Abstract
Background
Energy poverty, characterized by inadequate access to clean, reliable, and affordable energy, is a pressing issue in low- and middle-income countries (LMICs), with profound implications for health outcomes and healthcare utilization. In sub-Saharan Africa, including Ghana, energy poverty disproportionately affects rural and low-income households, exacerbating health disparities and limiting access to healthcare services. This study investigates the impact of energy poverty on health outcomes and healthcare-seeking behaviour. Specifically, it examines the likelihood of illness or injury, the duration of illness as an indicator of healthcare access, and the probability of consulting health practitioners, providing evidence to guide targeted policy interventions.
Methods
The study utilized data from the Ghana Living Standards Survey (GLSS 7) to examine energy poverty and health outcomes. The Multidimensional Energy Poverty Index (MEPI) was calculated to measure energy poverty among a sample of 7,460 households. Health outcomes were assessed based on the incidence and duration of illness or injury and healthcare-seeking behaviors. To analyze these relationships, econometric models such as binary logistic regression and Poisson regression were employed, with robustness checks conducted using the Lewbel Two-Stage Least Squares (2SLS) method to address potential endogeneity. The analysis was performed using STATA 18 software.
Results
The study finds that energy poverty is significantly associated with a higher probability of experiencing illness or injury (0.063, p < 0.01), while paradoxically linked to shorter illness duration (–1.299, p < 0.10). Among poor households, energy poverty significantly reduces the likelihood of consulting a healthcare practitioner (–0.036, p < 0.10). Lewbel 2SLS estimates confirm these patterns, showing a robust positive effect on illness incidence (0.181, p < 0.05) and a significant negative impact on healthcare utilization (–0.137, p < 0.10). These findings underscore energy poverty’s dual burden: increasing health risks while limiting timely access to care, particularly for the poor.
Conclusions
The findings highlight the critical effect of energy poverty on health outcomes in Ghana, particularly among the poorest and most vulnerable populations. Addressing energy poverty requires integrated policies that enhance access to clean and affordable energy while also improving healthcare services and public health education. Targeted interventions for vulnerable groups, including subsidies for clean energy technologies and energy-efficient appliances, are essential. This study calls for a nuanced approach that aligns energy access with public health goals, aiming to improve well-being and reduce health disparities in Ghana.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12889-025-23602-6.
Keywords: Energy poverty, Public health, Indoor air pollution, Multidimensional energy poverty index
Background
Energy poverty, a critical barrier to socio-economic development and public health improvement, has emerged as a significant policy concern worldwide [1]. Recognized for its central role in facilitating poverty reduction, economic transformation, and sustainable development, energy access and affordability are directly addressed under Sustainable Development Goal (SDG) 7, aiming for universal access to modern energy services by 2030 [2, 3]. The attainment of SDG 7 is pivotal not only for its intrinsic value but also for its instrumental role in advancing other SDGs related to health, gender equality, climate action, and economic growth [2].
Despite global strides towards improving energy access, sub-Saharan Africa remains significantly underserved, with an electrification rate of merely 43% and a staggering 78% of its population relying on solid biomass for cooking—a figure anticipated to rise with population growth [4]. This persistent energy poverty has profound implications for public health, exacerbating exposure to indoor air pollution, increasing vulnerability to respiratory diseases, and limiting access to healthcare services [5, 6]. The nexus between energy poverty and health is particularly concerning, as households without access to clean energy face heightened risks of illness, injury, and financial strain due to health-related expenditures [7, 8].
In Ghana, notable progress has been made in energy access, with an electrification rate of 86.3% achieved in 2016 and increased adoption of liquefied petroleum gas (LPG) for cooking [9]. However, these advancements mask persistent challenges, particularly in rural areas where reliance on solid biomass for cooking remains high [10, 11]. This reliance is a significant public health issue, contributing to indoor air pollution, which is responsible for approximately 16,600 deaths annually in Ghana [12]. Addressing this challenge requires a deeper understanding of how energy poverty affects health outcomes and healthcare-seeking behavior.
The literature on energy poverty in African contexts reveals a clear linkage between inadequate access to clean energy and a host of adverse health and social outcomes. Studies from Ghana, Uganda, and Nigeria illuminate the significant repercussions of energy poverty on physical and mental health, emphasizing a need for multifaceted policy interventions. One study identifies energy poverty as a determinant of respiratory infections among under-fives in Uganda [13], suggesting a convergence with findings by [14] which highlight the mental health toll of energy poverty on women in peri-urban African communities. This discourse was extended by demonstrating a quantitative association between energy poverty and mental health in Ghana, where energy deprivation severely escalates depression risks [15]. Additionally, another study reveals that energy poverty in Ghana leads to increased household health expenditures, with health insurance and remittances serving as buffers [16]. Further contextualizing these affects further, research on Nigerian women and students underscores the broader socio-economic consequences of energy poverty and emphasizes the importance of targeted energy policy reforms [17].
The relationship between energy poverty and health outcomes has been explored across various dimensions, revealing its significant effect on physical, mental, and economic well-being. In low- and middle-income countries (LMICs), the lack of clean energy access has been linked to health issues such as respiratory infections. Stevens et al. found that a 0.1 increase in the Multidimensional Energy Poverty Index (MEPI) increased the odds of acute respiratory infections (ARI) among children under five, particularly due to reliance on biomass for cooking and inadequate electrification [18]. Similarly, Khogali et al. [19] reported that electrification of healthcare facilities enhanced antenatal care, vaccination rates, and emergency services, with solar power proving critical for rural areas.
Mental health implications of energy poverty are equally critical. Lin and Okyere [20] highlighted that in Ghana, energy deprivation significantly increased depression risks, with the lack of household appliances being a major contributor. Zhang et al. [21] identified that women in rural China faced disproportionate mental health challenges due to energy poverty, worsened by intra-household labor inequalities.
Energy poverty also has substantial economic consequences. Nguyen-Phung and Le [22] found that energy poverty reduces health expenditures by reallocating resources to essentials such as food and daily necessities. Similarly, Lee and Yuan [23] emphasized the moderating role of urbanization in the energy poverty-health nexus, particularly in Sub-Saharan Africa, where urbanization helped mitigate adverse health impacts. Fan et al. [24] further demonstrated energy poverty’s role in amplifying health vulnerabilities, reducing populations’ resilience to public health crises, especially in high-income regions.
Additional studies provide critical insights into specific contexts. Irwin et al. systematically reviewed the health effects of electricity access in LMICs, finding that electrification improves mortality rates, disease burden, and access to healthcare, while unreliable electricity negatively impacts health outcomes. Wang et al. [25] showed that energy poverty in rural China significantly worsens health outcomes for older adults, with policy recommendations focused on enhancing energy services and targeted poverty alleviation. Furthermore, Katoch et al. [6]identified the adverse effects of energy poverty on health and education globally, advocating for economic improvement to achieve higher living standards.
