A Randomised Assessment of Trainee Doctors' Understanding and Interpretation of Diagnostic Test Results

Doctors are unfamiliar with diagnostic accuracy parameters despite routine clinical use of diagnostic tests to estimate disease probability.

Trainee doctors completed a questionnaire exploring their understanding of diagnostic accuracy parameters; ability to calculate post-test probability of a common surgical condition (appendicitis) and their perceptions on training in this area. To determine whether the method of information provision altered interpretation, trainees were randomised to receive diagnostic test information in three ways: positive test only; positive test with specificity and sensitivity; positive test with positive likelihood ratio in layman terms.

326 candidates were recruited across 30 training sessions. Trainees scored a median of three out of seven in questions concerning knowledge of diagnostic accuracy parameters. This was affected neither by training level (P = 0.737) nor by experience in acute general surgery (P = 0.738). 30 (11.8%) candidates correctly estimated post-test probability; with 86.6% overestimating this value. Neither level of training (P = 0.180) nor experience (P = 0.242) influenced the accuracy of the estimate. Provision of the ultrasound scan results in different ways was not associated with likelihood of a correct response (P = 0.857).

This study highlights the deficiencies in trainee doctors' understanding and application of diagnostic tests results. Most trainees over-estimated disease probability, increasing the risk of unnecessary intervention and treatment.