National trends of childhood vaccination coverage in Bangladesh: Understanding gaps from 2011, 2014 and 2017-18 Demographic and Health Surveys data


 Background

Vaccination is a safe and cost-effective strategy for protecting children from life-threatening diseases. This study aimed to investigate the changes over time in proportion of vaccination coverage across demographic subgroups in Bangladesh.
Method

Vaccination coverage for children 12–59 months of age was obtained from Bangladesh’s 2011, 2014, and 2017-18 Demographic and Health Surveys. Three multivariable binary logistic regression models with complete vaccination status at each survey data were used to generate adjusted odds ratios to uncover immunization variations across socio-demographic categories. Further, to detect potential geographic disparities, changes over time were analyzed at both the country and district levels.
Findings

Vaccination coverage grew from 88.1% in 2011 to 89.2% in 2017–2018, indicating a gap of more than ten percent for complete immunization coverage among children aged 12–59 months. Children aged 23–35 were more likely to have full vaccination compared to those aged 12–23 in 2011 (AOR = 1.6, 95% CI = 1.3–1.9), 2014 (AOR = 1.6, 95% CI = 1.3-2.0), and 2017-18 (AOR = 1.4, 95% CI = 1.1–1.7). The likelihood of full vaccination of children increased with maternal education with the highest odds among women with higher education in 2011 (AOR = 7.2, 95% CI = 4.0-12.9), 2014 (AOR = 4.0, 95% CI = 2.7–5.9), and 2017-18 (AOR = 5.8, 95% CI = 3.5–9.6) compared to those with no formal education. Children born to mothers who lived in urban areas were more likely to have full vaccination in 2011 (AOR = 1.5, 95% CI = 1.3–1.8), 2014 (AOR = 1.4, 95% CI = 1.2–1.7), and 2017-18 (AOR = 1.4, 95% CI = 1.1–1.7) compared to those who lived in rural areas. Children born to mothers who had at least 4 ANC visits were more likely to have full vaccination compared to those with less than 3 ANC visits in 2011 (AOR = 1.6, 95% CI = 1.2-2.0), 2014 (AOR = 1.5, 95% CI = 1.2-2.0), and 2017-18 (AOR = 1.6, 95% CI = 1.2–2.1 During these three surveys, Rangpur division had the highest vaccine coverage rate, while Sylhet division had the lowest vaccination coverage.
Conclusion

Although there was an improvement in these nationally representative surveys from 2011 to 2017-18, a portion of children still needs to be vaccinated to ensure full immunization coverage. To achieve 100% immunization coverage for all Bangladeshi children, policymakers must integrate vaccine programs with personalized health messaging and assurances of health safety for impoverished children and low-educated mothers. Increased institutional deliveries and prenatal care visits by mothers could assist increase their children's vaccination coverage.


