Although many studies have addressed the topic of stability versus change in depressive symptoms, few have further decomposed the change to continuous accumulation versus non-systematic state fluctuations or measurement errors. This further step requires a longitudinal follow-up and an appropriate stochastic model; it would, for example, evaluate the hypothesis that women accumulate more susceptibility events than men.

A linear stochastic differential equation model was estimated for a 16-year longitudinal course of depressive symptoms in the Young Finns community sample of 3596 participants (1832 women, 1764 men). This model enabled us to decompose the variance in depression symptoms into a stable trait, cumulative effects and state/error fluctuations.

Women showed higher mean levels and higher variance of depressive symptoms than men. In men, the stable trait accounted for the majority [61%, 90% confidence interval (CI) 48.9–69.2] of the total variance, followed by cumulative effects (23%, 90% CI 9.9–41.7) and state/error fluctuations (16%, 90% CI 5.6–23.2). In women, the cumulative sources were more important than among men and accounted for 44% (90% CI 23.6–58.9) of the variance, followed by stable individual differences (32%, 90% CI 18.5–54.2) and state fluctuations (24%, 90% CI 19.1–27.3).

The results are consistent with previous observations that women suffer more depression than men, and have more variance in depressive symptoms. We also found that continuously accumulating effects are a significant contributor to between-individual differences in depression, especially for women. Although the accumulating effects are often confounded with non-systematic state fluctuations, the latter are unlikely to exceed 27% of the total variance of depressive symptoms.