Environment as an enabling space for open defecation in pastoral communities: a mixed-methods study in Marsabit, Kenya

¹Sanitation Research Institute, Meru University of Science and Technology, Meru, Kenya, ²Department of Business studies, University of Embu, Embu, Kenya

^&Corresponding author
Grace Kasiva Eliud, Sanitation Research Institute, Meru University of Science and Technology, Meru, Kenya

Introduction    

Access to sanitation facilities is an important human right and a critical goal in the United Nations development agenda [1], yet millions of people globally still lack such a necessity. Despite global efforts towards achieving safe sanitation and the elimination of open defecation, a report by [2] showed that 43% of the global population accesses unsafe sanitation, including 419 million people who defecate in the open, with distinct disparities between developed and developing countries. Defecation in the open has been associated with reduced dignity and poor health outcomes among the affected communities [3]. However, the open defecation practice remains a significant challenge among pastoral communities, and the negative trends have been maintained for years despite the adoption of behavior change approaches such as community-led total sanitation. The community-led total sanitation (CLTS) approach has been widely used in generating demand for and utilization of latrines to stop open defecation and has been embraced by many countries as a national sanitation strategy [4]. The approach was developed by [5] and ignites a collective community-driven response to encourage toilet construction and attainment of open-defecation-free status in rural areas. In CLTS, communities are prompted through a series of mind triggers to analyze their sanitation status and to understand the consequences of open defecation, which encourages them to adopt toilets and become open defecation-free through stimulation of fear, shame, and disgust [6]. Since its inception in Bangladesh [5], CLTS has significantly spread across many regions in the Pacific, America, Asia, and Africa [7] and has often been used as a behavior change tool by rural sanitation development organizations [4]. However, the approach does not work equally well in every community [8], and sustaining improvements in rural sanitation has been challenging for reasons tied to specific contexts that are region-specific. Globally, Kenya has been ranked among the 26 countries contributing to high open defecation rates in the world, with 4.7 million people defecating in the open [9-10].

Although the rates are low in Kenyan counties like Siaya [11], Kitui, and Makueni [10-12], 85% of open defecation cases in Kenya are contributed by 16 counties, 10 from pastoral communities, including Marsabit [13], despite the rollout of CLTS in the counties. The figures suggest that CLTS efforts, particularly in pastoral communities, have been less impactful, demonstrating a trend of questionable endeavors at Kenyan county levels. The environmental context for pastoral communities could promote conditions that unintentionally dampen the spirit of latrine construction and promote open defecation. Addressing sanitation through the environmental context lens has, however, not been given the attention it deserves, especially in pastoral regions. Regions in pastoral communities are most often characterized by vast open land, mobility, temporary settlements, varying soil types, climatic conditions, and sparse population. In Mozambique, a study by [14] established that sanitation transformation has been quite challenging in rocky areas as communities exhibit lower latrine construction behaviours due to varying difficulties in ground excavation. Construction of latrines in such difficult-to-dig areas could require specialized tools and extra labour, increasing construction costs which communities may not be willing to incur. Zambia´s latrines built on sandy soils are prone to collapse, especially during heavy rains [15]. Frequent collapsing could make latrine rebuilding more costly, laborious and time-consuming. Soil types are important factors which sanitation programmers need to account for if durability and high performance of latrines adopted after CLTS is to be realized. The performance of latrine facilities could be highly impacted by climate issues, whose potential negative contribution to individuals´health outcomes has attracted high global attention [16]. The effects of climate change range from impairment of sanitation infrastructure, water and environmental contamination, and flotation of faeces in latrine pits in case of floods, which consequently pose public health risks to communities [17].

