Cold, dark and malnourished: a cross-sectional analysis of the relationship between energy poverty and household burden of malnutrition in sub-Saharan Africa

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • The study analyses data from multiple sub-Saharan African countries to examine the complex and multifaceted nature of the relationship between energy poverty and household burden of nutrition.

  • Computation of the different domains of energy poverty and the multidimensional energy poverty index adequately captures the multidimensional nature of energy poverty at the household level.

  • The use of different classifications of household burden of malnutrition provides a broad overview of assessing a wide range of the conditions of household burden of malnutrition.

  • The estimates of household burden of malnutrition may be lower than estimated due to the exclusion of men in the analysis because of the lack of data on anthropometric measurements for men.

  • There is the challenge of establishing causality, as the data capture a single point in time and do not inherently provide information about the temporal sequence of events.

Introduction

That malnutrition among women and children remains a public health concern in sub-Saharan Africa (SSA) is hardly contested by governments, non-governmental and development agencies. Malnutrition is broadly defined by the World Health Organization (WHO) as ‘deficiencies or excesses in nutrient intake, imbalance of essential nutrients or impaired nutrient utilisation’.1 Malnutrition can thus present as undernutrition or micronutrient deficiency on the one hand where there is inadequate nutrient intake or as overnutrition where there is excess nutrient intake. Overweight/obesity is the common form of overnutrition while wasting, stunting, underweight and micronutrient deficiencies are the common forms of undernutrition.1

Undernutrition could result from food insecurity where individuals do not eat nutritious food of good quality and in sufficient quantities to support optimal growth and physiological functions. Recent estimates reveal that compared with other regions of the world, SSA witnessed the highest percentage increase in food insecurity by 5.4% from 2019 to 2020 and the region has also experienced the slowest decline in stunting.2 The prevalence of childhood anaemia is estimated to be 64%, with an accompanying high morbidity and mortality.3 Additionally, the WHO acknowledges Africa to be facing a growing problem of obesity and considers it a ‘ticking time bomb’ if left unchecked.4 Thus, there is a continuous call for appropriate nutrition-sensitive interventions to address current malnutrition concerns—both under and overnutrition.

Malnutrition at the household level can be present in many forms, but the most common presentation is reported among women and children under five. The coexistence of women and children who are both malnourished in the same household increases the burden of malnutrition experienced at the household level. In the SSA context, although there is ample literature on the factors that influence childhood malnutrition and overweight/obesity among women, relatively little is known about how household energy influences these conditions of malnutrition. But evidence from research suggests that access to energy is known to be critical for the reduction of poverty, improvement in health, increased productivity, boosting competitiveness and promoting economic progress.5 6 Energy demand and consumption have and continue to increase significantly as countries become more developed, people become richer and the population increases.7 The fact that Goal 7 of the Sustainable Development Goals (SDGs) aims to ‘ensure universal access to affordable, reliable, sustainable and modern energy by 2030’ attests to the increasing recognition of the importance of energy to development by both national governments and international development organisations.8 Apart from the global socioeconomic development benefits of having access to clean energy, at the household and individual levels, there are numerous health benefits of using clean energy, particularly for women and children. An estimated 2.4 billion people who do not have access to clean fuels for cooking and heating are left exposed to harmful gas from open fires or stoves leading to the death of an estimated 3.8 million people—mainly women and children.9 10

Research on the relationship between energy/energy poverty and household health and well-being is skewed towards the global north, with research conducted in the SSA region at a budding stage, even though the region has the highest proportion of households that are energy poor. Hence, this study aims to examine the influence of household energy poverty on nutritional outcomes of households in SSA. This study contributes to knowledge in two unique ways. First, our findings contributes to our current knowledge of how energy poverty influences household well-being. The World Bank estimates that about a billion people, mostly in SSA and Asia, do not have access to daily electricity and acknowledges that the current state presents a major impediment to progress in development indicators such as health, education, food security, livelihoods and poverty reduction.11 To the best of our knowledge, except for a few studies,12 13 there has not been an empirical inquiry into the role energy poverty plays in household health and nutritional status within the SSA context. Second, most studies linking energy to health and well-being have often adopted unidimensional energy poverty measures. For example, Ghimire and Vatsa14 found that availability of electricity in the household was positively associated with overweight and obesity in women, while access to electricity reduces the likelihood of being underweight. Ahmed et al
15 also found a significant association between chronic childhood malnutrition and household biomass fuel (BMF) when they used BMF consumption as a measure of energy poverty.

