Protocol for a randomized control trial of the caregiver support intervention with Syrian refugees in Lebanon

The primary goal of this RCT is to evaluate the effectiveness of the CSI among Syrian refugee and host community caregivers in northern Lebanon.

This will be a parallel group superiority RCT with an intent to treat design, a 1:1 allocation ratio, and a waitlist control comparison group. The study will be conducted in two waves, due to the large number of participants. Recruitment for each wave will take place in the neighborhoods of greater Tripoli (see the “Setting” section), with wave 2 recruitment starting during implementation of wave 1. This will help us ensure a high quality of implementation and rigorous data collection at all assessment points.

The Government of Lebanon estimates that 1.5 million Syrian refugees, 54% of them children, are currently living in Lebanon, with about 1 million registered with the United Nations High Commissioner for Refugees (UNHCR) [22]. Sixty-nine percent of Syrian families in Lebanon live below the poverty line. This study will take place in Northern Lebanon, one of the most impoverished regions in the country. Specifically, the study will take place in the greater Tripoli/T5 region in the governorate of North Lebanon. As of January 2017, roughly 70,000 registered Syrian refugees were living in Tripoli; however, the actual number is widely understood to be much higher, as many Syrians have not registered as refugees with the UNHCR [22]. A 2017 study by the Feinstein Center found that approximately 75% of Syrians in Tripoli are living below the poverty line, with a monthly income below 180,000 LBP ($120). Their living conditions are generally precarious, work restrictions are increasingly stringent, and wages are among the lowest in Lebanon [8].

All assessments and intervention groups will be held in the community centers of community-based organizations (CBOs) with which War Child Holland collaborates, in or adjacent to the target communities. To be eligible to host the intervention, partner CBOs must have adequate space to conduct CSI groups with 12 adults, as well as a separate and adequately large space for childcare during the intervention and data collection.

Participants will primarily be Syrian refugee caregivers of children aged 3–12 years, with one index child per family. Up to 25% of the sample may be comprised of Palestinian refugee and/or Lebanese families, in keeping with the Lebanon government’s policy of ensuring that programming for refugees is also available to host communities. In families with more than one child in the 3–12 year age range, the research coordinator will select one child randomly by rolling a die. In the Standard Operating Procedure document on the selection of the index child, a table specifies the meaning of the die roll. For example, if a family has two eligible children, a roll of one–three means the younger child is selected, while four–six means the older children is selected. Estimated total sample size is 240 families and 480 parents/caregivers (see the “Statistical power and sample size” section).

The CSI is a nine-session weekly group intervention, co-facilitated by trained non-mental health professionals, who receive 6 days of training, three on-site observations with feedback, and weekly supervision. Groups are offered separately to women and men and are run with 10–12 participants. Table 1 lists session topics and corresponding modules, along with the stress management technique(s) taught in each session.

Sessions 1–4 are focused on strengthening caregiver wellbeing, with individual sessions on understanding and managing stress, disengaging from “thinking too much”—a culturally salient phenomenon that both indicates and exacerbates stress, roughly akin to rumination [23], and coping with anger and frustration, all while developing the group as a socially supportive setting. Sessions 5–8 focus on strengthening parenting under conditions of adversity (i.e., increasing awareness of the impact of stress on parenting, increasing positive parent–child interactions and the use of non-violent discipline methods, and reducing harsh parenting). Session 9 involves a review and closing of the intervention. In all but the final session, participants learn a new relaxation or stress management technique, drawn or adapted from the mindfulness and stress management practice. These techniques are also provided to participants in Arabic on mp3 files, which they can listen to on their smart phones or on mp3 players provided at the start of the program. Participants are encouraged to practice these activities at least three times each week. A considerable amount of time is spent at the start of each session reviewing the home practice and collectively problem-solving any barriers to practicing the techniques.

This study is employing a waitlist control as the comparison group, based on our successful use of the same design in the pilot RCT of the CSI. Participants in the waitlist control group in each wave of the study will be invited to participate in CSI groups that will begin shortly after the 3-month follow-up assessment has been completed.

Participant recruitment will be conducted in collaboration with the staff of the local CBOs with which War Child has collaborative relationships in each target community. Community breakfasts to announce the study, door to door recruitment—especially important for recruiting men—along with visits by research staff to settings where men commonly gather, flyers posted in local partner CBO offices, and word of mouth will all be used to recruit participants into the study. In order to recruit men/male caregivers, we will utilize the same strategies that proved effective in the pilot RCT: scheduling assessments and intervention sessions on days and at times that do not conflict with income-generation opportunities, training our recruitment team on the importance of men’s participation in the CSI, and crafting a recruitment message that emphasizes both the stress management and parenting foci of the intervention (Miller KE, Koppenol-Gonzalez GV, Arnous M, Tossyeh F, Chen A, Nahas N, et al.: Supporting families displaced by armed conflict: A pilot randomized controlled trial of the caregiver support intervention, submitted).

