The study’s framework was used to structure this analysis (see Table 3). Questionnaire responses were cleaned and re-coded to allow comparison across countries, where necessary and possible. They were then analysed using descriptive statistics in SPSS software. Routine data were plotted over time and if a small change in trend was visible, a segmented

regression analysis was conducted to formally test its statistical significance [20]. Ethical approval was gained from the London School of Hygiene and Tropical Medicine and from the study countries. The study was verbally described to participants, an information sheet SCH 900776 in vivo was provided and signed consent gained from all, prior to commencing data collection. 261 semi-structured interviews were conducted and

196 health facility questionnaires were completed (see Table 4). 245 interviews were recorded (94%) and 65 interviews were translated from Spanish, Amharic and Kinyarwanda into English. The new vaccines generally seemed to integrate well into existing health systems. The introductions were considered to have had no impact on many of the elements find more within the building blocks framework (see Table 5 for summary of findings). Of those effects that were identified, most were within the vaccination programme; very few effects on the broader health system were reported. Some effects (e.g. increased staff workload) were reported to be temporary, at the time of introduction only. Given space limitations, only key findings are discussed below. Despite many key informants and facility

respondents perceiving that the new vaccine introductions had increased coverage of other vaccines, especially in Kenya, Cameroon and Ethiopia, the routine data collected in all countries did not support these claims (see Fig. 1). The only exception was in the case of Mali (PCV), where uptake of the first pentavalent dose increased by about 40% (Fig. 1), although this effect was because not sustained over time. However it should also be noted that the analysis in Mali (PCV) was based on data from only 13 of the 27 included facilities, due to incomplete data being available in the remaining 14 facilities. The high demand for new vaccines may have encouraged those who had previously defaulted on existing routine vaccinations. This created an opportunity to check the vaccine status of those attending and, when necessary, administer missed doses. Although study participants reported isolated efforts to use the new vaccine to trace defaulters in this manner, no country demonstrated a systemic approach to this. No impact of the introduction on ANC service use was observed from routine data before and after the introductions. Study participants generally felt that the new vaccine introductions had not affected cold chain capacity for other vaccines or products, for a number of reasons.