Minding the Immunization Gap: Family Characteristics Associated with Completion Rates in Rural Ethiopia

Transcription

1 J Community Health (2010) 35:53 59 DOI /s ORIGINAL PAPER Minding the Immunization Gap: Family Characteristics Associated with Completion Rates in Rural Ethiopia Mary-Christine Sullivan Ayalew Tegegn Fasil Tessema Sandro Galea Craig Hadley Published online: 22 October 2009 Ó Springer Science+Business Media, LLC 2009 Abstract To examine risk factors for lack of immunization, we tested the impact of maternal, paternal, and household variables on child immunization status in children C1 year in a rural area of Ethiopia. Data collected by face-to-face interview on maternal, paternal, household and child variables from cross-sectional random sample community-based study on health and well-being in rural Ethiopia was used to test hypotheses on immunization status of children (n = 924). Bivariate and multivariate logistic regression models were used for two immunization outcomes: record of at least one vaccination, and record of DPT3, indicating completion of the DPT series. Complete data were available for 924 children C1 year of which 79% had at least one vaccination. Of those, 64% had DPT3/ Polio3; below recommended coverage level. Children were more likely to be vaccinated if the mother reported antenatal care (ANC), and less likely to be vaccinated if the mother had a history of stillbirth, and no opinion of health center. Children were more likely to have DPT3 if: mother had C1 year of education, mother reported ANC, or older paternal age. Children were less likely to have DPT3 in households with food insecurity and no maternal opinion of health center. The study had three findings with implications for immunization programming: (1) Mothers completing the recommended ANC visits is strongly associated with receiving at least one vaccination and with completing a vaccination series; (2) Maternal education is associated with a completed vaccination series; (3) Paternal characteristics may affect vaccination series completion. Keywords ANC DPT3 Immunization M.-C. Sullivan Hubert Department of Global Health, Rollins School of Public Health, Emory University, 1518 Clifton Rd., Atlanta, GA 30322, USA M.-C. Sullivan (&) Yale School of Nursing, 100 Church St. South, New Haven, CT 06519, USA mary-christine.sullivan@yale.edu A. Tegegn F. Tessema Department of Biostatistics and Epidemiology, Jimma University, P.O. Box 1348, Jimma, Ethiopia S. Galea University of Michigan School of Public Health, 109 Observatory St, Rm 3663, Ann Arbor, MI 48109, USA C. Hadley Department of Anthropology, Emory University, 1557 Dickey Dr., Atlanta, GA 30322, USA Introduction Child immunization is one of the most successful strategies for reducing child morbidity and mortality. Immunization requires a sustained, coordinated approach to facilitate delivery and acceptance at the community level and many resource-poor countries struggle to achieve vaccination coverage targets. The Expanded Programme on Immunization (EPI) works in many of these countries to provide the resources and technical assistance necessary to overcome logistical and cultural challenges. EPI estimates that its vaccination programs prevent the deaths of three million children each year, reducing the childhood burden of the targeted diseases by 13% since the mid-1970s [1]. The reasons for sub-par immunization rates are categorized into population and service delivery factors.

