Boosting refugee integration through psychological intervention

The low-intensity psychological intervention Problem Management Plus (PM+) was originally developed by the World Health Organisation (WHO) and the University of New South Wales to reduce psychological distress for people exposed to adversities. PM+ has proved to be an effective method for reducing mental health problems and improving the psychosocial functioning of people in crisis in diverse contexts. PM+ is a peer-based psychological intervention in which individuals without specific medical-therapeutic backgrounds are trained to become lay therapists (i.e., so-called «helpers»). Its peer-based approach overcomes socio-cultural and language barriers by having peers from the same cultural background deliver the intervention. This format enables PM+ to be cost-effective and readily scalable. The intervention consists of five weekly sessions covering four core components: stress management, problem management, behavioral activation, and strengthening social support. In this study, in addition to the original five sessions, participants will receive three booster sessions and access to a digital support tool. Each session is conducted one-on-one in person by a trained “helper” with one participant and lasts approximately 90 minutes. In addition, participants will receive three 30-45-minute telephone booster sessions 10, 22, and 34 weeks after the baseline assessment. These sessions will review how participants are using the PM+ skills and focus on troubleshooting any problems participants have in implementing such skills to cope with ongoing stressors. Participants will also receive access to online materials to practice PM+ strategies.

To shed light on the mechanism through which the treatment can affect economic decision-making for participants, we cross-randomize a trauma priming intervention that has been used in the literature in psychology and economics. We follow the approach used by Callen et al. (2014) to administer two emotional primes just before asking participants to make a sequence of risky investment choices (based on the global preferences survey, Falk et al. 2018). One half of respondents receive a positive prime that reads: “We are interested in understanding your daily experiences that make you happy or joyous. This could be anything, for example birth of child, marriage of a relative, or success in your job. Could you describe an event in the past year that caused you happiness?”. The other half of respondents receives a negative prime aimed at triggering a fearful recollection: “We are interested in understanding your daily experiences that may make you fearful or anxious. This could be anything, for example getting sick, experiencing violence, losing a job, etc. Could you describe one event in the past year that caused you fear or anxiety?”. The priming will be conducted only at the first follow-up survey (3 months after baseline) on a random subsample of respondents. The elicitation of time preferences (based on the global preferences survey, Falk et al. 2018) will follow that of risk preferences, so that we can also test whether the priming interacts with treatment in affecting intertemporal choices and patience.

2024-05-13

2027-05-30

The main outcomes of interest is:
- Immigrant Integration Index (IPL-24) measured at the 12-month endpoint

Immigrant Integration Index (IPL-24): To generate the aggregate integration index, we will use principal component analysis (PCA) to extract the first principal component from all IPL-24 questions.

We will additionally study the impact of the augmented PM+ on the following secondary outcomes:

1. Mental health: anxiety and depression (HSCL-25) measured 12 months after (this is the primary outcome with the same 12-month endpoint in the mental health work packages. We expect that much to most of the integration benefits of PM+ work through improving refugees’ mental health as measured with HSCL-25.

2. Six individual components of the Immigrant Integration Index (psychological, navigational, economic, linguistic, political, social).

3. Economic and educational situation: employment status, earnings, reliance on social welfare, enrolment in language courses, enrolment in any other school.

4. Health care utilization: prescription of health medication and utilization of health-related health care services, along with their associated costs. Where feasible, we distinguish in addition between mental and non-mental-health related outcomes.

In addition, for a random subsample of respondents in the treatment and control group (Subsample 1) who are exposed to the cross-randomized emotional primer, we will collect the following secondary outcome measures to better understand the potential underlying mechanisms through which improved mental health can affect economic decision-making:

5. Risk preferences (global preferences survey)
6. Time preferences (global preferences survey)

For risk and time preferences, we will exploit the cross-randomized emotional primes to study how the prime interacts with treatment exposure.

Because participation in the intervention is voluntary and the augmented PM+ consists of several components, it is also important to assess treatment take-up. We will analyze the take-up of PM+ based on the following measures: the number of completed PM+ and booster sessions and user data from the digital platform. In addition, we will collect survey data on the practice of PM+ strategies for a random subsample of respondents in the treatment and control group (Subsample 1), so we can estimate treatment effects on this outcome.

The secondary analyses will explore the evolution of treatment effects over time by making use of the survey data collected during the first two endpoints (3 and 6 months after baseline).

Components of the Immigrant Integration Index: Using the IPL-24 module, we will build a separate index for each of the six integration dimensions covered by IPL-24 (following Harder et al., 2018). For these outcome measures, we will extract the first principal component using polychoric principal component analysis (PCA) from the IPL-24 questions relevant for each dimension.

