- Email: [email protected]
Maternal determinants of timely vaccination coverage among infants in rural Bangladesh
Vaccine 32 (2014) 5514–5519
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Maternal determinants of timely vaccination coverage among infants in rural Bangladesh Lavanya Vasudevan a,b , Alain B. Labrique b,∗ , Sucheta Mehra b , Lee Wu b , Orin Levine b,c , Danny Feikin b,d , Rolf Klemm b , Parul Christian b , Keith P. West Jr. b a
Duke Global Health Institute, Durham, NC, United States Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United States c Bill and Melinda Gates Foundation, Seattle, WA, United States d Centers for Disease Control and Prevention, Atlanta, GA, United States b
a r t i c l e
i n f o
Article history: Received 1 June 2013 Received in revised form 16 May 2014 Accepted 26 June 2014 Available online 12 August 2014 Keywords: Timely vaccination Crude vaccination Expanded Program on Immunization (EPI) Bangladesh Timely vaccination score Maternal determinants Antenatal care Tetanus toxoid vaccination
a b s t r a c t Background: Timely vaccination, i.e., the receipt of all scheduled vaccinations in an age-appropriate fashion, is critical for the prevention of deadly diseases in infants and achievement of the UN Millennium Development Goal to reduce infant mortality. Infants, especially in rural or underprivileged settings often receive delayed vaccinations leaving them susceptible to vaccine-preventable illnesses early in the first year of life. In this study, we examined rates of timely vaccination among 24,435 infants born in Gaibandha and Rangpur rural districts of Bangladesh from 2001 to 2007. Methods: Vaccinations due by 14 weeks of age and administered through routine government immunization services were assessed using interviews with enrolled mothers between 11 and 18 weeks postpartum. We created a Timely Vaccination (TV) score to classify infants as vaccinated fully and on schedule (TV = 1) or not (TV = 0), and used multivariable logistic regression to identify maternal characteristics associated with infant’s timely vaccination status. Results: Our results suggest that only 19% of infants in this cohort received scheduled vaccinations on time by 11–18 weeks postpartum. Mothers’ engagement in paid employment [OR = 1.13, 95% CI: 1.03–1.23], receipt of tetanus toxoid vaccination [OR = 1.24, 95% CI: 1.11–1.38], history of antenatal care [OR = 1.22, 95% CI: 1.12–1.32], or higher socioeconomic status [OR = 1.07, 95% CI: 1.03–1.11] were positively associated with timely vaccination of their infants. Mother’s perception of small infant size at birth was negatively associated with timely vaccination [OR = 0.89, 95% CI: 0.82–0.97]. Conclusion: Timely vaccination coverage of infants in rural Gaibandha and Rangpur districts is extremely low. This analysis identifies important shortcomings associated with the 1-year vaccination benchmark of routine immunization performance and suggests the need for specific interventions based on potential maternal determinants as well as known system and programmatic barriers of timely vaccination among infants in rural Bangladesh. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction The role of vaccines in reducing childhood mortality is well documented [1–3]. Since the establishment of the Expanded Program on Immunization (EPI) in Bangladesh in 1979, the country has seen a rise in the coverage of routine immunization services, and in parallel, a steady decline in infant morbidity and mortality attributable to a reduction in vaccine-preventable diseases [3–6]. The EPI schedule in Bangladesh includes vaccinations against
∗ Corresponding author. Tel.: +1 443 287 4744. E-mail address: [email protected] (A.B. Labrique). http://dx.doi.org/10.1016/j.vaccine.2014.06.092 0264-410X/© 2014 Elsevier Ltd. All rights reserved.
tuberculosis, diphtheria, tetanus, pertussis, poliomyelitis, and measles in children less than a year old (2009 schedule, Supplementary Table 1) [4,7–9]. In 2009, the national coverage for these vaccines was estimated to be over 90% [7,8,10]. However, the mortality rates among Bangladeshi infants in 2011 remained high at 37 deaths per 1000 live born infants [10]. Routine immunization services in most countries, including Bangladesh, rely on the percentage of infants receiving each recommended vaccines by 12 months of age, but measured when the infant is 12–23 months of age, as an indicator of vaccination coverage success [5,9,11,12]. However, the potential drawback with using this crude coverage estimate is that it does not adequately capture departures from recommended vaccination schedules
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
during the first year of life [9,11–16]. Vaccine doses administered to infants too soon, or at incorrect intervals are included when determining crude coverage, even though these doses do not adequately protect the infant from vaccine-preventable diseases [9,11–16]. Not surprisingly, even in countries with high crude vaccination coverage, delays in administering timely vaccinations have been associated with an increased risk of morbidity and mortality among children [9,11,13–16]. Infants living in rural areas often face a combination of low vaccination coverage and delayed vaccination [1–3,9,11,17,18]. These critical vaccination gaps are further magnified when framed against a landscape of high disease burden, prevalent micronutrient deficiencies and vulnerabilities of low birth weight and preterm birth, present in a number of developing countries [19–21]. There are limited data describing specific factors that contribute to a complete, timely vaccination status among infants living in rural areas in Bangladesh [3–5,17]. While many studies have attempted to generalize the reasons for under-vaccination across the globe, determinants of vaccination can vary widely by societal contexts, necessitating country- and context-specific analyses of these determinants [2,9,14,22]. In this study, we used data from the Johns Hopkins Vitamin A (JiVitA-1) randomized community trial to determine the timeliness of vaccinations due in the first 14 weeks of life among infants residing in the Gaibandha and Rangpur rural districts of Bangladesh, as well as describe maternal characteristics associated with timely vaccination coverage in this region.
5515
score was calculated as the primary outcome as described below. Infants included in the study were classified into two sub-cohorts depending on whether the first postpartum interview was completed between 11 and 14 weeks (sub-cohort 1) or between 15 and 18 weeks (sub-cohort 2) after birth. Infants in sub-cohort 1 were eligible to receive 5 vaccine doses, namely 1 dose of BCG and 2 doses each of DPT and OPV and infants in sub-cohort 2 were eligible to receive 7 vaccine doses, namely 1 dose of BCG and 3 doses each of DPT and OPV. It should be noted that this method of classification allowed for different buffer times for the infant to receive any particular vaccine. For instance, an infant in sub-cohort 1 had a time frame of 11–14 weeks to receive the BCG vaccine, but only a time frame of 1–4 weeks to receive the second dose of OPV or DPT vaccine. Due to the high ownership of vaccination cards for children 12–23 months of age in Bangladesh (up to 67%), the TV score across both sub-cohorts was analyzed using reports from the government issued vaccination cards alone (Table 1) [5]. Infants receiving all due vaccine doses (either 5 or 7 vaccine doses depending on when the postpartum interview was conducted) between 11 and 18 weeks of age were assigned a TV score of 1 (Table 1). Infants who were not vaccinated at all or were partially vaccinated by the date of the first postpartum interview were assigned a TV score of 0. Mother’s recall was not used for calculating this score due to questions surrounding the validity of this method and associated recall bias for estimating vaccine coverage [25,26] (see Supplementary methods). 2.3. Statistical analysis
