Medline for Medical Students? Searching for the Right Answer

Objective: Does MEDLINE use, when added to more traditional sources of information, improve the accuracy of medical students' clinical decisions when compared to those obtained using traditional sources only? Design: Randomized control trial. Setting: McMaster University Faculty of Medicine, Undergraduate Program. Participants: The entire class of 101 medical students, class of 1998. Overall response rate on 9 items was 56% (510/909), with 35% (35/101) completing all 9 items. Intervention: All participants were randomized on each of the nine clincal scenarios for which the student could choose to apply, or refrain from applying, a proferred intervention. When randomized to the control arm, the student used traditional sources of information for decision-making. When randomized to the experimental arm, the student used MEDLINE searching in addition to more traditional sources of information, for decision-making. Main Outcome Measures: Prior to, and subsequent to the information search, the students indicated their comfort in using the proffered intervention on a seven point Likert scale. Results: Analyzed with one-way ANOVA, the mean rating post-search of the control non- MEDLINE arm was 2.94 (SD = 1.80) (where 1 = correct, 7 = incorrect) and of the experimental MEDLINE arm was 2.71 (SD = 1.81), (F(1,522) = 2.03, p = 0.15 n.s). The mean change of the control arm was 0.97 (SD = 2.04) and of the experimental arm was 1.008 (SD = 1.92), (F(1,511) = 0.04, p = 0.84n.s.). Conclusions: The addition of MEDLINE to more traditional answer-seeking behaviors by medical students does not translate into a beneficial impact on clinical decision-making.