The Latent-Variable Structure of the Addenbrooke's Cognitive Examination-Revised

Madeleine L Connolly,Stephen C Bowden,Leonie C Simpson,Malcolm Horne,Sarah McGregor
DOI: https://doi.org/10.1093/arclin/acz081
2020-02-20
Abstract:Objectives: To establish a theoretically justified factor structure for the Addenbrooke's Cognitive Examination-Revised (ACE-R). Methods: Our sample comprised 288 patients with Parkinson's disease (179 men and 109 women). The mean age of participants was 66.66 (SD = 8.93). Confirmatory factor analysis (CFA) was used to evaluate the test developers' five-factor model of the ACE-R, and alternative models as guided by the Cattell-Horn-Carroll (CHC) theory. Exploratory structural equation modeling (ESEM) was also employed to examine alternative factor structures to ensure that a good candidate model was not overlooked. Results: A three-factor CHC-guided CFA and a similar three-factor ESEM model both showed acceptable overall fit, and interpretable factor structures. The three-factor CFA model showed two factors of pure CHC constructs: acquired knowledge (Gc), and visuospatial ability (Gv), and one combined factor, namely, long-term memory retrieval, fluency, and working memory (Glr-Gsm). The three-factor ESEM model showed three factors essentially in line with the CFA results. Conclusion: The three-factor CHC-guided CFA model was selected as the best model to guide clinical interpretation of cognitive variables underlying ACE-R scores.
What problem does this paper attempt to address?