A study protocol to co-develop and evaluate interventions that strengthen healthcare resilience to extreme weather events in three African countries: the ASTRA study

Extreme weather events (EWEs) are becoming more frequent and intense due to climate change, disrupting healthcare systems and increasing risks for vulnerable populations. In low- and middle-income countries, these disruptions threaten progress in HIV and tuberculosis (TB) care by limiting access, straining services, and worsening health outcomes. Building health system resilience through environmentally sustainable infrastructure and processes, adequate and skilled workforce, and community engagement is essential to ensure continuity of care. In this four-year study (ASTRA), we aim to co-design, prioritise and evaluate interventions to strengthen community and health system resilience to EWEs for people living with HIV/TB. This multi-phase, mixed-methods study will be conducted in Mozambique, South Africa, and Zambia, countries which are highly vulnerable to EWEs such as storms, cyclones and flooding and bear a high burden of HIV and TB. Phase 1 will focus on understanding the context through stakeholder mapping, scoping interviews, a Delphi consensus process, and assessments of community and health facility vulnerability and adaptive capacity. We will carry out rapid ethnographic assessments [also known as Broad Brush Surveys (BBS)], spatial and statistical modelling, and a policy and programme analysis in this phase. Phase 2 will involve four sequential co-creation workshops in each country to produce prototype interventions that will be refined and prioritised for evaluation. In Phase 3, we will evaluate selected interventions using system dynamics modelling and economic analysis. Using systems dynamics modelling, we will model the final set of interventions, individually and in combination, to explore their potential effects on service delivery during EWEs. The economic evaluation will estimate the costs of inaction, assess the benefits of interventions through multi-criteria decision analysis, and determine their value for money to inform priority setting and policy action. Overall, the innovative use of participatory co-creation processes, system dynamics modelling, and economic evaluation, provides a model for designing and assessing adaptation interventions that are both evidence-based and locally relevant. ASTRA will inform national policy and planning and offer transferable knowledge for other climate-vulnerable settings facing similar challenges.

Introduction

Globally, the frequency, duration, and intensity of extreme weather events (EWEs), such as droughts, floods, and heat waves, have increased significantly, driven by anthropogenic climate change (Ebi and Hess, 2020; Watts et al., 2021). While the direct impacts of EWEs, including food and water insecurity, injuries, illness, and deaths, are well documented, their indirect effects on health systems and population vulnerability are often poorly described (Reed et al., 2022; Ortiz et al., 2022). Notably, EWEs can disrupt healthcare delivery, exacerbating pre-existing individual and population health risks by limiting access to care, affecting supply chains for essential medical resources, displacing populations, and increasing exposure to health threats such as malnutrition, food-, water- and vector-borne diseases (Naidoo et al., 2022). These disruptions place additional strain on health services due to sudden surges in demand, even while the health workforce is reduced by the same event. Over time, these indirect effects may lead to worse health outcomes, particularly for vulnerable populations, contributing to broader systemic challenges that are often difficult to quantify but critical to address (Iwuji et al., 2024; Orievulu et al., 2022).

As EWEs escalate, low- and middle-income countries (LMICs) face a disproportionate burden of impacts, and widening health inequalities (Iqbal et al., 2019). Many LMICs already bear the highest burden of HIV and TB, and there are concerns that the global and national targets set for controlling these epidemics are still out of reach (Joint United Nations Programme on HIV/AIDS, 2024). Weak health systems remain a key barrier to epidemic control, and now, amid a global health funding crisis, their fragility is even more evident (Brennan et al., 2014; Nichols et al., 2025; ten Brink et al., 2025). If left unaddressed, these EWE-related disruptions risk reversing progress in HIV and TB epidemic control, deepening health disparities, and hindering the achievement of universal health coverage.

Health system resilience is the ability to absorb shocks, maintain essential healthcare service functions, and adapt to future demands. It depends on proactive adaptations that enable healthcare systems to withstand, respond to, and recover from disruptions, ensuring continuity of care amid climate-related challenges (Witter et al., 2023). The World Health Organization (WHO) has identified key components essential for delivering safe and quality care during EWEs: a well-trained and supported health workforce; sustainable management of water, sanitation, and hygiene; reliable and sustainable energy systems; and resilient infrastructure, technologies, and operational processes that keep healthcare facilities functional (World Health Organisation, 2013). However, strengthening health systems alone is not enough. Broader adaptation efforts also need to address the vulnerability and adaptive capacity of communities that rely on these healthcare services (Carmen et al., 2022). Without integrating community perspectives, adaptation efforts risk misalignment with local realities, exacerbate inequities, and fail to protect those most affected by climate-related health threats (Dorji et al., 2023). A comprehensive approach that integrates these elements is critical to reducing the health impacts of climate change and securing/safeguarding essential services.

This study aims to co-design, prioritise and evaluate interventions deemed feasible to strengthen community and health system resilience to EWEs for people living with HIV/TB in Mozambique, South Africa, and Zambia. Our findings will provide additional knowledge for communities and health systems to prevent, prepare for and respond to current and emerging health risks associated with EWEs. Specifically, our objectives are to:

i. Systematically engage communities, health service implementers, and policymakers from health-determining sectors to prioritise needs and design interventions/adaptations to strengthen health service delivery during EWEs.

ii. Undertake community and health service delivery vulnerability and adaptive capacity assessments to establish health-related EWE vulnerability baselines.

iii. Develop system dynamics models to evaluate the impact of co-created interventions for improving the health system resilience

iv. Quantify the value-for-money of identified interventions.

The primary research questions linked to the objectives mentioned above are:

i. What is the impact of EWEs on the health and health service delivery, especially for people living with HIV and/or TB?

ii. What policies and actions can help improve the health care delivery and health outcomes of people living with HIV and/or TB during EWEs?

iii. What are the estimated costs and benefits of the proposed interventions co-developed in the study?

