The interplay of depression symptoms and physical activity: Bidirectional insights from 25-years of the Americans' changing lives panel

Abstract

Background

Depression symptomology (DSx) and insufficient Physical Activity (PA) are among the leading causes of illness, and major contributors to global public health burden. Reviews and meta-analyses indicate that DSx and PA cause each other, yet most studies conducted use data and analyses which cannot specify their bidirectional associations across the life course.

Method

The present study estimates dynamic panel models with fixed effects through structural equation models with full-information maximum likelihood estimation (ML-SEM) based on 5 waves (1986–2011) of the Americans' Changing Lives (ACL) panel. This is a nationally representative probability sample of 3499 non-institutionalized U.S. adults ages 25 and older in 1986. Respondents participated in an average of 3.29 waves, with 67.84% participating in at least 3 waves, and 27.26% participating in all 5 waves. Models adjust for age, partner status, social integration, activity limitations, and serious financial problems. Data are available from the ICPSR (4690).

Results

Estimates from ML-SEM reveal negative and significant bidirectional cross-sectional associations between DSx and PA. Models also indicate a negative and significant cross-lagged association from DSx to PA, but not from cross-lagged PA to DSx. On both sides, earlier levels are significantly associated with future levels of PA and DSx.

Conclusions

This study is the first to use a causal inference technique which adjusts for all time-invariant confounders while modelling the bidirectional linkages between depression symptoms and physical activity over 25-years of adulthood. It supports literature showing a consistent cross-sectional relationship, and advances understanding on how DSx earlier in the adult life course may influence PA as people age.