Abstract
Background
Mental, neurological, and substance use disorders are a leading cause of disability globally. In sub-Saharan Africa (SSA), underinvestment, centralised services, stigma, and workforce shortages create large treatment gaps. Digital delivery can expand access, yet sustainable models remain concentrated in high-income settings. University-industry (U-I) collaboration may bridge capability and scaling gaps, though differing cultures and timelines often impede progress. This study documents a hybrid U-I approach across design, development, and early evaluation of a digital mental health intervention in SSA, and extracts practical lessons.
Methods
We report a retrospective, multi-country case study of Inuka Coaching, a coach-guided, chat-based problem-solving therapy delivered via mobile/web. Four sequential phases integrated academic and pragmatic processes: (I) needs assessment using literature review, 49 interviews, and 7 focus groups across Kenya, Ghana, and Nigeria; (II) a 20-stakeholder Nairobi hackathon co-creating a minimum viable product (MVP), followed by assumption testing; (III) feasibility and acceptability testing with 13 clients and 5 coaches in Zimbabwe; and (IV) a pilot across 18 Nairobi primary care centres enrolling 80 adults, measuring Self-Reporting Questionnaire-20 (SRQ-20), Patient Health Questionnaire-9 (PHQ-9), and Generalized Anxiety Disorder scale-7 (GAD-7) at baseline, 4 weeks, and 3 months.
Results
Phase 1 produced structured design inputs for non-academic stakeholders, synthesized from the needs-assessment activities. Phase 2 delivered an MVP and a preliminary route-to-scale, shaped through iterative testing with stakeholders. Phase 3 produced a refined working prototype through usability testing. Phase 4 delivered a pilot evaluation protocol and dataset; the resulting clinical and feasibility findings are reported separately in an ethically approved publication. This case study focuses on the U-I collaboration process itself and does not present or discuss the results of the underlying research activities. Partnership dynamics were phase-aligned: industry led rapid iteration and productisation (Phases 2–3); academia led validity, measurement, and reporting (Phases 1 and 4). Key challenges included limited end-user involvement due to ethics timelines, divergent publishing and product schedules, and data-access restrictions.
Conclusions
This is the first documented hybrid U-I partnership for developing and evaluating a digital mental health intervention in SSA. Clear objectives, flexible phase-specific leadership, early assumption testing, proactive ethics planning, and dual dissemination can reduce research waste and support iterative development. Future collaborations may benefit from living-lab arrangements embedding end-users throughout, cross-disciplinary governance to bridge work cultures, and decision frameworks to manage cultural differences in low-resource settings.
Keywords: Digital mental health, university-industry collaboration, sub-Saharan Africa (SSA), mobile health, cross-sector collaboration
Highlight box.
Key findings
• This study documents the first hybrid university-industry (U-I) collaboration developing and evaluating a digital mental health intervention (Inuka Coaching) in sub-Saharan Africa (SSA).
• Effective collaboration required negotiating competing academic and industry values, with leadership strategically allocated across phases to leverage each partner’s strengths. Academic-led phases ensured rigour and a solid project foundation, while industry-led phases facilitated responsive product development.
What is known and what is new?
• Digital mental health interventions can improve access and outcomes, but most evidence originates from high-income settings and university-led development and evaluation. Few examples exist from SSA, and to our knowledge, none have documented a U-I collaboration.
• This manuscript provides the first empirical documentation of a hybrid U-I model in SSA, identifying enabling conditions, phase-specific leadership and integrated methods, through which academic and industry partners can bridge research and product development, and work towards scalability.
What is the implication, and what should change now?
• Policymakers, funders, and institutions should support cross-sector governance structures to enable scalable, evidence-based digital mental health solutions in low-resource settings.
• Hybrid U-I models should be adopted more widely, with this paper offering a replicable framework for effective collaboration that can be adapted and refined across contexts, especially low-and-middle-income contexts.
• Greater end-user involvement, robust ethics planning, and dual dissemination strategies targeting academic and industry audiences are needed to accelerate implementation and reduce research waste.
Introduction
Background
Mental, neurological, and substance use disorders contribute substantially to the global burden of disease, with depression and anxiety ranking among the leading causes of disability (1,2). In sub-Saharan Africa (SSA), the prevalence of common mental disorders (CMDs) is comparable to global averages, yet health systems face severe constraints in detection and care (3). Despite international recognition of mental health as a public health priority, SSA countries allocate less than 1% of their health budgets to mental health on average, and service delivery is often centralised in a few psychiatric hospitals (4,5). This underinvestment may lead to untreated mental disorders, which may result in an increased risk of suicide, chronic physical disease, and reduced productivity (2,6)
Treatment gaps and barriers
The treatment gap in SSA is among the largest worldwide, with up to 90% of individuals with CMDs and severe mental illness (SMI) not receiving treatment (7,8).
SMI patients receive priority within the limited system (services concentrated on tertiary/hospital-based care for severe cases) (8), while CMDs such as depression, anxiety are largely left to primary care or go unscreened entirely, given they’re rarely framed as emergencies (9). Several structural barriers drive this gap. Workforce shortages are severe, with many countries having fewer than one psychiatrist per million people (3), often heavily relying on psychiatric or general nurses to run mental health clinics, and who are also significantly understaffed (10). Existing specialists are concentrated in urban centers, leaving rural populations underserved. Mental health services are rarely covered by insurance, and out-of-pocket costs remain prohibitive (6). Stigma and cultural perceptions further deter individuals from seeking help, due to the fear of discrimination (11). The World Health Organisation projected that low-and middle-income countries (LMICs) require an estimated 6.4 billion US dollars to bridge the gap in human resources (4). Community-based models, such as task sharing to lay health workers, have shown promise in addressing these barriers (12,13). The Friendship Bench in Zimbabwe, where trained lay “grandmothers” deliver face-to-face problem-solving therapy (PST) in primary care and communities, demonstrated substantial reductions in depression and anxiety (14). However, implementing and scaling is resource-intensive, requires consistent supervision, and may not address barriers related to stigma or geographic access (15).
Promise of digital health
Digital technology offers opportunities to overcome structural barriers in mental health service delivery. Mobile phone ownership in SSA has expanded dramatically, with more than 80% of households now owning at least one mobile phone, and smartphone penetration increasing annually (16,17). This infrastructure makes digital interventions delivered through apps, short message service (SMS), or chat platforms potentially scalable at low cost (18). Digital mental health interventions can increase privacy and reduce stigma by allowing individuals to access care discreetly (19). Digital technologies can also extend reach to underserved populations, enable both synchronous and asynchronous interaction, and provide decision support for non-specialist providers (18,20). Evidence from high-income countries indicates that psychological treatment delivered by non-specialist or primary care providers achieves outcomes comparable to specialist outpatient care for adult depression, with no significant difference in response rates once setting-related factors are accounted for (18,21-24). Guided digital interventions similarly show effect sizes comparable to face-to-face formats for depression and anxiety, both in controlled trials (23) and in routine clinical practice (pooled g =1.18 for depression, g =0.94 for anxiety) (22). Guided self-help achieves efficacy similar to individual cognitive behavioural therapy (CBT), though with somewhat lower acceptability and higher dropout (21,24). However, this evidence derives almost entirely from high-income-country samples with reliable internet access and trained clinical or research-grade guidance. Its generalisability to SSA, where task-sharing typically relies on community health workers, intermittent connectivity, and resource-constrained infrastructure, has not yet been established. Most digital interventions are tested in high-income contexts, and evidence from SSA is sparse (25). Moreover, digital tools developed in LMICs are often donor-funded pilots that fail to transition to long-term, sustainable service models (26).
University-Industry (U-I) Collaborations
Hybrid collaborations between universities and industry have been proposed as a mechanism to accelerate digital product development, often referred to as U-I collaborations (27). There are two broad perspectives to approach U-I knowledge transfer in research (27). The first is ‘untargeted knowledge transfer’, which refers to a one-way knowledge flow between university and industry. The sources of information can be accessed through publications and conference proceedings. Knowledge derived from an academic institution is then applied to an unidentified target. The second approach is ‘targeted knowledge transfer’ between a university and one or more private partners, which refers to when the knowledge interaction gives the private partner the opportunity to access some level of knowledge appropriation (i.e., the direct capture and application of academically generated knowledge within a commercial product) (27). The current paper will focus on ‘targeted knowledge transfer’ involving a mixture of commercialisation and an intermediate approach to U-I collaboration (27,28).
The U-I partnership has been referred to as the cornerstone of innovation, in which transdisciplinary collaboration is seen as an essential component of innovation (29-31). While industry can contribute competences that can sustain, develop and scale services to reach a wider audience, academic research enables deeper insights in relation to user-need, and provides rigorous evaluation to ensure that interventions are acceptable, safe, and effective (32). This is important, especially considering that despite there being over 10,000 mental health platforms available to users, very few have undergone an external quality review or adequate evaluation (33,34).
University and industry collaboration challenges
Institutions in academic and commercial spheres have different cultures, values and goals that facilitate the production of different outputs (35). Critical differences in organisational drivers, leads to differences in the way that knowledge is generated (36). First, universities and industry differ in relation to the time-period perceived as appropriate for meeting a given goal (36,37). Universities adopt a long-term perspective to enable scientific depth and rigor when undertaking research. The primary objective of the academic researcher is to produce impactful scientific results that can be widely disseminated. On the other hand, industry, is more likely to focus on short-term knowledge generation processes, to aid rapid implementation of the results across many iterative cycles of product development (31,32,38).
