Data collection
The DHS collects data using a structured questionnaire that has several modules. For this study, data collected using household questionnaire, woman’s questionnaire and health facility questionnaire were used. The vaccination status of the eligible children was compiled from the mothers’ interview, vaccination card and health facility record.
The data were accessed by requesting the DHS program website [
34] after submitting the purpose of the study. We obtained the data for the years 2000, 2005, 2011 and 2016 from DHS. The mini-DHS 2019 full dataset is not yet publicly available.
Data analysis
The outcome variables were full immunization and inequality whereas household wealth index, mother’s educational status and place of residence were the independent variables.
A child is considered to be fully immunized when he/she had received one dose of Bacille Calmette-Guérin (BCG) vaccine, three doses of polio vaccine, three doses of the combined diphtheria, tetanus toxoid and pertussis (DTP3) vaccine, and one dose of measles vaccine [
30]. Each vaccine antigen variable was coded as “0” for those who didn’t receive the vaccine dose and “1” for those who have received the dose. Then these values were added to give the immunization status. The immunization status was recoded as “1” (full immunization) if the child had received all the recommended doses of the vaccines mentioned above and “0” (incomplete immunization) if the child has missed at least one of the vaccines doses mentioned above.
Health inequalities are observable differences in health indicators between subgroups of a population. Subgroups can be defined by geographic or socioeconomic factors such as economic status, education and place of residence [
35].
The full immunization coverage was calculated by dividing the number of children aged 12–23 months receiving full/basic immunization (one dose of BCG, three doses of polio, three doses of DTP3, and one dose of measles) by the total number of children aged 12–23 months participated in the survey.
DHS is a very broad survey that is conducted in multiple countries and cover plenty of indicators with a detailed design and methods; any data analysis using data from DHS need to consider that complexity. Thus, we used a complex data analysis statistical technique that takes in to consideration the nature of the data using “svy” command which is a Stata command for fits statistical models for complex survey data. It adjusts the results of a command for survey settings identified by svyset. Due to the non-proportional allocation of the sample to different regions and their urban and rural areas we need to ensure that the results are representative at the national and regional levels. Thus, sampling weights must be used to analyze the EDHS data. Since the DHS uses a two-stage stratified cluster sample, sampling weights were calculated based on sampling probabilities separately for each sampling stage and for each cluster in order to compensate for the unequal probability of selection between the different regions and their urban and rural areas as well as for non-response. A thorough explanation of the weighting procedure is found in the DHS reports [
28‐
32]. Projections for 2025 were made using a smoothed average using the data from DHS during 2000-2019. The smoothed average was established based on the average coverage reported in each DHS cycle and then weighted for the period 2000 to 2019. For this study, an annual increment in coverage was calculated by dividing the difference of the two consecutive survey coverage by the interval between the two surveys using the following formulas:
$$\mathrm{Annual}\ \mathrm{increment}\ 1\left(\mathrm{from}\ 2000-2005\right)=\frac{Full\ immunization\ coverage\ 2005-2000}{5}$$
$$\mathrm{Annual}\ \mathrm{increment}\ 2\ \left(\mathrm{from}\ 2005-2011\right)=\frac{Full\ immunization\ coverage\ 2011-2005}{6}$$
$$\mathrm{Annual}\ \mathrm{increment}\ 3\ \left(\mathrm{from}\ 2011-2016\right)=\frac{Full\ immunization\ coverage\ 2016-2011}{5}$$
$$\mathrm{Annual}\ \mathrm{increment}\ 4\ \left(\mathrm{from}\ 2016-2019\right)=\frac{Full\ immunization\ coverage\ 2019-2016}{3}$$
The projected national full immunization coverage for the year 2025 was calculated by adding the mean smoothed average for the period 2000-2019 multiplied by 6 on to the 2019 coverage, the calculation is shown below:
$$=\left(2019\ DHS\ coverage\right)+\left( mean\ for\ 2000-2019\ast 6\right)$$
$$=44.1+\left(1.58\ast 6\right)=53.6$$
Where mean is the smoothed average of the yearly increment in full immunization coverage from 2000 to 2019.
Six is the year difference between the last Ethiopian DHS in the year 2019 and the target year 2025.
The primary objective of the Ethiopia national EPI comprehensive multi-year plan (2016-2020) was to achieve at least 90% national coverage and 80% in every district with all vaccines by 202 0[
5]. This target is currently adjusted to 75% to be achieved by 2025 according to the Ethiopian Health Sector Transformation Plan II (2020/21-2024/25) [
36].
The inequality analysis was conducted using the WHO Health Equity Analysis Toolkit (HEAT) software
, Version 4.0 (beta) Geneva, World Health Organization,2020 [
37]
. Summary measures used to assess inequalities were relative concentration index and ratio. The ratio is a simple measure of inequality which do not account for the population share and measures the ratio of two population subgroups. Since place of residence has only two categories (urban and rural) in order to measure the relative inequality, ratio was used as a summary measure. The ratio is the immunization outcomes in the most-advantaged (urban) group divided by the immunization outcomes in the most-disadvantaged group (rural). The ratio takes a value of one if there is no inequality. For favorable health indicators like full immunization coverage, values greater than one indicates a concentration among the advantaged (urban) and values smaller than one indicates concentration among the disadvantaged subgroup (rural) [
38].
Since educational status and household wealth index have more than two subgroups with ordered dimensions, the relative concentration index was used as a summary measure to assess the inequality among the subgroups. The relative concentration index (RCI) is a complex measure of inequality which shows the health gradient across population subgroups by taking into account all population subgroups on a relative scale. The value of RCI is bounded between − 100 and 100 and takes the value of 0 when there is no inequality. Positive values indicate a concentration of the indicator among the advantaged (richest, secondary and higher education), while negative values indicate a concentration of the indicator among the disadvantaged (poorest, no education). The greater the absolute value of RCI, the higher the level of inequality [
38].
In order to add the 2019 data, we reconstructed the concentration curves and concentration indices for the years 2000-2019 using the XY (scatter) chart-type in Microsoft Excel. The inequality analysis was done for household wealth quintiles (lowest or poorest, second, middle, fourth, highest or richest), mother’s education status (no education, primary, secondary and above), and place of residence (rural/urban).
The concentration index “
Conc-I” was computed using the following formula:
$$Conc-I=\left(p1L2-p2L1\right)+\left(p2L3-p3L2\right)+\dots +\left(p\mathrm{T}-1L\mathrm{T}-p\mathrm{T}L\mathrm{T}-1\right)$$
where
p is the cumulative proportion of children,
L is the cumulative proportion of fully immunized children, and T is the number of socioeconomic groups.
When the concentration index increases or when the concentration curve moves away from the line of equality it shows greater inequality in distribution of the health variable of interest (full immunization), on the other hand, when the concentration curve gets closer to the line of equality it shows a lesser inequality in the distribution of the health variable.