"Because of these findings, we recommend that, until we have evidence from head-to-head trials favoring one treatment or the other, clinicians should make shared decisions, with patients, about which treatment modality to use," said Jeffrey Sonis, MD, MPH, lead author of the study and an associate professor in the departments of social medicine and family medicine in the University of North Carolina School of Medicine. "These decisions should be based on patient preferences with regard to several treatment characteristics, such as the frequency of visits and time between treatment initiation and the onset of beneficial effects."

The systematic review and meta-analysis was led by Sonis and is published in the December 2019 issue of the journal Psychiatry Research. Joan M. Cook, PhD, of Yale University is co-author of the study.

Most guidelines for the treatment of PTSD in adults recommend that trauma-focused psychotherapy be used as a first-line treatment and medications be used as a second-line treatment, Sonis said. In other words, they recommend that psychotherapy should be offered, preferentially, to adults with PTSD, over medication. Those guidelines base those recommendations upon the fact that the effect sizes (the magnitude of the treatment effects) from randomized trials that compare psychotherapy to controls (such as wait-list controls) are substantially larger than the effect size from trials comparing medications to placebo.

However, comparing effect sizes based on indirect comparison is a flawed approach, Sonis said, because methodological differences between psychotherapy trials and medication trials are known to be associated with the effect size. For instance, blinding or masking, which is routine in medication trials but impossible in psychotherapy trials is known to be associated with smaller effect sizes. Other methodological differences are also associated with effect size and may explain the difference in effect sizes for psychotherapy and medication trials.

"We believe that only head-to-head randomized trials comparing psychotherapy directly to medications should be used to determine the relative effectiveness of psychotherapy and medications for treatment of PTSD," Sonis said. "There have been other meta-analyses of head-to-head randomize trials for treatment of PTSD previously but several head-to-head trials had been published since the last meta-analysis. Therefore, we conducted a meta-analysis exclusively consisting of head-to-head randomized trials."

"We found that the best estimate of the effect, comparing psychotherapy and medications, was that there was no difference between the two. However, the 95% confidence interval was very wide, indicating that true effect may favor psychotherapy or it may favor medications. We concluded that there is still insufficient evidence to determine whether psychotherapy or medications were more effective for treatment of PTSD in adults," Sonis said.

Until there is clear evidence from head-to-head trials favoring one treatment or the other, Sonis said, clinicians should make shared decisions, with patients, about which treatment modality to use, based not on comparative effectiveness (which is inconclusive) but on patient preferences regarding the following factors:

• presumed mechanism of action
• what is required of the patient
• whether the patient will need to talk about the traumafrequency of visits
• time between treatment initiation and onset of beneficial effects
• duration of benefit after treatment cessation
• risk and adverse effects

"The bottom line is that while many researchers and clinicians believe that psychotherapy is more effective for treatment of PTSD than medications, our research shows that there is—as of right now—insufficient evidence from head-to-head trials to make that determination at this time," Sonis said. "Therefore, the choice of medication or psychotherapy as the initial treatment should be based on patient preferences for treatment characteristics and not on incorrect assumptions about which treatment is more effective."