Participant and Study Partner Prediction and Identification of Cognitive Impairment in Preclinical Alzheimer's Disease - Study Partner vs. Participant Accuracy

Abstract

Preclinical Alzheimer’s disease (AD) trials require participants to enroll with a study partner (SP), a person who can attend visits and report changes in the participant’s cognitive ability. We quantified the relative performance of SPs to assess participant cognition, cross­sectionally and longitudinally. We also stratified on amyloid status. We assessed participant and SP Everyday Cognition (ECog) scores and participant Alzheimer Disease Assessment Scale­cognitive subscale (ADAS13) data from 335 cognitively normal participant­ partner dyads in the AD Neuroimaging Initiative. We used random forest models and Linear Mixed Effects (LME) to model ADAS13 scores as a function of participant and/or SP ECog scores over time, focusing on 12­, 24­, and 48­month timepoints. LME models were adjusted for potential confounding factors, including APOE4 status, amyloid status, baseline age, years of education, and gender. In models predicting ADAS13 12 months into the future, the estimated mean variable importance (eMVI) associated with baseline SP ECog was the same as that associated with baseline participant ECog (eMVI=0.043, 95%CI 0.039, 0.047 for partner; eMVI=0.042, 95%CI 0.038, 0.045 for participant). For predicting ADAS13 48 months out, the eMVI associated with baseline SP ECog was slightly higher than that associated with baseline participant ECog (eMVI=0.145, 95%CI 0.132, 0.157 for partner; eMVI=0.134, 95%CI 0.125, 0.144 for participant; see Figure 1(a)). In cross­sectional models predicting 12­month ADAS13, the eMVI associated with SP ECog at 12 months and the eMVI associated with participant ECog at 12 months did not differ (eMVI=0.024, 95%CI 0.021, 0.027 for partner; eMVI=0.025 95%CI 0.023, 0.027 for participant). However, the eMVI associated with SP ECog at 48 months for predicting 48­month ADAS13 is twice as large as that associated with participant ECog at 48 months (eMVI=0.085, 95%CI 0.076, 0.092 for partner; eMVI=0.042, 95%CI 0.037, 0.48 for participant; see Figure 1(b)). We did not observe qualitative differences by amyloid status. Results suggest that, while baseline participant information reasonably predicts subsequent cognitive change, informants perform better at cross­sectionally recognizing cognitive status as observation time grows. Thus, partner information becomes increasingly important. As such, our results provide evidence to support the SP requirement to ensure trial data integrity.

Mary M. Ryan

Assistant Professor
mary [dot] ryan [at] wisc [dot] edu

My research interests include group sequential design and clinical trials, with applications in Alzheimer’s Disease biomarker discovery, as well as pragmatic and cluster randomized trials.