All questionnaire data will be gathered using tablets, using the software program Kobo, which allows questionnaires to be completed and uploaded without paper and pencil. Kobo is available free of charge from the Harvard Humanitarian Initiative (https://​www.​kobotoolbox.​org/​). Measures will be administered in Arabic to each parent/caregiver individually by trained and supervised research assistants. See Table 2 for an overview of all measures.

Three new measures were developed for this study: Caregiver stress, Stress management, and Parenting. The development and piloting of these measures is described in a forthcoming paper. Briefly, we decided to develop new measures for these key outcomes after we searched the literature extensively and were unable to identify measures that [1] were suitable for caregivers with children of such a broad age range (3–12 years); had been validated for use with Syrians, Lebanese, or Palestinians; and [3] were worded in ways that would be deemed acceptable in the target communities. All items on the new questionnaires were drafted in English (with Arabic terminology and idioms in mind), reviewed by a panel of experts, translated into Arabic and back-translated into English, with all discrepancies resolved through a consensus process among bilingual project staff. The items were then assessed for ease of understanding and cultural acceptability through a process of cognitive interviewing with groups of Syrian caregivers from the target community. All items were deemed acceptable and were readily understood. Minor wording changes were made to several items to ensure the intended meaning was conveyed. This same process of cognitive interviewing was also undertaken with the other questionnaires in the study. All measures, including the three newly developed questionnaires, were then administered to a group of 50 caregivers on two occasions, one week apart, to assess their internal consistency and test–retest reliability. The internal consistency and test–retest reliability of the three new measures ranged from acceptable to high, as described below.

Informed consent will be gathered at the baseline assessment, prior to gathering any data. The research coordinator will distribute a consent form in Arabic, and read it aloud to participants to ensure full comprehension regardless of their literacy levels. The research coordinator will allow time to address all questions and concerns participants may raise.

After caregivers have completed the baseline assessment, families will be randomized to the CSI or a waitlist control group. Randomization will be at the family level, to ensure that caregivers from the same family are not randomized to different arms of the study; 240 families will be randomized to the CSI or control arm. CSI groups will be held separately for women and men, yielding a total of 120 families and 240 caregivers per arm.

As we are running the study in several communities, a block randomization design will be used, using a participatory methodology implemented successfully in our pilot RCT, and adapted from the work of Mercy Corps and Yale University in their study of a life skills program for Syrian refugee children in Jordan [29]. At baseline assessment, after completing the questionnaires, one caregiver from each family will be asked by a research team staff member to draw a lollipop out of an opaque bag filled with an equal number of red and green lollipops to ensure an equal number of CSI and waitlist control participants. Caregivers will be told that after baseline data have been completed, a coin toss will determine the meaning of each color: one color will mean CSI and the other color will mean waitlist control. This process will be repeated in each of the communities where the study will be conducted, leading to an equal number of CSI and WLC families in each community and in the study as a whole. After the coin toss, done by a staff member of War Child Holland unaffiliated with the study, the outreach team will inform all participants of their group assignment and let CSI participants know the day and time of their weekly group sessions. A research team member will manage the lollipop selection, while a War Child staff member unaffiliated with the study will toss the coin.

The purpose of this two-step randomization process is to increase community buy-in to the randomization process, by demystifying it and giving participants an active role in the process. We successfully randomized participants in the pilot study in this way (there were no significant between-group differences on any variable following randomization). Moreover, participants understood the process and expressed a willingness to accept assignment to either the CSI or waitlist control arm. This willingness was confirmed by the high percentage of WLC participants who completed the post-intervention assessment (99%) in the pilot study.

A master list will be created that includes each family’s group assignment. This list will be kept in a secure location in the War Child Holland Tripoli office, with a copy in a similarly secure location in the War Child Office in Beirut. Only the research coordinator and the research implementation coordinator will have access to this master file during the study.

Questionnaire data will be gathered at three time points: baseline, post-intervention, and 3-month follow-up. Because the study will be implemented in two waves, there will be a total of six assessments, three in each wave. This is depicted in the SPIRIT figure chart in Fig. 2.

Participant flow, from enrollment through the 3-month follow-up assessment, can be seen in Fig. 3.

