Co-creating interventions to prevent mosquito-borne disease transmission in hospitals

1.2 Mbd transmission within health care environments in zanzibar

Malaria control initiatives coordinated by the government-led Zanzibar Malaria Elimination Programme (ZAMEP) during the early 2000s significantly reduced malaria prevalence in the archipelago (Björkman et al. 2019). Today, ZAMEP follows global guidelines for integrated vector management (IVM), set out by the WHO (Tanzanian Ministry of Health 2016). IVM is an evidence-based approach that combines multiple interventions with local stakeholders, government, vector control programmes and health care systems to reduce vector populations and prevent vector-borne diseases (World Health Organization 2012). Despite these successes, recent malaria outbreaks (Kooiman, 2025; Nachilongo 2024) and the increasing transmission of diseases such as dengue, yellow fever, and lymphatic filariasis in sub-Saharan Africa highlight the ongoing threat of mosquito-borne diseases (MBDs) in the region (Weetman et al. 2018). Owing to Zanzibar’s proximity with mainland Tanzania and popularity as a tourist destination, it is at risk of importing MBDs (Le Menach et al. 2011; Muller et al. 2025).

In MBD-prone areas, such as Zanzibar, the design of the built environment can influence the risk of MBD transmission (Lindsay et al. 2021; Mshamu et al. 2022; Seidlein et al. 2019; von Seidlein et al. 2017; Wilson et al. 2020). This can be through the creation of environments where mosquitoes can easily breed like unscreened water systems, building locations near breeding sites, openings in the building envelope where mosquitoes can enter or creating environments that are not conducive to adopting preventative measures like mosquito nets (Sloan Wood et al., in review).

Hospital design and usage patterns pose challenges for controlling disease-transmitting mosquitoes. In Zanzibar, hospitals are often situated in densely populated areas, providing abundant blood meals and breeding sites for mosquito species that thrive in the built environment (Figure 1) (Saleh et al. 2018; Saleh et al. 2020). Common breeding sites for mosquito species such as Aedes and Culex include artificial containers (Mboera et al. 2016; Ngingo et al. 2020; Saleh et al. 2018; Saleh et al. 2020) and unscreened water systems such as septic tanks (Chavasse et al. 1995; Jones et al. 2012). Open and crowded waiting areas expose visitors and staff to mosquito bites while increasing the availability of blood meals (Correia et al. 2012; Kampango, Furu, Sarath, Haji, Konradsen, Schiøler, Alifrangis, Weldon & Saleh 2021; Kirby et al. 2008). The building envelope often allows mosquito entry through unscreened or poorly sealed windows and doors (Donnelly et al. 2020; Jawara et al. 2018; Kaindoa et al. 2018; Musoke et al. 2018; Roberts et al. 2002).

Figure 1

Once inside, mosquitoes can move freely between wards when doors are left open for ventilation or operational tasks like cleaning and monitoring. Overcrowding due to bed shortages further heightens transmission risks, with some patients sharing beds (Falchetta et al. 2020). Hospitals may also act as transmission hotspots owing to their role in serving diverse communities and housing vulnerable patient populations in mosquito infested buildings (Almeida-Nunes et al. 2016; Ehelepola & Wijesinghe 2018). This risk is exacerbated by inadequate systems for detecting, managing, and controlling MBDs (Saleh et al. 2021; Saleh et al. 2022).

Common environmental malaria interventions conducted by ZAMEP and the WHO, such as indoor residual spraying (IRS), distribution of long-lasting insecticidal nets (LLINs) and insecticide-treated nets (ITNs) and larviciding (Tanzanian Ministry of Health 2016; World Health Organization 2012), present challenges in hospital settings. Continuous operations restrict LLIN and ITN use to ward patients, leaving out-patients and visitors in open waiting areas vulnerable to daytime-biting mosquitoes (Khatib et al. 2025). IRS disrupts hospital activities as rooms must be cleared before spraying. Larviciding requires consistent application and monitoring but may miss breeding habitats of other disease vectors, such as discarded water-holding containers or blocked roof gutters (Parker & Allan 2019). Additionally, global concerns over insecticide resistance and ecological impacts have prompted calls to reduce insecticide use (World Health Organization 2017).

