Development and Validation of Regression-based Neuropsychological Testing Norms for Peruvian adults to detect HIV-associated Neurocognitive Impairment

medRxiv [Preprint]. 2026 Feb 11:2026.02.09.26345550. doi: 10.64898/2026.02.09.26345550.

ABSTRACT

INTRODUCTION: Neurocognitive impairment (NCI) remains common among people living with HIV (PWH), particularly in low- and middle-income countries where accurate diagnostic tools are limited. In Peru, the lack of locally validated neuropsychological (NP) normative data in Spanish poses a major barrier to diagnosing HIV-associated NCI, especially among PWH who develop NCI at younger ages. This study aimed to develop regression-based NP norms for young and middle-aged Spanish-speaking adults in Lima, Peru and validate the norms in demographically similar PWH to improve diagnostic precision of HIV-associated NCI.

METHODS: A total of 164 healthy adults without HIV from Lima completed a comprehensive NP battery assessing memory, attention, executive function, and language, which are commonly affected in HIV-associated NCI. Multiple regression models were used to consider the influence of age, years of education, and sex on raw scores, yielding standardized demographically-adjusted norms for the population. The resulting norms were then applied to 310 PWH from Lima and then compared with previously published norms for Spanish speaking adults to evaluate performance differences.

RESULTS: Age and education were the strongest predictors of performance across tests, while sex had minimal influence. Compared to people without HIV, PWH had significantly lower educational attainment (mean 12.6 vs. 13.7 years) and exhibited significantly worse performance on normed scores of Benson Figure Copy, Benson Figure Delayed Recall, Color Trails 1 and 2, Hopkins Verbal Learning Test - Revised, and WAIS-III Digit Symbol Coding, Digit Span, and Symbol Search. There were statistically significant differences between T-scores on nearly all tests between our population-specific norms and previously published norms in both directions, indicating potential over- and under-detection errors when applying norms from non-local samples.

DISCUSSION: Our findings highlight the utility of locally derived norms in detecting subtle cognitive changes among young and middle-aged PWH compared with previously published norms for Spanish-speakers. Application of these norms reveals significant between-group differences that may go undetected using non-local normative data or raw scores. Future efforts should focus on rural norm development and inclusion of individuals with lower educational backgrounds in Peru and other Latin American countries.

PMID:41728318 | PMC:PMC12919123 | DOI:10.64898/2026.02.09.26345550