It is important to indicate that while various measures of household energy usage have been investigated, a household’s energy profile is delivered through a variety of technologies that span grid-based and off-grid systems, as well as a wide range of applications.16 Thus, investigating the effect of household energy using a unidimensional measure of energy poverty does not provide a wholistic measure of the phenomenon. In this study, we employ a more robust multidimensional framework of energy poverty, the Multidimensional Energy Poverty Index (MEPI) suggested by Nussbaumer et al which encompasses five dimensions of household energy poverty—lighting, cooking, services through household appliance, communication and entertainment.17 We further propose a conceptual model which illustrates the relationship between household energy poverty and household burden of malnutrition (see figure 1).

Figure 1
Figure 1

Conceptual model showing the mechanisms of impact of household energy poverty on household burden of malnutrition.

Pathways of the mechanism of influence of household energy poverty on household burden of malnutrition

Figure 1 shows the potential pathways through which energy poverty may influence a woman to be overweight/obese and/or children to be malnourished (stunted/wasted/anaemic or overweight) and further illustrates how children of obese mothers could be malnourished. The conceptual model also illustrates how energy poverty impacts the various conditions of malnutrition among women and children and the coexistence of these conditions in the same household. For example, on the one hand, inadequate access to clean energy results in the reliance on poor energy sources that are toxic and contribute to general health problems, particularly respiratory illness, stress and even mortality.18–20 There is also ample evidence that shows that stress and deprivation and other associated conditions lead to being overweight or obese21 22 via overeating, poor sleep quality and reduced activity. Therefore, in our conceptual model, we hypothesise that lack of clean energy at the household level could lead to stress and contribute to overweight and obesity among women in the household. On the other hand, a higher prevalence of stunting among households using unclean cooking fuels in some developing countries has been reported.23 Usually the fine particulate matter (PM2.5) released in households using biogas enters the lungs when inhaled, leading to increased respiratory issues in children and general morbidity which can hamper their growth.24–26 Loss of appetite, particularly in children, is associated with morbidity, while the later could also reduce usage of nutrients from foods consumed. Therefore, in our conceptual model, we hypothesise that household energy poverty impacts malnutrition among children in the household. Additionally, energy poverty, particularly in the domain of cooking (cooking fuel), storage (ownership of a refrigerator/freezer) and access to information on proper feeding and hygienic practices via communication/entertainment channels such as mobile phone, radio and television can affect feeding practices, thus contributing to malnutrition, especially among children.

In addition to these illustrated pathways, there are studies that have shown a link between maternal body mass index (BMI) and children’s nutritional status.27 28 Maternal BMI is an important determinant of the birth weight in infants29 30 and the birth weight of infants is known to influence children’s nutritional status even in later adult years. This linkage notwithstanding, the coexistence of maternal overweight/obesity with undernourished children in the same household is an important phenomenon associated with the current nutrition transition being experienced in low-income and middle-income countries31 which requires further investigation. And from the perspective of nutrition, health and well-being, the coexistence of women and children who are both malnourished in the same household increases the burden of malnutrition experienced at the household level. More importantly, energy poverty at the household level further exacerbates these conditions of malnutrition through the pathways illustrated in our conceptual model. The model therefore proves useful in examining the relationship between household energy poverty and household burden of malnutrition. The methods and analysis for the current paper is thus guided by our conceptual model.

Methods

Source of data

We used nationally representative cross-sectional data from the most recent Demographic and Health Survey (DHS) conducted across SSA. The data were downloaded from the Integrated Public Use Microdata Series (IPUMS) website.32 IPUMS extracts, transforms and loads data from numerous nationally representative surveys into a single view model using a data warehousing method. The IPUMS approach to data curation is possible because all DHS across participating countries are conducted using standardised survey design and data collection procedures. The DHS also follows a standard protocol for collecting data from respondents across participating countries.

Sample design and sample size

The DHS employs a multi-stage cluster sampling design in selecting respondents. Enumeration areas (clusters) are first selected from the census sampling frame. Households are then selected based on probability proportional to the size of the enumeration area. Heads of the selected households and women and men in the reproductive age (15–49 and 15–59 years, respectively) as well as children under the age of five in the selected households are eligible for interview. The sample for the current study is limited to countries with valid data on all variables of interest from the most recent DHS conducted in the country. The current study includes 18 SSA countries with available data from the recent demographic survey spanning 2008–2016. The analytical sample consists of 99 329 households from the 18 countries included in the study (see online supplemental appendix 1).

