Depressive symptom patterns in older adults: evidence from two national ageing cohorts.

BACKGROUND

Late-life depressive symptoms are heterogeneous, and total symptom scores may obscure symptom configurations with distinct prognostic implications. We aimed to identify reproducible depressive symptom patterns in older adults and test whether these patterns added information beyond CES-D total score.

METHODS

We conducted a harmonised longitudinal cohort analysis using the Health and Retirement Study (HRS) as the discovery cohort and the English Longitudinal Study of Ageing (ELSA) as an independent replication cohort. Baseline was each participant's first wave with complete CES-D-8 item data and sufficient information to construct five harmonised symptom domains. Latent class analysis was used to derive symptom patterns. Cross-cohort structural replication, baseline external validity, longitudinal associations and incremental value beyond baseline CES-D total score were evaluated.

RESULTS

The main LCA sample included 43,091 HRS participants and 19,037 ELSA participants. A 4-class solution was selected in both cohorts: Minimal Symptoms, Somatic-Motivational, Affective-Social and Global Distress. The active symptom patterns showed cross-cohort structural correspondence, with item-probability correlations of 0.980 for Somatic-Motivational, 0.828 for Affective-Social and 0.860 for Global Distress. Baseline external-validity profiles were also similar across cohorts. After adjustment for baseline CES-D total score, the Somatic-Motivational pattern was the most consistent prognostic pattern, showing higher odds of future elevated depressive symptom status in ELSA (OR, 1.39; 95% CI, 1.22-1.58) and HRS (OR, 1.53; 95% CI, 1.41-1.65), and higher risk of ADL worsening and mortality in both cohorts. Adding symptom pattern to score-only models improved likelihood-based fit across outcomes and identified observed risks above score-based expectations in the Somatic-Motivational pattern.

CONCLUSIONS

Reproducible depressive symptom patterns can be identified across two national ageing cohorts. The Somatic-Motivational pattern retained prognostic information beyond total depressive symptom severity, supporting symptom configuration as a complementary layer of late-life risk assessment.