Teaching Bayesian analysis to Emergency Medicine residents

Area of Science:

• Medical Education
• Evidence-Based Medicine (EBM)
• Emergency Medicine (EM)

Background:

• Accurate estimation of post-test probability for pulmonary embolism (PE) is crucial for effective clinical decision-making in Emergency Medicine.
• Existing performance gaps in applying key Evidence-Based Medicine (EBM) concepts, such as likelihood ratios, necessitate targeted educational interventions.

Purpose of the Study:

• To evaluate the impact of a brief didactic lecture on Emergency Medicine (EM) resident performance in applying a core Evidence-Based Medicine (EBM) concept.
• To assess the improvement in EM residents' ability to estimate post-test pulmonary embolism (PE) probability after an educational intervention.

Main Methods:

• A prospective, before-and-after study design was employed to assess EM resident performance.
• Residents' estimates of post-test PE probability were evaluated before and after a didactic lecture on Fagan's Nomogram and likelihood ratios (LRs).
• The study provided sensitivity and specificity data for computed tomography (CT) and D-dimer tests.

Main Results:

• The absolute percentage difference between resident estimates and true post-test PE probabilities decreased significantly from 14.5% to 4.5% after the intervention.
• This represents a 10% improvement in accuracy, achieving statistical significance (p = 0.002).
• Confidence intervals for the pre-intervention difference were 9.7%-19.9%, and for the post-intervention difference were 2.0%-6.8%.

Conclusions:

• A brief lecture is an effective method for teaching Evidence-Based Medicine (EBM) concepts to Emergency Medicine (EM) residents.
• The educational intervention demonstrably improved residents' accuracy in calculating post-test probability for pulmonary embolism (PE).