Introduction

Snakebite envenoming (SBE) is a neglected tropical disease (NTD) of major public health concern, particularly in tropical regions where environmental, geographical, and demographic factors intersect to elevate the risk of human-snake encounters [1]. Despite increasing global awareness, the true burden of SBE, especially in low- and middle-income countries (LMICs), remains elusive, hindering effective resource allocation and public health interventions [1,2].

In Africa, the estimated incidence of snakebite is alarmingly high, at one million bites per year [3], suggesting that the continent accounts for a substantial portion of global snakebite cases. Annual snakebite estimates vary from 100 to 650 bites per 100,000 inhabitants with mortality rates exceeding 10 deaths per 100,000 [4]. Factors such as rural living, agricultural activities, and limited access to healthcare exacerbate the risk of SBE in many African communities, however poor data availability complicates efforts to accurately assess the true extent of the problem.

Unravelling the true extent of SBE presents a complex challenge. Relying solely on health facility data offers only a partial view, that fails to capture the full spectrum of SBE’s impact on public health [5,6]. Given widespread underreporting, inadequate SBE surveillance systems and limited healthcare accessibility, particularly in rural areas, this narrow approach leads to a significant underestimation of the burden faced by affected communities, masking the broader socio-economic and health implications of SBE. This lack of comprehensive data impedes the development of effective intervention strategies, exacerbating the consequences of SBE in vulnerable populations. Addressing the challenge of quantifying SBE requires a comprehensive approach that overcomes the systemic barriers to accurate reporting and healthcare access in at-risk regions.

The challenges of poor data availability and limited resources are not unique to SBE but are common across the spectrum of NTDs. Given the impact this has for addressing NTD burden, the World Health Organization (WHO), in the strategy ‘A Road Map for Neglected Tropical Diseases 2021-2030’ [7] has called for integration across programmes in order to “improve [their] cost-effectiveness, coverage and geographical reach”. By consolidating efforts and resources for programme management, intervention delivery and disease monitoring, there is the potential to efficiently tackle the burden of multiple diseases simultaneously, maximizing impact and cost-effectiveness. Moreover, by collaborating across diseases that share characteristics, a more holistic approach to public health interventions becomes possible, emphasizing common preventive measures, promoting and improving treatment accessibility, and fostering community engagement.

Within this context, Kenya, situated in eastern Africa, faces a high burden from SBE and reported incidence and mortality figures underscore the urgency of addressing this issue. Annual incidence rates as high as 150 cases per 100,000 inhabitants have been reported in a community-based survey conducted in Kilifi County [8], though the vast majority of available Kenyan data still come from health facility reporting. A facility-based study conducted in the same county showed rates of just 44 cases per 100,000 inhabitants [9]. Another study conducted in Samburu County found a one year mortality rate of 28 deaths per 100,000 inhabitants [10]. Several other NTDs are also endemic in Kenya, and over the years considerable progress has been made in controlling several of these conditions, including lymphatic filariasis, onchocerciasis, trachoma, schistosomiasis, and soil-transmitted helminths (STH) [11]. Mass drug administration (MDA) is a public health intervention whereby a preventive chemotherapy is dispensed to all individuals within a defined population, irrespective of their disease status. Preventive chemotherapy, distributed through MDA, is one of the main interventions used to control and eliminate these NTDs [12]. With extensive coverage across the country [13] such platforms facilitate the collection of comprehensive data on disease prevalence and distribution, providing an opportunity to streamline resource allocation and delivery.

In light of this, our research sought to integrate snakebite surveillance into ongoing MDA control programmes for trachoma, schistosomiasis and STH within Kenya to acquire substantial data on community-level incidence of SBE and associated hospital seeking rates. We also sought to assess the costs of this integrated approach for capturing SBE burden data.

Methods

Results

A total of 13,117,754 persons from 17 counties were included in the MDA surveys, representing 27.6% of Kenya’s total population. 687,2155 (52.4%) respondents were female. The percentage of females in the sampled populations ranged from 48% to 59% (median 52%). The sampled county population size varied from 132,316–1,682,003 persons with a median coverage of 93% (min. 63%, max. 112%) of the total sub-county population as recorded in the 2019 census (Table 1). Two counties, Kajiado and Lamu, did not achieve the target coverage rate of 80%, achieving instead a coverage rate of 63% and 75% respectively. A second MDA intervention was organized in these counties but did not include the questions on snakebite for logistic reasons.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Mass Drug Administration (MDA) Coverage per county.

https://doi.org/10.1371/journal.pntd.0013732.t001

Snakebite incidence

A total of 4,667 cases was reported across the 17 counties; 52.6% male. Males had a slightly higher incidence rate of 39.3 cases per 100,000 inhabitants (95% CI 37.8-40.1), compared to females (32.2 cases per 100,000 inhabitants, [95% CI 30.9-33.6]). Females had an 18% lower chance of having a snakebite (RR = 0.82, 95% CI = 0.77-0.87).

