BACKGROUND: Kenya introduced a pentavalent vaccine including the DTP, Haemophilus influenzae type b and hepatitis b virus antigens in Nov 2001 and strengthened immunization services. We estimated immunization coverage before and after introduction, timeliness of vaccination and risk factors for failure to immunize in Kilifi district, Kenya. METHODS: In Nov 2002 we performed WHO cluster-sample surveys of >200 children scheduled for vaccination before or after introduction of pentavalent vaccine. In Mar 2004 we conducted a simple random sample (SRS) survey of 204 children aged 9-23 months. Coverage was estimated by inverse Kaplan-Meier survival analysis of vaccine-card and mothers' recall data and corroborated by reviewing administrative records from national and provincial vaccine stores. The contribution to timely immunization of distance from clinic, seasonal rainfall, mother's age, and family size was estimated by a proportional hazards model. RESULTS: Immunization coverage for three DTP and pentavalent doses was 100% before and 91% after pentavalent vaccine introduction, respectively. By SRS survey, coverage was 88% for three pentavalent doses. The median age at first, second and third vaccine dose was 8, 13 and 18 weeks. Vials dispatched to Kilifi District during 2001-2003 would provide three immunizations for 92% of the birth cohort. Immunization rate ratios were reduced with every kilometre of distance from home to vaccine clinic (HR 0.95, CI 0.91-1.00), rainy seasons (HR 0.73, 95% CI 0.61-0.89) and family size, increasing progressively up to 4 children (HR 0.55, 95% CI 0.41-0.73). CONCLUSION: Vaccine coverage was high before and after introduction of pentavalent vaccine, but most doses were given late. Coverage is limited by seasonal factors and family size.