Beside

Posts

Where to Launch Next: A Framework for Prioritizing Countries for Digital Depression Programs

Deciding where to launch a global health program is one of the most consequential decisions an NGO leader makes. Launch in the wrong context — insufficient digital infrastructure, hostile regulatory environment, existing well-resourced programs already serving the population — and you generate modest impact at high cost. Launch in the right context — large unmet need, viable delivery infrastructure, a regulatory environment that permits the model — and every dollar of organizational investment translates into substantially more impact.

For digital depression programs specifically, the decision involves a distinctive combination of factors that don’t apply to most other global health interventions. The technology-mediated delivery creates digital infrastructure requirements. The lay counselor model creates workforce requirements. The evidence-based content creates adaptation requirements. And the mental health context creates stigma and regulatory dimensions that don’t exist for bed net distribution or vaccination programs.

This post lays out the framework that Ambitious Impact developed for prioritizing countries for WHO Step-by-Step expansion, the ten countries that emerged from that analysis, and the practical considerations that should inform how organizations sequence geographic expansion in this space.

The Framework: What Matters for Country Prioritization

Country prioritization for digital depression programs requires assessment across five dimensions, each of which can be a constraint or an enabler depending on context.

1. Depression Burden

The most fundamental criterion is the size of the addressable problem. Countries with higher depression prevalence — whether due to conflict, poverty, displacement, or other structural factors — have more potential beneficiaries per unit of organizational investment. Depression burden is measured using Disability-Adjusted Life Years (DALYs) lost to depressive disorders, standardized by population.

High-burden contexts for depression include countries with significant conflict-affected populations (where depression rates are substantially elevated above non-conflict baselines), countries with high rates of poverty and economic precarity, and countries where structural factors — gender inequality, limited social support systems, rapid urbanization — contribute to elevated depression prevalence.

Burden alone doesn’t determine priority, but a context with low depression burden relative to other options should rarely be prioritized, regardless of how favorable other factors are. Impact per dollar is fundamentally constrained by addressable need.

2. Treatment Gap

Burden must be considered alongside existing provision. A country with high depression burden but well-developed, accessible mental health services represents a different opportunity — and a different level of additionality — than a country with comparable burden and near-zero existing provision.

The treatment gap — the proportion of people with depression who receive no treatment — provides a direct measure of unmet need. In most LMICs, the treatment gap is 75 percent or higher, but there is meaningful variation between countries. Countries where existing provision is genuinely scarce, and where a new program would be reaching populations with no current access to comparable care, represent higher-additionality opportunities than those where some provision exists, even if imperfect.

3. Digital Infrastructure

The WHO Step-by-Step model, as deployed by Kaya Guides, requires access to a smartphone and a WhatsApp account. This requirement is less restrictive than it might appear — WhatsApp penetration rates in many target countries are high, and smartphone ownership has expanded rapidly — but it is still a real constraint that varies significantly across and within countries.

Relevant metrics include: smartphone ownership rates (overall and by income quintile, gender, and urban/rural location), WhatsApp penetration and usage patterns, mobile data affordability and reliability, and electricity access for device charging. Programs that require continuous connectivity will be more constrained by infrastructure gaps than those that allow offline engagement or asynchronous delivery.

Digital infrastructure gaps are not static — they are closing rapidly in most LMICs — which means a context that is marginally below the threshold today may be viable within two to three years. Country prioritization should consider the trajectory of digital infrastructure development, not just the current state.

4. Regulatory Environment

Mental health is a regulated domain in virtually every country. The specific regulatory questions for a program like Step-by-Step include: What are the legal requirements for delivering psychological support? Can lay counselors legally conduct the support calls that are central to the guided self-help model? Are there requirements for clinical oversight or professional registration that would add cost or operational complexity? What are the data protection and privacy requirements for handling sensitive mental health information?

Regulatory environments for lay counselor delivery vary substantially. Some countries have explicit frameworks for community mental health worker roles that accommodate the Step-by-Step model. Others have no clear framework, creating legal ambiguity. A small number have restrictive requirements that would make lay counselor delivery of psychological support illegal or prohibitively complex.

Regulatory risk is not always a hard constraint — ambiguous environments can sometimes be navigated through engagement with regulatory authorities, and regulatory frameworks evolve. But organizations should enter new country contexts with a clear-eyed assessment of regulatory risk and a strategy for managing it, rather than discovering barriers after operational investment has been made.