2 54 J Community Health (2010) 35:53 59 Research shows maternal education or literacy to be the strongest and most consistent predictor of a child s vaccination outcomes, yet the reasons for this association remain elusive [2 8]. In an analysis of data from the demographic and health surveys (DHS) from 22 developing countries, maternal education remained significantly associated with child immunization after controlling for individual and community level variables, but a causal relationship could not be determined [9]. Local culture and knowledge of vaccination also influence individual acceptance and nonacceptance [10 14]. Even when vaccinations are provided free of charge, the time cost required for a trip to the health facility can be a barrier for receiving multiple vaccinations [15]. Child immunization research has focused on maternal influences rather than paternal, as women are primarily responsible for childcare in low-income countries, yet paternal characteristics may also influence vaccination outcome. One study of paternal influence on immunization from the Eastern Region of Ghana by Burgha et al. [16] found that the ability of a father to speak English was significantly associated with vaccination. Other studies noted the association of child vaccination status with paternal attitudes toward vaccination and paternal employment [17, 18]. Additional population factors include child sex, antenatal care (ANC), and social capital. In countries where a strong preference for the sex of the child exists there may also be gender bias in the decision to immunize a child [19 21]. ANC is a factor for child immunization outcomes in both high- and low-income countries. Studies in the U.S. found that women who completed the recommended ANC visits or had maternal-child provider continuity fared better on use of infant health care [22, 23]. Attendance at ANC services or delivery at a facility puts a mother in contact with the healthcare system, and may improve rates of child vaccination in developing settings, although this has not been extensively studied [2]. Many of the factors related above reflect social capital, as the connections within and between social networks can influence immunization outcomes. A study from the Butajira Demographic Surveillance Site in Ethiopia showed that lack of immunization was associated with under-five mortality rates, as well as low decision-making of women, and low social capital scores [24]. Service delivery of immunization includes quality of service provided by healthcare workers, immunization publicity and distribution, and education about side effects [5, 25]. A multi-country study in Asia and Africa in 1994 showed that the decision by mothers to bring their children to be vaccinated is influenced by perceptions and experiences with the quality of healthcare services. In Gurage zone, Ethiopia, researchers observed that vaccinations started late as mothers expected healthcare workers to start late, and healthcare workers expected the same of the mothers. Yet the vaccination session ended on time, resulting in some children not being vaccinated and unsafe injection practices. The availability of routine services in Ethiopia was also interrupted by the healthcare workers participating in national immunization events for weeks or months at a time [26]. Immunization in Ethiopia The recommended childhood vaccination schedule in Ethiopia consists of five vaccine contacts for BCG, DTP, OPV, and measles to be delivered by age 1. Most vaccinations are available routinely, from a static facility or weekly/monthly outreach immunization days. Polio immunizations are distributed by mass campaign [26]. The EPI program in Ethiopia is run by the Ministry of Health in cooperation with WHO, UNICEF, and other collaborators. WHO assists with technical assistance, planning, resources and social mobilization. Regional Health Bureaus are responsible for implementation. All vaccinations are provided free of charge [27]. A review of the EPI program in Ethiopia in 2001 identified a lack of supervision, high drop-out rates, inadequate number of trained health workers, and insufficient supplies as major obstacles [27]. Maintaining the cold chain remains a significant challenge for EPI in Ethiopia. A national inventory exercise in 2002 showed that only 66% of cold chain equipment was found functional at the time of inventory. Further complicating matters, 51 different manufacturers were found for the 4,833 pieces of equipment inventoried, making training and repair a challenge [27]. EPI has been working in Ethiopia since 1981, yet immunization rates in Ethiopia remain below the target of 80% coverage. DPT3 is a commonly used proxy measure for vaccination coverage and in 2006 only 20 districts of 85 had C90% DPT3 coverage, which is far below the EPI goal [28]. Childhood immunization is a priority for reducing the high mortality rate of deaths per 1000 live births for children under-five in Ethiopia and working towards the UN Millennium Development Goal of reducing child mortality by two-thirds by 2015 [29]. Therefore, the objective of this analysis was to test the impact of maternal, paternal, and household variables on child immunization status in children C1 year in a rural district of Ethiopia to identify family characteristics related to vaccination status and completion.

