Go Big or Go Home: Scaling Mental Health Solutions That Actually Work

Tom Osborn, Founder & CEO, Shamiri Institute

Scaling mental health solutions is the defining challenge of our generation. America spends more on mental health than any nation yet has some of the worst outcomes. Kenya has fewer than 500 psychiatrists serving 50+ million people. Traditional therapy was never designed for contexts with severe provider shortages and limited budgets. But what if peer-delivered group therapy could achieve 62% effect size, nearly identical to traditional US therapy's 64%, at $7.50 per person annually? At Shamiri Summit 2025, Tom Osborn revealed the "Four Enoughs" framework that enabled his organization to reach over 100,000 young people across Kenya. This isn't theory. It's a proven blueprint for solving problems at the scale they exist.

Why Even Rich Countries Can't Solve Mental Health

Mental health represents what Tom Osborn calls a "complex problem," one that even the wealthiest nations haven't cracked. America spends more per capita on mental health than any country yet consistently ranks among the worst for treatment outcomes and access. The challenge cuts across geographies because no one has truly figured out the sustainable formula.

In Kenya, the complexity multiplies exponentially. The country has fewer than 500 psychiatrists serving more than 50 million people. Traditional Western therapy assumes unlimited access to PhD-level psychologists or MD-level psychiatrists, with one-on-one sessions lasting 8-12 weeks at costs exceeding $100 per session. This model was never designed for contexts with severe provider shortages, limited budgets, and cultural barriers to seeking clinical care.

The standard response to resource constraints is to scale down ambitions and serve fewer people better. Tom Osborn's presentation at Shamiri Summit 2025 argued the opposite: complex problems demand solutions designed for scale from Day 1.

The Shamiri Model: Three Ideas That Created Something New

Shamiri Institute didn't invent task-shifting, brief interventions, or community-based care. But the combination created tiered psychotherapy designed specifically for resource-constrained contexts achieving outcomes comparable to traditional therapy at a fraction of the cost.

1. Task-Shifting (The 'Who')

Task-shifting means training non-professionals to deliver services traditionally reserved for specialists. Shamiri adapted this by training 18-22-year-old young people as peer therapy providers called Shamiri Fellows. These Fellows complete structured training programs and deliver group-based interventions under clinical supervision.

This directly addresses scarcity. If you don't have enough psychologists, train community members. It also reduces stigma by making mental health support feel peer-based rather than clinical. Research from Africa, India, Asia, and South America over the past two decades shows task-shifted mental health interventions can achieve outcomes comparable to specialist-delivered care.

2. Brief Evidence-Based Interventions (The 'What')

Simple, short interventions (4-6 sessions) can produce results comparable to 12 weeks of traditional CBT. Brevity enables scale. If you achieve meaningful outcomes in 4 sessions instead of 12, you serve 3x more people with the same resources.

Shamiri uses modules focused on growth mindset, gratitude, and value affirmation. Each intervention can be delivered in group format over 4 weekly sessions. Research demonstrates that teaching problem-solving skills or helping people make value-aligned decisions produces clinically significant outcomes without requiring extensive therapeutic relationships.

3. Community-Based Delivery (The 'Where')

Shamiri embeds mental health support within existing structures, schools in particular. This makes well-being less isolated and clinical, more integrated into social environments where young people already spend time.

By bringing services to where youth are rather than requiring them to seek specialized clinics, the model reduces barriers (transportation, stigma, time) and normalizes mental health conversations. Instead of waiting for depressed students to seek out a far-away clinic, Shamiri brings group therapy sessions to their school campus, led by Fellows who are only slightly older and from similar backgrounds.

The Innovation: Tiered Psychotherapy Architecture

The breakthrough wasn't any single idea. It was the system architecture combining all three:

Level 1: Peer Providers - Shamiri Fellows (18-22-year-olds) deliver group-based brief interventions (4 sessions) in schools and community centers. Each Fellow serves 30+ youth per cycle.

