Geographical distribution of cholera outbreaks in Kenya from 2015 to 2023: a comparative analysis

¹Department of Community Health, Epidemiology and Biostatistics, School of Public Health, Mount Kenya University, Thika, Kenya

^&Corresponding author
Waqo Gufu Boru, Department of Community Health, Epidemiology and Biostatistics, School of Public Health, Mount Kenya University, Thika, Kenya

Introduction    

Cholera has persisted as a critical global public health problem with its historical trajectory marked by six major pandemics resulting in millions of fatalities since the 19^th century. The disease continues to inflict a massive global burden, accounting for an estimated 1.3 to 4 million cases and 21,000 to 143,000 deaths annually [1]. Evidence demonstrates that this burden is unevenly distributed, disproportionately devastating vulnerable populations trapped in cycles of conflict, insecurity, poverty, and climate-induced disruptions [2]. Since 2021, global cholera incidence has surged dramatically, with Africa being significantly affected by the resurgence, accounting for 90% of all the reported cases in 2023 and experiencing a significant outbreak in 2024 with nearly 400,000 cases and over 7,000 fatalities [3,4]. In low- and middle-income countries such as Kenya, a key challenge in mitigating recurrent outbreaks is the lack of comprehensive hotspot mapping, which hinders targeted interventions and resource allocation [5].

Climate change profoundly compounds the challenges of disease eradication through events such as drought and floods, which trigger cholera outbreaks by destroying WASH infrastructure and contaminating water sources [6]. This ecological destruction prompts mass human displacement and sharp spikes in the transmission risk of cholera, as seen in the Horn of Africa and Malawi whereby over 36 million people were affected by drought, hence increasing the number of climate refugees [3,7]. Additionally, violence and political instability worsen these outbreaks by reducing access to clean water and impeding disease control and management efforts by government and NGOs [2,7].

In Kenya, Cholera is a significant public health challenge and is one of the diseases prioritized under the Kenya Integrated Disease Surveillance and Response (IDSR) strategy. Since the first case was detected in 1971 in the Turkana district, the country has battled cyclic and periodic epidemics that typically resurface every five to seven years [8]. National surveillance data from the Kenya Ministry of Health reveals a pattern of continuous annual outbreaks characterized by a major peak in 2015, which declined in 2016, followed by an increase from 2017 to 2019 [5,9]. These fluctuations capture the highly variable spatial and temporal burden of the disease. Recognizing this volatility, the Global Task Force on Cholera Control (GTFCC) emphasizes robust surveillance and case reporting as foundational pillars for successful disease elimination [10]. Despite these efforts, Kenya's most recent national hotspot mapping was based on surveillance data spanning only 2015 to 2019, which excludes the prolonged 2022 to 2024 outbreak period. Furthermore, existing studies report limited comparative analysis of how the geographical distribution of outbreaks has shifted across counties over time, which constrains the health sector in reprioritizing interventions [5]. Given that surveillance data serves as the baseline for strategic decision-making and the prioritization of targeted health interventions, this study sought to utilize epidemiological data from 2015 to 2023 to map high-risk cholera hotspots across selected counties in Kenya.

Methods     

Study design: this study adopted a retrospective study design. It assessed cholera surveillance data from 2015 to 2023 and the sociodemographic factors associated with cholera cases.

Study area: this research was conducted in Kenya.

Sample size determination and sampling technique: census sampling was employed in this study, including all reported cholera cases from all counties in Kenya from 2015 to 2023, allowing for a comprehensive and accurate assessment of the spatial and temporal distributions of cholera outbreaks. This approach ensured that no relevant data were excluded, which is crucial for mapping trends and identifying persistent or emerging hotspots.

Inclusion and exclusion criteria: the study included confirmed cholera cases reported in Kenya between 2015 and 2023 with complete geographic identifiers. Cases with missing location data, duplicate entries, or reported outside the study period were excluded, as shown in Table 1.

