Jacqueline Sin, Debbie Spain, Marie Furuta, Trevor Murrells, Ian Norman

Increasing evidence indicates that individuals who develop severe mental illness (SMI) are also vulnerable to developing post-traumatic stress disorder (PTSD), due to increased risk of exposure to traumatic events and social adversity. The effectiveness of trauma-focused psychological interventions (TFPIs) for PTSD in the general population is well-established. TFPIs involve identifying and changing unhelpful beliefs about traumatic experiences, processing of traumatic memories, and developing new ways of responding to cues associated with trauma. Little is known about the potential feasibility, acceptability and effectiveness of TFPIs for individuals who have a SMI and PTSD. To evaluate the effectiveness of psychological interventions for PTSD symptoms or other symptoms of psychological distress arising from trauma in people with SMI. We searched the Cochrane Schizophrenia Group's Trials Study-Based Register (up until March 10, 2016), screened reference lists of relevant reports and reviews, and contacted trial authors for unpublished and/or specific outcome data. We included all relevant randomised controlled trials (RCTs) which investigated TFPIs for people with SMI and PTSD, and reported useable data. Three review authors (DS, MF, IN) independently screened the titles and abstracts of all references identified, and read short-listed full text papers. We assessed risk of bias in each case. We calculated the risk ratio (RR) and 95% confidence interval (CI) for binary outcomes, and the mean difference (MD) and 95% CI for continuous data, on an intention-to-treat basis. We assessed quality of evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) and created 'Summary of findings' tables. Four trials involving a total of 300 adults with SMI and PTSD are included. These trials evaluated three active intervention therapies: trauma-focused cognitive behavioural therapy (TF-CBT), eye movement desensitisation and reprocessing (EMDR), and brief psychoeducation for PTSD, all delivered via individual sessions. Our main outcomes of interest were PTSD symptoms, quality of life/well-being, symptoms of co-morbid psychosis, anxiety symptoms, depressive symptoms, adverse events and health economic outcomes. 1. TF-CBT versus usual care/waiting list Three trials provided data for this comparison, however, continuous outcome data available were more often found to be skewed than unskewed, leading to the necessity of conducting analyses separately for the two types of continuous data. Using the unskewed data only, results showed no significant differences between TF-CBT and usual care in reducing clinician-rated PTSD symptoms at short term (1 RCT, n =13, MD 13.15, 95% CI -4.09 to 30.39,low-quality evidence). Limited unskewed data showed equivocal results between groups in terms of general quality of life (1 RCT, n = 39, MD -0.60, 95% CI -4.47 to 3.27, low-quality evidence), symptoms of psychosis (1 RCT, n = 9, MD -6.93, 95% CI -34.17 to 20.31, low-quality evidence), and anxiety (1 RCT, n = 9, MD 12.57, 95% CI -5.54 to 30.68, very low-quality evidence), at medium term. The only available data on depression symptoms were skewed and were equivocal across groups at medium term (2 RCTs, n = 48, MD 3.26, 95% CI -3.66 to 10.18, very low-quality evidence). TF-CBT was not associated with more adverse events (1 RCT, n = 100, RR 0.44, 95% CI 0.09 to 2.31, low-quality evidence) at medium term. No data were available for health economic outcomes. Very limited data for PTSD and other symptoms were available over the long term. 2. EMDR versus waiting listOne trial provided data for this comparison. Favourable effects were found for EMDR in terms of PTSD symptom severity at medium term but data were skewed (1 RCT, n = 83, MD -12.31, 95% CI -22.72 to -1.90, very low-quality evidence). EMDR was not associated with more adverse events (1 RCT, n = 102, RR 0.21, 95% CI 0.02 to 1.85, low-quality evidence). No data were available for quality of life, symptoms of co-morbid psychosis, depression, anxiety and health economics.3. TF-CBT versus EMDROne trial compared TF-CBT with EMDR. PTSD symptom severity, based on skewed data (1 RCT, n = 88, MD -1.69, 95% CI -12.63 to 9.23, very low-quality evidence) was similar between treatment groups. No data were available for the other main outcomes.4. TF-CBT versus psychoeducationOne trial compared TF-CBT with psychoeducation. Results were equivocal for PTSD symptom severity (1 RCT, n = 52, MD 0.23, 95% CI -14.66 to 15.12, low-quality evidence) and general quality of life (1 RCT, n = 49, MD 0.11, 95% CI -0.74 to 0.95, low-quality evidence) by medium term. No data were available for the other outcomes of interest. Very few trials have investigated TFPIs for individuals with SMI and PTSD. Results from trials of TF-CBT are limited and inconclusive regarding its effectiveness on PTSD, or on psychotic symptoms or other symptoms of psychological distress. Only one trial evaluated EMDR and provided limited preliminary evidence favouring EMDR compared to waiting list. Comparing TF-CBT head-to-head with EMDR and brief psychoeducation respectively, showed no clear effect for either therapy. Both TF-CBT and EMDR do not appear to cause more (or less) adverse effects, compared to waiting list or usual care; these findings however, are mostly based on low to very low-quality evidence. Further larger scale trials are now needed to provide high-quality evidence to confirm or refute these preliminary findings, and to establish which intervention modalities and techniques are associated with improved outcomes, especially in the long term.