Some of us think we have a problem in our suicide research and in our suicide prevention approaches. Actually there may be many problems with those (stay tuned for future blogs) but one of the concerns is the meaning of the statistics when it comes to the definition of “suicide attempt”.
A suicide attempt can be defined as a purposeful self-injury with the intent to die. A self-harm attempt on the other hand can be defined as a purposeful self-injury without the intent to die. Self-harm can be the result of many different factors, including but not limited to: difficulties with problem solving, difficulties with impulse control, copycat phenomenon, social or situational control, etc. Increasingly, research is showing that young people who self-injure may be substantially different from those who attempt suicide. So what does this mean?
Hospitals that use the ICD system (and that is all of them) tend to code self-injury as a suicide attempt. Even DSM at the time of this writing, does not allow for differentiation of self-injury from suicide attempt. Could it be that many of our statistics about suicide attempts are incorrect? Could it be that “truths” that we think we know – such as more females attempt suicide than males – may not be accurate but may be an artifact of not separating out self-harm attempts from suicide attempts?
And what about suicide prevention programs? Does a decrease in reported incidents of self-harm equal a decrease in real suicidal behaviour – that is suicide attempts? That does not mean that we should not try to bring down self-harm attempts, but it may mean that the methods useful for one outcome may not be useful at all for another outcome.
Actually, I think its time that we started to think more critically about what we mean when we use the words “suicide attempt”. Is it really a suicide attempt or is it a self-harm event? It is an important distinction. Both are important targets for interventions – public health type and clinical type. We need to separate them out in our statistics and we need to separate them out in our programs. Then we can get a better handle on what is actually happening and what we can do about it.