Effects of teaching evidence based medicine on physician assistant students' critical appraisal knowledge, self-efficacy and clinical application: A randomized controlled trial

Background -Teaching evidence-based medicine (EBM) has been widely incorporated into both physician and physician assistant (PA) education. However, gaps in the physician assistant EBM education literature include: limited grounding of EBM outcome measures in a model of EBM competence; limited use of randomized controlled designs; using outcome measures with unknown validity or reliability; uncertainty whether gains in the classroom translate into evidence seeking behavior in the clinic. Objective- Using a randomized controlled design, we explored whether providing EBM training to PA students has an effect on their EBM knowledge, skills, self-efficacy and evidence seeking behavior in a clinical situation.;Methods - Trial design- A single center, randomized, controlled, partially blinded, within and between groups trial. Participants- 61 self-selected first year physician assistant students at a mid-west Physician Assistant Program. Randomization- Participants were assigned to intervention and control (wait) groups by simple randomization. Blinding - Participants and course instructors were un-blinded; Scoring of outcome measures and data analysis were blinded Interventions - The intervention group received a standard PA didactic curriculum plus EBM training. The control (wait) group received the standard PA curriculum only. Outcome measures- The Fresno Test, a short answer validated measure of EBM knowledge and skills; the Self-Efficacy Scale, a validated Likert scale measure of EBM self-efficacy; and the Patient Encounter Clinical Application (PECA) test, a measure of evidence seeking behavior with a standardized (video) clinical patient scenario.;Results - Eighty-one first year PA students were eligible with sixty-one students self -selecting to participate. Thirty participants were randomized to the experimental group and thirty-one participants to the control group. One participant dropped out of the control group before data collection. Thirty students were analyzed in each group on an intention to treat basis.;Outcomes - On the Fresno, PA students who received EBM training (n = 30; M= 116.02; SD = 20.13) performed significantly better than PA students without EBM training (n = 30; M=91.20; SD = 19.16): independent t (58) = -4.89, p < .001 (95% CI = -34.97; - 14.66). Participants also performed significantly better on the Fresno after they received EBM training (M = 116.02; SD = 20.13) as compared to before they received EBM training (M = 92.17; SD = 20.80): dependent t = - 5.9, p < 0.001 (95 % CI = - 32.02; - 15.68). On the self-efficacy scale, PA students who received EBM training (n = 30; Mdn = 84.00; range = 59 -97) performed significantly better than PA students without EBM training (n = 30; Mdn = 55.5; range = 19 -84: U= 44.50, z = -6.00, p < .001). Participants performed significantly better on the self-efficacy scale after they receive EBM training (Mdn = 84; range = 59 -97) as compared to before they received EBM training (Mdn = 59.5; range 33 - 87): z = - 4. 78, p < 0.001. On the PECA, PA students who received EBM training perform no differently (n=30; Mdn = 8.9; range = .77 -45.71) than PA students without EBM training (n= 30; Mdn = 10; range = .91 - 52.5. U = 428.00, z = - 0.11, p > 0.05. Linear regression demonstrates that increasing Fresno pretest scores were significantly related to decreasing percent changes in the Fresno posttest scores: F - ratio =38.96; p < 0.001.;Conclusion -Teaching evidence based medicine to PA students appears to significantly improve their EBM knowledge, skills and self-efficacy, both within the same students (before and after) and when compared to PA students who have not received EBM training.;Improvements in EBM knowledge, skills and self-efficacy, do not appear to translate into significant improvement in EBM application in a simulated clinical patient encounter. Also, PA students with the strongest pre- intervention EBM knowledge and skills seemed to benefit least from EBM training suggesting a possible "law of diminishing return" for EBM education. Implications of these finding for EBM education and research are explored.