Time frame differences can be particularly problematic when developing digital products. In the past decade, technology has been growing at an exponential speed, which has largely been consistent with Moore’s Law, a model that posits that the information technology industry doubles in performance and functionality of digital electronics roughly every 2 years within a fixed cost (39). In academia, randomised controlled trials (RCTs) are the gold standard for evaluating healthcare interventions (40). However, on average, clinical trials can take up to 6–7 years to complete, or longer when developing new products or medicines (41). While RCTs are seldom used, pre–post design also requires a long duration, making the evaluation of continuous solution upgrades (e.g., after integrating technological updates, fixing bugs, etc.), commonly experienced in digital health products, highly challenging. It’s important to note that research designs that produce the highest level of evidence, do not always produce evidence that is transferable to practice settings (42,43).
There are methodological gaps in relation to evidence-supported agile approaches (32,44). In Figure 1, Guo and colleagues (32) map existing research approaches for ‘health digital solution evaluations’, as well as current methodological gaps, and found that pragmatic approaches may fill the gap for faster high quality evidence generation.
Figure 1.

Existing approaches for health digital solution evaluation, current methodological gap and emerging innovative pragmatic approaches to fill such gap, from Guo and colleagues (32). RCT, randomized controlled trial.
Evolving practices in addressing mental health using digital technologies, appears to warrant new approaches to prevent time-lags between research and implementation, and research waste, in which platforms developed and evaluated cannot be sustained (45). While there is hesitancy from academic universities to collaborate with industry due to different cultures, goals, and often due to ethical reasons and ability to conduct impartial research (46-48), there has been a recent shift in conversations, in which building environments that drive progress in digital mental health are increasingly becoming acknowledged by academics and funders.
This shift is also visible at a global governance level. World Health Organization (WHO)’s Global Initiative on Digital Health (GIDH) (49), launched in 2024 to accelerate the Global Strategy on Digital Health 2020–2025 (since extended to 2027) (50), brings governments, academic institutions, and industry partners together under a single coordinating structure aimed at aligning country level resources and knowledge exchange for digital health transformation, including mental health. Its objectives include assessing country needs, reducing fragmentation, and building local capacity for digital health technologies, which reflects the same rationale for cross sector collaboration examined in this paper, but applied at a multilateral scale rather than within a single partnership. Despite this, documented examples of how individual U-I collaborations function within, or contribute to, this wider effort remain scarce. GIDH illustrates the scale of coordinated global effort currently underway in this space, but offers little insight into the operational realities, governance arrangements, review timelines, and IP negotiations, that determine whether individual partnerships succeed (49,50). A Wellcome Trust commissioned project recently consulted stakeholders working in the field to uncover challenges and solutions to sustaining digital mental health solutions for therapeutic tools that improve public mental health (45). The report identified a number of barriers to creating environments that sustain, grow and scale digital programmes, due to gaps in academic resources and funding structures. Feedback suggested that pathways to market required a high level of expertise and responsiveness, owing to emerging technological advancements, economic conditions, regulatory processes and unpredictable markets, which academics are not well equipped to navigate. Moreover, academic teams do not usually have the right skills to support the development and commercialisation of digital mental health solutions. Often a wide range of expertise is needed to develop, maintain and scale digital solutions, such as software development, design, project management, data science, information governance, regulation, and implementation, which cannot be easily accessed in the academic space (45). In response to these limitations, U-I partnerships were proposed as a viable approach to bridging the skills gap in academia. However, a notable limitation related to this approach was the lack of successful examples that can be drawn on. This report therefore addresses calls by industry experts to demonstrate U-I collaborations (45).
Rationale and knowledge gap
Documentation of such collaborations in LMICs is limited, particularly in mental health. Most published examples come from high-income countries and focus on usability testing rather than cross-sector collaborations. There is therefore a critical gap in understanding how hybrid approaches function in resource-constrained contexts. Our study aims to contribute to understanding the processes and negotiations made by U-I partnerships to generate evidence. The rest of the paper will include a retrospective assessment of the U-I contribution to knowledge in the design, development, testing and evaluation of a coach-guided digital PST intervention for common mental health disorders (called Inuka Coaching) in the SSA region.
Despite growing interest in digital mental health partnerships across SSA and other low- and middle-income settings, comparatively few published studies document the U-I collaboration process itself, as distinct from the resulting intervention’s outcomes. PracticePal, a chatbot developed by Sangath and Dimagi to support delivery of a brief behavioural activation treatment for depression via community health workers in rural India, has recently been published as a development and pilot-testing study (51), but like much of the wider digital mental health literature, its contribution is intervention-focused rather than a documented account of how the academic-industry partnership itself was structured or sustained. U-I collaboration is a well-established practice in other sectors, including pharmaceuticals, biotechnology, and engineering, where formal partnership frameworks and governance benchmarks have developed over several decades (52). In global and mental health specifically, these remain comparatively novel arrangements: the underlying concepts are not new, but their application to this field is. This is important because collaborating with industry partners in health research is well documented to raise concerns about conflicts of interest that can affect the real or perceived objectivity of findings (53,54), yet there is often no established infrastructure to build and govern such relationships in global and mental health, nor shared knowledge of how these collaborations are likely to progress (52). Documenting our case study, including its challenges and how they were navigated, offers an opportunity for the field to learn from real experience and to anticipate similar issues in future partnerships, rather than each new collaboration starting from first principles.
To our knowledge, this paper is the first to report the use of a hybrid approach in developing a mental health intervention in the SSA region. By tracing four phases; formative needs assessment, development, feasibility pilot (prototype testing), and pilot delivery (pilot research study) (55), the study illustrates both the tensions and synergies of hybrid collaboration. In doing so, it contributes to the growing discourse on how to harness U-I collaboration and advance digital mental health in low-resource settings.
Objectives
To document a hybrid U-I approach across design, development, testing and early evaluation of a digital mental health intervention in SSA, and to extract practical lessons.
Methods
Procedure
This paper details four case studies retrospectively describing the process of generating knowledge through a hybrid U-I collaboration to develop a mobile app called Inuka Coaching. The Inuka intervention was developed across four phases of research (56), that include: (I) the identification and analysis of participant needs; (II) the generation of ideas and development of prototypes; (III) the testing and further development of prototypes; and (IV) a pilot evaluation of intervention, aligning with the WHO’s Monitoring and Evaluating Digital Health Interventions guide (55). Within each phase, we will document academic and industry research processes employed across features pertaining to the full research cycle (57) (see Figure 2 for a diagrammatic presentation of the documentation plan). This paper focuses on the process of U-I collaboration itself, rather than on the outcomes of the underlying research activities; it therefore does not present or discuss the specific findings of Phases 1–3.
Figure 2.

WHO monitoring and evaluation stage of digital intervention maturity (55) mapped to the research-cycle features tracked at each phase and the relative academic- versus industry-led emphasis (bars show the authors’ estimate of relative emphasis). WHO, World Health Organization.
Context
Inuka, which means (“arise” in Swahili) was conceived as an adaptation of Zimbabwe’s Friendship Bench (14), which showed efficacy in reducing depression and anxiety through lay health workers delivering PST in primary care. PST is a brief, structured psychological approach that helps clients identify a current problem, generate and weigh possible solutions, and implement an action plan. The initial collaboration that led to the development of Inuka was between Philips, a Dutch multinational health technology company (across different departments, including Philips design, research and the Africa incubator) and the University of Zimbabwe. Co-author R.v.D. led this project on behalf of Philips, while co-author D.C., Founder of the Friendship Bench, and Co-Founder of Inuka acted on behalf of University of Zimbabwe. The evaluation of the intervention was funded by Grand Challenges Canada.
In its present form, Inuka Coaching is available via a mobile app and can also be accessed using a web-platform. The service provides up to six online sessions lasting between 40 minutes (for the first session) and 20–30 minutes (for follow-up sessions). The key defining features of the Inuka coaching include: (I) a service-user coach matching service; (II) a self-assessment based on the validated Self Reporting Questionnaire-20 (SRQ-20) (58); (III) a coach facilitated PST intervention, (IV) referral pathways to clients with severe presentations and; (V) the promotion of psychological self-care through a self-assessment tool. The intervention also has quality assurance systems supporting the coach and supervisors to ensure quality support is provided (59). Table 1 provides further details on the principles of the intervention. Inuka Coaching is now independent of Philips and is a social enterprise owned by co-author R.v.D. (inuka.io). The desired outcome of the intervention is to prevent mental ill health through providing therapeutic effects, but it is not a medical device (59).
Table 1. A summary of the six principles of the Inuka Coaching intervention.
| Principle | Description |
|---|---|
| Matching | Person seeking psychological support (client) is matched with a service provider |
| Screening & support | The client is screened for common mental disorders using the SRQ-20 and is engaged in a structured conversation based on PST via a text-based chat with the service provider. PST is a brief, structured psychological therapy that helps clients address the specific life problems contributing to their distress (60). It follows a stepwise process: (I) identifying and clarifying the problem; (II) generating possible solutions; (III) weighing the pros and cons of each option; (IV) selecting and implementing a solution; and (V) reviewing the outcome |
| Decision support for Community Health Volunteers (see note for further details) | The service provider is provided with a set of structured scripts on what to say in each step and how to build rapport via chat. The application also requires the service provider to fill out an action plan in which the challenges of the client and the next steps are summarised |
| Referral | If the service provided is deemed insufficient by the client, or if the service provider deems the client’s presentation to be severe, the client can be referred to mental health professionals for diagnosis and treatment, with their consent |
| Promotion of psychological (self-)care | The Inuka app supplies information and self-assessment that is contextualised for local settings to create awareness and lower the barrier for seeking help |
| Quality assurance (QA) | Coaches and supervisors are supported by structured quality assurance systems, including supervision and review processes, to help ensure that the support provided to clients through the app meets required quality standards |
Note: Community Health Volunteers are lay people who are trained on the Inuka approach. CHV’s are provided with a set of structured scripts on what to say in each step and how to build rapport via chat. The application also requires the service provider to fill out an action plan in which the challenges of the client and the next steps are summarised. CHV, community health volunteer; PST, problem-solving therapy; SRQ-20, Self-Reporting Questionnaire-20.
