Feasibility, acceptability, and effectiveness of web-based and mobile PTSD Coach: a systematic review and meta-analysis

ABSTRACT Background: Timely and effective interventions can alleviate or prevent the development of clinical symptomatology in trauma-exposed individuals. However, due to limited access to these interventions, and/or stigma around accessing mental health services, there is an unmet need. Internet-and mobile-based interventions may help to address this need. Aims: This review aims to (i) synthesise the evidence on feasibility, acceptability, and effectiveness of the ‘PTSD Coach’ intervention (both web-based and mobile applications) in trauma-exposed individuals; (ii) evaluate the quality of this research; and (iii) identify challenges and recommendations related to ‘PTSD Coach’ intervention delivery. Method: Systematic database searches were conducted (PubMed/MEDLINE, PsycINFO, EMBASE, PLoS, Web of Science, PTSDpubs, Scopus, and clinical trial databases). Review inclusion was based on predefined inclusion criteria, and study quality was assessed with the mixed methods appraisal and the risk-of-bias tools for randomised trials. Where feasible, meta-analytical pooling of intervention effects on posttraumatic stress symptoms (PTSS) was conducted. Results: Seventeen manuscripts reporting on 16 primary studies were included with the majority evaluating a self-managed PTSD Coach mobile application intervention. Most studies were conducted in higher-income countries and females were over-represented. For both platforms, satisfaction and perceived helpfulness were generally high but type of smart device operating system was identified as an influence. The pooled effect size in symptom severity in the intervention group compared to the comparison group was not significant (standardised mean difference =  – 0.19) (95% CI − 0.41 to − 0.03, p = .09). Heterogeneity was not significant (p = .14; I2 = 40%). No study was excluded based on quality assessment. Conclusion: Findings support the feasibility and acceptability of ‘PTSD Coach’ in trauma-exposed individuals. However, evidence on the effectiveness on PTSS remains limited. More research is still needed in low-middle-income countries, particularly those in which supported ‘PTSD Coach’ interventions are evaluated in larger and more diverse samples.

• Most studies evaluated a self-managed PTSD Coach mobile application intervention in higherincome countries.

