Related Experiment Video
Updated: Jun 1, 2025

A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Effectiveness and utilization of a cognitive screening program for primary geriatric care
David P Salmon1,2, Anna Malkina3, Melanie L Johnson4
1Department of Neurosciences, University of California, San Diego, La Jolla, CA, 92093-0948, USA. dsalmon@ucsd.edu.
Background:
Effective detection of cognitive impairment in the primary care setting is limited by lack of time and specialized expertise to conduct detailed objective cognitive testing and few well-validated cognitive screening instruments that can be administered and evaluated quickly without expert supervision. We therefore developed a model cognitive screening program to provide relatively brief, objective assessment of a geriatric patient's memory and other cognitive abilities in cases where the primary care physician suspects but is unsure of the presence of a deficit.
Methods:
Referred patients were tested during a 40-min session by a psychometrist or trained nurse in the clinic on a brief battery of neuropsychological tests that assessed multiple cognitive domains. Short questionnaires covering subjective cognitive complaints, symptoms of depression, and medical history were also administered. Results were conveyed to a dementia specialist who reviewed them and returned their judgement of the validity of the cognitive complaint to the primary care provider. Retrospective medical records review was carried out for a random (stratified) half of the sample to determine how screening results were utilized. Screening tests were repeated after two years in a subset of 69 patients.
Results:
The 638 patients screened (mean age = 75.9 years; mean education = 14.9 years; 58% women) were classified by screening as having normal cognition (n = 177), depression (with possible cognitive changes; n = 115), mild cognitive impairment (MCI; n = 107), or dementia (n = 239). Classification accuracy was shown by high agreement with the eventual clinical diagnosis in the medical record (69%; Cohen's Kappa = .38; p < .001; 77% if MCI and dementia were collapsed; Cohen's Kappa = .58; p < .001) and longitudinal decline in cognitive test scores only in those initially classified as having MCI or dementia. Medical records documented discussion of screening results with the patient in 69% of cases (80% if MCI or dementia was detected) and often referral to a specialist (62%), new brain imaging (54%), or change in medication (58%) when screening indicated potential cognitive impairment.
Conclusion:
The cognitive screening program was well accepted by primary care providers as an efficient and effective way to evaluate concerns about cognitive decline in older adults.
More Related Videos
06:58Highlighting and Reducing the Impact of Negative Aging Stereotypes During Older Adults' Cognitive Testing
Published on: January 24, 2020
07:31A Computerized Functional Skills Assessment and Training Program Targeting Technology Based Everyday Functional Skills
Published on: February 13, 2020