Ethical review status
This retrospective case study draws on interviews, focus groups, and usability/feasibility testing conducted between 2015 and 2018 as part of intervention design, development, and evaluation activities. At the time, these activities were treated as product development and quality-improvement work rather than formal human-subjects research requiring ethics committee review. Despite a thorough search, the authors have not been able to locate institutional ethics or exemption documentation for Phases 1–3 from this period; we acknowledge this as a limitation of the present report. In light of this, the case study focuses on describing the U-I collaboration process itself and does not present or discuss the specific findings of Phases 1–3. For Phases 1 and 3, written informed consent was obtained from all participants before interviews, focus group discussions and usability testing, after the purpose of the activity and the voluntary nature of participation had been explained. No names were recorded in the data; participants were assigned codes, and quotations and personas reported in this paper are anonymised or composite. The Phase 4 pilot evaluation was approved by the Amref Ethics and Scientific Review Committee (KIAMBU/HRDU/AUTHO/2017/04/10/Njenga A, on 10th April 2017) and informed consent was taken from all individual participants before taking part. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
Statistical analysis
No statistical analyses were performed for this case study, which describes the collaboration process narratively. Phase 3 data were summarised descriptively (counts and frequencies) owing to the small sample size. Statistical analyses of the Phase 4 pilot are reported in the associated publication (59).
Results
Case study phases: hybrid academic-industry knowledge generation
The findings below describe the academic and industry research processes engaged in across the four phases, disaggregated by research-cycle feature (Tables 2,3) and mapped onto the WHO’s stages of intervention maturity, formative/needs assessment, development, feasibility pilot, and pilot delivery (Figure 2) (55). This process monitoring is distinct from the outcome evaluation reported separately in (59).
Table 2. Case study comparison of pragmatic and academic research, across phases of development and research-cycle feature.
| Research-cycle feature | Phase 1: identification of user needs | Phase 2: design and development | Phase 3: testing of the prototype | Phase 4: evaluating the intervention |
|---|---|---|---|---|
| WHO M&E stage of intervention maturity (55) | Formative/needs assessment | Development | Feasibility pilot | Pilot delivery |
| Objectives | ||||
| Academic | Generate evidence on mental health needs in the region | n/a | Generate evidence on the feasibility, acceptability, and delivery of the intervention | Generate evidence on the clinical effectiveness of Inuka, and understand the mechanisms underlying observed outcomes (process evaluation) |
| Pragmatic | Identify target audience | Develop a minimum viable product | Gain feedback on the experience and usability of the digital tool and intervention | Understand whether Inuka is feasible, acceptable, and reduces participants’ mental health complaints |
| Methods used (academic and pragmatic) | Literature review; qualitative individual interviews; focus group discussions | Direct observation (n=17); think-aloud testing (n=17); near-live testing (n=17); live testing (n=17) | Live testing (n=18); psychometric questionnaires (n=18); individual interviews (n=18) | Pilot cohort study design; recruitment of participants with common mental disorders; systematic delivery of intervention; pre-, post-, and 3-month follow-up assessments |
| Termination criteria (academic and pragmatic) | Saturation of data | When consensus was reached | Testing objectives satisfied | Pragmatic pilot sample; no formal sample size calculation (59) |
| Data analysis | ||||
| Academic | Qualitative thematic analysis by two independent coders | Academic review and interpretation of pragmatic analysis findings (no independent academic analysis conducted) | Academic review of statistical summaries and independent interpretation of findings to inform Phase 4 design (no independent academic analysis conducted) | Quantitative statistical data analysis |
| Pragmatic | n/a | Descriptive summaries; thematic analysis of observational field notes | Statistical summaries | n/a |
| Output | ||||
| Academic | n/a | n/a | Detailed data tables and results reporting | Detailed reporting of the project’s rationale, methodology, findings, and implications (discussion). |
| Pragmatic | Brief report with key findings from literature review and qualitative interviews | Summary reports from field notes | Summary reports from field notes | Summary of findings |
| Dissemination | ||||
| Academic | Brief report | Brief report | Brief report | Peer-reviewed journal |
| Pragmatic | Summary report for research team | PowerPoint slides | Summary report for research team | Summary of findings for research team and stakeholders |
| Time frame | ||||
| Academic & pragmatic (combined) | 5 months | 11 months | 6 months | 12 months |
The time frame reported per phase is combined across academic and pragmatic activities and does not separately capture the differing paces at which each typically proceeded within a phase (e.g., academic activities such as ethics review, coding, and formal write-up commonly extended beyond the pragmatic/product development timeline, reflecting differing institutional norms of rigor and speed). This table’s conceptualisation of academic and industry functions, and the distinction between outcome and process evaluation, draws on the UK Medical Research Council/National Institute for Health Research framework for developing and evaluating complex interventions (61) and related guidance on process evaluation of complex interventions (62). M&E, Monitoring and Evaluation; n/a, not applicable; WHO, World Health Organization.
Table 3. Participants, sites, and funding across each phase of the project.
| Activity | Sites | Participants | Funder |
|---|---|---|---|
| Phase 1: identification of user needs (n=49 interviews; n=45 focus group participants) | |||
| Individual semi-structured interviews | Kenya (n=22); Nigeria (n=15); Ghana (n=12) | Experts (n=6); practitioners (n=16); individuals with MH challenges (n=17); individuals without MH challenges (n=5); family members (n=5) | Philips Africa |
| Focus group discussion A | Nairobi, Kenya (n=7); Lagos, Nigeria (n=8) | In-patient and out-patient clients of a private psychiatric hospital (n=15) | Philips Africa |
| Focus group discussion B | Nairobi, Kenya (n=6); Lagos, Nigeria (n=6) | Participants without self-reported mental health challenges (n=12) | Philips Africa |
| Focus group discussion C | Nairobi, Kenya (n=6); Accra, Ghana (n=6); Lagos, Nigeria (n=6) | Mental health practitioners (n=6); staff of a private healthcare institution (n=12) | Philips Africa |
| Phase 2: design and development of prototype | |||
| Multi-stakeholder hackathon | Kenya (n=20) | Participants: web/app design (n=3); technological development (n=3); researchers (n=2); content experts (n=5); business stakeholders (n=5); patient and public involvement (n=2) | Philips Africa |
| Phase 3: testing of the prototype | |||
| Testing of the prototype | Zimbabwe | End-users (n=13); coaches (n=5) | Friendship Bench, Zimbabwe |
| Phase 4: evaluation of Inuka | |||
| Evaluation of Inuka | Kenya | Research participants (n=80) | Grand Challenges Canada |
Note: Individuals without self-reported mental health challenges were included as citizens of the study countries, so that we could explore differences in mental health needs between those with lived experience of mental health challenges and service use and those without. This allowed us to better understand the needs of the population we anticipated would be the end-users of a preventive intervention. “Private institution” refers to a private healthcare institution providing mental health services, distinct from public-sector provision. MH, mental health.
Table 3 summarises the research activities, settings, funding, and participants for each phase. As shown there, “private institution” refers to a private healthcare institution providing mental health services, distinct from public-sector provision; staff from this institution were included to capture perspectives from both delivery contexts. Phases were implemented across different countries because funding could not be secured from a single source, so site selection was driven largely by funding availability rather than by a priori design.
Throughout, “academic” and “pragmatic” refer to institutional norms of rigor and speed respectively, rather than distinct research disciplines. The account that follows is not intended to showcase novel methods, the individual activities described are established practice, but to document, to our knowledge for the first time in digital mental health or U-I partnerships in SSA, how such a partnership was actually conducted, providing context for the challenges and lessons discussed later.
Phase 1: formative needs assessment
The first phase focused on understanding the mental health needs of people living in low- and middle-income urban settings in SSA. Urban areas were selected primarily due to more consistent access to smartphones and reliable connectivity, which were necessary given the technology-based nature of the intervention. As a result, findings may not fully generalize to rural sub-Saharan African contexts, where infrastructure, access, and the broader service user journey may differ substantially. Recruitment bias was safeguarded against through independent review of the sampling strategy and eligibility. Philips, together with local research agencies in Nairobi, led the work using both academic and industry (pragmatic) approaches. The academic objective was to generate qualitative evidence on the nature and scope of mental health needs, while the pragmatic objective was to identify the target population for a digital mental health application. To achieve this, the team conducted a literature review and used its findings to shape semi-structured interviews and focus group discussions with a broad range of stakeholders, including service-users, informal mental health providers, and public and private care professionals. In total, 49 individual interviews and seven focus groups were carried out across Nairobi (in Kenya), Accra (Ghana), and Lagos (Nigeria). Participants were purposively sampled to capture a broad range of perspectives across service-user, informal-provider, and professional stakeholder groups. Interview and focus-group data were analysed thematically using a line-by-line coding approach, conducted independently by two coders to enhance reliability.
These activities highlighted the types of support required by people in urban Africa and helped map the service-user journey. The academic processes ensured a comprehensive understanding of mental health needs, while the pragmatic orientation focused on presenting the findings in an accessible format. Results were distilled into bullet point summaries, infographics, and visual reports that communicated key recommendations to primarily non-academic stakeholders. Outputs were packaged as bullet-point summaries, infographics, and visual reports for internal use in shaping the subsequent design phase. This case study does not present or discuss the specific findings of these activities, consistent with the scope described in section “Procedure”. Figure 3 provides an illustrative overview of the Phase 1 activities described above, together with an example of how discovery themes from stakeholder interviews and focus groups were translated into design implication (Figure 3).
Figure 3.

Overview of Phase 1 (formative needs-assessment) activities, alongside an illustrative extract of discovery themes and design implications from stakeholder interviews and focus groups, and an example participant quote. This case study does not present or discuss the full findings of Phase 1 (section “Procedure”); the material shown is illustrative only. SSA, sub-Saharan Africa.