Floods could damage weak sanitation facilities built on loose soils, fill them up and sweep the faecally contaminated water to environments, triggering outbreaks as was the case in Zimbabwe [18]. Availability of open land has been reported to influence acceptability of open defecation as a convenient practice, thus reducing motivation for latrine construction. Areas endowed with forests or open fields have been shown to offer a disabling environment for latrine construction [19]. In Tanzanian pastoral lands, a study by [20] established that forests surrounding pastoral homesteads offered a chance for residents to practice open defecation, as the forests provided sufficient privacy and cover. Residents in remote and inaccessible areas suffer multiple potential sanitation-related disadvantages due to challenges in identifying appropriate services and systems delivery mechanisms [21]. Long geographical distances between households have been reported to strain facilitators´ walking to implementation sites [20]. A study by [22] in Washington, DC found that remoteness made logistical and operational costs related to follow-ups after CLTS prohibitive. A similar report by [21] among Ethiopian and Kenyan pastoral communities found out that in extremely distant communities, communication and transport networks were unreliable, limiting community triggering and monitoring practices for eradication of open defecation after CLTS. When operations are costly or strenuous, CLTS implementers might ignore the difficult-to-reach regions, leaving their sanitation status helplessly poor. Given that environmental conditions differ with regions, this study aimed at examining the influence of environmental context on open defecation among Marsabit pastoral communities. It aimed to understand how environmental factors such as drought, floods, soil type, proximity to bushes or forests, and accessibility influenced adoption of CLTS.

Methods     

Study design and setting: the study used a cross-sectional descriptive survey design with a mixed methods approach using both qualitative and quantitative methods to collect data in two sub-counties in Marsabit, namely Saku and North Horr, which had the lowest and the highest open defecation rates [10-23], respectively, for comparison. The county was considered because it is among the pastoral communities contributing to increased national open defecation rates (KNBS, 2019).

Definition of study variables: environment refers to conditions and forces of the surrounding that affect behavior; pastoralists are people whose livelihood is centered on livestock keeping; enabling space is herein used to mean conditions that support behaviors such as open defecation. As defined by [24], open defecation is the human practice of leaving feces in fields, bushes, water bodies, or open land. The key variables used in the study are comprehensively defined in Table 1.

Study population: the study targeted residents of two purposively selected sub-counties in Marsabit County with the highest and lowest open defecation rates, namely, North Horr (Population= 71,445; households = 14,289) and Saku (Population = 79,170; households = 15,834). The two sub-counties were selected because the rates of open defection were distinctively different, with the highest being North Horr at 75% and the lowest, Saku, at 18.9% [10] for a comparison of CLTS-related factors. The study population included household heads, community leaders, government health officials, and implementing partners.

Sampling: the study used a random sample of 374 household heads, out of the intended 395 samples calculated using the [25] formula,

Data collection: household data was collected using a questionnaire and observation checklist. Qualitative data were collected from 2 focus group discussions consisting of CHPs and household community leaders based on sub-counties. Qualitative data were also collected from interviews with key sanitation experts and implementers. The data was audio recorded and notes were taken to supplement the audio.

Statistical analysis: quantitative data were analysed using descriptive and inferential statistics using the Statistical Package for the Social Sciences (SPSS) version 26. Logistic regression analysis was performed to establish the relationship between environmental factors and open defecation. Univariable analysis was performed to establish the influence of each independent variable separately on the dependent variable, and results were presented as unadjusted odds ratios with 95% confidence intervals (CI). Further, a multivariable analysis was conducted to establish the joint relationship of all the independent variables combined on open defecation, and findings were given as adjusted odds ratios at the 95% CI. The variables with a p-value less than or equal to 0.05 were regarded as significant contributors to open defecation. Audio from the recorded interviews and discussions was transcribed to generate text. The transcribed data and field notes were thematically analysed using the MAX Qualitative Data Analysis (MAXQDA) software as per [26] guidelines.

Ethical considerations: participation in the study was on a voluntary basis, and informed consent was given by participants to ascertain their willingness to take part in the study. A research permit was obtained from the Meru University of Science and Technology Institutional Research Ethics Review Committee (MIRERC) (Reference number: MU/1/39/28Vol.3 (189) and a data collection license was obtained from the National Commission for Science, Technology and Innovation (NACOSTI) [Reference number: NACOSTI/P/25/4173684]. Anonymity was observed to ensure that participants´ real identities were concealed and the data gathered was treated with utmost confidentiality to protect it from leaking to third unintended parties.