2. Methods 2.1. Study population The JiVitA-1 randomized community trial was conducted to study the effects of micronutrient supplementation on maternal mortality in the Gaibandha and Rangpur rural districts of Bangladesh [23]. This trial is described in depth elsewhere, but in brief, a total of 60,000 women of reproductive age were enrolled and prospectively followed, with 47,091 births occurring between 2001 and 2007 [20,23]. All data used in this study on maternal determinants of timely vaccination coverage were collected through household interviews of mothers at 3 months postpartum, referred to henceforth as the first postpartum interview [23]. Infants were assessed for their eligibility to be included into this analysis based on criteria established in the vaccine literature, including only live singleton births as well as consent and completion of the postpartum interviews (Fig. 1). Since we analyzed vaccinations due by 14 weeks of age (Supplementary Table 1), only those first postpartum interviews completed between 11 and 18 weeks of age of infant were included in the analysis (see Supplementary methods). Infants with no available vaccination information (by mother’s recall or vaccination card) from the postpartum interviews were excluded [24]. Infants whose mothers died or who were adopted were excluded from this analysis, limiting the inclusion to pairs of living infants and birth mothers [13]. 2.2. Outcome Supplementary Table 1 depicts the Bangladesh EPI schedule of vaccinations for infants by 1 year of age. According to this schedule, infants should receive one dose of the Bacillus Calmette-Guerin vaccine (BCG) at birth, followed by 3 doses each of the diphtheriapertussis-tetanus (DPT) vaccine and oral polio vaccine (OPV) by 14 weeks post-birth [4,7]. Since the dates on which the infants were vaccinated were not recorded in the postpartum interviews, we used the date of the postpartum interview as a proxy for the date of most recent vaccination. The infant’s Timely Vaccination (TV)
Multivariable logistic regression was used to identify associations between maternal characteristics and infant’s TV score, after adjusting for household and infant-specific characteristics. Characteristics previously shown to be associated with timely vaccination of infants or urban/rural differences in coverage of vaccination were identified from the literature and classified as maternal, household or infant-specific [2,4,9,27]. These were analyzed both independently and jointly for their association with the outcome (TV score) as described below. Maternal variables analyzed included the age of the mother (in years) when the pregnancy was detected, number of previous live births (parity), and number of previous pregnancies (gravidity). The wantedness of pregnancy was assessed based on whether the mother reported wanting to become pregnant then, wanting to wait until later or not wanting any more children. In addition, the number of years of formal education attained by the mother was also included. The mother’s employment status was assessed based on whether she was engaged in any form of paid employment. Visits to a doctor or health center specifically for antenatal check-up, and receipt of the tetanus toxoid (TT) vaccine were included as covariates in the analysis in addition to the mother’s perception of infant’s size at birth as large, medium or small. Household socioeconomic status was expressed using a living standards index as well as using a measure of the distance from the nearest market or town to the respondent’s household. In brief, the living standards index was created using principal component analysis of the following categories of variables collected during the enrollment of women into the JiVitA trial: land ownership, productive assets, dwelling characteristics, and durable assets [28]. Infant-specific characteristics analyzed included sex and morbidity. Report of morbidity was based on whether the infant experienced any of the following conditions in the first three months of life: cough, cold, difficulty breathing, high fever, diarrhea, or dysentery. Multivariable logistic regression models were constructed using manual selection of characteristics based on their association with the outcome in univariable models, Wald tests and 2 probabilities of groups of variables in multivariable models, and previously
5516
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
Assessed for eligibility (n = 47,091)a Excluded (n = 6403, 13.6%) • No postpartum interview date (n = 4774) • Infants died at age <3 months (n = 1112) • Not singleton birth (n = 439) • Mother died (n = 27) • Infant adopted (n = 51) Eligible infants (n = 40,688, 86.4%)
Excluded (n = 16,253, 34.5%) • Parents refused consent (n = 164) • Lost to follow up (n = 855)b • Missing all vaccination information (n = 12,564)c • Early/late first postpartum interview (n = 2670)d Infants analyzed at first postpartum interview at 11_18 weeks (n = 24,435, 52.0%)
Sub-cohort 1: Infants with first postpartum Interview after 11 weeks but prior to end of 14 weeks (n = 20,929, 44.4%)e
Sub-cohort 2: Infants with first postpartum Interview after 15 weeks but prior to end of 18 weeks (n = 3506, 7.6%)f
Fig. 1. Selection criteria for infants included in the analysis based on timing of postpartum interview. a Number of infants born in the JiVitA-1 cohort between 2001 and 2007 in Gaibandha and Rangpur rural districts of Bangladesh. b Number of infants lost to follow up including cases where the parents relocated away from the study site permanently (n = 103) as well as those where the infants were incompletely followed (n = 752). c Number of postpartum interviews where no vaccination information was collected. d Number of first postpartum interviews completed before 11 weeks or after the end of 18 weeks of age of infant. e Sub-cohort where first postpartum interview was completed after 11 weeks but prior to end of 14 weeks, and is expected to receive 5 vaccine doses, namely 1 dose of BCG and 2 doses each of DPT and OPV. f Sub-cohort where first postpartum interview was completed after 15 weeks but prior to end of 18 weeks, and is expected to receive 7 vaccine doses, namely 1 dose of BCG and 3 doses each of DPT and OPV.
Table 1 Definition and distribution of TV score among infants included in the analysis. Total number of infants assessed
Number of infants with vaccination cards
24,435
18,967
24,435
18,967
Timing of postpartum interview
Vaccines due
TVa score
Infant classification based on timeliness of vaccination
Number of infants (%)b
First postpartum interview after 11 weeks but prior to end of 14 weeks (sub-cohort 1) OR First postpartum interview after 15 weeks but prior to end of 18 weeks (sub-cohort 2)
BCG ≥1 dose DPT ≥2 doses OPV ≥2 doses BCG ≥1 dose DPT ≥3 doses OPV ≥3 doses
1
Vaccinated
3595 (19%)
First postpartum interview after 11 weeks but prior to end of 14 weeks (sub-cohort 1) OR First postpartum interview after 15 weeks but prior to end of 18 weeks (sub-cohort 2)
Criteria for TV score = 1 not met at first postpartum interview
0
Not vaccinated
15,372 (81%)
Abbreviations: TV, timely vaccination score. a TV score was determined based on information in infant’s vaccination card. b Percentage calculated as the number of infants vaccinated on time out of the number of infants with vaccination cards
220 (8.6%) 2560 Abbreviations: BCG, Bacillus Calmette Guerin; DPT, Diphtheria-Pertussis-Tetanus; OPV, Oral Polio Vaccine. a Percentage calculated as number of infants vaccinated out of the number of records where vaccine and dose information is available. b Infants not expected to receive more than 2 doses of DPT and OPV if first postpartum interview was conducted at 11–14 weeks
504 (14.8%) 3405 –b –b –b OPV ≥3 doses
–b
1084 (42.3%) 2560 1762 (51.8%) 3405 3655 (23.4%) 15,601 20,194 OPV ≥2 doses
8143 (40.3%)
2396 (93.6%) 2560 3266 (95.9%) 3405 13,963 (89.5%) 15,601 20,194 OPV ≥1 dose
19,192 (95%)
194 (7.7%) 2518 336 (9.6%) 3483 –b –b –b DPT ≥3 doses
–b
1072 (42.6%) 2518 1519 (43.6%) 3483 3546 (23.7%) 14,984 20,852 DPT ≥2 doses
7245 (34.7%)
2412 (95.8%) 2518 3000 (86.1%) 3483 13,740 (91.7%) 14,984 20,852 DPT ≥1 dose
16,711 (80.1%)
2664 3210 (91.6%) 3504 15,751 (98.4%) 16,004 18,780 (89.8%) 20,918
Number of interviews with vaccine and dose information Vaccinated n (%)a Number of interviews with vaccine and dose information
Vaccination card
Vaccinated n (%)a Number of interviews with vaccine and dose information
BCG ≥1 dose
Number of interviews with vaccine and dose information
Mother’s recall Mother’s recall
Vaccinated n (%)a
Sub-cohort 2: First postpartum interview at 15–18 weeks (n = 3506) Vaccine
Table 2 Coverage of vaccination at first postpartum interview among infants included in the analysis.
