Comparison of the neuropsychological-actuarial and clinical-consensus approaches to diagnosis of mild cognitive impairment in an ethnically diverse sample

<b>Objective:</b> Some studies show using neuropsychological-actuarial mild cognitive impairment (MCI) criteria may produce lower reversion rates (i.e. reverting from a cognitive diagnosis to no diagnosis) and higher associations with Alzheimer's disease biomarkers compared to conventional methods, but this has not been evaluated in a Hispanic sample. This study evaluated neuropsychological-actuarial MCI criteria performance in the Texas Alzheimer's Research and Care Consortium (TARCC) cohort compared to clinical-consensus diagnosis. <b>Method:</b> Data from 2,110 TARCC participants (47% Hispanic) were utilized. McNemar tests evaluated the performance of neuropsychological-actuarial versus clinical-consensus criteria in those who remained stable or converted to dementia over 3 annual visits, stratified by ethnicity (Hispanic vs. non-Hispanic). Chi-square analysis was used to compare frequencies of <i>APOE</i> ε4 allele positivity by diagnostic method. <b>Results:</b> Significantly more Hispanic participants were diagnosed with MCI using neuropsychological-actuarial criteria than were non-Hispanics, χ²(1) = 195.3, <i>p</i> < .001, <i>ϕ</i> = 0.32. Hispanic participants meeting neuropsychological-actuarial MCI criteria at baseline were more likely to revert at follow-up, χ²(1) = 10.04, <i>p</i> < .01, <i>ϕ</i> = 0.10. No differences in reversion rate were found between Hispanic and non-Hispanic individuals with clinical-consensus MCI diagnoses, χ²(1) = 0.38, <i>p</i> = .60. There was no association between <i>APOE</i> ε4 allele positivity and neuropsychological-actuarial diagnosis, while there was an association for clinical-consensus diagnoses <i>χ</i>²(1) = 15.1, <i>p</i> < .001, <i>ϕ</i> = 0.09. <b>Conclusions:</b> In TARCC, the clinical-consensus MCI diagnostic method produced fewer cases of reversion compared to the neuropsychological-actuarial method, particularly in Hispanic participants. This is consistent with recent research investigating the use of the neuropsychological-actuarial method in African American/Black individuals. Caution is warranted when using neuropsychological-actuarial criteria among individuals with diverse backgrounds.