Diagnosis in psychiatry: Biases and humility

To the Editor

The case by Weiß and Kluge (2020) raises salient matters for clinical practice, including cognitive biases, diagnostic approaches, and the medical training of psychiatrists.

Psychiatrists rely on expert intuition (type 1 thinking) and expert analytical skills (type 2 thinking) when making a diagnosis (Parker, 2018). Reliance solely on the former can result in clinical errors, as seen with Ms A where availability bias (the diagnosis of dissociation instead of cataplexy), anchoring bias and overshadowing (delayed consideration of a neurological disorder) and confirmation bias (adherence to the wrong preliminary diagnosis) may have occurred.

Some diagnostic errors are inevitable in mental healthcare where psychiatrists must make subjective clinical assessments under conditions of uncertainty. However, they may be minimised by empowering oneself with an understanding of heuristics and developing robust diagnostic approaches that guard against overreliance on these; maintaining a broad medical knowledge base and acknowledging our proneness to err; and seeking a second opinion early where indicated.

Psychiatric diagnosis is generally understood to be an exercise in Bayesian thinking or pattern matching. The Bayesian approach considers prior probability of a certain psychiatric diagnosis and adjusts this probability during the assessment to arrive at a final diagnosis. In some data-driven medical specialties, this is a powerful approach. However, in psychiatry where relevant probabilities are not readily available, this approach is problematic.

Pattern matching, checking off symptoms against a Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) list, is formulaic and can lead to premature closure on one diagnosis (applying Occam’s razor where the most parsimonious explanation of a patient’s symptoms is considered the best) or multiple diagnoses (applying Hickam’s dictum – the counterargument to Occam’s razor) due to expanding diagnostic categories in taxonomic and diagnostic tools.

Explanatory coherence is a mental and logical process that involves evaluating many pieces of evidence obtained during the patient assessment and considering how they support or contradict working hypotheses (Thagard and Larocque, 2018). This approach utilises a hypothesis-driven interview. It relies on adequate knowledge and clinical skills to construct a list of differential diagnoses and adequately assess for them. This leads me to a final reflection.

Failure to develop general medical knowledge and skills can contribute to diagnostic errors as seen with Ms A. In Australia, it is possible to commence psychiatry training immediately following internship. It is unlikely that doctors commencing psychiatric training in PGY2 (postgraduate year 2) have the requisite medical knowledge or skills to consider broader non-psychiatric differential diagnoses and manage comorbid medical illness in psychiatric patients.