Original Research

Improved cholera control in Kenya: A retrospective analysis of 2017–2019 in Nairobi and Homabay

Kyeng Mercy, Ganesh Pokhariyal, Noah T. Fongwen, Nicaise Ndembi, Lucy Kivuti-Bitok

Received: 24 July 2024; Accepted: 18 Sept. 2024; Published: 22 Nov. 2024

Copyright: © 2024. The Author(s). Licensee: AOSIS.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background: Kenya has recorded at least 38 678 cases and 695 deaths over the last decade, and costing on average $2.2 million annually. From 2014 to 2016, the country experienced one of the deadliest and largest outbreak. However, between 2017 and 2020, there was a decline in the number of reported cases and deaths.

Aim: This study seeks to reveal the investments made post the 2014–2016 outbreak and highlight existing gaps that need to be addressed to stop the resurgence of cholera outbreaks in Kenya.

Setting: The study was conducted in two counties: Homabay and Nairobi.

Methods: We used an observational study. Data were collected from 20 health facilities (involved in cholera control, during the study), 9 key informant interviews (KII) and 6 focus group discussions (FGDs).

Results: We found improvement such as: dissemination of standard operating procedures, aligned reporting system, field epidemiology programme, establishment of a public health emergency operating centre and improved partner coordination. On the other hand, 12 of the selected 20 (60%) facilities had no prior training before government financing and laboratory capacity was sub-optimal: 13 (65%) facilities had no prior training, 16 (20%) had no operational laboratory plan and 10 (50%) had inadequate laboratory test kits and reagents.

Conclusion: This study highlights that Kenya has experienced an improvement in specific core capacities.

Contribution: For Kenya to completely flatten the curve, there is need for more sustainable investment and government’s commitment in health system strengthening.

Keywords: Kenya; cholera; surveillance; response; evaluation.

Introduction

In the last 10 years, the world has seen a global rise in the number of cholera cases reported. Cholera continues to pose a significant threat to Kenya’s national health system and has caused at least 38 678 cases (including confirmed and clinically compatible cases) and 695 deaths over the last decade.¹ Within Africa, Kenya has experienced frequent cholera outbreaks, with more than 15 discrete outbreaks from 1971 to 2010. In 2015, Kenya accounted for 19% of the 172 454 reported cases and 7% of the 1304 globally reported deaths.² During 2014–2016, Kenya experienced one of the deadliest and largest cholera outbreaks since 2010, with at least 16 840 cases and 256 deaths reported from 30 of the 47 counties. This outbreak had far-reaching impacts on the quality of life and the country’s economy costing over $2.2m annually.³ Since then, the country has been reporting cholera cases almost annually in certain high-risk areas. Factors such as open defecation, high population density in urban slums, increased cross-border movement of persons from countries reporting cholera outbreaks amid humanitarian crises and the growing climate change with increased incidence of flooding were some of the key drivers of the outbreaks reported in Kenya.¹ The Global Task Force on Cholera Control launched an initiative to eradicate cholera by 2030 and developed a roadmap that focusses on: prompt detection and effective response; a multi-sectoral approach to detect cholera resurgence in the endemic areas; a productive partnership mechanism of coordination for surveillance, research, and monitoring and evaluation, resource mobilisation and advocacy. Aligning with this vision, the Kenyan government has targeted to reduce the cholera annual incidence to 0 per 100 000 by 2027 and have prioritised the administration of oral cholera vaccine (OCV) to identified hotspots to curb the spread of the disease.^2,4 Between 2017 and 2019, Kenya recorded a 16% decrease in cases compared to the 2014–2016 outbreak (Figure 1).⁴ This study seeks to understand the improvements made in the surveillance and response capacity post-2014–2016 that led to improvements in containing the outbreak despite health system challenges and competing health emergencies such as climate change.

FIGURE 1: Trend of cholera cases in Kenya from 2013 to 2023.

This evaluation focussed on the assessment of the ability of Kenya’s health system to detect and respond promptly to cholera cases during the 2017–2019 outbreak. The study builds on previous cholera capacity evaluations conducted in 2016, and aims to identify the current state of cholera investments in Kenya and assess the existing gaps in prevention and control efforts to mitigate the risk of future outbreaks.