5. Operational Feasibility

Even in contexts that score well on burden, treatment gap, digital infrastructure, and regulatory environment, operational realities can constrain program viability. Relevant factors include: the organization’s existing presence or partnerships in the country, the availability of local staff with the language skills and cultural competence to manage the program, the ability to recruit and train lay counselors from the target community, the stability of the operating environment (political, security, economic), and the availability of referral pathways for participants who need more intensive support than the program provides.

Operational feasibility considerations often argue for sequencing expansion in ways that allow organizational learning to compound. A second country that is culturally and operationally similar to the founding context is lower-risk than a second country that requires entirely new languages, counselor recruitment channels, and partnership relationships. The learning from context A reduces the cost and risk of entering context B when they share relevant features.

The Ten Priority Countries

Applying this framework, Ambitious Impact’s research identified ten countries as priority targets for WHO Step-by-Step expansion: Pakistan, China, Nigeria, Bangladesh, Indonesia, Egypt, Brazil, Tajikistan, Ethiopia, and Malaysia.

Each country has a distinctive profile that explains its ranking. A brief analysis of the key factors for each:

Pakistan scores highly on depression burden (elevated by conflict exposure, poverty, and high rates of gender-based adversity), treatment gap (mental health infrastructure is extremely limited, particularly outside major urban centers), and digital infrastructure (rapidly expanding smartphone penetration and high WhatsApp usage). The program has already been tested in a Pakistani RCT, reducing cultural adaptation requirements. Regulatory environment for lay counselors is manageable.

China presents a different profile: middle-income country with strong digital infrastructure, high WhatsApp-equivalent penetration (though regulatory complexity around foreign apps requires consideration), substantial depression burden driven by rapid urbanization and social isolation, and a significant treatment gap despite relatively better economic resources than other priority countries. A Chinese-language RCT of Step-by-Step has been completed, providing a head start on cultural adaptation.

Nigeria is the largest country in Sub-Saharan Africa by population, with a very large absolute depression burden, near-zero specialist mental health infrastructure, rapidly growing smartphone penetration, and a population that is predominantly young — a demographic that is both at elevated depression risk and particularly likely to engage with digital delivery.

Bangladesh combines high depression burden (driven by poverty, climate vulnerability, and demographic pressure), a large addressable population, growing smartphone penetration, and limited existing mental health provision. Geographic density — one of the world’s most densely populated countries — creates favorable economics for lay counselor recruitment and supervision.

Indonesia has the fourth-largest population in the world, high WhatsApp penetration, significant depression burden, and a growing digital health ecosystem that creates potential for partnership and integration with existing programs. Linguistic diversity (hundreds of regional languages) creates adaptation complexity that should be factored into sequencing decisions.

Egypt benefits from a completed Arabic-language RCT of Step-by-Step, making it one of the best-evidenced contexts for program deployment. Depression burden is significant, WhatsApp penetration is very high, and the urban population’s familiarity with digital tools supports engagement. An Egyptian-language program is effectively ready for deployment, reducing the adaptation investment required.

Brazil represents the largest opportunity in Latin America — a country with substantial depression burden, Portuguese-language adaptation requirements, high smartphone penetration, and significant geographic variation in mental health provision (urban areas have some access; rural and peri-urban areas have much less). Brazil’s regulatory environment is complex but navigable.

Tajikistan may surprise as a priority given its relatively small population, but it scores highly on the combination of very limited mental health infrastructure, high depression burden in a post-Soviet context with significant economic precarity, and a regulatory environment shaped by the Soviet healthcare legacy that includes community health worker roles compatible with lay counselor delivery.

Ethiopia has one of the largest absolute populations in Sub-Saharan Africa, very limited mental health provision, and a rapidly expanding digital infrastructure — still with significant gaps, particularly in rural areas, but improving. The Ethiopian context is higher-complexity than some other priority countries, but the scale of unmet need justifies the operational investment.

Malaysia is an upper-middle-income country with strong digital infrastructure, high WhatsApp penetration, and — despite relatively better economic development — a significant mental health treatment gap driven partly by stigma and partly by the concentration of specialist services in major urban centers. The economic profile makes Malaysia suitable for a model with slightly higher per-participant cost than the lowest-income contexts.

