I think you are probably correct that suicide is a decent proxy measurement. It is possible that it could be skewed by changes in how likely people are to honestly report suicide; as opposed to calling it an accident.