Validation of the Dutch-Flemish translated ABCD questionnaire to measure cardiovascular diseases knowledge and risk perception among adults

Abstract

Valid and reliable measurement of an individual's knowledge and risk perception is pivotal to monitor and evaluate the effectiveness of interventions aimed at preventing cardiovascular diseases (CVDs). The recently developed Attitudes and Beliefs about Cardiovascular Disease (ABCD) knowledge and risk questionnaire is shown to be valid in England. In this study, we evaluated the psychometric properties of the modified and Dutch (Flemish)-translated ABCD questionnaire using both the classical test and item response theory (IRT) analysis. We conducted a community-based survey among 525 adults in Antwerp city, Belgium. We assessed the item- and scale-level psychometric properties and validity indices of the questionnaire. Parameters of IRT, item scalability, monotonicity, item difficulty and discrimination, and item fit statistics were evaluated. Furthermore, exploratory and confirmatory factorial validity, and internal consistency measures were explored. Descriptive statistics showed that both the knowledge and risk scale items have sufficient variation to differentiate individuals' level of knowledge and risk perception. The overall homogeneity of the knowledge and risk subscales was within the acceptable range (>0.3). The exploratory and confirmatory factor analyses of the risk scale supported a three-factor solution corresponding to risk perception (F1), perceived benefits and intention to change physical activity (F2), and perceived benefit and intention to change healthy dietary habit (F3). The two parametric logistic (2-PL) and rating scale models showed that the item infit and outfit values for knowledge and risk subscales were within the acceptable range (0.6 to 1.4) for most of the items. In conclusion, this study investigated the Dutch (Flemish) version of the ABCD questionnaire has good psychometric properties to assess CVD related knowledge and risk perception in the adult population. Based on the factor loadings and other psychometric properties, we suggested a shorter version, which has comparable psychometric properties.