Factors associated with cholera in Kenya, 2008-2013

Introduction Kenya experienced widespread cholera outbreaks in 1997-1999 and 2007-2010. The re-emergence of cholera in Kenya in 2015 indicates that cholera remains a public health threat. Understanding past outbreaks is important for preventing future outbreaks. This study investigated the relationship between cholera occurrence in Kenya and various environmental and demographic factors related to water, sanitation, socio-economic status, education, urbanization and availability of health facilities during the time period 2008-2013. Methods The primary outcome analyzed was the number of cholera cases at the district level, obtained from the Kenya Ministry of Health's national cholera surveillance records. Values of independent variables were obtained from the 2009 Kenya Population and Housing Census and other national surveys. The data were analyzed using a zero-inflated negative binomial regression model. Results Multivariate analysis indicated that the risk of cholera was associated with open defecation, use of unimproved water sources, poverty headcount ratio and the number of health facilities per 100,000 population (p < 0.05). No statistically significant association was found between cholera occurrence and education, percentage of population living in urban areas or population density. Conclusion The Sustainable Development Goals and Kenya's blueprint for development, Kenya Vision 2030, call for access to sanitation facilities and clean water for all by 2030. Kenya has made important economic strides in recent years but continues to be affected by diseases like cholera that are associated with low socio-economic status. Further expansion of access to sanitation facilities and clean water is necessary for preventing cholera in Kenya.


Introduction
Cholera is an epidemic-prone diarrheal disease of global significance. The disease is endemic in Kenya  2015, the percentage of the overall population that used improved drinking water sources increased from 43% to 63% [14]. In urban areas, however, the percentage of population using improved water sources actually decreased from 92% to 82% during the same period. Only modest progress has been seen in Kenya with respect to sanitation. The percentage of the population using improved sanitation facilities increased from 25% in 1990 to 30% in 2015.

Methods
The investigation was based on a cross-district analysis of cholera  Table 1.  Table   2 for the 137 districts that were included in regression analyses. Univariate analysis using ZINB regression models suggested a statistically significant association (p < 0.05) between cholera occurrence and each of the factors investigated except for percentage of population living in urban areas. In Table 3 The risk of cholera was significantly higher in districts in the high poverty category compared to the low poverty category (ARR 3.29, 95% CI 1.21-8.94), but no significant difference in risk was seen between the mid and low poverty categories (ARR 0.82, 95% CI 0.32-2.10). Cholera risk was lower in districts with a high level of use of improved water sources compared to districts with a low level of use (ARR 0.43, 95% CI 0.21-0.89), but no significant difference was observed between the mid and low categories (ARR 0.96, 95% CI 0.48-1.91). Districts with a higher number of health facilities per 100,000 tended to have a higher number of reported cholera cases (ARR 1.06, 95% CI 1.02-1.10). The percentage of the population living in urban areas was the only factor that was a statistically significant predictor of zero inflation on its own and was therefore used in the zero inflation portion of the model. This factor had a negative relationship with zero inflation, suggesting that districts with a lower percentage of population living in urban areas had a higher probability of excess zeros (that is, not reporting cases).

Discussion
The results of this study suggest that districts in Kenya  This study also found an inverse relationship between cholera occurrence and use of an improved water source, suggesting a protective effect of improved water sources. This is similar to a finding by Leidner et al of a statistically significant relationship between cholera incidence and access to improved water sources [19]. An earlier study in Kenya, however, found no significant relationship, in multivariate analysis, between cholera incidence and the percentage of population without a piped water supply, which is one type of improved water source [9]. Access to improved water sources like piped water does not necessarily guarantee a clean water supply. Ensuring that improved water sources supply clean water is an important consideration. Although Kenya achieved an overall increase in percentage of population with access to improved water sources from 43% in 1990 to 63% in 2015, it must be noted that access in urban areas actually declined during this period. This is likely due to population growth in urban slums without adequate expansion of drinking water infrastructure. As efforts continue to expand access to clean water in rural areas, cholera outbreaks in The results of this study suggest that cholera burden is also associated with poverty. Other studies have found this association as well [11,12] and cholera is often referred to as a disease of poverty. This study did not attempt to investigate mechanisms by This supports similar results related to population density from a previous study in Kenya [9]. Observed patterns of cholera occurrence in Kenya indicate that cholera affects both urban and rural areas, including the most densely populated and least densely populated regions. This study also did not find an association between cholera occurrence and education. It is difficult to draw conclusions from this finding, however, given the high degree of colinearity that was observed between education, open defecation and poverty. This study is subject to several limitations. One is potential regional variations in surveillance capacity that could affect the quality of cholera surveillance data. This study attempted to mitigate potential variability in reporting across districts by using cholera data over several years. There is some uncertainty in

Competing interests
The authors declare no competing interests.

Authors' contributions
Gretchen Cowman designed the study protocol, collected data,     The following categories of facilities were not included: HIV counseling and testing center, nursing home, regional blood transfusion center, training institution, dental clinic, laboratory, radiology unit, health project, and facilities labeled as "not operational" or "pending opening."

Ministry of Health Master Facility
List, http://www.ehealth.or.keaccessed 23 Oct 2013 + "jabia" is a term used in Kenya to describe a traditional rainwater storage system Page number not for citation purposes 9