The purpose of this study was to determine if the Debriefing for Meaningful Learning Inventory© (DMLI) is a valid measure of the debriefing method Debriefing for Meaningful Learning© (DML).
Background
The evidence of improved learner outcomes through simulation debriefing has increased in the literature as the use of simulation has proliferated throughout nursing education (Hayden, Smiley, Alexander, Kardong-Edgren, & Jeffries, 2014). Because of the significance of the learning that occurs during debriefing (Shinnick, Woo, Horwich, & Steadman, 2011), the National Council of State Boards of Nursing (NCSBN) (Alexander et al., 2015) and the International Nursing Association for Clinical Simulation and Learning (INACSL Standards Committee, 2016) have recommended that debriefers receive formal training in a theory-based debriefing method to ensure consistent learning outcomes. However, it is unknown how debriefing training improves the ability to apply a debriefing method with learners. Indeed, because of the lack of valid instruments to measure the application of specific debriefing methods, it is not known how a debriefer enacts an evidence-based debriefing method during debriefing.
Instrument Development
DML is a theory-based debriefing method adopted widely across nursing education because of the improved clinical reasoning demonstrated with nursing students (Dreifuerst, 2012). DML facilitates guided reflective thinking using Socratic questioning through the phases of the six E's: engage, explore, explain, elaborate, evaluate, and extend. The Debriefing for Meaningful Learning Evaluation Scale© (DMLES) was developed as a behaviorally anchored rating scale. In psychometric testing, the DMLES demonstrated internal consistency (Cronbach’s alpha = 0.88), face validity, and content validity (scale-level CVI 0.92) (Author & Dreifuerst, 2016). The DMLI was developed from the DMLES to explore how a debriefer describes their understanding of the central concepts of DML, and their application of the behaviors of DML during debriefing. While the DMLES assesses observed DML debriefing behaviors, the DMLI was developed to further understand how a debriefer understands DML and subsequently applies the method with learners. Therefore, the 31-item DMLES was modified into the 57-item DMLI to comprehensively explore a debriefer’s understanding and application of DML. Five items of the DMLI assess the understanding of the DML central concepts with binary options of yes or no. Fifty-one DMLI items describe application of DML debriefing behaviors, scored with ordinal frequency options of always, sometimes, and never.
Methods
A confirmatory factor analysis was performed to determine if the 51 application items of the DMLI is a valid measure of DML. A latent class factor analysis (LCFA) was conducted since LCFA is the most common model-based clustering method used with discrete data (Dean & Rafferty, 2010). LCFA was used to identify latent case subtypes from the DMLI data set, and estimate parameter values for the model, thereby confirming the measurement theory of DML. The bootstrapping approach was used, requiring no assumptions of the DMLI data. To assess model fit, cross-classification frequencies were compared to the expected frequencies predicted by the model.
Sample
Known debriefers and members of INACSL were recruited for participation in this study. The DMLI was completed by 234 nurse educators who reported having received debriefing training, and facilitated simulation debriefing with baccalaureate prelicensure nursing students.
Results
The DFactor model within Latent GOLD© 5.1 (2015) was used to estimate cluster models within the DMLI data. The latent class approach to the DMLI data supported a six-class DFactor model, confirming the measurement theory of the six E’s of DML. The six class DFactor model provided a good fit to the DMLI data, L-squared = 7.0803 with 85 degrees of freedom; p = 0.298. The Bayesian information criterion (BIC) also indicated the preferred model was the six-class DFactor model (BIC = 6630.79). The factor loadings and commonalities were evaluated, and the factors were ordered according to R-squared, indicating how well the model predicts the DFactor score. R-squared for the six DFactors ranged from 0.77 to 0.91, indicating the amount of variance explained in the items of each DFactor.
Conclusions
LCFA was used to confirm the item groupings of the 52 DMLI application items. Each of the six DFactors correlated with the six E’s of DML; each of the 52 DMLI application items loaded onto one of the six DFactors. Findings from the LCFA demonstrated that the 52 DMLI application items did yield a model of good fit, indicating that the DMLI is a valid measure of the application of DML. This work is significant to the nursing profession by contributing a tested valid instrument for use in assessing the application of a debriefing method. Assessment of how evidence-based debriefing methods are translated into teaching practice is needed to advance the science of debriefing practice.