Introduction
Trauma exposure (TE) is a risk factor for the development of psychological distress and, in some individuals, the development of traumaand stressor-related, anxiety and depressive disorders (Adams et al., 2018;Levin-Rector et al., 2018;Suliman et al., 2014).The Diagnostic and Statistical Manual 5th edition (DSM-5) defines TE as experiencing, witnessing, or being confronted with actual or threatened death, serious injury, and/or sexual violence (American Psychiatric Association, 2013).Posttraumatic stress symptoms (PTSS) that can develop in response to a TE can be characterised in four clusters: (i) re-experiencing or intrusion, (ii) avoidance, (iii) negative cognition and mood, and (iv) arousal and reactivity (Weathers et al., 2013).Some trauma-exposed individuals may develop clinically significant PTSS meeting the threshold for posttraumatic stress disorder (PTSD) diagnosis, while others may present with sub-threshold PTSS (Cukor et al., 2010).Regardless, coping mechanisms are often overwhelmed and daily life activities can be negatively affected requiring prompt intervention (Morabito et al., 2020;Skeffington et al., 2020).
Timely and effective interventions can either alleviate or prevent the development of clinical symptomatology (Colizzi et al., 2020).However, many trauma-exposed individuals do not receive timely and effective assistance due to limited access and/or stigma around accessing mental health services (Becker & Kleinman, 2013;Semo & Mogga Frissa, 2020).In response, internet-and mobile-based interventions are increasingly used either as first-line treatment (e.g.whilst on waiting lists for professional mental health care) or as augmentation to mainstream therapeutic modalities (Owen et al., 2015;Price et al., 2014).
In 2011, a freely available self-managed mobile application-based intervention called PTSD Coach was introduced by the United States Departments of Veterans Affairs and Defense (VA/DoD) (Owen et al., 2015).The trauma-focused PTSD Coach mobile application aims to provide trauma-exposed individuals with psycho-education on trauma and PTSD, assist with managing and monitoring posttraumatic stress symptoms (PTSS), and encourage seeking treatment if so indicated.These aims are addressed through the four core sections of the mobile application: (1) 'Learn', (2) 'Track progress', (3) 'Manage symptoms', and (4) 'Get support.'Whereas 'Learn' provides the mentioned psycho-education, 'Track progress' refers to the brief assessments designed to assist individuals with symptom monitoring (e.g.rating anxiety level before and after accessing a specific symptom management tool).These assessments accompany the 21 evidenced-based tools under 'Manage symptoms.'Informed by both mindfulness and cognitive behavioural therapy principles, these easily accessible evidenced-based tools were designed to address eight core PTSD symptoms, namely, being reminded of the trauma, avoiding triggers, being disconnected from people, disconnected from reality, sad/hopeless, worried/anxious, angry, and unable to sleep.The mobile application's 'Get support' section includes a feature to customise contact details of applicable support services and safety plans when indicated.PTSD Coach does not require data post download nor collects identifiable data during usage and can be used independently as a self-management tool or as augmentative to other psychological support treatments (Kuhn et al., 2018).
A VA/DoD review summarised the primary research data of eight of their 20 available mobile application interventions, of which the well-documented PTSD Coach was one (Gould et al., 2019).The reviewers concluded that comparatively more research has been conducted on the 'PTSD Coach' and the 'Virtual Hope Box' mobile applications, and further evidence regarding effectiveness was needed.
Since its release, the PTSD Coach mobile application has been translated and culturally adapted in six countries (Australia, Canada, The Netherlands, Germany, Sweden, and Denmark) (Kuhn et al., 2018).At the time of the review by Kuhn et al. (2018), no systematic evaluation of PTSD Coach Australia, PTSD Coach Canada, and PTSD Coach Denmark mobile applications were conducted, and primary research had been conducted on SUPPORT Coach (The Netherlands) and PTSD Coach Sweden mobile applications.
Cumulatively, the developing research is promising as the results suggest high user satisfaction and acceptability, as well as self-reported PTSS improvement.Kuhn et al. (2018) concluded that even though the PTSD Coach mobile application is progressively expanding its reach, most of the research is conducted in higher-income settings with known higher smartphone ownership and higher mental health resource access.
In addition to the VA/DoD's research advances of the PTSD Coach mobile application (hereafter referred to as PTSD Coach MA) and general internet-based intervention development, a web-based 'PTSD Coach Online (PCO)' platform was developed and released in 2013 (National Center for PTSD, 2017).These two platforms are similar in that both provide easily accessible support based on user needs, are informed by mindfulness and cognitive behavioural therapy principles, and can be used as a self-management tool or augmentative to other support.However, the web-based platform includes additional psycho-educational videos with instructions for completing various computer-based interactive activities and requires an active internet connection.

Review aims
This mixed methods review of controlled and uncontrolled clinical studies on 'PTSD Coach' (web-based and mobile application platform) aims to (i) synthesise the evidence on feasibility, acceptability, and effectiveness in trauma-exposed individuals; (ii) assess the quality of this research; and (iii) identify challenges and recommendations related to PTSD Coach intervention delivery.For this review, feasibility and acceptability were determined by considering attrition rates, PTSD Coach MA/PCO usage, smart device ownership, and feedback provided by trauma-exposed study participants on perceived satisfaction/benefits and barriers/concerns.Effectiveness was determined by considering a reduction in PTSS evaluated with validated subjective or diagnostic measures (i.e.PTSD Checklist/Clinician-administered PTSD Scale).

Registration and protocol
This systematic review was registered with the International Prospective Register of Systematic Reviews (PROSPERO) database (CRD42022331731).It conforms with both the Cochrane Collaboration guidelines and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting standards (Liberati et al., 2009).

