Conditions-Based Learning Theory as a Framework for Comparative-Effectiveness Reviews: A Worked Example

Phenomenon: An evidence-informed era of medical education encourages the generation and use of comparative-effectiveness reviews, yet the reviews often conclude, curiously, that all instructional approaches are equally effective. Approach: We used a conditions-based learning theory to structure a review of the comparative-effectiveness literature on electrocardiogram instruction. We searched MEDLINE, EMBASE (Ovid), ERIC (Ovid), PsycINFO (Ovid), and CINAHL (EBSCO) from inception to June 2016. We selected prospective studies that examined the effect of instructional interventions on participants' knowledge and skill with electrocardiogram interpretation. Two reviewers extracted information on the quality of the studies, the effect of instruction on the acquisition of knowledge and skill, and instructional quality. Instructional quality is an index of the extent to which instruction incorporates 4 practices of Gagne's conditions-based learning theory: presenting information, eliciting performance, providing feedback, and assessing learning. Findings: Twenty-five studies (3,286 participants) evaluating 47 instructional interventions were synthesized. The methodological quality of most studies was moderate. Instructional quality varied: All interventions presented information and assessed learning, but fewer than half elicited performances or provided feedback. Instructional interventions that incorporated all 4 components improved trainees' abilities considerably more than those that incorporated 3 or fewer; respectively, standardized mean difference (SMD) = 2.80, 95% confidence interval (CI) [2.05, 3.55], versus SMD = 1.44, 95% CI [1.18, 1.69]. Studies that compared "innovative" to "traditional" types of instruction did not yield a significant pooled effect: SMD = 0.18, 95% CI [-0.09, 0.45]. Insights: The use of a conditions-based learning theory to organize the comparative-effectiveness literature reveals differences in the instructional impact of different instructional approaches. It overturns the unlikely, but common, conclusion that all approaches are equally effective.