In the pilot study, retention of the WLC group was facilitated by periodic WhatsApp text messages reminding WLC participants of the post-intervention assessment and the CSI intervention that would follow shortly thereafter. In the present study, there will be an additional three and a half month period due to the follow-up assessment before the WLC group receives the intervention. In order to maintain interest in the study, a separate social event will be held for the WLC and CSI groups in each wave. The event will be a gathering in a park or community center with food and music for a few hours, and will occur during the month following the post-intervention assessment.

A target of 480 caregivers (240 families) will be recruited in two waves. We initially calculated a sample size based on 90% power at 5% level of significance and an attrition rate of 20%, yielding a sample of 432 individuals or 216 families. However, because we aim to have 12 participants in each CSI group, we increased the sample size to 480 (240 families). Dividing the sample over two waves, with randomization into CSI and waitlist control in each wave, this sample size yields a total of ten CSI groups per wave (five for women and five for men), for a total of 20 CSI groups.

We will investigate the mediation process of different aspects of caregiver wellbeing and, through that, illuminate key basic knowledge about generalization of acquired skills in parenting through the CSI intervention. Some of the pathways of interest are illustrated in Fig. 4. If the efficacy analysis shows significant between-group differences in the mediators (caregiver stress (Caregiver Stress Scale), caregiver distress (K10), caregiver psychosocial wellbeing (WEMWBS), stress management), then we will use parametric regression models to:

Since all the measures are continuous, the indirect effects will be calculated by multiplying relevant pathways and bootstrapping will be used to produce valid standard errors for the indirect effects. All analyses will adjust for baseline measures of the mediators (Caregiver Stress Scale, caregiver distress (K10), stress management, outcome (Parenting scale) and putative measured confounders (e.g., socioeconomic status).

Prospective CSI facilitators must meet the following criteria to be accepted into the 6-day training:

CSI facilitators first participate in a 6-day training which combines didactic and experiential components, and includes extensive practice implementing the intervention within the training group. In addition to covering the specific content of the sessions (e.g., parental stress and wellbeing, relaxation exercises, coping with anger and frustration, child development, positive parenting), trainees also learn group facilitation skills through extensive role-playing and continuous feedback. The trainers complete a detailed competency checklist for each trainee at the start and completion of the training; trainees also complete a knowledge and attitudes assessment at the start and completion of the training. Only those trainees who are deemed sufficiently knowledgeable and skilled, based on a trainer-completed competency checklist and the pre-post knowledge assessment, are then invited to become CSI facilitators. During their first full implementation of the CSI, facilitators are observed a minimum of three times with the competency checklist completed on each occasion, and receive weekly supervision.

Childcare will be provided on-site for up to two children per participant. Trained and experienced animators (child care providers) trained by War Child Holland in child care and child safety will care for children, with two animators available during each session. The childcare sessions are not structured, and no formal activities are facilitated, in order to avoid providing any index children attending the childcare with a substantive psychosocial experience that might confound the results of the study. Various play materials and snacks will be provided, and the child care workers will ensure the safety and comfort of all children.

A fidelity checklist will be completed jointly by the co-facilitators immediately after each session of the CSI. The number of items on the fidelity checklist varies by session, as it covers all of the activities specific to each session. The research implementation coordinator and the CSI trainer and supervisor will conduct three on-site observations in each group and provide coaching to all new facilitators based on these observations. They will use the fidelity checklist and a supplementary competency checklist to guide their feedback. They will also meet with all pairs of facilitators weekly for supervision. Registers will be used to record attendance at all sessions of all CSI groups. These will be kept in the secure care of the research implementation coordinator. The research implementation coordinator and CSI trainer and supervisor will receive their own supervision weekly from War Child Holland’s Lebanon-based psychosocial coordinator, and biweekly supervision from a psychologist and faculty member the University of Balamand (Nahas).

The PI will hold weekly meetings with the research coordinator and research implementation coordinator to review all activities and issues and ensure fidelity to the research protocol.

To minimize contamination, randomization will be at the family rather than individual parent level, in order to avoid having parents/caregivers within the same family assigned to different arms of the study. To minimize contagion between the CSI and control arms, participants in CSI groups will be asked to avoid sharing program content, including the relaxation and stress management exercises, with anyone outside of their immediate household.

All assessments will be carried out by trained research assistants (see the “Research assistants” section for details about background and training of the research assistants). Efforts will be made to minimize interactions between the intervention and research teams during the study. As noted earlier, the rationale for keeping the RAs blind is explained to participants at each assessment, along with a request that they not reveal their group assignment. This was accomplished successfully during the pilot study, and we anticipate no difficulties replicating this in the present study.