Eligibility criteria
Only peer-reviewed primary research published in English was considered for the review.Specifically, this review considered quantitative, qualitative, or mixed methods primary research reporting on the feasibility, acceptability, and effectiveness of either 'PTSD Coach' platforms.We excluded secondary research papers such as narrative, scoping, and systematic reviews.We included studies utilising any of the two 'PTSD Coach' platforms as an intervention for trauma-exposed individuals (with and without a PTSD diagnosis) of all ages in any setting.A study was eligible for inclusion if feasibility, acceptability, or effectiveness was assessed (i.e. a minimum of any one of these domains was required for inclusion), and included methods (i.e.assessments/tools) of assessing the index domains).As this was a mixed methods systematic review, studies were included irrespective of the presence of a control group.Prior to the qualitative and quantitative (meta-analyses) synthesis, the primary reviewers independently conducted standardised searches, systematic review inclusion screening, methodological quality and risk of bias assessments, and data extraction (Moher et al., 2001;Viswanathan et al., 2018).The reviewers met to compare their independent findings and minor disagreements were resolved through discussion.An eligibility form was used to capture reasons for exclusion/inclusion (Additional File 1: Eligibility form).

Literature search
We conducted a systematic search between 19 and 26 July 2022 of the following electronic databases: PubMed/MEDLINE, PsycINFO, EMBASE, PLoS, Web of Science, PTSDpubs, and Scopus.Clinical trial registries (ClinicalTrials.gov; International Clinical Trials Registry Platform; Pan-African Clinical Trials Registry; International Standard Randomised Controlled Trial Number) were searched for additional relevant studies not identified by the initial database searches.No limitations on the date of publication were applied.References within the bibliographies of identified studies were manually searched for additional relevant studies.
We were unable to find primary research studies on the following three PTSD Coach MA versions: PTSD Coach Canada, CoachPTBS Germany, and PTSD Coach Denmark, which were included in a previous PTSD Coach MA review (Kuhn et al., 2018).Communication with the corresponding author of this review confirmed the accuracy of our search results.

Search strategy and study selection
The primary search was formulated and conducted in MEDLINE (PubMed) and translated to the other databases.The primary search included the following keywords (and MeSH terms): ('PTSD Coach' OR 'PTSD Coach Online') AND (stress OR trauma OR mental OR emotion).An information specialist (librarian) assisted with the identification of the correct search strategy and search terms.Figure 1 depicts the PRISMA Flow Diagram (McKenzie et al., 2021).The search identified (N = 267) records.We first applied non-primary research filters (i.e.reviews, newsletters, web articles, newsletter articles, reports, web pages, and guidelines) and removed 88 records.Thereafter, an additional 15 records were removed manually as they were identified as reviews or study protocols.After duplicates were removed (n = 91), 73 records were exported to Mendeley (reference management software) which was used to save and manage searches (Mendeley, 2009).
Further screening of titles and abstracts based on eligibility criteria (letter to the editor, not PTSD Coach-related primary research, systematic reviews, guidelines) led to the removal of 47 records, resulting in 26 full-text records remaining.An additional eight records were removed resulting in 17 records included in the review.Of the eight excluded records three were not conducted in trauma-exposed individuals; 'PTSD Coach' was not evaluated as an intervention (n = 4), and one was a review that included the primary studies already included.

Data extraction
The data extraction form used in the review can be accessed in Additional File 2: Data extraction form.The first reviewer extracted the relevant data from the included studies, while the second reviewer corroborated the extracted data.Extracted data included study characteristics; trauma-related data; study methodology; and outcome data.The main outcomes for which the data were sought included: (i) feasibility; (ii) acceptability; and (iii) effectiveness of the PTSD Coach MA and PCO on PTSS, respectively.

Quality assessment
Reviewers performed quality assessments independently on all manuscripts and resolved disagreements through discussion.The mixed methods appraisal tool (MMAT) was applied to all manuscripts (Hong et al., 2019).The MMAT is divided into five categories that evaluate the methodological quality of qualitative, quantitative (RCTs, non-RCTs, and descriptive), as well as mixed methods studies.Additionally, the risk-of-bias tool for randomised trials (ROB-2) was applied to the RCT data (Sterne et al., 2019).The ROB-2 evaluates the risk of bias due to (i) the randomisation process, (ii) deviations from the intervention, (iii) missing outcome data, (iv) measurement of the outcome, and (v) selection of the reported result.As depicted in Figure 2, the Robvis tool was used to create a traffic light plot to visualise the Rob 2 assessment results (McGuinness & Higgins, 2021).