Results     

General characteristics of participants: participants´ mean age was 43 years, standard deviation = 1.053, with those aged 38-47 years accounting for the highest percentage. Overall, females formed the largest proportion of the respondents. However, participants´ gender varied by sub-county, with 66.66% (118/180) of the participants engaged from North Horr being females compared to 59% (114/194) of those engaged from Saku sub-county. Christianity and Islam were the most common religions, accounting for 41.1% and 38.8%, respectively.

Adoption of community-led total sanitation: the study showed that the possibility of finding openly defecated fecal matter in North Horr sub-county was 3.523 times higher than in Saku (95% CI: 2.294-5.410, p=0.000). A comparison of CLTS implementation status revealed that the certification rate of villages as open-defecation-free was higher in Saku (50%) than in North Horr Sub-County (28%), despite having been triggered. The majority of the villages (69%) had been triggered in North Horr, with no progression to ODF status, compared to Saku (47%), indicating that CLTS efforts had been made for the sub-counties, but progress varied.

Drought on adoption of community-led total sanitation: findings showed that drought encounters had a high mean score for both sub-counties but with slightly more encounters reported in North Horr (Mean=4.51, SD=0.68) compared to Saku (Mean=4.34, SD=1.10). In the model showing unadjusted odds (Table 3), the likelihood of openly defecated fecal matter in drought-prone areas was significantly higher than in areas that did not experience drought (OR 1.151, 95% CI: 1.035-1.280, p=0.010).

Floods and adoption of community-led total sanitation: regarding the influence of floods, the majority of the respondents (79, 43.89%) from North Horr sub-county described it as having severe consequences compared to 53 (27.32%) from Saku sub-county. The influence of floods as well as land ownership on open defecation was statistically non-significant (p>0.05).

Soil type: North Horr sub-county´s types of soil were predominantly sandy (110, equivalent to 61.11%) or rocky (58, equivalent to 32.22%), with only a few areas characterized by clay or loam soil. However, for Saku, the composition of soil was mostly clay (106=54.64%) or loamy (79=40.72%), and very few areas had rocky or sandy soils. The unadjusted model showed that the type of soil remained statistically significant, with the possibility of finding openly left faecal matter (p<0.05) due to the varying difficulties of latrine construction and sustainability. Excavation of pits in rocky soils was described as most difficult and requiring improved tools and specialized labor, which could probably increase the effort or cost of constructing latrines, as reported: "It is hard and rocky where you cannot even dig until you get an excavator.ʺ(Household P5~FGD) "Like this is sand. No latrine can stand here for long. When rain comes, it will just go or sink.ʺ(Household P 4~FGD)

Presence of forests and bushes: forests, bushes, or open bare lands acted as convenient places for open defecation. Results from the unadjusted model with the odds of finding openly left faecal matter showed that the possibility was 1.333 times higher in areas where households resided near forests than in areas without forests, and the relationship was statistically significant (95% CI: 1.118-1.589, p=0.001).

Type of engagement and access to latrines while away from dwellings: when requested to indicate the activities they performed while away from households, herding cattle dominated in North Horr (67.22%). Crop farming was almost non-existent in North Horr, demonstrating that the land was quite arid. A more diversified pattern of activities, including looking for firewood, fetching water, farming, and business, among others, was noted in Saku sub-county. It was, however, observed that in both sub-counties, livestock keeping was predominant and the herders were away in the grazing lands. Distance to latrines was established as a major concern for the participants from both sub-counties, but of greater concern in North Horr compared to Saku sub-county. A higher response of 77 (42.78%) and 76 (42.22%) participants from North Horr sub-county indicated that when they had gone away from their households to perform the activities they had mentioned, latrines were completely inaccessible or not easily accessible compared to 22 (11.34) and 65 (33.51) in Saku sub-county. The odds of finding openly left faecal matter were 37.4% lower among people who could access latrines while away from households doing other activities than for those who had no access (OR 0.626, 95% CI: 0.551-0.710, p=0.000). In the model showing adjusted odds, all the variables that were significant in the unadjusted model such as type of soil; p<0.05, accessibility (adjusted OR=0.747, 95% CI: 0.642-0.869, p=0.000), drought (OR=1.333, 95% CI: 1.106-1.605, p=0.002) and proximity to forests (OR=1.642, 95% CI: 1.268-2.126, p=0.000) showed a statistically significant relationship with open defecation. Further, variables like floods and land ownership, which appeared statistically non-significant in the unadjusted model, were statistically significant in the adjusted model. Residents in areas which encountered severe floods had 1.250 higher odds of leaving faeces in the open compared to their counterparts. Land ownership was seen to lower the possibility of finding openly left faecal matter in the area (adjusted OR=0.660, 95% CI: 0.556-0.783, p=0.000).