3 J Community Health (2010) 35: Methods Study Setting Gilgel Gibe is located in Jimma zone within the state of Oromiya. It is 55 km from Jimma and 260 km from Addis Ababa. The communities in the Gilgel Gibe area are designated as rural with the exception of Asendabo district, which is urban (C. Hadley, personal communication, December 18, 2007). This is a rural area that relies primarily on subsistence agriculture. Jimma University conducts ongoing complete demographic surveillance in the Gilgel Gibe area to collect vital events data for a complete census of the population of more than 10,000 households in the area. Records are updated multiple times each year and there is high level of support from the community so refusals are rare. The Gilgel Gibe Growth and Development Study (GGGDS) is a cohort study of families in the demographic surveillance system. The GGGDS is concerned with adult mental health and child development. The baseline cohort for the GGGDS was a sample of households that had a child between the ages of 3 24 months from the universe of all births in Gilgel Gibe in the 2 years prior to the estimated start date of the survey. The study involves questionnaire and anthropometric information collected from the parents and developmental assessments conducted on their children. The data analyzed in this study was collected in June and July A structured questionnaire was developed and administered to participants by nine interviewers. Questionnaires and consent documents were developed in English and then translated and back translated into Amharic and Affan Oromifa by native speakers. Interviewers took a week-long training that included practice interviewing and role playing. Following the training, interviewers undertook a pilot study and the data was checked for consistency, outliers, and missing values. After the pilot, interviewers and investigators met to discuss experiences, issues, and finalize the questionnaire. The final questionnaires for men and women were administered separately to husbands and wives at the household in a private area. Written informed consent was obtained from all participants. This study was approved by the Institutional Review Boards of the University of Michigan, Jimma University, and Emory University. Variables Variables were categorized as maternal characteristics, maternal pregnancy healthcare and opinions, household attributes, and paternal characteristics and opinions. Previous literature has shown the importance of maternal characteristics for child immunization, and pregnancy healthcare reflects maternal utilization of health facilities. Household factors were included to look at socio-economic effects, and basic paternal variables were included since these variables have not been adequately examined in previous studies. Maternal variables included age, education, and pregnancy-related variables such as ANC visits, history of stillbirth, history of infant death, delivery location (facility vs. non-facility), and number of pregnancies in the last 5 years ([2 is a higher risk birth interval) were included as pregnancy healthcare and outcomes may influence healthcare-seeking for a child. Mothers were interviewed about the perceived quality of care at the nearest health post, health center, and hospital. A large number of these responses were Don t Know even though a scale was provided. This is consistent with low facility utilization in this area and we considered this variable as a proxy for interaction with the health facility such that those who have been to the facility will have an opinion and those who have not been answered Don t Know because they truly did not know about the quality of care. Household variables included food insecurity, socioeconomic status, and total children. Seven food security questions were taken from a validated food insecurity module which is used in both developed and developing countries [30]. Cronbach s alpha for the reliability of food security questions was.93 for mothers and.92 for father. Food security scores from the mother and father questionnaire had substantial agreement for having high food insecurity (defined as [4 affirmative responses) with a j =.66 and were highly correlated (Spearmans q = 0.71, P \.0001). Food security questions from the maternal/ child questionnaire were used for the food insecurity score variable. A subjective pictorial scale was used for socioeconomic status. A picture of a ladder with 10 rungs was presented and participants asked to rank their household standing relative to others in the community. Interviewers reported that respondents immediately grasped the concept of the ladder. Mother and father responses were highly