Pretest Scores Uniquely Predict 1-Year-Delayed Performance in a Simulation-Based Mastery Course for Central Line Insertion. Issue 3 (June 2018)

• Record Type: Journal Article
• Authors: Diederich, Emily; Thomas, Laura; Mahnken, Jonathan; Lineberry, Matthew
• Is Part Of: Simulation in healthcare. Volume 13:Issue 3(2018)
• Publication Date: 2018-06
• Subjects: Simulation -- Pretesting -- Mastery learning -- Testing effect -- Assessment for learning -- Procedural skill; Simulated patients -- Periodicals; 362.1
• DOI: 10.1097/SIH.0000000000000327
• Languages: English
• ISSNs: 1559-2332

Abstract

Abstract : Introduction: Within simulation-based mastery learning (SBML) courses, there is inconsistent inclusion of learner pretesting, which requires considerable resources and is contrary to popular instructional frameworks. However, it may have several benefits, including its direct benefit as a form of deliberate practice and its facilitation of more learner-specific subsequent deliberate practice. We consider an unexplored potential benefit of pretesting: its ability to predict variable long-term learner performance. Methods: Twenty-seven residents completed an SBML course in central line insertion. Residents were tested on simulated central line insertion precourse, immediately postcourse, and after between 64 and 82 weeks. We analyzed pretest scores' prediction of delayed test scores, above and beyond prediction by program year, line insertion experiences in the interim, and immediate posttest scores. Results: Pretest scores related strongly to delayed test scores ( r = 0.59, P = 0.01; disattenuated ρ = 0.75). The number of independent central lines inserted also related to year-delayed test scores ( r = 0.44, P = 0.02); other predictors did not discernibly relate. In a regression model jointly predicting delayed test scores, pretest was a significant predictor (β = 0.487, P = 0.011); number of independent insertions was not (β = 0.234, P = 0.198). Conclusions: This study suggests that pretests can play a major role in predicting learner variance in learning gains from SBML courses, thus facilitating more targeted refresher training. It also exposes a risk in SBML courses that learners who meet immediate mastery standards may be incorrectly assumed to have equal long-term learning gains.