Letters

Author's reply

BMJ 1996; 313 doi: http://dx.doi.org/10.1136/bmj.313.7061.884a (Published 05 October 1996) Cite this as: BMJ 1996;313:884
  1. Louis Appleby, Professor
  1. School of Psychiatry and Behavioural Sciences, Withington Hospital, Manchester M20 8LR

    EDITOR,—Simon Hatcher makes an important general point about statistical significance, but it does not apply to our study as we did not make the statement that he attributes to us. Our conclusion that risk factors for suicide in the general population do not identify people with mental illness who go on to commit suicide was based not on the absence of significant differences between cases of suicide and controls but on the fact that the rate of some putative risk factors was virtually identical in the two groups. We drew no conclusion about the duration of illness, for the reasons that Hatcher gives.

    We suspect that the difficulty in predicting suicide from risk factors is explained partly by findings such as ours and partly by the oversimplicity of risk factor models. Risk of suicide may be better understood as a balance of both risk and protective factors, with suicide occurring after a sequence of key events in which risk accumulates. Our research is now directed at identifying models of cumulative risk which may be the target of preventive intervention.