The University of Iowa College of Nursing, Iowa City, IA, USA.
George Washington University, School of Nursing, Washington, D.C. USA.
West J Nurs Res. 2021 Apr;43(4):374-380. doi: 10.1177/0193945920943898. Epub 2020 Jul 17.
The 18-item Mealtime Engagement Scale was developed to measure mealtime caregiving engagement with preliminary reliability and validity. Item Response Theory models (i.e., Rasch model and Rating Response Model) may provide insight into item functioning. This was a secondary analysis of 87 videotaped mealtime observations involving 18 nursing home staff providing care to residents with advanced dementia.In both models, item difficulties were distributed considerably along the latent trait and highly correlated. Two thirds of the 18 items were located at the moderate level of mealtime engagement. "Providing one-on-one assistance" was most frequently observed, and "re-approaching individual to continue meal" was least frequently observed. All items showed desirable fit to the models. The Rasch model had a significantly smaller deviance than the Rating Response Model, indicating an overall better fit. Findings provided preliminary support for item functioning and pointed out directions for item revisions. Future testing in larger diverse samples is needed.
18 项进餐参与量表旨在测量进餐期间的照护参与度,具有初步的信度和效度。项目反应理论模型(即 Rasch 模型和评分反应模型)可以深入了解项目的功能。这是对涉及 18 名养老院工作人员为患有晚期痴呆症的居民提供护理的 87 个进餐观察视频的二次分析。在这两种模型中,项目难度沿着潜在特质分布差异较大,且相关性很高。18 个项目中有三分之二位于中等水平的进餐参与度。“提供一对一的帮助”观察到的频率最高,“再次接近个人以继续用餐”观察到的频率最低。所有项目都与模型拟合良好。Rasch 模型的离差显著小于评分反应模型,表明整体拟合更好。研究结果初步支持了项目的功能,并指出了项目修订的方向。未来需要在更大的多样化样本中进行进一步的测试。