Related Experiment Video
Updated: Dec 1, 2025
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Development of a Brief Screen to Detect Cognitive Impairment in Older Adults: The QuickSort
Amie M Foran1, Jane L Mathias1, Stephen C Bowden2
1School of Psychology, Faculty of Health and Medical Sciences, University of Adelaide, Adelaide, South Australia, Australia.
Background:
Sorting tests detect cognitive decline in older adults who have a neurodegenerative disorder, such as Alzheimer's and Parkinson's disease. Although equally effective at detecting impairment as other cognitive screens (e.g. Mini-Mental State Examination (MMSE)), sorting tests are not commonly used in this context. This study examines the QuickSort, which is a new brief sorting test that is designed to screen older adults for cognitive impairment.
Design:
Observational cohort study.
Setting:
General community and inpatients, Australia.
Participants:
Older (≥60 years) community-dwelling adults (n = 187) and inpatients referred for neuropsychological assessment (n = 78). A normative subsample (n = 115), screened for cognitive and psychological disorders, was formed from the community sample.
Measurements:
Participants were administered the QuickSort, MMSE, Frontal Assessment Battery (FAB), and Depression Anxiety and Stress Scale-21. The QuickSort requires people to sort nine stimuli by color, shape, and number, and to explain the basis for their correct sorts. Sorting (range = 0-12), Explanation (range = 0-6), and Total (range = 0-18) scores were calculated for the QuickSort.
Results:
The Cognitively Healthy subsample completed the QuickSort within 2 minutes, 50% had errorless performance, and 95% had Total scores of 10 or greater. The likelihood of community-dwelling older adults and inpatients (n = 260) being impaired on either the MMSE or FAB, or both, increased by a factor of 3.75 for QuickSort Total scores of less than 10 and reduced by a factor of 0.23 for scores of 10 or greater.
Conclusion:
The QuickSort provides a quick, reliable, and valid alternative to lengthier cognitive screens (e.g., MMSE and FAB) when screening older adults for cognitive impairment. The QuickSort performance of an older adult can be compared with a cognitively healthy normative sample and used to estimate the likelihood they will be impaired on either the MMSE or FAB, or both. Clinicians can also use evidence-based modeling to customize the QuickSort for their setting.
