BMC Public Health BioMed Central Research article

Background Epidemiological investigations of recent outbreaks of vaccine preventable diseases have indicated that incomplete immunization was the major reason for the outbreaks. In Ethiopia, full immunization rate is low and reasons for defaulting from immunization are not studied well. The objective of the study was to identify the predictors of defaulting from completion of child immunization among children between ages 9–23 months in Wonago district, South Ethiopia. Methods Unmatched case control study was conducted in eight Kebeles (lowest administrative unit) of Wonago district in south Ethiopia. Census was done to identify all cases and controls. A total of 266 samples (133 cases and 133 controls) were selected by simple random sampling technique. Cases were children in the age group of 9 to 23 months who did not complete the recommended immunization schedule. Pre-tested structured questionnaire were used for data collection. Data was analyzed using SPSS 15.0 statistical software. Results Four hundred eighteen (41.7%) of the children were fully vaccinated and four hundred twelve (41.2%) of the children were partially vaccinated. The BCG: measles defaulter rate was 76.2%. Knowledge of the mothers about child immunization, monthly family income, postponing child immunization and perceived health institution support were the best predictors of defaulting from completion of child immunization. Conclusion Mothers should be educated about the benefits of vaccination and the timely administration of vaccines.


Background
Worldwide about 29,000 children under the age of five die every day, mainly from preventable causes. An Ethiopian child is 30 times more likely to die by his or her fifth birthday than a child in Western Europe [1]. Some of the deaths occur from illnesses like measles, malaria or tetanus. According to the demographic health survey (DHS, 2005) of Ethiopia, child mortality rate of the country was 132 per 1000 live births, which is one of the highest in the world. The child mortality rate in south Ethiopia was 142 per 1000 live births [2,3]. Immunization offered the greatest benefits for health, well being and survival of children than any other interventions [4]. From 1960-2002, a fifty percent reduction in under-five mortality was observed. Immunization programmes has saved the lives of nearly 4 million children [1][2][3]. Study showed that the cost to treat a vaccine preventable disease is 30 times more than the cost of the vaccine [5].
Epidemiological investigations of recent outbreaks of vaccine preventable diseases indicated that incomplete immunization was the major reason for the outbreaks. Moreover, a low immunization rate was the major reasons for many of the outbreaks of infectious diseases in the past two decades [6].
In 1980, the Ministry of Health of Ethiopia initiated the Expanded Program on Immunization (EPI). The objective of EPI in Ethiopia was to fully vaccinate 90% under-one children by the year 1990. The vaccination schedule of Ethiopia is based on the e recommendation of World Health Organization (WHO) for developing countries [7,8]. However, different literatures revealed that EPI schedule in Ethiopia is not completed as planned and full immunization rate is low (49.9%). As a result many children in Ethiopia do not get the benefits of immunization [9,10]. To increase the immunization coverage in Ethiopia, predictors of defaulting has to be identified through community based studies. Such community based studies are not available in Ethiopia. The objective of this study was to identify the predictors of defaulting from immunization among children between 9-23 months of age in Wonago district, south Ethiopia.

Study area
This unmatched case control study was conducted from February 15 to March 15, 2008 in Wonago district in south Ethiopia. Wonago is one of the densely populated districts in Ethiopia. The District has 17 rural and one urban Kebeles (small administrative units). The Kebeles in the district were stratified into urban and rural settings. Seven rural Kebeles were randomly selected by lottery method. The only urban Kebele, namely Wonago was also included in the study. The total population of the study Kebeles was 39600.

Participants
The study population were children in the age group of 9 to 23 months in the selected Kebeles. Cases were children in the age group of 9 to 23 months who did not complete the recommended schedule of immunization. Controls were defined as children who were in the age group 9 to 23 months and completed the recommended schedules of immunization which include Bacillus Calmette Guerin (BCG), pentavalent, polio and measles vaccines. Children who never got vaccination were excluded in the study. Sample size was calculated using EPI-Info 6.04 statistical software (Center for Disease Control and Prevention, Atlanta, 2005). The assumptions for the sample size calculation were: proportion of illiterate mothers or caretakers of 82% among the cases and 94.4% among controls [11], 80% power, 95% confidence interval, 10% non-response rate and a case: control ratio of 1:1. The total sample size was 266 (133 cases and 133 controls). Among many exposure variables, educational status was selected since it gave maximum sample size as compared to other variables. In the selected Kebeles, first census was done to identify all cases and controls. The total sample size was distributed to each Kebele based on proportional to size allocation (Table 1). Using the sampling frame of cases and controls from the census, 133 cases and 133 controls were selected by simple random sampling technique.
Prior to data collection, the questionnaire was pre-tested on 5% of the sample on similar population. The reliability of the items in the questionnaires was tested using Cronbach's alpha.