Level 2: Clinical Supervisors - Diploma-holders in psychology provide quality assurance and one-on-one support for complex cases. Each supervisor oversees 15-20 Fellows.

Level 3: Specialists - PhD-level psychologists and MD-level psychiatrists handle the most severe cases requiring specialized expertise. This level is reserved for 5-10% of cases.

This tiered approach solves the access problem by reserving scarce professional time for cases that truly need it, while empowering community members to handle the majority of mental health support needs through evidence-based protocols.

The Results: 75% of participants no longer met clinical criteria for depression after four weeks. 82% no longer met criteria for anxiety.

Comparison: Shamiri vs. Traditional Therapy

The Four Enoughs Framework: Making Solutions Scalable

Proving your model works (good enough) is only 25% of the scalability equation. Tom Osborn outlined three additional criteria every scalable solution meets:

Enough #1: Good Enough (Evidence-Based Outcomes)

Does your intervention produce clinically meaningful results? Shamiri conducted rigorous randomized controlled trials (RCTs) before scaling. Using common language effect sizes enabled direct comparison with traditional therapy, demonstrating 62% effect size vs. 64% for US therapy.

Don't scale until you have peer-reviewed evidence. Run RCTs in multiple contexts to prove replicability. Calculate cost per clinically meaningful outcome (Shamiri: $22 vs. hundreds for traditional therapy).

Enough #2: Big Enough (Addressable Market)

Even if your model works perfectly, is the population that can access it large enough to justify investment?

Shamiri's addressable market analysis considers geography (can the model operate in regions with necessary population density?), demographics (does the target population exist in sufficient numbers?), policy environment (will regulations allow implementation?), and talent availability (can you recruit enough providers?).

Map your ideal model against real-world constraints (infrastructure, literacy, geography, policy). Calculate intersection to determine true addressable market size.

Enough #3: Simple Enough (Partner Replicability)

Can organizations beyond yours implement your model without your direct involvement?

Shamiri partnered with other organizations to implement the model independently. Partner-led implementation produced similar trajectories to Shamiri-led implementation, proof the model transfers successfully.

Document every process. Train partner organizations. Measure their outcomes vs. your direct implementation. If partners can't replicate your results, your model isn't simple enough to scale beyond your organization.

Enough #4: Cheap Enough (Affordable at Scale)

Is your cost per beneficiary sustainable if government or beneficiaries themselves are paying?

Shamiri's Cost Evolution:

• 2018: $21 per person (unsustainable)

• 2024: $7.50 per person (750 Kenyan shillings)

• Target: $3 per person (enables direct parent/government payment)

But cost alone isn't enough. Cost-effectiveness matters more:

Cost per clinically meaningful outcome: Shamiri: $22 to move someone from clinically depressed to not depressed. Traditional therapy: $300-600 per outcome.

Cost per quality-adjusted life year (QALY): Shamiri significantly outperforms most healthcare benchmarks requiring $100,000+ per QALY.

Obsess over cost reduction through operational efficiency, technology automation, and process optimization. But also calculate and communicate cost-effectiveness to funders and governments who need ROI justification.

Why Ideas Accomplish Nothing Without Activities

Tom Osborn's foundational insight: An idea accomplishes nothing until it connects to measurable outcomes through specific behavioral changes and strategic activities.

Shamiri's mission is enabling young people to thrive. But how does tiered psychotherapy actually achieve that mission?

The Critical Behaviors That Need to Change

For Shamiri's model to work, several behaviors occur:

1. The Ministry of Education approves the program

2. Schools agree to host sessions

3. Young people sign up for therapy

4. Young people actually attend sessions

5. Shamiri Fellows deliver interventions effectively

If any link in this chain breaks, you achieve zero impact. Brilliant innovation means nothing if schools refuse participation or students don't show up.

The Drivers Behind Behaviors

Behaviors don't change spontaneously. Someone executes specific activities:

To get Ministry approval: Engage officials, secure MOUs, present research evidence, attend policy meetings, submit documentation.