Materials and methods

Setting

The study will be conducted in three Southern African countries: Mozambique, South Africa, and Zambia. Specifically, it will be conducted in Sofala Province (Mozambique), KwaZulu-Natal Province (South Africa), and Southern Province (Zambia). Including the three countries enables comparative insights and strengthens regional learning on how communities and health systems can prevent, prepare for, and respond to current and emerging health risks associated with EWEs. While the three countries share similar EWE exposures, public health priorities, and community vulnerabilities, they differ in health system capacity, governance structures, and levels of climate adaptation. By examining these experiences together, the study will identify common problems, showcase specific innovations for each context, and develop strategies from different countries to create effective, appropriate solutions for building health systems that withstand climate change in Southern Africa. All three countries regularly experience severe storms and floods that disrupt healthcare services. Mozambique is particularly vulnerable, with three to twelve cyclones forming annually in the Mozambique Channel (Charrua et al., 2020; Chilaule et al., 2024; Muleia et al., 2024). Sofala Province, one of the most cyclone-prone areas and the focus of this study, has been repeatedly affected by extreme weather (United Nations Office for the Coordination of Humanitarian Affairs, 2022; United Nations Office for the Coordination of Humanitarian Affairs, 2022). In April 2022, South Africa suffered its deadliest EWE in recent history, when KwaZulu-Natal (KZN) Province received over 300 mm of rainfall in 24 h, causing over 500 deaths and ZAR17 billion (GBP 850 million) in damages, including GBP 9.7 million in health facility losses (Mudefi, 2023; Grab and Nash, 2024). Historically affected by flooding and land degradation, Zambia’s Southern Province continues to experience extreme weather. In January 2022, flash floods and Cyclones Ana and Batsirai impacted over 19,000 people, exacerbating vulnerabilities in the region (Cliggett, 2000).

HIV and TB are major public health priorities in the three study countries, with substantial disease burdens and ongoing challenges in epidemic control (Joint United Nations Programme on HIV/AIDS, 2024). South Africa has the highest HIV burden, with 7.7 million people living with HIV in 2023; of these, 77% were receiving antiretroviral therapy, and 71% achieved viral suppression (Joint United Nations Programme on HIV/AIDS, 2024). In 2023, 212,000 new or relapsed TB cases were reported in the country, with 54% of the cases also living with HIV (World Health Organisation, 2024). In 2023, 2.4 million people were living with HIV in Mozambique: 86% were on ART, and 77% were virally suppressed (Joint United Nations Programme on HIV/AIDS, 2024). In the same year, 116,000 new or relapsed TB cases were recorded, with 23% of the cases also living with HIV (World Health Organisation, 2024). In 2023, Zambia had 1.3 million individuals with HIV: 95% on antiretroviral therapy and 92% virally suppressed (UNAIDS, 2024). In the same year, there were 54,000 new or relapsed TB cases reported, with 32% of the cases also living with HIV (World Health Organisation, 2024).

The public health systems in Mozambique, South Africa, and Zambia are structured similarly, with 3–4 levels of healthcare service provision (South Africa Department of Health, 2020; Zambia Ministry of Health, 2017; Mozambique Ministry of Health, 2014). The first level comprises primary care facilities that serve as the entry point and the first point of contact between users and the health system (i.e., primary health care and preventive health services). Notably, in all three countries, most of the population accesses healthcare services from primary healthcare facilities. District hospitals are the second level and provide routine surgical interventions and have greater diagnostic capacity than primary care facilities (e.g., X-ray facilities). Tertiary hospitals are the third level and are larger in terms of infrastructure, with advanced technologies for major surgeries. Provincial, central, and specialised hospitals constitute the upper levels of healthcare service provision (i.e., a broader range of specialised, curative, surgical and rehabilitative services). Health inequities are prevalent in all three countries. In Mozambique, more than half of the population walks for at least an hour to reach their nearest health facility. There are only 7 doctors per 100,000 people, a proportion that is among the lowest in the world (World Health Organisation, 2021). The private sector is limited but growing, mainly in big cities. Those who can afford private health care prefer to seek care in South Africa for complex medical issues. In South Africa, although the private sector is much smaller than the public sector, it accounts for about half of the health expenditures. About 79% of doctors work privately, leaving only 21% in the public sector (The healthcare system in South Africa, 2025). The distribution of healthcare resources is inequitable because it disproportionately favours private healthcare. In Zambia, there is inequitable access to basic health services between provinces and urban and rural areas. In urban areas, 99% of households are within 5 kilometres of a health facility compared to 50% in rural areas such as the Southern Province (Chankova and Sulzbach, 2006).

All three countries have a critical need for investments to improve readiness to deal with the effects of climate change and require urgent adaptation action. According to the Notre Dame-Global Adaptation Index (ND-GAIN) Country Index (Dame UoN, 2025), Mozambique is the 48th most vulnerable and the 17th least ready country to deal with the effects of climate change. Like Mozambique, Zambia has a high level of vulnerability to climate change but a low level of readiness, it is the 42nd most vulnerable country and the 45th least ready country to deal with the effects of climate change (Dame UoN, 2025). Relative to the other two countries proposed for this study, South Africa has lower vulnerabilities, it is the 100th most vulnerable and the 76th least ready country to deal with the effects of climate change (Dame UoN, 2025). However, for a country that is ranked among the most unequal in the world, improvements in readiness are essential to adapt to future challenges, especially for poor and underserved communities such as those found in rural KZN (McKeever, 2024; Statistics South Africa, 2024).

Study design

The ASTRA study is a multi-phase research project that incorporates both qualitative and quantitative methods, with each phase designed to build on and inform the next. This sequential and interdependent approach provides a structured pathway for understanding the study context, co-developing evidence-informed, contextually relevant interventions, and assessing their potential impact in specific settings. The study is organised into three phases, each with a specific focus and connections to the other phases. The overarching goal of Phase 1 activities is to identify opportunities to enhance health service delivery in the context of EWEs and to establish baseline data on health-related vulnerabilities linked to EWEs. The insights gained from this phase will inform the co-design activities in Phase 2. Then, the prototype interventions developed in Phase 2 will be evaluated in Phase 3. The study will draw on expertise from various disciplines, including social sciences, epidemiology, health economics, statistical modelling, environmental and climate sciences, and software development through collaboration with industry partners.

Data collection and analyses

In this section, we outline the three phases of the study, including the methods for data collection and analysis. Notably, some data collection methods within the same phase will be carried out concurrently. A study activity sequence schema is included in Supplementary material, Section 1 to illustrate the sequence of events according to the study’s work plan.