Quality control: to ensure the quality of data used, comprehensive data quality measures were implemented. This involved systematic validation checks and data cleaning protocols during the merging of cholera line list data. Additionally, to ensure accuracy and completeness of geographical identifiers, cross-referencing of counties and sub-counties with geocodes generated using Google Maps.

Data collection tools and procedures: epidemiological data were obtained from the National Cholera Case Line List maintained by the Division of Disease Surveillance and Response (DDSR) at the Ministry of Health, Kenya, covering the entire country during the study period. Data were collected using standardized cholera case line list forms developed by the DDSR, in accordance with the Kenya Integrated Disease Surveillance and Response (IDSR) Technical Guidelines. Cases were defined and classified based on the WHO/IDSR standard case definitions for cholera (suspected, probable, and confirmed cases), ensuring consistency in case ascertainment and reporting across all reporting units. A data abstraction form was used to extract key variables from the line list, including case identifiers, county and sub-county of reporting, date of onset, laboratory confirmation status, and outcome (recovered or deceased). Line lists were collected and collated from all counties and consolidated into a single national dataset. Data quality and validation were maintained through routine verification processes embedded within the IDSR framework, including cross-checking of line list entries against laboratory confirmation records where available. The consolidated line list was subsequently cleaned and georeferenced based on the geographical location of reported cases. County and sub-county population data, used as denominators for incidence calculations, were obtained from the Kenya National Bureau of Statistics.

Data analysis: the cholera case data from various sources were cleaned and checked for duplicates. Cases were entered into SPSS version 25 for descriptive analysis. Trends from 2015 to 2023 were analyzed to assess seasonality and geographic distribution on the basis of Mean Annual Incidence (MAI) and persistence. GIS coordinates of the patients' residential locations were used to construct a graphical map of the distribution of V. cholerae isolates using ArcGIS Pro. One-way ANOVA was used to examine temporal differences.

Ethical consideration: ethical approval and authorization to conduct the study were sought from the Mount Kenya University Ethical Review Committee (Ref: MKU/ISERC/4306) and the National Commission for Science, Technology, and Innovation (NACOSTI) (License no: NACOSTI/P/24/39710), respectively. Approval to access the cholera line list data was obtained from the Kenya Ministry of Health Division of Disease Surveillance and Response. Additionally, to ensure data privacy and protection, no personal identifiers, including names, addresses, and contact information, or any other directly identifying information were collected or extracted from the secondary data sources.

Results     

Distribution of cholera outbreak in Kenya by county, 2015 to 2023: From 2015 to 2023, a total of 42,850 suspected cholera cases were line-listed countrywide. Nairobi County had the highest cholera burden (24.8%), followed by Garissa (15.7%) and Wajir (10.2%), while Kitui (0.1%) and Narok Counties (0.1%) had the least burden. This data highlights temporal differences in cholera cases in Kenya. Nairobi County documented the highest burden, potentially because of the dense urban population, informal settlements with poor sanitation, and high mobility, which facilitates disease spread. In contrast, the low cholera burden in Kitui and Narok counties could potentially be due to lower population densities and underreporting of cholera cases. In regard to MAI analysis, Garissa County (163.8/100,000) had the highest values, followed by Wajir (104.7/100,000), Mandera (58.6/100,000), Tana River (48.4/100,000), and Nairobi County (23.4/100,000) (Figure 1). On the other hand, the five counties with the least cholera MAI included Nyeri, Kwale, Nandi, Kitui, and Narok.

Cholera annual incidence rate in Kenya from 2015 to 2023: as shown in Table 2, there was a significant temporal difference in cholera MAI between 2015 and 2023 (p= 0.047). The highest annual cholera incidence was in 2015, followed by 2023, whereas no cases were documented in 2019 and 2021. The standard deviations for MAI exceeded the corresponding means in several years (e.g., 2015: mean = 135.79, SD = 469.38), indicating a highly right-skewed distribution of cholera incidence across sub-counties, driven by a small number of high-burden areas such as Dadaab. Given this skew, median and interquartile range (IQR) values were also reported in Table 1 to provide a more robust summary of central tendency and dispersion.

