Individual-level exposure to disaster, neighborhood environmental characteristics, and their independent and combined associations with depressive symptoms in women

Soc Psychiatry Psychiatr Epidemiol. 2017 Sep;52(9):1183-1194. doi: 10.1007/s00127-017-1412-y. Epub 2017 Jun 27.

Abstract

Purpose: The severity of the stress response to experiencing disaster depends on individual exposure and background stress prior to the event. To date, there is limited research on the interaction between neighborhood environmental stress and experiencing an oil spill, and their effects on depression. The objective of the current study was to assess if the association between exposure to the Deepwater Horizon Oil Spill (DHOS) and depressive symptoms varied by neighborhood characteristics.

Methods: US Census data (2010) and longitudinal data collected in two waves (2012-2014 and 2014-2016) from female residents [N = 889 (Wave I), 737 (Wave II)] of an area highly affected by the DHOS were analyzed. Multilevel and individual-level negative binomial regressions were performed to estimate associations with depressive symptoms in both waves. An interaction term was included to estimate effect modification of the association between DHOS exposure and depressive symptoms by neighborhood characteristics. Generalized estimating equations were applied to the negative binomial regression testing longitudinal associations.

Results: Census tract-level neighborhood characteristics were not associated with depressive symptoms. Exposure to the DHOS and neighborhood physical disorder were associated with depressive symptoms cross-sectionally. There was no evidence of effect modification; however, physical/environmental exposure to the DHOS was associated with increased depressive symptoms only among women living in areas with physical disorder. Exposure to the DHOS remained associated with depressive symptoms over time.

Conclusions: Findings support the enduring consequences of disaster exposure on depressive symptoms in women and identify potential targets for post-disaster intervention based on residential characteristics.

Keywords: Depression; Disaster; Multilevel modeling; Neighborhood/place; Social epidemiology.

MeSH terms

  • Adolescent
  • Adult
  • Aged
  • Aged, 80 and over
  • Depression / epidemiology*
  • Disasters / statistics & numerical data*
  • Environmental Exposure / statistics & numerical data*
  • Female
  • Humans
  • Longitudinal Studies
  • Louisiana / epidemiology
  • Middle Aged
  • Residence Characteristics / statistics & numerical data*
  • Risk Factors
  • Young Adult