Phase 1: understanding context

The overarching goal of the first phase is to understand how EWEs impact HIV and TB healthcare delivery, identify strengths and gaps in local systems and policies, and inform the design of appropriate interventions. This phase will begin by identifying stakeholders who will be invited to participate in the various study components. Following this, we will conduct scoping interviews and utilise a Delphi process to gain an initial understanding of community and healthcare service delivery concerns related to EWEs. Additionally, we will assess the vulnerability and adaptive capacity of communities and healthcare facilities to EWEs.

Stakeholder mapping

For this study, stakeholders are individuals or groups who may influence or be affected by the implementation or outcomes of the study in any way. These include policymakers, health service providers, implementing organisations, community groups, civil society organisations, people living with HIV or TB, opinion leaders, government departments, and regulatory agencies. Initially, study staff in each country will identify potential stakeholders, drawing on those previously consulted during protocol development and engaging with community advisory boards (CABs). In addition, the study team will collaborate with community and government stakeholders to identify additional community advisory boards that exist in the communities that can be engaged to be part of the study. Stakeholders will be purposely selected to ensure diversity across stakeholder types. They will be informed about the various study components in which they may be invited to engage. Study staff will use a structured interview guide to assess and characterise identified stakeholders based on four key dimensions: their level of interest in the study, their relevance to its implementation or outcomes, their influence or decision-making power, and their level of support for the study (Franco-Trigo et al., 2020).

Scoping interviews with stakeholders

We will conduct scoping interviews with up to 10 purposively selected stakeholders per country to qualitatively explore stakeholders’ understanding of climate issues and their perceptions of the impact of EWEs on individuals, communities, and healthcare delivery. The scoping interviews are semi-structured exploratory interviews which aim to capture a broad range of stakeholder perspectives and contextual insights rather than achieve thematic saturation. Data adequacy will be assessed based on the diversity and relevance of information collected in relation to the interview objectives (Rahimi, 2024; Busetto et al., 2020). The same stakeholders will also participate in a two-round Delphi process to build consensus on priority issues in the study communities. The number of targeted stakeholders is consistent with the literature, which suggests panels of at least 10 participants achieve stable consensus while ensuring manageability across multiple rounds. Participants in these interviews will include local and provincial officials, policymakers, and civil society representatives from health- and health-determining sectors. Interviews will be conducted in English or the local language, at the participants’ preference. They will be conducted via phone, virtual meeting platforms, or in person. We will use rapid thematic analysis methods to develop context-specific statements reflecting the challenges raised by stakeholders (Renfro et al., 2022). These statements will form the basis of a two-round Delphi process to build consensus on priority issues in the study communities (Nasa et al., 2021). In the first round of the Delphi process, participants will be asked to rank the statements in order of relevance and priority of the impact of EWEs on individuals, communities, and healthcare delivery. After the first Delphi Round, the study team will analyse the rankings to identify areas of agreement and divergence. Average ranks will be assigned to the statements to determine their relative priority across all stakeholders. In the second Delphi Round, participants will receive anonymised summary feedback from Round 1, showing how the group ranked each statement. Stakeholders will be asked to re-rank a refined list of statements. A predefined threshold (e.g., 70% agreement on top-ranked items) will be used to determine consensus and to produce a prioritised list of agreed-upon statements (Niederberger et al., 2021).

Vulnerability and adaptive capacity assessments

To understand how communities experience and respond to EWEs, we will use a mix of complementary methods. Broad Brush Surveys will provide a rapid, structured approach to understanding communities’ environments, networks, and lived experiences that shape resilience. Time-series and spatial analyses will show how EWEs affect HIV and TB outcomes over time, adding quantitative evidence to the community insights. Health facility assessments will explore how services, infrastructure, and staff are affected and adapt during EWEs, while policy reviews will examine whether existing frameworks support continuity of care. Together, these approaches provide a connected view of vulnerability and adaptive capacity across community, facility, and policy levels.

Broad brush surveys (rapid ethnographic assessment of community vulnerability and adaptive capacity)

We will conduct community vulnerability assessments to (i) understand the impact of EWEs on HIV and TB healthcare delivery, (ii) identify strengths and gaps in local systems, policies, and capacities, and (iii) inform the design of responsive interventions. Using Broad Brush Surveys (BBS), we will gather data on key community features before focusing on healthcare and climate-related issues (Bond et al., 2023). The BBS approach will be adapted using the WHO Climate Change and Health: Vulnerability and Adaptation Assessment Framework and the DPSIR (Driving Forces, Pressures, State, Impact, Response) framework (World Health Organisation, 2013; Yee et al., 2012). The BBS provides a rapid yet systematic method for generating a holistic understanding of community contexts before a more focused inquiry. Data will be collected on four meta-indicators: physical features, social organisation, social networks, and community narratives (Table 1) (Bond et al., 2023). Fieldwork will span approximately 12 days per community and involve focus group discussions (FGDs), transect walks, structured observations, and in-depth interviews (IDIs) (Supplementary material, Sections 2, 3). FGDs will include 8–12 people, a number that allows diverse perspectives while ensuring adequate opportunities for participation among group members. Rapid thematic analysis of qualitative data will be conducted using a community profile template, enabling real-time data synthesis. Each country team will hold an analysis workshop to produce a short narrative report, layered maps, and visual outputs structured around meta-indicators and the impacts of EWEs on health.

Table 1

Table 1. ASTRA study data collection matrix and expected outputs.