Research methods and design

This observational study employed a mixed method of qualitative and quantitative data collection to evaluate cholera surveillance and response data at the county, facility and community levels. The quantitative and qualitative datasets were given equal weight in analysis and comparison. The study was a non-probability purposive sampling carried out in four sub-counties in Nairobi (Kibra, Embakasi, Dagoretti and Kasarani) and two in Homabay (Homabay town and Ndhiwa).

Study participants

Participants for the key informant interviews (KII) included health facility surveillance focal points (Integrated Disease Surveillance and Response Strategy [IDSR] focal point) or hospital officer in charge (where there is not a disease surveillance focal point) and district medical officers. A total of nine key informants were interviewed (four females and five males). Participants for the focus group discussions (FGD) included community health workers (community extension workers) and public health officers. Only those facilities involved in cholera surveillance within the study area (Dagoretti, Kibra, Embakasi, Kasarani, Ndhiwa and Homabay Town) were included in the health facility assessment.

Sampling procedure and justification

Health facilities were selected based on the following criteria: they were the first point of referral for all suspected cholera cases, they had a laboratory capacity to carry out cholera diagnosis and finally they had reported cases of both suspected and confirmed cholera for 2014–2016 and 2017–2019 cholera outbreaks in the Kenya health information system.⁵ A total of 20 healthcare facilities and 36 community healthcare workers were recruited from the sub-counties.

Data collection

Data on Kenya’s capacity to detect and respond to the 2017–2019 cholera outbreak were collected from the 20 health facilities using a questionnaire developed from three guidelines: IDSR, the updated guideline for evaluating surveillance systems and the Africa Centres for Disease Event-based Surveillance (EBS) framework. Nine KIIs and six FGDs were conducted to gain more insights on the capacity of Kenya’s surveillance system. Capacity areas were assessed using a consolidation of surveillance variables of three key pillars: system resources, system structures and core functions anchored on good governance and leadership.⁶

Data analysis

The data captured using questionnaires were programmed into Comcare software and exported to Stata 5 (STATA Corp [2017] STATA Statistical Software: Release 15. College Station, Texas, United States [US]: StataCorp LLC) for error checking and cleaning. Data from reviewed forms were additionally keyed in Microsoft Excel and exported to STATA for cleaning before analysis. All FGDs and KII were audio-recorded and transcribed into English. Transcripts were reviewed and a codebook was generated.

Ethical considerations

Ethical approval to conduct this study was obtained from the Kenyatta National Hospital University of Nairobi Ethics Review Committee (No. KNH/ERC/R/71). Written consent was also obtained from all the research participants.

Results

During the 2017–2019 outbreak, a total of 14 074 cholera cases and 266 deaths (case fatality rate [CFR]: 1.9%) were reported from 20 counties in Kenya. This represented a 16% decrease in the number of cases compared to the 2014–2016 outbreak where at least 16 500 cases were reported.⁷ During the 2017–2019 outbreak, a total of 333 cholera cases and 9 deaths (CFR: 2.7%) from cholera were reported from the two sampled counties^1,5,8: Nairobi (258 cases; 7 deaths) and Homabay (75 cases; 2 deaths). This represents an 85% decrease in the number of cases reported compared to the 2014–2015 period in these counties.^1,5,8 Table 1 provides an analysis of capacities across core areas, examining changes between the periods 2014–2016 and 2017–2019.

Surveillance governance

Placement of Kenya’s National Cholera Surveillance Unit and structure

During the 2017–2019 cholera outbreak, there was no functional national public health institute in place.⁹ However, the Government of Kenya had established the Division of Disease Surveillance and Response (DDSR) to build an effective, efficient public health surveillance and response system. At the regional level (County and sub-County), 70% of the responders indicated the presence of a trained Rapid Response Team (RRT), and Disaster Preparedness and Management Committees in charge of disease preparedness and response. In the health facilities, 60% had RRT and surveillance unit for monitoring, assessing and responding to cholera outbreaks.