Sequencing Expansion: Practical Considerations

Having a list of priority countries doesn’t answer the sequencing question: which to enter first, and in what order? Several principles should guide this decision for organizations in Kaya’s position.

Sequence for learning, not just impact. Earlier expansions should be chosen partly for their learning value — what will operating in this context teach us that will make subsequent expansions more efficient? Countries that share features with later priorities (language family, regulatory framework type, digital infrastructure profile) are valuable early entries because the learning compounds.

Minimize the number of simultaneous unknowns. Every new country entry involves multiple new unknowns — counselor recruitment, cultural adaptation, regulatory navigation, partnership development. Entering a new country that differs from the founding context on all of these dimensions simultaneously creates high operational risk. Sequencing to minimize the number of new unknowns per entry — choosing contexts that share language, or regulatory framework, or partnership network — reduces risk and improves the odds of a successful first cohort.

Build the referral pathway before scaling. The absence of referral infrastructure — clear pathways for participants who need more intensive support — is a safety and quality risk. Before scaling to significant participant volumes in a new context, organizations should invest in establishing at least basic referral relationships with whatever higher-level services exist, even if imperfect.

Consider the funding implications of geographic choice. Some contexts are more legible to specific funders than others. An organization that relies heavily on funders based in the UK may find that expansion to South Asia is more fundable than expansion to Latin America, not because of any difference in impact potential but because of funder familiarity and existing portfolio logic. Geographic sequencing should account for the funding implications of each choice, particularly for early-stage organizations where funding constraints are binding.

Don’t let the priority list become a constraint. The ten-country list is a starting point derived from the best available evidence, not a fixed mandate. If a compelling partnership opportunity, regulatory opening, or funding relationship creates a strong case for entering a country not on the list, the framework should inform — not override — that decision. The goal is maximum impact, and sometimes the path to maximum impact deviates from any pre-specified plan.

The Competitive Landscape

Country prioritization decisions should also consider the competitive landscape — whether other organizations are working effectively in the same space in the target context. The goal of NGO country selection should not be to compete with existing high-quality programs, but to extend provision to populations not currently being served.

In mental health in LMICs, the competitive landscape is generally sparse — most contexts have too little provision of any kind for “competition” to be a meaningful concern. But as the field develops and more organizations scale digital mental health programs, this will change. Maintaining awareness of what is being built in target contexts, and positioning expansion to be additive rather than duplicative, will become increasingly important.

Conclusion

Country prioritization for digital depression programs is a multi-dimensional analytical exercise that rewards rigor. The combination of depression burden, treatment gap, digital infrastructure, regulatory environment, and operational feasibility creates a complex decision space — but one that can be navigated systematically using available data.

The ten countries identified in the Ambitious Impact analysis represent a strong starting point for organizations working in this space. They are not equally ready for program launch — each has a distinctive profile of opportunities and challenges — and the sequencing decision should be made with as much care as the initial selection.

The ultimate measure of a country prioritization framework is not its analytical elegance but its impact: whether the populations in the countries selected are better served as a result of the decisions made. At $97 per DALY averted, each person reached in a well-chosen, well-executed country expansion represents exceptional value. The framework exists to maximize how many of those people can be reached.

To learn more about Kaya Guides’ expansion strategy and how to support the scaling of WHO Step-by-Step globally, visit besidehealth.org.

Related reading

Building the World’s First Nonprofit Around WHO Step-by-Step: Lessons from Kaya Guides’ Launch Year

Most nonprofits are founded by people who have lived experience of a problem, or who stumbled into a cause through circumstance. Kaya Guides was founded differently: it was selected, incubated, and launched by Ambitious Impact (also known as Charity Entrepreneurship), a research-driven incubator that uses systematic evidence review to identify the most cost-effective opportunities for new organizations to create impact.

The decision to build Kaya around WHO Step-by-Step was not intuitive or serendipitous. It was the product of a rigorous analysis of the global mental health landscape, the evidence base for different intervention types, the cost-effectiveness of delivery models, and the gap between existing provision and addressable need. That analytical foundation shapes everything about how Kaya was built — and provides a distinctive set of lessons for charity entrepreneurs and NGO leaders thinking about how to enter high-need, evidence-rich spaces.

