The prediction of dangerous behavior in emergency civil committment

The authors compared the records of 59 psychiatric patients involutarily committed to a Veterans Administration hospital on an emergency basis with those of a control group of 59 psychiatric patients with respect to the number of assaults noted during the first 45 days of hospitalization. The committed group had a .41 probability and the control group a .08 probability of committing an assault. The difference between the two groups was mainly accounted for by assaults that occurred during the first 10 days of hospitalization. The occurrence of an actual act of battery before admission did not predict assault in the hospital to a greater degree than did a verbal threat. The authors conclude that short-term clinical predictions of dangerousness predict assaultiveness in the hospital to a significant degree.