Intervention components, mediators, and mechanisms of change of Internet- and mobile-based interventions for post-traumatic stress disorder: protocol for a systematic review and meta-analysis

Post-traumatic stress disorder (PTSD) is a burdensome [1] and highly prevalent disorder with an estimated 6.8% lifetime prevalence rate of PTSD for the adult US population [2]. PTSD also ranks under the 15 mental and neurological conditions that are most threatening for disability-adjusted life years in Europe [3], but prevalence rates and exposure to potential traumatic events may vary between countries [4]. In DSM-5, PTSD is classified as among the group of trauma- and stressor-related disorders following exposure to a traumatic or stressful event with the diagnostic criteria including at least one stressor (e.g., exposure to threatened death, sexual violence), intrusion symptoms, avoidance, negative alterations in cognitions and mood, alterations in arousal and reactivity, a duration of at least 1 month, functional significance, and exclusion of other explanatory factors (e.g., substance use [5]). The current draft of the ICD-11 diagnostic criteria suggests a symptom cluster of re-experiencing, avoidance, and enduring perceptions of heightened current threat; furthermore, the upcoming ICD-11 will also allow the diagnosis of complex PTSD which is additionally characterized by problems in affect regulation, beliefs about oneself as diminished, defeated, or worthless, accompanied by feelings of shame, guilt, or failure related to the traumatic event and difficulties in sustaining relationships and in feeling close to others [6].

Psychological interventions to treat PTSD are numerous with exposure-based therapy, cognitive processing therapy (CPT), cognitive therapy (CT), therapies based on cognitive behavioral therapy (CBT) principles, eye movement desensitization and reprocessing (EMDR), and narrative exposure therapy showing at least moderate effects ranging from standardized mean differences (SMD) of − 2.0 to − 0.3 [7]. Patient preferences demonstrate a threefold higher preference for psychological treatments compared to medication [8]. Still, although there is a well-established evidence-base for effective psychotherapeutic face-to-face treatments for this condition [7, 9], only a fractional number of patients receive treatment [10‐12]. For example, in a survey interviewing US veterans with self-reported mental disorders (i.e., PTSD, generalized anxiety, major depression) after duty in Iraq and Afghanistan, only 23 to 40% stated that they received any professional help [10]. Another study examining asylum seekers in the Netherlands reports that 54% of the participants originating from Asian and African countries diagnosed with PTSD did not receive any treatment [13]. Reasons for the limited uptake of evidence-based treatments may be associated with fear of stigmatization or logistic barriers to treatment [10, 12]. Additionally, among those who actually do seek treatment, many do not receive evidence-based treatments or an adequate treatment dose [10].

Compared to traditional face-to-face intervention settings, using the Internet as a medium for the delivery of psychological interventions offers several advantages that might help to overcome treatment barriers [14]. Several meta-analyses showed their efficacy both for other mental disorders (e.g., depression or anxiety [15‐19]) and PTSD in particular [20‐22]. In addition, Internet- and mobile-based interventions (IMIs) are easy to access and offer users the opportunity to flexibly integrate the treatment into daily life, regardless of any limits in terms of space and time [23]. Moreover, users can set their own pace going through each session (potentially with automated feedback on their personal progress) as frequently or as quickly as they like [24]. Looking at PTSD in particular, online interventions can offer a unique and geographically independent treatment opportunity in underserved post-conflict areas [25]. It is also proposed that the anonymity in conduct of an online intervention for PTSD is one of its main advantages, particularly in some cultures (e.g., Arab cultures [25]). Correspondingly, one study has shown that anonymity, privacy, and confidentiality are the most frequently listed reasons for the use of the Internet for mental health support by survey respondents [26]. Additionally, studies also found IMIs on some mental health conditions such as depression to be scalable and cost-effective [14, 27, 28], a valid argument considering limited resources in many health care systems worldwide. Nevertheless, risks concerning IMIs (e.g., lacking adherence, non-detected negative outcomes, negative attitudes of both patients and clinicians towards these interventions, crisis management in case of anonymous users) should be carefully considered and comprehensively investigated [14].