This post draws on Kaya’s first year of operations — from launch in August 2023 through approximately mid-2024 — to examine what worked, what was hard, and what the organization’s early trajectory suggests about the challenges and opportunities of building at the frontier of global mental health.

Why Step-by-Step, and Why Now

Understanding Kaya’s launch requires understanding the analytical process that produced it. Ambitious Impact’s approach to identifying incubation opportunities involves several stages: systematic literature review across cause areas, cost-effectiveness modelling of candidate interventions, assessment of the organizational landscape (are effective organizations already doing this?), and evaluation of founding team fit.

The case for guided digital self-help for depression in LMICs emerged from this process for several reasons. The evidence base was unusually strong: five randomized controlled trials of WHO Step-by-Step, totalling more than 2,200 participants, producing consistent effect sizes comparable to face-to-face therapy. The cost-effectiveness was exceptional: approximately $97 per DALY averted at projected scale, competitive with malaria bed nets and other top-tier global health interventions. And the organizational landscape had a clear gap: no organization was systematically scaling the Step-by-Step model with the rigor and focus that the evidence warranted.

That last point is crucial. The decision to incubate a new organization rather than support an existing one rests on a gap analysis. If high-quality organizations are already working effectively in a space, the marginal return to incubating a new one is low. If the space is neglected — if the evidence exists but no one is doing the work — then a new, purpose-built organization can generate substantial additional impact.

Global mental health, and specifically guided digital self-help for depression in LMICs, was genuinely neglected. The WHO had developed Step-by-Step but lacked the mandate and infrastructure to operate a direct service delivery organization. Academic researchers had conducted the trials but were not in the business of scaling programs. Existing NGOs in the mental health space were mostly not working at the intersection of digital delivery, lay counselors, and rigorous evaluation that Step-by-Step represents. The gap was real.

Building the Model: The First Decisions

Kaya launched in India — a decision driven by a combination of factors: large depression burden, high WhatsApp penetration, a regulatory environment that allows lay counselor delivery of psychosocial support, and the ability to operate in English alongside local languages, which simplified early-stage operations before full localization infrastructure was in place.

The core operational model was established from the outset around the counselor-to-participant ratio that the evidence supports: one trained lay counselor per approximately 400 enrolled participants per year. Each counselor conducts the weekly 15-minute check-in calls that research shows are essential for differentiating guided from unguided self-help outcomes. The counselors are recruited from local communities, trained by Kaya, and supervised on an ongoing basis.

This model has direct implications for unit economics. The primary variable cost in the program is counselor time — a fixed number of hours per participant per week, regardless of how many total participants are enrolled. Technology costs (WhatsApp Business API access, participant management systems) are relatively low and scale favorably with volume. The implication is that Kaya’s per-participant cost should decline as the organization grows, with fixed overhead costs spread across a larger participant base.

By mid-2024, approximately 3,600 participants had been served. This is meaningful at the level of individual lives improved, but it is orders of magnitude below the potential scale of the program’s addressable population.

Lesson 1: The Counselor Model Is the Core Challenge

If there is one operational lesson that dominates Kaya’s first year, it is this: building and maintaining a high-quality lay counselor workforce is harder than it appears, and it is the central challenge to scale.

Recruiting counselors from local communities addresses several problems simultaneously: it keeps costs lower than professional alternatives, it provides cultural and linguistic proximity to participants, and it creates local employment. But it also creates challenges that professional workforce models don’t face.

Lay counselors are not career mental health workers. Many are taking on this role alongside other responsibilities. Turnover can be significant — people move, circumstances change, and the emotional demands of supporting people through depression are real even with only 15 minutes of contact per participant per week. Maintaining quality and consistency requires supervision infrastructure that itself requires skilled staff to deliver.

The supervision model matters as much as the training model. Initial training prepares counselors for the role, but ongoing supervision is what maintains quality, identifies counselors who are struggling or drifting from protocol, and provides a pathway for continuous improvement. Building this infrastructure — the supervisors, the supervision protocols, the quality monitoring systems — is less visible than the counselor training itself but at least as important.

For organizations entering this model, the implication is to invest in supervision infrastructure earlier than feels necessary. The cost of poor-quality counselor delivery isn’t just reduced program effectiveness — it’s reputational and ethical risk at a level that could compromise the organization’s ability to operate.

