Supplementary Material for: Utility of Individual Neurocognitive Domains of the Rowland Universal Dementia Assessment Scale to Detect Cognitive Impairment in an Urban Illiterate Population

Introduction: Brief cognitive tests like the RUDAS are useful for early detection of neurodegenerative diseases, especially in low-literacy populations in Low-and-Middle-Income Countries (LMICs). This study assesses the diagnostic performance of the RUDAS by subdomains to detect MCI and dementia in illiterate individuals in Lima, Peru
Methods: We performed a secondary analysis of a cohort of randomly selected illiterate individuals, recruited for validation of RUDAS in Lima, Peru. We compared the RUDAS domain and total scores to a battery of neuropsychological tests (FDS, BDS, ROCF and MINT) using a Spearman correlation coefficient. For diagnostic performance, we calculated the Area Under the Curve (AUC) for each RUDAS domain.
Results: We included 187 patients (64 controls, 60 MCI, and 63 dementia patients). Average RUDAS score was 23.87 ± 0.93, 20.43 ± 1.39 and 14.97 ± 2.21, respectively. Differences were statistically significant between groups for all domains, except for ideomotor praxis and visuospatial praxis. For convergent validity, most significant correlations from the bivariate analysis remained in the age-adjusted regression, particularly for praxis, memory, and judgment. The total RUDAS score had an AUC of 98% (CI95 96-100%) for both MCI and dementia. By domains, judgment, verbal fluency, and memory AUC ranging from 0.7 to 0.8.
Conclusions: Our study showed the RUDAS domains of visuospatial orientation and memory had an excellent diagnostic performance for the detection of dementia. However, further research is needed to validate the use of individual RUDAS items for assessing specific cognitive domains.