Health financing

Kenya just like majority of the African countries had not met its target of allocating at least 15% of government’s total budget in healthcare during both the evaluation periods. All African Union Member States pledged to this target of 15%, now termed the ‘Abuja declaration’.¹⁰ In 2013, the devolution policy in Kenya went into full force with the counties given the autonomy to mobilise and manage their resources. In the fiscal year 2013–2014, the county’s share of the total health budget was 54%. In the 2017–2018 fiscal year, their share increased to 63%, but decreased to 57% in 2018–2019.^11,12 This decrease had severe setbacks in the management of health emergencies at country level:

‘There were challenges and some of which needed financing, and you know that most of our activities, especially in health, I’m sorry to say, is majorly financed by partners. Sometimes, we also needed support supervision, and we still depended on them because of the transport challenges.’ (FGD E, sub-county surveillance officer, male)

Legislations, policy guidelines and standard operating procedures

During both periods, legislations and policies that govern disease surveillance and control including cholera were in place (Appendix 1). However, the level of implementation differed across both periods. The second edition of the IDSR by the World Health Organization (WHO) was in place prior to the 2014–2016 outbreak.¹³ Cholera was listed within this strategy as one of the priority diseases in the country. Before the onset of 2014–2016, less than 20% of facilities were actively implementing the IDSR. However, by 2017 all (100%) sub-counties had case definitions and 91% had standard operating procedures (SOPs) and guidelines in place. At the facility level, while 80% had a list of priority diseases, a standard case definition of cholera and guidelines on cholera waste management, 50% (n = 10/20) did not have SOPs for the coordination of communication across multiple sectors and stakeholders.

Event-based surveillance had not been fully entrenched into the surveillance structure and system then:

‘No. Actually, EBS is a new concept to us. It has been implemented in other areas such as Nakuru and Mombasa but in Nairobi, only a few of us have been sensitized on it; me and a few staff from Mbagathi and Mama Lucy hospitals.’ (KII K, sub-county county surveillance officer, male)

Resources

Trained workforce capacity

On trained workforce, for 2014–2016, there were no training needs assessment performed, and no staff had received any trainings on EBS. This performance was similar in 2017–2019 as 13 out of the selected 20 (65)% facilities had no training needs assessment conducted before the outbreak, 12 (60%) reported that there were no prior trainings of hospital health staff and community health promoters or supervisors on EBS and 11 (55%) indicated that there was no prior active engagement, sensitising and training of community leaders, health volunteers and other community members as appropriate, in the detection and reporting of cholera (Table 2):

‘Training needs are done after we have had an outbreak. Staffs are just donated from different departments.’ (KII W, sub-county surveillance officer, female)

‘In Kenya, training is not done as part of preparedness. Training is done after the outbreak has occurred; it is done during, not before.’ (KII D, health facility in-charge, female)

On average, confidence in skills for cholera surveillance and response stood at 34%, significantly lower than the gold standard of 80%.

Non-medical countermeasures

During the 2014–2016 outbreak, non-medical supplies scored between 60% and 100%. During the 2017–2019 period, 16 out of 20 (80%) had adequate drugs, materials and other products essential for cholera case management, while 10 out of 20 (50%) indicated that consumables such as laboratory testing reagents, personal protective equipment such as gloves, masks and gowns were not available and accessible by healthcare workers. On average, the resource availability stood at 60% of the requirements (Figure 2):

‘For us who were here, that is why we are saying that the PPEs and some equipment at the community level support World Vision. The support from partners was coordinated by counties.’ (FGD D, sub-county surveillance officer, male)

FIGURE 2: Account of resource availability.

Laboratory capacity

Less than 20% of facilities had guidelines for laboratory diagnoses prior to the 2014–2016 outbreak and furthermore the roles and responsibilities were not clear. Prior to the 2017–2019 outbreak, all the facilities had a guideline defining the roles and responsibilities, a list of reference laboratories in the country and an SOP on diagnostic testing. However, only 7 out of the 20 (35%) facilities had refresher and/or on-the-job training for laboratory staff before the outbreak and only 4 (20%) had an operational laboratory plan in place. None of the specified resources reached the 80% mark of adequacy. Specimen and material transport scored 40%, laboratory space 35%, equipment and staffing 25%, and reagents and consumables 15% (Table 3).