Lesson 2: Technology Is an Enabler, Not the Product

The shift from a dedicated app to WhatsApp delivery that occurred in later iterations of Step-by-Step reflects a broader lesson: the technology platform that works best is the one that participants actually use, regardless of how sophisticated it is.

WhatsApp has penetration rates above 80-90 percent in many of Kaya’s target markets. It doesn’t require a new download, a new account, or a learning curve. It works on low-end smartphones with limited data connections. And critically, it integrates the program into a communication tool that participants already use for family and social contact, reducing the sense of separateness that can make engaging with a mental health program feel like a stigmatized act.

The lesson for NGO leaders is to resist the pull of technology-forward thinking in program design. The most elegant app, the most sophisticated AI-driven chatbot, the most feature-rich digital platform is worth less than a solution that participants encounter where they already are. Technology should serve the program model, not define it.

This has a second implication for cost and maintenance. Purpose-built apps require continuous development investment to remain functional across operating system updates, device generations, and platform changes. WhatsApp delivery offloads much of this maintenance cost to Meta. For a small nonprofit operating in resource-constrained conditions, this operational simplification is valuable.

Lesson 3: Cultural Adaptation Is Non-Negotiable, But It Doesn’t Mean Reinventing the Wheel

Step-by-Step was developed by the WHO and has been tested in five different countries. It is not a context-neutral program — it was designed for LMIC contexts — but it still requires meaningful adaptation when deployed in a new setting.

Cultural adaptation is often framed as a choice between fidelity to the evidence-based protocol and responsiveness to local context. In practice, the distinction is not so stark. The core therapeutic content of Step-by-Step — the behavioral activation model, the problem-solving techniques, the cognitive restructuring approach — is not culturally specific. What requires adaptation is the packaging: the language used to describe depression and its treatment, the examples and scenarios used to illustrate concepts, the activities suggested for behavioral activation, and the way counselors are trained to communicate with participants from specific cultural backgrounds.

Kaya’s approach has been to maintain fidelity to the core protocol while investing in genuine local adaptation of the surrounding content. This requires people with both deep local knowledge and sufficient understanding of the therapeutic model to make adaptation decisions that preserve effectiveness. Finding and developing that combination of expertise is itself a significant organizational investment.

The broader lesson for charity entrepreneurs adapting evidence-based programs is: don’t assume that what looks like cultural context is actually relevant to outcomes. Be rigorous about distinguishing adaptations that protect effectiveness (maintaining the core therapeutic model) from adaptations that improve engagement and accessibility (language, examples, communication style). Both matter, but for different reasons, and conflating them can lead to adaptations that undermine the evidence base in the name of responsiveness.

Lesson 4: Theory of Change Clarity Is Operationally Valuable

One of the advantages of being incubated by an evidence-first organization is that Kaya launched with an unusually clear theory of change. The pathway from inputs (counselor time, technology, training) to outputs (completed program sessions) to outcomes (reduced depression symptoms) to impact (DALYs averted, WELLBYs gained) was specified before the first participant enrolled, with evidence-based estimates at each step.

This clarity is operationally valuable in ways that aren’t always obvious. When something goes wrong — completion rates are lower than expected, counselor quality varies, a particular cohort shows weaker outcomes — having a specified theory of change provides a diagnostic framework. Where in the causal chain is the problem? Is it at the output level (participants not completing sessions) or the outcome level (completers not showing symptom improvement)? The answer points to very different operational responses.

Organizations that launch without this clarity often can’t answer these diagnostic questions, which means they can’t improve efficiently. They know their metrics aren’t where they want them to be, but they don’t know where to intervene.

For charity entrepreneurs, the investment in rigorous theory of change development before launch — specifying not just the ultimate goal but each link in the causal chain, with evidence-based estimates of what each link should produce — pays operational dividends that compound over time.

Lesson 5: The First Year Is About Learning, Not Scale

3,600 participants in the first year is a meaningful number. It is not, however, the number that demonstrates Kaya has achieved its scaling mission. That mission is measured in the hundreds of thousands and millions of people who have access to effective depression treatment in settings where no alternative exists.

