HELP@APP: development and evaluation of a self-help app for traumatized Syrian refugees in Germany – a study protocol of a randomized controlled trial

Our interactive self-help app will rely on (i) evidence-based cognitive-behavioral therapy (CBT) for PTSD [24], and (ii) evaluated internet-based self-care approaches for the treatment of posttraumatic stress [25, 26]. Critically, meeting the specific needs of refugees as well as adapting the content in a culturally sensitive manner will be key to the development of our self-help app. Initially, interviews with approximately 5–10 target persons will thus be carried out to assess culturally sensitive aspects (e.g. different concepts of disease and disease management), that may be important for the intervention. To ensure full functionality and usability of the app, representatives of the target group will accompany all steps of the developmental process through regular consultation.

As can be seen in Fig. 1, the app will be set up modularly. Initially, participants will be guided through general user instructions and information about the association between war, escape, and posttraumatic stress will be provided (“onboarding”). Then, participants will be able to flexibly work through two modules (“trauma” and “resources”), which will consist of several submodules.

The “trauma”-submodules will address consequences of traumatization, especially focusing on posttraumatic stress symptoms as defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) [27]. In addition, symptoms of depression and anxiety will be introduced and explained. All submodules will provide (i) psychoeducational information (e.g. what are intrusions and flashbacks, what is somatization), and (ii) basic self-help techniques and skills with respect to symptom management (e.g. positive and negative coping strategies, cognitive restructuring, emotion regulation). We will also implement a self-test (PTSD Checklist for DSM-5/PCL-5) [28], that can be completed at any time in order to monitor symptom severity and to provide automated individually tailored feedback regarding progress and potential problems.

The “resources”-submodules will impart (i) psychoeducational information (e.g. about physical and mental needs, consequences of experienced traumata on self-esteem and relationships) as well as (ii) techniques and skills to strengthen and support own resources and self-care (e.g. setting goals, planning pleasant activities, gaining social support, relaxation techniques, mindfulness).

To maximize usability, content related materials will be provided in Arabic. Moreover, the app will additionally be complemented by video and audio sequences as well as interactive games and exercises. All submodules will be available at all times, giving users the possibility to flexibly switch between sections depending on their personal needs. The app will also provide a clearly visible emergency button with contact information regarding professional help in case of emergency. The app will be usable offline.

With respect to the technical implementation, the app development will take place in cooperation with the e-mental health program provider “Frühlingsproduktionen”.

Study participants will be Syrian refugees living in the urban area of Leipzig, Germany, who will be recruited using a multi-strategic approach [29]. If contacted refugees are willing to participate, written informed consent will be obtained. Trained native Arabic speaking study nurses will screen up to n = 1200 refugees for eligibility defined by the following inclusion and exclusion criteria, respectively.

Inclusion criteria:

Exclusion criteria:

The apps’ efficacy and usability as well as the overall associated cost effectiveness of the intervention will be evaluated in a prospective RCT with an IG and a CG (see Fig. 2). Following the initial screening and study group assignment, participants in the IG will gain access to the self-help app via person-specific log-in data (to avoid group contamination) for one month. During that time, the CG will obtain psychoeducational reading material about traumatization and posttraumatic stress (identical to the information delivered by the app) in form of a comprehensive brochure in Arabic.

Prior to randomization, participants will be pointed towards the required investment of time and effort necessary to complete the study in order to lower the risk for dropout.

Treatment adherence of the IG will be objectively assessed through the apps’ meta-data log-files that store information about the number of login-times and worked-through modules. Likewise, dropout assessment will be conducted to understand reasons for non-adherence.

Participants of both groups are assured to contact the study nurses in case of arising problems regarding current symptom burden, for example. Similar to the apps’ emergency button, the psychoeducational reading material will clearly state contact information regarding professional help in case of emergency.

Table 1 gives an overview of all outcomes and associated measures to be implemented throughout the entire study. Instruments that are not available in Arabic language will be translated according to recommendations of the “European Social Survey Programme” and “Cross-Cultural Survey Guidelines” respectively [33, 34]. Based on the TRAPD (Translation, Review, Adjudication, Pretesting, and Documentation) model, two independent and external translators will each provide drafts of a common forward and backward translation, which will subsequently be discussed and reviewed by an expanded group of bilingual experts. The approved versions will then be pretested. The entire procedure will be repeated, if necessary. This multi-step team translation approach not only ensures the reliable detection and elimination of problems of equivalence between the different questionnaire versions [35], but also takes into account the cultural specifics of the respective language. All other study materials (e.g. informed consent, brochure for the CG) will be translated by bilingual experts as well.

