Experimental Design
We refer to the primary Statistical Analysis Plan (SAP) of the Pamoja cRCT for full details on the experimental design, study population, sampling, timeline and data collection (https://www.isrctn.com/ISRCTN34179105). Here, we provide a summary of the main aspects. The cRCT uses a two-stage random sampling design. First, 18 zones (clusters) were selected within three refugee settlements in South-Western Uganda, based on (i) refugee presence from the Democratic Republic of the Congo (DRC) (target is more than 70% in overall sample); (ii) operational capacity of WCA or partners; (iii) absence of WCA evidence-based mental health interventions (e.g. TeamUp, BeThere, EASE) or comparable interventions by other organizations implemented in the 6 months prior to baseline; and (iv) local authority approvals.
The study included one eligible school per zone. Schools qualified if they (i) served ages 11-16 (at least 300); (ii) were willing to participate; (iii) could provide enrolment lists; (iv) were located more than 5 kilometers from a school in an adjoining cluster that had been enrolled in the study; and (v) were run by the government. In case of multiple eligible schools per zone, one school that met eligibility criteria was randomly selected.
The 18 schools were matched by number of students to form 9 pairs. Within each pair, one school was randomly assigned to the Pamoja intervention (treatment arm) and the waitlist control arm. Schools and families from zones in the control arm did not receive any of the Pamoja intervention components during the first year after the baseline assessment. They received the program upon completion of the endline assessments, one year after baseline. Treatment assignment of schools and classes was masked to the investigators, except the scientific coordinator, until endline data collection and data cleaning was completed.
The target population includes 11-to-16-year-old adolescents (at the time of baseline), living in one of three refugee settlements in South-Western Uganda – Kyangwali, Kyaka, and Nakivale. These settlements primarily host refugees from the DRC; 27-29% of their populations are below age 18.
For the Pamoja cRCT, a random sample of 252 adolescents, aged 11-16, was drawn from each school. Adolescents were randomly sampled from enrolment lists, stratified by gender (1:1). Eligibility criteria for adolescent participation included (i) living in the study zone; (ii) being aged 11 to 16 years old at the time of baseline; and (iii) speaking one of the four local languages (Kinyabwisha, Congolese Kiswahili, Runyankole or Runyoro/Rutooro).
The Pamoja cRCT study sample hence comprises 4,536 adolescents across 18 clusters (9 per arm), with 252 respondents per cluster, and their primary caregivers. Our study will be conducted using the same sampling frame and experimental design as the primary Pamoja cRCT.