. Undue emphasis on the normal pro-BNP may unnecessarily broaden the differential to include cryptogenic organizing pneumonia, resulting in unnecessary tests and delays in appropriate treatment.

References:

  1. Klein G. Naturalistic decision making. Hum Factors. 2008;50:456–60.
  2. McCormic JS. Diagnosis: the need for demystification. Lancet 1986;2:1434-5.
  3. Elstein AS, Schwarz A. Clinical problem solving and diagnostic decision making: selective review of the cognitive literature. BMJ 2002;324:729-32.
  4. Schmid HG, Norman GR, Boshuizen HP. A cognitive perspective on medical expertise: theory and implication. Acad Med 1990;65:611-21.
  5. Monteiro SM, Norman G. Diagnostic reasoning: where we’ve been, where we’re going. Teach Learn Med. 2013;25:26-32.
  6. Eva KW. What every teacher needs to know about clinical reasoning. Med Educ 2005;39: 98–106.