Abstract
OBJECTIVES: To investigate medical students' ability to interpret evidence, as well as their self-assessed understandability, perceived usefulness and preferences for design alternatives in an interactive decision support tool, displaying GRADE evidence summaries for multiple treatment options (Making Alternative Treatment CHoices Intuitive and Trustworthy, MATCH-IT). DESIGN: A combined randomised controlled trial and survey. Participants were presented with a clinical scenario and randomised to one of two versions of the MATCH-IT tool (A/B), instructed to explore the evidence and decide on a recommendation. Participants answered a questionnaire assessing interpretation, treatment recommendation self-assessed understandability and perceived usefulness before exposure to the other MATCH-IT version and asked questions on design preferences. SETTING: Online lecture in an evidence-based medicine (EBM) introductory course. PARTICIPANTS: 149 third-year medical students. 52% (n=77) had 6 months of clinical training and 48% (n=72) had preclinical training only. INTERVENTIONS: The MATCH-IT tool version A uses colour coding to categorise interventions by magnitude and direction of effects and displays all outcomes in a table on entry. Version B has no colour coding, and the user must decide which outcomes to display in the table. MAIN OUTCOME MEASURES: Interpretation of evidence, treatment recommendation, perceived usefulness and understandability, preference for format and design alternatives. RESULTS: 82.5% (n=123) of medical students correctly answered ≥4 out of 5 multiple choice questions assessing interpretation of data. 75.8% (n=114) of students made a treatment recommendation in accordance with an expert panel for the same clinical scenario. 87.2% (n=130) found the tool understandable while 91.9% perceived the tool as useful in addressing the clinical scenario. CONCLUSION: Medical students with no prior training in EBM can interpret and use the MATCH-IT tool. Certain design alternatives were preferred but had no bearing on interpretation of evidence or understandability of the tool.