Dataset of the impact of food insecurity on health outcomes in sub-Saharan Africa

Abstract

Objectives

The 2030 Sustainable Development Goals (SDGs) have goals and targets, including food insecurity and health outcomes. Hence, information about socioeconomic variables that determine the health outcomes of people is essential for health-related research, planning, and policy development. Therefore, this data paper aims to present (describe) the dataset of eight socioeconomic variables for 31 sub-Saharan Africa (SSA) countries from 2001 to 2018.

Data descriptions

The dataset was official information obtained via open online sources from the Food and Agricultural Organization (FAO), the World Bank (WB), and the United Nations Development Programme (UNDP). It included 558 observations and eight variables, such as life expectancy (LEXP), infant mortality (INFMOR), the prevalence of undernourishment (PRUND), average dietary energy supply (AVRDES), Gross domestic product per capita (GDPPC), general government health expenditure (GOVEXP), urbanisation (URBAN), and mean years of schooling (MNSCHOOL). Moreover, all the data estimation is conducted by Statistical Software (STATA) version 15. This paper achieved its intended objective with a detailed and understandable description.

Objectives

Several factors determine human health outcomes. Specifically, food insecurity and health outcomes are inseparable and are the agenda of the 2030 SDGs, researchers, and policymakers. Thus, awareness of the socioeconomic factors that affect people’s health outcomes is crucial to provide evidence-based findings and developing achievable and comprehensive policies. Therefore, this data paper aims to describe a dataset used in Beyene’s [1] article, which is freely and openly accessed at https://doi.org/10.6084/m9.figshare.22606442.v1 [2]. Specifically, this paper seeks to provide a detailed overview of the data sources, data collection methodology, and data quality. By doing so, researchers can better understand the context in which the data was gathered, its accuracy and its implications for the research.

There is evidence for the rationale for the research and collecting the data in relation to food insecurity and health in the context of SSA. According to Roser and Ritchie [3], food insecurity is highest in SSA nations, where approximately one-third of the population is classified as highly insecure. Besides, 11% (820 million) of the world’s population was malnourished in 2018, with SSA nations continuing to bear a more significant burden [3]. Likewise, Word Count [4] reported that the amount of people suffering from lack of food has grown since 2015; hence, meeting the SDGs of ending hunger by 2030 would be a major problem [4]. Since food insecurity is one of the determinantal factors for health outcomes, the author was motivated to collect the data and examine the implication of food insecurity on life expectancy and infant mortality in SSA countries.

Data descriptions

The dataset is secondary panel macro-level yearly data collected (via open online databases) from well-known institutions, such as FAO [5], the WB [6], and the UNDP [7]. The dataset has 558 observations and eight socioeconomic variables (for more details, see Data File 1 in Table 1). These variables are LEXP,¹ INFMOR,² PRUND,³ AVRDES,⁴ GDPPC, GOVEXP (% GDP), URBAN,⁵ and MNSCHOOL. Except for the mean years of schooling and average dietary energy supply, all variables were collected from WB. However, the mean years of schooling were obtained from UNDP, and the average dietary energy supply data were FAO data obtained from Knoema. The LEXP and INFMOR, dependent variables, are proxies of health outcomes and are transformed into natural logarithms.⁶ On the other hand, PRUND and AVRDES are proxy variables for food insecurity and are the target independent variables. All independent variables, except for GDPPC and MNSCHOOL, were measured in percentage. However, GDPPC is measured in constant 2010$ while MNSCHOOL is in the number of years (for more details, see Table 2 of a summary).

Table 1 Overview of data files/datasets

Table 2 Summary of the variables sources

The data were collected in several steps. First, the study selected the target and the control variables along with their proxies based on the existing literature. Second, to have long time-series data and a large sample size, the data was started in 1990 and included for all SSA countries. Third, the author searched for data from all available online sources; however, due to data availability for most variables and countries as well as for the sake of consistency, the data were collected only from the WB, UNDP, and FAO. However, the data for target variables, such as the prevalence of undernourishment and average dietary energy supply not available before 2001 in most SSA countries. Moreover, some countries have no data on the prevalence of undernourishment, even since 2001. Hence, since the balanced data is better than the unbalanced one, the author has excluded some countries with no or missing data and finally arrived at 31⁷ sampled SSA countries from 2001 to 2018. The collected data were recorded in an Excel sheet and imported to STATA 15 for analysis. Further, before the final estimation, the most efficient model was selected (for more details, see Additional file 1).

Limitations

Since the data collection was conducted at the end of 2020, the data ended in 2018. Thus, the data did not consider the recent global events, such as the Coronavirus disease of 2019 and the Russia-Ukrain war. In addition, the data were estimated using STATA 15. Moreover, the data was only focused on socioeconomic factors of health outcomes. Therefore, other scholars can extend the year until 2022, consider the environmental and political factors, and employ the latest statistical software (STATA 17).