1.2 Utilising co-creation and living labs to develop integrated interventions

Given the persistent threat of MBDs in the region, particularly in health care settings, the built environment’s contribution to increased MBD risk, and the challenges of implementing conventional control measures such as IRS, LLINs, ITNs, and larviciding in these contexts, there is a clear need for context-specific interventions (Monroe et al. 2020). WHO guidelines, which influence national policy, further emphasise involving diverse stakeholders, especially those most affected by MBDs, in designing and testing such interventions within real-world environments (World Health Organization 2020).

Living labs and co-creation offer complementary frameworks for designing context-specific health care interventions. A systematic review by Hossain et al., highlights significant variation in living lab definitions and applications (Hossain et al. 2019). Living labs are widely recognised as ‘user-centred open innovation ecosystems integrating research and innovation processes in real-life communities and settings’, as defined by the European Network of Living Labs (ENoLL 2025). Co-creation also has diverse definitions and is often conflated with terms like ‘co-design’ and ‘co-production’ (Vargas et al. 2022). In public health contexts, Vargas defines co-creation as ‘the collaborative approach of creative problem solving between diverse stakeholders at all stages of an initiative’ (Vargas et al. 2022).

Co-creation has proven effective in developing context-specific public health interventions that integrate a diverse stakeholder knowledge and needs (Osaki et al. 2024; Sequeira D’mello et al. 2024; Maaløe et al. 2018; Durand et al. 2014). In public health research, co-creation methods can offer other co-benefits for participants including knowledge integration, innovation and empowerment (Agnello et al. 2025). Living labs provide a framework for embedding the development, implementation and testing on these interventions within complex real-world settings like hospitals (Hossain et al. 2019).

While living labs and co-creation share many overlaps, two distinctions are important for hospital-based vector control interventions: living labs are inherently tied to testing within real-world environments (Bergvall-Kåreborn et al. 2009; Følstad 2008; Leminen et al. 2016; Mulder et al. 2008; Voytenko et al. 2016), whereas co-creation does not require this feature (Agnello et al. 2023; An et al. 2023; Longworth et al. 2024; Messiha et al. 2023; Vargas et al. 2022); co-creation mandates continuous multistakeholder engagement across all project phases, whereas living labs allow selective involvement at specific stages (Longworth et al. 2024; Messiha et al. 2023; Vargas et al. 2022). Given that hospital-acquired MBD transmission is influenced by the hospital environment and its use and operation by diverse stakeholders, both co-creation and living labs could offer valuable insights for sustainable mosquito control intervention design and testing.

The MBD-Free Project is a transdisciplinary research initiative to reduce hospital-acquired MBDs in Zanzibar through IVM interventions by modifying real-world hospital environments as in living labs, while employing co-creation to develop these interventions with stakeholders. This paper presents the co-creation process used to develop the first round of interventions in the (named omitted) project, in collaboration with staff, patients, and other stakeholders, initiated in December 2024 and implemented between July and September 2025.

The objectives of the study are:

1. To co-create integrated mosquito control interventions for implementation in hospital settings of Zanzibar.

2. To identify benefits and challenges of the co-creation process to inform future design of vector control interventions for health care settings.

2.1 Study sites

The study was conducted in four district hospitals of Unguja Island, Zanzibar, located 30 kilometres off Tanzania’s coast (Figure 1) and home to 1,346,332 people (National Bureau of Statistics 2022). The four district hospitals were built between 2023 and 2024 following the same design (Figure 1). These hospitals were constructed alongside six others of similar design and one regional hospital across the archipelago of Zanzibar. The four sites were purposefully selected as they represent urban, peri-urban and rural contexts in Zanzibar.

2.2 Study design

This study describes the co-creation process used to design and implement the first round of mosquito control interventions, conducted between June 2024 and September 2025. Environmental, entomological, and qualitative data were collected starting in June 2024, for one year prior to the planned start of intervention implementation. Interventions were allocated across four study hospitals and included two groups of co-created measures (I-1 and I-2). This study is part of the MBD-Free Project, which will include two further rounds intervention implementation across the study sites planned for 2026.

2.3 The co-creation process

The co-creation process to identify, develop, and implement integrated mosquito control interventions in the study hospitals, followed the PRODUCES+ guideline and recommendations for participatory methods and process evaluation (Agnello & Longworth 2022; Agnello, Ryom et al. 2024; Leask et al. 2019). PRODUCES+ is an evidence-based guideline and framework developed to provide systematic, step-by-step guidance for the co-creation, implementation, and evaluation of public health interventions through open collaboration and shared decision-making among stakeholders, aiming to improve public health outcomes. The process followed the four stages of PRODUCES+ co-creation: planning, conducting, evaluating and reporting, described in Figure 2. Evaluating and reporting took place after each workshop. The co-creation methods that workshop activities were based on were sourced from Agnello, Balaskas et al.’s 2024 Co-Creation Method Inventories and Friis’s 2015 Co-Creation Cards (Agnello, Balaskas et al. 2024; Agnello et al. 2025; Friis 2015).