However, given their presumed potential to overcome treatment gaps and advance mental health services on a large scale, a closer inspection of the evolving field of IMIs for PTSD is worthwhile. Gaining insight on the efficacy as well as on active intervention components (i.e., intervention components that are responsible for the efficacy and can therefore deemed active, e.g., human support, exposure, prompts, cognitive processing/restructuring) and mechanisms of change (i.e., processes or events responsible for the change of the outcome [29]) in IMIs may allow to further optimize outcomes and prevent possible adverse events [30‐33].

To test their efficacy, IMIs can be compared with active or inactive controls. Considering the confounding situation in the latter designs with specific ingredients (i.e., variables (e.g., therapeutic techniques) unique to a certain form of psychotherapy [34]) becoming mixed up with common factors (i.e., active ingredients contributing to the outcome while being common to all forms of psychotherapy [34, 35]), this systematic review and meta-analysis focusses on active controls. For example, in the context of IMIs for PTSD, one study [36] compares a self-guided web-based CBT intervention for cancer patients with web-based information-only with results favoring the intervention. Moreover, these active controls can be classified as either bonafide (i.e., treatments that were intended to be therapeutic [37]) or non-bonafide. Hereby, it is particularly proposed to compare IMIs with common factor controls as active controls (i.e., structural equivalent packages that do not contain the specific components of the IMI hypothesized as the active verum of the intervention [38]).

In contrast, so-called component studies that compare one treatment with a treatment in which one specific component is left out (i.e., dismantling studies) or added (i.e., additive studies) offer the possibility to examine active intervention components and assess their incremental (add on) effect sizes [39]. For instance, there is a study [40] comparing a self-guided web-based intervention for war veterans with (subthreshold) PTSD and hazardous alcohol use with peer support with the same intervention without peer support, whereas no between-group differences are reported. Nevertheless, to date, there is ample evidence for guidance as a beneficial feature of Internet-based interventions [16, 31] with pending evidence for incremental effects of other single components [16].

Lastly, mechanisms of change (i.e., the actual processes responsible for therapeutic change) can be indicated by mediators (e.g., functional cognitions, emotional regulation, or problem-solving skills), and thus, the study of mediators is an important first step to understand mechanism of change [29]. Hence, a quantitative assessment of changes within a well-established mediation analysis is necessary in this context. For example, one study [41] assessed the mediating role of self-efficacy beliefs on PTSD symptom reduction in an Internet-based self-efficacy intervention and reports that health and human services professionals experiencing increased levels of self-efficacy are more likely to show lower secondary traumatic stress at follow-up. Another study on the mechanisms of change in CPT and prolonged exposure therapy for PTSD assessed habituation and hopelessness [42] while a study on mechanisms of change in IMIs irrespective of disorders concluded that cognitive factors are an important mechanism of change while assuming causal effects without mediation analyses [33]. This leaves room for an explorative assessment of intervention components and mediators and mechanisms of change specific for PTSD. In line with previous publications [29, 43, 44], we hereby conceptualize mediators as intervening variables that may statistically account for the relationship between the independent variable and the outcome. In contrast, moderators are variables or characteristics that influence the magnitude (or direction) of the relationship between an intervention and outcome [29], thereby indicating on whom and under which conditions treatments have differential effects [44]. As such, moderators suggest that different processes are involved (i.e., “moderated mediation”), but conceptually represent no mechanisms of therapeutic change themselves [29, 44].

Additionally, adherence is an important aspect aside symptom severity, considering both the high drop-out rates in IMIs for PTSD in various trials (e.g., 51.2% post-intervention [41]) and the hypothesized dose-response relationship for the use of IMIs (e.g., [45, 46]), which suggests an underestimation of effects in case of low adherence. Thus, reviewing adherence data in this field is highly valuable in order to gather knowledge on intervention components contributing to increased adherence rates and thereby inform intervention development.