Supplemental material

Data collection

Data on household characteristics and individual demographic and socioeconomic characteristics, and health measures were collected as part of the DHS through face-to-face interviews conducted by trained enumerators using structured questionnaires. Data on various anthropometric measures including weight and height were collected and used to measure the nutritional status of women (ie, BMI) and children under five (ie, stunting, wasting, overweight). Anaemia among women and children was measured using their haemoglobin concentration levels.

Measurement of variables

Dependent variables

The outcome variables for our study include varied conditions of household burden of malnutrition measured using the nutritional status of women (in the reproductive age: 15–49 years) and children under five years in the surveyed households. The typologies of household burden of malnutrition examined include undernutrition, overnutrition, anaemia, double burden of malnutrition and triple burden of malnutrition. Household burden of malnutrition was derived based on the nutritional statuses of women and children in the surveyed households. The nutritional status indicators among children under five we examined included stunting, wasting, overweight and anaemia, while the nutritional status indicators for women included anaemia and BMI. Using these indicators, a child is stunted when his/her height-for-age z score is less than −2 and wasted when his/her weight-for-height age z score is less than −2.33 A child with Hb concentration of ≤110 g/L is anaemic.34 The nutritional status of women in the reproductive age (15–49 years) in the study households was classified in synchrony with the WHO’s categorisation of BMI as underweight (≤18.49 kg/m2); normal weight (18.50–24.99 kg/m2); overweight (25.00–29.99 kg/m2) and obese (≥30 kg/m2).35 The women were also classified as anaemic if their haemoglobin concentration was ≤100 g/L.36

Household burden of malnutrition was classified based on the nutritional status of the women and children in the household. A household with a child who is stunted and/or wasted and/or a woman is underweight is classified as an ‘undernutrition’ or ‘undernourished’ household. Similarly, a household suffers from overnutrition when there is a child and/or a woman who is overweight/obese in the household and such a household is classified as an ‘overnutrition’ or ‘over nourished’ household. Also, a household is classified as anaemic when either a child or a woman in the household, or both, are anaemic. A household is classified as having ‘double burden’ malnutrition when at least a child under five and a woman in the reproductive age in the household have both undernutrition and overnutrition, irrespective of who has which condition. When a household is anaemic, that is, either a child under five or a woman in the reproductive age in the household is anaemic or both are anaemic (see earlier definition of anaemia), and the household also has double burden of malnutrition, it results in a triple burden of malnutrition (ie, undernutrition, overnutrition and anaemia). Such a household is thus classified as having a triple burden of malnutrition. Figure 2 shows the classifications of the various conditions of household burden of malnutrition as applied in this study.

Figure 2
Figure 2

Classification of the different conditions of household burden of malnutrition.

Independent variables

The main independent variable in this study is household energy poverty. We assessed energy poverty in five domains and computed an MEPI (ie, energy deprivation) at the household level in accordance with the framework by Nussbaumer et al and further explained by Lin and Okyere.17 37 Household energy poverty was measured based on a household’s access to clean energy for lighting, cooking, household appliances (such as a refrigerator and or a freezer for services such as food storage), communication and entertainment. Household energy deprivation was assessed in each of the five domains and each domain was weighted by applying the suggested weights.17

This paper makes a significant contribution by comprehensively assessing energy deprivation at the household level across multiple dimensions. To accomplish this, we were guided by the MEPI framework introduced by Nussbaumer et al
17 and adopted by others such as Lin and Okyere.37 In measuring energy poverty (table 1), a household is considered energy poor if: (1) it is not connected to the grid or does not use solar or any other renewable energy source for electricity; (2) it uses BMFs such as firewood, charcoal, or residue for cooking or heating; (3) it does not have a refrigerator or deep freezer; (4) it does not have a TV or radio set; and (5) it does not own a mobile phone. In this paper, we assigned an equal weight for all indicators (0.20).

