Differential utility of immediate versus delayed memory measures for the identification of episodic memory...

Delayed recall measures are considered sensitive to episodic long-term memory impairment, and longer delays are argued to aid diagnostic accuracy. This meta-analysis examined whether delayed recall provides an advantage in identifying memory impairment relative to very short-delay or "immediate" recall. Eligible studies included a patient sample with suspected or known episodic memory impairment and a healthy control sample, both administered immediate and delayed recall trials. The standardized mean difference (Hedges' g) between patients and controls on immediate and delayed recall was calculated using multivariate random-effects meta-analysis. The meta-analysis included 290 studies, comprising 71,381 controls and 45,054 patients. Delayed recall ranged from 40 s filled delay to 6 weeks. Controls outperformed patients on immediate (g = 1.074, 95% CI [0.973, 1.176]) and delayed recall (g = 1.148, 95% CI [1.034, 1.261]). Patient-control differentiation was greater on delayed compared to immediate recall (g = -0.083, 95% CI [-0.126, -0.041]), though the difference was not clinically meaningful. Delay length (minutes) did not predict the delayed recall effect (β^ < 0.001) or moderate the difference between immediate and delayed recall (β^ = 0.020). Analyzed separately, word list, story, and shape recall measures showed no effect for increasing delay, and word list measures showed no delayed recall advantage. In mild cognitive impairment and Alzheimer's disease, delayed recall offered a small to medium advantage, but this was not present across all measures. All clinical subgroups analyzed showed no effect for increasing delay length. Further research is needed to clarify the utility of delayed recall. (PsycInfo Database Record (c) 2026 APA, all rights reserved).