To our knowledge, two meta-analyses in this field solely investigated telehealth interventions [47, 48], whereas three recent meta-analyses [20‐22] investigated Internet-based interventions for posttraumatic stress and found evidence for their efficacy. In particular, one meta-analysis [20] found effects ranging from SMD = − .66 to − .83 while including 13 studies with passive and 7 with active comparison conditions and an overall sufficient quality of included studies reported by the authors. Another meta-analysis [21] compared Internet-based CBT (iCBT) to both waitlist/TAU controls (SMD = − .71; 10 studies) and other interventions (SMD = − .28; 3 studies), with authors reporting a varying methodological quality of included studies. A third meta-analysis [22] with 33 included studies reported a pooled SMD of − .35 favoring Internet-based interventions independent of the comparison condition while including 18 active and 20 waitlist/TAU comparisons and reporting a slightly increased effect size when studies with reported high risk of bias were excluded (8 comparisons). However, none of these meta-analyses distinguished between bonafide and non-bonafide active controls, and to the best of our knowledge, there has been no systematic review or meta-analysis on IMIs for PTSD examining intervention components in additive or dismantling studies. The existing meta-analyses only provide indirect evidence on a possible larger effect of IMIs with therapeutic support compared to those without therapeutic support [21] and no evidence for responsible intervention components [20, 22] while including (predominantly) passive comparison conditions. The same holds true for mediators or mechanisms of change that have not been meta-analytically summarized so far. Moreover, even though the results indicate that Internet-based treatments are effective, interventions designed as smartphone applications were excluded in former reviews [20, 21].

Thus, this review and meta-analysis will provide a broader and updated overview of this field of research, focusing on intervention components and mediators. The proposed systematic review and meta-analysis aims to systematically review, critically evaluate, and statistically integrate the existing literature on IMIs for PTSD in order to answer the following research questions:

This review and meta-analysis is registered with the International Prospective Register of Systematic Reviews (PROSPERO, CRD42019130314) and will be reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [41]. The PRISMA-P checklist can be found in Additional file 1.

This systematic review and meta-analysis will add to previous research by summarizing, synthesizing, and discussing the existing literature on Internet- and mobile-based interventions for PTSD. The findings of this review will go beyond previous meta-analyses and systematic reviews in this field (e.g., [20‐22]), as results will extend to intervention components and mediators and mechanisms of change, while also covering mobile-based interventions. By identifying and quantifying the active intervention components and mediators and mechanisms of change, it may advance our understanding on how Internet- and mobile-based intervention take effect. This can help to develop more effective interventions and prevent possible adverse events. Moreover, results might assist in the verification and advancement of psychotherapeutic theories and the development of an extensive model on how IMIs for PTSD and psychotherapies in general might work [29, 72].

However, there are several limitations to the proposed systematic review and meta-analysis. One important limitation might be caused by substantial heterogeneity of included studies. The clinical, methodological, and statistical differences might limit the possibility of quantitatively pooling trials as well as the generalizability of findings. Previous meta-analyses on IMIs for PTSD have shown up to considerably high heterogeneity while suggesting differences in employed therapeutic techniques among others to be responsible (e.g., [22]). Moreover, only studies written in English will be included, which may lead to an overestimation of effects as trials with statistically significant results have been shown to be more likely published in English [73]. Furthermore, despite the attempts to include unpublished and non-significant studies, the proposed systematic review and meta-analysis might also be limited by publication bias.

In sum, given the substantial burden of disease associated with PTSD and the opportunities of IMIs for extending and augmenting available service supplies, the proposed review and meta-analysis is urgently needed and will substantially add to the current evidence. The results will inform intervention development by highlighting active intervention components and mediators responsible for therapeutic change, and extend the evidence base on the efficacy of IMIs for PTSD.