Quantitative individual and ecological vulnerability assessments

The quantitative vulnerability assessment consists of two parts: (1) analysis at the district or other available geographical unit level, and (2) analysis at the individual level. The first part involves mapping patterns of HIV and TB mortality for the study countries, to the extent that data are available, and creating a composite measure of vulnerability where data are available. Data will include demographic, socio-economic, environmental, and health system indicators, as well as weather and disaster data from national sources, EM-DAT (emdat.be), and global climate data repositories (see Supplementary material, Section 4). A health and climate change vulnerability index will be developed, drawing on socio-environmental indicators, to identify highly vulnerable areas using standard composite indicator methodology (Centre JR, 2008). Spatial and/or temporal patterns will be visualised using tables, graphs, and maps to guide further analysis and intervention planning. The composite indicator method allows us to provide a concise summary of the multidimensional vulnerability concept that can inform policymakers, the media and the general public. Composite indicators effectively summarise multi-dimensional and complex concepts and are useful/powerful tools for benchmarking, comparison, and tracking progress over time. Ultimately, they facilitate communication and provide a straightforward, policy-relevant summary to guide decision-making and resource allocation. In South Africa, where longitudinal data are available from a population-based cohort of over 150,000 individuals in KwaZulu-Natal, we will conduct individual-level analyses using Poisson regression and distributed lag models to estimate associations between EWEs and outcomes of interest (Gareta et al., 2021). These models will adjust for seasonal and long-term trends, while stratified analyses will explore differences by age, gender, and cause of death (Warner, 2015; Antonelli et al., 2024). Data preparation, analysis, and visualisations will be undertaken using Stata (StataCorp, College Station, TX, USA), R (R Foundation for Statistical Computing, Vienna, Austria), and QGIS (QGIS Development Team, Open Source Geospatial Foundation).

Health facility vulnerability and adaptive capacity checklists

We will purposively select four healthcare facilities in each study country and assess their vulnerability to EWEs, with a focus on two of the four components required for the provision of safe and quality healthcare during extreme weather events. These two components are the health workforce; and infrastructure, technologies, and service delivery processes, which were identified during stakeholder engagement to inform the development of the grant application. Facilities will include both previously affected and unaffected by EWEs, enabling comparison across varying exposure levels. The assessment will be conducted using the WHO Climate Change and Health Facility Checklist, with each component rated high, medium, or low vulnerability based on participant responses (World Health Organisation, 2013). Data will be captured electronically using REDCap and analysed with STATA (StataCorp, College Station, TX), R (R Foundation for Statistical Computing, Vienna, Austria), or similar statistical software. To complement the checklist findings, we will conduct in-depth interviews with 20—30 clinical and non-clinical staff per country, ensuring diversity in gender and years of experience. Interviews will explore participants’ perceptions of EWEs, preparedness, early warning systems, service disruptions, external support, and personal impacts. Additionally, two focus group discussions will be conducted in each country to further examine lived experiences of healthcare workers and adaptation strategies. Rapid thematic analysis will be used to identify contextual factors influencing the vulnerability and adaptive capacity of health facilities.

Policy and programme analysis

The second component of health facility vulnerability assessments will include an analysis of policies for TB, HIV, and disaster preparedness programmes to identify factors that influence vulnerability and coping capacity across key elements of healthcare service delivery for these two conditions. This includes evaluating whether HIV and TB treatment guidelines contain provisions to maintain uninterrupted care during EWEs and whether national health adaptation plans explicitly address HIV and TB care in the context of climate change. We will also examine how policies in other sectors, such as transport, housing, and energy, affect the continuity of HIV and TB services during such events. The policy analysis will take a retrospective approach through a review of existing policies and guidelines, and a prospective approach to identify opportunities to strengthen or introduce new policy measures (Lawlor, 1996). The analysis will be guided by the four main stages of the policy process: agenda setting, policy formulation, implementation, and evaluation (Lawlor, 1996).

Phase 2: co-creation of interventions

In this second study phase, we aim to develop, refine, and prioritise prototype interventions collaboratively with stakeholders, drawing on insights from Phase 1. To achieve this, we will conduct four sequential workshops in each country, using participatory and consensus-building approaches to ensure the interventions are contextually relevant and grounded in stakeholder input. The number of participants in each workshop will vary depending on the session’s purpose and scope. However, all workshops will have a minimum of 15 participants to ensure a variety of perspectives and encourage balanced discussions. The upper limit will be flexible, depending on the workshop’s objectives, stakeholder availability, and contextual factors in each country. This approach allows diverse viewpoints while maintaining manageable group dynamics that support effective participation.

Workshop 1

The first co-creation workshop in each country will include two parallel brainstorming sessions: one with community-based stakeholders and another with representatives from national, civil society, and international agencies. Each group will include 15–20 purposively selected participants, per country, and be facilitated by two trained study staff in each of the three countries. Facilitators will introduce key concepts such as climate change and EWEs to establish a shared understanding. Study staff will use a structured guide and participatory methods (e.g., buzz groups, round-robin discussions, problem trees), to explore participants’ perceptions of how EWEs impact HIV and TB service delivery (Duea et al., 2022). Findings from Phase 1, including scoping interviews and vulnerability assessments, will be presented to stimulate further discussion. Participants will then refine their inputs and reach a consensus on a consolidated list of challenges and potential interventions.

Workshop 2

The second workshop will combine participants (15–25 people) from the first parallel workshops into a joint workshop. During this workshop, the study staff will present the consolidated lists of EWE-related challenges and proposed interventions developed in Workshop 1. Study staff will use similar participatory and consensus-building methods to facilitate workshop participants’ review and prioritisation of potential interventions.

Workshop 3

The third co-creation workshop will include 20—30 participants per country, bringing together a balanced mix of stakeholders from the second workshop, health service implementers, policymakers from health-determining sectors (such as energy and housing), technical experts, and the study team. Participants will review and refine previously identified challenges and proposed interventions, discuss system constraints, and consider relevant WHO prototype interventions (Supplementary material, Section 5).

Workshop participants will also review two interventions recommended by stakeholders during the study’s planning phase. The first a priori intervention focuses on strengthening the capacity of the health workforce to anticipate, respond to, and recover from the impact of EWEs. A total of 150 health workers (50 in each country) will be trained through a climate change and health module, delivered online with one topic released every other month. Online training for healthcare workers in Mozambique will include official Portuguese translation, and the training materials will be translated into Portuguese to enhance understanding. Participants will be divided into two cohorts of 75, and the course will be evaluated through participant feedback and pre- and post-course assessments. The course will be promoted through health departments and ministries, targeting a diverse mix of healthcare workers, including clinicians, programme managers, facility-based staff, and community health workers across various levels of the healthcare system.

The second planned intervention is the Community Resilience Map, which will be implemented only in South Africa, based on stakeholder recommendations to include it as a potential intervention. This real-time decision-making tool enables data input from communities and decision-makers to support disaster response. Initially deployed during the April 2022 KwaZulu-Natal floods (Naidoo et al., 2022), the Community Resilience Map will be adapted to the current study setting and expanded to include data input from healthcare facilities. Additional enhancements to be discussed during the workshop include an interactive geographical information systems (GIS) map, real-time dashboards tailored to decision-maker needs, artificial intelligence-enabled communication, threaded conversations, alerts, and a directory of organisations with detailed profiles.