Table 1

Dimension and indicators of the Multidimensional Energy Poverty Index

Control variables

Based on extant literature on the factors that influence the nutritional status of households and their members, the covariates considered in this study include: (1) characteristics of children under five in the household such as their age and sex38; (2) characteristics of women in the reproductive age in the household including their age, marital status and level of education39; and (3) household factors such as the total number of people in the household (ie, household size), and the characteristics of the household head (ie, age and sex). Additionally, household sanitation was considered as improved when the sanitation facility used by the household was linked to a public sewer or septic system, or when it comprised a pour-flush latrine, a ventilated improved pit latrine or a basic pit latrine. A household is considered to have access to improved water source if the household has access to public water systems, boreholes, safeguarded dug wells, protected springs or rainwater as their primary source of drinking water.40 41

Methods of analysis

All analyses were conducted using the Stata statistical software package, V.4.2 (2017; StataCorp, College Station, TX, USA). Statistical significance was set at the 5% α-level (p<0.05). The data were weighted for analysis to avoid inaccurate estimates of the primary sampling unit used during sample selection. The characteristics of the study sample were described using means and percentages. Logistic regression analysis was used to examine the relationship between household energy poverty and household burden of malnutrition, where ‘1’ represents a household with a condition of malnutrition and ‘0’ otherwise. Altogether, five conditions of household burden of malnutrition were examined. These were: (1) undernutrition (ie, undernourished household=‘1’ where a child under five and or woman in the reproductive age in the household is undernourished and households without undernutrition=‘0’ where there is no undernourished child or women in the household hold), (2) overnutrition (ie, over-nourished household=‘1’ where a child under five or a woman in the reproductive age or both are over nourished and household without overnutrition=‘0’ where there is no child or woman with overnutrition in the household), (3) anaemia (ie, anaemic household=‘1’ where a child under five or a woman in the reproductive age in the household or both are anaemic and non-anaemic household=‘0’ there is no anaemic woman or child in the household), (4) double burden malnutrition (ie, double burden malnutrition household=‘1’ where a child under five is undernourished and a woman in the reproductive age is over nourished or vice versa and ‘0’ otherwise) and (5) triple burden of malnutrition (triple burden malnutrition household=‘1’ where a child under five is undernourished and a woman in the reproductive age is over nourished (or vice versa) and a child under five and or woman in the reproductive age is anaemic and ‘0’ otherwise). The following equation was estimated in this study:

Embedded Image

where, Embedded Image

is the condition of household malnutrition i
; Embedded Image represents the household energy poverty k in household j
; Embedded Image , Embedded Image and Embedded Image represent a vector of characteristics for woman k
, child i and household j
, respectively. Embedded Image signifies the randomly distributed error term; β represents the coefficients of the regressors and α is the constant term.

Patient and public involvement statement

This study uses secondary data from publicly available data that can be accessed from the website of IPUMS DHS. The study does not involve collecting primary data from patients. There was also no patient or public involvement in the development of the research questions, the methods or the results.

Results

Characteristics of study sample

The households in the study had 6.57 members on average (table 2). The average age of household heads was 38.70 years. A little over three-quarters (77.17%) of the households were headed by males. A little less than 6 out of 10 households (55.58%) used improved sources of drinking water, and a similar proportion (58.11%) used unimproved toilet facilities. Also, the proportion of households belonging to higher wealth quintiles decreased as wealth quintiles increased. About three-quarters of the households in the study were in rural areas, and a little over a quarter were in urban areas. The average age of children under five in the households was 28.16 months, with children between 24 and 59 months constituting more than half of the total sample of children under five years. Male children constituted slightly more than half of the sample of children under five years. Women of reproductive age in the households in the study sample were 29.08 years old on average, with those in the 25–29 years age group constituting the highest proportion and the 15–19 years age group constituting the least proportion. Nearly nine out of ten of the women were married or cohabiting at the time of the survey, while about two out of five had attained primary level of education and only about 2% had completed tertiary or higher level of education.

Table 2

Distribution of the study sample by sociodemographic characteristics of the household and members of the household (women 15–49 years and children under five years)

Household energy poverty

The results show different levels of deprivation across the various domains of energy poverty (table 3). About three out of four households in the study are energy poor in terms of lighting, meaning these households do not have access to clean energy for lighting. Similarly, more than nine out of ten households are energy poor in terms of cooking fuel. These households used non-renewable sources of energy such as wood, charcoal and animal residue, for cooking. Only about 7% of the households used clean/modern fuel for cooking. In terms of household assets, a little less than one-tenth of the households did not own a refrigerator or freezer; about 45% did not own a television or radio; and 38% did not own a mobile phone. Considering the MEPI, which ranged from 0.00 for households that are least energy poor to 1.00 for households that are most energy poor, the households in the study fell a little above the middle of the spectrum of the index with an average energy poverty index of 0.69 (table 3).

