The Impact of Lifting Barriers to Mental Health Care for Refugees: Evidence from a Randomized Controlled Trial

This intervention conducts a randomized controlled trial to evaluate interventions aimed at improving access and utilisation of mental health services among recently recognized refugees in the Paris region, with a particular focus on demand-side obstacles. It takes advantage of the AGIR programme (Accompagnement Global et Individualisé des Réfugiés), which provides individualized support to improve refugees’ integration into employment and housing. More precisely, this research project tests whether peer support and simplified access to mental healthcare, such as assistance with appointment scheduling, contribute to putting refugees on the path to better integration.

2025-05-26

2026-06-30

1. Mental health service utilisation (short-term outcomes)
2. Mental health status (intermediate outcomes)
3. Healthcare knowledge: (intermediate outcomes)
4. Socioeconomic integration (longer-term outcomes)

Outcomes. Our analysis will focus on three families of outcomes including:

1. Mental health service utilisation (short-term outcomes): This list of outcomes aims to capture the effectiveness of our intervention in increasing
refugees’ uptake of mental healthcare.14 We construct a set of healthcare access measures for refugee i, capturing both the extensive margin (dummy
variable) and the intensive margin (continuous counts) across three dimensions:
• Contact: Whether, and how frequently, the individual engaged with any of the four partner healthcare facilities or the helpline.
• Schedule: Whether, and how often, the individual scheduled an appointment at a partner facility, either directly or via the helpline.
• Attend: Whether, and how many times, the individual attended a scheduled appointment at a partner facility.

For each of the three dimensions, namely contact, scheduling, and attendance, we also construct a variable Speed, measuring the number of months between treatment assignment and the first recorded contact, scheduled appointment, or attendance. This allows us to capture the speed of engagement, which can provide additional insights into the effectiveness of the intervention in facilitating quicker access to services.

To capture attendance outside of our partner facilities, we augment the Attend variable with data from the endline questionnaire, including a dummy variable equal to one if the refugee has consulted any mental health professional in the past year (extensive margin), as well as the number of times the
refugee has seen a mental health professional (intensive margin).

2. Mental health status (intermediate outcomes): This set of outcomes aims to capture the effectiveness of our intervention in improving refugees’ men-
tal health. We construct a series of self-reported mental health measures for refugee i, either measured in the endline survey or computed as changes be-
tween the baseline and endline surveys:
• PTSD and C-PTSD symptoms: measured through the International Trauma Questionnaire (ITQ) items.
• Self-rated improvement in health: Categorical response to a question on whether overall health has improved over the past year.
• Self-rated improvement in mental health: Categorical response to a question on whether mental health has improved over the past year as
a result of receiving mental health therapy.
• Self-rated general health: Categorical response to a question assessing general health status.
• Health satisfaction: Categorical response to a question on the level of satisfaction with current health.

3. Healthcare knowledge: (intermediate outcomes): This set of outcomes aims to capture the effectiveness of our intervention in improving refugees’ under-
standing of the French healthcare system. We construct a series of self-reported measures of mental health knowledge for refugee i, drawn either from the endline survey or computed as changes between the baseline and endline surveys:
• General knowledge: Categorical response to a question assessing the respondent’s knowledge of the French healthcare system.
• Self-scheduling capacity: Categorical response to a question on how easy it is for the respondent to obtain an appointment at a healthcare
facility when needed.

4. Socioeconomic integration (longer-term outcomes): This set of outcomes aims to capture the effectiveness of our intervention in improving refugees’
economic assimilation. We construct a series of employment and housing measures for refugee i, drawn from either the endline survey or the administrative dataset. These outcomes are measured either at the end of the treatment period or as changes between the treatment date and the end of the project.
• Employment status: Dummy variable indicating whether the participant is currently employed.
• Activity type: Set of dummy variables indicating the nature of employment (full-time, part-time, training etc.).
• Employment satisfaction: Categorical variable indicating the level of satisfaction with current employment status.
• Employment stability: Variable indicating the number of different jobs held by the refugee during the past year.
• Housing status: Dummy variable indicating whether the participant has a stable housing.
• Housing type: Set of dummy variables indicating the nature of housing (private, emergency shelter, social housing, etc.).
• Housing satisfaction: Categorical variable indicating the level of satisfaction with current housing situation.
• Housing stability: Variable indicating the number of different housing arrangements occupied by the refugee during the past year.