To get schools to sign up: Conduct outreach campaigns, meet with principals, demonstrate value proposition, create scheduling systems, provide implementation support.

To get students to attend: Student messaging campaigns, peer influence activation, convenience optimization using psychological motivators (belonging, status, fairness, safety).

To train Fellows: Build recruitment pipelines, create training curricula, develop quality assurance systems, provide ongoing supervision.

From One-Time Actions to Sustainable Systems

These aren't one-time activities. They're ongoing processes. Speaking to one school doesn't create sustainable behavior change. Building systems that continuously drive the right behaviors is what creates scale.

Instead of manual outreach to individual schools, Shamiri built automated school matching algorithms, templated partnership agreements, scheduled touchpoint systems, and feedback loops where school satisfaction scores inform continuous improvement.

For every critical behavior in your theory of change, identify: (1) What specific activities drive that behavior, (2) Who is responsible for executing those activities, (3) What systems/technology can make those activities scalable, (4) What metrics indicate those activities are working.

AI + Automation: From Thousands to Millions

Having designed an innovative, scalable model with proven outcomes, the final question remained: How do we actually reach millions of people?

Tom Osborn's answer: Technology is the accelerator, not the solution.

Technology Application #1: Operational Efficiency

Technology automates manual processes: Recruitment streamlines how Shamiri identifies, screens, and onboards Fellows. Training uses digital tools enabling standardized delivery while reducing cost. Payment systems ensure Fellows are compensated reliably through mobile money integration. Supervision technology allows supervisors to oversee more Fellows with better quality assurance.

Over the past year, Shamiri built a technology stack reducing operational overhead by 40%, enabling the same staff to support 60% more Fellows.

Technology Application #2: AI for Precision Care

How do you match each person to the right intervention, at the right intensity, delivered in the right format? AI analyzes individual characteristics (symptoms, preferences, contexts) to recommend: which tier (peer-led Level 1, clinical Level 2, specialist Level 3), which intervention module (growth mindset, gratitude, value affirmation), and which delivery format (in-person groups, digital, hybrid).

Better matching improves outcomes while using resources more efficiently. Instead of defaulting everyone to Level 1, AI identifies the 20% who need Level 2 earlier, preventing crisis escalation.

Technology Application #3: AI-Powered Provider Supervision

How do you supervise 500+ peer providers at scale without sacrificing quality?

Becoming a great therapist happens in two stages: Training (learning theory and techniques) and Frontline Delivery (actually delivering sessions, learning from experience). The second stage is where providers develop true expertise, but it's also where they need supervision the most.

Shamiri spent 12 months building an AI supervision tool that: analyzes session recordings to identify what providers do well and where they struggle, provides personalized feedback to each provider, identifies patterns across providers to improve training curricula, and enables one supervisor to support 40+ Fellows (vs. 15 without AI).

If one supervisor can oversee 40 Fellows instead of 15, the same supervision capacity serves 2.6x more beneficiaries without hiring additional clinical staff.

Key Takeaways for Practitioners

Start with context: Don't import solutions wholesale. Adapt proven ideas to your specific constraints.

Test rigorously: Run RCTs to prove your model works before scaling. Common language effect sizes allow comparison across interventions.

Design for scale from Day 1: Group-based delivery, peer providers, and community integration aren't "nice-to-haves." They're enablers of reach.

Obsess over cost: Drive unit costs down aggressively, but also measure cost-effectiveness to justify investment.

Partner to validate simplicity: If others can't implement your model, it won't scale beyond your organization.

Embrace technology strategically: Use it for efficiency first, then for innovation (AI-enhanced care).

Measure obsessively: Track which activities drive behaviors that produce impact, then double down on what works.

About Shamiri Summit 2025: Wires & Signals took place November 13-14, 2025 at Glee Hotel, Nairobi, Kenya, bringing together mental health innovators, practitioners, policymakers, and researchers to explore the systems we build, the messages they carry, and the collective futures they create for youth wellness across Africa.