Risk preferences: The measure of risk preferences is based on the staircase risk task used by Falk et al. (2018), which relies on five interdependent choices. The task will be unincentivized. In each question, participants decide between a sure choice and a lottery. The resulting risk measure is obtained following the tree logic provided by Falk et al. (2018) in their online appendix.

Time preferences: The measure of time preferences is based on the staircase time task used by Falk et al. (2018), which relies on five interdependent choices. The task will be unincentivized. In each question, participants decide between a sum of money today or a sum of money in 12 months. The resulting patience measure is obtained following the tree logic provided by Falk et al. (2018) in their online appendix.

Recruitment and informed consent

The study will focus on refugees and asylum seekers (RAS) who speak at least one of 12 languages (German, English, French, Arabic, Farsi, Kurdish, Tigrinya, Turkish, Ukrainian, Russian, Tamil and Pashto), covering approximately 80% of individuals with a refugee background in Switzerland (State Secretariat for Migration, 2020). Study participants will be recruited through the implementation project SPIRIT in various local sites in the cantons that agreed to participate in the experimental study BRIGHT. The implementation project SPIRIT offers PM+ as a standard low-intensity mental health service to eligible participants in various local sites in the six asylum regions of Switzerland (State Secretariat for Migration, 2023). The cantons that agreed to participate in the experimental study currently include Schwyz, Thurgau, and Zurich. The recruitment will target RAS who reside in these cantons and are aged 16 or older.

Eligibility to participate in the experimental study will be assessed during an initial screening meeting held in person by a trained assessor who is fluent in the participant’s mother tongue. The screening and all subsequent assessments will rely on a digital platform set up for the project. Eligibility will be based on several inclusion and exclusion criteria (see the supporting document “Eligibility criteria” for a detailed description of the eligibility criteria). The main inclusion criterium is based on the score obtained on the Kessler Psychological Distress Scale (K10; Kessler et al., 2002), a brief clinically validated screening questionnaire that assesses general psychological distress in the past 30 days. This study will use a score equal to or greater than 20 as an indication of moderate to high levels of psychological distress. Participants with a K10 score equal to or higher than this cut-off value will be considered eligible to participate. In addition to the K10 score, several exclusion criteria will be assessed in the screening meeting. These include:

- significant cognitive or neurological impairment measured through specific tools developed by the WHO and integrated into the PM+ manual (impairment questionnaire);
- acute medical conditions or severe mental disorders (e.g., psychotic or substance-abuse disorders) measured through specific tools developed by the WHO and integrated into the PM+ manual (impairment questionnaire, World Health Organization WHO, 2016);
- acute risk of suicide as measured by the Suicidal Ideation Attribution Scale (SIDAS; van Spijker et al., 2014) and the Thoughts of Suicide Questionnaire (World Health Organization WHO, 2016).

The study will rely on two parallel channels to recruit participants. First, interested RAS can register for the initial screening meeting online, by filling out a form on the website of our field partners or on the central landing page of the SPIRIT project website. Second, several field partners will provide contact lists of RAS they are working with. The study coordinators of the field partners will then try to schedule a screening meeting with all persons on the contact lists. The combination of these two approaches is aimed at increasing the representativeness of the study sample by reducing self-selection while allowing all interested RAS to participate in the initial screening.

During the screening meeting, interested RAS will be asked to provide their consent to participation. The informed consent will entail a two-step procedure: a written informed consent form for the screening and baseline assessment, provided at the beginning of the screening meeting; a second informed consent form for taking part in the PM+ program and in follow-up assessments, provided only to eligible participants. At the beginning of the screening session, the assessor will cover the content of the first IC form: they will explain the nature of the study, its purpose, the steps involved, the expected duration, the data confidentiality, the potential risks and benefits and any discomfort it may entail. Each participant is informed that participation in the study is voluntary, that they may withdraw from the study at any time, and that withdrawal of consent will not affect their subsequent medical assistance and treatment. The subjects are informed that they can ask any question during the study and consult with family members, friends, their treating physicians, or other experts before deciding on their participation in the study. Enough time will be given to the subjects to make an informed decision about their participation in this study. Only if participants agree to and sign the IC form will they proceed to the screening and baseline assessment.

