P695 - USING ITEM RESPONSE THEORY TO IDENTIFY ESSENTIAL ITEMS OF SCREEN-14 TO DETERMINE NUTRITION RISK IN COMMUNITY-LIVING OLDER ADULTS

P695

USING ITEM RESPONSE THEORY TO IDENTIFY ESSENTIAL ITEMS OF SCREEN-14 TO DETERMINE NUTRITION RISK IN COMMUNITY-LIVING OLDER ADULTS

H. Keller^1,2,*, M. Jackson¹, M. Holmes³, M. Mourtzakis¹, S. Kirkpatrick³

¹Kinesiology and Health Sciences, University of Waterloo, ²Schlegel-UW Research Institute for Aging, ³School of Public Health Sciences, University of Waterloo, Waterloo, Canada

Rationale: Identification of nutrition risk in older adults is beneficial to improving nutritional status and preventing adverse outcomes.  Simple, efficient screening tools are needed. SCREEN-14 is currently 14 main and 3 sub-items, which is too long for routine screening in primary care. The purpose of this study was to use Item Response Theory to develop a shorter version of SCREEN-14.

Methods: An online version of SCREEN-14 was completed by 20,093 Canadians. The Graded Response Model, suitable for items with ordinal response options, was used to identify items based on their ability to differentiate between levels of nutrition risk. Items that provided greater measurement value were considered for inclusion in two candidate abbreviated versions: one derived from the full sample and one based on respondents aged 75 and older. This latter model was more inclusive of key risk factors and was selected by the research team for use in model testing. 

Results: The abbreviated version included eight items: weight change, intentionality of weight change, skipping meals, appetite, vegetable and fruit intake, fluid intake, difficulty chewing, and challenges with grocery shopping. The Test Information Function indicated that the eight selected items provided strong measurement precision across levels of nutrition risk, particularly among individuals at higher risk (i.e., those with lower scores).  Model fit indices confirmed adequacy of the abbreviated version (M2 = 2247.87 df=20, p<0.001; RMSEA= 0.075; CFI= 0.905; TLI=0.868).

Conclusion: This study identified items to produce a shortened version of the SCREEN-14 that appropriately scaled respondents according to their level of nutritional risk. This abbreviated version requires further testing and specifically criterion validation with a dietitian assessment of risk but holds promise as a potential short tool for use in clinical practice.

Disclosure of Interest: None declared