Challenges with the availability of Rapid Diagnostic Test (RDT) kits were reported during the 2017–2019 outbreak similar to the challenges encountered during the 2014–2016 period:

‘The issue with rapid diagnostic test kits was that they were never enough. This is a big problem. For example, we are dealing with the outbreak today, but we last tested a case several months ago.’ (KII K, health facility in-charge, female)

Unlike the RDT, the culture test takes 72 h (3 days) or up to about 2 weeks and was used to confirm the disease.

Mechanisms and system structures

During the 2014–2016 outbreak, there were no clear roles and responsibilities among stakeholders involved in the response. However, responders confirmed this was in place during the 2017–2019 outbreak. These included information flow (Figure 3)¹⁴, multi-sectoral coordination mechanism, laboratory and surveillance networks, cross-border and partner coordination and the linkage of laboratory systems with community and health facility surveillance.

FIGURE 3: Integrated Disease Surveillance and Response (IDSR) Data Flow in Kenya.

Coordination and networks

The emergency operations centre was in place only during the 2017–2019 period and was activated for the first time which played a key role in coordinating the response at the national level. Partner coordination at the facilities stood at 60% with 40% indicating that there were no partner coordination mechanisms in their facilities during both periods. There were no available frameworks nor resources for partner engagements during both the periods:

‘The system consists of well-prepared human resources who are able to detect, notify and investigate an outbreak in a timely manner. However, other resources intended for coordination and investigation were limited.’ (KII H, sub-county surveillance officer, male)

In addition, responders underscored concerns about sustainability after the partners left:

‘I’m suggesting that they [health partners] should also be thinking about sustainability … At the end of their projects term, everything stalls.’ (FGD D, community health worker, female)

There was no multi-sectoral coordination meeting before this outbreak, and other sectors such as water and the environment were not optimally engaged:

‘It [water department] was not that strong compared to what we have right now. Right now, there is good coordination …’ (KII N, sub-county surveillance officer, male)

During both the periods, there were no cross-border coordination mechanisms established for cholera because of limited information of cholera outbreaks related to cross-border movements. Sixty per cent (n = 12/20) of facilities acknowledged existence of laboratory networks, which facilitated sample referrals, testing and interpretation of results.

Data management and information flow

During the 2017–2019 period, there was a clear guideline on how data were managed and reported from the community to the national level. However, this was not in place during the 2014–2016 period (Appendix 2). The information flow highlights that a suspected case of cholera is detected at the communities and facilities and then reported to the sub-county disease surveillance team which manages the case at the facility closest to the source with capacity.

System performance

During the 2017–2019 period, the overall system performance was at 73.7%, with simplicity having the highest score at 85.4% and stability the least score of 56.1%. The system sensitivity, usefulness, acceptability, data quality, positive predictive value, timeliness and simplicity scores were 83.1%, 83.5%, 69.4%, 68.1%, 69.3%, 69.8% and 85.4%, respectively (Figure 4). It was widely agreed that the system’s ability to detect diseases was considerably good; however, the system was not able to respond to diseases on time:

‘The surveillance system works but timely resource availability hinders prompt and swift response.’ (KII D, sub-county surveillance officer, male)

FIGURE 4: System attributes.

Discussion

Our findings demonstrate that there were indeed some improvements in the country’s capacity to detect, investigate and respond to the cholera outbreak comparatively.⁵ The rollout of the IDSR, surveillance and response guidelines, and SOPs and case definitions to the lower levels improved community engagement and prompt detection and reporting of cholera cases by community health workers. Even though the performance of the timeliness (for detection) indicator stood at 68.9%, this would have been better if the EBS system was fully operational.^15,16,17 Other studies conducted in Kenya demonstrated that case detection in areas improved when SOPs, guidelines and case definitions are in place, and the community health workers are trained on the same as opposed to areas where these are missing.¹⁸ The availability of a cholera response plan just before 2017 also guided the response and facilitated coordination among stakeholders and partners⁴ Unlike in 2014, when there were no clear roles and responsibilities in reporting, the healthcare workers attested that they were clear with their reporting lines during the 2017–2019 outbreak. This finding aligns with another study from Nigeria which demonstrated that preparedness plans significantly improved early detection for Ebola and shortened the response time by the local authorities. The absence of such plans during the 2014–2016 period could have negatively reduced the effectiveness and speed in containing the outbreak.¹⁹ During both the periods, there was no harmonised mechanism for reporting at the community level, which further delayed standardising data across different communities and tracking cases in time. Some community health workers reported using a hard copy of the reporting form, while others reported collecting information through phone calls. Such a reporting mechanism has been proven to lead to incomplete and unharmonised reports and standards, delaying decision making and resource allocation during emergencies.¹⁸ These variations have now been addressed through the digital tool ‘m-Daura’ which is now serving as a platform for all community reports and EBS.¹⁷