These studies collectively underscore the cascading effects of energy poverty, encompassing heightened disease risks, amplified mental health challenges, and significant economic strains. Building on this body of literature, our study seeks to achieve two primary objectives. First, it examines how energy poverty affects health outcomes, specifically the incidence and duration of illnesses or injuries, as well as healthcare-seeking behaviors. Second, the study focuses on vulnerable populations, examining how energy poverty disproportionately affects the poorest households and exacerbates existing health disparities. By addressing these objectives, the research aims to provide actionable insights for policymakers and stakeholders, advocating for targeted interventions that mitigate the adverse health effects of energy poverty and promote equity in energy and healthcare access.
The paper progresses by detailing the methods and data employed for analysis, followed by a presentation of empirical findings and discussions. A robustness check is conducted to verify the results. The conclusion synthesizes these insights, offering final reflections on the study’s implications for understanding and addressing energy poverty’s effect on health.
Materials and methods
Data
This study utilizes data from the Ghana Living Standards Survey (GLSS), a nationally representative survey that offers extensive details on household demographics, health outcomes, income, education, and various other economic indicators. Conducted by the Ghana Statistical Service, the GLSS is instrumental in providing a broad spectrum of data crucial for assessing living conditions and informing policy decisions in Ghana. Our analysis is based on data from the seventh wave (GLSS 7), carried out in 2016/2017. The GLSS 7 covers a wide array of topics relevant to household well-being, including questions on health status, access to healthcare services and energy usage. This data allows for an in-depth examination of energy poverty and its implications for health outcomes and healthcare utilization among Ghanaian households. After to address missing observations, a sample of 7,570 households was used for the analysis.
Variables
Conceptualization and measurement of energy poverty
The construct of energy poverty is pivotal to our analysis, capturing the extent to which households are constrained in accessing sufficient, reliable, and affordable energy services. Literature on energy poverty acknowledges both objective and subjective frameworks for its measurement, each offering distinct insights into the phenomenon’s dimensions and implications [26, 27].
Objective measures traditionally include the energy expenditure-income ratio, delineating energy poverty as the share of household income allocated to energy or fuel costs. A higher ratio signifies greater energy poverty, with some studies adopting a 10% cut-off of income spent on energy to classify households as energy poor [27, 28]. Subjective measures, conversely, often reflect material deprivation, especially in colder seasons, operationalized through binary indicators of a household’s reported inability to adequately heat their home due to financial constraints [26].
Given the data limitations in capturing subjective measures of energy poverty within the GLSS and the complexity of energy poverty in developing contexts, our study employs the MEPI. The MEPI integrates both objective and subjective dimensions of energy poverty, resonating with the economic conditions and clean energy adoption rates in developing countries [29–31].
The multidimensional energy poverty index (MEPI)
The MEPI framework encompasses five dimensions and six indicators, reflecting a household’s access to cooking, lighting, connected household appliances, entertainment/education, and communication technologies as seen in Table 1. Drawing from the multidimensional poverty approach of the Oxford Poverty and Human Development Initiative [32], inspired by Amartya Sen’s theories on capabilities and deprivations, the MEPI provides a comprehensive measure of energy poverty.
Table 1.
Dimensions, indicators, and weights for multidimensional energy poverty
| Dimension | Indicator (Weight) | Variables | Deprivation cut-off |
|---|---|---|---|
| Cooking | Modern cooking fuel (0.205) | Type of cooking fuel | Any fuel use besides electricity, LPG, kerosene, natural gas or biogas |
| Indoor pollution (0.205) | Food cooked on stove or open fire (no hood/chimney), indoor, if using any fuel beside electricity, LPG, natural gas, or biogas | True | |
| Lighting | Electricity (0.20) | Has access to electricity | False |
| Services provided by means of household appliances | Household appliances ownership (0.13) | Has a fridge | False |
| Entertainment/Education | Entertainment/Education appliances ownership (0.13) | Has a radio/television | False |
| Communication | Telecommunication means (0.13) | Has a phone land line/mobile phone | False |
Source: Adopted from: [30]
Weighting within the MEPI is designed to reflect the relative importance of each dimension. The cooking and lighting dimensions receive additional emphasis due to their fundamental role in household energy needs, with cooking assigned a slightly higher weight to signify its critical importance in developing countries’ contexts. Consequently, indicators within the cooking dimension are weighted at 0.205 each, lighting at 0.20, and the remaining dimensions at 0.13 each, ensuring an equitable distribution of weights across the spectrum of energy needs [29, 32].
The MEPI calculates a household’s energy deprivation score as a weighted sum of deprivations, ranging from 0 (no deprivation) to 1 (maximum deprivation). A threshold of 0.33 is employed to classify a household as energy poor, indicating a significant level of energy [30]
 |
Where 
is the household energy deprivation score, 
= 1 if a household is deprived in indicator I and 
= 0 if otherwise. 
is the weight attached to indicator I with 
This study applies the MEPI to the GLSS 7 data, leveraging its comprehensive coverage of household demographics, economic indicators, and energy use patterns to identify energy-poor households in Ghana. The application of MEPI facilitates a nuanced understanding of energy poverty’s multidimensional nature, enabling targeted analyses of its effects on health outcomes and healthcare-seeking behaviours within the Ghanaian context.
Health outcomes
The GLSS 7 provides a suite of variables acting as proxies for health outcomes, crucial for analyzing the effect of energy poverty on health status and healthcare access. Key among these is the incidence of illness or injury, identified through binary indicators that denote whether a household member has experienced health issues within a recall period, offering a direct snapshot of household health status. The duration of these health issues is also recorded, highlighting the severity and potential economic repercussions, such as lost workdays and the necessity for extended medical treatment. Health Seeking Behaviour variables further enrich this analysis by indicating whether medical advice was sought, the healthcare provider type, and the care location, illuminating patterns in healthcare-seeking behavior and service accessibility.
Control variables
Our analysis examines the interplay between energy poverty and health outcomes, integrating a set of control variables chosen for their theoretical relevance and empirical significance, as highlighted in the literature [16, 26]. These variables are instrumental in accounting for a broad spectrum of influences on health outcomes that operate independently of energy poverty, thereby ensuring the integrity of our findings.
This study used a range of demographic and socioeconomic control variables, including the age of the household head, household size, sex of household head and urban residence. In addition, categorical variables were included for the level of education of the household head and marital status. Crucially, we also controlled for health insurance status, captured as a binary indicator reflecting whether any household member was covered by health insurance, given its potential influence on healthcare-seeking behavior. A description and measurement of all variables used in the study are provided in supplementary file Table S1.