By combining academic depth with pragmatic dissemination, Phase 1 not only clarified the psychological support needs of target populations but also translated these insights into actionable design inputs for the development of the Inuka digital intervention.
Phase 2: development
Phase 2 focused on transforming the insights from Phase 1 into a working digital intervention. The central objective was to explore whether a mobile solution could feasibly enhance or even replace traditional face-to-face psychological support, and to produce a minimum viable product (MVP) that could be tested and refined. This phase was strongly shaped by pragmatic research needs, but academic partners were integrated to ensure the research methodology would enable validity.
The process began with the Hacking Mental Health hackathon, held on 10 November 2015 at iHub, an innovation hub and hacker space for Nairobi’s technology community. The event brought together 20 stakeholders, including technologists, software developers, Philips design and research teams, academic partners, content experts, service-users, and business representatives. The hackathon aimed to collaboratively design a prototype of a digital self-help application inspired by the Friendship Bench (14) model of PST. This venture enlisted local organisations in the region (e.g., Dr Monk in Kenya, Kenya Counselling and psychological association, iHub Nairobi, Philips Research Kenya, University of Zimbabwe in Zimbabwe, University of Nairobi in Kenya, etc.). Participants were purposively selected and invited by the organising team based on their relevant expertise and stakeholder role, drawing on the local organisations listed above, rather than through an open call.
Over the course of the event, hackathon participants worked collaboratively to immerse themselves in the context of mental health needs identified in Phase 1. Activities included mapping the user-experience journey of a digital intervention, creating storyboards, sketching wireframes, and building a mock-up prototype, which aligned with the industry approach of conducting research. In addition to technical design work, the team also developed a preliminary business case, which considered potential scalability and sustainability of the intervention in low-resource settings.
Following the hackathon, assumption testing was undertaken to examine the feasibility and acceptability of the prototype. This included A/B landing page experiments designed to test user preferences, evaluate different design hypotheses, and gauge early interest in the service. Philips design teams applied rapid co-creation techniques, proprietary design processes (defined as a unique method, system, or manufacturing technique owned exclusively by a single company or individual, which others cannot legally use, copy, or sell without permission), and design thinking to accelerate iteration cycles, while academic partners advised on study design and data collection to enhance the robustness of the testing. This process reflected cross-sector collaboration, defined here as structured, collaborative engagement between academic, industry, and community/end-user stakeholders in the design and testing of an intervention, as distinct from formal co-design, in which end-users act as design partners throughout the process. Beyond A/B testing, Phase 2 also progressively tested the prototype through direct observation (researchers observing participants using the prototype without intervention), think-aloud testing (participants verbalising their thoughts while completing tasks), near-live testing (trials under conditions closely approximating real-world use but with researcher support), and live testing (test with the prototype used independently in real-world conditions), as summarised in Table 2. Figure 4 provides an illustrative overview of the Phase 2 hackathon and assumption-testing activities described above, including the composite personas used during the flow-experience mapping procedure.
Figure 4.

Overview of Phase 2 (development) activities, including the Hacking Mental Health hackathon and assumption-testing methods, illustrated with two anonymised composite personas used during the flow-experience mapping procedure. Prototype iterations from this phase are shown in Figure 5.
Outputs from Phase 2 included a mock-up prototype of the Inuka application, an initial set of user-journey maps, and a draft business case for scaling. Key insights from the hackathon and testing were documented in PowerPoint presentations and shared with the wider research and business teams. Although the approach was dominated by pragmatic methods emphasising speed and usability, academic involvement ensured that the work remained grounded in established principles of research and evaluation (e.g., use of inclusion criteria, validated tools for exploring feasibility, qualitative interviews and analysis to understanding the experiences of the user and delivery agent). Monthly review meetings, which included representatives from research, design, and business teams, provided a structured forum for validating assumptions and refining the prototype as development progressed.
This phase illustrated the complementary value of industry and academic approaches, that is, industry methods allowed for fast-paced iteration and product development, while academic oversight safeguarded against methodological weaknesses that might undermine the eventual credibility of the intervention.
Phase 3: feasibility pilot (usability testing)
Prototype testing was conducted in Zimbabwe. The third phase focused on testing the feasibility and acceptability of the Inuka prototype in a real-world setting. This work was conducted in Zimbabwe between March and September 2016, and again in January 2017, alongside the ongoing Friendship Bench pilot. Insights from Phase 2 had suggested that a brief, human-supported intervention [as defined by the WHO Classification of Digital Health Interventions (63)], delivered by a trained lay coach rather than a fully automated programme, offered the strongest business case, and this testing phase was designed to explore how the prototype could be used in practice.
The academic objective of this phase was to generate evidence on feasibility and acceptability, while the pragmatic objective was to capture insights into user experience and usability. To meet these aims, the study recruited 18 participants (i.e., 13 intended end-users and 5 coaches) through convenience sampling. End-users were required to be over 18 years of age, seeking services from a primary health care facility in Zimbabwe, able to provide informed consent, literate in English, comfortable with text-based chat, and available to attend at least two sessions. Coaches were community grandmothers already trained and active within the Friendship Bench programme, rather than being recruited via a separate eligibility process.
Once enrolled, participants used the prototype under the guidance of trained coaches testing the application on a device provided by the research team. Patients completed a brief online survey after their first session, covering self-reported symptom severity and initial impressions of the app, while both clients and coaches completed system usability surveys. Phase 3 was designed to assess brief usability and acceptability rather than clinical outcomes; no baseline assessment was collected, and formal pre-post outcome testing was not undertaken in this phase (preliminary clinical outcomes were subsequently assessed in Phase 4). Following the second session, one-on-one in-depth interviews were conducted with both groups to further explore their experiences with the intervention. Industry partners contributed by applying usability-focused methods such as screen-capture analysis to observe how participants navigated the app.
Data from this phase were analysed descriptively, as the small sample size prevented the use of inferential statistics, with qualitative and quantitative data reported together. The analysis was disseminated as an internal non-peer reviewed summary report, that was used internally to inform iterative refinements to the app’s usability and the coach-guided process ahead of Phase 4. Together, these approaches built a clearer understanding of how the prototype functioned in practice, and identified key components, such as the screening tool, the self-help modules, and the coach-user interaction, that required further development before large-scale evaluation. Figure 6 provides an illustrative overview of the Phase 3 activities described above, including example usability-driven refinements and anonymised participant feedback.
Figure 6.

Overview of Phase 3 (feasibility pilot/usability testing) activities in Zimbabwe, illustrated with example usability-driven refinements and anonymised participant feedback. Findings are analysed descriptively [section “Phase 3: Feasibility pilot (usability testing)”]; participant identifiers shown are anonymised codes, not names.
Figure 5 illustrates four early iterations of the Inuka application interface during Phase 2 (design and development of the prototype), from the initial self-automated “digital wellbeing” concept (v0.1) through to a human coach-delivered, chat-based model (v0.2 onward). Each iteration reflects specific feedback from hackathon participants and assumption testing, most notably the decision to move away from a fully automated chatbot toward lay coach-delivered support. These screenshots reflect the application’s state during its 2015–2016 development period only; the app has continued to evolve since, and the interface shown here is not representative of the current, publicly available version of Inuka at the time of this paper’s publication.
Figure 5.

Selected iterations of the Inuka application interface during Phase 2 (design and development of the prototype), showing the shift from a self-automated concept (v0.1) to a human coach-delivered, chat-based model (v0.2 onward) between 2015 and 2016.
Phase 4: pilot delivery (pilot study)
The final phase focused on piloting the Inuka Coaching intervention in a larger, real-world setting to assess its feasibility, usability, and a preliminary clinical assessment (59). From March 2017 to March 2018, a prospective cohort study was conducted across 18 primary healthcare centres in Nairobi, Kenya.
Eighty adults were recruited into the study. As this was a pilot/feasibility evaluation, a formal sample size calculation was not conducted, consistent with common practice for pilot studies (59). Eligibility criteria included scoring 8 or higher on the Self-Reporting Questionnaire-20 (SRQ-20) (58), being, aged 18 years or older, conversant in written and spoken English, and familiar with the use of smartphones. Participants provided informed consent before taking part. Each participant was offered four PST sessions (the current service offers up to six) delivered through the Inuka mobile application. Inuka app is described in Table 1 in further details. Sessions were coach-guided and delivered via chat by trained community health volunteers, with the first session lasting about 40 minutes and follow-up sessions lasting 20–30 minutes. As part of Phase 4, clinical measures were selected for administration at three time points: baseline, 4 weeks post-intervention, and 3 months after baseline. These included the SRQ-20 (58), validated in Kenya and widely used to screen for CMDs; the Patient Health Questionnaire-9 (PHQ-9) (64), for depression; and the Generalized Anxiety Disorder scale-7 (GAD-7) (65), for anxiety. The academic objective of this phase was to test the feasibility of administering these measures within a real-world service-delivery setting, and to identify and resolve logistical challenges in doing so. In parallel, the pragmatic objective was to establish whether the service could be delivered feasibly at the larger scale represented by this phase, compared with Phase 3. This case study documents the process of designing and delivering this evaluation; the resulting clinical and feasibility findings are reported in the associated pilot publication (59). Figure 7 provides an illustrative overview of the Phase 4 activities described above, including participant recruitment, intervention delivery, and outcome-measurement time points; clinical and feasibility findings from this phase are reported separately in the associated peer-reviewed publication (59).
Figure 7.

Overview of Phase 4 (pilot delivery) activities in Nairobi, Kenya, showing participant recruitment, intervention delivery, and outcome-measurement time points. Clinical and feasibility findings from this phase are reported separately in ref. (59) and are not duplicated here. GAD-7, Generalized Anxiety Disorder scale-7; PHQ-9, Patient Health Questionnaire-9; PST, problem-solving therapy; SRQ-20, Self-Reporting Questionnaire-20.
