Nicely Done Piece of Analysis

Nicely Done Piece of Analysis

Maybe I'm just in a really good mood, but this article Steven P. Segal and Philip M. Burgess, "Effect of Conditional Release From Hospitalization on Mortality Risk," Psychiatric Services, 57:1607-13 [November 2006] really impressed me. It is an analysis of the effects on mortality risk for psychiatric patients under what Victoria, Australia calls a community treatment order, or what we would call involuntary outpatient commitment (IOC) in the United States. It wasn't just that the results were encouraging (a 14% reduction in noninjury-related deaths compared to similar psychiatric patients who were simply released), but I was very impressed with the various methods that they used to look for selection bias problems as well.

One of the problems with studies of this nature is that they are often comparing psychiatric patients who were under IOC with patients that were not--and those under IOC tended to be substantially different in either history of violence, severity of illness. This study seems to have done a nice job of making sure that even though the populations in aggregate are different, that they are comparing the results for similar groups in both populations.