Commentary on Raghuraman et al. (2021). On the Long-Term Efficacy and Effectiveness of Narrative Exposure Therapy
In the past years, several systematic reviews and meta-analyses have been published assessing the eﬀectiveness of narrative exposure therapy [NET; (Lely et al., 2019; Raghuraman et al., 2021; Siehl et al., 2021; Wei and Chen, 2021)]. The meta-analyses had diﬀerent aims and came to diﬀerent conclusions about the eﬀectiveness of NET. Lely et al. (2019) and Wei and Chen (2021) focused on between-treatment eﬀects, post intervention, comparing NET with active and non-active control- treatment-conditions. In addition to comparing between-treatment eﬀects Raghuraman et al. (2021) and Siehl et al. (2021) assessed also the temporal stability of the eﬀects. The latter two studies investigated the reduction of symptom severity of posttraumatic stress disorder (PTSD) and the percentage of PTSD diagnoses over several follow-up periods. Raghuraman et al. (2021) indicated a medium standardized mean diﬀerence (SMD) in favor of NET in comparison to active and inactive control groups in the long-term and no beneﬁt regarding PTSD diagnoses. The authors cautioned against using the existing evidence to inform policies and guidelines. In contrast, Siehl et al. (2021) found a large SMD in favor of NET compared to active or inactive control groups in the long-term. They reported an improvement of eﬀectiveness over time when analyzing active control groups and concluded that NET is an eﬀective treatment approach in post-conﬂict settings and refugee populations, highlighting the high external validity of the trials. Acknowledging the signiﬁcant eﬀorts of both author groups to select, code, and analyze the existing evidence, we aim to clarify potential underlying reasons for the diﬀerences between the two meta-analyses. The purpose of this commentary is two-fold: (a) discuss more generally ways to assess the quality of a treatment, such as NET, that is used in a broad range of contexts and (b) more speciﬁc diﬀerences between the two meta-analyses in (1) selecting and analyzing strategies, and (2) potential coding errors.