With the first consent form, all subjects will be asked to consent to the following (“IC for the screening and baseline assessment”):

1. Participation in the screening and baseline assessments
2. Sharing their external health insurance provider data
3. Linking their survey data and linked insurance provider data with administrative data (data on the asylum process, characteristics of the asylum seekers, population data, AHV contributions, and income) collected by government agencies (State Secretariat for Migration, Federal Statistical Office, Central Compensation Office)

Subjects eligible to receive the intervention will additionally be asked to consent to the following (“IC for the intervention”):

4. Being randomized into the treatment (PM+ now) or control (PM+ later) group
5. Voluntarily participating in the PM+ intervention
6. Being invited to multiple follow-up assessments


Randomization samples and treatment assignment

The study will rely on a randomized phase-in design. All eligible participants who meet the inclusion criteria, do not meet any of the exclusion criteria, and provide consent to participate in the BRIGHT experimental study, will be randomly assigned to one of the two intervention groups:

1. Treatment group (Group A) which receives the augmented PM+ from the beginning: the first PM+ session will be scheduled to take place within two weeks from the baseline assessment.
2. Control group (Group B) which receives the augmented PM+ 12 months later: the first PM+ session will be scheduled to take place approximately 12 months after the baseline assessment.

This experimental design allows for studying the causal effects of the augmented PM+ intervention over the course of 12 months by comparing treated participants with a pure control group. At the same time, all eligible participants will be able to benefit from the intervention, as those assigned to the control group gain access to it after 12 months.

Participants in Group A and Group B will then be randomly assigned to one of the following two groups:

1. Subsample 1 which additionally receives questions on risk and time preferences (global preferences survey) and questions on the practice of PM+ strategies.
2. Subsample 2 which additionally receives questions on self-efficacy and prolonged grief.

Within Subsample 1, respondents from the Treatment (Group A) and Control (Group B) groups will be cross-randomized to receive either the positive or the negative emotional prime before the risk and time preference questions.


Outcome Data Collection

The research team will collect outcome data from surveys and administrative data sources (see supporting documents “Survey questions and schedule” and “Administrative data sources”). The first (screening) survey will be administered before randomization. This first survey will be administered by an assessor fluent in the study language selected by the participant. This survey will serve the purpose of assessing the eligibility of respondents to participate in the study. Randomization of eligible participants who consent to participate in the experimental study takes place at the end of the screening survey and is followed by the baseline assessment.

The baseline assessment is self-administered online while the assessor stays present and remains available to address any questions that may arise as participants complete the questionnaires. The administration of the baseline assessment is conditional on participants not meeting any of the exclusion criteria. Participants with a K10 score below 20 who do not meet any of the exclusion criteria will still receive the baseline assessment but will then no longer be part of the study. Their baseline data, complemented with register data, will be used to conduct additional analyses based on non-experimental causal inference techniques, such as RDD.

Subsequently, follow-up assessments for participants in the two experimental groups will occur according to a pre-specified schedule (3 – 6 – 12 months after the baseline survey) and will be self-administered online as a default. At the end of the screening and baseline meeting, assessors will indicate whether a respondent requires assistance for the follow-up surveys, in which case the surveys will be administered either in person or online with the provision of assistance from an assessor.

Not available

Randomization done by statistical software

Individual

No

N/A

For the experimental study, we aim to recruit 1,500 refugees and asylum seekers. The target number refers to the total number of participants who meet all eligibility criteria and consent to participate in the experimental study. To meet this target, a large proportion of RAS residing at local sites will be screened. The randomization will not be clustered, but to ensure a balanced assignment of units to the treatment and control groups over time, we will implement a block randomization based on blocks of 48 subjects. Within each block of 48 subjects, 24 will be randomly assigned to the treatment group and 24 to the control group. The block randomization is implemented by the digital platform set up for the study. Since conducting a pilot study was not possible, we expect implementation issues, particularly at the beginning of the study. We, therefore, keep the option to drop the observations from the first 40 participants who enrolled in the study (treatment and control). The associated data will be discarded and used as equivalent to pilot data to identify and fix any implementation issues.

Of the projected 1500 participants to be recruited for the study, half will be randomly assigned to the intervention group (Group A; N = 750) and half to the control group (Group B; N = 750). The intervention group receives the augmented PM+ intervention right after the screening, while the control group receives it 12 months later.

Of the projected participants in the treatment Group A, half will be randomly assigned to Subsample 1 (N=375) and half to Subsample 2 (N=375). The same applies to Group B. Participants in Subsample 1 and Subsample 2 receive a different subset of additional questions, which are described in the Experimental Design section.

Within Group A and Group B, participants in Subsample 1 will be cross-randomized to receive either the positive or the negative emotional prime before the risk and time preferences questions.