Framing the first year as primarily a learning exercise — not a failure to scale, but a successful acquisition of the knowledge needed to scale well — is both accurate and strategically important. The learning from year one includes: what participant recruitment channels work in the Indian context; how long counselor training takes to produce reliable quality; what the actual completion rate distribution looks like and what predicts completion; where the technology friction points are; what the per-participant cost actually is relative to projections.

This learning is the essential foundation for year two growth. An organization that tried to scale from launch without accumulating this learning would scale its problems alongside its reach. The discipline to build carefully, measure rigorously, and grow only as fast as the organizational learning can support is a hallmark of well-run NGOs — and one of the lessons that the charity incubation model is particularly well-positioned to instil.

What Comes Next

Ambitious Impact’s research identified ten priority countries for Step-by-Step expansion: Pakistan, China, Nigeria, Bangladesh, Indonesia, Egypt, Brazil, Tajikistan, Ethiopia, and Malaysia. Each represents a combination of high depression burden, sufficient digital infrastructure, and limited existing provision that makes it a strong candidate for Kaya’s model.

Geographic expansion is the primary vector for impact growth over the next several years. Each new country requires the full stack of operational investment — counselor recruitment and training, cultural adaptation, technology localization, health system partnerships, and regulatory compliance. The organizational capacity to manage this complexity while maintaining program quality is the central organizational development challenge Kaya faces.

For the field more broadly, Kaya’s trajectory in its second and third years will be informative for the question that all promising early-stage NGOs eventually face: can the model that worked in the founding context be replicated at scale, across contexts, without sacrificing the quality and fidelity that made it work in the first place? The evidence suggests it can. The execution is what will determine whether that evidence is realized.

To learn more about Kaya Guides, its work, and how to support its expansion, visit besidehealth.org.

Related reading

The Lay Counselor Model: Delivering Evidence-Based Mental Health Care at Scale Without Psychiatrists

The arithmetic of mental health care in low- and middle-income countries is unforgiving. There are an estimated 300 million people worldwide living with depression. In low-income countries, there are approximately 1.1 mental health workers per 100,000 people. Training a psychiatrist takes a decade. A clinical psychologist, five to seven years. Even an entry-level community mental health worker, in the conventional model, requires months of supervised training before they can safely deliver evidence-based psychological interventions.

The math doesn’t work. You cannot close a 75 percent treatment gap using a workforce model that requires years of specialist training per worker, when the countries with the greatest need have the least capacity to produce or retain those specialists.

The lay counselor model is the response to this arithmetic. It doesn’t try to replace specialists. It tries to reconfigure the delivery system so that the components of evidence-based care that don’t require specialist training can be delivered by people who receive focused, structured preparation over weeks rather than years — freeing specialists to focus on the cases that genuinely require their expertise.

This post examines what the lay counselor model is, what the evidence shows about its effectiveness, how it’s operationalized in the WHO Step-by-Step program, and what the implementation challenges are for organizations trying to build at scale.

The Task-Shifting Logic

Task-shifting — the delegation of specific health care tasks from higher-skilled to lower-skilled workers — has been well-established in other areas of global health before it was applied to mental health. Community health workers delivering malaria treatment, vaccination, and basic maternal and child health services have been a feature of primary health care in LMICs for decades. The evidence that appropriately trained and supervised community workers can deliver specific, well-defined health interventions safely and effectively is not in dispute.

The application to mental health has been slower, partly because of the persistent (and not entirely wrong) view that psychological interventions require the kind of nuanced clinical judgment that takes years of training to develop, and partly because the evidence base for specific manualized interventions that could be delivered by non-specialists was limited until relatively recently.

The development of manualized, structured psychological programs — with clear protocols, defined session content, and explicit decision rules for handling common difficulties — has changed this calculus. When the therapeutic content is structured and deliverable via a manual or digital interface, the role of the human support provider shifts: from “clinician delivering therapy” to “supporter facilitating engagement with a structured program.” These are different jobs with different training requirements.

The lay counselor in WHO Step-by-Step is not delivering therapy. They are providing the brief, consistent human contact that research shows makes the difference between guided and unguided self-help outcomes. Their job is to listen, to encourage, to notice if someone seems to be deteriorating and needs referral, and to help participants work through practical barriers to engagement. This job does not require a clinical degree. It requires good communication skills, cultural competence, basic psychoeducation, and the ability to follow a protocol.