Timely vaccination status was analyzed in 24,435 infants of ages 11–18 weeks residing in Gaibandha and Rangpur rural districts of Bangladesh from 2001 to 2007 (Fig. 1). In this cohort, 20,929 first postpartum interviews were conducted between 11 and 14 weeks post-birth (sub-cohort 1). These infants were assessed for the receipt of at least 5 scheduled vaccine doses (at least 1 dose of BCG and 2 doses each of DPT and OPV). In addition, 3506 first postpartum interviews were conducted between 15 and 18 weeks post-birth. These infants were assessed for the receipt of at least 7 scheduled vaccine doses (at least 1 dose of BCG and three doses each of DPT and OPV). Out of the 24,435 infants, 18,967 had government issued vaccination cards. Of the infants who had vaccination cards, 3595 (19%) were completely vaccinated on time (TV score = 1), while 15,372 infants (81%) were not completely vaccinated on time (TV score = 0), based on their age at the time of the first postpartum interview (Table 1). To ensure consistency of findings, the vaccination status of infants immunized at the first postpartum interview was compared by whether the data were obtained through mother’s recall as opposed to government issued vaccination cards and further classified by dose and type of vaccine administered (Table 2). Data for BCG vaccine showed ≥90% coverage according to mother’s recall as well as government issued vaccination cards (Table 2). According to vaccination cards, a small fraction of infants (n = 176, 0.9%) received more than the single recommended dose of BCG vaccine (Supplementary Table 2). At the first postpartum interview, the coverage for DPT and OPV was successively lower for the second and third doses, with timely vaccination coverage for the third dose at less than 10% according to vaccination cards at the first postpartum interview (Table 2). Mothers were more likely to report the receipt of four or more doses of DPT and OPV by their infants in contrast to the government-issued vaccination cards (Supplementary Table 2) suggesting that mother’s recall may not be a very reliable source of vaccination history of their infants in this setting. Univariable logistic regression revealed higher odds of a TV score of 1 if a mother had higher number of years of education [OR = 1.02, 95% CI: 1.01–1.03, p-value < 0.001], was employed in a paying job [OR = 1.14, 95% CI: 1.05–1.25, p-value = 0.002], received the TT vaccine when pregnant [OR = 1.27, 95% CI: 1.14–1.42, pvalue < 0.001], or sought antenatal care when pregnant [OR = 1.31, 95% CI: 1.21–1.41, p-value < 0.001] (Table 3). The perception of the size of the infant at birth as small by the mother [OR = 0.87, 95% CI: 0.80–0.95, p-value = 0.002] was negatively associated with a TV score of 1. Proximity of the household where the infant resided to a weekly market [OR = 0.86, 95% CI: 0.78–0.94, p-value = 0.001] was negatively associated with a TV score of 1, while proximity to a town did not show a significant association. Finally, the living standards index [OR = 1.12, 95% CI: 1.07–1.15, p-value < 0.001] was positively associated whereas sex of the infant and morbidity prior to the first postpartum interview did not show any association with a TV score of 1. The age of the mother, parity, gravidity, and wantedness of pregnancy did not show significant associations with the outcome. In the final multivariable logistic regression model, characteristics positively associated with timely vaccination included
Sub-cohort 1: First postpartum interview at 11–14 weeks (n = 20,929)
3. Results
Vaccination card
Vaccinated n (%)a
published associations with infant vaccination [2,4,9,27]. Biological relevance and statistical significance (i.e., OR greater than or less than 1 and p-value < 0.05) were considered when examining association of the characteristics with a TV score of 1. The final model selection was based on a combination of Hosmer–Lemeshow goodness-of-fit tests, Akaike Information Criteria, as well as the total number of variables used to generate the model. All statistical analyses were performed using Stata version 12 (Stata Corp, College Station, TX).
5517
2626 (98.6%)
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
5518
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
Table 3 Results of univariable and multivariable logistic regression evaluating the association of maternal, household and infant characteristics with TV score of infants in rural Bangladesh. Predictors
Unadjusted OR
Age Parity Gravidity Wantednessa Years of educationb Paid employment TT vaccinationc Sought antenatal care Perceived size of infant at birthd Large or medium Small Household location Neither Within 0.5 km of weekly market Within 0.5 km of town Living standards indexe Sex Morbidityf
Adjusted 95% CI
p-value
OR
95% CI
p-value
1.01 0.99 1.00 0.99 1.02 1.14 1.27 1.31
1.00–1.01 0.96–1.01 0.97–1.01 0.92–1.06 1.01–1.03 1.05–1.25 1.14–1.42 1.21–1.41
0.07 0.37 0.44 0.81 <0.001* 0.002* <0.001* <0.001*
– – – – – 1.13 1.24 1.22
– – – – – 1.03–1.23 1.11–1.38 1.12–1.32
– – – – – 0.008* <0.001* <0.001*
REF 0.87
0.80–0.95
0.002*
REF 0.89
0.82–0.97
0.009*
REF 0.86 0.90 1.12 0.99 1.06
–
–
–
0.78–0.94 0.54–1.48 1.07–1.15 0.92–1.06 0.94–1.20
0.001* 0.67 <0.001* 0.82 0.35
1.07 – –
1.03–1.11 – –
0.001* – –
Abbreviations: OR, odds ratio; CI, confidence interval; TT, tetanus toxoid. * Indicates statistically significant p-value (<0.05). a Whether pregnancy was desired at the time of detection. b Years of formal schooling c Whether mother received tetanus toxoid vaccination as part of antenatal care. d Mother’s perception e Living Standards Index incorporates type of material in household floor, walls and roof, and ownership of durable assets. f Classified as morbid if the infant ever experienced one or more of the following conditions anytime between birth and the time of the first postpartum interview: cough, cold or difficulty breathing, diarrhea, dysentery, high fever.
paid employment [OR = 1.13, 95% CI: 1.03–1.23, p-value = 0.008], TT vaccination [OR = 1.24, 95% CI: 1.11–1.38, p-value < 0.001] and antenatal care seeking behavior [OR = 1.22, 95% CI: 1.12–1.32, pvalue < 0.001] among mothers. The perceived size of infant at birth as small [OR = 0.89, 95% CI: 0.82–0.97, p-value = 0.009] was negatively associated with a timely vaccination while the living standards index [OR = 1.07, 95% CI: 1.03–1.11, p-value = 0.001] was positively associated with the outcome. 4. Discussion In this study, we analyzed maternal characteristics that influence timely vaccination status in the early months of life of infants in Gaibandha and Rangpur rural districts of Bangladesh. A 2011 survey of timeliness of vaccination in Bangladesh reported timely BCG vaccination (at birth) in only 14.5% of the infants, with a median delay of 7 weeks [9]. Our analysis suggests that vaccination delays are not limited to BCG administration alone. In the representative, rural population of 24,435 infants analyzed in this study, only 19% of infants received all doses of scheduled vaccines (BCG, DPT and OPV) due by 14 weeks of age (Table 1). In addition, timely vaccination coverage of DPT and OPV in this cohort were lower than reported national statistics, which measure vaccination coverage by 12 months [4,7,10]. Specifically, coverage was successively lower for the second and third recommended doses of the DPT and OPV vaccines, with fewer than 10% of all infants receiving all three doses of DPT and OPV by 14 weeks (Table 2), consistent with findings from other studies [3,9,14,29]. These results suggest that infants receive catch-up vaccinations between 15 and 52 weeks of age. Previous studies examining inequities in vaccination coverage in Bangladesh reported lower vaccination rates among girls, children of poorer and less educated parents, those living in rural or isolated areas or urban slums, and children from ethnic minority groups [6,17]. Our results suggest that classic determinants of infant vaccination status such as age, education, or parity of the mother, sex of infant, and household sizes, which may be
