Survival Analysis Methods To Estimate Incidence, Survivorship And Risk Factors For Brucellosis In Livestock In Kajiado East Subcounty, Kajiado County, Kenya

Abstract

Brucellosis is a common bacterial zoonotic disease that is caused by various Brucella species and mostly affects cattle, goats, sheep, pigs and dogs. It results in significant economic losses and human sufferings. Manifestation in livestock is mainly through abortions, retained placentas, premature births, infertility and reduced milk production. Despite the disease being successfully controlled in many developed regions, it is still of major public health importance in sub-Saharan Africa. In Kenya, there is limited data on incidence of brucellosis in livestock. The main objective of this study was to estimate the incidence rates, disease risk probabilities and the risk factors associated with time to brucellosis infection in different livestock species in Kajiado County. Multistage sampling technique was used whereby in the first stage 4 out of 17 locations were selected randomly, followed by proportionate simple random sampling of herds in the selected locations. Stratified random sampling was used in selected herds to identify animals that were enrolled into the study. A cohort of 1369 sheep, 1711 goats and 709 cattle from 500 compounds were enrolled into the study and followed up for 9 months. At the animal herd level, risk factors that were assessed included production system, mixing with other herds, contact with wildlife and breeding system. At individual animal level, data was collected on breed, age, sex, species, breeding status and breeding system. Blood samples were collected from enrolled animals at enrolment and on each of the two follow-up visits. Sera was tested for antibodies against Brucella using competitive Enzyme-Linked Immunosorbent Assays (ELISA). Semiparametric and nonparametric survival analysis techniques were used to explore risk factors associated with time to brucellosis infection in different livestock species. Disease incidence rates were calculated and xiii survival probabilities compared using the log rank method. Cumulative incidence was 1.7%, 0.7% and 0.3% in cattle, sheep and goats respectively. Incidence rates for infection with brucellosis were highest in cattle with a rate 7 times that of goats, experiencing worse survival that sheep and goats. On bivariate analysis, the hazard was 1.5 times higher in cattle that were more than 2 years old compared to the under 2 years old though the finding was not statistically significant. Similarly, for sheep the hazard was about 2 times for sheep more than 6 months old. On multivariable Cox proportional hazard regression, there was marginal statistically significant association between natural breeding system and brucellosis infection (p=0.05) in cattle, while sheep that were raised under semi-zero grazing system had an increased hazard for brucellosis infection (p =0.0001). More brucellosis incidence studies for Brucella infection are required to aid in the understanding of transmission dynamics and therefore inform prevention and control programs in pastoralist settings