A randomized controlled trial of SAFMEDS to improve chest radiograph interpretation among medical students

Rationale and objectives: Junior doctors find chest radiograph (CXR) interpretation challenging, and commonly make diagnostic errors. The aim of this study was to evaluate the effectiveness of SAFMEDS in teaching un-dergraduate medical students to identify important chest abnormalities in radiology imaging. Materials and methods: A pragmatic randomized controlled trial design was utilized. Third-year medical students were randomly assigned to an intervention group (n = 20), who received the SAFMEDS intervention, or a control group (n = 20), who did not receive the intervention. Three participants (one intervention, two control) with-drew from participation. Percentage accuracy in CXR interpretation was assessed at three timepoints (baseline, post-test, and retention). A series of one-way between-subjects' analyses of covariance, with percentage accuracy at the pre-test timepoint entered as the covariate, were conducted. Results: Large effect sizes of the SAFMEDS intervention were observed at post-test and retention (eta(2) = 0.67, eta(2) = 0.58 respectively), with the intervention group demonstrating significantly higher percentage accuracy in CXR interpretation as compared to the control group at both timepoints. Intervention group performance decreased by 5.26% (SD = 9.80) from post-test to retention. Twelve intervention group participants (63.2%) met the fluency criteria. Despite large effect sizes, there were no significant differences in post-test or retention performance between fluent and non-fluent participants (eta(2) = 0.17 and eta(2) = 0.2 respectively). Conclusion: SAFMEDS offers an effective adjunct to usual teaching. Future research could focus on examining the effect of fluency with a larger sample, in addition the expansion of the SAFMEDS approach, and its' application to other health profession populations.