Researchers assessed predictors of suicidal events and non-suicidal self-harm in a group of 334 moderately to severely depressed adolescents (ages 12-18) who had not responded to at least eight weeks of SSRI treatment. Participants were switched to one of four treatments: a different SSRI, venlafaxine, a different SSRI plus cogni- tive-behavior therapy (CBT), or venlafax- ine plus CBT. Initially, the researchers examined suicidal events and non-suicidal self-harm through spontaneous reports from participants. Halfway through the study, because of concerns raised by the FDA about antidepressants and suicidality, researchers switched to using a more sys- tematic method of assessing suicidality and self-harm, asking specific questions about suicidal intent and behavior on a weekly basis. (Not surprisingly, researchers detected that participants reported signifi- cantly more suicidal events (20.9% vs. 8.8%) and non-suicidal self-injury events (17.6% vs. 2.2%) when the more systematic method of evaluating suicidality was utilized.) Generally, there were no significant differ- ences between treatment groups; however, patients whose baseline suicidal ideation was higher than average were significantly more likely to experience a suicidal event on venlafaxine during the study than patients receiving SSRIs (37.2% vs. 23.3%). Regardless of treatment, patients with any of three characteristics – a history of non- suicidal self-injury, severe drug use, or serious family conflict – were more likely to experience suicidal events during the study (Brent DA et al., Am J Psychiatry 2009;166:418-426).