This might seem to be a good study, but it isn’t. It concludes that the SADPERSONS scale for suicide risk assessment is not much good. However the end points for testing the sensitivity of the scale were hospitalization or acts of self harm and those are not good endpoints.
the SADPERSONS scale tool is intended to identify those who might be at risk for suicide–suicide gestures and all the other conduct that is characteristic of personality disorders may or may not be an endpoint against which to identify suicide risk.
Probably there is no reliable psych test for this area of risk assessment–completed suicides are done by people with a plan, a real plan, and will they reveal it to a inquiry by a provider?
But part of the problem of assessing for sensitivity and specificity is that there is acting out and then there is suicide –as in successful. We all know the two entities are quite different.
I would distinguish between the tight lipped older man with bad stresses in his life and the immature teenager–wouldn’t you? Which one is likely to be getting hospitalizations or make a suicide gestures like taking pills with rescue available or cutting the arms? Which one is more at risk for committing suicide with an effective, sudden and irreversible means like a gun?