BACKGROUND: Symptom checklists used in clinical research of gastroparesis have not been evaluated for reliability or validity. The Gastroparesis Cardinal Symptom Index (GCSI) is a reliable, validated measure whereby patients score three groups of symptoms: nausea/vomiting, postprandial fullness/early satiety and bloating. AIM: The purpose of this study was to translate the GCSI into Hebrew and then validate the language of the translation for use in future studies. METHODOLOGY: The measure was first translated into Hebrew and then back-translated to English, followed by a comparison of the two versions. The English and Hebrew versions were then completed by 40 healthy bilingual volunteers, using a cross-over design, with half answering the English version first while the other half replied to the Hebrew version first. RESULTS: The internal consistency for both measures was moderate, the English slightly lower than the Hebrew (Cronbach's alpha 0.67 vs. 0.72). Intra-class correlation values were larger than 0.7 for all but one of the symptoms (item 3, "vomiting"), although a Wilcoxian signed-rank test found this correlation to be of borderline significance (z=-1.63, p=0.50). CONCLUSIONS: The Hebrew translation of the GCSI is valid for use in clinical research, although further psychometric testing is needed to test its value for this purpose. This manuscript describes the stages of measurement, translation and language validation, the difficulties that researches face and possible solutions when researches choose to use a measure which was developed in a different language.
BACKGROUND: Symptom checklists used in clinical research of gastroparesis have not been evaluated for reliability or validity. The Gastroparesis Cardinal Symptom Index (GCSI) is a reliable, validated measure whereby patients score three groups of symptoms: nausea/vomiting, postprandial fullness/early satiety and bloating. AIM: The purpose of this study was to translate the GCSI into Hebrew and then validate the language of the translation for use in future studies. METHODOLOGY: The measure was first translated into Hebrew and then back-translated to English, followed by a comparison of the two versions. The English and Hebrew versions were then completed by 40 healthy bilingual volunteers, using a cross-over design, with half answering the English version first while the other half replied to the Hebrew version first. RESULTS: The internal consistency for both measures was moderate, the English slightly lower than the Hebrew (Cronbach's alpha 0.67 vs. 0.72). Intra-class correlation values were larger than 0.7 for all but one of the symptoms (item 3, "vomiting"), although a Wilcoxian signed-rank test found this correlation to be of borderline significance (z=-1.63, p=0.50). CONCLUSIONS: The Hebrew translation of the GCSI is valid for use in clinical research, although further psychometric testing is needed to test its value for this purpose. This manuscript describes the stages of measurement, translation and language validation, the difficulties that researches face and possible solutions when researches choose to use a measure which was developed in a different language.