Abstract
<p>The purpose of this project is to characterize neurocognitive functioning and correlates of neurocognitive functioning, as well as to examine the diagnostic accuracy of relatively promising neurocognitive tests in two underserved sociodemographic groups living with HIV: monolingual Spanish-speaking Hispanics and English-speaking African-Americans, in order to better understand HIV-Associated Neurocognitive Disorder (HAND) within these groups. Examining HAND in underserved groups is paramount to the development of effective HAND screening and treatment algorithms in clinics that serve these groups, such as the AIDS Clinical Research Unit (ACRU) at the University of Miami. The current project was drawn from an initiative to develop a HAND screening algorithm for this clinic, such that patients with milder forms of HAND can be identified. The current study examined the performance of convenience samples of English-speaking African-Americans (n = 38) and monolingual Spanish-speaking Hispanics (n = 50) on a variety of easy-to-use, sensitive neurocognitive tests we previously collected at the ACRU: Grooved Pegboard (GP), Trail Making Test (TMT), Action Fluency (AF), and the Hopkins Verbal Learning Test-Revised (HVLT-R). I used these tests and medical chart review to classify HAND via the 2007 Frascati diagnostic criteria, the most up-to-date criteria for classifying HAND. Furthermore, I examined the association between an array of psychosocial, medical, and behavioral factors and HAND classification, which may aid in informing HAND screening and testing algorithms. Finally, I also examined the sensitivity and specificity of promising measures, Action Fluency (which tests the ability to generate novel verbs) and Trail Making Test, in detecting HAND, in order to elucidate whether these free and easy-to-administer measures should be added to screening batteries to detect HAND within these groups. Using logistic regression analyses, I analyzed associations between medical, behavioral, and psychosocial correlates and HAND. Perceived stress and poorer sleep quality were found to be at least marginally associated with increased odds of HAND, while social support from friends was found to be associated with decreased odds of HAND. Using HAND diagnosis via GP, HVLT-R, and medical chart review as the gold standard, I calculated the sensitivity and specificity of the AF and TMT, as well as performed ROC analyses to evaluate the overall diagnostic accuracy of these measures. Both the AF and TMT measure did not demonstrate adequate sensitivity (> 70%) in detecting HAND in the overall sample, as well as within each of the sociodemographically distinct groups. Furthermore, the AF and TMT demonstrated poor AUC estimates (> .70), indicating poor overall diagnostic accuracy. Though the sample size was limited, results indicate that psychosocial factors such as stress, sleep quality, and social support from specific sources may aid in identifying individuals at risk for or living with HAND. These results also indicated poor diagnostic accuracy for the AF and TMT in the study samples. Future work should evaluate these measures using larger samples and more extensive testing batteries.</p>