> The data collection process for reported suicides would be fairly stable > and consistent. A suicide is a subset of all unnatural death (UD) cases. > These are difficult to manipulate, unlike some other crime statistics, due > to the requirement to dispose of a dead body after following the legal > process. All UD cases routinely go for autopsy and suicide as a mode of > death gets established at the end of the inquest.
I'd read a report (sadly, can't find it) that much of the suicide or household accident cases went unreported earlier and suicides were reported as illnesses to avoid autopsy which is slowly being fixed. I remember in 2001 when someone I know tried to slit his wrist (largely for show I think, he used a car key and managed only get seriously scratched) - when we went to the hospital they told us not to admit because the cops had become stringent and required registration for every attempted suicide case etc. But the quantum some psychiatrists think it really is 95 per lakh rather than the reported 9 to 11 - http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2913651/) which, if true, means that my argument that data collection is better has yielded only marginally better data :) Either ways - I concede that the suicide rate has increased - we are at 11.4 officially versus, 5.8 per lakh in 1981. More than doubled in thirty years!
