Re-defining ‘clinical reasoning’

Clinical reasoning is an archetypal feature of the medical profession that has garnered increasing attention in recent years, due to the high prevalence of diagnostic errors in real-world clinical practice.¹ As Professor Ian Scott aptly described in a previous article, flaws in clinical decision-making ‘are (often) not due to incompetence or inadequate knowledge, but (rather attributed) to frailty of human thinking under conditions of complexity, uncertainty, and pressure of time’.²

While medical practitioners will readily agree on the importance of clinical reasoning to be taught and assessed in medical education,³ ironically, most will be stumped when asked to provide a proper definition of clinical reasoning as a concept. In 1991, Kassirer and Kopelman defined clinical reasoning as ‘strategies… to combine and synthesise diverse data into … diagnostic hypotheses, make the complex trade-offs between the benefits and risks of tests and treatment, and formulate plans for patient management’.⁴ Two decades later, clinical reasoning remains vaguely defined as ‘thinking and decision-making processes associated with clinical practice’,⁵ with some also recognising the need to consider ‘both conscious and unconscious cognitive operations interacting with contextual factors’.⁶ As Yazdani correctly pointed out, despite decades of research since the 1980s, clinical reasoning has remained a ‘challenging, promising, complex, multidimensional, mostly invisible, and poorly understood process’.⁷ This naturally creates a major problem with downstream development and implementation of curricular pedagogy.

Therefore, in this article, we sought to revisit the concept of clinical reasoning by discussing it through three lenses: first, as a practice-defining model construct underpinning medical decision-making; second, as a learnable skill attained through deliberate practice and dialectical instruction in naturalistic settings; and, third, as the ultimate medicolegal yardstick for the professional standard of care.

In practical terms, clinical reasoning is a highly sophisticated process by which physicians arrive at medical decisions in real-world practice, which involves cognitive processing of clinical information, psychoemotional cues and situational/contextual attributes.

To date, the most popular theory of clinical reasoning is Kahneman's dual-process model of intuitive (thinking fast) and analytical (thinking slow) thought patterns.⁸ In brief, intuitive thought processes are fast, unconscious and reliant on heuristics or pattern recognition, whereas analytical thinking is slow, conscious, reasoned and systematic.⁸ However, in recent years, some have argued that clinical reasoning is predominantly intuitive due to the exigencies and time pressures of clinical practice,⁹ or exists along a continuum of cognitive patterns with differing judgement speeds and resource dependence.¹⁰ Regardless, it is nonetheless important to understand the intrinsic differences in cognitive thought processes in clinical reasoning, which can affect decisional accuracy and efficiency.

For example, in clinical practice, expert physicians who are experienced diagnosticians often have ‘illness scripts’ (mental representations of disease conditions and manifestations) that they can quickly draw upon to make ‘spot diagnoses’ intuitively.⁹ However, when discrepancies between an initial intuitive response and incoming external cues occur, a seasoned clinician would then revert to a more systematic approach and recalibrate his/her diagnostic considerations.¹¹ In fast-paced, high-acuity clinical settings, clinical heuristics are often employed for its efficiency and may be adequate to achieve decision-satisficing objectives.^{12, 13} Examples of heuristics include medical mnemonics to remember causes of clinical presentations, Occam's razor (finding a unifying diagnosis), Sutton's law (choose investigations that are most likely to yield the diagnosis; ‘go where the money is’), common conditions occur commonly (‘if you hear hoof beats, don't think zebras’), ruling out of worst-case scenario (ROWS; always come up with dangerous differential diagnoses first to exclude actively) or Casablanca strategy (evaluating for standard ‘suspects’ in common clinical manifestations).^{2, 14} Yet the use of heuristics need be carefully balanced with the risk of cognitive errors such as biases and sub-optimal decision-making. For example, certain medical mnemonics (e.g. the 5 ‘Fs’ (fat, fertile, forty, female and fair) risk factors for gallstone disease) can perpetuate implicit bias or premature diagnostic closure by failing to consider other differential diagnoses,¹⁵ while Occam's razor is not always true when there can be multiple concurrent diagnoses explaining patient's clinical manifestations, especially in those with complex multimorbidities. Importantly, adopting a Casablanca strategy of sending a battery of ‘standard’ tests for clinical presentations such as fever without relying on clinical judgement can lead to inappropriate/over-testing, which is undesirable.

In addition, emotions experienced in clinical encounters can easily influence rational decision-making, such as natural affective responses to patients' experiences or expressed emotions (e.g. empathy, countertransference or distress), inter-disciplinary tensions (e.g. conflicting professional opinions, power differentials) or dealing with ethical/moral dilemmas.¹⁶ For example, in physicians who care for seriously/terminally ill patients, it was found that unexamined emotions could lead to incoherent care goals and poor medical decision-making.¹⁶ In the ethical realm, Haidt's social intuitionist model suggests that intuitions/emotions may directly underlie moral reasoning as he found that individuals commonly perceive/feel that an action is wrong before coming up with explanations to justify why this is so.¹⁷