What Lay Counselors Actually Do in Step-by-Step

In the WHO Step-by-Step model, lay counselors conduct weekly 15-minute telephone check-in calls with each participant throughout the five-week program. The structure of each call follows a protocol:

Opening and rapport building. A brief, warm opening that establishes continuity from the previous call and creates a safe space for the participant to share how they’re doing.

Progress review. A check-in on how the participant has been engaging with the program since the last call. Have they completed the current session? Have they tried the suggested activities? What did they find helpful, and what was difficult?

Problem-solving. If the participant encountered barriers to engagement — technical difficulties, competing responsibilities, emotional resistance — the counselor helps them think through how to address those barriers before the next session.

Encouragement and normalization. Affirmation of the participant’s efforts, normalization of any difficulties they’ve experienced, and encouragement to continue.

Safety check. A routine, non-alarmist check on whether the participant is experiencing any thoughts of self-harm or significant deterioration. If concerning signals are present, the counselor follows a defined escalation protocol to connect the participant with appropriate support.

Close and next steps. A clear close that sets up the next call and encourages the participant to engage with the next session before then.

Fifteen minutes. Five elements. A protocol that a well-trained lay counselor can follow consistently across hundreds of calls. The simplicity is the point — it’s what makes the model scalable.

The Training and Supervision Infrastructure

The effectiveness of lay counselors depends entirely on the quality of their training and the robustness of the supervision infrastructure that supports them. This is the most commonly underestimated element of lay counselor programs and the most common source of quality degradation as programs scale.

Kaya Guides’ training model for lay counselors covers several domains:

Program knowledge. Counselors learn the content and logic of each of the five Step-by-Step sessions in depth — not just what participants will encounter, but why the program is structured as it is, what the therapeutic rationale for each element is, and what the common questions and difficulties are. This knowledge enables counselors to answer participant questions confidently and to notice when a participant’s response to session content is atypical.

Communication skills. Active listening, empathic responding, open questioning, and motivational interviewing techniques. The ability to create a safe, non-judgmental space in a 15-minute phone call is a learnable skill, but it requires deliberate practice and feedback.

Boundary management. Lay counselors are not therapists, and the program works best when they don’t try to be. Training includes explicit instruction on the limits of the counselor role — what they are there to do (support engagement, provide encouragement, monitor wellbeing) and what they are not there to do (provide clinical assessment, offer advice outside the program content, maintain contact outside the scheduled call structure).

Safety protocols. Recognition of warning signs — suicidal ideation, psychotic symptoms, severe functional deterioration — and clear escalation pathways. This is non-negotiable and must be trained to a level of reliability that doesn’t depend on counselor judgment in the moment.

Self-care and secondary traumatic stress. Working with people in depression, even briefly and non-clinically, carries emotional costs. Counselors who receive 400 calls per year from people describing their struggles are at risk of cumulative emotional burden if they don’t have tools for managing it. Training in self-care and supervision structures that include space for counselors to process their own experience are essential for workforce sustainability.

Ongoing supervision — typically weekly group supervision sessions facilitated by a more senior staff member — provides a space for case review, quality monitoring, skill development, and counselor support. The quality of supervision is as important as the quality of initial training; it is the mechanism by which training translates into sustained performance.

One Counselor, 400 Participants: The Scale Mathematics

A single lay counselor working full-time and conducting 15-minute calls can support approximately 400 enrolled participants per year. The calculation is straightforward: five calls per participant (one per session per week), 15 minutes each, equals 75 minutes of direct counselor time per participant. At 400 participants, that’s 500 hours of direct call time per year — approximately 10 hours per week, leaving time for preparation, supervision, documentation, and administrative tasks within a standard working week.

At Kaya’s projected scale of $1.3 million per year total program cost, and with one counselor per 400 participants, counselor compensation is a significant but not dominant cost driver. The remainder covers technology, training, supervision, program management, and organizational overhead.

The implication for program economics is important: the per-participant cost of human support in this model is low — a few dollars per participant for the full five-week program — and the counselor ratio means that adding participants has a predictable and manageable cost structure. Unlike specialist-dependent models, where adding capacity requires adding expensive professionals, the lay counselor model scales linearly with a lower-cost input.