Introduction
Promoting the health and wellbeing of children is one of the greatest responsibilities that nations, regions, communities and households have had to bear over the years. Protecting children against diseases constitutes one of the fundamental concerns in the promotion of children's health and wellbeing. This could come in varied ways including the provision of safe drinking water, meeting the nutritional needs of children, as well as vaccinating children against preventable diseases [1,2]. This study emphasizes childhood vaccination as a technique for improving children's health and well-being.
Vaccinations are widely acknowledged as one of the safest and most cost-effective ways to protect children against infectious diseases such as tuberculosis and measles [3]. Thus, childhood vaccination has been increasing over the past decades [4]. For example, the vaccination of children against diphtheria-tetanus-pertussis (DTP3) increased astoundingly from a global coverage of 20% in 1980 to 85% in 2019 [5]. More profound is the evidence that vaccination averts between 2-3 million deaths attributable to vaccine-preventive diseases such as Diphtheria, Pertussis, Tetanus, and Measles among children under-ve every year [6,7].
Although the world has seen remarkable improvements in childhood vaccination, achieving complete coverage over time remains an important public health concern [8]. Not every child is getting vaccinated.
For instance, 19.4 million infants did not receive basic vaccination as at the end of 2019 [9,10]. Most of these de ciencies in childhood vaccination coverage are recorded in low-and-middle-income countries (LMICs). Evidence from Ethiopia suggests that in 2018, more than 1.2 million children were not vaccinated with the rst dose of measles vaccines [11]. The WHO asserts that, in remote rural areas of LMICs, only 1 out of 20 children have access to vaccination [12].
The consequences of not achieving complete childhood vaccination cannot be underestimated. Vaccination provides an opportunity to avert millions of deaths and a host of vaccine-preventable diseases among children [3]. Within the framework of the WHO, children who miss scheduled vaccinations for any reason due to health facility problems such as canceled vaccination schedules or vaccine stock-outs are categorized as having incomplete vaccination [13]. Denying children access to a complete dose of vaccines would be catastrophic as a countless number of children will die or develop some form of disabilities [14]. As such, it is imperative to understand the nuances that characterize childhood vaccination coverage over time.
Available evidence suggests that there are several factors that in uence the uptake of vaccination for children under-ve. For instance, a qualitative study by Jalloh et al. [15] indicates that perceived beliefs about the side effects coupled with concerns about receiving multiple vaccines on the same day were signi cant barriers to the uptake of childhood vaccination and its coverage. Also, other studies from Ghana [16] and Nigeria [17] suggest that urban residency and higher maternal education are associated with higher likelihoods of complete childhood vaccination coverage.
Since 1979, the Government of Bangladesh (GOB) has started vaccinations against six preventable diseases (tuberculosis; diphtheria, pertussis, and tetanus; polio; and measles) through the Expanded Program on Immunization (EPI) [18]. According to the Bangladesh Immunization guidelines, children who have received one dose of the vaccine against tuberculosis, Bacille Calmette-Guerin (BCG), three doses of a pentavalent vaccine (DPT, Hib, and HepB), three doses of the polio vaccine (excluding the polio vaccine given at birth), and one dose of the measles and rubella vaccine are considered as fully vaccinated, if they would miss any of the recommended doses they will be considered as partially vaccinated [18,19].
Bangladesh as a country has attained signi cant heights in reducing childhood mortality; this is seen in the country's capacity to meet the Millennium Development Goal 4 [20]. Through the implementation of the WHO's Expanded Programme on Immunization (EPI), Bangladesh was able to commit su ciently towards the promotion of childhood vaccination coverage which saw a sustained impact on childhood mortality, reducing it from 133 deaths per 1000 live births in 1993 to 46 deaths per 1000 in 2014 [20]. In a bid to augment efforts toward childhood vaccination coverage, the Bangladeshi government came up with three nationwide supplementary immunization activities (SIAs) from 2000 to 2016 [21]. These initiatives resulted in a signi cant decline in the incidence of measles, from 14 745 incident cases in 2010 to 972 in 2016 [15,21]. Nevertheless, Bangladesh faces high levels of childhood mortality which places the country among the top ten countries with the highest rates of childhood mortality, with vaccinepreventable diseases being the causes of these mortalities [15,22]. This makes Bangladesh an opportune context to understand childhood vaccine coverage and its concomitant factors.
Bangladesh relies on composite estimates based on administrative coverage data gathered from healthcare providers, population-based household surveys, and governmental agencies [20]. However, due to the incompleteness and mistakes associated with the original collection of data on childhood immunization in Bangladesh, such estimates are frequently incorrect [20]. As a result, utilizing a nationally representative survey provides much more clarity and strong data to investigate the factors that in uence vaccination coverage [23]. Using nationally representative data from the BDHS in 2011, 2014, and 2017-18, the study aims to track the vaccination status of children aged 12 to 59 months and examine the factors that in uence full immunization coverage.

Study population
The current study utilized three recent nationally representative cross-sectional Bangladesh demographic health survey data (BDHS) from 2011 to 2017-18. Bangladesh is a south-eastern Asian country that has administratively divided into eight divisions: Dhaka, Chittagong, Khulna, Mymensingh, Rajshahi, Rangpur and Sylhet. The survey included both urban and rural households from all administrative regions. The data was collected using two-stage strati ed cluster sampling design of the household. At the rst stage, enumeration areas were selected with probability proportional to sizes like 672 in 2017-18, 600 in both 2014 and 2011 BDHS respectively. After getting the enumeration area, on average 30 households were selected from each cluster using systematic sample selection. Detailed descriptions of the DHS sampling design were published with the BDHS survey report [19,24,25].