Using participatory methods, each country team will develop theories of change for the selected interventions using two causal loop diagrams (CLDs): one illustrating factors influencing the demand for health services among people living with and TB, and the other focusing on supply-side determinants during EWEs. The CLDs will be used to visualise system dynamics, feedback loops, and delays. The workshop will conclude with prioritising prototype interventions and policy options through consensus-based approaches.

Workshop 4

The fourth workshop will bring together the cross-country study team and a small group of experts (n = 4) from Workshop 3 to consolidate findings from all three countries. The focus will be on reviewing and finalising the prototype interventions, CLDs, intervention dynamics, system constraints, and facilitators identified in previous workshops.

Phase 3: evaluating the potential effectiveness and benefit of prototype interventions

To assess how the co-created interventions could strengthen HIV and TB services during EWEs, we will use system dynamics modelling and economic evaluation. System dynamics modelling was selected because it allows simulation of the potential effects of interventions in complex, real-world systems where direct testing may not be feasible. The economic evaluation will combine cost analysis and multi-criteria decision methods to capture financial, social, and equity considerations. Together, these approaches provide a practical and evidence-based way to understand which interventions could offer the most significant potential impact and sustainability.

System dynamics modelling

We will use system dynamics modelling to assess the potential impact of co-created prototype interventions, health workforce capacity building and the Community Resilience Map, on the demand and supply of HIV and TB services during EWEs in the study communities (Supplementary material, Section 6) (Semwanga et al., 2016). The model will be developed using the STELLA simulation package (ISEE Systems, Lebanon, NH, USA) and informed by data from Phases 1 and 2, along with additional information on service availability, general service readiness, and HIV and TB service readiness (Semwanga et al., 2016). To test, verify, and refine the model, a stakeholder workshop will be held with 20—25 participants, including the study team, technical experts, and policymakers from health and related sectors. Participants will simulate policy interventions, provide suggestions to improve usability and visual presentation and validate the model’s structure and assumptions. The final set of interventions will be modelled individually and in combination to explore their potential effects on service delivery during EWEs.

Economic evaluation

The economic evaluation will include three key components: estimating the costs of inaction and the proposed interventions, assessing the benefits of interventions using multi-criteria decision analysis, and evaluating the value for money of each intervention (Supplementary material, Section 7). Costs of inaction, such as premature death, productivity loss, displacement, and strain on health systems, will be estimated using primary data from earlier study phases, financial records, expert input, and published literature from government and research institutions (Organization WH, 2013; Tan-Torres Edejer et al., 2003). Costing will adopt a provider perspective (and patient perspective where possible), applying a 3% discount rate over a 10-year horizon (Tan-Torres Edejer et al., 2003). Costs may include implementation, maintenance, operational, and transition costs, and, where relevant, costs of enabling actions identified through system dynamics modelling.

Multi-criteria decision analysis will assess the benefits of interventions across several criteria, including feasibility, ethical, social, and equity considerations (Thokala et al., 2016). Workshops will be held in each country with 20—30 participants, including community members, policymakers, experts, and study staff. Participants will define benefit-related criteria and assign weights reflecting stakeholder values. A MCDA-based framework will be developed in MS Excel, following the principles of the Analytic Hierarchy Process approach to determine criteria weights and alternative scores (Hansen and Devlin, 2019; Saaty, 1984). Additional analyses will be performed to test the reliability and sensitivity of the weights and scoring. The final step of the economic evaluation will include computing and visualising interventions on a value-for-money chart (Figure 1), comparing benefits and costs to identify those with the highest potential impact.

Figure 1

Figure 1. Overview of linkage between study phases and work packages, ASTRA study.

Data management

Each country will collect and manage its own data in secure repositories hosted on encrypted servers with restricted, password-protected access. Audio recordings, transcripts, and related documents will be stored on secure institutional servers with daily backups and audit trails. Qualitative data, such as IDI and FGD transcripts, will be typed in Microsoft Word, password-protected, and stored separately from any personally identifiable information. Each participant will be assigned a unique study identifier, and the national principal investigator or their designee will maintain a separate, encrypted linkage file that connects these identifiers to personally identifiable information. This linkage file will not be shared or stored with the research data. Hard-copy materials, including signed consent forms and field notes, will be kept in locked, fireproof cabinets accessible only to authorised personnel. Data will be retained for 5 to 10 years after publication or completion of the project, after which all electronic and hard-copy data will be securely destroyed in accordance with institutional and national data protection guidelines.

Discussion

The ASTRA study has the potential to bridge knowledge gaps at individual, community, and policy levels by co-producing evidence that strengthens the resilience of healthcare systems in the face of EWEs. By integrating findings from community and health system vulnerability assessments, the interventions developed will be grounded in context-specific realities and guided by a health equity lens (Singh et al., 2023). The study supports several Sustainable Development Goals (SDGs), including SDG 3 (Good Health and Well-being), SDG 10 (Reduced Inequalities), SDG 13 (Climate Action), and SDG 17 (Partnerships for the Goals) (Lee et al., 2016). Notably, the study contributes to building local research capacity and strengthening the health workforce to prepare for and respond to climate-related shocks.

Key anticipated outcomes include improved awareness among communities, health providers, and policymakers about the health impacts of EWEs and the co-development of resilience plans for study communities. Through participatory workshops and stakeholder engagement, the study design has the potential to mitigate fragmented responses observed during previous EWEs (Gooding et al., 2022). The study findings will provide decision-makers with evidence to support policy development and resource allocation for strengthening climate-resilient health service delivery. Although the primary focus is on HIV and TB, strengthened systems are expected to have spillover benefits for other conditions, including non-communicable diseases.

There are limitations to consider. The health systems in the study countries remain fragile, with challenges such as weak infrastructure, limited human resources, and poor coordination at national and district levels (Karamagi et al., 2024). These may affect participation, timely implementation, and the translation of research into action. Additionally, given the unpredictable nature of EWEs, it is possible that study sites may not experience such events during the study period. To address this limitation, we will rely on historical data and simulation approaches to assess the potential impact of proposed interventions. In line with the country’s ethics guidelines, participants will be reimbursed for their time and expenses incurred (e.g., travel costs) related to study participation.

