Development and psychometric analysis of the Diabetes Device Confidence Scale for school nurses.

Abstract

School nurses need to be equipped to help students with diabetes devices. No existing tools assess school nurse self-efficacy in using devices. To develop and evaluate the psychometric properties of a novel scale to measure school nurse confidence with diabetes devices. We generated a list of items with community partners and examined logical validity. We then revised and distributed the item set to school nurses in Pennsylvania to examine aspects of structural validity, convergent validity, and internal consistency reliability. We used item response theory to refine the scale. Facilitated discussion with collaborators generated an initial list of 50 potential items. Based upon the item-content validity index, we revised/eliminated 13 items. School nurses (n= 310) in Pennsylvania completed an updated 38-item scale; the majority had experience with insulin pumps or continuous glucose monitors. Exploratory factor analysis identified a one-factor solution, suggesting a unidimensional scale. We eliminated 13 additional items based upon significant inter-item correlation or skewed response patterns. Item response theory did not identify additional candidate items for removal. Despite a high degree of redundancy (Cronbach's alpha > 0.90), we retained all remaining items to maximize the future utility of the scale. The 25-item, unidimensional Diabetes Device Confidence Scale is a new tool to measure school nurse confidence with diabetes devices. This scale has future clinical, programmatic, and research applications. Combined with knowledge assessments, this scale can serve to evaluate school nurse device use readiness, assess training gaps, and tailor interventions.