correlated, and we chose to use the mother s response as it was more conservative and dichotomized it at the median to High SES and Low SES. Total children was included in its original form. Children should receive all vaccinations by age 1. The dichotomous outcome variables for children C1 year were receipt of At least one vaccination (of BCG/Polio0; DPT1/Polio1; DPT2/Polio2; DPT3/Polio3; measles) and DPT3, indicating completion of DPT series. DPT3 is used to measure immunization coverage and as a proxy for the strength and quality of immunization services [15]. Univariate and bivariate analysis of key dependent and independent variables was completed to assess the prevalence of maternal, paternal, household and child

4 56 J Community Health (2010) 35:53 59 characteristics. Logistic regression models were used to assess the relationship between predictor variables and child vaccination outcomes by adding variables by category. Finally, a Generalized Estimating Equation (GEE) model was used for each outcome to account for interviewer effects and differences between kebeles. Statistical significance was set at \0.05. All statistical tests were carried out in SAS 9.1. Results There were n = 1016 maternal/child questionnaires completed and n = 960 paternal questionnaires completed. Characteristics of the study sample are shown in Tables 1 and 2. The vaccination rate for children C1 year in Gilgel Gibe area (n = 924) was 79% for having at least one vaccination. Vaccination cards were produced by 51% of those C1 year reporting vaccination (n = 370). Of those reporting vaccination, 97% had received BCG/Polio0, 64% DPT3/Polio3, and 54% measles. Complete data was available for n = 924 children and n = 847 households. Women had a mean education level of 0.0 years and 16% had at least 1 year of education. During the last pregnancy, 14% of women attended the WHO recommended 4? ANC visits [31]. For mothers, 46% had no opinion of the quality of care at the nearest health center. In bivariate analyses, all of the following were significantly associated with greater likelihood of vaccination at P \.05: women with at least 1 year of education; completed recommended 4? ANC visits during last pregnancy; a history of stillbirth; a history of infant death; facility delivery for last birth; mothers having no opinion of nearest health center; fathers with at least 1 year of education; fathers having no opinion of health center; and child sex. No household variables were significant in bivariate analysis but were included in the models as socio-economic variables may modify the relationships between other variables and immunization. Maternal age and having more than 2 pregnancies in 5 years were not significant but may also affect the mother s ability to have her child immunized on time. Paternal variables and child sex were included to detect any influence on vaccination status. Using a GEE model to account for correlation between interviewers assigned to kebeles, a child with at least one immunization was more likely to have a mother who completed ANC (OR = 5.24, 95% CI: 1.55, 17.67), while those with a history of stillbirth (OR = 0.43, 95% CI: 0.19, 0.96) and no maternal opinion of health center (OR = 0.55, 95% CI: 0.42, 0.72) were less likely to be immunized. A subset analysis using the final model for children who had received At least one vaccination was then used to model the outcome variable of DPT3 (n = 624), to determine any differences in characteristics associated with those who completed the DPT3 vaccination series and defaulters. For completion of a vaccination series using the proxy of DPT3 dose, children with mothers with [1 year of education (OR = 2.40, 95% CI: 1.57, 3.66), who completed the recommended ANC visits (OR = 2.89, 95% CI: 1.93, 4.34), and had older fathers (OR = 1.03, 95% CI 1.00, 1.05) were more likely to be immunized. Children with a mother with no opinion of the health center (OR = 0.71, 95% CI 0.57, 0.87), and a household that was food insecure (OR = 0.70, 95% CI: 0.57, 0.87) were less likely to be immunized. These results are summarized in Table 3. Lastly, propensity score matching was used to verify the strong relationship between ANC visits and the vaccination outcomes [32]. When matched on additional variables, mothers attending recommended ANC visits were significantly more likely to have their child vaccinated (McNemar S = 72.43, P \.0001) or to have completed a vaccination series (McNemar S = 29.39, P \.0001). Discussion The goal of this research was to assess the family factors which affect child immunization status in this rural area of Ethiopia. We found that children with at least one vaccination were more likely to have mothers