During the 2017–2019 outbreak, there was an improvement in community engagement, and the community health workers could easily get relevant information from the community about cholera cases. This facilitated the penetration of the Ministry of Health (MoH) in hard-to-reach communities and facilitated early detection even in areas without proper hospitals, which aligns with the same results obtained from a similar study conducted in Haiti.²⁰ Such workforce capacity could further be leveraged in the roll out of OCV in cholera hotspots. Even after a surge experienced in 2015, the Kenyan MoH established the Public Health Emergency Operating Centre (PHEOC) in 2016 only, and it was activated for the first time in 2017.^21,22 Given the critical role PHEOC plays in coordinating outbreaks, this key milestone was a huge achievement in coordinating and controlling the outbreaks during 2017–2019 outbreak.²¹ A study conducted in Pakistan revealed the improvement in coordination and response to dengue when the PHEOC was established and activated.²³ The decentralisation of the PHEOC could further improve outbreak coordination and response at sub-national levels while also lessening the burden on the national PHEOC during widespread outbreaks, as demonstrated by Nigeria and Uganda.^21,24

Multisectoral collaboration was weak during both periods. The MoH, in collaboration with line ministries and partners, only convened the first national task force forum for the fight against cholera in July 2018. This was one of the setbacks in improving water hygiene and sanitation, which is a critical pillar in the control of the disease as demonstrated by previous publications.²⁵

Laboratory systems and networks have been shown to be very instrumental in confirming cases during outbreaks for timely case management and response. However, the ability of countries to decentralise testing for advanced techniques such as culture and polymerase chain reaction (PCR) has always been a challenge. Even after this was highlighted as one of the challenges in the 2014–2016 period, the confirmatory tests were only conducted at the central level with a turnaround time of approximately 2 weeks. Such delays have been shown to sustain transmission chains in the communities by other studies.^26,27 Rapid diagnostic test kits have been shown by other studies to bridge this gap by improving case detection, reducing the risk of misdiagnosis and inappropriate treatment at the community.²⁸ However, while there was capacity at the sub-counties to perform RDTs, the kits were often limited in stock, leaving the communities to rely on suspected case definitions to detect them. A study conducted on the evaluations for sensitivity and cross-reactivity of Crystal VC (RDT test kits) used during the 2017–2019 outbreak, demonstrated a sensitivity of 97.5%, and specificity of 100% with no cross-reactivity; however, they were limited in stock.²⁹ The Africa Centres for Disease Control and Prevention (CDC) has identified this as a critical gap on the African continent and is now supporting local manufacturing of laboratory test kits for priority diseases. The deployment of Field Epidemiology and Laboratory Training Program (FELTP) during the 2016–2017 played a key role in the control of the outbreak in urban cities. While there was significant support at the national level, the counties and sub-counties were not well-manned. Most of the staff deployed to the field at the sub-counties were mobilised by partners whose work ceased when the funding dwindled. Several countries also observed this scenario during sustained outbreaks such as coronavirus disease 2019 (COVID-19).³⁰ In addition, the Kenyan health system relies mainly on voluntary commitment of community health assistants to detect and report health threats which has not been sustainable over the years.^31,32 The formalisation of the engagement of community health workers as well as their formal integration into the healthcare system has shown to improve workforce development for emergency reponse.³³

