Validating a French-language version of Health Communication Assessment Tool: Content validity, test-retest reliability, interrater agreement, and implications for further development

Introduction
The assessment of the communication skills of nurses in clinical settings bears both practical and educational significance. Despite the significance of communication skills to realize high-quality patient-care, the development and validation of such an instrument applicable in different cultural contexts is still lacking and faces certain challenges. Additionally, the assessment of nurse-patient communication, which is dynamic and context-bounded, requires a fine-grained instrument, while not compromising the complexities of the skills under assessment.
The aim of the present study was to test a French-language version of the Health Communication Assessment Tool (HCAT) (Campbell et al., 2013), on a French-speaking sample in Belgium against different measures of content relevance, validity, and reliability. Based on cognitive fluency perspective, we developed three simulated scenarios registering three levels of skill performance, including low, medium, and high demonstration. Thus, the psychometric properties of the HCAT instrument can be thoroughly examined.
Methods
Ten experts in communication and 52 nurse educators participated in the study. Firstly, the instrument was translated from English to French using back-translation technique. After semantic equivalence was agreed upon, the experts rated the relevance of the 22 items making up the HCAT, resulting in the removal of one culturally-irrelevant item. Subsequently, the experts and nurse educators rated the communication skills of the nurse in three scenarios. To ensure test-retest reliability, the nurse educators rated the scenarios twice within a four-week interval. The experts’ ratings were used as the gold standard when examining the reliability of the nurse educators’ ratings.
Results & Discussion
Exploratory and confirmatory factor analyses showed that the instrument could distinguish three aspects of communication skills, i.e. professional presentation, empathy, and trust building. Two-way random intraclass correlation coefficients (ICCs), ranging from 0.917 to 0.982, revealed that interrater agreement was excellent for consistency and absolute average-measures. However, at absolute single-measure, an ICC=0.425 suggested adequate interrater reliability for the medium-performed scenario, which decreased in low-performed (ICC=0.348) and high-performed (ICC=0.176) scenarios. Further, more than half of the items displayed poor test-retest reliability.