Econometric models
In our study, we deploy an econometric framework to investigate the effect of energy poverty on two primary outcomes: health outcomes and healthcare-seeking behavior. This framework is designed to accommodate the nuanced dynamics of energy poverty’s effects, while also addressing potential endogeneity issues.
Model 1: determinants of health outcomes
This model aims to assess the effect of energy poverty on various health outcomes within households. Given the multifaceted nature of health outcomes, which can range from self-reported health status to more objective measures such as incidence and duration of illnesses, we employ a suite of econometric techniques tailored to the specific nature of each health outcome variable.
For binary health outcomes (the incidence of illness or injury), we employ a logistic regression model. To provide more intuitive and directly interpretable results, marginal effects are computed instead of odds ratios. Marginal effects indicate the change in the probability of the outcome occurring 
for a one-unit change in the independent variable, holding all other explanatory variables constant [33].
The logistic regression model for household head 
is given by:
 |
1 |
Where Yi is the binary outcome variable indicating the occurrence of a health event (e.g., illness, injury) for individual i, with 1 indicating the presence of the event and 0 otherwise. EPi represents the energy poverty status of the household, measured by the Multidimensional Energy Poverty Index (MEPI) or an equivalent continuous measure. Xi is a vector of individual and household characteristics (age, sex, marital status, household size). Zi includes location-specific variables and other relevant covariates such as urban/rural residence that could influence health outcomes. α, β, γ, δ are the parameters to be estimated. The marginal effect of energy poverty 
on the probability of the health event 
is derived as:
 |
2 |
This calculation reflects the change in the likelihood of experiencing a health event due to a one-unit change in energy poverty status, providing policy-relevant insights on probabilities rather than abstract odds.
For count data which measures the number of healthcare visits, a Poisson regression model or a negative binomial regression (if overdispersion is present in the data) framework can be applied [34–36]. The model for individual i for count data is specified as:
 |
3 |
Where, λi is the expected count of the health-related event (number of healthcare visits) for individual i. The model includes energy poverty as the key independent variable, operationalized through the MEPI. Control variables include demographic factors (age, gender, marital status, education) and other relevant covariates identified through literature review.
We conducted a subgroup analysis focusing exclusively on poor households, as defined in the GLSS7 dataset. Households are classified as “poor” when their total annual consumption expenditure per adult equivalent falls below the national poverty line established in GLSS7. Following the Ghana Statistical Service’s methodology, we compare each household’s consumption per adult equivalent to the 2012/13 upper poverty threshold of GH₵1,314.0 per adult equivalent per year; households below this threshold are included in the poor-household subgroup. This approach ensures consistency with the official poverty profile and allows our results to be interpreted in the context of Ghana’s nationally recognized consumption‐based poverty monitoring [37].
Model 2: healthcare seeking behaviour
This model assesses how energy poverty affects the decision to seek medical advice or treatment after an illness or injury. We use the same binary logistic regression framework introduced in Model 1, replacing the health-event outcome with a care-seeking indicator (1 if care was sought, 0 otherwise). MEPI remains our key explanatory variable, and we include the identical set of individual, household, and location controls from Sect. 2.3.1. We report average marginal effects to show how a unit change in energy poverty status alters the probability of seeking healthcare. By reusing the established framework, we streamline our analysis and emphasize the substantive difference between falling ill and actively pursuing medical care.
Results
Descriptive statistics
Table 2 presents the distribution of key categorical variables across energy poverty status. Among the 7,460 observations used in the regression sample, 24.53% of households are classified as energy poor. The incidence of illness or injury is reported for 14.59% of deprived households, compared to 20.57% of non-deprived households. Regarding healthcare utilization, 10.55% of deprived households consulted a health practitioner, compared to 12.01% of non-deprived households.
Table 2.
Summary statistics for categorical variables
| Energy Poverty Status | |||
|---|---|---|---|
| Variables | Deprived | Not Deprived | Overall |
| Energy Poverty | |||
| Deprived | - | - | 1,830 (24.53%) |
| Not Deprived | - | - | 5,630 (75.47%) |
| Health outcomes | |||
| Incidence of illness/injury | |||
| Yes | 267 (14.59%) | 1,158 (20.57%) | 1,425 (19.10%) |
| No | 1,563 (85.41%) | 4,472 (79.43%) | 6,035 (80.90%) |
| Consulted Health Practioner | |||
| Yes | 193 (10.55%) | 676 (12.01%) | 869 (11.65%) |
| No | 1,637 (89.45%) | 4,954 (87.99%) | 6,591 (88.35%) |
| Control Variables | |||
| Gender | |||
| Male | 990 (54.10%) | 2,864 (50.87%) | 3,854 (51.66%) |
| Female | 840 (45.90%) | 2,766 (49.13%) | 3,606 (48.34%) |
| Location | |||
| Urban | 1,497 (81.80%) | 2,212 (39.29%) | 3,709 (49.72%) |
| Rural | 333 (18.20%) | 3,418 (60.71%) | 3,751 (50.28%) |
| Education | |||
| No education | 4 (0.22%) | 14 (0.25%) | 18 (0.24%) |
| Primary | 122 (6.67%) | 1,650 (29.31%) | 1,772 (23.75%) |
| Secondary | 1,008 (55.08%) | 3,553 (63.11%) | 4,561 (61.14%) |
| Post-Secondary | 696 (38.03%) | 413 (7.34%) | 1,109 (14.87%) |
| Marriage | |||
| Married | 711 (38.85%) | 1,620 (28.77%) | 2,331 (31.25%) |
| Not married | 1,119 (61.15%) | 4,010 (71.23%) | 5,129 (68.75%) |
| Health insurance | |||
| Yes | 1,492 (81.53%) | 4,381 (77.82%) | 5,873 (78.73%) |
| No | 338 (18.47%) | 1,249 (22.18%) | 1,587 (21.27%) |
Values represent frequencies with percentages in parentheses. The energy poverty status categories (“Deprived” and “Not Deprived”) correspond to households classified based on the Multidimensional Energy Poverty Index (MEPI). Percentages are calculated within each energy poverty status group, and the “Overall” column includes the aggregated frequencies and percentages for the entire sample of 7,570 households
In terms of demographic and socioeconomic characteristics, 54.10% of household heads in deprived households are male, while the figure is 50.87% among the non-deprived. The proportion of rural households is markedly lower among deprived households (18.20%) than among the non-deprived (60.71%). Educational attainment also varies substantially: 38.03% of deprived household heads have post-secondary education, compared to 7.34% of those in non-deprived households. Marriage rates are higher in deprived households (38.85%) than in non-deprived households (28.77%). Health insurance coverage is high across both groups, with 81.53% of deprived and 77.82% of non-deprived households having at least one insured member.