We hypothesize that the app intervention reduces posttraumatic stress symptomatology (primary outcome), quantified using the Posttraumatic Diagnostic Scale for DSM-5 (PDS-5 [30]). The PDS-5 is a well-established self-report measure based on the diagnostic criteria of the DSM-5, which allows the assessment of posttraumatic stress symptom severity.

Regarding secondary outcomes, we moreover expect the intervention to reduce symptoms of depression, anxiety, and somatization. Positive intervention-related effects are also expected regarding quality of life, self-efficacy, perceived stigmatization, social support, ambiguous loss, and posttraumatic growth.

Use of resources in terms of health demands will be assessed using a cost book (bottom-up approach). Furthermore, we aim to gather information on user acceptance and usability of the self-help app using the System Usability Scale (SUS) [36] and Technology Acceptance Model (TAM3 [37]), respectively.

Last, meta-data stored in the app’s log-files will provide quantitative information about the use of the app, e.g. number of days used, number of worked-through modules, or operating time.

After positive screening, study participants will be enrolled. Measurements are scheduled directly before the intervention (baseline, T0/pre), immediately after the intervention (T1/post, one month later), and three month after the intervention (T2/follow-up) to test for medium-termed treatment effects. A detailed overview of the schedule of enrolment, interventions and assessments is given in Fig. 3.

Anticipated sample size was calculated based on recent research evaluating the efficacy of telemedical-based treatment of posttraumatic stress symptoms [20, 21, 38]. Given a moderate between-group effect at T1 (Cohen’s d = 0.5), a significance level of α = 0.05 and a statistical power of 1-β = 0.90, optimal sample size for estimating a significant treatment effect is n = 140. Expecting a drop-out rate of approximately 40% (e.g. due to change in residence status), n = 94 participants have to be included additionally, yielding an ideal baseline sample of n = 234 participants at T0 [39].

A multi-strategic approach to recruit Syrian refugees living in the urban area of Leipzig, German, will be targeted. This includes using established contacts as well as building new contacts with facilities frequented by Syrian refugees (e.g. refugee related clubs, associations and organizations, psychosocial centers, information centers, language schools, community colleges, universities, registration authorities, hospitals, clinics, primary care practices, psychotherapeutic practices, etc.). These facilities act as multipliers to reach out to potential study participants. Likewise, participants will be recruited from local reception centers and other forms of collective accommodation centers. We will use social media, instant messaging and online tools to spread information about the study. The local media will be involved. Bilingual flyers and posters will be displayed all across the city in appropriate public spots. During the recruitment phase, we will have an information booth at local events that might attract Syrian refugees. Importantly, since Syrian refugees are well connected with each other, we aim to reach additional participants of the Syrian community by using a snowball sampling approach, e.g. we ask participants to spread information about the study within their personal social network [29].

Participants identified as eligible during screening will subsequently be randomly assigned to the IG or CG. Group assignment will follow a 1:1 ratio utilizing randomized permuted blocks of six, stratified by age and sex. This procedure ensures both balance in sample size across groups and control of important covariates. An external, independent statistician will generate the randomization block lists with a respective computer program (“blockrand”-package [40] written for R [41]). The different lists will be coded by this person so that the study coordinator, responsible for individual group allocation, remains blind to the lists’ strata identity.

Besides randomization, the data analysts, who will conduct the primary analysis concerning the hypothesized group differences (IG vs. CG) in our primary and secondary outcome measures (see above), will be blind to group assignment.

All assessments will be carried out in form of structured face-to-face interviews led by trained native Arabic speaking study nurses. The structured interviews will be present in form of paper-and-pencil assessments that the study nurses will complete to collect data on sociodemographics, primary and secondary outcomes using standardized questionnaires, as detailed above. Additionally, data will be collected to monitor possible negative effects and reasons for dropout (see below).

Data on the app usage will be collected through meta-data log-files that store anonymous information about the number of login-times and worked-through modules.

Data entry will take place successively and timely after collection using a statistical package. However, a data completeness and consistency check will take place beforehand. In case of missing answers to items, reasons for missing information will be clarified and efforts undertaken to reach completion. De-identified data will be entered in a data mask which will be password-protected and stored locally, with access granted only to study personnel who have signed data protection wavers. Anonymous log-file data of app usage will be collected, stored and transferred using SSL-technology to ensure data encryption and protection. In fact, specific arrangements regarding data protection (e.g. separation of evaluation and program, no real names for program login) will be defined by a separate data protection concept.