Finally, situational, contextual or environmental parameters may influence clinicians' decision-making, such as time, resource or information constraints,¹⁸ or current stressors, temporally recent experiences and other specific challenges that could trigger bias and irrational thinking.¹⁹ Recently, two major situativity theories – situated cognition (SitCog) and ecological psychology (EcoPsych) – have emerged to describe the dynamic interplay between individual cognitive reasoning processes and the environment.^{20, 21} In SitCog, ‘the outcome of cognition is based on the specifics of the situation’, where patient factors, provider factors and the context/environment of the clinical encounter collectively influence eventual decision-making.²⁰ In EcoPsych, however, cognition is seen as ‘a product of the relationship between the individual clinician and their surroundings’, where supportive elements (known as ‘affordances’) may trigger useful perception or actions by clinicians who have the necessary abilities (known as ‘effectivities’) to act on the affordances.²⁰

Since the 1980s, psychologist Gary Klein and his team have performed groundbreaking work in naturalistic decision-making by studying career professionals working in high-stakes, time/resource-constrained and volatile environments.²² They found that these expert professionals were only able to make good decisions under such circumstances based on a combination of intuitive and analytical cognitive processes by accumulating enough real-world practice experience.²² Klein subsequently summarised this in a recognition-primed decision model, where expert decision-makers can intuitively recognise a problem, generate expectancies and mental simulations of how different courses of action will likely pan out and, then, based on incoming signals/cues that match expectancies or generate discrepancies, actively adapt and recalibrate thinking.²²

In the same vein, the traditional forms of clerkship training involving direct patient care and interactions are indispensable in cultivating clinical reasoning prowess. Broadly, clinical clerkship attachments are most effective when trainees are immersed in an iterative process of direct observation, deliberate practice and receiving feedback, with further consolidation of learning achieved through dialectical methods of instruction such as Socratic/clinical questioning,²³ chart review or chart-stimulated recall²⁴ and mini-CEX (clinical evaluation exercise) assessments.²⁵ In recent years, simulation-based medical education with standardised patients and objective structured clinical examinations has become increasingly popular pedagogical modalities as they are practical, convenient, provide a safe learning environment and enable standardisation and objectivity in assessments. Nonetheless, simulated practice cannot, by itself, replace real-world experiences of dealing with time, resource or information constraints, physical and emotional demands or situational volatility.

While adverse clinical outcomes (e.g. treatment complications, patient morbidity and mortality) are often the triggers for a malpractice complaint or lawsuit to be filed, the medicolegal process of evaluating the merits of the case ultimately hinges on the quality of clinical decision-making. This means that from a strictly medicolegal perspective, the professional standard of care is adjudged by the clinical reasoning processes that led to specific medical decisions. The Bolam-Bolitho test is the classical medicolegal framework applied to evaluate malpractice/negligence claims based on whether the medical decision or action undertaken is supported by the opinions of a respected body of peers and whether the proclaimed standard of care by the defendant doctors is truly logical and reasonable in the given clinical context, evidence-based and passes a risk–benefit analysis with competing alternatives.²⁶

To illustrate this, we look at a recent example of a medical malpractice case that was brought before the General Division of the High Court of Singapore, where the plaintiff alleged negligence in the clinical care of an elderly patient who developed a sudden in-hospital cardiac arrest leading to demise.²⁴ In this case, the court ruled in favour of the defendants because they were able to clearly articulate, with proper clinical documentation, the rationale for their medical decisions (for instance, the decision to hold off a cardiology referral for urgent coronary angiogram was justified on the basis of a clinical impression of type 2 myocardial infarction precipitated by sepsis as opposed to a type 1 athero-occlusive cause of myocardial infarction), which were supported by fellow professional experts and deemed reasonable and logical in the given clinical context.²⁷

In another example, this time involving a case of professional misconduct, a psychiatrist had issued a number of prescriptions that did not comply with the Ministry of Health (MOH) guidelines, and the patient passed away 4 days after the last of these prescriptions was issued.²⁸ The Court of Three Judges of the General Division of the High Court of Singapore noted that there were situations where medical professionals can depart from MOH guidelines, but they must prove that: (i) they considered the rationale for the guidelines and concluded that the deviation was justified after conducting a risk–benefit analysis; (ii) the decision is objectively defensible in the circumstances, as per the prevailing test for medical negligence; and (iii) the deviation is discussed with the patient and consent has been obtained (at least in certain circumstances).²⁸ The court highlighted that the medical practitioner bears the evidential burden of proof here as this ‘strikes the correct balance between protecting the well-being of the patient from practices which are known to carry a high level of risk, and affording medical practitioners the flexibility to innovate and tailor their treatment to the specific needs and challenges of each case with which they are confronted’.²⁸ While the court accepted the defendant's explanations in respect of some of his deviations from the guidelines, it rejected his explanations for prescribing multiple concurrent benzodiazepines despite knowing the patient was taking opioid analgesics, as well as prescribing mirtazapine and zolpidem CR beyond their respective maximum dosages. His explanations were criticised for being overly general (e.g. that the patient had a complex psychiatric condition, or that he knew the patient's ‘functioning’) and for not identifying what the specific benefits were and how they outweighed the risks in that case – in other words, the court found fault with the quality of his clinical reasoning.²⁸

In summary, clinical reasoning is the practice-defining feature of the medical profession, which is characterised by sophisticated cognitive processes, intertwined with psychoemotional elements and situational/contextual attributes that collectively influence decision-making. The cultivation of clinical reasoning in medical training is achieved through deliberate practice and dialectical instruction in naturalistic settings, but can be supported with simulated learning. Finally, the importance of clinical reasoning is underpinned by its function as the medicolegal yardstick of the professional standard of care.