Cognitive assessment through telehealth platforms may soon become more precise and accessible for older adults, potentially transforming how mental health services reach underserved populations. The convergence of digital health infrastructure with artificial intelligence creates new opportunities for early cognitive screening in primary care settings.
Researchers analyzed data from 1,143 older adults across 11 healthcare sites using the HERMES Digital Platform between 2022 and 2025. Their machine learning model examined 83 potential risk factors to predict Mini-Cog cognitive assessment scores, ultimately identifying eight key features strongly associated with low cognitive performance. The Random Forest classifier used Elastic Net regularization to refine feature selection, demonstrating how algorithmic approaches can distill complex behavioral and demographic data into actionable clinical insights.
This work represents a meaningful step toward scalable cognitive health monitoring, particularly for geographically isolated or economically disadvantaged older adults who face barriers to traditional psychiatric care. The integration of tele-neuropsychology with primary healthcare addresses a critical gap in mental health infrastructure. However, the study's effectiveness depends heavily on digital literacy among target populations and the reliability of remote cognitive assessments compared to in-person evaluations. While machine learning models show promise for risk stratification, their clinical utility requires validation across diverse populations and healthcare systems. The approach appears most valuable as a screening tool rather than a diagnostic replacement, offering a practical method to identify individuals who warrant further comprehensive cognitive evaluation.