Data Processing and analysis
The collected data were entered into computer and analyzed using SPSS version 15.0 statistical software. The data was cleaned for inconsistencies and missing values. Different predictors of defaulting from immunization were assessed using logistic regression model. In the first model, the effects of socio-demographic variables were assessed. The effect of maternal and child health services were assessed in the second model. In the third and fourth model, perception and knowledge related variables were assessed. From all the models, variables which had significant association with dependent variable (P < 0.05) were entered to the final regression model using stepwise logistic regression technique.

Operational definitions
The following operational definitions were used:

Defaulter
If the child missed at least one of the recommended vaccines, he/she was considered as defaulter.

Complete immunization
Complete immunization was considered if the child took all the recommended vaccines including BCG, pentavalent, polio and measles vaccine by the age of 23 months.

Perception
Perceived health institution support, perception on the presence barrier of immunization, and perception about immunization were assessed using Likert Scale questions. Mean score for each constructs was computed and dichotomized into positive and negative. If respondents scored below the mean, he/she would be labeled as having negative perception.

Dropout rate (DOR)
The rate difference between the initial vaccine (BCG or Pentavalent I) and the final vaccines (Pentavalent III or Measles).

Ethical consideration
Ethical clearance was obtained from the Ethical review Committee of Jimma University. Before data collection, written consent was obtained from the respondents.

The effect of Maternal and child health services on defaulting
There was a significant association between utilization of postnatal care (PNC) service and completion of child immunization (p-value < 0.05). Mothers who did not use PNC service after delivery of the child under study were 6 times more likely to have defaulter children than mothers who did use PNC services, [OR = 5.8 (95% CI: 1.68, 27.29)]. Mothers who did not postpone vaccination schedule were 98% times less likely to have defaulter children as compared with mothers who ever postponed vaccination schedule ( Table 2).  (Table 3).

Knowledge as predictors of defaulting
Knowledge of mothers or immediate caretakers about schedule of vaccines had significant association with completion of immunization (p-value < 0.001). Mothers who did know the schedules of vaccine were 3 times more likely to vaccinate their children fully than mother who didn't know vaccine schedule, [OR = 3 (95% CI: 1.4, 6. 3)]. Mothers or immediate care taker who did not know the benefits of immunization in preventing the occurrence of epidemic were 6.4 times more likely to have defaulter children than mother who knew the benefits, [OR = 6.4, (95% CI: 0.43, 9.53)]. Knowledge about measles and polio vaccines were also significantly associated with completion of child immunization after controlling the effect of other variables (Table 4).

Independent predictors of defaulting
In the final logistic regression model, we have identified that monthly family income, postponing of child immunization schedule, perceived health institute support, knowledge about measles and benefits of vaccination   (Table 5).

Discussion
All the study Kebeles had access to immunization services monthly at outreach and weekly at health institutions. However, significant proportion of children (42%) didn't complete the recommended immunization schedule. The dropout rate in our study was higher as compared to a study conducted in Kenya [12] which was 22.6%. However, it is lower than that of the national EPI cluster survey result (35.6%) [13]. Determinants of receipt of immunization are complex and interwoven. This study identified several socio-behavioral factors which affected child immunization. The most important variables that predicted defaulting from immunization were perceived health institution support, monthly family income, postponing child immunization schedule, and knowledge of the mothers/caretakers about the benefit of immunization. Of the socio demographic characteristics of the  respondents, only monthly family income was found to be predictor of defaulting from completion of child immunization in this study. A similar study by Renstein showed that income had consistently affected receipts of immunization [14]. Other socio-demographic variables were not associated with defaulting. Mothers or immediate caretakers who had negative attitude about health institution support were 2.3 times more likely to have defaulter children than mothers or immediate caretakers who had positive attitude. Similar finding was obtained from other study which showed that the barriers of completion of child immunization were poor knowledge, attitude and perception of health institution support [15]. This study had identified that maternal knowledge about immunization was one of the major reasons for defaulting. Other literatures had similar findings elsewhere [16,17]. It was shown that postponing child immunization schedule was a predictor of completion of child immunization. Similarly, a facility based study conducted in Nigeria showed that the most common reason given by respondents for defaulting of child immunization were illnesses of children and postponing immunization [18].
We tried to assess many predictors of defaulting from child immunization using a community based case control study. We had identified that knowledge and perception of the mothers had contributed for defaulting from child immunization. However, our study didn't address the effect of lack of access to health care and shortage of trained human resources as predictors of defaulting from child immunization. Lack of infrastructure and adequate human resources might be the root causes of defaulting from child immunization. Our study had also some limitations which included recall bias where mother might forget the immunization status of their children and misclassification. Due to forgetfulness of the mothers/caretakers about the status of immunization of their children, cases and controls could be misclassified.

Conclusion
It is concluded that behavioral factors like knowledge, attitude and perception of mother had strong association with defaulting from child immunization. Monthly family income and postponing of child immunization schedule were also the other independent predictors of defaulting. Mothers should be educated about the benefits of vaccination and the timely administration of vaccines.