Figure 2

2.4 Stage one: planning

The planning stage involved framing the study using Leask et al. 2019’s PRODUCES framework (Leask et al. 2019), stakeholder sampling and recruitment and needs assessment. Before initiating the co-creation process, project facilitators attended a three-day training workshop in Zanzibar in July 2024. This workshop established a shared understanding of co-creation among the interdisciplinary team, familiarised facilitators with the PRODUCES+ guideline, and initiated the planning stage.

2.4.3 Needs assessment

A comprehensive needs assessment (Altschuld & Lepicki 2009) was undertaken using mixed methods. Methods were identified by the research team that could provide data related to environmental factors and behaviours that could increase risk of MBD transmission, identified in the literature review (Sloan Wood et al., in review). This comprised of semi-structured interviews with patients and staff, non-participant observations, architectural surveys, environmental measurements and entomological data collection. Interviews across four hospitals, guided by a project-developed script (Supplemental data file 2: Interview Guides), explored mosquito-prone areas, behaviours influencing human–mosquito contact such as window and door use, environmental factors shaping these behaviours, and current mosquito control practices; when possible, staff participated in walkalong interviews (Kusenbach 2003), enabling them to demonstrate relevant hospital spaces and challenges. Regular observational visits provided insights into environmental and behavioural factors linked to MBD transmission risk, including occupant density, ventilation patterns, waste management and maintenance practices (O’Cathain et al. 2019). Architectural data, gathered by a qualified architect using surveys, construction drawings, drone imagery and photography, identified building and site features affecting MBD risk, such as window screening, orientation, shading, and spatial planning. Environmental measurements, including CO₂ concentration, temperature, humidity, door-opening frequency and thermal imaging, were used to assess conditions influencing ventilation and occupant behaviour, while entomological surveys and mosquito trap collections across the hospital grounds documented species presence and populations. Figure 3 shows some of the needs identified through this process.

Figure 3

2.5 Stage two: conducting

Four types of workshop were conducted during the conducting stage. Workshop activities are described in Table 2 and Figures 4 and 5. Methods for all workshops are sourced from Agnello, Balaskas et al.’s 2024 Co-Creation Method Inventories and Friis’s 2015 Co-Creation Cards (Agnello, Balaskas et al. 2024; Agnello et al. 2025; Friis 2015).

Figure 4

Figure 5

The introduction workshop (W1) introduced co-creators to the project team, project focus and the co-creation process. The intervention ideation workshop (W2) focused on stakeholder needs identified through the needs assessment, to ideate mosquito control interventions and how these could be tailored to the hospital environment. Intervention development workshops (W3 and W4) took interventions ideated in the previous workshops and gathered feedback on sketch proposals to refine the designs further. Implementation workshops (W5) included co-creators in the implementation and construction of the interventions through a communal planting day. To enhance a shared sense of ownership as outlined by Leask et al. (2019), participants received T-shirts and diaries featuring a custom logo and bilingual slogan – ‘I’m co-creating mosquito-free hospitals’/‘Tuweke mazingira ya hospitali bila ya mbu’ – to foster team identity (Figure 5A). WhatsApp groups were created for each hospital to facilitate ongoing communication.

Qualitative data were collected using custom templates filled out by participants during workshop and pinned to walls for group discussion (Figure 5C and E; Supplemental data file 4: Co-creation templates) or via online questionnaires. Facilitators reviewed templates during workshops to clarify ambiguous points. Templates were photographed, translated from Swahili to English and transcribed. Sticky dot votes were counted and recorded. All templates are available to download (https://zenodo.org/communities/mbdfree/records?q=&l=list&p=1&s=10&sort=newest).

2.6 Stage three: evaluating

The evaluation stage assessed workshop quality and adherence to co-creation principles through participant exit surveys (Supplemental file 5: Exit surveys), closing circle discussions, participant interviews (Supplementary file 2: Appendix. Interview guides), a post-workshop team self-assessment conducted by facilitators, and an evaluation guide completed by an observer (Agnello, Ryom et al. 2024).