Other findings from the qualitative study: Figure 1 summarises the themes that emerged on the relationship between open defecation and environmental context and how the practice relates to environmental contamination and disease exposure. This section outlines other environmental factors that facilitated the practice of open defecation in the area.

Aridity and lack of tools for digging latrine pits: while Saku sub-county´s land would support agriculture, a significant portion of North Horr had a dry environment which least favored crop farming and discouraged the need to purchase farming tools which could be used to dig pits for latrines, as indicated by a participant who said: "There are no jembes, it is so dry. No one goes to the farm so we don´;t have anything to dig toilets with. Unless the doctor gives us a way forward. ʺ(Household P4- FGD) Another interviewee reported that:"They will just do a small pit, probably using local equipment. Those cannot dig deep; they don´t have jembes. ʺ(Health Officer 2-In-depth interviews)

Strong winds, latrine collapse and open defecation: stronger winds were mostly observed in North Horr sub-county, which were most concentrated in areas near known deserts such as Chalbi and other arid parts of the sub-county such as Turbi. In (Figure 2 A, B) show latrines whose superstructures had been destroyed by strong winds in Turbi. The evidence of wind destruction was shown by the torn and damaged design of the iron sheets from latrines´ superstructures. The influence of the forces of nature on sanitation was emphasized by respondents in the qualitative study as follows: "A lot of wind affects also the latrines get washed off.ʺ(Household P4~FGD) "In this desert, you noticed the wind could remove your clothes, especially in Turbi. That is what it does to latrines. See, the one which is fallen there. It is wind!ʺ(CHA~1 In-depth interviews)

Effect of heat and increased temperatures on community-led total sanitation adoption: increased temperatures, observed to be highest in North Horr compared to Saku sub-counties affected CLTS adoption. The hot and arid environment was reportedly causing faeces to dry quickly, which reduced smell and visible contamination by flies, thus lowering perceived sanitation risks. One respondent noted that: "After 10 minutes the faeces dry up, the temperatures are very high in North Horr compared to Saku.ʺ(CHA 1~In-depth interviews) Another respondent explained that: "It dries up so quickly. If you do CLTS, the experiments, when it is dry, it will not have the expected impact. It is not smelling in the first place, and you know the target is smell and flies.ʺ(Health officer 2~In-depth interviews)

Land vastness, logistical challenges and open defecation: the land in North Horr was vast and had a smaller population compared to Saku; the unoccupied areas acted as convenient places for open defecation. Participants in the interviews highlighted that: "The land is vast one person can defecate here, another person there.ʺ(Household P 3~FGD) "I will show you a bush that people know is for open defecation. You will see for yourself. Most of them have no latrines, so they go there.This bush is easy for open defecation, because this one will defecate here and another place, so this one favours.ʺ(CHP 1~Indepth interviews) On the other hand, bushes were sources of materials for latrine construction, which suggested that Saku sub-county had unlimited choices of latrine construction materials after CLTS compared to North Horr, as noted by a respondent that: "Those people even if they dug a latrine, will they have the trees to cover the latrines? It is hard to get them in North Horrʺ(CHA 1~In-depth Interviews). CLTS implementation activities that discouraged open defecation were affected by logistical challenges due to unreachable households. This was confirmed in the qualitative study where a respondent indicated that implementers would concentrate on the easy-to-reach areas and leave the areas that were far from towns. "Going to the nearest town from here will cost you KSh. 2000 because you will have to go on a motorbike. There are no vehicles.ʺ(NGO 1~In-depth interviews) "NGOs will implement CLTS in areas which are near towns, reachable areas. Inside there is very far; there are no good roads.ʺ(CHA 1~In-depth interviews) "Most of Saku´s areas are near the town; that is why you can see the difference. Areas, even in Saku, that are far from towns, and you know North Horr is worse, are rarely reached. Some places are also dangerous; one can be attacked.ʺ(NGO 2~Indepth interviews)