For all the (12) counties where schistosomiasis and STH MDA was conducted, snakebite cases were divided into three age groups (1–4 years, 5–14 years, and 15+). Of the total 1,655 cases, 108 snakebite survivors were aged 1–4, 352 were aged 5–14 and 1,195 were aged 15 or above. The lowest snakebite incidence rate was reported in children 1–4 years old at 5.6 cases per 100,000 inhabitants (95% CI 4.6-6.8). Children aged 5–14 reported a snakebite incidence rate of 9.2 (95%CI 8.3-10.2) cases per 100,000 inhabitants and those aged 15 and above had a rate of 19.5 cases per 100,000 inhabitants (95%CI 18.4-20.7).

The overall incidence rate of snakebite in all the counties surveyed was 35.6 (95% CI 34.6 – 36.6) cases per 100,000 inhabitants/year at risk. Incidence rates varied greatly between the counties (Fig 1). Turkana County had the highest incidence of 412.9 snakebites per 100,000 inhabitants/year (95% CI 393.6 – 432.8), closely followed by Baringo County (410.9 [95% CI 380.1-443.5]). Vihiga County recorded the lowest incidence rate of 3.7 snakebites per 100,000 inhabitants/year (95% CI 2.2-5.3-4.7). There was a very weak negative correlation (rho -0.14) between subcounty level incidence rates of snakebite and the population density.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Between county variation in snakebite incidence in Kenya.

Base map: GADM database of Global Administrative Areas, version 4.1 (www.gadm.org). Used under license for academic, non-commercial use (https://gadm.org/license.html). Maps created in QGIS version 3.40 (QGIS Development Team. QGIS Geographic Information System. Open-Source Geospatial Foundation Project. http://qgis.osgeo.org).

https://doi.org/10.1371/journal.pntd.0013732.g001

Table 2 presents the snakebite counts and incidence by county. Inverse weighting was applied based on MDA coverage to provide estimated total case numbers for the target population in each county if 100% coverage was not met.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Population density and snakebite incidence per county, as reported by community members over the past twelve months.

https://doi.org/10.1371/journal.pntd.0013732.t002

Snakebite incidence varied less within counties than between counties. The maps of Turkana County and Kakamega County show examples of within-county variation in snakebite incidence in high and low incidence settings respectively (Figs 2 and 3). Detailed maps presenting sub-county and county level incidence rates for all 17 counties are provided in S1–S15 Files.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Within county variation in snakebite incidence in Turkana County (A high incidence setting).

Base map: GADM database of Global Administrative Areas, version 4.1 (www.gadm.org). Used under license for academic, non-commercial use (https://gadm.org/license.html). Maps created in QGIS version 3.40 (QGIS Development Team. QGIS Geographic Information System. Open-Source Geospatial Foundation Project. http://qgis.osgeo.org).

https://doi.org/10.1371/journal.pntd.0013732.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Within county variation in snakebite incidence in Kakamega County (A low incidence setting).

Base map: GADM database of Global Administrative Areas, version 4.1 (www.gadm.org). Used under license for academic, non-commercial use (https://gadm.org/license.html). Maps created in QGIS version 3.40 (QGIS Development Team. QGIS Geographic Information System. Open-Source Geospatial Foundation Project. http://qgis.osgeo.org).

https://doi.org/10.1371/journal.pntd.0013732.g003

Comparison with previous studies

The data presented in Table 3 below illustrates how our study, which captured community-level incidence of snakebite through MDA integration, provides significantly higher estimates of incidence in several counties compared to prior published hospital-based studies and household level surveys. For example, in Baringo County, comparing our community-based estimate of 410.9 cases per 100,000 people per year with the hospital-based studies (e.g., 67.9/100,000 by Coombs et al. and 6.7/100,000 by Ochola et al.). In contrast, our lower incidence estimates in counties such as Kakamega and Busia are consistent with previous findings.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Snakebite incidence rate in counties covered in this study compared to incidence in published literature.

