Depressive Symptoms and All-Cause Mortality in a Nationally Representative Longitudinal Study With Time-Varying Covariates

Abstract

OBJECTIVE: To examine the relationship between depressive symptoms and all-cause mortality in a longitudinal study with a nationally representative sample. Research has shown that depressive symptoms increase mortality risk, but results have been inconclusive regarding the role of physical health conditions in the relationship. This study asks whether the association between depressive symptoms and mortality exists independent of contemporaneous physical health conditions, is spurious because of prior physical health conditions, or is mediated by later physical health conditions. METHODS: Data are drawn from the Americans' Changing Lives Study, a sample of 3617 noninstitutionalized Americans aged 25 years or older. Respondents were interviewed in 1986, 1989, 1994, and 2002. Depressive symptoms (Center for Epidemiologic Studies Depression Scale [CES-D]), physical health, and confounders were measured at each wave. Mortality status was ascertained yearly through 2007. Discrete time hazard models with time-varying covariates were used to estimate the association between CES-D scores and mortality. RESULTS: Between 1986 and 2007, 1411 survey respondents died. Depressive symptoms were associated with mortality after adjusting for stress, coping characteristics, social support, and health behaviors (odds ratio [OR] = 1.23, 95% confidence interval [CI] = 1.11-1.36, p < .001). However, the association became nonsignificant after accounting for contemporaneous physical health conditions (OR = 1.06, 95% CI = 0.95-1.17, p = .31). Prior physical health conditions did not explain the association (OR = 1.24, 95% CI = 1.11-1.39, p < .001). The association between lagged depressive symptoms and mortality was mediated by later physical health conditions (p = .94). CONCLUSIONS: Study findings support the mediation hypothesis. The effect of depressive symptoms on mortality is mediated by later physical health.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 30 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.