Remote digital cognitive assessment in a trial-ready Alzheimer's disease cohort: A scalable approach for early intervention studies.

INTRODUCTION: Early detection of cognitive decline in Alzheimer's disease (AD), particularly in preclinical stages, is critical for evaluating therapeutic interventions. Traditional cognitive assessments often require lengthy in-person visits, and may therefore limit scalability for younger, trial-ready populations for primary and secondary prevention studies. We evaluated the feasibility, reliability, and validity of high-frequency remote digital cognitive assessments in individuals with autosomal dominant AD (ADAD). METHODS: One hundred twenty-three mutation carriers and non-carriers from the Dominantly Inherited Alzheimer Network Trials Unit from 20 international sites (Ages 19-66 years) completed remote assessments via personal smartphones, prompted four times daily for 7 days (≈ 3 minutes per session), and conventional in-clinic cognitive testing at baseline. Adherence, between-person reliability, test-retest reliability (intraclass correlation coefficients [ICCs]), construct validity (confirmatory factor analysis), and sensitivity to clinical disease progression were evaluated. RESULTS: Remote measures demonstrated excellent reliability (ICCs > 0.90 after just 10 sessions) and strong construct validity, with tasks loading onto memory, attention, and executive function domains. A composite of the remote tests slightly outperformed a traditional composite at separating mutation carriers from non-carriers. Average adherence to the study protocol was 42%, lower that observed in previous studies of older adults. DISCUSSION: High-frequency remote cognitive assessment is feasible, reliable, and valid in a relatively young international ADAD cohort. This approach offers substantial utility for clinical trials, including improved accessibility and enhanced reliability, supporting its integration into early-intervention studies.