Table 3

Distribution of households by energy poverty status

Household burden of malnutrition

The results in table 4 show that about 45% of the households were undernourished households. Such households had either a child and or a woman who is undernourished (ie, stunted/wasted and underweight, respectively) in the household. Similarly, a little over one out of five of the households were overnutrition households, meaning a child under five years in the household was overweight or a woman aged 15–49 years in the household was overweight or obese or both. Again, about two-thirds of the households were anaemic households, that is, either a child under five or a woman in the reproductive age in the household, or both are anaemic. The combination of the various conditions of malnutrition among women and children in the households in the study translates into nearly 7% and 5% of the households being double and triple burden malnutrition households, respectively. Among children under the age of five, about one out of three were stunted, while less than one in ten were wasted or overweight. However, about six out of ten were anaemic. Among women in the reproductive age, a little over two-thirds were in the normal BMI category, but a little over one-tenth were underweight, while all together, about one out of five are overweight or obese.

Table 4

Prevalence of malnutrition among the study sample

Household energy poverty and household burden of malnutrition

Table 5 (section A) shows the results of the binary logistic regression analysis of household energy poverty and household burden of malnutrition using the MEPI as the indicator of household energy poverty. After controlling for other covariates (ie, household, and individual characteristics and country fixed effects), increasing household energy deprivation was associated with a higher likelihood of a household being an undernourished, double and triple burden malnutrition household (OR=1.859, 1.282, 1.324, respectively). Increasing household energy poverty was, however, associated with a lower likelihood of the household being over nourished (OR=0.451).

Table 5

Binary logistic regression analysis of the relationship between Multidimensional Energy Poverty Index (A) and domains of energy poverty (B) and household burden of malnutrition

The results for the various domains of energy poverty (table 5, section B) show that deprivation in any dimension of energy poverty was associated with an increase in the likelihood of households being undernourished. Also, except for energy poverty with respect to cooking (ie, using biofuels), deprivation in all other domains of household energy was associated with a decrease in the likelihood of a household experiencing overnutrition. Household deprivation in the domain of entertainment and access to mobile phones were associated with a lower likelihood of the household being anaemic. Again, household energy deprivation in terms of lightning was associated with a lower likelihood of the household experiencing a double burden of malnutrition. However, using biofuels and not having a refrigerator were associated with an increased likelihood of households having double and triple burden malnutrition. The results of the full model for the MEPI as well as the full model results of the various domains of energy poverty are shown in online supplemental appendices 2 and 3, respectively.

Discussion

Energy poverty is fast being recognised as a public health concern particularly in developing countries and while energy poverty in itself is a developmental challenge, there are also far-reaching consequences for nutrition, health and well-being in general. To the best of our knowledge this study is the first to examine the influence of different domains of energy poverty and a multidimensional index for energy poverty (ie, a composite score derived from the different dimensions) on different measures of household burden of malnutrition in the SSA context. Overall, our results indicate that energy poor households are more likely to suffer from undernutrition but less likely from overnutrition. The finding that energy poverty is associated with an increased likelihood of household undernutrition is consistent with the findings of previous studies.42 43 This may be because household energy poverty impacts food preparation and storage which in turn influences nutritional status. For example, lack of access to modern cooking fuels, such as gas or electric stoves in a household, often leads to the use of solid fuels such as wood or charcoal, which are not only unsustainable but also have negative health impacts.44 45 Often, inhaling smoke generated from such solid fuels causes respiratory and other health issues.46 Additionally, the energy required for cooking can make it difficult for households to afford adequate food.47 Given that in many SSA countries, energy is required to have access to clean water, the lack of the former results in a higher risk of waterborne diseases and other health issues, which ultimately contributes to poorer nutrition among members of the household thus impacting the household burden of malnutrition as illustrated in our conceptual model.

Our finding on the relationship between energy poverty and overnutrition is contrary to that observed in previous studies.48 Earlier studies conducted predominantly in European countries49–51 have shown that energy poverty has a negative impact on mental health (such as undermining self-efficacy, self-esteem, life satisfaction and depression) which then increases the likelihood of one being obese via overeating or the increased intake of high-calorie foods. Probable explanations for the observed result in the current study could be that among SSA populations, households that are energy poor are predominantly in rural areas with less access to the high-calorie foods and thus have a decreased likelihood of being obese in general and not only because of being energy poor but also because of poverty in socioeconomic terms. Secondarily, based on the theory of habituation, which posits that individuals become acclimatised to a stimulus and their response to it decreases over time, energy poverty in SSA countries may be less likely to lead to mental health issues as compared with Western countries. Consequently, the proposed pathway linking energy poverty to obesity in the Western context is unlikely in the SSA context.

