Analytical Appraisal of Validity and Reliability Metrics in Dementia Knowledge Assessment Instruments: Consequences for Clinical Nursing Practice in Vietnam.

PURPOSE: To validate and compare the psychometric properties of the Alzheimer's Disease Knowledge Scale Vietnam Version (ADKS-V), Dementia Knowledge Assessment Tool Version Two (DKAT2-V), and 27-item Dementia Knowledge Assessment Scale Vietnam Version (DKAS-V). METHOD: A cross-sectional study surveyed 360 community-dwelling older adults and 394 primary care nurses in Ho Chi Minh City. RESULTS: The scales demonstrated strong content validity (.86 to .91), convergent validity (r = .698 to .770), and reliability (Kuder-Richardson 20 [KR-20] = .765 to .820, ω = .793 to .849). Unidimensional confirmatory factor analysis (CFA) fit was poor. In contrast, target-population-specific orthogonal bifactor CFA models confirmed superior fit (comparative fit index = .939 to .981, root mean square error of approximation = .031 to .049) over unidimensional structures. High general factor strength (ωH = .571 to .811) and negligible specific subscale variance (mostly <.15) justified single composite scoring. CONCLUSION: All three Vietnamese versions are culturally appropriate and robust. We recommend DKAT2-V for nursing education, ADKS-V for community screening, and DKAS-V as a universal epidemiological tool.