Structural intervention at one bridge decreases the overall jumping suicide rate in Victoria, Australia

Abstract Aims There is clear evidence that installing safety barriers is effective in preventing jumping suicides from high-risk bridges with only moderate displacement to other nearby bridges. However, the impact of barriers on jumping suicides across broader geographical areas is not well understood. We examined patterns in jumping suicides across the state of Victoria, Australia, after a safety barrier was installed at the West Gate Bridge which, before the installation of the barrier, was the site of approximately 40% of Victoria’s jumping suicides. Methods We used negative binomial regression analyses on Victorian data from 2000 to 2019 to compare rates of jumping suicides at the West Gate Bridge, other bridges and non-bridge jumping locations before, during and after the West Gate Bridge barrier installation. We conducted linear regression analyses to examine whether the distance travelled from the deceased’s usual residence to the location of their jumping suicide changed between the before, during and after barrier installation periods. Results After installation of the barrier, there were no jumping suicides at the West Gate Bridge (rate ratio [RR] = 0.00, 95% credible intervals [95% Cr] = 0.00–0.0001) and there was strong evidence that the rate of jumping suicides at all locations declined by 65% (RR = 0.35, 95% Cr = 0.22–0.54). At other bridges, there was also evidence of a reduction (RR = 0.31, 95% Cr = 0.11–0.70), but there was no evidence of a change at non-bridge locations (RR = 0.74, 95% Cr = 0.39–1.30). Conclusion After installation of the safety barrier at the West Gate Bridge, jumping suicide in Victoria decreased overall and at other bridges, and did not appear to change at non-bridge locations. Our findings show that when barriers are installed at a site responsible for a disproportionately high number of jumping suicides, they are not only highly effective at the site where the barriers are installed but can also have a prevention impact beyond the immediate locale at similar sites.