In terms of dissemination, this phase reflected the hybrid nature of the project. Initially, results were shared with funders and partners that include both Philips (cross sector team) and the University of Zimbabwe through a short, industry-style report that emphasised key outcomes and practical implications. Subsequently, the data were fully analysed, written-up according to academic standards, and published in a peer-reviewed journal, ensuring that the findings contributed to the scientific literature (59). However, efforts to publish the research were conducted approximately 2 years after the completion of the study.
Phase 4 was notable in that it adhered to academic procedures at every stage: design, evaluation, and dissemination, while still producing outputs valued by non-academic audiences. This balance underscored the strength of the hybrid model, demonstrating how a digital mental health intervention could be both rigorously evaluated and pragmatically advanced toward implementation.
Discussion
Key findings
This case study is the first to document the design, development, and pilot evaluation of Inuka Coaching, a guided digital PST intervention based on the Friendship Bench approach (14). Inuka Coaching was developed through a U-I partnership in SSA. Four phases were used: (I) formative needs assessment; (II) development; (III) feasibility pilot, and (IV) pilot delivery (pilot study). Academic processes were more prominent in Phases 1 and 4, ensuring greater validity, reproducibility, and generalisability, whereas pragmatic processes dominated in Phases 2 and 3, facilitating rapid prototyping and efficient iteration. Overall, the findings suggest that both academic and pragmatic approaches were crucial for creating a balance between speed of development while retaining validity. For example, academic processes strengthened the scientific foundation and credibility of Inuka Coaching, while pragmatic methods reduced costs associated with the time and resources associated with following stricter academic processes, accelerated development, and prevented research waste by uncovering feasibility issues early. The balance of approaches was flexible across phases, reflecting the objectives and constraints at each stage. The 3-year partnership across these studies demonstrated that U-I models can be effectively implemented in LMIC contexts. An important caveat for the rest of the discussion is that the value of the U-I partnership was reported based on relational factors that focused on building a sustainable working relationship, rather than performance against other, more objective metrics. As such, these values represent self-reported experiences only, and cannot be independently or internally validated. It should also be noted that this paper will lean more heavily on discussing the costs to academic rigor, in light of stricter tenets associated with reliability and validity in the academic sector.
Strengths and limitations
Few documented examples exist of U-I collaborations in developing digital mental health tools, and, to our knowledge, none have been reported in the SSA region. In high-income settings, such partnerships often emphasise usability testing or scaling commercially viable apps, rather than developing culturally sensitive interventions using cross-sector collaboration (31,57). Academic institutions, by contrast, typically focus on pilots or evaluations, but rarely progress toward implementing a business model (45). This paper presents an approach that combined cross-sector, collaboration with local stakeholders, iterative prototyping, and pilot evaluation, generating lessons that can help guide future U-I collaborations. Local partners and end-users shaped the intervention throughout, enhancing cultural sensitivity and contextual validity. Participatory design methods (56), including a Nairobi hackathon, demonstrated the value of multidisciplinary collaboration. Since 2019, the Inuka-Friendship Bench-University of Zimbabwe-London School of Hygiene & Tropical Medicine (LSHTM) collaboration has advanced from early digital development to feasibility testing, service innovation and large-scale evaluation. A pilot of Inuka during coronavirus disease 2019 (COVID-19) in Zimbabwe demonstrated the feasibility and acceptability of delivering Friendship Bench-based PST through Inuka and informed refinements in digital supervision, accessibility, and quality assurance (66). This work has since progressed to a Wellcome-funded programme to optimise the intervention and evaluate its effectiveness, implementation, and costs in low-resource settings.
This case study has several limitations. The collaboration described was small and organically developed around shared goals at the start of product development, and may not reflect larger, more formal U-I partnerships with different governance structures, intellectual property arrangements and expectations, or in which a shared goal was not the precursor to the partnership (67). As a retrospective account, the analysis reflects the perspectives of authors with oversight but does not systematically include all team members, especially the cross-sectors collaboration from the industry-end, which may have limited the depth and breadth of reporting in this case report. Another limitation relates to the level of rigor adopted in the academic process. While the paper often presents a dichotomous framing of academic rigor versus the fast pace offered by U-I collaboration, the reality was more nuanced. In attempting to balance the respective values of each sector, the collaboration often produced trade-offs between competing standards, ultimately reducing both academic rigor and the speed of industry iteration, compared to if the research had been conducted within either sector independently. For example, patient and public involvement (PPI) despite being a gold standard in mental health research (68), only 2 of the 20 hackathon participants represented PPI, and their specific contribution to the resulting design was not independently recorded or evaluated alongside the other stakeholder groups. A related limitation was our reliance on opportunistic funding. Each phase followed the priorities of whoever funded it, rather than the recommendations of the phase before, so lessons from one phase could not always be carried into the next. For example, Phase 1 explored mental health needs with a broad range of stakeholders, including people with and without self-reported mental health difficulties, family members, practitioners, and inpatients at a private psychiatric hospital, and found that a preventive, technology-based approach targeting CMDs offered the best fit in terms of need, safety, and feasibility, while a scalable digital intervention was less suited to clinical populations. However, later funding was tied to different communities and countries in SSA. As a result, the intervention was tested and piloted with populations that did not fully match those who had informed its design, and there was no opportunity to repeat the needs assessment or adapt the intervention locally. Although the formative research suggested that mental health needs were similar across the four countries involved, the intervention may not transfer easily between them, and further adaptation is likely to be needed before implementation in any given country. Cultural fit also varied between countries; for example, users in Zimbabwe felt that the Swahili name of the intervention (Inuka) was less appropriate in their context. Transferability may also be limited within countries. Recruitment was concentrated in urban settings, where access to smartphones and reliable connectivity, both of which a technology-based intervention depends on, is more consistent. The findings may therefore not generalise to rural areas of SSA, where infrastructure, access to technology, and the service-user journey are likely to differ substantially. Such contextual factors are a key consideration in the Medical Research Council (MRC) framework for developing and evaluating complex interventions (61). Methods for understanding context were also applied inconsistently. User flow mapping was introduced in Phase 2 but not repeated as the intervention moved to new settings. In addition, no theory of change was developed at any phase. A theory of change sets out the causal mechanisms and assumptions linking intervention activities to intended outcomes and is recommended in the development of complex interventions; without one, it is difficult to establish which components of the intervention, or of the collaboration that produced it, are responsible for any observed effects. Subsequent work has begun to address these gaps by carrying out adaptation work ahead of prospective testing and mapping the user journey in context (66), representing a first attempt to understand context and user needs more systematically.
Finally, the age of the work must be acknowledged. The intervention design took place 2015–2018, with the Phase 4 pilot published in 2021 (59). The case study was written later, due to pragmatic reasons around funding to help disseminate the findings (59). A decision was made to publish the development process due to a lack of available examples of U-I collaboration, and to get past taboo associated with working in Industry in academia, despite significant inefficiencies and research waste associated with digital innovations in mental health (45,54). We also felt it was an important contribution to the literature to provide a case study of some of the realities of these collaborations, while offering solutions grounded in retrospective reflection and existing guidance, helping to frame expectations at a time when U-I collaborations are expected to increase (45).
Explanations of findings
While the partnership achieved its intended goals, it also encountered a number of challenges. In early phases, engagement of people with lived experience of mental illness was constrained by ethical review requirements (69), limiting opportunities for co-design with the target population (e.g., in Phase 2). Input was often confined to limited feedback sessions (e.g., saying) or usability testing (e.g., doing) rather than sustained cross-sector collaboration across iterative cycles. It should be noted that while Inuka Coaching is not a medical device, obstacles to accessing people with mental health problems was a challenge. Future U-I collaboration may adopt “living lab” model, in which end-users are embedded throughout design and testing in real-world settings (70). Such an approach can strengthen U-I collaborations by providing trusted, sustained access to communities, ensuring that user perspectives inform both rapid iteration and academic evaluation
Differences in priorities between sectors created recurring tension. On one hand, researchers in the team wanted to employ rigorous research methodologies and analytics, which while more time intensive, would have garnered more accurate data-points to derive more robust outcomes in usability, acceptability and feasibility. On the other hand, web designers, user-interface, user-experience and front-end developers were keen to engage in agile learning through a process of rapid testing, failing and adaptation after an MVP was developed. Such differing perspectives were especially prominent during Phases 2 and 3. This presented a challenge for the team, because while it was understood that the rapidly changing digital landscape required the team to move fast (71,72), the team equally required the use of robust research methodologies to reliably demonstrate safety, acceptability and intervention outcomes (72). Differences in opinion were facilitated by project leadership, which helped to strategically prioritise research tasks. While an equal contribution between U-I collaboration may be deemed as preferential, long-term collaborations may benefit from adopting a university or industry-dominant leadership in relation to relevant phases or tasks. For instance, the literature on U-I collaborations more broadly suggest that the use of prototypes and models may help to resolve vagueness around the objectives and provide opportunities to connect new and existing knowledge (36).
Different expectations arose around timelines for publishing research. Pragmatic research often adopts short reports with high-level recommendations that can be communicated relatively rapidly, preparing research for peer-reviewed journals typically takes longer to prepare and publish. Over the past 30 years, the average time taken between submission and acceptance was 100 days, not including the time to prepare the manuscript (73). Moreover, in a multidisciplinary team in which different people have been involved in different phases of the study, it can also be time consuming to collate the necessary information to produce a publication, especially when access to data are restricted by data protection procedures. To take an example, the data for phase four could only be accessed and analysed by a member of the Philips team, in line with ethical approvals sought at the time limiting the pace of progress of preparing the manuscript for publication. Institutional review board (IRB) approvals could also not be accessed following the project as the project has been closed, and staff working on the project had moved outside of the organisation and were not able to help.