Synthesis
We provide a general description of eligible manuscripts and participants.Thereafter, we describe the interventions used in each study, including the relevant outcome data.We then summarise findings based on the feasibility, acceptability, and effectiveness of the intervention on PTSS.To synthesise data from RCTs on the effectiveness on PTSS changes, we conducted a meta-analysis to calculate the pooled effect size based on primary outcome data (i.e.PTSD Checklist -PCL) using Review Manager (RevMan) 5.4.1 (Higgins et al., 2022).Means and standard deviations were used for the continuous outcomes.A randomeffects model was used to estimate summary effect sizes (ES) with related confidence intervals (CI of 95%).The I 2 statistic was used to evaluate the heterogeneity of the included studies (Higgins et al., 2022).Lastly, we report the results of the quality assessments.

Intervention data
Table 2 summarises the intervention characteristics and outcomes of the included primary studies.Eight of the primary studies (50%) were conducted in community samples (American, South African, Swedish, and Egyptian) followed by four (25%) in veteran samples.The remainder were conducted in the general population (n = 2; 12.5%), among healthcare professionals (n = 1; 6.25%), and motor vehicle accident emergency care patients (n = 1; 6.25%).
Intervention duration varied from three days to 16 weeks with the two multi-country studies not specifying intervention duration.Apart from the multicountry studies that included a one-year observation period, post-intervention follow-up time points for the remainder of the studies varied from immediately after PTSD Coach MA/PCO use to 24 weeks.
Two primary studies reported specifics on smart device ownership (Kuhn et al., 2014;Miner et al., 2016), while either ownership or access to a smart device can be deduced from study inclusion criteria from the remainder of the studies.

Feasibility and acceptability
For this review, feasibility and acceptability were determined by considering attrition rates, PTSD Coach MA/PCO usage, smart device ownership, and feedback provided by the trauma-exposed study participants about perceived satisfaction/benefits and barriers/concerns.

Attrition and usage
In the primary RCT studies, attrition rates ranged from 10% to 23% at post-intervention follow-up (Bröcker et al., 2022;Hensler et al., 2022;Kuhn et al., 2017;Miller-Graff et al., 2021;Miner et al., 2016;Possemato et al., 2016;van der Meer et al., 2020).Attrition rates ranged from 0% to 8% at the post-intervention follow-up in the four pre-post Notes.Bröcker paper consisted of two pilot randomised controlled trials (a = Pilot 1; b = Pilot 2) and each pilot was assessed independently; 1 No information regarding allocation concealment was provided, however, due to self-reported outcome data this likely did not result in bias; 2 Outcome assessors were not blinded to the intervention provided, however, since the outcomes were self-reported the outcome data was not influenced; 3 Taking into account explanations 1,2 ; 4 Taking into account explanation 2 ; 5 No specific information regarding the presence/absence of baseline imbalances provided; primary studies (Cernvall et al., 2018;Hiratsuka et al., 2019;Kuhn et al., 2014;Tiet et al., 2019).Owen et al. (2015) indicated that regular use of the MA decreased from 61.1% after the initial download, to 46% at one month, 28% at 3 months and 10.6% at approximately 12 months post-download.This decline in regular MA usage appeared to be influenced by the type (iOS vs. Android) of smart device operating system (OS) that was used, with Android users being less likely to use the MA at one month post download than iOS users (Owen et al., 2015).Hallenbeck et al. (2022) reported that while most downloads (73.7%) were from iOS users they showed significantly lower regular MA usage levels than Android users.Furthermore, they report that 72% of all users engaged with the MA at any point, with regular MA usage declining from 87% the first day post download to 10% and 0.69% one month and one year post download (Hallenbeck et al., 2022).
In contrast, Tiet et al. (2019) reported that 70% of users engaged with the MA one month post download, and 97% and 72% two and three months later, respectively.However, the influence of smart device OS type was not reported.The latter study as well as other primary study data suggest that both less study attrition rates and higher MA engagement are positively influenced by a type of support (i.e.telephonic/clinician support) (Bröcker et al., 2022;Hiratsuka et al., 2019;Pacella-LaBarbara et al., 2020;Possemato et al., 2016;Tiet et al., 2019).
Some participants found the structure (Cernvall et al., 2018;Hensler et al., 2022;Owen et al., 2015) and content (Shakespeare-Finch et al., 2020) of the MA problematic.Difficulties and confusion with the PCO website were also reported (Ellis et al., 2022;Hiratsuka et al., 2019).This links to the barrier of the inability to customise some features of the MA (Kuhn et al., 2014;Owen et al., 2015;Shakespeare-Finch et al., 2020), which was also noted as a barrier to PCO use (Ellis & Miller-Graff, 2021).
General technical glitches were another barrier experienced by MA users (Cernvall et al., 2018;Owen et al., 2015).The technical problems appear to be influenced by the type of smartphone device OS used, with Android users reporting significantly more technical problems than iOS users (Owen et al., 2015).For one user, the technical difficulties resulted in increased psychological distress (Owen et al., 2015).Other participants also mentioned potential psychological distress as a barrier (Hensler et al., 2022;Shakespeare-Finch et al., 2020).Concerning PCO, computer literacy (Bröcker et al., 2022) and the inability to use the website on a smartphone (Hiratsuka et al., 2019) hindered use.
Other barriers to feasibility and acceptability of the MA included differences in expectation (Cernvall et al., 2018;Hensler et al., 2022), the perceived need for MA use (Pacella-LaBarbara et al., 2020;Shakespeare-Finch et al., 2020), and insufficient time availability (Pacella-LaBarbara et al., 2020).Dissatisfaction with the content was noted by some MA users (Hensler et al., 2022) as well as by some PCO participants (Ellis et al., 2022).
Lastly, a few MA participants suggested integrating other support sources rather than using the MA as a stand-alone intervention (Cernvall et al., 2018;Owen et al., 2015;Shakespeare-Finch et al., 2020).Linked hereto is that most PCO participants noted the lack of a human element as a barrier (Ellis et al., 2022).

