Format

Send to

Choose Destination
J Affect Disord. 2011 Mar;129(1-3):275-81. doi: 10.1016/j.jad.2010.08.010. Epub 2010 Sep 15.

Decomposing the association of completed suicide with air pollution, weather, and unemployment data at different time scales.

Author information

1
Department of Psychiatry, Chu-Tung Veterans Hospital, Hsin-Chu County, Taiwan. ccyang@physionet.org

Abstract

BACKGROUND:

Research has implicated environmental risk factors, such as meteorological variables, in suicide. However, studies have not investigated air pollution, known to induce acute medical conditions and increase mortality, in suicide. This study comprehensively assesses the temporal relationship between suicide and air pollution, weather, and unemployment variables in Taipei City from January 1 1991 to December 31 2008.

METHODS:

This research used the empirical mode decomposition (EMD) method to de-trend the suicide data into a set of intrinsic oscillations, called intrinsic mode functions (IMFs). Multiple linear regression analysis with forward stepwise method was used to identify significant predictors of suicide from a pool of air pollution, weather, and unemployment data, and to quantify the temporal association between decomposed suicide IMFs with these predictors at different time scales.

RESULTS:

Findings of this study predicted a classic seasonal pattern of increased suicide occurring in early summer by increased air particulates and decreased barometric pressure, in which the latter was in accordance with increased temperature during the corresponding time. Gaseous air pollutants, such as sulfur dioxide and ozone, were found to increase the risk of suicide at longer time scales. Decreased sunshine duration and sunspot activity predicted the increased suicide. After controlling for the unemployment factor, environmental risks predicted 33.7% of variance in the suicide data.

CONCLUSIONS:

Using EMD analysis, this study found time-scale dependent associations between suicide and air pollution, weather and unemployment data. Contributing environmental risks may vary in different geographic regions and in different populations.

PMID:
20828830
DOI:
10.1016/j.jad.2010.08.010
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center