Table 3 reports summary statistics for continuous variables. The average energy poverty score is 0.715 for deprived households and 0.160 for non-deprived households. Illness duration averages 6.26 days among the deprived and 6.15 days among the non-deprived. The mean age of household heads is slightly lower in deprived households (32.81 years) compared to non-deprived ones (34.65 years). The average household size is 4.15 for deprived households and 2.92 for non-deprived households.
Table 3.
Summary statistics for continuous variables
| Energy Poverty Status | |||
|---|---|---|---|
| Variables | Deprived | Not Deprived | Overall |
| Energy Poverty | |||
| Energy Poverty Score |
Mean = 0.715 SD = 0.155 |
Mean = 0.160 SD = 0.063 |
Mean = 0.579 SD = 0.276 |
| Health outcomes | |||
| Illness Duration |
Mean = 6.26 SD = 4.29 |
Mean = 6.15 SD = 4.16 |
Mean = 6.24 SD = 4.26 |
| Control Variables | |||
| Age |
Mean = 32.81 SD = 17.51 |
Mean = 34.65 SD = 14.54 |
Mean = 33.26 SD = 16.85 |
| Age Squared | Mean = 1383.22, SD = 1458.52 |
Mean = 1411.73 SD = 1237.93 |
Mean = 1390.21 SD = 1407.58 |
| Household Size |
Mean = 4.15 SD = 2.83 |
Mean = 2.92 SD = 1.98 |
Mean = 3.85 SD = 2.70 |
Values represent the mean and standard deviation (SD) of each variable within the categories of energy poverty status (“Deprived” and “Not Deprived”). The “Overall” column includes the aggregated mean and SD for the entire sample of 7,570 households. The Energy Poverty Score reflects the multidimensional deprivation index, where higher scores indicate greater energy poverty. Illness Duration is measured in days. Age is reported in years, and Age Squared is included to capture non-linear effects in the analysis. Household Size represents the total number of members in a household
Effect of energy poverty on health outcomes
The analysis, as presented in Table 4, focuses on the effect of energy poverty on health outcomes across three main dimensions: the incidence of injury/illness, the duration of such illnesses/injuries, and the consultation with health practitioners. Notably, energy poverty significantly influences these health-related outcomes.
Table 4.
Effect of energy poverty on health outcomes
| Model 1 | Model 2 | ||
|---|---|---|---|
| Variables | Incidence of injury/illness | Duration of illness/injuries | Consulted HP |
| Energy poverty | 0.063*** | –1.299* | –0.004 |
| (0.021) | (0.706) | (0.017) | |
| Age | –0.002 | 0.022 | –0.003*** |
| (0.001) | (0.039) | (0.001) | |
| Age2 | 0.000*** | 0.000 | 0.000*** |
| (0.000) | (0.000) | (0.000) | |
| Hhsize | –0.005** | –1.131*** | –0.001 |
| (0.002) | (0.076) | (0.002) | |
| Gender | |||
| Female | 0.011 | –0.295 | 0.027*** |
| (0.009) | (0.318) | (0.007) | |
| Location | |||
| Rural | 0.056*** | 0.510 | 0.017* |
| (0.010) | (0.355) | (0.009) | |
| Education | 0.030 | –0.760 | –0.032 |
| Primary | (0.085) | (2.305) | (0.078) |
| 0.027 | –0.699 | –0.031 | |
| Secondary | (0.085) | (2.285) | (0.078) |
| 0.015 | –0.255 | –0.035 | |
| Post-Secondary | (0.086) | (2.328) | (0.079) |
| Mariage Status | –0.056*** | –1.017*** | –0.016* |
| Married | (0.011) | (0.351) | (0.009) |
| Health Insurance | |||
| Yes | 0.012 | 0.213 | 0.040*** |
| (0.011) | (0.374) | (0.008) | |
| Observations | 7,460 | 1,425 | 7,460 |
Coefficients represent marginal effects derived from regression models. Model 1 examines the probability of experiencing an injury or illness, Model 2 assesses the duration of illnesses or injuries (measured in days), and the final column evaluates the likelihood of consulting a healthcare practitioner. Robust standard errors are presented in parentheses. Statistical significance levels are denoted as ***p < 0.01, **p < 0.05, and *p < 0.1. Positive coefficients indicate an increase in probability or duration, while negative coefficients reflect a decrease. Variables include demographic and socioeconomic controls, with “Energy Poverty” as the primary independent variable. Observations reflect the sample size used in each model
In Model 1, energy poverty is significantly associated with a higher probability of experiencing illness or injury (coefficient = 0.063, SE = 0.021, p < 0.01). In Model 2, energy poverty is significantly associated with a shorter duration of illness or injury (coefficient = − 1.299, SE = 0.706, p < 0.10). In Model 3, the association between energy poverty and consultation with a healthcare practitioner is negative but not statistically significant (coefficient = − 0.004, SE = 0.017).
For control variables, household size is negatively associated with both illness incidence (–0.005, SE = 0.002, p < 0.05) and illness duration (–1.131, SE = 0.076, p < 0.01). Age squared is positively and significantly associated with all three outcomes, while age itself shows negative associations with illness consultation (–0.003, SE = 0.001, p < 0.01). Female household heads are significantly more likely to consult a healthcare practitioner (0.027, SE = 0.007, p < 0.01). Rural residence is significantly associated with increased likelihood of illness (0.056, SE = 0.010, p < 0.01) and healthcare consultation (0.017, SE = 0.009, p < 0.10). Marriage is significantly associated with a lower probability of both illness (–0.056, SE = 0.011, p < 0.01) and consultation (–0.016, SE = 0.009, p < 0.10). Health insurance coverage is significantly associated with increased consultation with a healthcare practitioner (0.040, SE = 0.008, p < 0.01), but not with illness incidence or duration.
Effect of energy poverty on health outcomes for poor households
Table 5 presents the marginal effects of energy poverty on health outcomes for a subsample of 1,333 poor households. In Model 1, energy poverty has a negative but statistically insignificant association with the incidence of illness or injury (coefficient = − 0.005, SE = 0.078). In Model 2, energy poverty is associated with a reduction in illness duration (coefficient = − 1.762, SE = 1.709), but the effect is not statistically significant. In Model 3, energy poverty is significantly associated with a lower probability of consulting a healthcare practitioner (coefficient = − 0.036, SE = 0.020, p < 0.10).
Table 5.
