Time series analysis of alcohol consumption and suicide mortality in the United States, 1934-1987.

OBJECTIVE: The objective is to analyze time series data on per capita alcohol consumption and suicide mortality between 1934 and 1987. Specifically, the aim is to test the hypothesis that increases in per capita consumption are associated with increases in suicide mortality overall and in gender and age subgroups. This study also examines whether the effect of per capita consumption on subsequent rates of suicide is influenced by other aggregate factors, particularly unemployment rates, per capita income and divorce rates. METHOD: This analysis examines the temporal structure of suicide mortality, alcohol consumption and covariate time series data in the United States, 1934-1987, using the Box-Jenkins method. RESULTS: Bivariate associations between alcohol consumption and suicide rates were not significant. However, when unemployment was included in the model, increases in per capita alcohol consumption were significantly related to increases in suicide overall, for men and women, and for the young (under age 40) and middle-aged (40 to 59 years), but not for those over age 60. CONCLUSIONS: The use in modeling of other known covariates of suicide rates, particularly unemployment, affects the magnitude of the relationship between alcohol consumption and suicide. Consistent with other studies, unemployment was significantly related to suicide and was shown to confound the relationship between alcohol and suicide. The effects of changes in alcohol consumption on changes in suicide rates differ by gender and age group, which demonstrates that focusing on the total population alone can mask divergent effects that can cancel each other out when subgroups are combined.