Effectiveness
Effectiveness was determined by considering a reduction in PTSS evaluated with validated subjective or diagnostic measures (i.e.PCL / CAPS-5).Where the included studies had more than one follow-up point, data from baseline to post-intervention was used.For the meta-analysis, we included the six PTSD Coach MA RCTs that reported on PCL outcome data (Hensler et al., 2022;Kuhn et al., 2017;Miner et al., 2016. Pacella-LaBarbara et al., 2020;Possemato et al., 2016;van der Meer et al., 2020).Due to variability in study design and outcome measures (PCL-C, PCL-S and PCL-5) a random effects model and standardised mean difference (SMD) were used to calculate effect sizes.Some studies reported a significant decrease in total PCL scores between baseline and post-intervention in the intervention group but not in the waitlist control groups (Kuhn et al., 2017;Miner et al., 2016), while others reported a significant decrease in both study groups in the respective RCTs, namely self-managed PTSD Coach versus clinician supported PTSD Coach, and access to versus no access to SUPPORT Coach (Possemato et al., 2016;van der Meer et al., 2020).However, the overall greater decrease in symptom severity in the intervention group compared to the comparison group (pooled effect size) was not significant, SMD = -0.19(95% CI −0.41 to −0.03, p = .09).Heterogeneity was not significant (p = .14;I 2 = 40%).
The larger (N = 87) PCO RCT with treatment or waitlist control groups reported a non-significant decrease in total PCL scores from baseline and post-intervention between conditions (Cohen's d = -0.14)(Miller-Graff et al., 2021).
Another small (N = 10) PTSD Coach MA RCT reported a significant decrease in total CAPS-5 score (Reliable change index = < 1.96) in all participants from baseline to post-intervention in the counsellorsupported arm, and in two participants in the selfmanaged arm (Bröcker et al., 2022).
Pre to post-intervention PCL data also showed a non-significant decrease in total PCL scores from baseline to post-intervention (Cernvall et al., 2018;Tiet et al., 2019), however of interest is the reported significant decrease in the PCL re-experiencing subscale score (p = .035)(Tiet et al., 2019).Although outside the scope of this review, it is worth noting that some studies (Cernvall et al., 2018;Hensler et al., 2022;Kuhn et al., 2017;Miller-Graff et al., 2021;Possemato et al., 2016;Tiet et al., 2019) reported on the intervention effects on depression as a secondary outcome.The potential broader benefit of 'PTSD Coach' on depressive symptoms as a secondary outcome was supported by some studies (Hensler et al., 2022;Kuhn et al., 2017;Possemato et al., 2016;Tiet et al., 2019), but not by others (Cernvall et al., 2018;Miller-Graff et al., 2021).

