NZ Government's trend analysis of hospitalised self-harm is misleading


Correspondence to:
Professor John Langley, Injury Prevention Research Unit, Department of Preventive and Social Medicine, Otago University, PO Box 913, Dunedin, New Zealand. Fax: +64 3 479 8337; e-mail:


Objective: The aim of this paper is to demonstrate that the trends published in the New Zealand (NZ) Government's 2006 Suicide Trends document for hospitalised self-harm are misleading.

Methods: Analysis of incident self-harm events resulting in hospitalisation and reference to published material on injury outcome indicators for the NZ Injury Prevention Strategy (NZIPS).

Results: The significant increase in rates of self-harm hospitalisation presented in Suicide Trends from 1989 to a large extent reflect changes in recording practice rather than any change in self-harm in the community. Indicators with significantly fewer threats to validity suggest there has been little, if any, increase in the incidence of self-harm. The authors of Suicide Trends did not adequately specify how they defined a case and, moreover, their methods were not consistent with those used for the NZIPS indicators.

Conclusions and Implications: The methodological challenges to producing valid indicators for the purposes of measuring trends in important non-fatal injury are substantial. Unless we accept that the usual methods of measuring trends in non-fatal injury are misleading and commit to taking up the challenge to produce and use better indicators, we will continue to run the risk of misleading ourselves and the public.