Depending on the findings, we may also investigate the mechanisms through which the intervention operated, using a set of additional quantitative and qualitative outcomes measured in the endline survey. These include detailed questions on why individuals did not seek mental health support, and how they obtained an appointment if they did.

None

Refugees enrolled in the AGIR programme meet regularly (at least every two months) with their assigned social workers. These meetings will serve as an opportunity to inform refugees about the project, obtain their consent, and implement the intervention. Refugees already enrolled in the project, as well as those entering it, will be assigned to one of four treatment arms, as described below. All treatments will be implemented on the same day, in the refugee’s native language, immediately following the provision of informed consent. We describe below in detail these four treatment arms:

(C) Control Group: Basic information. Participants receive a standardised information guide in their language listing available mental health services in the Paris region. It includes information on the name, address, nearest metro station, opening hours, and method of consultation (with or without appointment, or both), as well as telephone numbers and email addresses. A “RISE Partner” stamp is also added to the leaflet, allowing refugees to easily identify services that are affiliated with the project.
(G1) Treatment 1: Basic information and discussion with a peer-helper. In addition to the standard information guide (C), participants have a one-on- one confidential discussion with a peer helper in their language. Peer-helpers are recruited from a university diploma in mediation and interpretation in a migration context offered by INALCO (Paris). The interview focuses only on mental health issues and mental health care structures tailored to refugees in Paris. Project staff do not have access to this discussion or its content. As supported by the literature review, this treatment conjectures that face-to-face interactions between refugees and trained peer supporters can reduce sociocultural barriers, self-stigma, and mistrust towards mental health professionals. The peer helper team was recruited with attention to gender balance,
ensuring that, whenever possible, female refugees are supported exclusively by same-sex peer helpers. The languages supported by the team of peer helpers are similar to those used in the videos.

(G2) Treatment 2: Basic information and access to a helpline. In addition to the standard information guide (C), participants are granted access to a helpline that provides personalised support from a peer-helper who speaks their language, specifically to assist with booking appointments at mental health facilities. The phone number for this helpline is provided by attaching a label to the back of the information leaflet. There is one telephone number per refugee language. Figure 5 depicts this label. The sticker is placed in front of the refugee when the information guide is handed over to increase the salience of this treatment. The same individuals who provide peer support will be responsible for answering the helpline and scheduling appointments at a facility for the refugee who calls, using a phone provided to them. It should also be noted that this assistance is limited to scheduling the first appointment only. As supported by the literature review, this treatment conjectures that providing telephone assistance leading to appointments at appropriate care facilities can eliminate structural barriers to accessing mental health care.

(G3) Treatment 3: Basic information, discussion with a peer-helper and access to a helpline. In addition to the standard information guide (C),
participants receive both the peer-helper discussion (G1) and helpline access (G2). This treatment is thus a combination of the two previous treatments and is based on the hypothesis that these two approaches are complementary. By simultaneously addressing both sociocultural and structural barriers to accessing care, it aims to improve the ability of refugees to seek treatment.

Not available

In office by a computer

Stratified randomisation at the individual level (strata is defined as group of nationality x gender)

No

Around 2000

Around 2000

Around 500 per treatment arm. 4 groups.

For primary outcomes, employment and autonomous housing, we can detect effects of between 8.5 and 10.5 percentage points under Scenario 1 (1733-2167 individuals) and between, 6.8 and 8.4 percentage points under Scenario 2 (2734-3417). These magnitudes appear sufficient given historical trends: ELIPA2 (2023) documents 20 and 24 percentage point increases in employment and housing, respectively, over three years (2019-2022).