Quality of included papers
The results of the MMAT and ROB-2 are presented in Table 3 and Figure 3, respectively.In general, the quality of the manuscripts was deemed acceptable based on the different study designs.The main concerns were a lack of sufficient information concerning allocation concealment (n = 3) and blinding to the intervention by outcome assessors (n = 6), however, since the outcomes were self-reported the outcome data was not influenced; insufficient information on baseline group differences (n = 2); and small sample sizes (n = 3) that may have affected representativeness.Reviewers queried a discrepancy found between the reported frequencies by ethnicity in Kuhn et al. (2017) where the reported numbers exceeded the number of enrolled participants (Table 1, pg. 270).The quality assessment of this study was not adversely affected since the author clarified that participants self-identified with more than one ethnic category.

Discussion
This systematic review provided a synthesis of the feasibility, acceptability, and effectiveness of PTSD Coach MA and PCO data in trauma-exposed individuals and highlighted intervention challenges and recommendations for future research.Seventeen manuscripts reporting on 16 primary studies were included in the review.Findings support the feasibility and acceptability of 'PTSD Coach' in trauma-exposed individuals, however evidence for its effectiveness is currently limited.
Generally, users were satisfied with both 'PTSD Coach' platforms, however, the type of smart device OS (i.e.iOS or Android) affected MA use and satisfaction with a suggestion to develop OS specific versions of the MA (Owen et al., 2015).Perceived helpfulness for both 'PTSD Coach' platforms was generally high, with type of smart device OS use again playing a role.Differences in experience of and preference for either iOS or Android are frequently debated among users and developers (Győrödi et al., 2017;Kaur Ubhi et al., 2017;Wang & Godfrey, 2013) Like other PTSD intervention studies, this review found that most of the studies used the PCL to monitor intervention effects on PTSS and one study used the CAPS-5 (Steubl et al., 2021).Some PTSD Coach MA studies reported a significant reduction in PTSS in the intervention group (Kuhn et al., 2018;Miner et al., 2016).However, as seen in Figure 3, the pooled effect size was not significant.Tiet et al. (2019) was the only study that reported intervention effects on PCL subscales in which they found a significant decrease in the re-experiencing subscale.This level of subscale analysis can inform which PTSD symptom clusters are more sensitive to 'PTSD Coach', and possibly aid in either a more targeted intervention approach or intervention refinement.
Most studies were conducted in HICs.This may be due to 'PTSD Coach' originating in a HIC, and comparatively higher internet access and smart device ownership compared to LMICs (Campbell et al., 2015;Kuhn et al., 2018;Stork et al., 2013).Additionally, significantly higher treatment-seeking behaviour (53.5%) in HICs compared to LMICs (22.8%) likely contributed (Koenen et al., 2017).Aside from the 'PTSD Coach' platforms, other internetand mobile-based interventions for PTSD were also mostly conducted in HICs (Steubl et al., 2021).
The over-representation of females in the 14 studies that reported on sex is interesting since 'PTSD Coach' was originally developed for veterans (mostly men) (Harrington et al., 2019).However, the over-representation of females agrees with general PTSD research indicating a higher prevalence of PTSD reported by women, who are also more likely to seek treatment (Koenen et al., 2017;Olff, 2017).Furthermore, the lack of specification of ethnicity in the studies is noteworthy.It is important to expand this research to other groups considering 'PTSD Coach' research has suggested a difference in treatment response rates in subgroup analyses (Pacella-LaBarbara et al., 2020), further highlighting the importance of stratifying by sex/ethnicity.
While the quality of the included papers can generally be deemed acceptable, future studies should be more rigorous in terms of reporting on allocation concealment and blinding of outcome assessors, providing sufficient baseline information, and including larger sample sizes.
In conclusion, most of the studies evaluated a selfmanaged 'PTSD Coach' intervention with a select few evaluating an augmented (i.e.weekly tip messages/ phone calls and clinician/counsellor) intervention delivery.The studies that evaluated a supported 'PTSD Coach' intervention generally seemed to have less attrition, higher engagement, and increased perceived helpfulness and satisfaction (Bröcker et al., 2022;Hiratsuka et al., 2019;Possemato et al., 2016Possemato et al., , 2017;;Tiet et al., 2019).