https://doi.org/10.1371/journal.pntd.0013732.t003

Health facility attendance following snakebite

In the 12 counties where SCH/STH MDA was conducted, health facility seeking rate was disaggregated by gender and age groups. There was no notable difference between the health facility seeking rate after a snakebite between the male (63%, 554 of 882) and female (67%, 520 of 773) population in these counties. Following a snakebite, the health facility seeking rate was 51% (n = 55) for children aged 1–4.The rate was 67% (n = 236) in those aged 5–14, and 65% (n = 783) in individuals aged 15 and older). The proportion of snakebite survivors reporting having sought health facility care varied across all the 17 counties, from 3% in Trans Nzoia County to 86% in Kajiado County (median overall 66%). Aside from West Pokot, in all counties the number of snakebite survivors reported having attended a health facility were lower than the data reported in the Kenya Health Information System (KHIS) for the same region and period (Table 4). Table 4 compares the number of health facility visits reported for snakebite in the MDA surveys to snakebite case numbers reported through the KHIS.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Comparison between snakebite related health facility (HF) visits based on the MDA exercise and data reported in the KHIS data. Weighting was done at the sub county level. *Percentages <100 indicate lower number of hospital visits as reported in the communities compared to numbers in HFs.

https://doi.org/10.1371/journal.pntd.0013732.t004

Discussion

A lack of geographic and demographic data on snakebite envenoming incidence within countries has been noted as a barrier to effective distribution of antivenom [22,23]. Our collaboration with NTD MDA campaigns has provided the most extensive data on community snakebite incidence in Kenya to date. Between January 2020 and August 2023, we obtained high quality data from 17 counties, accounting for 27.6% of the country’s population. The work uniquely captures the considerable variation in incidence rates across an endemic country, and in doing so increases our knowledge of incidence in both high and relatively lower burden regions; the latter are often overlooked as studies tend to focus on areas where snakebite envenoming is a known risk [22,24,25]. Incorporating snakebite data capture within the MDA campaigns has provided rich epidemiological data from which policy makers can effectively plan resource allocation for snakebite treatment and prevention.

We found an overall snakebite incidence of 36 cases/100,000 people/year, with figures as high as 413/100,00/year in Turkana, 411/100,000/year in Baringo and 176/100,000/year in West Pokot [9,18,20,26–30].

The data we have obtained from 17 of the 47 Kenyan counties to date shows a clear geographical pattern, with the highest estimated incidence rates clustered in counties to the northwest. We found a weak negative correlation between county-level incidence and population density and within counties (between sub-counties). Snakebite risk generally tends to be higher in sparsely populated rural areas [31–34]. That our study shows a weak correlation likely reflects the multiple factors accounting for the risks at subcounty level. Population density is just one of several spatially varying factors accounting for the geographical variation in snakebite risk, with climate and environmental factors affecting snake habitat suitability [26], and sociodemographic factors affecting human-snake interaction [32,35,36]. Prevention and control efforts should reach remote rural populations with typically the poorest access to healthcare services [21,37]. In future work, conducting geospatial modelling would allow us to further utilise this data, to model and predict risk across a larger area.

Within the MDA campaigns, we collected data on the proportion of snakebite survivors who reported attending a health facility. Our results show considerable variation in the proportion reporting healthcare attendance across the counties, ranging from as low as 3% in Trans Nzoia to 86% in Kajiado. There are many reasons documented for why patients might not seek care in formal health facilities, including long travel distances, high cost of treatment compared to traditional or self-treatment and local traditions and beliefs [20,38], however we did not observe a correlation between healthcare attendance and population density. Due to the limited number of questions we were able to ask, it was not possible for us to capture detail on the reason for patients not seeking health facility care. With attendance below 70% in 11/17 counties, there is potential benefit in working with communities in these areas in order to find ways to improve healthcare access and use.

There were discrepancies when comparing the number of cases reporting health facility attendance compared to the number of snakebite cases reported by the KHIS for the same period, with estimates for 12/17 counties below 70% of KHIS figures and a range from 0.9% of the KHIS figures in Trans Nzoia to 250% in West Pokot. We expected some variation between these figures, with our estimates weighted for the county population to account for incomplete MDA coverage rates, however the degree of concordance varies substantially across counties. It is possible that KHIS figures reflect multiple reporting of cases, as patients may initially seek care at a local primary healthcare facility and are subsequently referred to higher tier facilities for antivenom administration or further management [39], or patients may be re-recorded as a snakebite case on re-attendance for follow-up care. It is also possible that the KHIS figures for one location capture patients who have travelled out of or been referred from their home locality to receive appropriate treatment. The implication of this is twofold. Firstly, standardising the collection and reporting of snakebite case information by healthcare facilities has the potential to considerably improve the utility and reliability of this routine data source for understanding snakebite burden across the country. Secondly, given patients may not seek care at the facility in their sub-county of residence, both survey data and KHIS reports should be considered in antivenom stock planning.