This arithmetic is what makes the $97/DALY cost-effectiveness figure achievable. Remove the lay counselor model and replace it with therapist-delivered support, and the cost per participant increases by an order of magnitude. Remove the human support entirely and you lose the guided/unguided effect size differential. The lay counselor is not a cost-cutting compromise — it is the design element that makes the economics and the outcomes simultaneously achievable.

Quality at Scale: The Central Challenge

The lay counselor model works at small scale when supervision is intensive and close. The challenge at large scale is maintaining quality when the supervisor-to-counselor ratio becomes less favorable and when counselors are geographically distributed across multiple contexts.

Quality degradation in lay counselor programs typically follows predictable patterns. Protocol drift — gradual deviation from the specified call structure, often in the direction of counselors providing advice or support that goes beyond their defined role — is common and can be both ineffective and harmful. Compassion fatigue, if not addressed in supervision, can produce counselors who are going through the motions rather than genuinely engaging. And in contexts with high counselor turnover, the organizational knowledge embedded in experienced counselors is lost and must be rebuilt, creating quality troughs during onboarding periods.

Technology-assisted quality monitoring offers partial solutions. Call recording (with participant consent) enables supervisors to review counselor performance and identify drift before it becomes entrenched. Structured documentation of call content enables pattern analysis across counselors and cohorts. Automated flags for unusual patterns — calls that are significantly shorter than the protocol specifies, participants who are not completing sessions but haven’t triggered a safety concern — can direct supervisor attention efficiently.

But technology doesn’t replace the relational dimension of supervision. Counselors need to feel supported, not surveilled. The supervision relationship is itself a model for the counselor-participant relationship — a safe, non-judgmental space for reviewing performance, addressing difficulties, and continuing to develop. Organizations that manage quality through monitoring alone, without investing in the relational quality of the supervision relationship, tend to see high counselor turnover and the quality problems that come with it.

Referral Pathways: The Missing Infrastructure

The lay counselor model is designed for mild-to-moderate depression — the large majority of people who would benefit from psychological support. It is explicitly not designed for severe depression, active suicidality, psychotic features, or significant psychiatric comorbidity. These presentations require specialist care.

The safety protocols built into Step-by-Step counselor training address the identification side of this: counselors are trained to recognize warning signs and escalate. What is often missing — and what represents a significant gap in lay counselor program design — is the referral destination. Where does a counselor escalate to when they identify someone who needs more intensive support?

In high-income countries, this question has a relatively clear answer: the GP, the emergency department, the crisis line, the community mental health team. In many LMIC contexts, the answer is far less clear. Specialist mental health services may be distant, expensive, or simply non-existent. Community hospitals may not have psychiatric beds. Crisis lines may not be available in the participant’s language.

Building referral pathways is not primarily a clinical challenge — it’s a systems and partnership challenge. It requires establishing relationships with whatever higher-level services exist in a given context, developing clear protocols for warm handoffs, and being honest with participants about what is and isn’t available. In some contexts, it may require advocacy for the development of referral infrastructure that doesn’t yet exist.

Organizations scaling lay counselor programs must invest in referral pathway development as a first-order priority, not an afterthought. The safety of participants depends on it, and the credibility of the program depends on its ability to demonstrate that it takes safety seriously.

The Evidence Summary

The evidence for the lay counselor model in guided digital self-help for depression is, by the standards of global mental health, strong. Five RCTs of WHO Step-by-Step — each of which used lay counselors rather than professional therapists as the human support component — produced pooled effect sizes of 0.78, comparable to face-to-face therapy. The human support was provided by community members with weeks of training, not years.

This is not a compromise. It is a demonstration that the therapeutic effect of guided self-help programs derives primarily from the structured content and from the presence and consistency of human contact, not from the clinical qualifications of the person providing that contact. Within the specific, well-defined role that lay counselors are trained to fill, clinical training is not a prerequisite for effectiveness.

What the evidence also shows — consistently — is that the absence of human support produces substantially worse outcomes. The lay counselor is not a luxury or an enhancement. They are the mechanism by which digital content produces guided rather than unguided self-help effects. Any organization attempting to scale digital mental health for depression in LMICs without this component is working against the evidence.

To learn more about the Kaya Guides model and how lay counselors are scaling mental health care globally, visit besidehealth.org.

Related reading