Effect of energy poverty on health outcomes for poor households
| Model 1 | Model 2 | ||
|---|---|---|---|
| Variables | Incidence of injury/illness | Duration of illness/injuries | Consulted HP |
| Energy Poverty | –0.005 | –1.762 | –0.036* |
| (0.078) | (1.709) | (0.020) | |
| Age | –0.000 | –0.105 | –0.002 |
| (0.003) | (0.077) | (0.002) | |
| Age2 | 0.000 | 0.002** | 0.000 |
| (0.000) | (0.001) | (0.000) | |
| Hhsize | –0.007* | –1.287*** | –0.004 |
| (0.004) | (0.127) | (0.003) | |
| Gender | |||
| Female | –0.000 | –0.363 | 0.002 |
| (0.022) | (0.567) | (0.016) | |
| Location | |||
| Rural | 0.046 | 0.686 | –0.004 |
| (0.030) | (0.844) | (0.023) | |
| Education | |||
| Primary | — | – | –0.036* |
| (0.020) | |||
| Secondary | — | 1.105** | — |
| (0.524) | |||
| Post-Secondary | — | –0.082 | — |
| (1.579) | |||
| Mariage Status | |||
| Married | –0.098*** | –0.489 | –0.064 |
| (0.028) | (0.715) | (0.052) | |
| Health Insurance | |||
| Yes | –0.007 | –0.710 | 0.015 |
| (0.028) | (0.667) | (0.018) | |
| Observations | 1,320 | 270 | 1,333 |
Coefficients represent marginal effects derived from regression models. Model 1 examines the probability of experiencing an injury or illness, Model 2 assesses the duration of illnesses or injuries (measured in days), and the final column evaluates the likelihood of consulting a healthcare practitioner. Robust standard errors are provided in parentheses. Statistical significance levels are indicated as ***p < 0.01, **p < 0.05, and *p < 0.1. Positive coefficients indicate an increase in probability or duration, while negative coefficients reflect a decrease. The analysis is restricted to poor households, with “Energy Poverty” as the primary independent variable and controls for demographic and socioeconomic factors. Observations reflect the sample size used in each model
Within this subsample, household size remains significantly negatively associated with both illness incidence (–0.007, SE = 0.004, p < 0.10) and duration (–1.287, SE = 0.127, p < 0.01). Age squared is positively associated with illness duration (0.002, SE = 0.001, p < 0.05), while age itself shows no significant effects. Being married is significantly associated with a lower probability of illness (–0.098, SE = 0.028, p < 0.01). Primary education is negatively associated with healthcare consultation (–0.036, SE = 0.020, p < 0.10), while secondary education is positively associated with illness duration (1.105, SE = 0.524, p < 0.05). No significant association is found between health insurance coverage and any of the three health outcomes among poor households.
Robustness check and extensions
Energy-poverty status may be endogenous in our health-outcome equations for several reasons. First, reverse causality is plausible: households suffering poor health could reduce fuel purchases or forgo appliance use to cover medical expenses, inflating measured energy poverty. Second, omitted-variable bias may arise if unobserved factors, such as household health preferences, caregiver time constraints, or local health-infrastructure quality, simultaneously influence both energy choices and health. Third, measurement error in our MEPI indicator (misreporting fuel use or appliance ownership) can attenuate coefficients or even reverse their sign. Finally, selection bias could occur if the subset of households with complete energy and health data is non-random.
To address these concerns without external instruments, we implement the Lewbel [38] heteroskedasticity-based two-stage least squares (2SLS) estimator, which constructs internal instruments from model residuals under minimal distributional assumptions [38]. This method, which leverages internally generated instruments based on a heteroskedastic covariance restriction, was selected to address potential biases and endogeneity issues, thereby ensuring the reliability of our conclusions. The Lewbel [38] 2SLS framework, noted for its robustness in the absence of external instruments, has been extensively utilized in recent literature [16, 39–41]. The estimates in our analysis are consistent with our earlier findings, affirming the effect of energy poverty on health outcomes in Ghana.
Lewbel two stage least squares
Table 6 presents the Lewbel 2SLS results for the entire sample, showing significant positive relationships between energy poverty and the incidence of injury/illness, and a noteworthy negative relationship on the likelihood of consulting healthcare providers, The analysis is consistent with the baseline regression: while energy poverty correlates with higher incidences of health issues, it also appears to decrease the probability of seeking medical advice.
Table 6.
Results from Lewbel two stage least squares
| Model 1 | Model 2 | ||
|---|---|---|---|
| Variables | Incidence of injury/illness | Duration of illness/injuries | Consulted HP |
| Energy Poverty | 0.181** | −0.740 | −0.137* |
| (0.0894) | (1.655) | (0.0734) | |
| Controls? | Yes | Yes | Yes |
| External IV | No | No | No |
| Internal IV | Yes | Yes | Yes |
| N | 7,502 | 1,438 | 7,500 |
| R-squared | 0.019 | 0.019 | 0.009 |
| F-statistics | 20.20 | 3.03 | 13.27 |
Coefficients represent marginal effects derived using the Lewbel Two-Stage Least Squares (2SLS) estimation method. Model 1 estimates the probability of experiencing an injury or illness, Model 2 examines the duration of illnesses or injuries (measured in days), and the final column assesses the likelihood of consulting a healthcare practitioner. The “Energy Poverty” variable represents the primary independent variable, with internally generated instruments used to address potential endogeneity. Control variables were included in all models, though external instruments were not applied. Robust standard errors are reported in parentheses. Statistical significance levels are denoted as ***p < 0.01, **p < 0.05, and *p < 0.1. Observations represent the sample size for each model, and R-squared values indicate the proportion of variance explained. The F-statistics reflect the strength of the instruments in the first stage. Positive coefficients suggest an increase in the respective outcome, while negative coefficients
Table 7 extends this analysis to poor households, underscoring the differential effect of energy poverty on this vulnerable segment. Results indicate a more moderate but still significant relationship between energy poverty and health outcomes. Notably, the findings suggest that energy poverty among the poor influences health-seeking behaviours differently, pointing to the need for targeted interventions that address the unique challenges faced by these households in accessing healthcare services.
Table 7.