Conclusion

As climate-related health risks intensify globally, the need for studies like ASTRA to produce scalable, sustainable, and equitable solutions has never been more critical. This work establishes a foundation for future research that connects climate science, public health, and systems thinking to pursue more resilient and inclusive health systems. Its innovative use of system dynamics modelling, economic evaluation, and participatory co-creation processes provides a model for designing and assessing adaptation interventions that are both evidence-based and locally relevant. ASTRA will inform national policy and planning and offer transferable knowledge for other climate-vulnerable settings facing similar challenges.

Author contributions

TM: Conceptualization, Methodology, Visualization, Writing – original draft, Writing – review & editing. KO: Conceptualization, Methodology, Project administration, Writing – original draft, Writing – review & editing. VB: Conceptualization, Funding acquisition, Methodology, Project administration, Writing – original draft, Writing – review & editing. MS: Conceptualization, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing. EG: Conceptualization, Formal analysis, Writing – original draft, Writing – review & editing. TK: Conceptualization, Formal analysis, Writing – original draft, Writing – review & editing. SK-G: Project administration, Writing – original draft, Writing – review & editing. LN: Writing – original draft, Writing – review & editing. MC: Writing – original draft, Writing – review & editing. SK: Formal analysis, Writing – original draft, Writing – review & editing. TC: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. RC: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. EO: Methodology, Writing – original draft, Writing – review & editing. CH: Conceptualization, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing. JS: Conceptualization, Writing – original draft, Writing – review & editing. DK: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. SB: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. AR: Conceptualization, Writing – original draft, Writing – review & editing. NM: Conceptualization, Software, Writing – original draft, Writing – review & editing. CO: Conceptualization, Software, Writing – original draft, Writing – review & editing. SC: Conceptualization, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing. CI: Conceptualization, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. This study was supported by the National Institute for Health and Care Research (NIHR204828).

Acknowledgments

The authors express sincere gratitude to the members of the Scientific Advisory Board from Mozambique, South Africa and Zambia for their invaluable guidance and thoughtful input throughout the development and refinement of the study concept and design. Their expertise and support have been instrumental in shaping the direction and rigour of this work.

Conflict of interest

Authors NM and CO were employed by the company Pista Ventures (Pty) Ltd, Durban, South Africa. Author CO was employed by the company Black Box Design and Development (Pty) Ltd, Durban, South Africa.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The authors declare that Gen AI was used in the creation of this manuscript. We used generative AI (Grammarly) for copy editing, including the improvement of grammar, readability, and clarity of procedures.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fclim.2025.1679392/full#supplementary-material

References

Antonelli, J., Wilson, A., and Coull, B. A. (2024). Multiple exposure distributed lag models with variable selection. Biostatistics 25, 1–19. doi: 10.1093/biostatistics/kxac038,

PubMed Abstract | Crossref Full Text | Google Scholar

Bond, V, Nel, M, Simuyaba, M, Chirwa, T, Viljoen, L, Chiti, B, et al., The broad-brush survey approach. A set of methods for rapid qualitative community assessment 2023 Pretoria: Statistics South Africa.

Google Scholar

Brennan, A., Browne, J. P., and Horgan, M. (2014). A systematic review of health service interventions to improve linkage with or retention in HIV care. AIDS Care 26, 804–812. doi: 10.1080/09540121.2013.869536,

PubMed Abstract | Crossref Full Text | Google Scholar

Busetto, L., Wick, W., and Gumbinger, C. (2020). How to use and assess qualitative research methods. Neurol. Res. Prac. 2:14. doi: 10.1186/s42466-020-00059-z,

PubMed Abstract | Crossref Full Text | Google Scholar

Carmen, E., Fazey, I., Ross, H., Bedinger, M., Smith, F. M., Prager, K., et al. (2022). Building community resilience in a context of climate change: the role of social capital. Ambio 51, 1371–1387. doi: 10.1007/s13280-021-01678-9,

PubMed Abstract | Crossref Full Text | Google Scholar

Centre JR (2008). Handbook on constructing composite indicators: Methodology and user guide Joint Research Centre: Handbook on constructing composite indicators:methodology and user guide: OECD publishing; 2008.

Google Scholar

Chankova, S, and Sulzbach, S: Zambia health services and systems program. Occasional paper series Human resources for health 2006(1). Bethesda: Abt Associates Inc. 1, 5.

Google Scholar

Charrua, A. B., Bandeira, S. O., Catarino, S., Cabral, P., and Romeiras, M. M. (2020). Assessment of the vulnerability of coastal mangrove ecosystems in Mozambique. Ocean Coast. Manage. 189:105145. doi: 10.1016/j.ocecoaman.2020.105145

Crossref Full Text | Google Scholar

Chilaule, S. M., Macuacua, X. V., Mabica, A. P., Miranda, N. A., Pereira, H. d S., Gudo, E. S., et al. (2024). Natural disasters’ impact on water quality and public health: a case study of the cyclonic season (2019–2023). Pollutants 4, 212–230. doi: 10.3390/pollutants4020014

Crossref Full Text | Google Scholar

Cliggett, L. (2000). Social components of migration: experiences from Southern Province, Zambia. Hum. Organ. 59, 125–135. doi: 10.17730/humo.59.1.f29132613q2k543p

Crossref Full Text | Google Scholar

Dame UoN: Notre Dame Global Adaptation Initiative (ND-GAIN). In.; 2025.

