The differential role of socioeconomic status dimensions in depressive symptoms of aging adults: data from the Hamburg City Health cohort Study

Background: Socioeconomic status (SES) has consistently been associated with depressive symptoms, however, it remains unclear which subset of SES variables is most relevant to the development of depressive symptoms. This study determined a standardized SES-Index to test the relationship of its sub-dimensions with depressive symptoms.

Methods: HCHS data (N = 10,000; analysis sample n = 8,400), comprising participants 45+ years of age, was used. A standardized approach to quantify SES was employed. Depressive symptoms were quantified using the Patient Health Questionnaire-9 (PHQ-9). Using multiple linear regression models, PHQ-9-scores were modeled as a function of age and sex, and (1a) total SES-Index score versus (1b) its three sub-dimension scores (education, occupational status, income). Models were compared on explained variance and goodness of fit. We determined risk ratios (RR, concerning a PHQ-9 sum score ≥ 10) based on (low, middle, high; 2a) SES-Index scores and (2b) the sub-dimension scores, with groups further differentiated by sex and age (45-64 versus 65+). We distinguished between the total SES-Index score and its three sub-dimension scores to identify relevant SES sub-dimensions in explaining PHQ-9-variability or risk of depression.

Results: Among all regression models (total explained variance 4-6%), income explained most variance, but performance of the SES-Index was comparable. Low versus high income groups showed the strongest differences in depressive trends in middle-aged females and males (RRs 3.57 and 4.91). In older age, this result was restricted to females (RR ≈ 2). Middle-aged males (versus females) showed stronger discrepancies in depressive trends pertaining to low versus high SES groups. In older age, the effect of SES was absent. Education was related to depressive trends only in middle-aged females and males. In an exploratory analysis, marital status and housing slightly increased model fit and explained variance while including somatic symptoms lead to substantial increases (R² _adj = 0.485).

Conclusion: In line with previous research, the study provides evidence for SES playing a significant role in depressive symptoms in mid to old age, with income being robustly linked to depressive trends. Overall, the relationship between SES and depressive trends appears to be stronger in males than females and stronger in mid compared to old age.