Following the intervention and within 3 weeks after the post-intervention data collection, focus groups discussions will be conducted by the research coordinator and a trained research assistant with all groups of CSI participants (male and female separately), to examine perceptions of the intervention, perceived impact on personal wellbeing and parenting, and barriers to accessing the intervention and to utilizing intervention content. The focus groups will also explore participants’ experience of completing the assessment battery (e.g., level of difficulty, perceived burden).

Participants will be assigned a study ID, and their names will not appear on any data collected. Questionnaire data will be uploaded directly by the research coordinator to the secure KOBO server, from which it will be downloaded to WCH’s own secure server and entered into SPSS (quantitative) or NVivo (qualitative). All data will be backed-up securely at the WCH Amsterdam office, and all data and analysis files will be securely password protected and encrypted. For all questionnaire data, range and consistency checks will be performed within days post data-entry, which will occur immediately following each wave of data collection. Any queries identified will be resolved promptly by the trial management team. A detailed WCH data management policy will serve as guidance on all data management issues. Focus group notes will be taken verbatim in Arabic, translated into English, and entered into NVivo, where they will be coded and analyzed by the research team.

We will monitor the occurrence of AEs using the War Child Holland (WCH) Adverse Events Reporting Procedure. A data safety and management committee (DSMC) will provide oversight on AEs and safety protocols for the study. The DSMC will be composed of individuals to be named prior to the start of the trial, none of whom will be a part of this study. The key terms of reference for the DSMC will be to review reports of AEs (within 48 h of notification). Reviews should occur on a 12-monthly basis via Skype. The purpose of the DSMC is to monitor the occurrence of AEs and where required make decisions on further actions to be taken to determine whether AEs are likely to be related or unrelated to the intervention. The DSMC will have the mandate to recommend stopping the study if it is determined that the risks of the study outweigh the benefits.

All AEs reported spontaneously by participants or observed by the investigators or other staff members will be recorded by the research team on WCH’s Adverse Event reporting form. AEs can be detected by anyone as all study staff will be trained in their detection and management. Intervention facilitators will initially discuss AEs reporting with their supervisor, who will report this to the research coordinator. Research assistants will initially discuss AE reporting with the research coordinator. The research coordinator will subsequently share all AEs with the PI, who will be responsible for ensuring appropriate responses to all AEs. All AEs will be reported to the DSMC. This will occur within 24 h for AEs. The chair or nominated person from the DSMC will review AEs within 48 h and the DSMC will review all AEs monthly and where necessary to determine whether AEs are likely to be related or unrelated to the intervention.

Any participant may elect to discontinue their participation in the study at any point for any reason, including participation in the intervention. Any participant who becomes repeatedly disruptive to the intervention may be asked to discontinue participation in the intervention, but not the study.

A small reimbursement (approximately 5 USD) of transportation costs of caregivers and children to the organization offices and/or community centers will be provided for assessment sessions. Transportation will be provided for attendance at assessment and intervention sessions. Refreshments will also be provided at both intervention and assessment sessions. This compensation is considered minimal and in line with regular WCH implementation programming, and was chosen in order to avoid difficulties of jealousy within the community amongst non-eligible families, and potential coercion of families to participate in sessions and assessments.

A team of 12 research assistants will be trained for a period of at least 5 days prior to the baseline assessment of wave 1. This group will include a sub-group of experienced research assistants who worked on the CSI pilot RCT, as well as newly trained RAs. Research assistants are native Arabic speakers, preferably with experience serving refugee and vulnerable communities, and will be recruited in Lebanon based on recommendations from local colleagues. Training will cover research ethics, child safeguarding, handling adverse events, building rapport with participants, communication with children, troubleshooting, and the use of the tablets, among other topics, and will include extensive practice administering all questionnaires to be used in the study.

This research protocol has been approved by an institutional review board at the University of Balamand in Tripoli, Lebanon. Written (or witnessed, if the participant is not literate) informed consent by caregivers will be mandatory for enrollment in all research activities. We will protect the confidentiality of personal data principally through procedures to separate study data and participant identifiable data. No children will be involved as participants in this study.

The results of this project will be published in English in peer-reviewed journals (we will aim to publish in open access journals). They will be disseminated in Arabic and English to key stakeholders through reports and presentations. If the CSI is shown to be effective, an adaptation manual will be developed in the final months of the grant period to facilitate the implementation by War Child in other countries where the organization is active, and by other institutions and non-governmental organizations interested in using the CSI in their programming.

Any substantive amendments to this trial protocol will be submitted to the institutional review board committee at the University of Balamand that approved the original protocol. Such amendments will also be submitted to ISRCTN for inclusion in the registered protocol.