Google Scholar

Dorji, T., Morrison-Saunders, A., and Blake, D. (2023). Understanding how community wellbeing is affected by climate change: evidence from a systematic literature review. Environ. Manag. 72, 568–586. doi: 10.1007/s00267-023-01833-w,

PubMed Abstract | Crossref Full Text | Google Scholar

Duea, S. R., Zimmerman, E. B., Vaughn, L. M., Dias, S., and Harris, J. (2022). A guide to selecting participatory research methods based on project and partnership goals. J. Particip. Res. Methods. 3. doi: 10.35844/001c.32605

Crossref Full Text | Google Scholar

Ebi, K. L., and Hess, J. J. (2020). Health risks due to climate change: inequity in causes and consequences: study examines health risks due to climate change. Health Aff. 39, 2056–2062. doi: 10.1377/hlthaff.2020.01125,

PubMed Abstract | Crossref Full Text | Google Scholar

Franco-Trigo, L., Fernandez-Llimos, F., Martínez-Martínez, F., Benrimoj, S., and Sabater-Hernández, D. (2020). Stakeholder analysis in health innovation planning processes: a systematic scoping review. Health Policy 124, 1083–1099. doi: 10.1016/j.healthpol.2020.06.012,

PubMed Abstract | Crossref Full Text | Google Scholar

Gareta, D., Baisley, K., Mngomezulu, T., Smit, T., Khoza, T., Nxumalo, S., et al. (2021). Cohort profile update: Africa Centre demographic information system (ACDIS) and population-based HIV survey. Int. J. Epidemiol. 50, 33–34. doi: 10.1093/ije/dyaa264,

PubMed Abstract | Crossref Full Text | Google Scholar

Gooding, K., Bertone, M. P., Loffreda, G., and Witter, S. (2022). How can we strengthen partnership and coordination for health system emergency preparedness and response? Findings from a synthesis of experience across countries facing shocks. BMC Health Serv. Res. 22:1441. doi: 10.1186/s12913-022-08859-6,

PubMed Abstract | Crossref Full Text | Google Scholar

Grab, S. W., and Nash, D. J. (2024). A new flood chronology for KwaZulu-Natal (1836–2022): the April 2022 Durban floods in historical context. S. Afr. Geogr. J. 106, 476–497. doi: 10.1080/03736245.2023.2193758

Crossref Full Text | Google Scholar

Hansen, P., and Devlin, N. (2019). “Multi-criteria decision analysis (MCDA) in healthcare decision-making” in Oxford research Encyclopedia of economics and finance. Geneva: World Health Organisation

Google Scholar

Iqbal, U., Rabrenovic, M., and Li, Y.-C. (2019). Health care quality challenges in low-and middle-income countries, vol. 31: Oxford University Press, Int J Qual Health Care 31:165–165.

Google Scholar

Iwuji, C. C., McMichael, C., Sibanda, E., Orievulu, K. S., Austin, K., and Ebi, K. L. (2024). Extreme weather events and disruptions to HIV services: a systematic review. The Lancet HIV 11, e843–e860. doi: 10.1016/S2352-3018(24)00186-3,

PubMed Abstract | Crossref Full Text | Google Scholar

Joint United Nations Programme on HIV/AIDS. 2024 2024 global AIDS report — The Urgency of Now: AIDS at a Crossroads.

Google Scholar

Karamagi, H. C., Afriyie, D. O., Charif, A. B., Sy, S., Kipruto, H., Moyo, T., et al. (2024). Mapping inequalities in health service coverage in Africa: a scoping review. BMJ Open 14:e082918. doi: 10.1136/bmjopen-2023-082918,

PubMed Abstract | Crossref Full Text | Google Scholar

Lawlor, E. F. (1996). Policy change and learning: An advocacy coalition approach. Journal of Policy Analysis and Management. 1996, 15:110–121.

Google Scholar

Lee, B. X., Kjaerulf, F., Turner, S., Cohen, L., Donnelly, P. D., Muggah, R., et al. (2016). Transforming our world: implementing the 2030 agenda through sustainable development goal indicators. J. Public Health Policy 37, 13–31. doi: 10.1057/s41271-016-0002-7

Crossref Full Text | Google Scholar

McKeever, M. (2024). Social stratification and inequality in South Africa. Sociol. Compass 18:e13173. doi: 10.1111/soc4.13173

Crossref Full Text | Google Scholar

Mozambique Ministry of Health Strategic plan for the Helth sector 2014–2019 2014 Maputo: Mozambique Ministry of Health.

Google Scholar

Mudefi, E. (2023). Disaster management ‘deeds’ in the context of April 2022 KwaZulu-Natal floods: a scoping review. Int. J. Disaster Risk Reduct. 98:104122. doi: 10.1016/j.ijdrr.2023.104122

Crossref Full Text | Google Scholar

Muleia, R., Maúre, G., José, A., Maholela, P., Adjei, I. A., Karim, M. R., et al. (2024). Assessing the vulnerability and adaptation needs of Mozambique’s health sector to climate: a comprehensive study. Int. J. Environ. Res. Public Health 21:532. doi: 10.3390/ijerph21050532,

PubMed Abstract | Crossref Full Text | Google Scholar

Naidoo, K., Manyangadze, T., and Lokotola, C. L. (2022). Primary care disaster management for extreme weather events, South Africa. African J. Primary Health Care Family Med. 14:3778. doi: 10.4102/phcfm.v14i1.3778,

PubMed Abstract | Crossref Full Text | Google Scholar

Nasa, P., Jain, R., and Juneja, D. (2021). Delphi methodology in healthcare research: how to decide its appropriateness. World J. Methodol. 11, 116–129. doi: 10.5662/wjm.v11.i4.116,

PubMed Abstract | Crossref Full Text | Google Scholar

Nichols, B. E., Geng, E. H., Moakley, E., Phillips, A. N., Imai-Eaton, J. W., Stover, J., et al. (2025). Rapid development of an online tracker to communicate the human impact of abruptly halting PEPFAR support. J. Int. AIDS Soc. 28:e26433. doi: 10.1002/jia2.26433,

PubMed Abstract | Crossref Full Text | Google Scholar

Niederberger, M, and Köberich, SNetwork D Coming to consensus: the Delphi technique Oxford University Press 2021 Eur. J. Cardiovasc. Nurs. 20 692–695 doi: 10.1093/eurjcn/zvab059

Crossref Full Text | Google Scholar

Organization WH: Climate change and health: a tool to estimate health and adaptation costs. In: Climate change and health: A tool to estimate health and adaptation costs. edn.; 2013.