Role of waterways: it was observed that there was a presence of various waterways formed during heavy rains, either along roads, away from or in forests, locally termed as laggas, which guaranteed privacy for open defecation. A participant in the qualitative study noted as follows:"That is a lagga, a waterway, you see they are so many. They are everywhere. Open defecation happens there. See, this is not even very dry, it hides them from being seen.ʺ(CHP 2~ In-depth interviews) Another respondent indicated that: "That water from laggas goes to rivers. It is contaminated with the faeces that people leave there. Those are hiding places; they hide you from being seen.ʺ(Household P4~FGD)

Rocks as hiding places and enablers of open defecation: where there were no bushes, the environment, especially in North Horr sub-county, still seemed to be convenient for open defecation due to the presence of big rocks where people could hide and defecate (Figure 3). Participants from the qualitative study noted that: "The rocks, you see the rocks, they are even hot and favors anal cleansing. I mean the environment really encourages the practices.ʺ(Health officer 2~In-depth interviews) "There is no latrine for use on your way to North Horr. The vehicles have to stop over so that passengers can go to the ´latrine´. There is no latrine; you will jump out, you know, a Land Cruiser; we hide behind rocks and do our thing. The rocks are big; no one can see you.ʺ(Household P3~FGD)

Discussion     

Findings on mean age of 43 was a result of the engagement of the young people and the old in rearing and protecting cattle away from households. Females were mostly found at the households doing household chores because males were in the grazing fields, given the fact that the areas were pastoral lands. The study identified environmental aspects that acted as enabling conditions for open defecation in the pastoral communities in Marsabit, Kenya. The environmental conditions seemed to be more favourable in Saku sub-county compared to North Horr sub-county, which promoted the varying possibilities of open defecation in the areas. The factors enabling open defecation spanned from climatic conditions such as drought, floods, heat, and strong winds, soil types, open bare land, land ownership, rocks, and waterways. Rainfall in the pastoral regions was often unpredictable, leading to vulnerability to drought. Failed rainfall and drought promoted migration in areas without latrines in search of water and pasture for livestock. Drought episodes were associated with loss of livestock and food shortages, which shifted the communities´ attention from constructing latrines to meeting basic needs and caring for their livestock. Further, prolonged periods of drought least favoured agriculture and consequently the need to purchase farming tools such as jembes, which in most cases would be the tools for latrine excavation. Hiring people with excavation tools was an alternative, which could have been costly for such a community that was struggling to survive. A previous study in Kajiado, Kenya, by [27] related lack of excavation tools to low latrine construction, thus increasing open defecation among the Maasai pastoral communities. Floods damaged latrines or even made them inaccessible and unusable, or avoided because of the risk of sliding or falling in pits. Ideally, CLTS champions for adoption of even the simplest toilets, whose vulnerability to flood impacts was high. Destruction of latrines by floods could suggest the need for frequent reconstruction, which could be very expensive, especially for low-income communities.

The findings were in line with results obtained by [28], who attested to the influence of floods on reducing the uptake of CLTS in Mozambique, and by [29] in Burkina Faso, where sustainability of latrines remained low in flood-prone areas. The type of soil affected the durability of latrines and relapse to open defecation. Rocky soils were difficult to excavate using simple tools, while sandy soils were too loose to support strong latrines (Figure 4). The findings concurred with results obtained from Zambia by [30]. Winds carried away latrines since, in deserts or arid areas, there were no trees to block the destructive impact. The results implied that use of very local materials for latrine construction, as mostly emphasized in CLTS, would be impractical in such areas due to the influence of the winds. Strong winds could have serious implications on sustained latrine ownership and use, which extend beyond damaged superstructures to reduced safety and privacy of users. People could avoid using the facilities in fear of getting hurt by latrine structures blown by wind, and of being suddenly seen inside a latrine after the wind has blown it. Previous researchers in Malawi [31] and in Burkina Faso [29] have also acknowledged the impact of strong winds on latrine sustainability and open defecation. Community-led total sanitation strategies on disgust and shame particularly triggered through exercises that demonstrate the movement of flies from faeces to food. The rapid drying impact of increased heat on faecal matter reduced moisture and smell that ideally would attract flies, which are critical sensory and observable cues when implementing CLTS. Dry faeces therefore produced minimal smell and attracted fewer or no flies as expected during triggering exercises, which weakened the intended emotional impact of CLTS on encouraging communities to stop defecating in the open.

