Do all psychological treatments really work the same in posttraumatic stress disorder?

Abstract

A recent meta-analysis by Benish, Imel, and Wampold (2008, Clinical Psychology Review, 28, 746-758) concluded that all bona fide treatments are equally effective in posttraumatic stress disorder (PTSD). In contrast, seven other meta-analyses or systematic reviews concluded that there is good evidence that trauma-focused psychological treatments (trauma-focused cognitive behavior therapy and eye movement desensitization and reprocessing) are effective in PTSD; but that treatments that do not focus on the patients' trauma memories or their meanings are either less effective or not yet sufficiently studied. International treatment guidelines therefore recommend trauma-focused psychological treatments as first-line treatments for PTSD. We examine possible reasons for the discrepant conclusions and argue that (1) the selection procedure of the available evidence used in Benish et al.'s (2008)meta-analysis introduces bias, and (2) the analysis and conclusions fail to take into account the need to demonstrate that treatments for PTSD are more effective than natural recovery. Furthermore, significant increases in effect sizes of trauma-focused cognitive behavior therapies over the past two decades contradict the conclusion that content of treatment does not matter. To advance understanding of the optimal treatment for PTSD, we recommend further research into the active mechanisms of therapeutic change, including treatment elements commonly considered to be non-specific. We also recommend transparency in reporting exclusions in meta-analyses and suggest that bona fide treatments should be defined on empirical and theoretical grounds rather than by judgments of the investigators' intent.

Fig. 1

Comparison of individual non-directive treatments with trauma-focused CBT programs. Intent-to-treat analyses for percent remitted (loss of PTSD diagnosis) with treatment. The study marked with an arrow was selected for Benish et al.'s (2008) meta-analysis. Abbreviations: WL/EDU = waitlist or psychoeducation; SUP/PCT = supportive or present-centred therapy; CBT = trauma-focused CBT. Bl03 = Blanchard et al. (2003). Br03 = Bryant et al. (2003). E09 = Ehlers et al. (2009). F91 = Foa et al. (1991). McD05 = McDonagh et al. (2005). N04 = Neuner et al. (2004). Sch07 = Schnurr et al. (2007).

Fig. 2

Effect sizes for changes in PTSD symptoms with treatment for Brom et al.'s (1989) study and for trauma-focused CBT programs (PTSD following trauma in adulthood). In line with Brom et al., effect sizes are based on completers. To ensure the comparison is fair, only studies with similar or lower drop-out rates as in Brom et al. are shown (drop-out rates in parentheses below). However, the pattern would be the same if all trials were shown. Effect sizes were calculated as the pre-post difference in PTSD symptom scores, divided by the pooled standard deviation. Abbreviations: EXP = exposure therapies, CBT = cognitive behavior therapies, Desens = Trauma desensitization, Hypno = hypnotherapy, Psychodyn = psychodynamic therapy; B89 = Brom et al. (1989, 11%). Br08 = Bryant et al. (2008, 17%). E03 = Ehlers et al. (2003, 0%), E05 = Ehlers, Clark, Hackmann, McManus, and Fennell (2005, 0%), E09 = Ehlers et al. (2009, 3%). Fe99 = Fecteau and Nicki (1999, 17%). F99 = Foa et al. (1999, 8%). K03 = Kubany, Hill and Owens (2003; 5%). M98 = Marks et al. (1998, 13%). L02 = Lee et al. (2002, 8%). Ro05 = Rothbaum et al. (2005, 13%). T99 = Tarrier et al. (1999, 11%). V94 = Vaughan et al. (1994, 8%).