who attended ANC. Children of mothers with a history of stillbirth, or who had no opinion of health center were less likely to be vaccinated. Further analysis on vaccinated children using the outcome of DPT3, showed that a slightly different set of variables were important for completing a vaccination series: children with a mother who had at least 1 year of education, reported high frequency of ANC visits, and greater paternal age, were more likely to complete a vaccination series. Children from households with food insecurity and with no maternal opinion of health center were less likely to complete a vaccination series. This study was limited by the variables available in the GGGDS questionnaire and relied on self-report. It may be subject to recall bias and same-source bias as mothers answered questions pertaining to the index child. There was no evidence that socio-economic status was associated with vaccination, but food security is likely a more appropriate measure in this rural and resource poor area. Having or not having an opinion of the nearest health center, was treated as proxy for interaction with the health center, and the high percentage of people without an opinion is consistent with a low rate of facility utilization in this area. Clustering by interviewer was done to reduce any differences between kebeles introduced by

5 J Community Health (2010) 35: Table 1 Descriptive characteristics of study population Variable Mean or % SD Range Age Father (n = 950) Mother (n = 1,011) Child C1 year (n = 924) 22.2 (months) (months) Education (years) Male (n = 949) Female (n = 1,012) BMI (kg/cm 2 ) Father (n = 950) Mother (n = 1,007) Total number of children (n = 1,007) Self-perceived socio-economic status (1 10) Father (n = 950) Mother (n = 1,003) Food insecurity past 3 months (n = 1,006) Household assets (n = 945) Distance to nearest health post (km) (n = 1,007) Distance to nearest health center (km) (n = 1,000) Distance to nearest hospital (km) (n = 999) Child Sex C1 year (n = 924) Male 482 (52%) Female 441 (48%) Self-perceived SES (n = 1,016) High SES (C3) 729 (72%) Low SES (\3) 287 (28%) Food Security (n = 1,016) Food insecure ([4) 395 (39%) Not food insecure (B4) 621 (61%) Distance to nearest health center (n = 1,016) Live in Asendabo 107 (11%) Not live in Asendabo 909 (89%) Table 2 Vaccination history of children [1 year in Gilgel Gibe (n = 924) Variable Yes No Any vaccination (n = 919) 729 (79%) 190 (21%) Vaccination card (n = 729) 370 (51%) 359 (49%) BCG (n = 680) 657 (97%) 23 (3%) DPT1 (n = 680) 642 (94%) 38 (6%) DPT2 (n = 676) 545 (81%) 131 (19%) DPT3 (n = 675) 433 (64%) 242 (36%) Measles (n = 675) 362 (54%) 313 (46%) between-interviewer differences. This research is crosssectional and cannot determine causality, but the results may help in determining areas for future research. Apart from service delivery factors, family and household characteristics do influence child vaccination status and series completion. Maternal education was not a predictor for receipt of at least one vaccination but may be due to the very low levels of education among women in the study area; the mean education level in this area is 0.0 years. Consistent with the idea that contact with the healthcare center is important is the finding that Mother having no opinion of the health center was significantly negatively associated with vaccination, though this finding was not found for paternal opinions. History of having a stillbirth requires further qualitative research to determine its practical significance but may be related to a pattern of non-attendance at ANC. Paternal education, which was higher than women s, was close to \.05 but not significantly associated. From this analysis, it appears that prior maternal healthcare interaction and opinions are most important for determining whether a child in Gilgel Gibe will receive at least one vaccination.

6 58 J Community Health (2010) 35:53 59 Table 3 Final regression results for predicting vaccination outcomes based on family characteristics Bold text indicates significant at \.05 Variables Completing a vaccination series requires a commitment to accessing healthcare multiple times within the child s first year. Therefore, modeling DPT3 rather than any vaccination, as an outcome is essential for determining actual vaccination coverage and default rates. Using the DPT3 outcome, our research is consistent with others that found maternal education, and ANC visits to be significant predictors [2 8, 12, 16, 22, 23, 33 36]. We were unable to use the proximity variable in the DPT3 model but it worth noting that only 1 child who lived in Asendabo did not complete the DPT series, which suggests that proximity