Answers are grounded in this paper and cite their sources.
Background Trauma-focused psychotherapies (TFPs) effectively treat post-traumatic stress disorder (PTSD), yet patients show variable response or discontinue treatment. Numerous augmentation strategies have been developed to address these barriers, but the field lacks a framework that extends beyond mechanistic categories as well as a meta-analytic synthesis of their effects.Objective To develop a classification framework for non-pharmacological augmentation strategies of first-line TFP for PTSD and to meta-analytically evaluate whether these strategies improve treatment outcomes.Method We conducted a systematic review and meta-analysis of randomised controlled trials (RCTs) comparing TFP augmented with a non-pharmacological intervention versus TFP alone or an active control (i.e. sham intervention or dose increase). Eight databases were searched to July 2026. Augmentation strategies were qualitatively categorised. Primary outcomes were PTSD symptom severity and response rate; dropout was a secondary outcome.Results Fifty RCTs met inclusion criteria. Their augmentations were categorised within a three-level framework comprising targets, mechanisms, and implementation forms. Of these, 39 trials (40 comparisons; N = 2,853) were meta-analysed and showed no significant overall effects (symptom severity: g = –0.10, p = .062; response: RR = 1.04, p = .111; dropout: RR = 0.94, p = .489). One subgroup targeting sleep disturbance showed an add-on effect.Conclusions The present findings suggest that augmenting TFP with non-pharmacological interventions does not necessarily lead to incremental efficacy at the overall level, even though small benefits were observed for sleep-focused strategies. Our three-level framework provides a structured approach for conceptualising this heterogeneous field. The absence of a robust overall effect should not be mistaken for evidence that augmentation cannot be effective.
No takes yet. Share an insight, caveat, or question.
Langhammer et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: