We conducted this study in Turkana County, Kenya. Turkana has a total area of 77,000 km² (Figure 1). Pastoralists make up the majority of the 1,256,152 people who live in Turkana (Turkana County Government, 2023a). Other livelihoods practiced include fishing, artisanal mining, crop farming, and small-wage labor (Turkana County Government, 2023a). Livestock herds are large and diverse, with an estimated 2.8 million cattle, 6.7 million sheep, 6.9 million goats, close to 1 million camels, 600,000 donkeys, and an estimated 160,000 Indigenous poultry (Griffith et al., 2024; Kenya National Bureau of Statistics, 2015).

The climate in Turkana is classified as arid and semi-arid (ASAL) with mean annual rainfall ranging from 400–600 mm. Vegetation is shaped by rainfall and soil characteristics, consisting of annual grasses and shrubs in the plains and perennial grasses and trees in the highlands (Olang, 1984). Larger rivers that flow into Lake Turkana support acacia and other socio-ecological important species.

Since devolution, service delivery is under the purview of the Turkana County Government, supported by national ministries and NGO development partners (Griffith et al., 2020a). Health and development indicators remain poor: The under-five mortality rate is 65 per 1,000, infant mortality is 39 per 1,000 live births, and global acute malnutrition is at emergency levels of 34.8% (Global Burden of Disease Collaborative, 2024). OH activities are coordinated through the County One Health Unit (COHU) chaired by the health, veterinary, and environment department directors, alongside designated focal persons from each department (Turkana County Government, 2023b).

Social life in Turkana can be characterized by familial structures and mobility patterns. The immediate family makes up the awi (i.e., homestead) that can stay in one place (e.g., in a village or the ere–home area), or join an adakar (ngadakarin pl.). Adakar are groups of households that come together to graze their livestock, for example, migrating to access dry season grazing areas. These groups have designated leaders, often an emuron (seer), ex-chief, or important businessman and council of elders that make decisions (Akabwai, 1992). Adakar will often disband during the rainy season when people and livestock return to their ere. In highly insecure areas along the border, these groups of adakar are referred to as arigan (or arum rum in Turkana South; C. Bode, personal communication, May 7th, 2026) and remain mobile year-round for protection and access to water and pasture that is often across international borders. In this study, we used the term “kraal” to refer to these mobile groups, distinguishing them from villages, which constitute more settled administrative units, consistent with local usage across the county.

This study employed a mixed-methods approach combining participatory FCM and GT qualitative analysis to examine gendered conceptualizations of OH in Turkana County, Kenya. We purposively sampled participants based on defined inclusion criteria. Eligible participants were 18 years or older and involved in daily livestock management and livelihood activities. No restrictions were placed on livestock species, and the range of species participants managed emerged through the mapping process itself.

Data were collected across one village and five kraals spanning three subcounties (Turkana West, Lokichogio, and Loima; Figure 1). We specifically targeted kraals to capture the perspectives of transboundary pastoralists along the border with South Sudan and Uganda. These communities represent some of the most underserved and mobile populations in the county and are the focus of OHIs like Kimormor (Griffith et al., 2023). Nanam village was included to provide comparison with a more settled population and was selected opportunistically as the research team was stationed there during a Kimormor activity.

To ensure representation, we aimed for gender parity across small-group mapping sessions and diversity of age (e.g., elders, adults, youth) where possible. We conducted a total of 15 small-group mapping sessions, consisting of seven women’s groups and eight men’s groups (Table 1). Each group produced one fuzzy cognitive map that was subsequently aggregated into a gendered map for analysis.

We collected data over a six-week period in January and February 2024 through community small-group mapping sessions. Sessions were facilitated in English and Kiswahili, with real-time translation to and from Nga’turkana by trained trilingual field translators. Recordings were subsequently translated into English and transcribed by trilingual research assistants. All participants provided informed verbal consent after the purpose, procedures, voluntary nature, and data handling were explained. Participants were also provided the PI’s and local team members’ contact information for questions or concerns.

Each session lasted approximately an hour and was audio-recorded for accuracy and transcribed for further analysis in Delve (www.delvetool.com).

Ethical approval for human subjects research was obtained through the University of Nairobi Biosafety, Animal Use and Ethics Committee (REF: FVM BAUEC/2023/447), the Kenyan National Commission for Science, Technology, & Innovation (NACOSTI; License No: NACOSTI/P/23/30716), and the Tufts University Human Subject Institutional Review Board (IRB ID 00004210). Prior to each mapping session, all participants were informed about the purpose, procedures, data handling, and confidentiality of the study. Verbal consent was documented and witnessed by a member of the research team.

Figures 3, 4 present the aggregated fuzzy cognitive maps for women and men, respectively. Node size reflects degree centrality (i.e., importance), while arrow width represents edge weight (i.e., strength of perceived relationships). These figures provide a visual overview of system composition and connectivity. Quantitative comparisons of map structure and component importance are presented below.

To examine the gendered conceptualization of OH, we assessed the structural characteristics of each map (Table 2). The aggregated women’s cognitive map contained 66 components and 515 connections, with a density of 0.12 and an average of 7.8 connections per component. Of the 66 components, 32 functioned as driver nodes (outgoing connections only), 0 as receiver nodes (incoming connections only), and 34 as ordinary nodes (both incoming and outgoing connections).

The men’s cognitive map contained 55 components and 410 connections, with a density of 0.14 and an average of 7.5 connections per component. Of the 55 components, 28 were driver nodes, 2 were receivers, and 25 were ordinary nodes. Drivers made up approximately 48.5% of the women’s map and 50.9% of the men’s map. Ordinary nodes accounted for 51.5% and 45.5% of the respective maps.

Women and men shared thirty-eight components across their aggregated cognitive maps. To compare the relative importance of these concepts, we calculated standardized centrality scores (z-scores) for each gender group. Standardization allowed us to control for differences in map size and structure, enabling meaningful comparisons across gender. Figure 5 presents z-scores for all shared components, highlighting differences in the relative centrality of each concept within women’s and men’s maps.

Men and women shared the same top ten most central components, including human health, livestock health, nutrition, animal source foods (ASFs), veterinary services, health services, foodstuffs, pasture resources, income, and water resources (Figure 5; Table 3). However, among these components, women prioritized animal source foods, access to veterinary services, and foodstuffs. In contrast, men prioritized income, water resources, and access to health services (Table 3).

Figure 6 shows the difference in standardized scores (men minus women), illustrating the degree to which each shared component was relatively more (or less) central in one group’s map than the other. Women prioritized labor, animal source foods (ASFs), migration, gender inequity (primarily represented as workload differences), foodstuffs, firewood, drought and education. In contrast, men prioritized human resources, income, water resources, sick animal consumption and gold mining. These divergent priorities are explored in more detail through GT analysis below.

GT analysis revealed how men and women perceive and respond to socioecological conditions, and the health and wellbeing consequences they emphasized. Here, we report four categories that highlight key gendered differences in how men and women conceptualize the OH system.

Women’s larger maps with more connections underscore a more integrated systems perspective rooted in daily labor and care across multiple domains of household, livestock management, and the environment. Their framing of livestock disease through downstream effects on caregiving and household nutrition, for example, reflects not only knowledge of how disease is a significant constraint on milk production (Adane et al., 2012; Griffith et al., 2021; Nthiwa et al., 2019), but also a deeper relational commitment to animal wellbeing–an ethic of care that extends across species (de La Bellacasa, 2017; Yurco, 2024b). This contrasted with how men described OH relationships–for example, talking about livestock as the source of their livelihoods. Interestingly, women placed more importance on education, likely reflecting the role that it plays in livelihood diversification and in some instances, increased access to income and decision-making authority in pastoral communities (Baird et al., 2024).

Men’s smaller, denser map, in contrast, reflects a more governance and risk-based understanding of OH, organized around landscape dynamics, access to natural resources, and control of livestock–for example, linking rainfall to livestock density and disease transmission (VanderWaal et al., 2017). Anbacha and Kjosavik (2021) similarly found among Borana pastoralists in Ethiopia that men give more attention to governance and conflict stressors due to their traditional roles and responsibilities. It is worth noting, however, that men’s herding decisions are ultimately driven by livestock nutrition and health needs–a form of traditional ecological knowledge that shapes mobility patterns and resource governance (Sharifian et al., 2023). Nevertheless, men’s framing in our study operated at the level of herd and resource management rather than the household and interspecies relationships emphasized by women.

Insecurity further illustrates these gendered perspectives. While men described conflict as shaping livestock mobility and resource governance decisions, women experienced it as compounding their caregiving burdens by restricting access to water, wild fruits, and the loss of donkeys (Pike, 2019; Schilling and Werland, 2023).

Women in our study demonstrated extensive knowledge of livestock diseases encountered through daily caregiving. However, similar to previous studies, we found that women had less knowledge of zoonotic diseases compared to men (Babo et al., 2022; Tschopp and Kidanu, 2024). Men also placed much higher importance on the risk from consuming sick animals. Interestingly, in a previous participatory epidemiology study in Turkana using mixed-gender focus groups, zoonotic diseases were identified as an important community concern (Griffith et al., 2024), suggesting that group composition and social dynamics may influence what knowledge is expressed in participatory settings.

This difference reflects situated knowledge (Haraway, 1988), which argues that all knowledge emerges from embodied, material engagement with the world. Men’s routine engagement with formal OH institutions, likely exposes them to zoonotic disease information, and knowledge transfer from veterinarians in pastoral communities has been shown to primarily target men (Babo et al., 2022). Women in pastoralist communities, and more broadly across Kenya (Kyotos et al., 2022), face additional barriers in accessing animal health services and information due to gendered power dynamics and social norms, including dependence on men’s permission and restrictions on interactions with male service providers (Kamel et al., 2025; Mutambo et al., 2025). In a companion study, we found that women CADRs and community health promoters (CHPs), i.e., frontline animal and human health workers, demonstrated in-depth knowledge of zoonotic diseases and prevention methods, supporting that institutional access and training, not capacity, explains this knowledge gap (Griffith et al., 2026b).

Pastoralist women’s exclusion from formal and customary governance spaces for rangeland management, mobility decisions, and resource allocation (Jode and Flintan, 2020; Kristjanson et al., 2014; Yurco, 2024b) is further reflected in Turkana’s institutions. Customary decision-making is mediated through male-dominated institutions such as the Tree of Men (Turkana County Government, 2024), while men overwhelmingly fill political, administrative, and professional roles in formal governance (International Labour Organization, 2022). This illustrates how structural exclusion from governance spaces renders women’s extensive care-based knowledge largely invisible within OH frameworks–a form of epistemic oppression in which marginalized groups’ knowledge is systematically devalued regardless of its relevance or quality (Dotson, 2014). As a result, men’s focus on zoonotic disease risk aligns with the dominant OH narrative in Turkana and Kenya more broadly (Turkana County Government, 2023b; Zoonotic Disease Unit, 2021). This reflects how institutional power shapes whose knowledge becomes legible within OH (Weiszhar et al., 2025). It is also worth noting that men may have emphasized these factors partly in response to the research context, perceiving the veterinary and OH background of the research team as signaling an interest in zoonoses. This dynamic itself reflects men’s familiarity with veterinary and public health professionals.

In order to overcome this gap, increasing women’s participation in formal and customary governance structures to at least 50% is an important first step (African Union, 2019), but these efforts must be grounded within collective pastoral institutions to avoid unintentionally reinforcing marginalization (Jode and Flintan, 2020). Participatory rangeland management, multistakeholder innovation platforms, and community conversations from the HEAL project offer promising models for pastoral areas–creating structured, legitimate spaces where women’s priorities can shape OH agendas (Flintan and Eba, 2023; Mumin et al., 2026). Training women frontline workers and closing the gender professional gap in pastoral areas would further position women as legitimate OH knowers and decision-makers.

OH has been criticized for being too reductionist and biomedically focused, neglecting the environment and failing to consider political, cultural, social, and economic factors that influence multispecies health and wellbeing (Braam et al., 2024; Davis and Sharp, 2020; Meisner et al., 2024; Van Patter et al., 2023). Our findings demonstrate that pastoralists already hold a deeply holistic understanding of health that encompasses not only socioecological linkages but also a relational duty of care for the health of the whole system (Wingett et al., 2025), underscoring why a community-based approach is essential (Puppo et al., 2025).

Women’s perspectives–what we term care-based OH–would represent a paradigm shift: rather than focusing solely on biomedical and surveillance systems in pastoral contexts, it would target everyday relational and material conditions that sustain health and wellbeing across people, animals, and environments (Kaiser-Grolimund et al., 2025). Critically, this would include women’s own physical health as a direct consequence of their caregiving labor–a dimension that is largely absent from current OH frameworks. When gender and maternal health are recognized within OH, they are typically framed around access to health services or infectious disease risk rather than the physical toll of daily and seasonal labor (Cataldo et al., 2023; Griffith et al., 2023; Mor et al., 2024). The health and wellbeing consequences of caregiving labor have received more attention in the resilience literature (Ash et al., 2025; Fernández-Giménez et al., 2025; Walker et al., 2022), but care-based OH would go further and actively target these burdens as upstream determinants of health.

In practice, this would mean reducing the physical burden of water collection and other daily tasks, supporting access to veterinary and health services as part of household caregiving responsibilities and animal welfare, and mitigating the labor, nutritional, and safety challenges that accompany drought and migration. Prioritizing donkey welfare and health would simultaneously reduce women’s physical burdens, expand economic opportunities, and improve multispecies wellbeing (Johnston, 2026; Valette, 2014).

