Foot-and-mouth disease (FMD) is endemic in Kenya where serotypes A, O, SAT1 and SAT2 are frequently encountered. Despite the importance of the dairy industry and the frequent reporting of disease, the epidemiology of FMD and field-based vaccine effectiveness has been poorly described in these endemic settings. Additionally, the disease impact has been inadequately characterised, despite the importance of such information when allocating scarce resources for animal health in national disease control strategies. The objectives of this doctoral thesis were to gain field experience of FMD in endemic settings and to use appropriate outbreaks to assess the vaccine effectiveness, gather evidence to optimise the use of vaccines and inform national policy, and to estimate disease impact. Outbreaks on two large-scale dairy farms located within Nakuru County, Kenya, were investigated and detailed descriptions of the outbreaks are presented. Both farms regularly used locally produced, aqueous adjuvanted, non-NSP purified quadrivalent (A, O, SAT1, SAT2) vaccine every 4-6 months. The first attended outbreak was caused by serotype SAT2 and evidence was found of limited or no vaccine effectiveness. At the second outbreak, due to serotype O, there was evidence of increasing protection with increasing number of doses. The reasons behind the vaccine poor effectiveness are discussed and are likely to include poor match with the field strain and inappropriate schedules in youngstock. Virus neutralisation test data were made available from the vaccine manufacturer who sample animals on farms using routine prophylactic vaccination. The influence of maternally derived antibody on the response to vaccination was investigated with these data and recommendations on vaccine schedules and future research priorities are made based on the evidence presented. On the farm that had SAT2, analysis of the disease impact was performed using individual animal data. Longitudinal analysis of individual milk yields utilising generalised estimating equations and an autoregressive variance structure to account for the correlation of yields for individual animals was performed. Predictions of 305-day milk yields were made based on previous lactations in the same herd. Despite a clear herd level impact, no difference was found between recorded clinical FMD cases and non-cases. More detailed analysis revealed significant reductions among older animals in earlier stages of lactation but younger cows were able to recover sufficiently so that no overall impact was seen. The impact of clinical disease on the rate of clinical mastitis and culling was analysed utilising a historical cohort approach with survival analysis over a 12-month period after the commencement of the outbreak. Hazard ratios (HR) were generated using Cox regression accounting for nonproportional hazards by inclusion of time-varying effects. There was good evidence of an increased rate of mastitis in the first month after the onset of the outbreak (HR=2.9, 95%CI 0.97-8.9, P=0.057) although the effect on culling was less clear. The implications of these findings for policy and further research are discussed.