Cholera annual incidences by sub-county: out of the 42,850 suspected cholera cases countrywide cases the five high-burden counties, i.e., Garissa, Wajir, Mandera, Tana River and Nairobi Counties, accounted for 63.4% of all the cases (27,169). Furthermore, there were spatial and temporal differences in cholera incidence, as shown in Figure 2. In 2015, Wajir East sub-county had the highest annual cholera incidence (2663.25/100,000). In 2016, 2017, 2018, 2022, and 2023, Dadaab Sub County had the highest annual cholera incidence of 640.85/100,000, 577.11/100,000, 1287.27/100,000, 783.65/100,000, and 611.50/100,000, respectively while in 2020 Embakasi East sub-county had the highest annual cholera incidence (365.29/100,000). The consistently high incidence in Dadaab sub-county is alarming. The sub-county hosts large refugee camps, and thus the data suggest that humanitarian crises and conditions in refugee or IDP settlements are key drivers of cholera outbreaks. The shift of high incidence to Embakasi East sub-county in 2020 implies that cholera hotspots can change over time; thus, public health responses need to be flexible and adaptive. In Ijara, North Horr, Saku, and Laisamis, there were no reported cholera cases between 2015 and 2023. The absence of documented cholera cases could potentially be due to effective ongoing control measures or underreporting due to limitations in surveillance.

Mean annual incidence (MAI) for cholera outbreaks in sub-counties in Nairobi, Garissa, Wajir, Mandera and Tana River counties: using MAI and persistence, four sub-counties, Dadaab, Embakasi East, Embakasi West, and Kibra, were classified as high priority, as shown in Table 3 and Figure 3. Using the same criteria, five sub-counties-Bura, Wajir East, Wajir South, Fafi, and Garsen-were classified as medium to low priority. Additionally, 8 and 17 sub-counties were classified as medium and low priority, respectively.

Association between socio-demographic characteristics and vital status: Table 4 shows a significant association between gender and vital status (χ² = 19.734, df=1, p< 0.001). The majority of cholera cases (67.6%) who succumbed to cholera were male, while a third (32.4%) were female. The findings suggest that men are disproportionately affected by cholera, potentially due to factors such as poor health-seeking behavior and low levels of hygiene as compared to females. Additionally, age had a significant association with vital status (t=-3.968, df=12658, p<0.001). The mean age of cholera patients who died was higher (29.52 ± 22.34 years) as compared to that of those who were alive (23.88 ± 19.92 years).

Discussion     

Kenya recorded 42,850 suspected cholera cases between 2015 and 2023, with Nairobi and Garissa counties bearing the highest burden and Kitui and Narok the least. The disproportionately high burden in Nairobi County is likely driven by inadequate sanitation infrastructure, compromised open drainage systems, and extreme population density within informal urban settlements such as the Kibera and Mathare slums [11]. Furthermore, Nairobi's status as the nation's capital city and primary transit hub acts as a critical vector for population mobility, potentially facilitating the rapid transmission and dissemination of the pathogen across regions [12]. High burden in ASAL counties and humanitarian zones may potentially be due to climate change parameters such as drought and floods, which affect water quality and consequently result in huge reliance on contaminated water sources [13]. Mean annual incidence (MAI) data revealed marked disparities in cholera burden, with Garissa (163.8/100,000) and Wajir counties bearing the highest, and Kitui and Narok the least. Poor WASH infrastructure, influx of refugees, and climate shocks could explain the high MAI in Garissa County [14]. Additionally, cross-border transmission and a nomadic way of life could be associated with the high MAI in Wajir and Mandera counties [15].

Five high-burden counties i.e. Nairobi, Garissa, Wajir, Mandera, and Tana River, accounted for a substantial 63.4% (27,169) of all nationwide cholera cases between 2015 and 2023, indicating a persistent spatial and temporal cholera burden in Kenya. This disproportionate distribution shows that cholera transmission is not uniform in the country but is mainly present in hotspot areas, mainly in arid and semi-arid lands (ASALs) and highly populated urban settings such as Nairobi. Further, the lower MAI in Kitui and Narok counties indicates the spatial differences in cholera transmission. The study findings align with a Kenyan study which documented that cholera is mainly present in areas with poor water sanitation and hygiene (WASH) infrastructure, high population, and in refugee settlements [5]. The spatial understanding is important for enhancing public health interventions and focusing resources such as vaccination and improvements of WASH infrastructure in high-burden counties for purposes of cholera control and consequent elimination.

