Vaccination Week in the Americas, 2011: an opportunity to assess the routine vaccination program in the Bolivarian Republic of Venezuela

Our coverage survey was conducted on August 13–14, 2011. For MOH and PAHO, it was considered an operational study and not research. The protocol was reviewed by the human subjects coordinator for CDC’s Center for Global Health, and was considered a programmatic evaluation and exempt from institutional board review of human subjects research. The survey’s target area included the municipality of Bolivar, subdivided into the areas of San Antonio del Táchira (the municipality’s capital and predominately urban area), Palotal (periurban area) and a rural area consisting of Juan Vicente Gómez and Isaías Medina Angarita parishes. Children were eligible to participate if they were less than 6 years of age during April and May 2011.

We calculated a sample size of 267 children eligible to be vaccinated during VWA 2011 assuming 50% of eligible children would receive at least one vaccine, precision of ±6%, and probability of achieving precision of 0.95. We used binomial distribution to calculate a sample size of 720 children less than 6 years of age, based on the assumption only 40% of target age children would be eligible for at least one vaccine and to ensure 90% probability of identifying the target of 267 eligible children.

We used a stratified 30-cluster survey design, proportionally allocating clusters to the 3 areas based on population size (22 in San Antonio del Táchira, 5 in Palotal, and 3 in the rural area). In each stratum, we selected clusters (barrios) with replacement and probability of selection proportional to size. We used population estimates from MOH and included 24 households with at least one child less than 6 years old in each cluster. We selected an area within the cluster using a 10x10 grid superimposed over a map of the cluster and chose one cell within the grid via simple random sampling. The first house within that cell was also chosen at random and from there, we visited all subsequent households to the right, until 24 households with at least one child less than 6 years of age had been surveyed.

A total of 44 interviewers were trained who were health workers^b from Bolivar. Survey teams consisted of two interviewers and one supervisor who accompanied two or three teams. Survey teams were assigned to areas outside of their normal jurisdiction and collected data on all children in the target age range within a participating household.

Survey teams administered a standardized questionnaire to any available caregiver (e.g. mother, father, grandmother, etc.), to collect information about the caregiver’s education, household demographics, type of health center routinely visited, awareness about a vaccination activity conducted in their community during April or May 2011, and, when applicable, the caregiver’s actions in response to that knowledge. Survey teams transcribed information in each child’s vaccination card (doses received and dates of administration). When a card was not available, teams reviewed the health center’s vaccination records; teams did not collect caregivers’ recollections about specific vaccine doses received, but did record responses to questions about the campaign and communication. We trained data entry staff who entered data into an EpiInfo® (CDC, Atlanta, GA, USA) database. In the field, accuracy of data entry was reviewed by the technical focal point from the national immunization program.

We used SAS v9.2 (SAS Institute, Inc., Cary, NC, USA), SUDAAN® v10 (RTI International, Research Triangle Park, NC, USA), and R v3.0 (R Foundation for Statistical Computing, Vienna, Austria) to clean and analyze data. We restricted vaccination-related analyses to children with information available through vaccination cards (and for a small minority, health center records). We determined each child’s eligibility to receive a specific vaccine dose during VWA 2011 based on their date of birth and the information recorded regarding specific doses received. We calculated proportions of children who received any vaccine doses, specific vaccine doses, or all vaccine doses for which they were eligible during VWA 2011 (April 23, 2011 – May 31, 2011). We used all vaccine doses transcribed, regardless of when they were administered, to estimate overall coverage of specific vaccines at the time of the survey, by five 1-year age cohorts. We estimated vaccination coverage, 95% confidence intervals (CI), and Chi-square tests to compare coverage across age cohorts, taking into account the stratification, the first stage cluster sampling, and stratum-level weights (estimated target population in each stratum divided by sample size in each stratum) in SUDAAN. We used R survey package [10] to estimate time to vaccination and 95% CIs (Aalen [hazard-based] estimator) for each dose of pentavalent vaccine^c (Penta) and first dose of measles-mumps-rubella vaccine (MMR1). The results are presented in graphs as 1 minus time to vaccination to show estimated change in probability of being vaccinated through 24 months of age (Penta) and 36 months of age (MMR1). To show uncertainty around the estimated curve, 95% CIs for probability of vaccination are calculated and plotted at each time point that an event (i.e., child is vaccinated) occurs.

We surveyed caregivers of 839 children less than six years of age in 698 households, including 513 (74%) households in the urban area of San Antonio del Táchira, 114 (16%) in the semi-urban Palotal, and 71 (10%) in the rural area. Most (84%) were mothers; 8% were grandmothers. Fifty-nine percent of caregivers reported having completed high school or university education. Fifty-two percent of eligible children were female.