Results from Lewbel two stage least squares (Poor Households)
| Model 1 | Model 2 | ||
|---|---|---|---|
| Variables | Incidence of injury/illness | Duration of illness/injuries | Consulted HP |
| Energy Poverty | 0.0445** | −0.409 | −0.0245 |
| (0.0217) | (0.525) | (0.0183) | |
| Controls? | Yes | Yes | Yes |
| External IV | No | No | No |
| Internal IV | Yes | Yes | Yes |
| Observations | 1,315 | 266 | 1,312 |
| R-squared | 0.025 | 0.020 | 0.018 |
| F-statistics | 20.18 | 3.04 | 12.25 |
Coefficients represent marginal effects derived using the Lewbel Two-Stage Least Squares (2SLS) method, focusing on poor households. Model 1 examines the probability of experiencing an injury or illness, Model 2 evaluates the duration of illnesses or injuries (measured in days), and the final column assesses the likelihood of consulting a healthcare practitioner. The “Energy Poverty” variable is the key independent variable, with internally generated instruments used to mitigate endogeneity concerns. Control variables are included in all models, while no external instruments were applied. Robust standard errors are reported in parentheses. Statistical significance is denoted as ***p < 0.01, **p < 0.05, and *p< 0.1. The R-squared values represent the proportion of variance explained by the models, and F-statistics indicate the strength of the instruments in the first stage. Positive coefficients indicate an increase in the respective outcome, while negative coefficients suggest a reduction
Discussion
This study seeks to examine the relationship between energy poverty and health outcomes in Ghana. Specifically, the study investigated how energy poverty influences the incidence of illness or injury, duration of illness, and healthcare-seeking behavior, with further emphasis on disparities among poor households. The study reveals notable disparities between energy-poor and non-deprived households in Ghana, both in terms of health outcomes and demographic characteristics. Energy-poor households constitute 24.53% of the sample and report a lower incidence of illness or injury (14.59%) compared to non-deprived households (20.57%), as well as slightly lower healthcare consultation rates (10.55% versus 12.01%). Interestingly, energy-poor households are predominantly urban (81.80%), whereas non-deprived households are mostly rural (60.71%). These deprived households also tend to have younger household heads (32.81 years compared to 34.65 years), smaller family sizes (4.15 versus 2.92), and a higher proportion of members with post-secondary education (38.03% versus 7.34%).
Demographic and socio-economic factors further amplify the health inequalities associated with energy poverty. Rural households, which face higher rates of energy deprivation, exhibit greater illness incidence and lower healthcare utilization, suggesting that geographical location plays a critical role in access to energy and health services. Similarly, education level strongly influences health outcomes and care-seeking behavior, with lower attainment linked to adverse effects. These findings underscore the complex and multidimensional nature of energy poverty, emphasizing its central role in shaping population health and reinforcing structural inequities that hinder the achievement of universal healthcare and inclusive energy access.
The findings of this study contribute to the discourse on the relationship between energy poverty and health outcomes in LMICs, building upon existing literature while offering new insights. Energy poverty’s influence on health-seeking behavior and health outcomes has been extensively documented, but our results introduce additional layers of complexity. For instance, Okorie et al. [42] and Irwin et al. [43] highlight that energy poverty exacerbates health risks and diminishes healthcare utilization due to financial and logistical barriers. Our findings echo these patterns, showing that while energy-deprived households in Ghana exhibit heightened illness risks, their healthcare-seeking behaviors are often constrained by structural inequities such as healthcare access and affordability.
Existing research has emphasized the role of energy poverty in exacerbating physical injuries and respiratory illnesses, with studies by Khogali et al. [19] and Ballesteros-Arjona et al. [44] focusing on mental and respiratory health risks in LMICs and OECD countries, respectively. While our findings align with this, we add critical geographical context, showing how rural Ghanaian households are disproportionately impacted by energy poverty, compounding health disparities due to limited access to clean energy and healthcare infrastructure.
Gender dimensions of energy poverty are also well-documented, with Oliveras et al. [45] and Zhang et al. [21] underscoring its disproportionate effects on women’s mental health and healthcare expenditures. In line with these studies, our research finds that female-headed households in Ghana demonstrate higher engagement in healthcare-seeking behaviors, suggesting that women often shoulder the responsibility of health management in energy-deprived contexts.
Economic consequences of energy poverty are explored in studies such as Nguyen-Phung and Le [22], which reveal reduced health expenditure in Vietnam due to financial strain. Similarly, Abbas et al. [46] argue that financial insecurity in energy-deprived households influences health behaviors. Our findings corroborate these economic dimensions, highlighting how energy-poor households in Ghana often underreport illness durations, potentially due to untreated or undiagnosed conditions caused by financial constraints.
The importance of targeted policy interventions to mitigate the health effects of energy poverty has been proposed by Middlemiss et al. [47] and Pan et al. [48]. Our findings reinforce this need, emphasizing the necessity for interventions that not only enhance energy access but also address structural healthcare barriers, especially in rural and low-income settings. Additionally, the nonlinear dynamics of energy poverty and health outcomes explored by Kose [49] and Pondie et al. [50] resonate with our findings, which demonstrate that the impact of energy poverty varies significantly by demographic and geographic context.
Respiratory health risks highlighted in studies like Stevens et al. [18] are also supported by our findings, but we expand this to a broader spectrum of outcomes, including injury incidence and healthcare utilization. Furthermore, delayed care-seeking behaviors identified by Carrere et al. [51] and Clarke-Deelder et al. [52] are contextualized in our research, showing how socioeconomic stratification amplifies barriers to healthcare access in energy-deprived communities.
Policy recommendations
Addressing the intertwined challenges of energy poverty and public health in Ghana necessitates comprehensive policy interventions that span multiple sectors. Firstly, enhancing access to clean and affordable energy, especially in rural and underserved communities, is crucial. Policies should focus on subsidizing cleaner energy technologies and improving infrastructure to mitigate the health implications of energy poverty. Simultaneously, an integrated approach that combines energy policy with public health strategies is essential. This could include initiatives that cater to both the energy needs and health vulnerabilities of the populace, such as promoting clean cooking technologies to curb indoor air pollution.
Moreover, vulnerable, and impoverished households require targeted support to overcome the disproportionate effect of energy poverty. Strategies might encompass targeted subsidies for energy expenses, the provision of energy-efficient appliances, and improved access to healthcare services to alleviate the adverse health effects of energy scarcity. Additionally, investing in public health education is vital. Community awareness campaigns about the health risks of energy poverty and the benefits of clean energy adoption can foster informed decision-making. These campaigns should also emphasize the importance of seeking medical treatment for energy-related health issues, ensuring that public health education is a cornerstone of the broader strategy to combat energy poverty. Together, these measures can significantly contribute to improving health outcomes and advancing sustainable development in Ghana.
Limitation of the study
This study has some limitations. First, the reliance on self-reported data for illness incidence and healthcare utilization may introduce recall bias and underreporting, particularly in energy-deprived households. Second, the analysis focuses on Ghana, which may limit the generalizability of findings to other low- and middle-income countries with varying energy and healthcare landscapes. Additionally, unmeasured confounders such as cultural practices and informal healthcare use could influence the results.
