Time-to-immunization varied by location as well: children in Kilifi Township received each dose of pentavalent vaccine earlier than their peers in rural areas. However, the hypothesis that improved physical access to vaccine clinics increases the timeliness of immunization was not substantiated by our data. This finding may stem from a number of factors. First, travel time to vaccine clinics varied little within the Epi-DSS. Maximum pedestrian and vehicular travel times to vaccine clinics were less than 3 h and less than 2.5 h, respectively, with 75% of children residing less than 72 min on foot and less than 42 min by vehicle from a clinic. In this context, traveling to clinics may not impose

a significant burden on families or hamper timely immunization. Second, we were unable see more to account for several factors that may confound the association between time-to-immunization and physical access to care. We employed sublocation-level maternal education as a proxy for socio-economic status and were therefore unable to reflect inequalities in socio-economic status within sublocations, which may be associated both with distance to clinics and timing of immunization. Further, we were unable to Sorafenib mouse account for family size or birth intervals in our model. Parity and birth intervals may affect time-to-immunization and are likely to vary with distance to clinics; they may therefore be important

confounders as well [9] and [30]. We have previously shown that travel time is a barrier to hospital admission in the Kilifi Epi-DSS (J Moïsi, submitted). Assuming no residual already confounding, the absence of a relationship between timeliness of vaccination and distance to clinics in this analysis suggests that programmatic differences between immunization and hospital service

delivery play an important role in service utilization. Programmatic factors contributing to high immunization coverage may include the decentralized provision of immunization services, the perceived high quality of these services, or the focus on proactive outreach efforts via Supplementary Immunization Activities (SIAs) and mobile clinic team activities. Measles and polio SIAs were conducted in Kilifi District in the second half of 2006, but should have no effect on pentavalent vaccine coverage since the vaccine was not delivered through this mechanism. Outreach via mobile teams was donor-funded, localized, and sporadic during the study period yet may have contributed to high coverage as well. While no variations in time-to-immunization were seen with travel time to vaccine clinics, other key predictors of immunization rates were identified in this study. At a given age, children were 14% less likely to be immunized with pentavalent vaccine during the rains than during the dry season: the rainy season coincides with the harvest and impedes travel, even for short distances.