A first look at undergraduate nursing students' knowledge of evidence-based practice using the evidence-based practice knowledge assessment in nursing

Abstract

<p>Session presented on Friday, July 24, 2015:</p> <p><strong>Purpose:</strong> The quantitative relationship between self-reports of competence and objectively measured performance is generally &lt; r = .3 and in a striking number of studies, inverse relationships have been observed. With the slow rate of adoption of EBP in clinical environments, an objective measure of nurses' EBP knowledge is clearly needed. Here we report initial validity and reliability evidence for a new objective measure of EBP knowledge, the Evidence-Based Practice Knowledge Assessment in Nursing (EKAN).</p> <p><strong>Methods:</strong> Seven subject-matter experts reviewed the candidate items, culminating in a final item pool of N = 75 items with an S-CVI = .94. Rasch modeling (1PL item-response theory [IRT]) with jMetrick (Meyer, 2014) was used to evaluate psychometric performance on the theorized unidimensional trait of EBP knowledge. Candidate items were tested in N = 200 undergraduate nursing students from two large Midwestern nursing education programs. Subjects were M = 24.8 (SD = 5.3) years old and 90.5% female.</p> <p><strong>Results:</strong> For the final, 20-item EKAN, mean difficulty was .19 (Range -2.0 - 2.8), weighted mean square infit was 1.01 (Range .95 - 1.06), standardized weighted mean square infit was .33 (Range -.7 - 1.6), unweighted mean squares outfit was 1.02 (Range .93 - 1.14), standardized unweighted mean squares outfit was .34 (Range -1.08 - 2.00). For the scale, the item separation index was 7.05 and the person separation index was 1.66. Item reliability was .98; person reliability was .66. These values reflect strong item performance but highlight trait homogeneity in the subject pool. EKAN scores ranged from 5 - 16 (of a possible 20); M = 10.4 (SD = 2.31). A known-group effect was observed (M = 10.01 vs. 11.47; t = -2.53, p = .01) when comparing scores from subjects recently exposed to vs. unexposed to prior EBP, research, or statistical coursework.</p> <p><strong>Conclusion:</strong> The 20-item EKAN showed strong evidence of trait unidimensionality and desirable scale psychometrics when evaluated using the Rasch model. Additional studies are in progress among groups possessing a greater range of EBP knowledge to provide additional validity evidence. The EKAN is an efficient objective EBP knowledge measure available to educators and researchers.</p>