The development of the Inuka Coaching took place in an industry setting and as a result academic values were not always at the forefront of Phases 1–3 of the project. One study found that the quality of research is less important for establishing the collaboration than the applicability or ‘readiness to use’ knowledge that will be produced (74), which generally aligned with how different U-I research approaches were applied within our project. The prioritisation of different research approaches was also applied in relation to the funders’ requirements. For example, Phase 4 was funded by Grand Challenges Canada, whose scope of funding was better aligned with academic approaches to research. Trust and communication is reported to be key in overcoming cognitive and cultural differences in relation to working timeframes. The literature also suggests that intermediaries may be helpful in resolving differences that may emerge (67).
Comparison with similar research
The findings echo previous calls for hybrid approaches in digital health evaluations (45). Guo et al. argued that novel evidence-generation methods are needed to balance the speed of technology development with the rigor of academic research, in which the adoption of pragmatic research were adopted (32). Our case study supports this, showing how assumption testing, hackathons, and MVPs can coexist with structured evaluation when leadership is flexibly assigned and when partnerships are long-term. Academic involvement was operationalized across the study design, data collection, analysis, and interpretation stages. This included input on the sampling strategy and choice of outcome measures, review of data collection instruments and training of field staff, development of coding framework and statistical analysis plan, and challenge of initial interpretations to guard against overstated claims and bias. Methodological oversight was iterative across Phases 3 and 4. Findings from Phase 3 were reviewed by academic partners, whose feedback directly informed refinements to the Phase 4 design, helping to safeguard against methodological weaknesses such as recruitment bias (through review of the sampling strategy and eligibility criteria), measurement error (through review of the data collection instruments and choice of outcome measures, and training of field staff), and overstated interpretation of pragmatic findings (through independent review of the coding framework and statistical analysis plan, and academic challenge of initial interpretations before dissemination).
The U-I collaboration on this project was not contractual, with the exception of funding in which roles within the project were negotiated based on funder needs, rather, the partnership was formed through establishing common interests. The mental health care gap, and the pervasive stigma experienced in LMICs presented an opportunity for academics who wish to understand how to bridge the care gap through technology, while social entrepreneurship aimed to establish digital services with the potential to solve unmet mental health needs. At present, this collaboration is perceived to be one that will continue in the longer term (66), due to a shared purpose and mutual respect for what each sector has to bring to the partnership. U-I relationships that are predefined contractually have been found to limit trust (31), while having a long-term relationship has been found to increase the likelihood of a successful collaboration, when compared to short-term relationships (29). An informal approach to collaboration may have been effective in this project for two key reasons. First, the working group was small, and roles were clearly defined and did not overlap (75). The project at the time was largely funded by Philips Research, and therefore leadership was attributed in relation to the sector responsible for funding the phases. However, larger projects involving a higher number of stakeholders may benefit from formal approaches to agreeing on responsibilities and leadership. A strong example is Kenya’s “Maker Hub”, which brings together local makers (such as innovators and engineers), maternal, newborn and child health (MNCH) practitioners, and biomedical engineers to co-design, prototype and test low-cost, open-source medical devices that can be made locally. The aim is to improve the supply of MNCH equipment for Kenyatta National Hospital and the facilities that refer patients to it. The Hub was started by the NGO Concern Worldwide, funded by the Bill & Melinda Gates Foundation and the Philips Foundation, and overseen by a committee chaired by the Ministry of Health. Partners from several sectors produced four device prototypes: the Kenya Bureau of Standards evaluated two, and one went through clinical testing (76). Although funding has changed between phases, having the Ministry of Health manage the Hub appears to keep product development on track.
Consistent with our findings, another commonly reported challenge is friction around institutional review board (IRB) processes, which several studies identify as the greatest source of delay and misalignment between partners (77-80). One case study of U-I collaboration in digital health research found that industry partners were often unaware of IRB requirements and timelines. As a result, the academic partner had to act as a formal liaison with the IRB and negotiate priority review just to keep pace with industry expectations (77).
A second U-I case study, on reducing risk for cardio-oncology patients, reported similar concerns and proposed a structural solution: identifying future lines of investigation in advance, so that IRB approval could be secured more quickly and partner trust built before tackling more complex consent and data-collection requirements (78). This tension extends beyond single case reports. A mixed-methods study of UK digital mental health partnerships, which used stakeholder consultation and consensus-building methods to identify principles for effective industry-academic partnerships, found university bureaucracy (e.g., obtaining institutional approvals and following strict processes) and slow contracting to be core structural barriers, alongside cultural differences in how institutions balance speed and rigour (79). Similarly, a retrospective analysis of digital health small and medium-sized enterprises’ (SMEs’) expectations and experiences of U-I collaboration highlighted friction around time and effort. It suggested managing expectations through a review at the start of a project to clarify institutional processes and the support available (80).
Implications and actions needed
The case studies presented in this paper offer several transferable lessons for U-I collaborations in digital mental health:
Clear objectives and flexible sector-based leadership. Industry leadership is best suited to address objectives relating to early prototyping, while academic leadership is critical for evaluation phases responsible for assessing acceptability, feasibility and effective at reducing mental health complaints of the final product. However, this structure should be applied flexibly, allowing both sectors to meet goals associated with their discipline. For example, while industry may be best placed to lead product development phases, there should still be scope to produce outputs that justify academic involvement within these phases of U-I collaboration. For example, this can include planning and documenting development processes that can contribute to the wider literature using academic principles, while still prioritising rapid product development. Respective sector needs should be communicated early stages, to manage expectation (45).
Engage end-users from the outset. Ethics approval processes are often time-consuming and can restrict engagement with vulnerable populations during iterative, responsive work. A living lab model embeds end-users throughout design and real-world testing (70), and can be paired with ethical clearance that anticipates end-user involvement at each phase, reducing the delays associated with returning to institutional review boards each time end-user input is needed (77).
Dual or phased dissemination across an extended time period. Funders and implementers require rapid, high-level summaries, while peer-reviewed publications establish credibility in the scientific community. It is therefore important to factor in a dual or phased dissemination plan, in which rapid reporting can be used during early intervention development, while a sufficient amount of time and resources are allocated to designing research methods, and preparing peer reviewed publications.
Developing shared goals over an extended period. Having a common goal enabled cross-sector discussions in which trade-offs between agility and rigour could be negotiated constructively. Partnerships may therefore benefit from agreeing on a shared, cross-cutting goal that extends beyond any single project phase, reducing friction arising from differing sectoral values. This is critical because some industry partners end collaboration abruptly once a product phase is complete. This can cut off academic access to data, personnel and materials such as ethical approvals, particularly when project staff are reassigned or leave the organisation. Agreeing in advance on continued access to relevant data, and on dissemination timelines that meet academic standards, is therefore important.
Governance bodies should assist U-I collaborations. While an informal partnership was effective for developing Inuka Coaching, larger partnerships may require third-party governance that can suggest U-I collaborations based on shared goals, as well as help navigate cultural and operational gaps between universities and industry. Such intermediaries may act as boundary spanners, ensuring that academic priorities (rigor, evidence, ethical oversight) and industry priorities (agility, usability, scalability) were aligned around common goals, and mediate tensions. We recommend developing a dedicated digital mental health ecosystem, that brings together universities, start-ups, NGOs, funders and governments to ensure interventions are both evidence-based and scalable (76,80).
Agreeing evidence benchmarks at the outset. U-I partners should agree early in the collaboration on the evidence standards required at each stage of development, such as the level of evaluation needed before moving from prototype to pilot or from pilot to scale-up. Making these benchmarks explicit can help partners anticipate where trade-offs between speed and rigour are likely to arise and negotiate them constructively. Over time, such benchmarks could form the core of a practical toolkit to guide U-I teams in aligning research and development processes with the needs of both sectors.
Conclusions
This case demonstrates that hybrid U-I collaborations can successfully design, develop and evaluate culturally sensitive digital mental health interventions in SSA. Having a shared a common goal, clear objectives, flexible leadership and mutual respect allowed academic rigour and industry agility to complement each other. This balance, however, required ongoing negotiation and, at times, trade-offs: the depth of academic rigour and the pace of industry iteration were both reduced compared with what either sector might achieve alone. Future research should focus on developing decision-making frameworks or toolkits to guide U-I teams in balancing speed and scientific rigor, thereby supporting the creation of sustainable, scalable digital health solutions in low-resource settings.
Supplementary
The article’s supplementary files as
Acknowledgments
We thank all stakeholders whose input shaped the intervention across all phases of development.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The activities conducted during Phases 1–3 (2015–2018), including interviews, focus group discussions, and usability/feasibility testing, were undertaken as part of intervention design, development, and evaluation and were considered product development and quality improvement activities at the time. Despite a thorough search, the authors have not been able to locate institutional ethics or exemption documentation for Phases 1-3 from this period. For Phases 1 and 3, written informed consent was obtained from all participants before interviews, focus group discussions and usability testing, after the purpose of the activity and the voluntary nature of participation had been explained. No names were recorded in the data; participants were assigned codes, and quotations and personas reported in this paper are anonymised or composite. The Phase 4 pilot evaluation was approved by the Amref Ethics and Scientific Review Committee (KIAMBU/HRDU/AUTHO/2017/04/10/Njenga A, on 10th April 2017) and informed consent was taken from all individual participants before taking part. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
Footnotes
Funding: The activities described in this case study were funded by Philips Africa/Philips Research (Phases 1–2), Friendship Bench, Zimbabwe (Phase 3), and Grand Challenges Canada (Phase 4). Preparation of this manuscript received no specific funding.
