Best strategies for reducing the suicide rate in Australia: Response to Pirkis

To the Editor

We welcome Prof Pirkis’ commentary on our article and appreciate her support for our approach to informing communities as to the optimal evidence-based approaches to reducing suicides in Australia. We note that Prof Pirkis adopted a different formula to calculating Population Preventable Fraction (PPF). As noted in the paper, the approach adopted the formula provided by Rockhill et al. (1998), also presented in Kleinbaum et al. (1982) as formula 9.13. The formula for PPF, as per Kleinbaum et al. (1982)

$$PPF=\frac{p\left(1-RR\right)}{p\left(1-RR\right)+RR}$$

where p is the estimated proportion of suicide attempt/death cases exposed to the strategy and RR is the risk ratio for the effect of that strategy. This approach is appropriate for use when risk ratio (RR) and case fraction (i.e. proportion of cases exposed, reflecting our estimates of prevalence) are available. It is difficult to comment on the alternative formula used by Prof Pirkis, given that her formula was not cited and is not used elsewhere in literature presenting PPF estimates (e.g. Omer et al., 2011).

Regarding our estimates of RR, we noted that these are the best available estimates and that there is an absence of robust data available for many of the strategies. Although RR may not be consistent for the whole population, we endeavoured to identify programmes that were tested in settings consistent with each strategy.

We acknowledge that it is difficult to account for existing efforts in suicide prevention. However, we did not assume baseline of zero across all strategies as Prof Pirkis suggests. For example, psychosocial treatments were assumed to increase from a baseline of 31.6% to 50% as indicated in the method. We agree that considerable and valuable efforts have been conducted to promote media guidelines in Australia, but note that the impact of this strategy on suicide deaths (even assuming the low baseline provided by the cited Austrian example) is likely to be relatively small. There has been considerable resistance to school-based suicide prevention programmes in Australia, so implementation of this strategy is starting from a very low baseline. We concur that data on means restriction are complicated by the range of potential interventions available, and should be tailored to the specific needs of individual communities.

There are many challenges in preventing suicide. Although we are unlikely to halve the suicide rate in the coming years, using available evidence to identify the most impactful suicide prevention methods in Australia, and working together at the community level to implement these methods is likely to have greatest impact.