significant in other settings, are not strongly predictive of infant vaccination status in the rural setting of our study population [2,27]. Interestingly, previous utilization of healthcare services such as antenatal care and TT vaccination by the mother were strongly associated with timely vaccination status of the infant. An analysis of data from 166 Demographic and Health Surveys from 67 countries by the Swiss Tropical and Public Health Institute yielded similar results [30]. In addition, inequities in antenatal care coverage have been reported between urban and rural areas in Bangladesh in parallel to differing rates of vaccinations suggesting a strong link between the two [5]. Socioeconomic status of the infant’s family was also a strong determinant of whether infants were vaccinated on time, as has been previously reported [5,24]. This is also reflected in the finding that mothers engaged in paid employment were more likely to get their infants vaccinated according to schedule. A link between socioeconomic status and education status of the mother has been proposed [3], but mother’s education status was not strongly associated with a child’s vaccination status in this analysis. Finally, small ‘perceived size of infant at birth’ was negatively associated with vaccination status. Others have suggested that delaying the age of first vaccination to 28 days or greater is associated with higher risk of non-completion of the vaccination schedule [11,14]. Our results reiterate the importance of identifying and counseling families who may be at risk of delaying vaccinations due to small perceived size of infant at birth. The JiVitA-1 cohort used in this analysis is reflective of the rural population of Bangladesh [23]. However, there are several limitations to this analysis. The JiVitA-1 cohort was established to study the effects of vitamin A supplementation on reducing pregnancy-related maternal and infant mortality and information on infant vaccination status was recorded as part of the household surveys conducted during the study [20,21,23]. However, specific vaccination-related information such as the dates of vaccination, reasons for missed vaccinations, or previously known predictors of vaccination were not collected actively during the study. Hence, our
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
analysis is limited to the existing types and quantities of available data. For instance, we used the date of the postpartum interviews as a proxy for the timeliness of vaccinations recommended by 14 weeks of age. As specified in the methods section, the use of postpartum interview dates as a proxy for dates of vaccination allowed for different buffer times for the infant to receive any particular vaccine. Hence, within a sub cohort, infants interviewed earlier had a shorter window of opportunity to get vaccinated on time as opposed to infants interviewed later. Some factors that may potentially be very strong predictors of infant vaccination status in rural locations such as accessibility of health care services, or programmatic factors pertaining to vaccine worker training, attitude, missed outreach and advocacy opportunities or inadequate supply of vaccines were not studied [4,17,22,27]. In addition, we did not distinguish between completely unvaccinated and partially vaccinated infants [11]. This study highlights the need to improve timeliness of vaccinations among infants in rural Bangladesh, especially during the early months of life. We concur with other reports in the literature that timely vaccination coverage is a better performance metric for routine immunization services than crude vaccination rates alone [9,11–16]. Illustrative solutions worth exploring may include timely identification and counseling of mothers whose infants are not vaccinated on schedule. Digital or mobile phone based recordkeeping systems that can prioritize and alert vaccinators to such cases may have greater utility over paper-based systems in addressing these constraints. Findings from this study may be used to guide local, regional and national programs aiming to improve timely coverage of vaccinations in rural populations. Funding source This work was funded partly through a Grand Challenges Explorations Phase I grant to AL and LV. The funding source had no involvement in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. The field trial and associated studies were supported through Grant No. 614 (Global Control of Micronutrient Deficiency) from the Bill and Melinda Gates Foundation, Seattle WA; Cooperative Agreement HRN-A-0097-00015-00 and Global Research Activity GHS-A-00-03-00019-00 between the Office of Health, Infectious Diseases and Nutrition, US Agency for International Development (USAID), Washington, DC; the USAID Mission, Dhaka, Bangladesh; the Sight and Life Global Nutrition Research Institute, Baltimore, MD; the Micronutrient Initiative/Department of Foreign Affairs, Trade and Development, Ottawa, Ontario; and the Nutrilite Health Institute, Access Business Group LLC, Buena Park, CA; and the Ministry of Health and Family Welfare, Government of Bangladesh, Dhaka. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine. 2014.06.092. References [1] World Health Organization. State of the world’s vaccines and immunization; 2009.
5519
[2] Rainey JJ, Watkins M, Ryman TK, Sandhu P, Bo A, Banerjee K. Reasons related to non-vaccination and under-vaccination of children in low and middle income countries: findings from a systematic review of the published literature, 1999–2009. Vaccine 2011;29:8215–21. [3] Breiman RF, Streatfield PK, Phelan M, Shifa N, Rashid M, Yunus M. Effect of infant immunisation on childhood mortality in rural Bangladesh: analysis of health and demographic surveillance data. Lancet 2004;364:2204–11. [4] Khan MM, Yoder RA. Technical Report No. 24. Expanded program on immunization in Bangladesh: cost, cost-effectiveness, and financing estimates, 2012. Bethesda, MD: Partnerships Heal Reform Proj Abt Assoc Inc.; 1998. p. 1–78. [5] National Institute of Population Research and Training (NIPORT), Mitra and Associates, and ICF International. Bangladesh Demographic and Health Survey 2011. MA, USA: Dhaka, Bangladesh/Calverton; 2013. [6] Rubayet S, Shahidullah M, Hossain A, Corbett E, Moran AC, Mannan I, et al. Newborn survival in Bangladesh: a decade of change and future implications. Health Policy Plan 2012;27(Suppl. 3):iii40–56. [7] World Health Organization Regional Office for South-East Asia. Bangladesh 2009 EPI Fact Sheet; 2010. [8] World Health Organization. Immunization Profile – Bangladesh; 2011. [9] Akmatov MK, Mikolajczyk RT. Timeliness of childhood vaccinations in 31 low and middle-income countries. J Epidemiol Community Health 2011:5. [10] UNICEF, World Health Organization. Immunization Summary. A statistical reference containing data through 2011; 2013. [11] Bicaba A, Haddad S, Kabore M, Taminy E, Feletto M, Fournier P. Monitoring the performance of the Expanded Program on Immunization: the case of Burkina Faso. BMC Int Health Hum Rights 2009;9(Suppl. 1):S12. [12] Expanded Programme on Immunization of the Department of Immunization Vaccines and Biologicals. Training for mid-level managers (MLM) 7. The EPI coverage survey. Geneva, Switzerland: EPI; 2008. [13] Fadnes LT, Nankabirwa V, Sommerfelt H, Tylleskär T, Tumwine JK, Engebretsen IMS. Is vaccination coverage a good indicator of age-appropriate vaccination? A prospective study from Uganda. Vaccine 2011;29:3564–70. [14] Sadoh AE, Eregie CO. Timeliness and completion rate of immunization among Nigerian children attending a clinic-based immunization service. J Health Popul Nutr 2009;27:391–5. [15] Murray CJL, Shengelia B, Gupta N, Moussavi S, Tandon A, Thieren M. Validity of reported vaccination coverage in 45 countries. Lancet 2003;362:1022–7. [16] Luman ET, McCauley MM, Stokley S, Chu SY, Pickering LK. Timeliness of childhood immunizations. Pediatrics 2002;110:935–9. [17] Chowdhury AMR, Salam Bhuiya A, Mahmud S, Abdus AKM, Karim F. The immunization divide: who gets vaccinated in Bangladesh? J Health Popul Nutr 2003;21:193–204. [18] World Health Organization. MDG 4: reduce child mortality; 2013. [19] West Jr KP, Christian P, Katz J, Labrique A, Klemm R, Sommer A. Effect of vitamin A supplementation on maternal survival. Lancet 2010;376:873–4 [author reply 874]. [20] West Jr KP, Christian P, Labrique AB, Rashid M, Shamim AA, Klemm RD, et al. Effects of vitamin A or beta carotene supplementation on pregnancy-related mortality and infant mortality in rural Bangladesh: a cluster randomized trial. J Am Med Assoc 2011;305:1986–95. [21] Klemm RD, Labrique AB, Christian P, Rashid M, Shamim AA, Katz J, et al. Newborn vitamin A supplementation reduced infant mortality in rural Bangladesh. Pediatrics 2008;122:e242–50. [22] Favin M, Steinglass R, Fields R, Banerjee K, Sawhney M. Why children are not vaccinated: a review of the grey literature. Int Health 2012;4:229–38. [23] Labrique AB, Christian P, Klemm RD, Rashid M, Shamim AA, Massie A, et al. A cluster-randomized, placebo-controlled, maternal vitamin A or beta-carotene supplementation trial in Bangladesh: design and methods. Trials 2011; 12:102. [24] Tadesse H, Deribew A, Woldie M. Predictors of defaulting from completion of child immunization in south Ethiopia, May 2008: a case–control study. BMC Public Health 2009;9:150. [25] Suarez L, Simpson DM, Smith DR. Errors and correlates in parental recall of child immunizations: effects on vaccination coverage estimates. Pediatrics 1997;99:e3. [26] Ramakrishnan R, Rao TV, Sundaramoorthy L, Joshua V. Magnitude of recall bias in the estimation of immunization coverage and its determinants. Indian Pediatr 1999;36:881–5. [27] Weiss WM, Winch PJ, Burnham G. Factors associated with missed vaccination during mass immunization campaigns. J Health Popul Nutr 2009;27:358–67. [28] Gunnsteinsson S, Labrique AB, West Jr KP, Christian P, Mehra S, Shamim AA, et al. Constructing indices of rural living standards in Northwestern Bangladesh. J Health Popul Nutr 2010;28:509–19. [29] Hanson CM, Furrer E, Schwalbe N, Berkley S. Prediction of immunisation performance. Lancet 2013;381:349–50. [30] Hilber AM, Bosch-Capblanch X, Schindler C, Beck L, Sécula F, Mckenzie O, et al. Gender and immunisation. In: Summary report for SAGE. 2010.
