Remote neuropsychological assessment using the UDS v3.0 T‐Cog: Preliminary data among participants at the Penn ADRC

Abstract

AbstractBackgroundDue to the COVID‐19 pandemic, collection of the Uniform Data Set (UDS) at Alzheimer’s Disease Research Centers (ADRCs) transitioned to remote administration via the T‐Cog. Feasibility of data collection and dissemination of normative data in individuals across the cognitive spectrum will be necessary for accurate utilization of remotely‐captured neurocognitive performance and classification of cognitive function.MethodData was included from individuals who completed a remote T‐Cog protocol during their annual NACC evaluation. Additional data included demographics, informant‐reported functioning, self‐reported depression and consensus diagnosis. Descriptive and general linear model (GLM) statistics were used to characterize the overall sample and test for differences in performance across diagnostic groups.ResultA total of 151 participants with normal cognition (NC; N=103), mild cognitive impairment (MCI; N=32) and Alzheimer’s disease (AD; N=16) were administered the T‐Cog (Age = 75.2+7.2, Education=16.3+3.1; 62% Female; 75% White). Data were collected via telephone (68%), video‐conference (29%), or a combination (3%). Time to completion ranged from 38‐120 minutes (M=58+14 minutes), with NCs showing a significantly lower (p<0.05) mean completion time (54 minutes) than those with MCI (65 minutes) and AD (69 minutes). Across the battery, 80% of responses were deemed “extremely valid”, 19% were “questionably valid” and <1% were “invalid”. Contributors to poor validity included hearing impairment (34%), distractions (16%), interruptions (6%), lack of effort/disinterest (9%), fatigue (3%), emotional issues (6%), or other reasons (25%). Diagnostic groups varied on tests of global cognition and functioning, as measured by the MoCA‐Blind and Global CDR, with the highest performance among NC, followed by MCI, and AD (p‐values<0.01). Nearly all T‐Cog measures of language, memory, and executive functioning abilities showed expected differences across diagnostic categories (covarying for age, sex and education; p<0.05). Comparisons to prior in‐person UDS neurocognitive testing across all diagnostic groups will also be presented.ConclusionPreliminary feasibility and performance data on the T‐Cog reveals expected differences in T‐Cog performance. Comparisons to within‐subject in‐person testing and deviations from in‐person norms will be presented to further explore the validity of remote cognitive testing and identify tests that show particular promise in discriminating between diagnostic groups.