Men’s governance and risk perspective complements women’s care-based framing by articulating key SES relationships across conflict, livestock mobility, and service delivery. Excluding either perspective would produce an incomplete understanding of how the OH system functions. Therefore, gender is not only an equity concern but a fundamental analytical entry point for OH research and effective practice (Galiè et al., 2024; Garnier et al., 2025).

Reflexivity is defined as “a set of continuous, collaborative, and multifaceted practices through which researchers self-consciously critique, appraise, and evaluate how their subjectivity and context influence the research process” (Olmos-Vega et al., 2023). Central to this process is positionality–the ways in which the researcher’s social identity, worldview, and assumptions shape how knowledge is produced and interpreted (Holmes, 2020).

This study is shaped by the positionality of the first author (EFG), a veterinarian and white male researcher from the Global North. His training in veterinary science and public health reflects a largely biomedical orientation, which may have subtly guided aspects of the FCM process, for example, by focus on infectious disease pathways and prioritizing structural over relational components of OH.

Knowledge co-production was also shaped by how participants and researchers perceived each other. Community members referred to EFG as an mzungu (i.e., a white person of European descent), a Kiswahili word that emerged through colonial encounters and today carries postcolonial meanings tied to the humanitarian and development state (Christian and Namaganda, 2023). Many initially assumed he was affiliated with an NGO or faith-based organization and might provide material assistance. Such expectations can create a dynamic where participants feel pressure to provide “correct” answers or emphasize needs in hopes of further support. We attempted to mitigate this through clear informed consent procedures, while members of the research team who work for the county government could help channel concerns appropriately.

Several participants also expressed broader skepticism toward research, with one woman stating: “Now that he is here, he should help us. Not just asking my problems and going back” (MS 15). This reflects well-documented forms of research fatigue in marginalized communities when research does not lead to tangible change (Patel et al., 2020). In response to early feedback, particularly from women asking why we focused mainly on challenges, we revised our facilitation guide to explicitly incorporate community-driven solutions.

Despite these complexities, several participants described the process as meaningful, emphasizing learning and dialogue rather than extraction. As one group of young women noted, “We are grateful for this occasion as we have learned a lot from you. We have not gone to school, but we are learning from you and the questions you have asked” (MS 09). These reflections underscore both the ethical responsibilities and the potential of participatory research to foster mutual learning, even within asymmetrical power relations.

This study has several limitations that should be acknowledged. First, it focused on gender differences but did not systematically examine how gender intersects with other social identities, including age, marital status, and religion that shape experiences and access to resources and power (i.e., intersectionality; Crenshaw, 2013), nor did it explicitly assess gendered livestock species ownership and management, although patterns were largely inferred from what men and women described during the mapping sessions. The study also did not capture intra-household dynamics, as group mapping of the broader system rather than the household was our unit of analysis. Although GT provided important contextual detail, ethnographic or formal gender analysis tools may be better suited to capture both intra-household dynamics and livestock ownership patterns in future studies.

Second, our study did not evaluate how county governments, NGOs or customary institutions include women’s contributions in practice. Future research should examine the institutional conditions, including in OH planning and implementation, under which women’s knowledge is not only heard but taken seriously. This includes testing approaches that foster gender-transformative change rather than symbolic participation.

Third, time and resource constraints meant we could not present the findings to participants, which would have helped to verify the findings and represents an important next step. Fourth, FCM studies ideally continue mapping sessions until concept saturation is reached—the point at which new sessions no longer produce novel concepts. Sarmiento et al. (2024a) recommend 12–15 maps to reach saturation, which we were unable to achieve due to time and resource constraints.

Finally, this study focused on pastoralist communities and did not include other livelihood groups present in Turkana County, such as fishing communities, or peri-urban/urban populations, which limits the representativeness of findings across the county’s diverse population. More broadly, findings are specific to Turkana’s pastoral context and comparative research examining gendered OH perspectives across pastoral communities in East Africa and beyond would strengthen their generalizability.

We found that pastoralist men and women conceptualize OH through distinct, yet complementary pathways shaped by gendered labor, institutional access, and power relations. Women’s care-based, relational understanding of OH highlights socioecological dynamics that are often overlooked in dominant, technocratic OH approaches. Our integrated FCM-GT approach makes these forms of knowledge visible and actionable. Embedding care-based perspectives within OH policy and practice is therefore essential for developing interventions that are not only equitable, but better align with pastoral life in the rangelands.