Google Scholar

Orievulu, K., Ayeb-Karlsson, S., Ngwenya, N., Ngema, S., McGregor, H., Adeagbo, O., et al. (2022). Economic, social and demographic impacts of drought on treatment adherence among people living with HIV in rural South Africa: a qualitative analysis. Clim. Risk Manag. 36:100423. doi: 10.1016/j.crm.2022.100423,

PubMed Abstract | Crossref Full Text | Google Scholar

Ortiz, A. M. D., Chua, P. L., Salvador, D. Jr., Dyngeland, C., Albao, J. D. G. Jr., and Abesamis, R. A. (2022). Impacts of tropical cyclones on food security, health and biodiversity. Bull. World Health Organ. 101, 152–154. doi: 10.2471/BLT.22.288838,

PubMed Abstract | Crossref Full Text | Google Scholar

Rahimi, S. (2024). Saturation in qualitative research: an evolutionary concept analysis. Int. J. Nurs. Stud. Adv. 6:100174. doi: 10.1016/j.ijnsa.2024.100174,

PubMed Abstract | Crossref Full Text | Google Scholar

Reed, C., Anderson, W., Kruczkiewicz, A., Nakamura, J., Gallo, D., Seager, R., et al. (2022). The impact of flooding on food security across Africa. Proc. Natl. Acad. Sci. USA 119:e2119399119. doi: 10.1073/pnas.2119399119,

PubMed Abstract | Crossref Full Text | Google Scholar

Renfro, C. P., Rome, Z., Gatwood, J., and Hohmeier, K. C. (2022). Use of rapid assessment procedures when analyzing qualitative data in pharmacy research. Res. Soc. Adm. Pharm. 18, 2249–2253. doi: 10.1016/j.sapharm.2021.05.013,

PubMed Abstract | Crossref Full Text | Google Scholar

Saaty, T. L. (1984). “The analytic hierarchy process: decision making in complex environments” in Quantitative assessment in arms control: Mathematical modeling and simulation in the analysis of arms control problems Eds. A, Rudolf Huber, and K, Reiner. (Boston: Springer), 285–308.

Google Scholar

Semwanga, A. R., Nakubulwa, S., and Adam, T. (2016). Applying a system dynamics modelling approach to explore policy options for improving neonatal health in Uganda. Health Res. Policy Systems 14, 1–17. doi: 10.1186/s12961-016-0101-8,

PubMed Abstract | Crossref Full Text | Google Scholar

Singh, D. R., Sah, R. K., Simkhada, B., and Darwin, Z. (2023). Potentials and challenges of using co-design in health services research in low-and middle-income countries. Glob. Health Res. Policy 8:5. doi: 10.1186/s41256-023-00290-6,

PubMed Abstract | Crossref Full Text | Google Scholar

South Africa Department of Health Referral policy for south African health services and referral implementation guidelines 2020. Pretoria: South Africa Department of Health

Google Scholar

Statistics South Africa. Poverty in South Africa: Findings from general household survey, 2019 & 2022. In.; 2024.

Google Scholar

Tan-Torres Edejer, T, Baltussen, R, Adam, T, Hutubessy, R, Acharya, A, et al., Making choices in health: WHO guide to cost-effectiveness analysis. 2003. Geneva: World Health Organisation

Google Scholar

ten Brink, D., Martin-Hughes, R., Bowring, A., Wulan, N., Burke, K., Tidhar, T., et al. (2025). Impact of an HIV budget crisis: modelling HIV infections and mortality in 26 countries. doi: 10.1016/S2352-3018(25)00074-8

Crossref Full Text | Google Scholar

The healthcare system in South Africa (2025). Available online at: https://www.expatica.com/za/healthcare/healthcare-basics/healthcare-in-south-africa-105896/

Google Scholar

Thokala, P., Devlin, N., Marsh, K., Baltussen, R., Boysen, M., Kalo, Z., et al. (2016). Multiple criteria decision analysis for health care decision making—an introduction: report 1 of the ISPOR MCDA emerging good practices task force. Value Health 19, 1–13. doi: 10.1016/j.jval.2015.12.003,

PubMed Abstract | Crossref Full Text | Google Scholar

United Nations Office for the Coordination of Humanitarian Affairs Mozambique: Tropical cyclone Gombe flash update no.6 2022 Maputo:United Nations Office for the Coordination of Humanitarian Affairs.

Google Scholar

United Nations Office for the Coordination of Humanitarian Affairs Mozambique: Tropical storms ana and Dumako flash update no.10 in. 2022 Maputo:United Nations Office for the Coordination of Humanitarian Affairs.

Google Scholar

Warner, P. (2015). Poisson regression. J. Family Plan. Reproductive Health Care 41, 223–224. doi: 10.1136/jfprhc-2015-101262,

PubMed Abstract | Crossref Full Text | Google Scholar

Watts, N., Amann, M., Arnell, N., Ayeb-Karlsson, S., Beagley, J., Belesova, K., et al. (2021). The 2020 report of the lancet countdown on health and climate change: responding to converging crises. Lancet 397, 129–170. doi: 10.1016/S0140-6736(20)32290-X,

PubMed Abstract | Crossref Full Text | Google Scholar

Witter, S., Thomas, S., Topp, S. M., Barasa, E., Chopra, M., Cobos, D., et al. (2023). Health system resilience: a critical review and reconceptualisation. Lancet Glob. Health 11, e1454–e1458. doi: 10.1016/S2214-109X(23)00279-6,

PubMed Abstract | Crossref Full Text | Google Scholar

World Health Organization Regional Office for Europe: Climate change and health: A tool to estimate health and adaptation costs. World Health Organization Regional Office for Europe; 2013

Google Scholar

World Health Organisation: The state of the health workforce in the WHO African region - 2021. Brazzaville: World Health Organisation Regional Office for Africa.

Google Scholar

World Health Organisation (2024). Global tuberculosis report 2024 Geneva: World Health Organisation.

Google Scholar

Yee, S. H., Bradley, P., Fisher, W. S., Perreault, S. D., Quackenboss, J., Johnson, E. D., et al. (2012). Integrating human health and environmental health into the DPSIR framework: a tool to identify research opportunities for sustainable and healthy communities. EcoHealth 9, 411–426. doi: 10.1007/s10393-012-0805-3,

PubMed Abstract | Crossref Full Text | Google Scholar

Zambia Ministry of Health Zambia National Health Strategic Plan 2017–2021 2017 Lusaka: Zambia Ministry of Health.

Google Scholar