Validity and reliability of the Turkish version of three screening tools (PYMS, STAMP, and STRONG-kids) in hospitalized children

• Published in: Clinical nutrition ESPEN Vol. 39; pp. 96 - 103
• Main Authors: Pars, Hatice, Açıkgöz, Aylin, Erdoğan, Beyza Doğanay
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01-10-2020
• Subjects: Malnutrition; Screening tools; Translations; Validation; Validation; Malnutrition; Translations; Screening tools
• ISSN: 2405-4577; 2405-4577
• DOI: 10.1016/j.clnesp.2020.07.011

There are different malnutrition screening tools developed in the English language, but limited data available on the validity of their versions translated into other languages. The aim of this study is to construct a Turkish version of PYMS, STRONG-kids, and STAMP and to determine their validity and reliability in a pediatric population.
The present study was conducted at two stages. First, the process of translating PYMS, STRONG-kids, and STAMP into the Turkish language was performed. In the second stage, the validity and reliability were evaluated in the new version of the tools. The anthropometric assessment was used as the reference standard in evaluating the validity of criteria. The reliability was observed through inter-rater and intra-rater agreement.
202 children in total were involved in the present study, 42 of whom took part in the reliability phase. The inter-rater agreement between one dietitian and one nurse was kappa (κ) = 0.955 (95% CI 0.904–1.000) for PYMS, 0.901 (95% CI 0.828–0.974) for STRONG-kids, and 0.963 (95% CI 0.912–1.000) for STAMP (almost perfect agreement). Sensitivity of the PYMS, STRONG-kids, and STAMP were 96.8%, 87.1%, and 77.4%, respectively. Specificity of the PYMS, STRONG-kids, and STAMP was found to be 65.0%, 30%, and 60.0%, respectively.
The inter-rater and intra-rater reliability of the three tools were determined to be good. While PYMS was found to be more sensitive in detecting the malnutrition risk when compared to the other tools, STRONG-kids had the lowest specificity according to the agreement between tool results and anthropometric assessment.