to health center affects completion rates. That food insecurity and not SES is related to completing a vaccination series may indicate that food insecurity is a more accurate measure of the effect socio-economic status on health outcomes and more closely captures resource availability in seasonal subsistence settings. The first main finding from this research is that completing the recommended ANC visits is strongly associated with receiving at least one vaccination and with completing a vaccination series. The second finding is that maternal education is significantly associated with completing a vaccination series, which is consistent with previous literature on this topic [2 8]. History of recommended ANC visits was strongly associated with child vaccination status in all outcomes and may be further indication that a mother s contact with the healthcare system will carry over to her child s healthcare. The third finding is that paternal factors such as age, education, and opinions of the nearest health center which have not been adequately examined in the literature might also impact completion of a vaccination series and deserve consideration in future immunization research. These findings suggest that a more family-oriented approach may be the key to improving child immunization outcomes. Education is a strong predictor for immunization in general, but in a setting where education levels are low, ANC may be a better place to engage women and perhaps even men in healthcare. ANC can provide an opportunity to educate about immunization and to follow-up with young children. There is no single bridge to be built to overcome the immunization gap for children; rather many smaller bridges are needed for families to be able to access and utilize immunization services. This research identifies potential areas for construction. Acknowledgments The authors are grateful to the local staff of GGGDS and colleagues at Jimma University for their support and work during data collection and manuscript preparation, and to the community of Gilgel Gibe for their on-going support and participation in this study. Funding for the Gilgel Gibe Growth and Development Study (GGGDS) study is provided by the Michigan Interdisciplinary Center on Social Inequality, Mind and Body (R24 HD047861) and start up funds. Conflict of Interest: None declared. References At least one vaccination (N = 847) DPT3 completion (N = 624) Age mother 1.00 (0.95, 1.05) 0.97 (0.94, 0.99) Education mother 1.07 (0.71, 1.62) 2.40 (1.57, 3.66) Risky pregnancy interval 1.08 (0.65, 1.81) 0.96 (0.69, 1.33) ([2 pregnancies in 5 years) Recommended 4? ANC visits 5.24 (1.55, 17.67) 2.89 (1.93, 4.34) History of stillbirth 0.43 (0.19, 0.96) 2.25 (0.97, 5.23) History of infant death 1.22 (0.67, 2.23) 1.29 (0.76, 2.21) Facility delivery 5.97 (0.87, 41.13) 0.86 (0.60, 1.22) Mother has no opinion of health center 0.55 (0.42, 0.72) 0.71 (0.57, 0.87) Food insecurity 0.94 (0.63, 1.40) 0.70 (0.57, 0.87) Self-perceived economic status 0.86 (0.66, 1.13) 1.06 (0.60, 1.87) Total children 0.98 (0.84, 1.15) 0.97 (0.89, 1.06) Age father 0.99 (0.97, 1.01) 1.03 (1.00, 1.05) Education father 1.52 (0.94, 2.44) 1.48 (0.94, 2.34) Father has no opinion of health center 0.96 (0.68, 1.34) 1.41 (0.97, 2.06) Male child 0.70 (0.44, 1.11) 0.81 (0.63, 1.04) 1. The World Bank. (2003). Public health at a glance: Immunization. Retrieved Dec 19, 2008 from A0E6J0XVO0. 2. De, P., & Bhattacharya, B. N. (2002). Determinants of child immunization in four less-developed states of North India. Journal of Child Health Care, 6(1), Basu, A. M., & Stephenson, R. (2005). Low levels of maternal education and the proximate determinants of childhood mortality: A little learning is not a dangerous thing. Social Science & Medicine, 60(9), Parashar, S. (2005). Moving beyond the mother child dyad: Women s education, child immunization, and the importance of

7 J Community Health (2010) 35: context in rural India. Social Science & Medicine, 61(5), Alemu, W., Woldeab, M., & Meche, H. (1991). Factors influencing non-attendance in the immunization of children in 3 selected regions, Ethiopia, July Ethiopian Medical Journal, 29(2), Elliott, C., & Farmer, K. (2006). Immunization status of children under 7 years in the Vikas Nagar area, North India. Child Care Health and Development, 32(4), Nath, B., Singh, J., Awasthi, S., Bhushan, V., Kumar, V., & Singh, S. (2007). A study on determinants of immunization coverage among months old children in urban slums of Lucknow district, India. Indian Journal of Medical Sciences, 61, Chhabra, P., Nair, P., Gupta, A., Sandhir, M., & Kannan, A. T. (2007). Immunization in urbanized villages of Delhi. Indian Journal of Pediatrics, 