Conclusion
This study highlights the significant relationship between energy poverty and health outcomes, emphasizing its disproportionate effect on vulnerable households in Ghana. The findings reveal that energy-deprived households experience higher rates of illness and injury, shorter illness durations potentially reflecting untreated conditions, and logistical barriers that limit healthcare utilization. While the study underscores the need for targeted policy interventions to mitigate the health effects of energy poverty, it recognizes the complexities of addressing structural inequities in energy access and healthcare availability. Future research should expand into broader contexts and longitudinal frameworks to validate these findings and inform sustainable, equitable energy and health policies.
Supplementary Information
Acknowledgements
Not applicable.
Abbreviations
- 2SLS
Two-Stage Least Squares
- EPA
Energy Poverty Alleviation
- GIMPA
Ghana Institute of Management and Public Administration
- GLSS
Ghana Living Standards Survey
- HP
Health Practitioner
- IV
Instrumental Variable
- LPG
Liquefied Petroleum Gas
- MEPI
Multidimensional Energy Poverty Index
- N
Number of observations
- NIHR
National Institute for Health and Care Research
- OPHI
Oxford Poverty and Human Development Initiative
- SDG
Sustainable Development Goal
- STATA
Statistical Analysis Software
Authors’ contributions
J.K.A.D, M.A.A, K.N.B., and W.O. collaboratively conceptualized the study. M.A.A. developed the methodology and, along with W.O., performed the formal analysis. M.A.A. took the lead in writing the manuscript, with J.K.A.D., K.N.B., and W.O. contributing to the review and editing process. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Data availability
The data was collected by Ghana Statistical Service and the authors, thus, do not have the permission to share the data. However, access to the data can be requested from Ghana Statistical Service at Microdata Home (statsghana.gov.gh).
Declarations
Ethics approval and consent to participate
Ethical approval for this study was not required, as the analysis is based on fully anonymized secondary data obtained from the Ghana Statistical Service (GSS), specifically the Ghana Living Standards Survey Round 7. According to the Ghana Microdata Dissemination Policy, only anonymized Public Use Files (PUFs) are released to researchers, with all direct and indirect personal identifiers removed to protect participant confidentiality. Under this policy, the use of such anonymized datasets for research purposes does not require additional Institutional Review Board (IRB) approval [53]. Permission to use the data was formally granted by GSS. All analyses in this study were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Raghutla C, Chittedi KR. Energy poverty and economic development: evidence from BRICS economies. Environ Sci Pollut Res. 2022;29(7):9707–21. [DOI] [PubMed] [Google Scholar]
- 2.United Nations. Sustainable Development: The 17 Goals. 2020. Available from: https://sdgs.un.org/goals. Cited 2024 Apr 1.
- 3.United Nations Development Program. Policy Brief. 2018. Available from: https://sustainabledevelopment.un.org/content/documents/17480PB8.pdf. Cited 2024 Apr 1.
- 4.International Energy Agency. Energy Access Outlook: From Poverty to Prosperity. 2017. Available from: https://scholar.google.com/scholar_lookup?title=Energy%20Access%20Outlook%3A%20From%20Poverty%20to%20Prosperity&author=International%20Energy%20Agency&publication_year=2017. Cited 2024 Apr 1.
- 5.Liu H, Hu T. Energy poverty alleviation and its implications for household energy consumption and health. Environ Dev Sustain. 2023;26(4):10063–83. [Google Scholar]
- 6.Katoch OR, Sharma R, Parihar S, Nawaz A. Energy poverty and its impacts on health and education: a systematic review. Int J Energy Sect Manage. 2024;18(2):411–31. [Google Scholar]
- 7.Brown H, Vera-Toscano E. Energy poverty and its relationship with health: empirical evidence on the dynamics of energy poverty and poor health in Australia. SN Bus Econ. 2021;1(10):139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jessel S, Sawyer S, Hernández D. Energy, poverty, and health in climate change: A comprehensive review of an emerging literature. Front Public Health. 2019;7. [DOI] [PMC free article] [PubMed]
- 9.World Bank. Access to electricity (% of population) - Ghana. 2023. Available from: https://data.worldbank.org/indicator/EG.ELC.ACCS.ZS?locations=GH. Cited 2024 Apr 1.
- 10.Kipkoech R, Takase M, Ahogle AMA, Ocholla G. Opportunities and challenges in ghana’s renewable energy sector. Discover Appl Sci. 2024;6(10):530. [Google Scholar]
- 11.Bukari D, Tuokuu FXD, Suleman S, Ackah I, Apenu G. Ghana’s energy access journey so far: a review of key strategies. Int J Energy Sect Manage. 2021;15(1):139–56. [Google Scholar]
- 12.Inkoom D, Crentsil AO. Estimation of Indoor Air Pollution and Health Impacts Due to Biomass Burning in Rural Northern Ghana. 2015.
- 13.Terfa ZG, Ahmed S, Khan J, Niessen LW. Household microenvironment and Under-Fives health outcomes in uganda: focusing on multidimensional energy poverty and women empowerment indices. Int J Environ Res Public Health. 2022;19(11):6684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shupler M, Baame M, Nix E, Tawiah T, Lorenzetti F, Saah J, et al. Multiple aspects of energy poverty are associated with lower mental health-related quality of life: A modelling study in three peri-urban African communities. SSM - Mental Health. 2022;2:100103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lin B, Okyere MA. Does energy poverty affect the well-being of people: evidence from Ghana. Sustain Prod Consum. 2021;28:675–85. [Google Scholar]
- 16.Bukari C, Broermann S, Okai D. Energy poverty and health expenditure: evidence from Ghana. Energy Econ. 2021;103:105565. [Google Scholar]
- 17.Famewo AS, Uwala VA. Socio-economic impacts of rural energy poverty on women and students in Esa-Oke, Nigeria. J Sustain Environ Manage. 2022;1(2):84–93. [Google Scholar]
- 18.Stevens M, Yang-Huang J, Nieboer D, Zhou S, Osman KA, Raat H, et al. Multidimensional energy poverty and acute respiratory infection in children under 5 years of age: evidence from 22 low-income and middle-income countries. J Epidemiol Community Health (1978). 2023;77(11):687–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Khogali A, Ahmed A, Ibrahim M, Karrar K, Elsheikh M, Abdelraheem E, et al. Building power-ful health systems: the impacts of electrification on health outcomes in LMICs. Psychol Health Med. 2022;27(sup1):124–37. [DOI] [PubMed] [Google Scholar]
- 20.Lin B, Okyere MA. Multidimensional energy poverty and mental health: Micro-Level evidence from Ghana. Int J Environ Res Public Health. 2020;17(18):6726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Zhang Z, Linghu Y, Meng X, Yi H. Is there gender inequality in the impacts of energy poverty on health? Front Public Health. 2022;10. [DOI] [PMC free article] [PubMed]
- 22.Nguyen-Phung HT, Le H. Energy poverty and health expenditure: empirical evidence from Vietnam. Soc Sci. 2024;13(5):253. [Google Scholar]
- 23.Lee CC, Yuan Z. Impact of energy poverty on public health: A non-linear study from an international perspective. World Dev. 2024;174:106444. [Google Scholar]
- 24.Fan Y, Döring T, Li S, Zhang X, Fang M, Yu Y. Energy poverty and public health vulnerability: A multi-country analysis. Sustain Dev. 2024;32(5):5161–80. [Google Scholar]
- 25.Wang Y, Wu Y, Wang C, Li L, Lei Y, Wu S et al. The impact of energy poverty on the health and welfare of the middle-aged and older adults. Front Public Health. 2024;12. [DOI] [PMC free article] [PubMed]
- 26.Awaworyi Churchill S, Smyth R. Ethnic diversity, energy poverty and the mediating role of trust: evidence from household panel data for Australia. Energy Econ. 2020;86:104663. [Google Scholar]
- 27.Boardman B. Fixing fuel poverty. Routledge; 2013.
