Suicide Prevention and Method Restriction: Evaluating the Impact of Limiting Access to Lethal Means Among Young Australians

Abstract

Given the finite resources allocated to suicide prevention, it is necessary to direct resources into interventions that are most likely to have an impact. This article tests for possible impacts on youth suicides of a cost-intensive Australian policy change (increased firearms restriction) that limited access to a means of suicide. Suicide rates by different age groups and methods were examined for structural breaks, using Zivot-Andrews and Quandt tests. No breakpoint was found in firearm suicide among Australian youth around the time of the 1996 legislative changes. Method restriction in the form of firearms legislation could not be tied to a corresponding impact on youth suicide.

Keywords:

• firearms
• method restriction
• suicide
• youth

Notes

¹This was based on data from the Australian Bureau of Statistics, which may differ slightly from AIHW data. For discussion, see Elnour and Harrison (2009).

Note. **Significant at the 1% level.

*Significant at the 5% level.

Note. **Significant at the 1% level.

*Significant at the 5% level.

N.S: No significant break.

²To discern whether the existing downward trend (apparent from visual inspection of Figure 1) in firearm suicide among the 35 to 44 year age group increased (i.e., “sped up”) after the structural break, further analysis was undertaken to compare the rate of change pre- and post-1997 (via an interaction term in a regression model). A battery of analyses was used, including Poisson regression, negative binomial regression, and simple linear regression. These did not support the hypothesis of a significant difference in the rate of change pre- and post-1997 (firearm suicides declined per year, on average, 5% pre- and 6% post-1997), which suggests the absence of any sustained impact of the structural break on the pre-existing trend for that age group.

Note. N.S = No significant break.

Italicized text indicates a ZA breakpoint that was not found to be significant using a Quandt test.

³Details are available upon request.