74(2), Desai, S., & Alva, S. (1998). Maternal education and child health: Is there a strong causal relationship? Demography, 35(1), Streefland, P. H., Chowdhury, A. M. R., & Ramos-Jimenez, P. (1999). Patterns of vaccination acceptance. Social Science & Medicine, 49(12), Gedlu, E., & Tesemma, T. (1997). Immunization coverage and identification of problems associated with vaccination delivery in Gondar, north west Ethiopia. East African Medical Journal, 74(4), Berhanel, Y., Bekele, A., & Tesfaye, F. (2000). Immunization (EPI) in Ethiopia: Acceptance, coverage, and sustainability. Ethiopian Medical Journal, 38(Suppl 1), Topuzoglu, A., Ay, P., Hidiroglu, S., & Gurbuz, Y. (2007). The barriers against childhood immunizations: a qualitative research among socio-economically disadvantaged mothers. European Journal of Public Health, 17(4), Bhattacharyya, K., & Murray, J. (2000). Community assessment and planning for maternal and child health programs: A participatory approach in Ethiopia. Human Organization, 59(2), USAID. (2003). Immunization essentials: A practical field guide. Office of Health, Infectious Diseases, and Nutrition, Bureau for Global Health. 16. Brugha, R. F., Kevany, J. P., & Swan, A. V. (1996). An investigation of the role of fathers in immunization uptake. International Journal of Epidemiology, 25(4), Torun, S. D., & Bakirci, N. (2006). Vaccination coverage and reasons for non-vaccination in a district of Istanbul. BMC Public Health, 6, Chowdhury, A. M. R., Bhuiya, A., Mahmud, S., Salam, A., & Karim, F. (2003). Immunization divide: Who to get vaccinated in Bangladesh? Journal of Health Population and Nutrition, 21(3), Choi, J. Y., & Lee, S. H. (2006). Does prenatal care increase access to child immunization? Gender bias among children in India. Social Science & Medicine, 63(1), Bishai, D., Suzuki, E., McQuestion, M., Chakraborty, J., & Koenig, M. (2002). The role of public health programmes in reducing socioeconomic inequities in childhood immunization coverage. Health Policy and Planning, 17(4), Borooah, V. K. (2004). Gender bias among children in India in their diet and immunisation against disease. Social Science & Medicine, 58(9), Butz, A. M., Funkhouser, A., Caleb, L., & Rosenstein, B. J. (1993). Infant health-care utilization predicted by pattern of prenatal-care. Pediatrics, 92(1), Gill, J. M., Saldarriaga, A., Mainous, A. G., & Unger, D. (2002). Does continuity between prenatal and well-child care improve childhood immunizations? Family Medicine, 34(4), Fantahun, M., Berhane, Y., Wall, S., Byass, P., & Hogberg, U. (2007). Women s involvement in household decision-making and strengthening social capital crucial factors for child survival in Ethiopia. Acta Paediatrica, 96(4), Nichter, M. (1995). Vaccinations in the third-world a consideration of community demand. Social Science & Medicine, 41(5), Streefland, P. H., Chowdhury, A. M. R., & Ramos-Jimenez, P. (1999). Quality of vaccination services and social demand for vaccinations in Africa and Asia. Bulletin of the World Health Organization, 77(9), World Health Organization. (2008). Expanded Programme on Immunization: Ethiopia. Retrieved April 2, 2008 from World Health Organization. (2007). Immunization Profile: Ethiopia. Retrieved April 2, 2008 from globalsummary/immunization/countryprofileresult.cfm. 29. Central Stastical Agency [Ethiopia] & ORC Macro. (2006). Ethiopia demographic and health survey Addis Ababa, Ethiopia and Calverton, Maryland, USA: Central Statistical Agency and ORC Macro. 30. Perez-Escamilla, R., Segall-Correa, A. M., Maranha, L. K., Sampaio, M. D. A., Marin-Leon, L., & Panigassi, G. (2004). An adapted version of the US department of agriculture food insecurity module is a valid tool for assessing household food insecurity in Campinas, Brazil. Journal of Nutrition, 134(8), Villar, J., & Bergsjo, P. (2002). WHO antenatal care randomized trial: Manual for the implementation of the new model. Geneva: WHO. 32. D Agostino, R. (1998). Biostatistics tutorial: Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Statistics in Medicine, 17(19), Xie, J. P., & Dow, W. H. (2005). Longitudinal study of child immunization determinants in China. Social Science & Medicine, 61(3), Luman, E. T., McCauley, M. M., Shefer, A., & Chu, S. Y. (2003). Maternal characteristics associated with vaccination of young children. Pediatrics, 111(5), Racine, A. D., & Joyce, T. J. (2007). Maternal education, child immunizations, and public policy: Evidence from the US national immunization survey. Social Science & Medicine, 65(8), Kidane, T., & Tekie, M. (2003). Factors influencing child immunization coverage in a rural District of Ethiopia, Ethiopian Journal of Health and Development, 17(2),