Diagnosis of early dementia by the Double Memory Test: encoding specificity improves diagnostic sensitivity and specificity

Abstract

Objective: To compare the Double Memory Test (DMT) with standard memory tests in the diagnosis of early dementia.

Background: Diagnosis of dementia requires memory impairment, but few memory tests coordinate acquisition and retrieval to optimize encoding specificity for high sensitivity and specificity. The DMT was developed to improve early diagnosis.

Design: We compared the discriminative validity of the DMT, Paired Associates (PA), and Logical Memory (LM) memory tests in a nested case-control study of 30 cases of early dementia and 90 controls matched for age, education, and sex.

Methods: The DMT includes memory tests with (CCR) and without (ICR) encoding specificity. Both tests use category cues to elicit retrieval, but CCR optimizes encoding specificity because the same cues are used for acquisition and retrieval. ICR does not because category cues are used only for retrieval. We used conditional logistic regression to estimate diagnostic sensitivity and specificity.

Results: The median BIMC of dementia cases was 10, indicating mild dementia. CCR had much higher sensitivity (93%) and specificity (99%) than ICR (53%, 94%), PA (68%, 91%), and LM (48%, 92%). CCR had the greatest advantage in the mildest cases.

Conclusions: CCR has substantially higher sensitivity and specificity for diagnosis of early dementia than memory tests that do not coordinate acquisition and retrieval. Superior discrimination by CCR is due to an encoding specificity deficit in dementia that increases the difference in recall by cases and controls. CCR is an efficient test with excellent discriminative validity that should facilitate diagnosis of early dementia.