Training of healthcare workforce (laboratory, community health assistants, and volunteers and clinicians) was still considered one of the issues in both periods and it led hospital-acquired cholera infections in 2015.⁵ In 2017, most of the staff were seconded by partners and training was not properly conducted. The ripple effect is that, these system gaps translated to sub-optimal overall system performance revealed through the attributes. Of the 10 attributes (data quality, completeness, simplicity, timeliness, stability, usefulness, sensitivity, acceptability, representativeness, and positive predictive value) assessed, 3 (usefulness, sensitivity and simplicity) had acceptable performances of > 80%. The system stability scored lowest (56.1%) which was because of the uncertainty of staff contracts and unavailability of MoH computers for data capture. The breakdown of such devices usually leads to a disruption of workflow, given that it relies on the mercies of the staff for repairs.³⁴ Health financing remains the driving force to ensure workforce sustainability and sustainable health financing. The total spending of health increased from $3476m accounting for 5.2% of the country’s gross domestic product (GDP)³⁵ in 2015 to 8% in 2018,³⁶ which is still below the 15% target of the Abuja declaration to which Kenya is a signatory. On the other hand, there was a decrease in budgetary allocations in health promotion and infectious disease prevention from 16 398m KES in 2015 to 7701m KES in 2015 and 6093m KES in 2015 to 5505m KES in 2017, which partly accounted for the poor water, sanitation, and hygiene (WASH) performance in counties and sub-counties without alternative funding source. The operationalisation of the Kenya National Public Health Institute will further strengthen the country’s health system governance and mobilise resources for disease prevention, detection and response. This would further strengthen key pillars such as workforce development, surveillance, laboratory, and emergency preparedness and response.

Limitations

We noted several limitations in this study. Firstly, the sampling deployed was non-probability purposive sampling in selecting the counties and the facilities thus generalising the result to other countries might not be feasible. However, the qualitative aspect of the study provides deep insights that could support other countries with improving cholera preparedness, surveillance and response. Secondly, there were issues with recall bias given that some of the staff who were actively involved in both outbreaks had moved to other duties. Retrieving historic data was also challenging because the system depended on paper-based reporting during that period. We were able to support the findings with other data retrieved from hospital registers as well as national weekly reports which we could find on some websites. Lastly, we were not able to evaluate the effectiveness of the deployment of the RDTs during both periods which limited our understanding of the impact of its use between both the outbreaks.

Conclusion

The Republic of Kenya has experienced an improvement in specific core capacities in cholera surveillance and response after 2014 to 2016 leading to decreased cases and deaths registered between 2017 and 2019. Improvements in areas such as the development and dissemination of guidelines and SOPs, defining information flow (including roles and responsibilities) across administrative levels and actors, activating the FELTP and establishment of a PHEOC were some major successes registered partner engagement and devolution that improved partner coordination. However, there remains sub-optimal performance in government financing and allocation of funds towards addressing public health emergencies. Establishing the National Public Health Institute will be an opportunity for such an agency to lead resource mobilisation and workforce development to support all the pillars of emergency response. Given the devolution of governance in Kenya, there is also an urgent need to explore the decentralisation of PHEOC and reference laboratories to reduce detection and confirmation time. The issue of shortage of RDT was a critical gap that can also be addressed via investment in local manufacturing. While there was no evidence for cross-border cholera transmission then, there is now ongoing cross-border transmission of cholera in the Horn of Africa and thus a need to strengthen the cross-border coordination to improve information sharing and joint response planning in border areas.

Acknowledgements

The authors would like to thank Moses Otieno for facilitating some of the focus group discussions in Nairobi and all the focal points from the different facilities who took part in the study. We would also like to acknowledge the insights provided by Dr. Philip Ngere (National event-based surveillance focal point) and for his valuable contributions. We also acknowledge the contributions of all key informants as well as participants of the focus group discussions.

The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.

K.M. developed the original idea for this study including the study design and methodology. K.M., N.T.F. and L.K.-B. analysed the data. K.M., G.P., N.N., and L.K.-B. reviewed and improved the article. All authors provided feedback and approved the content of the final article.

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

The views and opinions expressed in this article are those of the authors and are the product of professional research. The article does not necessarily reflect the official policy or position of any affiliated institution, funder, agency or that of the publisher. The authors are responsible for this article’s results, findings and content.

References

Appendix 1

Appendix 2