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Maternal determinants of timely vaccination coverage among infants in rural Bangladesh Lavanya Vasudevan a,b , Alain B. Labrique b,∗ , Sucheta Mehra b , Lee Wu b , Orin Levine b,c , Danny Feikin b,d , Rolf Klemm b , Parul Christian b , Keith P. West Jr. b a
Duke Global Health Institute, Durham, NC, United States Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United States c Bill and Melinda Gates Foundation, Seattle, WA, United States d Centers for Disease Control and Prevention, Atlanta, GA, United States b
a r t i c l e
i n f o
Article history: Received 1 June 2013 Received in revised form 16 May 2014 Accepted 26 June 2014 Available online 12 August 2014 Keywords: Timely vaccination Crude vaccination Expanded Program on Immunization (EPI) Bangladesh Timely vaccination score Maternal determinants Antenatal care Tetanus toxoid vaccination
a b s t r a c t Background: Timely vaccination, i.e., the receipt of all scheduled vaccinations in an age-appropriate fashion, is critical for the prevention of deadly diseases in infants and achievement of the UN Millennium Development Goal to reduce infant mortality. Infants, especially in rural or underprivileged settings often receive delayed vaccinations leaving them susceptible to vaccine-preventable illnesses early in the first year of life. In this study, we examined rates of timely vaccination among 24,435 infants born in Gaibandha and Rangpur rural districts of Bangladesh from 2001 to 2007. Methods: Vaccinations due by 14 weeks of age and administered through routine government immunization services were assessed using interviews with enrolled mothers between 11 and 18 weeks postpartum. We created a Timely Vaccination (TV) score to classify infants as vaccinated fully and on schedule (TV = 1) or not (TV = 0), and used multivariable logistic regression to identify maternal characteristics associated with infant’s timely vaccination status. Results: Our results suggest that only 19% of infants in this cohort received scheduled vaccinations on time by 11–18 weeks postpartum. Mothers’ engagement in paid employment [OR = 1.13, 95% CI: 1.03–1.23], receipt of tetanus toxoid vaccination [OR = 1.24, 95% CI: 1.11–1.38], history of antenatal care [OR = 1.22, 95% CI: 1.12–1.32], or higher socioeconomic status [OR = 1.07, 95% CI: 1.03–1.11] were positively associated with timely vaccination of their infants. Mother’s perception of small infant size at birth was negatively associated with timely vaccination [OR = 0.89, 95% CI: 0.82–0.97]. Conclusion: Timely vaccination coverage of infants in rural Gaibandha and Rangpur districts is extremely low. This analysis identifies important shortcomings associated with the 1-year vaccination benchmark of routine immunization performance and suggests the need for specific interventions based on potential maternal determinants as well as known system and programmatic barriers of timely vaccination among infants in rural Bangladesh. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction The role of vaccines in reducing childhood mortality is well documented [1–3]. Since the establishment of the Expanded Program on Immunization (EPI) in Bangladesh in 1979, the country has seen a rise in the coverage of routine immunization services, and in parallel, a steady decline in infant morbidity and mortality attributable to a reduction in vaccine-preventable diseases [3–6]. The EPI schedule in Bangladesh includes vaccinations against
∗ Corresponding author. Tel.: +1 443 287 4744. E-mail address: [email protected] (A.B. Labrique). http://dx.doi.org/10.1016/j.vaccine.2014.06.092 0264-410X/© 2014 Elsevier Ltd. All rights reserved.
tuberculosis, diphtheria, tetanus, pertussis, poliomyelitis, and measles in children less than a year old (2009 schedule, Supplementary Table 1) [4,7–9]. In 2009, the national coverage for these vaccines was estimated to be over 90% [7,8,10]. However, the mortality rates among Bangladeshi infants in 2011 remained high at 37 deaths per 1000 live born infants [10]. Routine immunization services in most countries, including Bangladesh, rely on the percentage of infants receiving each recommended vaccines by 12 months of age, but measured when the infant is 12–23 months of age, as an indicator of vaccination coverage success [5,9,11,12]. However, the potential drawback with using this crude coverage estimate is that it does not adequately capture departures from recommended vaccination schedules
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
during the first year of life [9,11–16]. Vaccine doses administered to infants too soon, or at incorrect intervals are included when determining crude coverage, even though these doses do not adequately protect the infant from vaccine-preventable diseases [9,11–16]. Not surprisingly, even in countries with high crude vaccination coverage, delays in administering timely vaccinations have been associated with an increased risk of morbidity and mortality among children [9,11,13–16]. Infants living in rural areas often face a combination of low vaccination coverage and delayed vaccination [1–3,9,11,17,18]. These critical vaccination gaps are further magnified when framed against a landscape of high disease burden, prevalent micronutrient deficiencies and vulnerabilities of low birth weight and preterm birth, present in a number of developing countries [19–21]. There are limited data describing specific factors that contribute to a complete, timely vaccination status among infants living in rural areas in Bangladesh [3–5,17]. While many studies have attempted to generalize the reasons for under-vaccination across the globe, determinants of vaccination can vary widely by societal contexts, necessitating country- and context-specific analyses of these determinants [2,9,14,22]. In this study, we used data from the Johns Hopkins Vitamin A (JiVitA-1) randomized community trial to determine the timeliness of vaccinations due in the first 14 weeks of life among infants residing in the Gaibandha and Rangpur rural districts of Bangladesh, as well as describe maternal characteristics associated with timely vaccination coverage in this region.
5515
score was calculated as the primary outcome as described below. Infants included in the study were classified into two sub-cohorts depending on whether the first postpartum interview was completed between 11 and 14 weeks (sub-cohort 1) or between 15 and 18 weeks (sub-cohort 2) after birth. Infants in sub-cohort 1 were eligible to receive 5 vaccine doses, namely 1 dose of BCG and 2 doses each of DPT and OPV and infants in sub-cohort 2 were eligible to receive 7 vaccine doses, namely 1 dose of BCG and 3 doses each of DPT and OPV. It should be noted that this method of classification allowed for different buffer times for the infant to receive any particular vaccine. For instance, an infant in sub-cohort 1 had a time frame of 11–14 weeks to receive the BCG vaccine, but only a time frame of 1–4 weeks to receive the second dose of OPV or DPT vaccine. Due to the high ownership of vaccination cards for children 12–23 months of age in Bangladesh (up to 67%), the TV score across both sub-cohorts was analyzed using reports from the government issued vaccination cards alone (Table 1) [5]. Infants receiving all due vaccine doses (either 5 or 7 vaccine doses depending on when the postpartum interview was conducted) between 11 and 18 weeks of age were assigned a TV score of 1 (Table 1). Infants who were not vaccinated at all or were partially vaccinated by the date of the first postpartum interview were assigned a TV score of 0. Mother’s recall was not used for calculating this score due to questions surrounding the validity of this method and associated recall bias for estimating vaccine coverage [25,26] (see Supplementary methods). 2.3. Statistical analysis