Outcome assessment
The outcome of interest in this study was to look over the Child vaccination status among children aged 12-59 months in Bangladesh. The government of Bangladesh started the Expanded Program on Immunization (EPI) against six preventable diseases: tuberculosis, especially the extrapulmonary forms (Bacille Calmette-Guérin [BCG] vaccine); diphtheria, pertussis, and tetanus (DPT vaccine); poliomyelitis (oral polio vaccine [OPV]); and measles (measles vaccine) among children in 1979 [24]. Table 1 shows the basic vaccine schedule by time. Children aged greater than 12 months are considered to be fully vaccinated if they have the BCG vaccine, three doses of polio, three doses of DPT and one dose of measles at any time before the survey. Partially vaccinated were de ned as lacking any dose of the basic vaccination. While those who failed to take the recommended doses of vaccine were categorized as "none". Vaccination coverage information was collected in two ways from the vaccination card or from the mother's verbal report. For logistic regression, binary vaccination status was recoded as fully vaccinated and not fully vaccinated (merging partially vaccinated and no vaccinated). Hepatitis B vaccine (1-3 dose), Haemophilus in uenzae type B vaccine, inactivated polio vaccine (IPV) was not included in the current study.

Explanatory variables
According to the guidance of reviewed literature and the availability of the variables, several demographic and health variables were included in this current analysis [3,26,27]. The included variables for this study are child age, mother's age, mother's education, antenatal care (ANC) visit, place of residence, division, gender, place of delivery, number of children. Children aged 12-59 months were selected to conduct the current study who were categorized as 12-23 months, 23-35 months, 36-47 months, and 48-59 months. Maternal age was categorized as less than 24 years, 24-34 years, and above 34 years [28,29]. The household wealth index was calculated using principal component analysis of the different household assets, including oor type, lighting source, cooking fuel, main drinking water source, stool disposal facility, stool disposal site, ownership of telephone, internet access, and household head's education, etc [25] The wealth index of the household of the respondent was recategorized as poor (poorest or poorer), middle, and rich (richer, richest). Media access is an important factor for vaccination coverage which was measured by asking mothers about the number of times they read a newspaper, listen to the radio, and watch television. Adding these variables media access was recategorized as less than once a week or at least once a week. For huge non-response missing information for the variable place of delivery and antenatal care visits, those respondents were categorized in the other category. Place of delivery was recategorized as home, health facility delivery (public and private health care facility), and others (others, missing). Respondents were categorized into at least 4 visits, less than 4 visits and missing (who does not respond). Maternal education was also categorized into no education, primary education, and secondary or higher. Bangladesh has administratively divided in eight divisions: Dhaka, Chittagong, Khulna, Mymensingh, Rajshahi, Rangpur and Sylhet. Mymensingh division is created in 2015 composing some districts from northern part of Dhaka division [30]. That's why information of Mymensingh division was not available separately in 2011 and 2014 BDHS survey data.

Statistical analysis
Statistical analysis was conducted considering the children aged 12-59 months in the three recent BDHS data sets 2011, 2014 and 2017-18. To assess the changes in all basic vaccination coverage across Bangladesh, vaccination status among children aged 12-59 months for each vaccine (full, partial and none) among different survey datasets were calculated with chi-square analysis using the three survey datasets to assess the differences. The study used descriptive statistics to show the distribution of the respondents by their characteristics and the differences in the coverage between categories were tested using chi-square analysis. To explore the relationship and changes between children's age and vaccination coverage, the prevalence of fully vaccinated was plotted against their birth year which helps to assess possible cohort effects on vaccination status with a signi cant test (chi-square test) for trend.
The graphical presentation was done in the two populations: rst who have a vaccination card and then, for children with a vaccination card or had information from mother's recall. Additionally, to explore the spatial distribution and changes of this distribution over time of vaccination coverage among children aged 12-59 months in the three recent survey year maps of the rate of change of fully vaccinated over time children were created for three survey datasets separately which shows the changes in the rate of fully vaccinated children within each division. The rate of change in the vaccination coverage over time within each division was calculated using the formula: Further, three multivariable binary (fully vaccinated and not fully vaccinated) logistic regression model was used separately to estimate odds ratios with 95% con dence intervals for socioeconomic factors' association with full vaccination using three different survey datasets. A 5% signi cance level was used for each analysis in this study. All analyses were performed using the statistical package SPSS (version 23.0) and STATA (version16.0). Spatial analysis along with maps was generated using ArcGIS (version 10.5).