The classification of sub-counties into high, medium, and low priority levels using MAI and levels of persistence is critical in guiding implementation of targeted cholera control projects in hotspot regions. The present study identified Dadaab, Embakasi East, Kibra, and Embakasi West sub-counties as high priority areas. Dadaab is a well-known refugee settlement and is regularly affected by a high cholera burden due to poor humanitarian conditions such as overcrowding, poor WASH infrastructure, and regional instability characterized by cross-border movement and influx of refugees [16]. For instance, conflict and drought in Somalia and flooding in South Sudan resulted in a significantly high number of refugees trooping to the Dadaab refugee camp in 2022, with considerable cholera cases being reported among new arrivals [17]. Further evidence suggests that climate change, which is characterized by severe droughts and flooding, propagates the spread of cholera [18].

Embakasi East, Embakasi West and Kibra, which were also identified as high-priority sub-counties, are densely populated urban informal settlements in Nairobi County. These informal settlements are persistently affected by cholera outbreaks, mainly caused by poor urban planning, poor sanitation, and inadequate access to safe drinking water [9]. Evidence suggests that crowded living areas, such as slum dwellings, contribute to cholera outbreaks [19]. Further, a World Bank report indicated that, as the population grows, poor urban dwellers are pushed into informal and low-income settlements with little or no sanitation and water [20]. For instance, informal urban settlements such as Kibra are faced with poor sanitation, congestion, poverty among inhabitants, and unreliable sewerage systems, which increase the risk of the spread of cholera [21]. The findings underscore the need for context-specific interventions which may include improvement of WASH infrastructure in refugee settings and enhanced waste control and management in urban slums.

A considerable number of medium-priority sub-counties, that is, two in Wajir Sub County, two in Tana River County, and one in Garissa County, were also identified. While these sub-counties experience low MAI and persistence, there is still a need for preparedness and disease surveillance. The occurrence of cholera in medium priority areas could be linked to environmental triggers such as floods, drought, or movement of people from highly endemic neighboring sub-counties. For instance, the influx of refugees into Wajir County from Somalia and Ethiopia could explain the identification of the county as a cholera hotspot [5]. Furthermore, the lack of potable drinking water due to drought and flooding in the Tana River could potentially explain the high incidence and persistence of cholera in this county [22].The mapping of cholera hotspots is critical in enabling public health initiatives to optimize resource distribution, hence ensuring that high-priority areas receive intensive and sustained multi-sectoral interventions. Additionally, medium and low priority areas may benefit from increased surveillance, rapid response, and targeted preventive measures to prevent future occurrence of cholera outbreaks [5].

Interestingly, Ijara sub-county in Garissa County did not report any cases between 2015 and 2023. This observation could be potentially due to government Oral Cholera Vaccination interventions in Garissa County, among other counties [23]. Furthermore, the lack of cholera cases could also be due to underreporting, health promotion education, and the distribution of water treatment solutions by non-governmental organizations (NGOs) operating in the County [24]. Evidence suggests that cholera is a neglected and under-reported disease, potentially due to limitations in surveillance systems, stigma, and fear of the effect on trade and tourism [2]. Age was significantly associated with vital status, whereby the mean age of cholera patients who died was higher than the mean age of those who were alive. This observation could be potentially associated with differences in exposure patterns, immune responses, and health-seeking behaviors across age groups. For instance, immunosenescence, a phenomenon characterized by a decrease in immune response with an increase in age, has been reported in humans starting in middle age [25]. This phenomenon has been associated with the increased severity of infectious diseases, such as cholera and influenza, and the increased prevalence of cancer [26,27]. For instance, a Nigerian study reported high cholera case fatality rates in older adults [28]. Additionally, evidence suggests that delayed healthcare seeking behavior among elderly individuals due to dismissal of cholera symptoms as common diarrhea, poor economic status and physical immobility may explain the study findings [29].