2. Methods 2.1. Study population The JiVitA-1 randomized community trial was conducted to study the effects of micronutrient supplementation on maternal mortality in the Gaibandha and Rangpur rural districts of Bangladesh [23]. This trial is described in depth elsewhere, but in brief, a total of 60,000 women of reproductive age were enrolled and prospectively followed, with 47,091 births occurring between 2001 and 2007 [20,23]. All data used in this study on maternal determinants of timely vaccination coverage were collected through household interviews of mothers at 3 months postpartum, referred to henceforth as the first postpartum interview [23]. Infants were assessed for their eligibility to be included into this analysis based on criteria established in the vaccine literature, including only live singleton births as well as consent and completion of the postpartum interviews (Fig. 1). Since we analyzed vaccinations due by 14 weeks of age (Supplementary Table 1), only those first postpartum interviews completed between 11 and 18 weeks of age of infant were included in the analysis (see Supplementary methods). Infants with no available vaccination information (by mother’s recall or vaccination card) from the postpartum interviews were excluded [24]. Infants whose mothers died or who were adopted were excluded from this analysis, limiting the inclusion to pairs of living infants and birth mothers [13]. 2.2. Outcome Supplementary Table 1 depicts the Bangladesh EPI schedule of vaccinations for infants by 1 year of age. According to this schedule, infants should receive one dose of the Bacillus Calmette-Guerin vaccine (BCG) at birth, followed by 3 doses each of the diphtheriapertussis-tetanus (DPT) vaccine and oral polio vaccine (OPV) by 14 weeks post-birth [4,7]. Since the dates on which the infants were vaccinated were not recorded in the postpartum interviews, we used the date of the postpartum interview as a proxy for the date of most recent vaccination. The infant’s Timely Vaccination (TV)
Multivariable logistic regression was used to identify associations between maternal characteristics and infant’s TV score, after adjusting for household and infant-specific characteristics. Characteristics previously shown to be associated with timely vaccination of infants or urban/rural differences in coverage of vaccination were identified from the literature and classified as maternal, household or infant-specific [2,4,9,27]. These were analyzed both independently and jointly for their association with the outcome (TV score) as described below. Maternal variables analyzed included the age of the mother (in years) when the pregnancy was detected, number of previous live births (parity), and number of previous pregnancies (gravidity). The wantedness of pregnancy was assessed based on whether the mother reported wanting to become pregnant then, wanting to wait until later or not wanting any more children. In addition, the number of years of formal education attained by the mother was also included. The mother’s employment status was assessed based on whether she was engaged in any form of paid employment. Visits to a doctor or health center specifically for antenatal check-up, and receipt of the tetanus toxoid (TT) vaccine were included as covariates in the analysis in addition to the mother’s perception of infant’s size at birth as large, medium or small. Household socioeconomic status was expressed using a living standards index as well as using a measure of the distance from the nearest market or town to the respondent’s household. In brief, the living standards index was created using principal component analysis of the following categories of variables collected during the enrollment of women into the JiVitA trial: land ownership, productive assets, dwelling characteristics, and durable assets [28]. Infant-specific characteristics analyzed included sex and morbidity. Report of morbidity was based on whether the infant experienced any of the following conditions in the first three months of life: cough, cold, difficulty breathing, high fever, diarrhea, or dysentery. Multivariable logistic regression models were constructed using manual selection of characteristics based on their association with the outcome in univariable models, Wald tests and 2 probabilities of groups of variables in multivariable models, and previously
5516
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
Assessed for eligibility (n = 47,091)a Excluded (n = 6403, 13.6%) • No postpartum interview date (n = 4774) • Infants died at age <3 months (n = 1112) • Not singleton birth (n = 439) • Mother died (n = 27) • Infant adopted (n = 51) Eligible infants (n = 40,688, 86.4%)
Excluded (n = 16,253, 34.5%) • Parents refused consent (n = 164) • Lost to follow up (n = 855)b • Missing all vaccination information (n = 12,564)c • Early/late first postpartum interview (n = 2670)d Infants analyzed at first postpartum interview at 11_18 weeks (n = 24,435, 52.0%)
Sub-cohort 1: Infants with first postpartum Interview after 11 weeks but prior to end of 14 weeks (n = 20,929, 44.4%)e
Sub-cohort 2: Infants with first postpartum Interview after 15 weeks but prior to end of 18 weeks (n = 3506, 7.6%)f
Fig. 1. Selection criteria for infants included in the analysis based on timing of postpartum interview. a Number of infants born in the JiVitA-1 cohort between 2001 and 2007 in Gaibandha and Rangpur rural districts of Bangladesh. b Number of infants lost to follow up including cases where the parents relocated away from the study site permanently (n = 103) as well as those where the infants were incompletely followed (n = 752). c Number of postpartum interviews where no vaccination information was collected. d Number of first postpartum interviews completed before 11 weeks or after the end of 18 weeks of age of infant. e Sub-cohort where first postpartum interview was completed after 11 weeks but prior to end of 14 weeks, and is expected to receive 5 vaccine doses, namely 1 dose of BCG and 2 doses each of DPT and OPV. f Sub-cohort where first postpartum interview was completed after 15 weeks but prior to end of 18 weeks, and is expected to receive 7 vaccine doses, namely 1 dose of BCG and 3 doses each of DPT and OPV.
Table 1 Definition and distribution of TV score among infants included in the analysis. Total number of infants assessed
Number of infants with vaccination cards
24,435
18,967
24,435
18,967
Timing of postpartum interview
Vaccines due
TVa score
Infant classification based on timeliness of vaccination
Number of infants (%)b
First postpartum interview after 11 weeks but prior to end of 14 weeks (sub-cohort 1) OR First postpartum interview after 15 weeks but prior to end of 18 weeks (sub-cohort 2)
BCG ≥1 dose DPT ≥2 doses OPV ≥2 doses BCG ≥1 dose DPT ≥3 doses OPV ≥3 doses
1
Vaccinated
3595 (19%)
First postpartum interview after 11 weeks but prior to end of 14 weeks (sub-cohort 1) OR First postpartum interview after 15 weeks but prior to end of 18 weeks (sub-cohort 2)
Criteria for TV score = 1 not met at first postpartum interview
0
Not vaccinated
15,372 (81%)
Abbreviations: TV, timely vaccination score. a TV score was determined based on information in infant’s vaccination card. b Percentage calculated as the number of infants vaccinated on time out of the number of infants with vaccination cards
220 (8.6%) 2560 Abbreviations: BCG, Bacillus Calmette Guerin; DPT, Diphtheria-Pertussis-Tetanus; OPV, Oral Polio Vaccine. a Percentage calculated as number of infants vaccinated out of the number of records where vaccine and dose information is available. b Infants not expected to receive more than 2 doses of DPT and OPV if first postpartum interview was conducted at 11–14 weeks
504 (14.8%) 3405 –b –b –b OPV ≥3 doses
–b
1084 (42.3%) 2560 1762 (51.8%) 3405 3655 (23.4%) 15,601 20,194 OPV ≥2 doses
8143 (40.3%)
2396 (93.6%) 2560 3266 (95.9%) 3405 13,963 (89.5%) 15,601 20,194 OPV ≥1 dose
19,192 (95%)
194 (7.7%) 2518 336 (9.6%) 3483 –b –b –b DPT ≥3 doses
–b
1072 (42.6%) 2518 1519 (43.6%) 3483 3546 (23.7%) 14,984 20,852 DPT ≥2 doses
7245 (34.7%)
2412 (95.8%) 2518 3000 (86.1%) 3483 13,740 (91.7%) 14,984 20,852 DPT ≥1 dose
16,711 (80.1%)
2664 3210 (91.6%) 3504 15,751 (98.4%) 16,004 18,780 (89.8%) 20,918
Number of interviews with vaccine and dose information Vaccinated n (%)a Number of interviews with vaccine and dose information
Vaccination card
Vaccinated n (%)a Number of interviews with vaccine and dose information
BCG ≥1 dose
Number of interviews with vaccine and dose information
Mother’s recall Mother’s recall
Vaccinated n (%)a
Sub-cohort 2: First postpartum interview at 15–18 weeks (n = 3506) Vaccine
Table 2 Coverage of vaccination at first postpartum interview among infants included in the analysis.
