# Diagnostic accuracy and metacognition in dermatology: A cross-sectional analysis of confidence and decision-making

## TL;DR

This study explores how confidence and decision-making in dermatology vary with experience and affect diagnostic accuracy.

## Contribution

The study identifies how metacognitive skills and experience influence dermatologic diagnostic accuracy and decision efficiency.

## Key findings

- Diagnostic accuracy and confidence increase with dermatologic experience.
- Board-certified dermatologists were most accurate when responding quickly, unlike residents.
- Medical students showed overconfidence and poor alignment between confidence and decisions.

## Abstract

Diagnostic accuracy in dermatology requires both visual expertise and metacognitive skills such as confidence, calibration, and decision-making under uncertainty. Miscalibration may lead to diagnostic error, unnecessary testing, and unsafe management.

Assess how confidence, response time, and decision behavior vary across levels of dermatologic experience and how these factors relate to diagnostic accuracy and efficiency.

This cross-sectional study included 68 participants (medical students, resident physicians, and board-certified dermatologists) who completed a multiple-choice diagnostic task with 50 diverse dermatologic images. Participants selected a diagnosis option, rated their confidence, and decided regarding additional inspection.

Diagnostic accuracy and confidence increased with experience. Board-certified dermatologists were most accurate when responding quickly, but not after longer deliberation, which did not hold true for residents. Medical students displayed significant overconfidence and poor alignment between confidence and decisions. Across all groups, 12% of melanomas were dismissed/overlooked.

Small sample size limits subgroup comparisons. A simulated setting may not fully capture clinical complexity.

Metacognitive skills differ by experience and influence diagnostic accuracy and efficiency. Training should support calibration and adaptive decision-making.