The Development and Evaluation of an Alternative Abbreviated Nutrition Risk Screening Tool Using Item Response Theory

BACKGROUND AND OBJECTIVES: Brief tools are needed for nutrition risk screening in clinical practice and community settings. A previously developed short tool, SCREEN-8, although widely used in practice and research, may not be specific enough, resulting in overidentifying risk. This study aimed to identify the best-minimum set of items from SCREEN-14 (Seniors in the Community Risk Evaluation for Eating and Nutrition) to determine nutrition risk in older adults. RESEARCH DESIGN AND METHODS: Data from participants who had completed an online version of SCREEN-14 (n = 20,093) were available for analysis. Item Response Theory (IRT), using the Graded Response Model, was applied to identify a subset of eight items for all participants, as well as those over the age of 75 years, with the resulting item sets compared. The eight items identified for the 75+ year subsample were selected for consideration in the abbreviated version; scoring remained consistent with SCREEN-8 for comparison. RESULTS: The proportion at risk based on SCREEN-8 was 42.1% whereas the proportion at risk with SCREEN-IRT was 27.7%. Items in SCREEN-IRT demonstrated a good model fit (RMSEA = 0.075; CFI = 0.905). SCREEN-8 also demonstrated a good/acceptable model fit (RMSEA = 0.064; CFI = 0.913). DISCUSSION AND IMPLICATIONS: A new 8-item version of SCREEN was created with IRT. This tool has different items than SCREEN-8 and resulted in a lower proportion of adults as potentially at nutrition risk. Further research on criterion validity of SCREEN-IRT is needed.