With respect to household’s multidimensional energy poverty, the results from the current study show that the more energy poor a household is the higher the likelihood of the household having a double burden or triple burden of malnutrition. However, analysis of the domains of energy poverty shows some nuances. For instance, while having access to a refrigerator in a household can positively impact access to a more diverse and nutritious diet, including iron-rich foods, which can help reduce the risk of anaemia and undernutrition, it may also lead to an increase in the availability and consumption of high-calorie foods, potentially resulting in overnutrition. Thus, the need for examining probable mediating factors between specific dimensions of energy poverty and nutrition outcome is not only very important for future studies but also useful for providing context specific insights. Additionally, numerous other factors, such as dietary practices, religious and cultural factors influence nutrition related behaviours and practices and may thus present an opportunity for the increased incidence of double burden or triple burden of malnutrition in a household as these factors impact the nutritional status (ie, undernutrition, overnutrition and anaemia) of the members of the household.

The findings of this study make useful contributions to knowledge in several ways. First, the nuances in the findings of the present study shed light on the relationship between household multidimensional energy poverty as well as its different components and household burden of malnutrition. The findings suggest that while there is a relationship between the MEPI and household burden of malnutrition, a focus on the domains of household energy poverty is equally important as the index may not reveal which domains of household energy is/are lacking and therefore need to be addressed. Second, this study is unique in that it uses nationally representative data from SSA to explore household energy poverty as a contributing factor to household burden of malnutrition. Considering that majority of research reported by previous studies on this subject are seen in the developed countries context, these results from the SSA context highlight the need for further research to unpack the nuances identified in this study. These strengths notwithstanding, this study is not without limitations. First is the use of secondary data with limited measures for computing the dependent variables. In the present study, anthropometric measures and haemoglobin concentration levels were available for women in the reproductive age and children under five but not men. The household burden of malnutrition therefore excludes men. However, considering that households typically include men, the exclusion of men in the measurement of nutritional status indicators does not give a full picture of the household burden of malnutrition. We therefore recommend that population-level surveys such as the DHS and nutrition assessment surveys collect data on anthropometric measures for males in addition to females and children under five so that the full complement of household burden of malnutrition can be examined. Second, some variables were not included in the data and so these variables were not included in our analysis. For example, the DHS did not have questions to allow for the examination of household indoor air population as a component of energy poverty. Including this indicator in our analysis as has been done in other studies that explored the multidimensional aspects of household energy poverty could have enriched our results. Third, although the DHS collects consistent variables across participating countries, some variables may be missing or collected differently, leading to a sample size that may vary when conducting pooled data analysis, as seen in the present study. Despite these challenges, the study provides valuable insights into the influence of household energy poverty on household burden of malnutrition in SSA.

Conclusion

The findings of this study reveal conclusions about the relationship between household energy poverty and household burden of malnutrition. First, there is a significant association between multidimensional energy poverty and household burden of malnutrition, specifically undernutrition, overnutrition, double burden malnutrition and triple burden malnutrition in SSA. Additionally, the different dimensions of household energy poverty showed significant association with different conditions of household burden of malnutrition. These findings suggest that the gains from addressing energy poverty in SSA could be twofold, ensuring access to clean, sustainable and renewable energy as targeted by the SDGs while also reducing the household burden of malnutrition. Additionally, addressing the different dimensions of household energy poverty will ultimately result in multidimensional energy security for households. Governments of SSA countries must therefore implement targeted interventions that aim to increase access to clean and renewable energy in all dimensions of household energy needs thus addressing domain specific energy poverty and ensuring multidimensional energy security while also tackling malnutrition. Specific policy interventions should target increasing access to modern energy for lighting and cooking to maximise the gains of ensuring access to clean and sustainable energy while also addressing malnutrition.

Data availability statement

Data are available in a public, open access repository. This study analysed secondary data which was accessed and downloaded from the IPUMS website on written request and approval.

Ethics statements

Patient consent for publication

Ethics approval

This study used secondary data from the most recent Demographic and Health Survey conducted in the study countries with valid data. Ethical approval for the demographic and health surveys was granted by ORC Macro Institutional Review Board. Individuals who are willing to participants in the survey give voluntary informed consent before being interviewed.

This post was originally published on https://bmjopen.bmj.com