Scientific challenges underlying adaptation of existing cognitive tests or development of new tests to address cultural and linguistic diversity

Abstract

AbstractBackgroundHigh quality cognitive assessments may be a requisite part of care for people with any disease associated with cognitive changes. While extensive research has been devoted to development of reliable and valid cognitive assessment in some cultures, the development of cognitive assessments that address cultural and linguistic diversity has been relatively neglected.MethodThis talk will provide a conceptual overview of the principal scientific issues involved in developing culturally‐sensitive and scientifically robust assessments of cognition. The issues will be illustrated with examples of successful application of these principles, together with limitations. Recent research highlighting the risks of over‐interpretation of normative comparisons across cultural or ethnic diversity will be highlighted.ResultTwo broad scientific precepts underpin development of any new cognitive test. The first of these involves establishing construct validity, that is, the scientifically defensible measurement of target cognitive functions. Secondly, a normative basis for individual assessment is desirable.Regarding the first of these requirements, establishing construct validity can most efficiently triangulate from established cognitive measures, for example, informed by the best available model of cognitive abilities, namely the Cattell‐Horn‐Carroll model. Using the empirical methods of convergent and discriminant, any new test can be evaluated for convergence on construct estimation, within the limits of measurement error, with an established test of the target construct developed in another culture. The best available methods for this purpose involve multi‐group confirmatory factor analysis, which allows evaluation of measurement invariance (MI). If established, MI permits generalization of construct validity research across cultures.Regarding the second broad requirement, representative normative samples are ideal for individual assessment, but are expensive. An alternative is to use an individual baseline approach where repeat assessment within any one patient provides the control standard against which to monitor change, for example, in response to progressive disease. Methods for evaluating the reliability or precision of repeat assessments will be briefly delineated.ConclusionProviding rigorous cross‐cultural assessment of dementia and other cognitive disorders entails addressing significant scientific challenges. However, all of challenges fall within established methods that are readily available to researchers.