Probability error in diagnosis: the conjunction fallacy among beginning medical students

Area of Science:

• Cognitive Psychology
• Medical Education
• Decision Science

Background:

• The conjunction fallacy is a cognitive error where individuals estimate the probability of a conjunction of events to be higher than the probability of a single event.
• Physicians are known to be susceptible to this fallacy, impacting clinical judgment and probability estimation.
• Understanding the prevalence of this fallacy in medical students is crucial for early intervention.

Purpose of the Study:

• To investigate the susceptibility of beginning medical students to the conjunction fallacy.
• To determine if medical students exhibit similar cognitive biases in probability estimation as practicing physicians.
• To identify potential educational strategies to address this cognitive error in medical trainees.

Main Methods:

• A cohort of 134 medical students participated in a written exercise involving probability assignments to symptoms and symptom combinations.
• Case vignettes were used to present clinical scenarios, including a case of the common cold.
• Rates of conjunction rule violations were calculated and analyzed across demographic and educational subgroups.

Main Results:

• Nearly half (47.8%) of the beginning medical students demonstrated the conjunction fallacy.
• No significant differences in violation rates were observed between male and female students.
• Prior educational background and statistics coursework did not significantly alter the prevalence of the conjunction fallacy.

Conclusions:

• A substantial proportion of medical students are prone to the conjunction fallacy, indicating a potential for cognitive bias in early medical training.
• While present, the rate of the conjunction fallacy among students was lower than previously reported for physicians, possibly due to less reliance on "representativeness" heuristics.
• Educating medical students about the conjunction fallacy offers a promising approach to reduce this predisposition and improve probability estimation skills.