Floor‐effect in the performance of the Spanish‐language MoCA by cognitively unimpaired Peruvian individuals with low educational background

• Published in: Alzheimer's & dementia 2024-12, Vol.20 (S3), p.n/a
• Main Authors: Custodio, Nilton, Calandri, Ismael Luis, Montesinos, Rosa, Custodio, Belen, Malaga, Marco, Chambergo‐Michilot, Diego, Bustamante‐Paytan, Diego, Aguero, Katherine, Verastegui, Graciet, de la Colina, Adrián Noriega
• Format: Article
• Language: English
• Subjects: Clinical Manifestations
• ISSN: 1552-5260; 1552-5279
• DOI: 10.1002/alz.090704

Background The Montreal Cognitive Assessment (MoCA) stands as a prominent cognitive impairment screening tool, finding widespread use globally and existing in official versions across 14 languages, including Spanish. Despite this, the challenges emerge due to the extensive variations within the Spanish language, which is not only the fourth most spoken language worldwide but also possesses significant geographic diversity, particularly evident in regions like Peru. Here, the intersection of regional nuances, low educational backgrounds, and culturally distinct tasks complicates the application of a standard MoCA version. We aimed to assess MoCA performance in individuals with both high and low education levels from an urban region of Peru to investigate the applicability of MoCA in low‐educated population. Methods A total of 40 cognitively healthy individuals with low education (less than 6 years, mean = 3.8 ±1.8) and 40 individuals with higher education (more than 6 years, mean = 14.1 ±3.5) participated in the study. We assessed them with the Spanish version 7.0 of MoCA and estimated total score and domain indices. We fitted linear regression models to evaluate age‐ and sex‐adjusted differences between groups. We examined the presence of a floor effect in the low education group, defining it as the test’s minimum possible score surpassing the mean minus three standard deviations. This observation indicates a lack of sensitivity in capturing normal inter‐individual variation at lower values. To delve deeper, we employed a three‐parameter Item Response Theory model (IRT) to estimate the difficulty of each test item. Results We observed significant differences in the total MoCA score (b = 9.3, 95% CI: 7.7,11, p