Timely vaccination status was analyzed in 24,435 infants of ages 11–18 weeks residing in Gaibandha and Rangpur rural districts of Bangladesh from 2001 to 2007 (Fig. 1). In this cohort, 20,929 first postpartum interviews were conducted between 11 and 14 weeks post-birth (sub-cohort 1). These infants were assessed for the receipt of at least 5 scheduled vaccine doses (at least 1 dose of BCG and 2 doses each of DPT and OPV). In addition, 3506 first postpartum interviews were conducted between 15 and 18 weeks post-birth. These infants were assessed for the receipt of at least 7 scheduled vaccine doses (at least 1 dose of BCG and three doses each of DPT and OPV). Out of the 24,435 infants, 18,967 had government issued vaccination cards. Of the infants who had vaccination cards, 3595 (19%) were completely vaccinated on time (TV score = 1), while 15,372 infants (81%) were not completely vaccinated on time (TV score = 0), based on their age at the time of the first postpartum interview (Table 1). To ensure consistency of findings, the vaccination status of infants immunized at the first postpartum interview was compared by whether the data were obtained through mother’s recall as opposed to government issued vaccination cards and further classified by dose and type of vaccine administered (Table 2). Data for BCG vaccine showed ≥90% coverage according to mother’s recall as well as government issued vaccination cards (Table 2). According to vaccination cards, a small fraction of infants (n = 176, 0.9%) received more than the single recommended dose of BCG vaccine (Supplementary Table 2). At the first postpartum interview, the coverage for DPT and OPV was successively lower for the second and third doses, with timely vaccination coverage for the third dose at less than 10% according to vaccination cards at the first postpartum interview (Table 2). Mothers were more likely to report the receipt of four or more doses of DPT and OPV by their infants in contrast to the government-issued vaccination cards (Supplementary Table 2) suggesting that mother’s recall may not be a very reliable source of vaccination history of their infants in this setting. Univariable logistic regression revealed higher odds of a TV score of 1 if a mother had higher number of years of education [OR = 1.02, 95% CI: 1.01–1.03, p-value < 0.001], was employed in a paying job [OR = 1.14, 95% CI: 1.05–1.25, p-value = 0.002], received the TT vaccine when pregnant [OR = 1.27, 95% CI: 1.14–1.42, pvalue < 0.001], or sought antenatal care when pregnant [OR = 1.31, 95% CI: 1.21–1.41, p-value < 0.001] (Table 3). The perception of the size of the infant at birth as small by the mother [OR = 0.87, 95% CI: 0.80–0.95, p-value = 0.002] was negatively associated with a TV score of 1. Proximity of the household where the infant resided to a weekly market [OR = 0.86, 95% CI: 0.78–0.94, p-value = 0.001] was negatively associated with a TV score of 1, while proximity to a town did not show a significant association. Finally, the living standards index [OR = 1.12, 95% CI: 1.07–1.15, p-value < 0.001] was positively associated whereas sex of the infant and morbidity prior to the first postpartum interview did not show any association with a TV score of 1. The age of the mother, parity, gravidity, and wantedness of pregnancy did not show significant associations with the outcome. In the final multivariable logistic regression model, characteristics positively associated with timely vaccination included
Sub-cohort 1: First postpartum interview at 11–14 weeks (n = 20,929)
3. Results
Vaccination card
Vaccinated n (%)a
published associations with infant vaccination [2,4,9,27]. Biological relevance and statistical significance (i.e., OR greater than or less than 1 and p-value < 0.05) were considered when examining association of the characteristics with a TV score of 1. The final model selection was based on a combination of Hosmer–Lemeshow goodness-of-fit tests, Akaike Information Criteria, as well as the total number of variables used to generate the model. All statistical analyses were performed using Stata version 12 (Stata Corp, College Station, TX).
5517
2626 (98.6%)
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
5518
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
Table 3 Results of univariable and multivariable logistic regression evaluating the association of maternal, household and infant characteristics with TV score of infants in rural Bangladesh. Predictors
Unadjusted OR
Age Parity Gravidity Wantednessa Years of educationb Paid employment TT vaccinationc Sought antenatal care Perceived size of infant at birthd Large or medium Small Household location Neither Within 0.5 km of weekly market Within 0.5 km of town Living standards indexe Sex Morbidityf
Adjusted 95% CI
p-value
OR
95% CI
p-value
1.01 0.99 1.00 0.99 1.02 1.14 1.27 1.31
1.00–1.01 0.96–1.01 0.97–1.01 0.92–1.06 1.01–1.03 1.05–1.25 1.14–1.42 1.21–1.41
0.07 0.37 0.44 0.81 <0.001* 0.002* <0.001* <0.001*
– – – – – 1.13 1.24 1.22
– – – – – 1.03–1.23 1.11–1.38 1.12–1.32
– – – – – 0.008* <0.001* <0.001*
REF 0.87
0.80–0.95
0.002*
REF 0.89
0.82–0.97
0.009*
REF 0.86 0.90 1.12 0.99 1.06
–
–
–
0.78–0.94 0.54–1.48 1.07–1.15 0.92–1.06 0.94–1.20
0.001* 0.67 <0.001* 0.82 0.35
1.07 – –
1.03–1.11 – –
0.001* – –
Abbreviations: OR, odds ratio; CI, confidence interval; TT, tetanus toxoid. * Indicates statistically significant p-value (<0.05). a Whether pregnancy was desired at the time of detection. b Years of formal schooling c Whether mother received tetanus toxoid vaccination as part of antenatal care. d Mother’s perception e Living Standards Index incorporates type of material in household floor, walls and roof, and ownership of durable assets. f Classified as morbid if the infant ever experienced one or more of the following conditions anytime between birth and the time of the first postpartum interview: cough, cold or difficulty breathing, diarrhea, dysentery, high fever.
paid employment [OR = 1.13, 95% CI: 1.03–1.23, p-value = 0.008], TT vaccination [OR = 1.24, 95% CI: 1.11–1.38, p-value < 0.001] and antenatal care seeking behavior [OR = 1.22, 95% CI: 1.12–1.32, pvalue < 0.001] among mothers. The perceived size of infant at birth as small [OR = 0.89, 95% CI: 0.82–0.97, p-value = 0.009] was negatively associated with a timely vaccination while the living standards index [OR = 1.07, 95% CI: 1.03–1.11, p-value = 0.001] was positively associated with the outcome. 4. Discussion In this study, we analyzed maternal characteristics that influence timely vaccination status in the early months of life of infants in Gaibandha and Rangpur rural districts of Bangladesh. A 2011 survey of timeliness of vaccination in Bangladesh reported timely BCG vaccination (at birth) in only 14.5% of the infants, with a median delay of 7 weeks [9]. Our analysis suggests that vaccination delays are not limited to BCG administration alone. In the representative, rural population of 24,435 infants analyzed in this study, only 19% of infants received all doses of scheduled vaccines (BCG, DPT and OPV) due by 14 weeks of age (Table 1). In addition, timely vaccination coverage of DPT and OPV in this cohort were lower than reported national statistics, which measure vaccination coverage by 12 months [4,7,10]. Specifically, coverage was successively lower for the second and third recommended doses of the DPT and OPV vaccines, with fewer than 10% of all infants receiving all three doses of DPT and OPV by 14 weeks (Table 2), consistent with findings from other studies [3,9,14,29]. These results suggest that infants receive catch-up vaccinations between 15 and 52 weeks of age. Previous studies examining inequities in vaccination coverage in Bangladesh reported lower vaccination rates among girls, children of poorer and less educated parents, those living in rural or isolated areas or urban slums, and children from ethnic minority groups [6,17]. Our results suggest that classic determinants of infant vaccination status such as age, education, or parity of the mother, sex of infant, and household sizes, which may be