Strengths and limitations
To our knowledge, this is the first review to combine quantitative and qualitative studies on both 'PTSD Coach' platforms, resulting in a richer picture of the currently available research.However, only including peer-review published research may be a limitation, as we may have missed useful information included in unpublished research.Additionally, the relatively small number of eligible RCTs and small sample sizes limits the evidence on whether PTSD Coach is an effective intervention for reducing PTSS.Nonetheless, this is the first 'PTSD Coach' review that includes a meta-analysis indicating an increase in both general research advances in this area and specifically in effectiveness data as was a need suggested by previous reviews (Gould et al., 2019;Kuhn et al., 2018).

Future research
These findings suggest that more research is needed in LMICs, particularly that which evaluates supported 'PTSD Coach' interventions in larger and more diverse (i.e.sex and ethnicity) samples.Additionally, objective monitoring of PTSS (i.e.blinded monitoring with diagnostic measures such as the CAPS-5) in conjunction with self-report measures is needed.Although many studies have been conducted since the last review (Kuhn et al., 2018) more research is needed to increase evidence for the effectiveness of 'PTSD Coach' as an intervention for PTSS.A focus on improvements related to PTSD symptom clusters is also recommended to further inform intervention effectiveness and/or improvements.The evaluation of PTSD Coach intervention effects on co-morbid disorders such as anxiety and depression may also prove insightful.Other authors have suggested that 'PTSD Coach' be evaluated as an augmented intervention (i.e.supported) (Miller-Graff et al., 2021;Miner et al., 2016) or augmentative intervention (i.e.alongside psychotherapy) (van der Meer et al., 2020) as part of future research agendas.
In sum, findings support the feasibility and acceptability of 'PTSD Coach' in trauma-exposed individuals.However, evidence on the effectiveness on PTSS remains limited; and especially in LMIC where there is limited access to these interventions, and/or stigma around accessing mental health services more research is needed.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow chart of included studies.From: Page et al. (2021).

Figure 2 .
Figure 2. Intervention effects on changes in PTSS severity post-interventioncomparison intervention versus control group.
with corresponding author rationale for reported ethnicity numbers resulting in N = 135 rather than N = 120, participants could self-identify with more than one category, which they did; 2 Dutch citizen = 96.9% and non-Dutch citizen = 3.1%; 3 Men, other, or preferred not to answer; a = Pilot 1; b = Pilot 2; F = Female; FGD = Focus Group Discussion; Int.= Intervention; M = Male; NS = Not Specified; PTSD = Posttraumatic Stress Disorder; Quali.= Qualitative; Quanti.= Quantitative; RCT = randomised controlled trials; TAU = treatment as usual.

Figure 3 .
Figure 3. ROB-2 Quality appraisal of included articles.Notes.Bröcker's paper consisted of two individual randomised controlled trial pilots and each pilot was assessed independently.

Funding
E.B is supported by the National Research Foundation

Table 1 .
Study and baseline sample characteristics.

Table 2 .
Intervention description and outcome data.
Notes:1Original PTSD Coach Mobile App version available from the app store; 2 Non-public, research version of PTSD Coach Mobile App; 3 Post-intervention is the first follow-up reported when there are more than one follow-up time points; 4 12 and 16-week follow-up not reported on due to possible confounding factors described by authors; 5 PTSD Coach Sweden; 6 Health is Our Tradition: Balance and Harmony after Trauma (Fully adapted PTSD Coach Online); 7 Changed score from 4 to 12 weeks reported, since post-traumatic stress symptoms keyed to index trauma not assessed at baseline. 8PTSD Coach Australia; 9 SUPPORT Coach

Table 3 .
MMAT Quality appraisal of included articles.