2.7 Stage four: reporting

Harvest reports summarising feedback from the intervention development workshop and feedback workshops were created for each hospital, shared with co-creators via WhatsApp groups and re-presented at subsequent workshops. A comprehensive evaluation is planned for the end of the MBD-Free Project.

2.8 Positionality statement

The authors include architects, global health researchers, co-creation specialists, entomologists, public health practitioners from Zanzibar, Tanzania, Denmark, the UK and the USA. All have experience in public health interventions, largely in the study context or Tanzania. The authors acknowledge that their diverse professional and cultural backgrounds, including insider and outsider perspectives, may influence the study. To reduce bias, inclusive stakeholder engagement, conducting co-creation workshops in local languages, and reflexive evaluation were prioritised throughout the project.

3.1 Co-creation process evaluation

Exit surveys from workshops W1 and W2 showed that 98% of participants (51/52) agreed or strongly agreed that their knowledge could help make hospitals safer from mosquitoes, and 100% (52/52) learned more about where mosquitoes are found. All participants reported increased understanding of ways to prevent mosquito bites in the hospital (79.2% strongly agreed, 20.8% agreed), and 96.2% (51/53) felt their experiences could further improve these methods. Team self-assessment feedback led to practical adjustments, such as clarifying handwriting. Interview responses echoed these findings, highlighting participant pride, learning about mosquito risks, and intent to apply new knowledge both inside and outside the hospital:

4.1 Benefits

Adhering to PRODUCES+ guidance provided the co-creation process with a clear structure; workshops were framed around defined problems and prioritised stakeholder needs as a core component of intervention design; continuous evaluation and improvements were embedded into the process, and outcomes were shared with participants in the reporting stage to ensure transparency.

Co-creation and living lab methods enabled both the adaptation of well-known vector control interventions identified in the literature review, such as insect-proofing septic tanks, improving doors, and repairing window screening, to the specific hospital environment and the development of more novel approaches such as distributed exterior waiting areas, personalised bed fans, and shading through planting that directly address needs emerging from the co-creation process, such as overheating in wards and the desire for more shaded gardens. The process also resulted in the combination of multiple interventions, as recommended by WHO guidelines for vector control (World Health Organization 2017). Other co-creation studies have reported participant resistance to novel ideas and a preference for conventional approaches (Bovill et al. 2016; Moe et al. 2025) but this was largely absent in our study.

Resistance was only partially observed when interventions increased staff workload; for example, gardening staff initially expressed concern over additional watering required for new garden waiting areas. These concerns were resolved through discussion about the benefits, maintenance planning, and financial compensation. Notably, the gardener subsequently took initiative by planting extra shaded trees beyond the original intervention scope. Co-creation also expanded the reach of the interventions, enabling them to address MBD risks both within and beyond hospital boundaries. For example, discarded plastic and glass containers, which are difficult to dispose of and often become mosquito breeding sites in Zanzibar (Kampango, Furu, Sarath, Haji, Konradsen, Schiøler, Alifrangis, Saleh & Weldon 2021) were repurposed as building materials in our designs (Figure 8). Specific features, such as integrating games into tabletops in these materials, were directly tailored to needs identified during stakeholder workshops. This was made possible through the snowballing of the co-creation network, connecting architects with local material suppliers, entomologists who highlighted the risks posed by improper disposal, and end users who shared essential qualitative insights about gaps in current hospital design.

Figure 8

Collaboration with a diverse array of hospital stakeholders, including recommendations for end-user groups often left out of the intervention development process such as patients (Laurisz et al. 2023) and maintenance staff (World Health Organization 2012), helped capture both technical expertise and valuable lived experience. This supports calls for a more pluralistic integration of knowledge to address climate sensitive challenges, like the increasing burden of MBDs, to ‘co-create solutions that are well adapted to local realities and that are both scientifically sound and socially robust’ (Tschersich & Kok 2022: 363). There was little observed evidence of non-participation due to power hierarchies between participants, a common concern in participatory intervention development in health care and vector control (Dambach et al. 2024; Laurisz et al. 2023; O’Shea et al. 2019). Engagement was observed across participants in groups with mixed backgrounds and reflected in the evaluation feedback. The involvement of end users in this process challenges persistent criticisms of top-down intervention design in vector control (Dambach et al. 2024) and reflects evolving best practice in integrated vector management that prioritises sustained, context-specific and participatory engagement (World Health Organization 2017).

