This article first appeared in the British Medrcal]ournol (Vol. 310. he 1995: 1479-14801. The’bMJ Publishing Group ig the copyright holder of this article and it is reproduced with their kind permission. Some people believe strongly that talking through traumatic or stressful experiences may help the psychological recovery of those who have suffered psychological wounding. This belief has led to military psychiatrists providing immediate interventions at the front line that are intended to heal and return soldiers to activity (Soloman & Bendenishty 1986), and it has fuelled psychological debriefing methods such as Mitchell’s critical incident stress debriefing (Dunning 1988). Debriefing programmes have expanded rapidly, reflecting a powerful social movement that sees them as meeting the needs of workers in emergency services, victims of disasters, and those who have been affected by trauma and violence in wider society (Pennebaker & Susman 1988; Ersland et al 1989; Shapiro Kunkler 1990). Yet, although military psychiatry has succeeded in returning troops to action, outcome studies suggest that soldiers who are repeatedly traumatised by returning to combat may suffer even higher rates of severe and chronic posttraumatic stress disorder in the longer term (Soloman & Bendenishty 1986). The proponents of debriefing agree that it needs evaluation, yet there is a dearth of systematic evaluation and outcome studies. In one of the few systematic evaluations, many subjects reported two weeks after debriefing that they had found it helpful and felt less stressed (Robinson & Mitchell 1993). Nevertheless, 41% of the group of chiefly female welfare workers (n-65) and almost 6% of the group of chiefly male emergency workers (n-102) reported the traumatic experience to be ‘still having considerable or great impact on them,’ with a high frequency of cognitive and other symptoms. Symptoms were not, however, assessed systematically with standard measures. Randomised controlled trials of the effectiveness of debriefing have not been reported, although a few studies include comparison with a group that was not debriefed. A study of fnefighters two weeks after they dealt with a hotel fire in Norway showed that most of those who attended debriefing reported that it had helped and increased their self confidence. Their scores for intrusive thoughts and avoidance behaviour measured by the impact of events scale (Horowitz et al 1979) were no different from those of the group who had simply talked to their colleagues informally (Hytten & Hasle 1989). Nearly half of a group of emergency workers surveyed one year after attending serious bus crashes still reported considerable symptoms, and 13% thought that they would probably not recover (Griftiths & Watts 1992). Those who had been debriefed (182 of 285) had significantly higher scores for morbidity and distress on the general health questionnaire (Goldberg & Hillies 1979) and the impact of events scale. These findings provide little evidence that the debriefing, even though perceived as helpful, was effective in preventing negative outcomes. The group with high distress might have been worse without it, but the study was unable to show this. Similar findings were obtained in a longitudinal study of 195 people who had helped after an earthquake, of whom 62 had been debriefed and 133 had not (Kenardy et al in press). Screening for degree of stressfulness, threat, and psychological exposure over the subsequent two years found a general decrease in symptoms, with less improvement over time among those who had been debriefed, even though 80% rated the debriefing as helpful. Clearly, neither perceived helpfulness nor experience of debriefing was associated with more positive outcomes. These studies were not controlled trials. The groups may not have been comparable, perhaps experiencing different stressors (loss rather than trauma), having uncertain roles, and having more welfare or counselling functions, for which the debriefing model was inappropriate, or perhaps being more distressed at the outset. But it is conceivable that debriefing may exaggerate the traumatic process (Kenardy et al in press; Watts R 1994) or even be associated with a delayed presentation. In addition, exposure to informational social support, which forms part