Ethical approval
This study used a secondary data analysis of publicly available survey data from the MEASURE DHS program (https://www.dhsprogram.com). Participants in the DHS gave informed consent at the time of participation in the original survey. These DHS survey reports are publicly available; datasets are accessible upon application. We requested the data set, and permission was granted to download and use the data for this study. Table 1 shows the basic vaccination administration schedule for children under 12 months in Bangladesh.   Table   2).

Results
In Fig 1, probable cohort effects were depicted using the entire vaccination rate, which was   In the current study, as children's age rise, so does the prevalence of full vaccination coverage, and age is substantially associated with full vaccination coverage in all survey periods (all p<0.05). Maternal age and maternal education had also a significant association with children's full vaccination coverage. The highest percentage of full vaccination coverage among children was observed when their mothers' age was between 24-34 years. Full vaccination coverage was also increased over the period with the increases of maternal educational level. In the present study, most of the children who got full vaccination were from urban areas (90.3%, 86.5%, and 89.3% respectively in three waves of BDHS). The coverage of full vaccination was also significantly associated with the media access, place of delivery, ANC visit (all p<0.05; Table 3). The adjusted regression models of the present study demonstrate that age of the child, maternal age and education, wealth index, place of residence, media access, ANC visit of the mother and number of children were significantly associated with the coverage of the full vaccination in three different waves of BDHS. The full vaccination coverage was found to be increased significantly among the older child in all three BDHS. Maternal age was found to be significantly associated with the full vaccination coverage in two-time points except in 2011.
With the increase of the age of the mother, the vaccination coverage increased, the coverage of full vaccination was higher among the children with mothers aged > 34 years, the likelihood of getting full vaccination were 1.5 (AOR=1.5, 95% CI: 1.1, 2.0) and 1.8 (AOR=1.8, 95% CI: 1.1, 3.2) times higher compared to children of mothers aged < 24 years in 2014 and 2017-18, respectively ( Table 4).
The study found that mothers who have had primary, secondary, and higher education, their  (Table 4).
In the present study, place of delivery was also significantly associated with childhood full  (Table 4).

