Social determinants of health and long-term profiles of subjective memory ratings in middle-aged and older adults: A population-based longitudinal study.

OBJECTIVES: To characterize long-term profiles of subjective memory ratings and examine their associations with cumulative, domain-specific, and individual social determinants of health (SDoH) among middle-aged and older adults. METHODS: We analyzed six waves of the China Family Panel Studies (2012-2022) comprising 14,938 adults aged ≥45 years. Group-based trajectory modeling was used to characterize longitudinal profiles of subjective memory ratings. Logistic regression estimated associations with baseline SDoH, with restricted cubic splines assessing dose-response shape and false-discovery-rate correction addressing multiple testing. Sensitivity analyses incorporated baseline memory, prospective follow-up, objective cognition, missing data, attrition, and complex sampling. A binary Ising network was estimated as a secondary analysis. RESULTS: Two profiles were identified and differed predominantly in their overall subjective-memory rating levels, with similar rates of change over time. Each additional adverse SDoH was associated with higher odds of the persistently lower subjective-memory rating profile after multivariable adjustment (OR, 1.23; 95% CI, 1.20-1.25), with attenuation after additional adjustment for the 2012 memory rating (OR, 1.16; 95% CI, 1.13-1.19). The association persisted when profile classification was restricted to 2014-2022 (OR, 1.15; 95% CI, 1.13-1.18). The cumulative association was nonlinear (P < 0.001), and one of 10 cross-domain interactions remained after false-discovery-rate correction. Subjective memory ratings correlated weakly with objective cognitive measures. In the network analysis, lower education had the largest expected influence estimate, although bootstrap comparisons did not support a statistically distinct hierarchy among the highest-ranking nodes. CONCLUSIONS: Greater social disadvantage was associated with persistently lower subjective memory ratings, highlighting enduring differences in subjective-memory rating levels over time. These findings indicate enduring social patterning of subjective memory from midlife into later life while underscoring the distinction between subjective ratings and objective cognitive performance.