Initial diagnostic hypotheses bias analytic information processing in non-visual domains.

Area of Science:

• Medical Education
• Cognitive Psychology
• Clinical Reasoning

Background:

• Initial diagnostic hypotheses are known to bias automatic information processing.
• The effect of initial hypotheses on analytic information processing remains less understood.
• Understanding this bias is crucial for improving diagnostic accuracy in clinical practice.

Purpose of the Study:

• To investigate the impact of an initial diagnostic hypothesis on analytic information processing in medical residents.
• To assess how clinical experience, specifically the frequency and urgency of alternative diagnoses, influences diagnosis selection.

Main Methods:

• 19 subspecialty medical residents diagnosed three clinical presentations (dyspnoea, headache, chest pain) during an objective structured clinical examination.
• Participants were randomly assigned cases with either a correct or incorrect initial diagnostic hypothesis.
• For incorrect hypotheses, the alternative diagnoses varied in clinical frequency and urgency.

Main Results:

• Correct initial hypotheses were always retained, while incorrect hypotheses were retained in only 10.9% of cases.
• Correct initial hypotheses led to correct diagnoses in all instances, versus 65.2% for incorrect hypotheses (Risk Ratio 1.5).
• Neither the clinical frequency nor the urgency of alternative diagnoses was associated with the selection of a different diagnosis.

Conclusions:

• An initial diagnostic hypothesis demonstrably biases analytic information processing during diagnosis.
• The data used to refute an initial hypothesis appear to be the primary driver for selecting an alternative diagnosis.
• Future research should focus on strategies to generate more accurate initial hypotheses or mitigate diagnostic bias.