Concurrent diagnostic efficiency and 8-year measurement stability of the weighted subjective memory complaints questionnaire-revised.

Abstract

BackgroundSubjective memory complaints (SMC) are a potential early indicator of dementia, but whether their measurement properties hold over years of repeated assessment is unexamined. The weighted Subjective Memory Complaints Questionnaire-Revised (SMCQ-R) showed higher cross-sectional accuracy than the original SMCQ, but its performance over extended follow-up requires verification.ObjectiveTo evaluate whether SMCQ-R maintains measurement fidelity over 8 years relative to SMCQ and Mini-Mental State Examination (MMSE).MethodsWe analyzed 2057 community-dwelling older adults without baseline major depressive disorder from the Korean Longitudinal Study on Cognitive Aging and Dementia (KLOSCAD), followed biennially for 8 years (5 assessments). Concurrent discriminative performance was evaluated via AUC, Cohen's d, signal-to-noise ratio (SNR). Sensitivity analysis excluded participants with incident depression during follow-up, and separately used all available data.ResultsAll three measures demonstrated excellent cross-sectional validity (mean AUC: SMCQ-R = 0.883, SMCQ = 0.889, MMSE = 0.884). SMCQ-R showed numerically larger effect sizes than SMCQ (Cohen's |d|: 1.884 versus 1.742, p = 0.154) and higher SNR (1.659 versus 1.485, p = 0.327), neither significant. Both subjective measures showed significant SNR reductions with incident depression, whereas MMSE remained robust.ConclusionsSMCQ-R showed a significant AUC decline over 8 years (slope = -0.0079/year, p = 0.024), but maintained clinically adequate accuracy (AUC ≥ 0.844), similar to SMCQ (0.883 vresus 0.889), despite a 36% reduction in item burden. Using all available data, this decline was not an artifact of the complete case restriction, and the MMSE discriminated at least as well as SMCQ-R at every assessment. SMCQ-R is regarded as an efficient complement to the MMSE rather than evidence of its superiority.