- 28.Bouzarovski S, Petrova S. A global perspective on domestic energy deprivation: overcoming the energy poverty–fuel poverty binary. Energy Res Soc Sci. 2015;10:31–40. [Google Scholar]
- 29.Adusah-Poku F, Takeuchi K. Household energy expenditure in ghana: A double-hurdle model approach. World Dev. 2019;117:266–77. [Google Scholar]
- 30.Nussbaumer P, Bazilian M, Modi V. Measuring energy poverty: focusing on what matters. Renew Sustain Energy Rev. 2012;16(1):231–43. [Google Scholar]
- 31.Crentsil AO, Asuman D, Fenny AP. Assessing the determinants and drivers of multidimensional energy poverty in Ghana. Energy Policy. 2019;133:110884. [Google Scholar]
- 32.Alkire S, Foster J. Counting and multidimensional poverty measurement. J Public Econ. 2011;95(7–8):476–87. [Google Scholar]
- 33.Onukwugha E, Bergtold J, Jain R. A primer on marginal Effects—Part I: theory and formulae. PharmacoEconomics. 2015;33(1):25–30. [DOI] [PubMed] [Google Scholar]
- 34.Gebeyehu FG, Geremew BM, Belew AK, Zemene MA. Number of antenatal care visits and associated factors among reproductive age women in Sub-Saharan Africa using recent demographic and health survey data from 2008–2019: A multilevel negative binomial regression model. PLOS Global Public Health. 2022;2(12):e0001180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Abu Bakar NS, Ab Hamid J, Mohd Nor Sham MSJ, Sham MN, Jailani AS. Count data models for outpatient health services utilisation. BMC Med Res Methodol. 2022;22(1):261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Le DD, Gonzalez RL, Matola JU. Modeling count data for health care utilization: an empirical study of outpatient visits among Vietnamese older people. BMC Med Inf Decis Mak. 2021;21(1):265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ghana Statistical Service. Poverty Trends in Ghana. 2018.
- 38.Lewbel A. Using heteroscedasticity to identify and estimate mismeasured and endogenous regressor models. J Bus Economic Stat. 2012;30(1):67–80. [Google Scholar]
- 39.Churchill SA, Marisetty VB. Financial inclusion and poverty: a Tale of forty-five thousand households. Appl Econ. 2020;52(16):1777–88. [Google Scholar]
- 40.Koomson I, Villano R, Hadley D. Effect of financial inclusion on poverty and vulnerability to poverty: evidence using a Multi-Dimensional measure of financial inclusion. SSRN Electron J. 2020.
- 41.Zhang Q, Posso A. Microfinance and gender inequality: cross-country evidence. Appl Econ Lett. 2017;24(20):1494–8. [Google Scholar]
- 42.Okorie DI, Lin B. Association of energy poverty and catastrophic health expenditure. Energy. 2022;253:124108. [Google Scholar]
- 43.Irwin BR, Hoxha K, Grépin KA. Conceptualising the effect of access to electricity on health in low- and middle-income countries: A systematic review. Glob Public Health. 2020;15(3):452–73. [DOI] [PubMed] [Google Scholar]
- 44.Ballesteros-Arjona V, Oliveras L, Bolívar Muñoz J, Olry de Labry Lima A, Carrere J, Martín Ruiz E, et al. What are the effects of energy poverty and interventions to ameliorate it on people’s health and well-being? A scoping review with an equity lens. Energy Res Soc Sci. 2022;87:102456. [Google Scholar]
- 45.Oliveras L, Artazcoz L, Borrell C, Palència L, López MJ, Gotsens M, et al. The association of energy poverty with health, health care utilisation and medication use in Southern Europe. SSM Popul Health. 2020;12:100665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Abbas K, Xie X, Xu D, Butt KM. Assessing an empirical relationship between energy poverty and domestic health issues: A multidimensional approach. Energy. 2021;221:119774. [Google Scholar]
- 47.Middlemiss L, Stevens M, Ambrosio-Albalá P, Pellicer-Sifres V, van Grieken A. How do interventions for energy poverty and health work? Energy Policy. 2023;180:113684. [Google Scholar]
- 48.Pan L, Biru A, Lettu S. Energy poverty and public health: global evidence. Energy Econ. 2021;101:105423. [Google Scholar]
- 49.Kose T. Energy poverty and health: the Turkish case. Energy Sources Part B: Econ Plann Policy. 2019;14(5):201–13. [Google Scholar]
- 50.Messie Pondie T, Engwali FD, Ongo Nkoa BE, Noubissi Domguia E. Energy poverty and respiratory health in Sub-Saharan africa: effects and transmission channels. Energy. 2024;297:131158. [Google Scholar]
- 51.Carrere J, Peralta A, Oliveras L, López MJ, Marí-Dell’Olmo M, Benach J, et al. Energy poverty, its intensity and health in vulnerable populations in a Southern European City. Gac Sanit. 2021;35(5):438–44. [DOI] [PubMed] [Google Scholar]
- 52.Clarke-Deelder E, Afriyie DO, Nseluke M, Masiye F, Fink G. Health care seeking in modern urban LMIC settings: evidence from lusaka, Zambia. BMC Public Health. 2022;22(1):1205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Ghana Statistical Service. Ghana Statistical Service Microdata Dissemination Policy. 2016; https://www2.statsghana.gov.gh/docfiles/Microdata_Policy/Ghana%20Microdata%20Dissemination%20Policy.pdf?.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data was collected by Ghana Statistical Service and the authors, thus, do not have the permission to share the data. However, access to the data can be requested from Ghana Statistical Service at Microdata Home (statsghana.gov.gh).