There is limited snakebite data available from Kenya, but our estimates suggest higher rates than those identified from previous hospital-based studies [18]. Incidence estimates based on hospital data are known to underestimate true figures, as many snakebite patients, particularly in remote rural locations, seek treatment from traditional healers rather than using the formal healthcare system, whilst those developing complications from severe envenoming may die before they reach a health facility [2,20,26]. Notably, there was a discrepancy between the counties reporting the highest number of cases through the MDA surveys and the KHIS, with only two of the top five counties by MDA data (Turkana and Kilifi) in the top five for KHIS reports. In this context, our data provides an important addition to the literature from community based-studies which seeks to provide a more complete reflection of the snakebite burden across Kenya. There is relatively little high quality snakebite incidence data available more broadly from sub-Saharan Africa with which to compare our findings. A review of the global burden of snakebite estimated an upper incidence for the East Africa region of 22.61/100,000 [2], though it notes the paucity of data available to base this on. Elsewhere, the majority of data comes from small community studies conducted in the West and Central Africa regions and our highest incidence rates fall within the middle to upper range of these figures, similar to those found in studies from Cabo Delgado, Mozambique (352/100,000) [27] and Benue Valley, Nigeria (497/100,000) [28] but lower than very high rates found in areas including Cameroon (665/100,000) [29], and Bandafassi, Senegal (915/100,000) [40].

Our study successfully integrated two key questions on snakebite burden across several large MDA surveys, allowing us to rapidly obtain information on this condition for a quarter of Kenya’s population for less than £3000 in addition to existing costs for the MDA [6,41]. A systematic review identified integrated NTD strategies to be cost effective [42]. In line with the WHO’s call for greater integration across NTD programmes, to ‘improve [their] cost-effectiveness, coverage and geographical reach’ [41], this study highlights the considerable benefits from collaboration across NTD programmes. Integration often combines activities in health education, management or control. Our study illustrates how combining different types of activities can be beneficial; elimination efforts for established NTDs with data collection on basic epidemiology for the more recently recognized NTD snakebite envenoming. As an NTD with very little formally published data [2], such approaches have the potential to provide significant benefits for understanding snakebite epidemiology across high-burden countries.

This study has limitations that should be considered. Firstly, there are factors which may have impacted the accurate reporting of snakebite. It is possible that recall bias may have led to snakebite cases being misclassified either within or outside of the one-year time frame of interest. Where this has occurred, it is likely to have affected data collection similarly across counties and therefore the variation between counties is unlikely to have been affected. It is also possible that the cultural perceptions of snakebite may have affected how forthcoming some participants were in providing information; social stigma associated with snakebite [43] could have led some respondents not to disclose their experience, particularly if data collection was conducted in a public setting. Conversely, in settings in which community engagement on healthcare matters is strong, due to regular NGO presence or research activity, residents may have been more inclined to report on healthcare matters. Secondly, working within the confines of existing NTD programmes meant that data collection was limited to areas in which MDA activities were planned, which did not always cover the whole county. Finally, while this approach to data collection has provided important information on snakebite risk, the nature of integrating with existing programmes meant that the range of information we were able to gather about cases was limited and could not include (i) discriminating bites by venomous vs non-venomous snakes and (ii) data on the burden of mortality from snakebite. To further inform prevention and control strategies, gathering detailed information from priority counties on the circumstances surrounding snakebite episodes, the management received and outcomes from the bite will be important.

The data collected has enabled us to identify those areas that should be prioritised for snakebite prevention and control, and integrating with MDA programmes that are planned over the upcoming years will provide an opportunity across selected counties to both fill the gaps in data collection and repeat assessments in order to better understand the dynamics of snakebite burden over time.

Conclusion

Our collaboration with NTD MDA campaigns has yielded comprehensive data on snakebite incidence in Kenya, revealing significant nationwide variation and offering vital insights into both high and lower snakebite-risk regions. This informs resource allocation for treatment and prevention. Integrating with MDA programmes that are planned over the upcoming years will provide an opportunity across selected counties to both fill the gaps in data collection and repeat assessments in order to better understand the dynamics of snakebite burden over time. Integrated approaches, capitalizing on existing platforms for multiple NTDs, demonstrate significant cost savings while aligning perfectly with the WHO roadmap for NTD control and elimination.

Supporting information

Acknowledgments

We extend our sincere gratitude to the State Department of Public Health and Professional Standards and the Vector Borne and Neglected Tropical Diseases Unit for their invaluable support and collaboration throughout this study. We also thank the numerous stakeholders, including African Institute for Health and Development (AIHD), Operation Eyesight, Sight Savers, Fred Hollows Foundation, Amref Health Africa, Children Without Worms (CWW), Evidence Action, Child Investment Fund Foundation (CIFF), ARISE Fund, ALMA Foundation and END Fund, for their continuous support and contributions towards the MDA exercise. Additionally, we acknowledge the Ministries of Health in the participating counties for their cooperation and assistance in facilitating data collection. Our heartfelt appreciation goes to the CDDs whose dedication and hard work made this research possible.