Gender had a significant influence on vital status among the cholera cases, whereby the majority of cholera patients (67.6%) who had succumbed to cholera were male, while close to a third (32.4%) were female. The findings indicate a notable gender disparity in cholera mortality, with males being disproportionately affected compared to females. Similarly, according to a study assessing global trends of cholera between 1990 and 2019, the rate of cholera mortality in the African region was higher in males (12.29/ 100,000) than in females (10.53/100,000) [30]. A Zimbabwean study also reported that men had higher odds of dying from cholera than did women [31]. During the 2018 cholera pandemic in Nigeria, sex was significantly associated with clinical outcomes, whereby the number of males who succumbed to cholera (54.72%) was higher than that of females (45.28%) [32]. This study is subject to a few limitations. The analysis was based on secondary data from the National Cholera Case Line List, which may be affected by underreporting and gaps in data quality across counties and over time. For instance, the absence of cases in 2019 and 2021 shows interruption of transmission or gaps in surveillance and reporting. Additionally, since incidence was analyzed at county and sub-county level, the findings may potentially be subject to ecological fallacy, as aggregate-level patterns may not accurately show individual-level risk factors or transmission dynamics within the administrative units.

Conclusion     

This study demonstrates distinct geographical and demographic disparities of cholera burden in Kenya between 2015 and 2023, with just five high-burden counties accounting for nearly two-thirds (63.4%) of all cases nationwide. While Nairobi County accounted for the highest overall burden (24.8%), Garissa County registered the highest Mean Annual Incidence (MAI) of 163.8 per 100,000, underscoring severe cholera vulnerabilities within arid and semi-arid lands (ASAL), humanitarian settings, informal urban settlements, and specific demographics, including males and older individuals. These localized variations, particularly the alarmingly high MAI in Dadaab sub-county, underscore the need to shift from a generalized nationwide response to a proactive, targeted strategy. To align with the Global Task Force on Cholera Control (GTFCC) 2030 Roadmap, healthcare stakeholders must prioritize data-driven resource allocation, enhanced cholera surveillance, robust water, sanitation, and hygiene (WASH) infrastructure, and targeted health education within these high-priority cholera hotspots.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Conceptualization; Waqo Gufu Boru and Alfred Owino Odongo. Methodology; Waqo Gufu Boru, Alfred Owino Odongo, John Kariuki. Validation; Waqo Gufu Boru, Alfred Owino Odongo, John Kariuki. Data collection; Waqo Gufu Boru. Data curation; Waqo Gufu Boru and Alfred Owino Odongo. Formal analysis & drafting paper; Waqo Gufu Boru. Review & Editing of the manuscript; Waqo Gufu Boru, Alfred Owino Odongo, John Kariuki.

Acknowledgments     

The authors express their gratitude to the research participants for their enthusiastic participation.

Tables and figure     

Table 1: sampling frame

Table 2: cholera Mean Annual Incidence among patients aged 1 to 92 years in Kenya by Year (2015 and 2023) (n=42,850 cases)

Table 3: classification of cholera cases by sub-county using MAI and cholera persistence in Nairobi, Garissa, Wajir, Mandera and Tana River counties among patients aged 1 to 92 years between 2015 to 2023 (n= 27,169 cases)

Table 4: association between socio-demographic factors and vital status among cholera cases documented in Nairobi, Garissa, Wajir, Mandera and Tana River counties among patients aged 1 to 92 years between 2015 to 2023 (n= 27,169 cases)

Figure 1: cholera mean annual incidence (MAI) in Kenya by County, 2015 to 2023

Figure 2: cholera annual incidence in Kenya by Sub-County between 2015 to 2023

Figure 3: cholera hotspot map for Nairobi, Garissa, Wajir, Mandera and Tana River counties in Kenya based on MAI and persistence data between 2015 and 2023

References