Introduction

On 1^stJanuary 2016, the world officially began implementation of the action plan based on Sustainable Development Goals (SDGs). Goal three aims to ensure health and well-being for all people of all ages by improving reproductive, maternal and child health.¹

Study of the Horn of Africa countries is especially important, as the situation remains disastrous.²The Horn of Africa region is plagued by a set of complex,³often interrelated factors including environmental degradation, climate-related disasters such as droughts and floods.^4-5

Multiple factors hinder access to and utilization of health services in the Horn of Africa. These factors include lack of a functional health system, geographical accessibility, financial barriers and limited availability of services.^1-2

In particular, this study aims to construct a synthetic indicator of maternal and child health to enable comparison between five countries in 2017 in the Horn of Africa.

The index also allows to study the impact of each variable individually so as to determine disparities in the variables associated with the SDGs for each country. Additionally, the research explores the relative impact of each variable by using the correction factor.

Method

The methodological approaches used in this study are based on the construction of a synthetic index that follows Pena's method (DP₂).⁶ The DP₂ provides an ideal solution to the problems involved in devising a synthetic indicator, particularly those related to aggregation and weighting of simple indicators and information duplicity.⁷

The DP₂ measures the distance between the issue studied in each country and a fictitious base reference. We take as reference a theoretical country that obtains the worst values for the variables studied⁷.

The DP₂ from country j is defined as follows:

where d_i=|x_ji- x_*i| is the distance between the value of variable i in country j and the reference base. The reference base comprises the results from an imaginary country which reflects the worst possible scenario for all the variables where X_*= (x_*1, x_*2,..., x_*n) coincides with the minimum vector. The reference base would therefore be attributed a value of zero in the synthetic indicator.⁷n is the number of variables, (_i is the standard deviation of variable i, and (1-R²_i,_i-1,...,₁) is a “correction factor”⁶that avoids redundancy.^{2,6 7 8-9}The coefficient of determination, R²_i,_i−1,...,1,is the determination coefficient in regression X_i over X_i-1, X_i-2,...,X₁, which is already included, with R²₁=0. Put differently, the coefficient measures the part of the variance of each variable explained by the linear regression estimated using the preceding variables.²The ordering of the variables corresponds to their relative weight measured in terms of linear correlation with the final synthetic indicator.⁷The input order of the variables is determined by an algorithm that reaches convergence when the indicator fulfils a number of desirable properties.^{2,6 7 8-9}

It is also possible to establish an order or hierarchy based on the amount of information that each variable contributes to the DP₂. To determine this, we construct the Ivanovic Discrimination Coefficient (IDC).^{10 11-12}This indicator finally shows us the amount of information provided by the i-th variable. It can range from 0 (in the event that the values of Xi are identical in all the countries) to 2 (in the event that the variable has total discriminatory power, that is, when the amount of information varies greatly across territories). Thus, the closer the IDC value is to 2, the more useful information it provides to explain the differences in the level of child and maternal health in the countries studied.^11-12

Data were collected from the work of the United Nations Statistical Commission, which created the Inter-agency and Expert Group on SDG Indicators (IAEG-SDGs). In particular, we used five variables of child and maternal health associated with the goal 3¹(Table 1), using as a reference the detailed information contained in a set of variables set out under the SDGs in the Report 2018,¹³which provide a more extensive and more reliable set of statistics on the SDG 3 of Horn of Africa. The countries included into the analysis were Ethiopia, Kenya, Somalia, Eritrea and Djibouti. To guarantee fulfilment of the properties of the synthetic indicator, we multiply specific variables whose increase implies a worsening of the child and maternal health by −1.

The year of analysis is 2017, but for those variables where information was not available for that date, the nearest year was taken as an alternative. This has occurred in the variable “Attended by skilled health personnel, percentage”, whose available information is from 2016.

Results

Constructed from the variables included in Table 1, the result is shown in Table 2, which ranks the five countries by level of child and maternal health.

The resulting classification (Table 2) shows, first, a distance of almost 5 points between the best-positioned country (Djibouti) and the worst-positioned (Somalia) in 2017. These results indicate a relatively high disparity between the countries analysed.

The results show that Djibouti made the greatest progress toward the goals for child and maternal health, with a distance of 4.56 from the baseline (Table 2). It was followed by Kenya (3.46), which accounts for 29% of the total population of the Horn of Africa.

Taken together, Somalia and Ethiopia account for nearly 70% of the population of the Horn of Africa. They, in contrast, are the countries with the worst theoretical scenarios (Table 2).

If we analyse the results obtained for the variables with the greatest inequality in intercountry values (IDC),¹⁰the most discriminating variable is “Attended by skilled health personnel, percentage” (Table 1). The second-most-discriminating variable is “Maternal mortality ratio per 100,000 live births”.

In addition, by means of correction factors, the synthetic indicator DP₂ only includes the new information from each variable.⁹ In particular, the variable “Maternal mortality ratio per 100,000 live births” contains all of its information, so the corresponding correction factor is 100%, as a result of being most closely correlated with it (Table 1).

Discussion and conclusion

The DP₂ method shows territorial disparities in child and maternal health in the Horn of Africa in 2017. We obtained a difference of 4.56 units between Djibouti and the reference value. Djibouti achieved a higher level of child and maternal health, but it accounts for only 0.5% of the total Horn of Africa population. At the opposite extreme, Somalia registers extremely low values in the set of partial indicators.

Priority must be given to interventions to address the variables that have greater power to explain the differences in the values between countries relative-primarily the variable “Attended by skilled health personnel, percentage”.

The differing values of these variables suggest that progress in maternal health is uneven throughout the Horn of Africa, while fewer territorial differences exist in the variables associated with child health as defined in the SDGs.

In summary, delivery of health services is greatly in need of improvement, especially in Somalia and Ethiopia, and there is an urgent need to increase the number of health workers throughout the region to lower maternal and infant mortality.¹⁴

In general, the DP₂ classification for these countries differs from that made by the Human Development Index (HDI) for countries with low human development in 2017 (Table 2). In this sense, our analysis takes into account a range of SDGs variables, some of which are not included in the HDI.