Among caregivers surveyed, 90% reported using public outpatient and hospital settings for routine vaccination services, while 5% utilized private institutions and 5% reported other institutions. Among 698 households surveyed, 216 (31%, 95% CI 25-38%) respondents reported having heard about a vaccination activity during April or May 2011. Among these 216, 152 (70%, 95% CI 59-81%) reported having heard about a general campaign, while only 20 (9%, 95% CI 5-18%) identified VWA 2011 specifically. The most common sources of information among those aware of a recent vaccination activity were radio (28%, 95% CI 17-39%) and the local health center (20%, 95% CI 13-29%). When caregivers were asked how they responded after learning about a vaccination activity, 63% (95% CI 54-73%) reported bringing their child to the health center with a vaccination card, 7% (95% CI 2-19%) said they brought their child without a vaccination card, 25% (95% CI 17-34%) indicated that they did not bring their child to the vaccination site and 5% (95% CI 2-10%) of responses were classified as “other”.

Among 839 surveyed children, 784 (93%) had a vaccination card, and for an additional 6 (1%) children, vaccination data were available from clinic records. Among all 790 children with available vaccination data, 528 (67%, 95% CI 63-71%) were age-eligible to receive at least one vaccine dose during VWA 2011 (Table 2). Among these, 126 (24%) received at least one dose for which they were eligible, and 69 (13%) received all indicated doses. Among the 126 children who received at least one vaccination, 57 (45%) did not receive all vaccinations for which they were eligible. Among 18 children eligible to receive Bacillus Calmette-Guérin vaccine (BCG) and 13 eligible for the first dose of hepatitis B, 16 (89%) and 11 (85%) received BCG and hepatitis B, respectively, representing the largest proportion of eligible children to receive an indicated vaccine dose. By contrast, among children eligible to receive yellow fever vaccine (N = 127), the second booster dose of oral polio vaccine (OPV2) (N = 45), or the second MMR dose (MMR2) (N = 87), only 3 (2%), 1 (2%), and 4 (5%), respectively, received the recommended vaccine(s).

BCG coverage estimates were at least 95% for each 1-year age cohort; however, coverage with birth dose of hepatitis B vaccine was lower, with the lowest being 81% (95% CI 71-88%) in children aged 24–35 months (Table 3).

Coverage with Penta3 exceeded 90% in most cohorts, and was generally higher than coverage with OPV3, which ranged from 86% in 12–23 month-olds to 96% in 48–59 month-olds. MMR1 coverage was only 72% (95% CI 63-80%) among children aged 12–23 months, although in older cohorts it was approximately 90%. Similarly, yellow fever vaccine coverage estimates were higher (92-95%) in older age cohorts but significantly lower in younger age cohorts, with coverage estimates of 62% (95% CI 54-70%) among children aged 12–23 months and 82% (95% CI 75-87%) in children aged 24–35 months. In contrast, second dose rotavirus vaccine coverage estimates were highest in the two youngest age cohorts (51% [95% CI 41-62%] among 12–23 month-olds and 52% [95% CI 43–60] among 24–35 month-olds), but were significantly lower in the three older age cohorts: among children aged 60–71 months, coverage was only 21% (95% CI 14-29%). Coverage with first DTP and OPV booster doses generally increased with increasing age, from 59% (95% CI 52-66%) (DTP) and 64% (95% CI 55-72%) (OPV) among children aged 24–35 months to 74% (95% CI 62-83%) (DTP) and 81% (95% CI 69-90%) (OPV) among children aged 48–59 months. First booster coverage for children aged 60–71 months continued to increase for DTP at 77% (95% CI 67-85%) but fell slightly for OPV to 79% (95% CI 66-88%). There were statistically significant differences in coverage between birth cohorts for the following vaccine doses: hepatitis B birth dose, Penta3, MMR1, yellow fever vaccine, second rotavirus dose, and first booster doses of DTP and OPV vaccines.

Approximately 90% of children received the first dose of pentavalent vaccine (Penta1) between 2 and 4 months of age (Figure 1); however, timeliness of subsequent doses decreased, even though coverage remained high. For example, most cohorts only achieved 50-70% coverage with Penta3 by the end of the recommended time of 6 to 8 months of age, although coverage by 8 months of age was only about 40% in the 36–47 month group. By 24 months of age, all 5 cohorts had reached 80-95% coverage with Penta3.

There was more variability between cohorts in MMR1 coverage and timeliness of vaccination (Figure 2). The CI for the 12–23 month olds indicates coverage at 15 months was between 45 and 63%. The curves for the 23–35, 36–47, and 48–59 month groups indicate higher coverage by 15 months. MMR1 coverage gradually increased for all cohorts over the following two years.