A study by [32] in Ethiopia also noted similar findings. Such findings signified the need for adapted strategies of shame-dependent interventions to promote their effectiveness in regions characterized by high temperatures. Bushes, waterways (laggas), and rocks acted as convenient spaces that provided natural privacy for open defecation. Long stretches without public latrines for use by passengers contributed to helpless open defecation. The main economic activity for most households was cattle rearing, which encouraged open defecation in the grazing lands. Engagement in other activities such as looking for water or firewood also encouraged open defecation given the absence of sanitation facilities in the points. The reason behind low accessibility of households in North Horr compared to Saku was that in North Horr, households were sparsely distributed due to the vastness of the land, as opposed to Saku, where most of the communities were densely concentrated in certain areas, not far from each other. A similar study in Washington DC by [22] established that low accessibility of households demoralized implementers of CLTS from visiting the households for mobilization, monitoring, or follow-up. Implementers could be demoralized due to the need to incur high costs going to the places. The findings therefore signaled that CLTS implementation activities that aim at discouraging open defecation were skewed as a result of logistical challenges due to unreachable households.

Conclusion     

Performance of CLTS in pastoral communities is limited by environmental aspects that create a convenient space for open defecation and make latrine construction difficult. Implementers of CLTS and practitioners should adapt triggering techniques to account for challenges related to the environment as opposed to assuming one-size-fits-all approaches. Further, there is a need for implementers to work closer with researchers to co-design operational behavior change frameworks that sustainably address sanitation and promote open-defecation-free pastoral communities. In addition, the county government of Marsabit, in collaboration with the Ministry of Health, should provide public latrines in major stopovers to minimize open defecation cases during transit.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Idea conceptualization and study design: Grace Kasiva Eliud, Lilian Mukiri Kirimi, Kirema Nkanata Mburugu, Robert Gitunga Muriungi. Data collection, analysis and and final writeup: Grace Kasiva Eliud, Lilian Mukiri Kirimi, Kirema Nkanata Mburugu, Robert Gitunga Muriungi. Critical review and manuscript editing: Grace Kasiva Eliud, Lilian Mukiri Kirimi. Planning and execution of the study: Grace Kasiva Eliud, Lilian Mukiri Kirimi, Kirema Nkanata Mburugu, Robert Gitunga Muriungi. Supervision and validation: Lilian Mukiri Kirimi, Kirema Nkanata Mburugu, Robert Gitunga Muriungi. All the authors have read and agreed to the final manuscript.

Acknowledgments     

Sincere appreciation to the study guides and the team who contributed to the success of the data collection exercise.

Tables and figures     

Table 1: definition of variables used in the study conducted in Marsabit, Kenya

Table 2: distribution of study participants recruited from the study in North Horr and Saku sub-counties, Marsabit in 2025 (N=39 for qualitative data and 374 for quantitative data)

Table 3: odds of open defecation from logistic analysis on the data obtained from in Marsabit (n=374)

Figure 1: themes that emerged from the qualitative study conducted among purposively selected interviewees from North Horr and Saku sub-counties, Marsabit in 2025, N=39

Figure 2: latrines destroyed by strong winds in North Horr sub-county in Marsabit; photos taken in Bubisa-Turbi Ward, North Horr Sub-County, marsabit by: Researchers (2025)

Figure 3: rocks used as open defecation sites captured in a stopover in North Horr sub-county, Marsabit: photo taken in North Horr by: Researchers (2025); A) shows rocks used for open defecation during transit in North Horr, where there were no latrines; B) shows a vehicle carrying passengers, which had made a stop for passengers to relieve themselves in places without latrines

Figure 4: latrine constructed on loose sandy soil in Marsabit; photo taken in North Horr sub-county by: Researchers (2025)

References