Association between delayed pentavalent vaccination and immunisation drop-out in rural western Kenya: findings from a cross-sectional survey

AbstractBackgroundThe identification of determinants of vaccination coverage and delays is important because these predictors identify subsets of the population that can be targeted with public health interventions. This study's aims were to determine vaccination coverage and timeliness estimates and to identify risk factors for not being vaccinated or receiving late vaccinations in rural western Kenya.MethodsThe study was set in 120 villages within the Kenya Medical Research Institute and Centers for Disease Control and Prevention (KEMRI/CDC) surveillance system in rural western Kenya. In March, 2013, we did a cross-sectional survey of women in the study district who had infants aged 12–23 months. We excluded data from women who did not have their child's immunisation booklet and could provide only a verbal report of their child's immunisation history. We calculated risk factors for delayed and non-receipt of immunisation using binomial regression with log link. Infants were categorised as “delayed” if immunisation was received more than 28 days from schedule. Analyses were restricted to first and third dose of pentavalent vaccine (diphtheria, tetanus, and pertussis [DTP], hepatitis B, and Haemophilus influenzae type b), measles vaccine, and full immunisation (BCG, three doses of polio, three doses of pentavalent, and measles vaccines).FindingsWe surveyed 2495 women and included data from 1748 women (70%). We excluded data from 747 women (30%) who did not have their child's immunisation booklet card present. Immunisation coverage for pentavalent 1, pentavalent 3, measles, and full immunisation were 99%, 95%, 83%, and 80%, respectively. The proportions of infants with delayed pentavalent 1, pentavalent 3, and measles immunisations were 10%, 24%, and 29%, respectively. Risk factors for not receiving vaccination and delayed vaccination were, overall, similar when estimate trends and confidence intervals were taken in to account. Specifically, infants who received pentavalent 1 late were more likely than children who had the first pentavalent vaccine within schedule to not receive pentavalent 3 (adjusted RR 5·60 [95%CI 3·75–8·37]), measles vaccine (1·47 [1·11–1·94]), and full immunisation (1·87 [1·51–2·32]). Adjusted analyses for predictors of immunisation booklet present at time of survey versus verbal report found that women living closer to the health facility, women with younger infants, and those with male infants were more likely to have an immunisation booklet present at time of survey.InterpretationImmunisation coverage for pentavalent 1, pentavalent 3, and measles vaccines was high, but many children receive their vaccinations more than 28 days after the recommended scheduled date. Since late pentavalent 1 vaccination was associated with pentavalent series drop-out, not being immunised against measles, and not being fully immunised, interventions that target infants with delayed pentavalent 1 vaccination may improve overall immunisation coverage. A potential limitation of this study was that almost a third of mothers were excluded because they could provide only a verbal report of immunisation history instead of their child's immunisation booklet. If there were substantial differences in characteristics between those with an immunisation booklet and those without, and these characteristics were associated with vaccination coverage and delay, the reported estimates may be susceptible to bias.FundingThe Bill and Melinda Gates Foundation