Adopting a living lab approach, in which interventions were developed within the hospitals in collaboration with end users (ENoLL 2025), ensured the hospital environment and day-to-day operations were foregrounded in intervention design. This integration was seen in workshop activities (Table 2) such as the walkthroughs, which encouraged participants to critically reflect on how their environment and practices influence MBD transmission risks. Prolonged stakeholder engagement and continuous process evaluation allowed for iterative improvements in workshop and intervention design, following Moore et al.’s recommendation to embed continuous evaluation within co-creation processes to produce outcomes better aligned with stakeholder opinions and needs (Moore et al. 2015), and enabled earlier intervention stages to be revisited in light of new insights revisited based on new insights (Agnello & Longworth 2022). The comprehensive qualitative feedback captured an extensive data set to inform the intervention design both in this project and future related projects, and could also mitigate risks associated with implementing novel interventions prior to their implementation (Liedtke et al. 2012).

The process facilitated knowledge exchange and capacity-building across traditional silos, as reflected in participant feedback and interviews, changes to practices observed during ongoing visits to the hospitals, and the use of the co-creation groups to exchange co-beneficial knowledge, observed in the co-creation WhatsApp groups. Notably, subsequent renovations in one hospital, led by previous co-creation participants, included the adoption of door brushes and self-closing mechanisms, and waste dumping sites were cleared following a co-creation waste management workshop. This supports the evidence that co-creation fosters capacity-building, empowers stakeholders to take ownership of solutions and expands professional networks (Vargas et al. 2022), and that the co-creation process itself can be seen as an intervention (Leask et al. 2019: 12). Incorporating structural partners as participants, such as governmental engineering departments, creates opportunities for integrating vector control measures into future health care infrastructure in the region, a strategy that is more resource-effective than retrofitting existing buildings (Obonyo et al. 2019).

4.2 Challenges

The study echoed challenges documented in co-creation and living lab projects in public health research, health care and vector control. Practical coordination difficulties arose in aligning busy participant schedules, religious practices, material procurement, and approvals, requiring a high degree of organisational flexibility. The process also carried notable resource demands. Financial support for the MBD-Free Project was one of the most critical enablers of the co-creation process, a challenge common among co-creation projects (Agnello et al. 2025), raising questions about its feasibility without external funders. Longitudinal engagement and repeated workshops were labour- and time-intensive (Dambach et al. 2024). However, continuous hospital visits also helped build rapport and networks, aiding recruitment and strengthening relationships between participants and facilitators. Frequent staffing changes and the large number of employees restricted dissemination among staff and key decision makers, sometimes posing risks to implementation (Brown et al. 2025). The time-intensive demand on participants may have reduced the likelihood of some stakeholders, especially senior decision makers, participating (Vargas et al. 2022).

Securing regulatory approvals for interventions required for health care settings from stakeholders who could not participate in the process reduced the co-creation group’s decision-making power (Gilbert et al. 2025; Martin et al. 2024) and caused project delays. This challenge was partly mitigated through additional meetings, albeit with additional time costs for facilitators. Ensuring the process did not impact clinical care necessitated adaptability; for example, one workshop had to be rescheduled on the day owing to a hospital emergency requiring participating staff. Workshop activities also had to be tailored to avoid impacting patient privacy; for instance, ward modifications were demonstrated in empty rooms or via visual materials when patients were present.

Co-creation and living lab projects working with ecologies and non-human living stakeholders, such as in MBD prevention, recommend the inclusion of specialists like ecologists to represent the interests of nature (Aniche et al. 2024; Diana & Christian 2024), as done in Hospital C. This approach could help highlight environmentally detrimental vector control practices before implementation, a criticism of historic approaches such as oiling water sources and insecticide use (van den Berg 2009).

Co-creation involves distributing decision-making power to a broader group of co-creators, shifting control from research team to other stakeholders and challenging traditional designer researcher roles (Amorim & Ventura 2023; Farrington 2016). Co-creation also creates highly context-specific designs, which may not be easily replicated across different sites. Such context-dependence can create challenges in achieving intervention consistency and translating co-created solutions to new settings. In our study, having hospitals with the same design supported some alignment, but changes in local context and user groups are likely to necessitate different approaches in subsequent rounds of intervention. An in-depth evaluation of the entire co-creation process and intervention evaluation for all three rounds of interventions is planned once the project concludes in 2027.