All acquired data will be examined with respect to potential inconsistencies and missing values, which will be replaced using multiple imputation methods [42], if appropriate. To detect potential systematic biases, a drop-out analysis will be performed. Internal consistencies of the newly translated inventories will be calculated to evaluate their reliability. All subsequent analyses will be treated as intention-to-treat or completer-analyses. Data-analysts will be blinded for group allocation.

To assess the potential health economic benefits, we will perform analyses of the (i) actual costs and (ii) the ICER of the intervention. Costs will be quantified by the frequency of how often physicians and therapists are consulted (including days spent in a hospital) and valuated using current standard costs [43]. The app development expenses will be treated as economically reasonable overhead costs and added accordingly. To assess the interventions’ economic efficiency, the ICER will be calculated as the ratio between the difference of the actual costs and the difference of the averaged treatment effects (indexed by either quality-adjusted life years or primary outcome scores) between IG and CG. Respective confidence intervals will be derived by non-parametric bootstrapping. To visualize the ICERs statistical uncertainty, cost-effectiveness-acceptance curves will be constructed, indicating the probability of an acceptable ICER in relation to willingness to pay for an effect. Alternatively, net benefit regressions [44] can be applied to estimate the respective group specific net benefits while accounting for potentially confounding variables such as age, sex, and education.

The results of the study will be reported according to the guidelines of the Consolidated Standards of Reporting Trials (CONSORT) statement [45].

A data monitoring committee (DMC) will not be established for several reasons. First, periodic inspection does not seem justified as trial duration is short (4 weeks). Second, overall risk associated with the self-help intervention is considered low, and therefore, the likelihood for the need to modify or discontinue the trial is considered insignificant. However, precautions have been taken to minimize the potential of any harm towards individual study participants, as detailed below.

Preventively, we will only enroll study participants without severe posttraumatic and/or depressive symptomatology and acute suicidality. The self-help program targets individuals with mild to moderate symptom severity. Despite these precautions, the occurrence of negative effects during the trial in individual cases cannot be entirely ruled out. Therefore, potential negative effects will be monitored through a standardized assessment procedure at follow-up 1 (post-intervention) and follow-up 2 (four months after baseline). Negative effects are defined as adverse events (AE) and severe adverse events (SAE). Specifically, AE comprise i) an increase in target symptoms (posttraumatic and depressive symptomatology, suicidality), ii) the occurrence of novel psychological or physical symptoms, and iii) any negative events, all of which may or may not be associated with trial participation, in both IG and CG. Participants will be questioned whether any AE is thought to emerge from trial participation. SAE are defined as severe negative events that occur during study participation and result in adverse reactions requiring some form of high intensity treatment (i.e., deliberate self-harm, suicide attempt, life-threatening events, non-elective or extended hospitalization, an event causing chronic or serve disability) or fatality, including suicide. SAE will be surveyed and reported regardless if they are deemed related to the trial or not. The principle investigator (SRH) will be immediately informed about the occurrence of any AE and SAE. Likewise, a consensus conference (incl. the principle investigator and clinical investigators) will be held immediately to decide on appropriate measures in response to AE and SAE. Moreover, data collected on AE and SAE will be analyzed and reported as part of the study outcomes.

Practically, during the trial, study participants of both groups (IG and CG) will be provided with detailed information regarding emergency hotlines and professional help, if needed, i.e., in case of unexpected symptom deterioration. The IG will have access to this information via an emergency button leading to a help menu within the app. The CG receives the same information as part of the psychoeducational reading material that is handed out at baseline. In addition, study participants of both groups will be encouraged to contact our study nurses if they need further support to find emergency help.

Auditing will take place in form of independent reviews of the data collection across all assessment waves. Specifically, 5 % of the questionnaires at baseline, follow-up 1 and follow-up 2, respectively, will be randomly drawn and inspected regarding their degree of matching with the database input. Source data verification will be performed independently from investigators and study sponsors by commissioned, external statisticians.

In case of necessary protocol amendments, we will describe all relevant details/changes and associated rationales.

The results of the RCT will be published in international, peer-reviewed journals. Moreover, we plan to present our findings at international and national conferences and meetings relevant to the field. To reach an even broader publicity, we aim to cooperate with the largest health insurance provider in Germany (AOK - Federal Association of Local Health Insurance Funds), national institutions concerned with migration and integration as well as stakeholders/protagonists of refugee care centers. We will additionally arrange several information events for experts from the field of psychiatry and psychology as well as for trauma-specific experts and representatives of psychosocial counselling centers. Lastly, we aim to share our results via public media channels. We aim to make the app freely available to everyone.