significant in other settings, are not strongly predictive of infant vaccination status in the rural setting of our study population [2,27]. Interestingly, previous utilization of healthcare services such as antenatal care and TT vaccination by the mother were strongly associated with timely vaccination status of the infant. An analysis of data from 166 Demographic and Health Surveys from 67 countries by the Swiss Tropical and Public Health Institute yielded similar results [30]. In addition, inequities in antenatal care coverage have been reported between urban and rural areas in Bangladesh in parallel to differing rates of vaccinations suggesting a strong link between the two [5]. Socioeconomic status of the infant’s family was also a strong determinant of whether infants were vaccinated on time, as has been previously reported [5,24]. This is also reflected in the finding that mothers engaged in paid employment were more likely to get their infants vaccinated according to schedule. A link between socioeconomic status and education status of the mother has been proposed [3], but mother’s education status was not strongly associated with a child’s vaccination status in this analysis. Finally, small ‘perceived size of infant at birth’ was negatively associated with vaccination status. Others have suggested that delaying the age of first vaccination to 28 days or greater is associated with higher risk of non-completion of the vaccination schedule [11,14]. Our results reiterate the importance of identifying and counseling families who may be at risk of delaying vaccinations due to small perceived size of infant at birth. The JiVitA-1 cohort used in this analysis is reflective of the rural population of Bangladesh [23]. However, there are several limitations to this analysis. The JiVitA-1 cohort was established to study the effects of vitamin A supplementation on reducing pregnancy-related maternal and infant mortality and information on infant vaccination status was recorded as part of the household surveys conducted during the study [20,21,23]. However, specific vaccination-related information such as the dates of vaccination, reasons for missed vaccinations, or previously known predictors of vaccination were not collected actively during the study. Hence, our
L. Vasudevan et al. / Vaccine 32 (2014) 5514–5519
analysis is limited to the existing types and quantities of available data. For instance, we used the date of the postpartum interviews as a proxy for the timeliness of vaccinations recommended by 14 weeks of age. As specified in the methods section, the use of postpartum interview dates as a proxy for dates of vaccination allowed for different buffer times for the infant to receive any particular vaccine. Hence, within a sub cohort, infants interviewed earlier had a shorter window of opportunity to get vaccinated on time as opposed to infants interviewed later. Some factors that may potentially be very strong predictors of infant vaccination status in rural locations such as accessibility of health care services, or programmatic factors pertaining to vaccine worker training, attitude, missed outreach and advocacy opportunities or inadequate supply of vaccines were not studied [4,17,22,27]. In addition, we did not distinguish between completely unvaccinated and partially vaccinated infants [11]. This study highlights the need to improve timeliness of vaccinations among infants in rural Bangladesh, especially during the early months of life. We concur with other reports in the literature that timely vaccination coverage is a better performance metric for routine immunization services than crude vaccination rates alone [9,11–16]. Illustrative solutions worth exploring may include timely identification and counseling of mothers whose infants are not vaccinated on schedule. Digital or mobile phone based recordkeeping systems that can prioritize and alert vaccinators to such cases may have greater utility over paper-based systems in addressing these constraints. Findings from this study may be used to guide local, regional and national programs aiming to improve timely coverage of vaccinations in rural populations. Funding source This work was funded partly through a Grand Challenges Explorations Phase I grant to AL and LV. The funding source had no involvement in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. The field trial and associated studies were supported through Grant No. 614 (Global Control of Micronutrient Deficiency) from the Bill and Melinda Gates Foundation, Seattle WA; Cooperative Agreement HRN-A-0097-00015-00 and Global Research Activity GHS-A-00-03-00019-00 between the Office of Health, Infectious Diseases and Nutrition, US Agency for International Development (USAID), Washington, DC; the USAID Mission, Dhaka, Bangladesh; the Sight and Life Global Nutrition Research Institute, Baltimore, MD; the Micronutrient Initiative/Department of Foreign Affairs, Trade and Development, Ottawa, Ontario; and the Nutrilite Health Institute, Access Business Group LLC, Buena Park, CA; and the Ministry of Health and Family Welfare, Government of Bangladesh, Dhaka. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine. 2014.06.092. References [1] World Health Organization. State of the world’s vaccines and immunization; 2009.
5519
[2] Rainey JJ, Watkins M, Ryman TK, Sandhu P, Bo A, Banerjee K. Reasons related to non-vaccination and under-vaccination of children in low and middle income countries: findings from a systematic review of the published literature, 1999–2009. Vaccine 2011;29:8215–21. [3] Breiman RF, Streatfield PK, Phelan M, Shifa N, Rashid M, Yunus M. Effect of infant immunisation on childhood mortality in rural Bangladesh: analysis of health and demographic surveillance data. Lancet 2004;364:2204–11. [4] Khan MM, Yoder RA. Technical Report No. 24. Expanded program on immunization in Bangladesh: cost, cost-effectiveness, and financing estimates, 2012. Bethesda, MD: Partnerships Heal Reform Proj Abt Assoc Inc.; 1998. p. 1–78. [5] National Institute of Population Research and Training (NIPORT), Mitra and Associates, and ICF International. Bangladesh Demographic and Health Survey 2011. MA, USA: Dhaka, Bangladesh/Calverton; 2013. [6] Rubayet S, Shahidullah M, Hossain A, Corbett E, Moran AC, Mannan I, et al. Newborn survival in Bangladesh: a decade of change and future implications. Health Policy Plan 2012;27(Suppl. 3):iii40–56. [7] World Health Organization Regional Office for South-East Asia. Bangladesh 2009 EPI Fact Sheet; 2010. [8] World Health Organization. Immunization Profile – Bangladesh; 2011. [9] Akmatov MK, Mikolajczyk RT. Timeliness of childhood vaccinations in 31 low and middle-income countries. J Epidemiol Community Health 2011:5. [10] UNICEF, World Health Organization. Immunization Summary. A statistical reference containing data through 2011; 2013. [11] Bicaba A, Haddad S, Kabore M, Taminy E, Feletto M, Fournier P. Monitoring the performance of the Expanded Program on Immunization: the case of Burkina Faso. BMC Int Health Hum Rights 2009;9(Suppl. 1):S12. [12] Expanded Programme on Immunization of the Department of Immunization Vaccines and Biologicals. Training for mid-level managers (MLM) 7. The EPI coverage survey. Geneva, Switzerland: EPI; 2008. [13] Fadnes LT, Nankabirwa V, Sommerfelt H, Tylleskär T, Tumwine JK, Engebretsen IMS. Is vaccination coverage a good indicator of age-appropriate vaccination? A prospective study from Uganda. Vaccine 2011;29:3564–70. [14] Sadoh AE, Eregie CO. Timeliness and completion rate of immunization among Nigerian children attending a clinic-based immunization service. J Health Popul Nutr 2009;27:391–5. [15] Murray CJL, Shengelia B, Gupta N, Moussavi S, Tandon A, Thieren M. Validity of reported vaccination coverage in 45 countries. Lancet 2003;362:1022–7. [16] Luman ET, McCauley MM, Stokley S, Chu SY, Pickering LK. Timeliness of childhood immunizations. Pediatrics 2002;110:935–9. [17] Chowdhury AMR, Salam Bhuiya A, Mahmud S, Abdus AKM, Karim F. The immunization divide: who gets vaccinated in Bangladesh? J Health Popul Nutr 2003;21:193–204. [18] World Health Organization. MDG 4: reduce child mortality; 2013. [19] West Jr KP, Christian P, Katz J, Labrique A, Klemm R, Sommer A. Effect of vitamin A supplementation on maternal survival. Lancet 2010;376:873–4 [author reply 874]. [20] West Jr KP, Christian P, Labrique AB, Rashid M, Shamim AA, Klemm RD, et al. Effects of vitamin A or beta carotene supplementation on pregnancy-related mortality and infant mortality in rural Bangladesh: a cluster randomized trial. J Am Med Assoc 2011;305:1986–95. [21] Klemm RD, Labrique AB, Christian P, Rashid M, Shamim AA, Katz J, et al. Newborn vitamin A supplementation reduced infant mortality in rural Bangladesh. Pediatrics 2008;122:e242–50. [22] Favin M, Steinglass R, Fields R, Banerjee K, Sawhney M. Why children are not vaccinated: a review of the grey literature. Int Health 2012;4:229–38. [23] Labrique AB, Christian P, Klemm RD, Rashid M, Shamim AA, Massie A, et al. A cluster-randomized, placebo-controlled, maternal vitamin A or beta-carotene supplementation trial in Bangladesh: design and methods. Trials 2011; 12:102. [24] Tadesse H, Deribew A, Woldie M. Predictors of defaulting from completion of child immunization in south Ethiopia, May 2008: a case–control study. BMC Public Health 2009;9:150. [25] Suarez L, Simpson DM, Smith DR. Errors and correlates in parental recall of child immunizations: effects on vaccination coverage estimates. Pediatrics 1997;99:e3. [26] Ramakrishnan R, Rao TV, Sundaramoorthy L, Joshua V. Magnitude of recall bias in the estimation of immunization coverage and its determinants. Indian Pediatr 1999;36:881–5. [27] Weiss WM, Winch PJ, Burnham G. Factors associated with missed vaccination during mass immunization campaigns. J Health Popul Nutr 2009;27:358–67. [28] Gunnsteinsson S, Labrique AB, West Jr KP, Christian P, Mehra S, Shamim AA, et al. Constructing indices of rural living standards in Northwestern Bangladesh. J Health Popul Nutr 2010;28:509–19. [29] Hanson CM, Furrer E, Schwalbe N, Berkley S. Prediction of immunisation performance. Lancet 2013;381:349–50. [30] Hilber AM, Bosch-Capblanch X, Schindler C, Beck L, Sécula F, Mckenzie O, et al. Gender and immunisation. In: Summary report for SAGE. 2010.