Studies in academic research centres with selected patients have shown that several cognitive behaviour therapies are effective in the treatment of PTSD following traumas affecting individuals or small groups. Little information is available on the extent to which these positive findings will generalize to more routine clinical settings with less selected patients or to a trauma that affects a whole community. The present study addresses these generalization issues. A consecutive series of 91 patients with PTSD resulting from a car bomb which exploded in the centre of Omagh, Northern Ireland in August 1998 were treated with cognitive therapy, along lines advocated by Ehlers and Clark (2000). There were no major exclusion criteria and 53% of patients had an additional axis I disorder (comorbidity). Therapists were NHS staff with heavy caseloads and modest prior training in CBT for PTSD. A brief training in specialist procedures for PTSD was provided. Patients received an average of eight treatment sessions. Significant and substantial improvements in PTSD were observed. Degree of improvement was comparable to that in previously reported research trials. Comorbidity was not associated with poorer outcome, perhaps because comorbid patients were given more sessions of treatment (average 10 vs 5 sessions). Patients who were physically injured improved less than those who were not physically injured. Overall, the results indicate that the positive findings obtained in research settings generalize well to a frontline, non-selective service.