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-2026-0020/coif). R.v.D. is a co-founder of Inuka and holds shares in the company. R.v.D. led the development and evaluation of the Inuka intervention described in this paper, as part of a project led by Philips; Philips’ involvement ended when the project closed. R.v.D. reports that Philips Research and Grand Challenges Canada funded the Inuka project in the time period of the manuscript, but not anymore for many years. D.C. is a co-founder of Inuka and serves as its coaching quality lead as an unpaid advisor and holds shares in the company. The other authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-2026-0020/dss
References
- 1. GBD 2017 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018;392:1789-858. Erratum in: Lancet. 2019 Jun 22;393(10190):e44.. 10.1016/S0140-6736(18)32279-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Walker ER, McGee RE, Druss BG. Mortality in Mental Disorders and Global Disease Burden Implications. JAMA Psychiatry 2015;72:334-41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Bruckner TA, Scheffler RM, Shen G, et al. The mental health workforce gap in low- and middle-income countries: a needs-based approach. Bull World Health Organ 2011;89:184-94.. 10.2471/BLT.10.082784 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.World Health Organization. Human resources for mental health: Workforce shortage in low-and middle- income countries. Geneva; 2011 [cited 2019 Jan 2]. Available online: https://www.who.int/hrh/resources/observer8/en/
- 5. Patel V, Saxena S, Lund C, et al. The Lancet Commission on global mental health and sustainable development. Lancet 2018;392:1553-98. Erratum in: Lancet. 2018 Oct 27;392(10157):1518.. 10.1016/S0140-6736(18)31612-X [DOI] [PubMed] [Google Scholar]
- 6. Vigo D, Thornicroft G, Atun R. Estimating the true global burden of mental illness. Lancet Psychiatry 2016;3:171-8.. 10.1016/S2215-0366(15)00505-2 [DOI] [PubMed] [Google Scholar]
- 7. Atewologun F, Adigun OA, Okesanya OJ, et al. A comprehensive review of mental health services across selected countries in sub-Saharan Africa: assessing progress, challenges, and future direction. Discov Ment Health 2025;5:49.. 10.1007/s44192-025-00177-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Charlson FJ, Diminic S, Lund C, et al. Mental and substance use disorders in Sub-Saharan Africa: predictions of epidemiological changes and mental health workforce requirements for the next 40 years. PLoS One 2014;9:e110208.. 10.1371/journal.pone.0110208 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Komu CK, Ngigi M, Ambrose MJ. Barriers and Facilitators to Accessing Mental Health Services for Adults in Sub-Saharan Africa: A Systematic Review. Ment Heal Sci 2025;3:e70006. [Google Scholar]
- 10.Institute of Medicine (US) Forum on Neuroscience and Nervous System Disorders. Mental, Neurological, and Substance Use Disorders in Sub-Saharan Africa: Reducing the Treatment Gap, Improving Quality of Care: Workshop Summary. In: Systems of Care for MNS in Sub-Saharan Africa. Washington (DC): National Academies Press (US); 2010 [cited 2026 Jul 17]. Available online: https://www.ncbi.nlm.nih.gov/books/NBK53427/ [PubMed]
- 11. Gureje O, Lasebikan VO, Ephraim-Oluwanuga O, et al. Community study of knowledge of and attitude to mental illness in Nigeria. Br J Psychiatry 2005;186:436-41.. 10.1192/bjp.186.5.436 [DOI] [PubMed] [Google Scholar]
- 12. Ryan GK, Bauer A, Endale T, et al. Lay-delivered talk therapies for adults affected by humanitarian crises in low- and middle-income countries. Confl Health 2021;15:30.. 10.1186/s13031-021-00363-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Hoeft TJ, Fortney JC, Patel V, et al. Task-Sharing Approaches to Improve Mental Health Care in Rural and Other Low-Resource Settings: A Systematic Review. J Rural Health 2018;34:48-62.. 10.1111/jrh.12229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Chibanda D, Weiss HA, Verhey R, et al. Effect of a Primary Care-Based Psychological Intervention on Symptoms of Common Mental Disorders in Zimbabwe: A Randomized Clinical Trial. JAMA 2016;316:2618-26.. 10.1001/jama.2016.19102 [DOI] [PubMed] [Google Scholar]
- 15. Le PD, Eschliman EL, Grivel MM, et al. Barriers and facilitators to implementation of evidence-based task-sharing mental health interventions in low- and middle-income countries: a systematic review using implementation science frameworks. Implement Sci 2022;17:4.. 10.1186/s13012-021-01179-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.GSMA. The Mobile Economy Sub-Saharan Africa 2024. Chicago; 2024.
- 17.Silver L. Smartphone Ownership Is Growing Rapidly Around the World, but Not Always Equally. Pew Research Center. 2019.
- 18.Doukani A. Social Inclusion and mHealth. In: Liamputtong P, editor. Handbook of Social Inclusion. Springer, Cham; 2022:1-24. [cited 2022 Dec 25]. Available online: https://link.springer.com/referenceworkentry/10.1007/978-3-030-48277-0_109-4 [Google Scholar]
- 19. Naslund JA, Aschbrenner KA, Araya R, et al. Digital technology for treating and preventing mental disorders in low-income and middle-income countries: a narrative review of the literature. Lancet Psychiatry 2017;4:486-500.. 10.1016/S2215-0366(17)30096-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Fairburn CG, Patel V. The impact of digital technology on psychological treatments and their dissemination. Behav Res Ther 2017;88:19-25.. 10.1016/j.brat.2016.08.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Cuijpers P, Miguel C, Ciharova M, et al. Psychological treatment of adult depression in primary care compared with outpatient mental health care: A meta-analysis. J Affect Disord 2023;339:660-75.. 10.1016/j.jad.2023.07.011 [DOI] [PubMed] [Google Scholar]
- 22. Etzelmueller A, Vis C, Karyotaki E, et al. Effects of Internet-Based Cognitive Behavioral Therapy in Routine Care for Adults in Treatment for Depression and Anxiety: Systematic Review and Meta-Analysis. J Med Internet Res 2020;22:e18100. Erratum in: J Med Internet Res. 2025 Dec 29;27:e69127.. 10.2196/18100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Cuijpers P, Noma H, Karyotaki E, et al. Effectiveness and Acceptability of Cognitive Behavior Therapy Delivery Formats in Adults With Depression: A Network Meta-analysis. JAMA Psychiatry 2019;76:700-7. Erratum in: JAMA Psychiatry. 2019 Sep 1;76(9):986. Erratum in: JAMA Psychiatry. 2022 Feb 1;79(2):180.. 10.1001/jamapsychiatry.2019.0268 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Karyotaki E, Ebert DD, Donkin L, et al. Do guided internet-based interventions result in clinically relevant changes for patients with depression? An individual participant data meta-analysis. Clin Psychol Rev 2018;63:80-92.. 10.1016/j.cpr.2018.06.007 [DOI] [PubMed] [Google Scholar]
- 25. Fu Z, Burger H, Arjadi R, et al. Effectiveness of digital psychological interventions for mental health problems in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Psychiatry 2020;7:851-64. Erratum in: Lancet Psychiatry. 2021 Nov;8(11):e21.. 10.1016/S2215-0366(20)30256-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Kaboré SS, Ngangue P, Soubeiga D, et al. Barriers and facilitators for the sustainability of digital health interventions in low and middle-income countries: A systematic review. Front Digit Health 2022;4:1014375.. 10.3389/fdgth.2022.1014375 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Hermans J, Castiaux A. Knowledge Creation through University-Industry Collaborative Research Projects. The Electronic Journal of Knowledge Management 2007;5:43-54. [Google Scholar]
- 28. Poyago-Theotoky J, Beath J, Siegel DS. Universities and Fundamental Research: Reflections on the Growth of University–Industry Partnerships. Oxford Rev Econ Policy 2002;18:10-21. [Google Scholar]
- 29.Borrell-Damian L. University-industry partnerships for enhancing knowledge exchange. Brussels; 2009 [cited 2020 Nov 20]. Available online: https://eua.eu/downloads/publications/collaborative doctoral education university-industry partnerships doc-careers project.pdf
- 30. Gill KS. Knowledge Networking in Cross Cultural Settings. IFAC Proceedings Volumes 2000;33:269-75. [Google Scholar]
- 31. Wallin J, Isaksson O, Larsson A, et al. Bridging the gap between university and industry: The Mechanisms for innovative efficiency. International Journal of Innovation and Technology Management 2014;11:1440005. [Google Scholar]
- 32. Guo C, Ashrafian H, Ghafur S, et al. Challenges for the evaluation of digital health solutions-A call for innovative evidence generation approaches. NPJ Digit Med 2020;3:110.. 10.1038/s41746-020-00314-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.King DR, Emerson MR, Tartaglia J, et al. Methods for Navigating the Mobile Mental Health App Landscape for Clinical Use. Curr Treat Options Psychiatry 2023. [Epub ahead of print]. doi: 10.1007/s40501-023-00288-4. [DOI] [PMC free article] [PubMed]
- 34. Nicholas J, Larsen ME, Proudfoot J, et al. Mobile Apps for Bipolar Disorder: A Systematic Review of Features and Content Quality. J Med Internet Res 2015;17:e198.. 10.2196/jmir.4581 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Watson D. Misunderstanding Modern Higher Education: Eight “category mistakes”. Oxford; 2012 Feb [cited 2021 Apr 6]. Available online: https://www.hepi.ac.uk/2012/02/29/misunderstanding-modern-higher-education-eight-category-mistakes/
- 36. de Wit-de Vries E, Dolfsma WA, van der Windt HJ, et al. Knowledge transfer in university–industry research partnerships: a review. J Technol Transf 2019;44:1236-55. [Google Scholar]
- 37. Ghauri P, Rosendo-Rios V. Organizational cross-cultural differences in the context of innovation-oriented partnerships. Cross Cultural & Strategic Management 2016;23:128-57. [Google Scholar]
- 38. Clemensen J, Larsen SB, Kyng M, et al. Participatory design in health sciences: Using cooperative experimental methods in developing health services and computer technology. Qual Health Res 2007;17:122-30.. 10.1177/1049732306293664 [DOI] [PubMed] [Google Scholar]
- 39. Shalf J. The future of computing beyond Moore's Law. Philos Trans A Math Phys Eng Sci 2020;378:20190061.. 10.1098/rsta.2019.0061 [DOI] [PubMed] [Google Scholar]
- 40. Kendall JM. Designing a research project: randomised controlled trials and their principles. Emerg Med J 2003;20:164-8.. 10.1136/emj.20.2.164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.PhRMA. Biopharmaceutical Research and Development: The process behind new medicine. 2015 [cited 2020 Nov 22].