Discussion
The Sustainable Development Goal encourages countries and governments to take steps to ensure that their national immunization programs are fully vaccinated by 2030 [31]. Therefore, in order for Bangladesh to assess its progress toward full childhood immunization, it is critical to comprehend the factors that in uence vaccination coverage over time. The goal of this study was to track the vaccination status of children aged 12 to 59 months and examine the factors that in uence full immunization coverage. Overall, our ndings show that the percentage of those who have had complete vaccination has increased signi cantly over time. This is substantiated by a recent Bangladeshi study that demonstrated a 20% increase in complete immunization coverage over time [20]. This increment in full vaccination coverage may possibly be attributable to conscious health policies such as the three nationwide supplementary immunization activities (SIAs) from 2000 to 2016 which were rolled out by the Bangladeshi government [21]. Despite the fact that full vaccination status for all vaccines increased signi cantly over time, BCG continuously had the greatest full vaccination coverage. This is consistent with the ndings of Boulton et al., who found that BCG had the highest full vaccination coverage compared to the other vaccinations [20]. This conclusion could be explained by the fact that, unlike other vaccines that are given after a few weeks (such as OPV and DTP) or months (such as measles), BCG is given at birth, minimizing the risk of not getting immunized [32].
We also observed some divisional variations from our spatial analysis of the distribution of change rate in relation to childhood vaccination coverage. It can be observed from our study that, between 2011 to 2014, a positive change rate was observed in Dhaka division while the worst situation was found in Sylhet division. This is in agreement with the ndings of Sheikh et al. [18]. This higher likelihood of incomplete vaccination in the Sylhet division was linked to the remote hilly and riverine nature of the area coupled with the fragile communication system of this area [18]. However, between 2014 to 2017-18, all divisions experienced an increase in vaccination status except Dhaka division, with the highest improvement in change rate being recorded in the Sylhet division.
Our ndings also reveal that the factors associated with childhood vaccination coverage over time spread across child, maternal and contextual factors. Concerning the child factors, our analyses revealed that the current age of the child was signi cantly associated with full vaccination coverage. Thus, the likelihood of full vaccination increased with increased age in children. Hence, older children were more likely to be fully vaccinated. The ndings align with evidence from DR Congo [3]. First, childhood vaccinations are scheduled and for that matter, it is expected that older children will be at a higher likelihood of being fully vaccinated. Moreover, Bangladesh has had a fair share of mass vaccination programs over the years [33].
Such mass vaccinations, in the perspective of Alfonso et al. [3], lead to catch-up vaccination with age, thereby predisposing older children to a higher possibility of being fully vaccinated.
In relation to maternal factors and childhood vaccination coverage, our ndings revealed that maternal age and education were signi cantly associated with vaccination coverage over time. This result is substantiated by earlier studies conducted in Bangladesh [34,35], DR Congo [36], and Ethiopia [6]. Our ndings revealed that children born to mothers older than 34 years were 1.8 times more likely to be fully vaccinated as compared to those born to mothers younger than age 24. Often, mothers younger than age 24 are inexperienced and therefore may not be knowledgeable about the timing and relevance of ensuring that their children receive the full dose of vaccinations. Hence, explaining the lower odds of full vaccination coverage among children born to mothers younger than age 24. Children born to mothers who had formal education had higher likelihoods of full vaccination coverage. A plausible justi cation of the effect of maternal education on childhood vaccination coverage could be that mothers with formal education are more likely to be knowledgeable about childhood vaccination, its bene ts, as well as being aware of the schedules [37].
At the contextual level, urban residence, delivering at a place having health facilities, and attending at least 4 ANC visits were signi cantly associated with increased full vaccination coverage. This nding is consistent with studies from Bangladesh [3], as well as studies from Ethiopia [37] and Ghana [2].
Delivering at a health facility and attending at least 4 ANC visits provides mothers with the opportunity to be exposed to health education and sensitization messages about the importance of ensuring full vaccination of their children, as well as providing parents with vaccination schedules [37]. Another possible explanation for the ndings is that mothers who receive more ANC visits may gain satisfaction with healthcare access which could potentially translate into higher vaccination coverage [38].
Concerning the urbanicity of our ndings, it could be argued from the perspective that urban areas have a substantial proportion of health professionals and health facilities; therefore, more children in the urban areas are likely to have access to full vaccination coverage compared to their counterparts in the rural areas [16,39].

Strengths and limitations
Our study has several strengths. The dataset employed in this study was nationally representative, ensuring that the ndings may be applied to children throughout Bangladesh. In addition, the DHS dataset is a validated, repeatable, standardized, and very thorough survey. This, combined with the rigorous statistical procedures used, ensures the validity of our ndings and the study's replicability in other scenarios. Nonetheless, the study has several intrinsic limitations, which should be taken into account when interpreting our ndings. Causation cannot be established because of the BDHS's use of cross-sectional design. Furthermore, information about the status of child vaccination is dependent on either immunization cards or women's self-reports; hence, there is the possibility of recall bias, which could lead to an under-or overestimation of vaccination coverage.

Conclusion
According to the study, children's immunization rates have risen over time. The national trend of childhood immunization rates showed an increasing pattern as a result of various known child, maternal, and demographic factors. To increase childhood immunization in Bangladesh, we believe that EPIs, SIAs, and other full-childhood vaccine interventions should focus on children aged 12-23 months, those born to mothers with no formal education, and younger mothers (less than 24 years). In addition, if Bangladesh is to attain 100% immunization coverage, the government must step up efforts to increase facility births and ANC attendance.