Memory Test Performance of People with Subjective Cognitive Decline Recruited From Different Settings: A Systematic...

Subjective cognitive decline (SCD) is defined as self-experienced cognitive decline without objective impairment on standardised tests. Research suggests SCD may show subtle impairment on detailed neuropsychological assessment and might therefore indicate the earliest stage of neurodegeneration. This review (PROSPERO: CRD42023382096) seeks to determine whether group differences in memory task performance between people with and without SCD exist. The review included studies since 2014 comparing episodic memory performance between people with and without SCD; where people with SCD were recruited exclusively from community or medical settings. Studies providing data for people with mild cognitive impairment (MCI) were included in a separate meta-analysis comparing SCD and MCI. A systematic search was conducted (PsycINFO, Web of Science, MEDLINE, CINAHL, PubMed on 11th August 2023). Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies tool. 1,815 records were identified, of which 45 met inclusion criteria and were included in a random-effects meta-analysis (SCD N = 5,948, Non-SCD N = 8,468). Twenty-one studies additionally provided data for an MCI group (SCD N = 1,034, MCI N = 2,119). Results indicated people with SCD performed significantly worse than people without SCD (Hedges' g = -0.19, 95% CI = -0.26, -0.11) and significantly better than MCI participants (Hedges' g = 1.20, 95% CI = 0.94, 1.46). For both meta-analyses there was significant between-study heterogeneity. There was a significant risk of publication bias for the meta-analysis comparing SCD to Non-SCD. These results suggest detailed memory assessment may be sensitive to reduced objective memory performance in SCD. However, it is unclear whether the small effect size has clinical significance. Higher quality and larger studies are needed to rule out the influence of moderating factors on memory performance.