Probing the smoking-suicide association: do smoking policy interventions affect suicide risk?

Abstract

Introduction: Smokers exhibit elevated risk for suicide, but it is unknown whether smoking interventions reduce suicide risk. We examined whether state-level policy interventions-increases in cigarette excise taxes and strengthening of smoke-free air laws-corresponded to a reduction in suicide risk during the 1990s and the early 2000s. We also examined whether the magnitude of such reductions correlated with individuals' predicted probability of smoking, which would be expected if the associations stemmed from changes in smoking behavior.

Methods: We paired individual-level data on suicide deaths from the U.S. Multiple Cause of Death files, years 1990-2004, with living population data from the same period. These were linked with state data on cigarette excise taxes and smoke-free air policies. Utilizing a quasiexperimental analytical approach, we estimated the association between changes in policy and suicide risk. To examine whether associations correlated with individuals' probability of smoking, we used external survey data to derive a predicted probability of smoking function from demographic variables, which was then used to stratify the population by predicted smoking prevalence.

Results: Cigarette excise taxes, smoke-free air policies, and an index combining the two policies all exhibited protective associations with suicide. The associations were strongest in segments of the population where predicted smoking prevalence was the highest and weaker in segments of the population where predicted smoking prevalence was the lowest, suggesting that the protective associations were related to changes in smoking behavior.

Conclusion: These results provide support for the proposition that population interventions for smoking could reduce risk for suicide.

Figure 1.

Ratio of suicide rate to national average in 10 states that implemented the strongest tobacco control policies between 1990 and 2004 (black), and 10 states that implemented the least strong policies during this period (gray). Strength of policy is gauged by change in the policy index described in Methods section, which combines excise taxes and smoke-free air scores. Trend lines are 5-year moving-window averages of the plotted ratio. Please note that because suicide rates are based on population counts (as opposed to statistical sampling), confidence intervals are not applicable.