Global Health
Crisis Methods for Crisis Settings: Flexing Our Randomised Controlled Trials to Flexibly Test Behavioural Interventions in War Sydney Tucker* Sydney Tucker Tucker Tucker Tucker Tucker Tucker Tucker Tucker sydney.tucker@spi.ox.ac.uk
Background
Two-thirds of the world’s children live in conflict-affected countries. War-time threats present challenges for rigorously evaluating interventions – including loss-to-follow-up and ethical challenges such as urgency to deliver potentially beneficial interventions without delay. Epidemiologists must develop rigorous yet flexible methodologies to identify effective solutions to protect children in war.
Methods
We conducted a two-armed parallel-group randomised controlled trial among N=350 displaced Palestinians in Jordan. We utilised rolling recruitment with a randomised roll-out design. Participants were 1:1 randomised to Hope Groups – a mental health, parenting, and violence prevention support group intervention – or waitlist control. The primary analysis estimates the effectiveness of Hope Groups post-intervention on parent mental health and violence against children. Our primary model is a two-level generalised linear mixed-effects model nesting person within timepoint to estimate incidence rate ratios, with confidence intervals estimated via wild bootstrapping. Building on theory from Rolled-Entry Matching, we conduct a secondary analysis estimating the intervention’s effects at 3, 6, and 12 months follow-up, comparing previous intervention participants to newly enrolling untreated participants at each timepoint.
Results
We observed high response rates: 91.2% of participants completed endline surveys. After the intervention, the control arm’s parental depression and anxiety rate was 1.85 (95%CI=1.52,2.17) times the rate in the intervention arm. Similarly, the control arm’s rate of violence against children was 1.41 (95%CI=1.16,1.72) times the intervention arm’s rate. Our innovative long-term follow-up design shortened control arm wait-times by 4-8x compared to traditional RCT approaches. Follow-up results are pending and will be reported at SER.
Discussion
In an era of polycrisis, developing innovative impact evaluation methods is a public health priority.