Pioneering research on the path to mental healthcare for all
We believe rigorous research is the foundation of scalable, effective care. Continuous learning allows us to scale with fidelity and rigor—enabling transformative care models that can reach millions.
From evidence to impact, continuously
We believe the future of mental healthcare depends on systems that can both prove and improve what works. Rigorous research allows us to establish causal evidence of what is effective. Continuous real-world learning allows us to scale those models without losing fidelity or quality. Our Evidence Engine establishes effectiveness. Our Optimization Engine learns from real-world delivery—allowing us to scale care with fidelity, rigor, and continuous improvement.
Evidence Engine
Establishes what works
We use rigorous, gold-standard research to answer foundational questions: what works, for whom, and why. Through randomized trials, long-term follow-up, and careful measurement of clinical outcomes, we generate causal evidence that defines effective care.
Optimization Engine
Learns how to scale what works
We continuously test and refine how care is delivered in real-world settings. Through rapid experimentation and data-driven iteration, we learn how to scale proven interventions without losing fidelity, quality, or impact.
The four conditions for mental healthcare to reach scale
Through our Evidence Engine, we study not just what works—but whether it can scale. We organize this learning around four questions. Together, they define the conditions required for mental healthcare to reach millions—and test whether our model can meet them.
Good Enough
Does it work?
We start with the most fundamental question: does this intervention create real, meaningful change? Through rigorous research, we establish whether effects are statistically significant, clinically meaningful, and sustained over time.
Across multiple randomized controlled trials, Shamiri demonstrates clinically meaningful reductions in depression, with outcomes comparable to traditional one-on-one therapy delivered by experts. After just four sessions, 80% of participants no longer meet the threshold for clinically elevated depression and anxiety—effects that are sustained across the full school year.
*Data from gold-standard randomized controlled trial of the Shamiri model, published in JAMA Psychiatry.
How we know
- ✓Pre-registered trials with long-term follow-up
- ✓Standardized, validated clinical instruments
- ✓Clinically meaningful thresholds to assess recovery
Learning how to scale what works
Proving what works is only the first step. The harder challenge is learning how to deliver it—across systems, at scale, and without losing quality. Through our Optimization Engine, we continuously test, adapt, and improve how care is delivered in the real world. These are some of the questions that we are answering right now to shape the future of scalable mental healthcare.
Can governments deliver care effectively?
Real scale happens when governments and partners can implement the Shamiri model effectively. We are currently testing this through ongoing work with government partners. Early evidence suggests that the key to unlocking governments as implementers lies in building the right enabling infrastructure—training systems, supervision, and ongoing support. With these in place, governments can deliver care at scale with comparable outcomes.
Can supervision be automated without losing fidelity?
We believe automation and AI can play a critical role in maintaining fidelity as our model scales. Early evidence shows that AI-assisted supervision can match human fidelity scores while significantly reducing cost—pointing to a scalable path for sustaining quality.
Do mental health interventions drive broader long-term life outcomes?
Early findings suggest strong links between improved mental health and gains in functioning. Ongoing work is quantifying impacts on education, employment, livelihoods, and long-term health.
Can we intervene earlier—and effectively?
New pilots are exploring delivery for younger adolescents, testing whether earlier intervention can shift long-term trajectories.
Eight years of building the science of scalable mental healthcare
Over the past eight years, we have built one of the largest evidence bases for youth mental health in sub-Saharan Africa. Our work has progressed from early pilots to large-scale trials—each stage designed not only to prove what works, but to understand how it can scale. This is a journey from efficacy to impact: from demonstrating clinical outcomes, to understanding mechanisms, to testing delivery at scale, and now to building systems that can reach entire populations.
12+
Trials
18K+
Participants
35+
Publications
2018
Shamiri 1.0 RCT
First pilot demonstrating that a brief, lay-provider intervention can improve adolescent mental health.
2019
Shamiri 2.0 RCT
First large-scale RCT published in JAMA Psychiatry, showing strong clinical outcomes, with ~80% recovery sustained over time.
2021
Shamiri 3.0 RCT
Multi-arm RCT identifying what drives improvement—moving from “does it work” to “why it works.”
2022
Dissemination RCT 1.0
First dissemination study demonstrating that partners can deliver the model with comparable outcomes across contexts—evidence of scalability.
2023
Dissemination Trial 2.0
Second trial demonstrating that partners can deliver the model with comparable outcomes across contexts—evidence of scalability.
2024
Livelihoods & long-term impact
Expanding beyond clinical outcomes to test links between mental health, education, and livelihoods.
2025
Current frontier
Beyond mental Testing government delivery, AI-assisted supervision, and early-adolescent adaptation—building the systems required for population-level scale. — TVET/livelihoods
Explore the full portfolio
35+ papers, datasets, white papers, and policy products
Featured research
Randomized controlled trial of the Shamiri intervention
A randomized trial of 413 Kenyan adolescents found that Shamiri, a 4-week lay-provider-delivered intervention, produced greater reductions in depression and anxiety symptoms at posttreatment (d = 0.35 for depression, d = 0.37 for anxiety) and sustained effects through 7-month follow-up.
Pilot randomized trial of Shamiri in Kibera
A pilot RCT of 51 Kenyan adolescents in an urban slum found that the Shamiri intervention produced greater reductions in depression (d = .32) and anxiety (d = .54) compared to study skills control, and improved academic performance.
Cluster-randomized trial of three single-session interventions
A cluster RCT of 895 Kenyan adolescents found that the values intervention significantly reduced anxiety symptoms in the full sample (d = 0.31), while both values and growth mindset reduced anxiety in the clinical subsample.