Adaptive instruction and learner interactivity in online learning: a randomized trial

The purpose of this study was to evaluate two online instructional design features, namely adaptation to learner prior knowledge and use of questions to enhance interactivity in online portrayals of physician–patient encounters, in the context of instructing surgical specialists to deliver perioperative tobacco interventions. An online learning module on perioperative tobacco control was developed, in formats incorporating permutations of adaptive/non-adaptive and high/low interactivity (i.e., 2 × 2 factorial design). Participants (a national sample of US anesthesiology residents) were randomly assigned to module format. Primary outcomes included tobacco knowledge, time to complete the module, and self-efficacy in delivering tobacco interventions. One hundred fourteen residents completed the module, which required a median of 60 min (interquartile range 49, 138). The difference in post-module tobacco knowledge score was similar for adaptive and non-adaptive formats [mean difference 0.3 of 10 possible (95% CI − 0.3, 1.0), p = 0.25] but time was shorter for the adaptive format [− 7 min (95% CI − 14, 0), p = 0.01] and knowledge efficiency (knowledge score divided by time) was higher [0.08 units (95% 0.03, 0.14), p = 0.004]. The level of interactivity had no significant effect on self-efficacy [− 0.1 on a 5-point scale (95% CI − 0.3, 0.1), p = 0.50] in delivering tobacco interventions (both outcomes using 5-point scales). Adapting online instruction to learners’ prior knowledge appears to improve the efficiency of learning; adaptation should be implemented when feasible. Adding features that encourage learner interaction in an online course does not necessarily improve learning outcomes.