Evaluating the accuracy of brief cognitive assessments for the early detection of cognitive impairment in adults across primary and specialty care settings: A systematic review and meta‐analysis
Jennifer Loveless, Lauren Pilcher, Corey J. Bolton, Dylan X. Guan, Uzochukwu Imo, Chimkaha G. A. Ngubo, Mary Beth McAteer, Olivia Urquhart, Subhashini K Rangarajan, Lorie S. Kim, Katarina Sako, Belinda Tang, Yara Yakoub, Luiza S. Machado, Fred Ko, Dorene Rentz, Sarah Pahlke, Simin Mahinrad, Heather Snyder, Malavika P. TampiAbstract
BACKGROUND
This systematic review and meta‐analysis aimed to assess the accuracy of brief (i.e., ≤15 minutes), freely available cognitive tests for the early detection of cognitive impairment (including mild cognitive impairment and dementia) in adults 55 years and older in primary and specialty care.
METHODS
A systematic search of MEDLINE, Embase, and Cochrane Library databases was conducted to identify studies published between January 1999 and August 2024 that evaluated the accuracy of brief, freely available cognitive tests for early detection of cognitive impairment. The tests included 5‐Cog, Eight‐item Informant Interview to Differentiate Aging and Dementia (AD8), General Practitioner assessment of Cognition (GPCOG), Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE), Mini‐Cog, Memory Impairment Screen (MIS), Montreal Cognitive Assessment (MoCA), Quick Dementia Rating System (QDRS), Rowland Universal Dementia Assessment Scale (RUDAS), and Saint Louis University Mental Status (SLUMS). Though widely used, the Mini‐Mental State Examination did not meet eligibility criteria and was not considered. Accuracy measures were extracted from eligible studies and pooled where possible. Risk of bias was assessed using Quality Assessment of Diagnostic Accuracy Studies 2, and certainty of evidence was assigned according to the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach.
RESULTS
Forty‐seven studies evaluating 10 index tests, including short versions of IQCODE and MoCA, were included. Test accuracy varied widely among tests and data pooling was not possible for most analyses. Studies conducted in both primary and specialty care were identified for AD8, Mini‐Cog, MIS, MoCA, QDRS, Short IQCODE, and SLUMS, while evidence for IQCODE, short MoCA, and RUDAS was limited to a single setting. No studies meeting eligibility criteria were found for 5‐Cog and GPCOG. Across all tests, in primary care, sensitivity ranged from 0.17 to 0.98, and specificity ranged from 0.10 to 0.98. Across all tests and assuming the same prevalence in specialty care, sensitivity ranged from 0.73 to 0.97, and specificity ranged from 0.45 to 0.90.
CONCLUSION
This systematic review provides a synthesis of the evidence on the accuracy of brief, freely available cognitive assessment tools for the early detection of cognitive impairment in primary and specialty care. Results from primary care settings informed the development of evidence‐based recommendations presented in the accompanying clinical practice guideline on the use of brief cognitive tests in primary care.