Abstract
The 20-item Caregiver Contribution to Heart Failure Self-care (CACHS) was developed and tested in a series of preliminary studies providing evidence for the validity of this instrument in Canada. The purpose of this study was to assess the initial psychometric testing of a translated version of the 20-item CACHS in a European sample using classical test theory and item response theory (IRT) methods. This study is a preplanned subanalysis of data from a multisite, descriptive study of caregivers (n = 277; mean age, 52.7 years; 70.4% female) of Italian patients with heart failure. The CACHS-Italian version demonstrated adequate validity and reliability using classical test theory methods. Using IRT methods, test information function was unimodal with more item information in the negative versus positive regions of the latent trait. Most items displayed moderate to high discrimination with acceptable levels of difficulty. Differential item functioning, which determines whether caregivers from different groups with equal amounts of the latent trait have different expected item responses, differed based on caregivers' biological sex as did the IRT discrimination "[alpha]" and difficulty "[beta]" parameters. Overall, the CACHS-Italian version demonstrated similar psychometric properties and excellent reliability as demonstrated in the Canadian sample. However, unlike the original Canadian English version, the CACHS-Italian version showed significant nonuniform differential item functioning, discrimination, and difficulty based on caregivers' biological sex. Ongoing examination of the CACHS-Italian version, particularly in men, is needed as is further linguistic validation using the Canadian CACHS-English version in non-North American English-speaking countries.