Related Experiment Video
Updated: Jul 3, 2026
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Initial specialist validation of clinical decision support recommendations from a machine learning-enabled digital
Ali Jannati1,2, Claudio Toro-Serey1, Marissa Ciesla1
1Linus Health, Inc., Boston, MA, United States.
Frontiers in Neurology
|July 2, 2026
Summary
Primary care providers can use the Core Cognitive Evaluation (CCE) to detect cognitive impairment in older adults. Expert review found CCE-generated clinical decision support appropriate for guiding patient care and referrals.
Area of Science:
- Neurology
- Geriatrics
- Digital Health
Background:
- The advent of disease-modifying therapies for Alzheimer's disease (AD) necessitates scalable tools for primary care providers (PCPs) to detect cognitive impairment.
- The Linus Health Core Cognitive Evaluation (CCE) is a tablet-based tool integrating the Digital Clock and Recall (DCR) and Life and Health Questionnaire (LHQ) to provide clinical decision support (CDS) and decision-tree pathways.
Purpose of the Study:
- To assess the content validity of the CCE-derived CDS and decision-tree pathways for PCPs managing older adults with suspected cognitive impairment.
- To evaluate the appropriateness of CCE recommendations for PCP workup and referral decisions.
Main Methods:
- A retrospective specialist content-validity study was conducted using a modified RAND/UCLA Appropriateness Method.
- Five cognitive/behavioral neurologists rated CDS recommendations and pathway parts for patients aged ≥55 using a 1-9 scale.
- Expert agreement was quantified using Intraclass Correlation Coefficients (ICC).
Main Results:
- All recommendations for cognitive impairment met the appropriateness threshold (median score ≥7).
- Seven borderline/impaired DCR pathways were deemed appropriate (median 7-8).
- Two pathways, one for cognitively unimpaired individuals and another for anti-amyloid therapy referral, fell below the threshold (median scores 6 and 5, respectively).
- Moderate agreement was found per patient (ICC [2,k]=0.61), with lower agreement for individual diagnostic concerns (ICC [2,k]=0.25).
Conclusions:
- Cognitive neurologists found CCE-derived CDS appropriate for PCP use in evaluating and referring older adults with suspected cognitive impairment.
- The study supports the initial content validity of assessment-linked CDS.
- Refinement priorities include low-risk pathways and emerging therapy referrals, with future studies planned for PCP appropriateness and implementation.