- 42. Mohr DC, Weingardt KR, Reddy M, et al. Three Problems With Current Digital Mental Health Research . . . and Three Things We Can Do About Them. Psychiatr Serv 2017;68:427-9.. 10.1176/appi.ps.201600541 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Gilbody S, Littlewood E, Hewitt C, et al. Computerised cognitive behaviour therapy (cCBT) as treatment for depression in primary care (REEACT trial): large scale pragmatic randomised controlled trial. BMJ 2015;351:h5627. Erratum in: BMJ. 2016 Jan 12;352:i195.. 10.1136/bmj.h5627 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Pawloski PA, Brooks GA, Nielsen ME, et al. A Systematic Review of Clinical Decision Support Systems for Clinical Oncology Practice. J Natl Compr Canc Netw 2019;17:331-8.. 10.6004/jnccn.2018.7104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Batty R, Florescu A, Gulliksen E, Wong A. Catalytic environments to drive progress in digital mental health research. London; 2024 [cited 2025 Apr 17]. Available online: https://wellcomeopenresearch.org/documents/9-301
- 46. Siegel DS, Waldman DA, Atwater LE, et al. Commercial knowledge transfers from universities to firms: Improving the effectiveness of university-industry collaboration. J High Technol Manag Res 2003;14:111-33. [Google Scholar]
- 47. Lee YS. The sustainability of university-industry research collaboration: An empirical assessment. J Technol Transf 2000;25:111-33. [Google Scholar]
- 48.Sharma RS. Role of Universities in Development of Civil Society and Social Transformation. Proceedings of International Academic Conferences 2604181, International Institute of Social and Economic Sciences. 2015. Available online: https://ideas.repec.org/p/sek/iacpro/2604181.html
- 49.WHO. Global Initiative on Digital Health. 2024 [cited 2026 Aug 6]. Available online: https://www.who.int/publications/m/item/global-initiative-on-digital-health
- 50.WHO. Global strategy on digital health 2020-2025. Geneva; 2021 [cited 2026 Aug 6]. Available online: https://www.who.int/publications/i/item/9789240020924.
- 51. Agrawal R, Monteiro K, Narasimhamurti NS, et al. A Conversational Agent (PracticePal) to Support the Delivery of a Brief Behavioral Activation Treatment for Depression in Rural India: Development and Pilot-Testing Study. JMIR Form Res 2025;9:e73563.. 10.2196/73563 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Meslin EM, Rager JB, Schwartz PH, et al. Benchmarks for ethically credible partnerships between industry and academic health centers: beyond disclosure of financial conflicts of interest. Clin Transl Med 2015;4:36. Erratum in: Clin Transl Med. 2016 Mar;5(1):4.. 10.1186/s40169-015-0077-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Lo B, Field MJ. Conflict of Interest in Medical Research, Education, and Practice. [Internet]. Institute of Medicine (US) Committee on Conflict of Interest in Medical Research, Education, and Practice. Washington (DC): National Academies Press (US); 2009 [cited 2026 Aug 8]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK22926/ [PubMed]
- 54. Besley JC, McCright AM, Zahry NR, et al. Perceived conflict of interest in health science partnerships. PLoS One 2017;12:e0175643. Erratum in: PLoS One. 2018 Aug 10;13(8):e0202392.. 10.1371/journal.pone.0175643 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.WHO. Monitoring and evaluating digital health interventions. Geneva; 2016 [cited 2026 Jul 28]. Available online: https://www.who.int/publications/i/item/9789241511766
- 56. Clemensen J, Rothmann MJ, Smith AC, et al. Participatory design methods in telemedicine research. J Telemed Telecare 2017;23:780-5.. 10.1177/1357633X16686747 [DOI] [PubMed] [Google Scholar]
- 57. Mann DM, Chokshi SK, Kushniruk A. Bridging the Gap Between Academic Research and Pragmatic Needs in Usability: A Hybrid Approach to Usability Evaluation of Health Care Information Systems. JMIR Hum Factors 2018;5:e10721.. 10.2196/10721 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.World Health Organization. A user’s guide to the Self-Reporting Questionnaire (SRQ-20). Geneva; 1994.
- 59. Doukani A, van Dalen R, Valev H, et al. A community health volunteer delivered problem-solving therapy mobile application based on the Friendship Bench 'Inuka Coaching' in Kenya: A pilot cohort study. Glob Ment Health (Camb) 2021;8:e9. Erratum in: Glob Ment Health (Camb). 2023 Sep 15;10:e54.. 10.1017/gmh.2021.3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60. Cuijpers P, de Wit L, Kleiboer A, et al. Problem-solving therapy for adult depression: An updated meta-analysis. Eur Psychiatry 2018;48:27-37.. 10.1016/j.eurpsy.2017.11.006 [DOI] [PubMed] [Google Scholar]
- 61. Skivington K, Matthews L, Simpson SA, et al. A new framework for developing and evaluating complex interventions: update of Medical Research Council guidance. BMJ 2021;374:n2061.. 10.1136/bmj.n2061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62. Moore GF, Audrey S, Barker M, et al. Process evaluation of complex interventions: Medical Research Council guidance. BMJ 2015;350:h1258.. 10.1136/bmj.h1258 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 63.World Health Organisation. Classification of Digital Health Interventions. Geneva; 2018 [cited 2019 Aug 6]. Available online: http://who.int/reproductivehealth/topics/mhealth/en/
- 64. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001;16:606-13.. 10.1046/j.1525-1497.2001.016009606.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65. Spitzer RL, Kroenke K, Williams JB, et al. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166:1092-7.. 10.1001/archinte.166.10.1092 [DOI] [PubMed] [Google Scholar]
- 66. Dambi J, Norman C, Doukani A, et al. A Digital Mental Health Intervention (Inuka) for Common Mental Health Disorders in Zimbabwean Adults in Response to the COVID-19 Pandemic: Feasibility and Acceptability Pilot Study. JMIR Ment Health 2022;9:e37968.. 10.2196/37968 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 67. Corley EA, Boardman PC, Bozeman B. Design and the management of multi-institutional research collaborations: Theoretical implications from two case studies. Res Policy 2006;35:975-93. [Google Scholar]
- 68. Gray-Burrows KA, Willis TA, Foy R, et al. Role of patient and public involvement in implementation research: a consensus study. BMJ Qual Saf 2018;27:858-64.. 10.1136/bmjqs-2017-006954 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69. Goodyear-Smith F, Jackson C, Greenhalgh T. Co-design and implementation research: challenges and solutions for ethics committees. BMC Med Ethics 2015;16:78.. 10.1186/s12910-015-0072-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70. Fotis T, Kioskli K, Sundaralingam A, et al. Co-creation in a digital health living lab: A case study. Front Public Health 2022;10:892930.. 10.3389/fpubh.2022.892930 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71. Fleming TM, de Beurs D, Khazaal Y, et al. Maximizing the Impact of e-Therapy and Serious Gaming: Time for a Paradigm Shift. Front Psychiatry 2016;7:65.. 10.3389/fpsyt.2016.00065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 72. Torous J, Hsin H. Empowering the digital therapeutic relationship: virtual clinics for digital health interventions. NPJ Digit Med 2018;1:16.. 10.1038/s41746-018-0028-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 73. Powell K. Does it take too long to publish research?. Nature 2016;530:148-51.. 10.1038/530148a [DOI] [PubMed] [Google Scholar]
- 74. Muscio A, Pozzali A. The effects of cognitive distance in university-industry collaborations: Some evidence from Italian universities. J Technol Transf 2013;38:486-508. [Google Scholar]
- 75. Morandi V. The management of industry-university joint research projects: How do partners coordinate and control R&D activities?. J Technol Transf 2013;38:69-92. [Google Scholar]
- 76. Ayah R, Ong'ech J, Mbugua EM, et al. Responding to maternal, neonatal and child health equipment needs in Kenya: a model for an innovation ecosystem leveraging on collaborations and partnerships. BMJ Innov 2020;6:85-91.. 10.1136/bmjinnov-2019-000391 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 77. Ford KL, Moore SL, Zhou S, et al. Advancing evidence-based digital health through an innovative research environment: an academic-industry collaboration case report. Mhealth 2019;5:37.. 10.21037/mhealth.2019.08.08 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 78. MacLeod J, Abdelrahim M, Painter S, et al. Ten step academic-industry digital health collaboration methodology: A case-based guide for digital health research teams with the example of cardio-oncology. Front Cardiovasc Med 2022;9:982059.. 10.3389/fcvm.2022.982059 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 79. Hall SS, Hastings O, Prentice KM, et al. Principles of Industry-Academic Partnerships Informed by Digital Mental Health Collaboration: Mixed Methods Study. JMIR Ment Health 2025;12:e77439.. 10.2196/77439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 80. Austin D, May J, Andrade J, et al. Delivering digital health: The barriers and facilitators to university-industry collaboration. Heal Policy Technol 2021;10:104-10. [Google Scholar]
