Structural Inequality in Clean Fuel Access and Acute Respiratory Infection Among Nigerian Children: An Intersectional Multilevel Analysis

This study of nearly 29,000 Nigerian children reveals that while polluting cooking fuels are initially linked to acute respiratory infections, the risk is predominantly shaped by intersecting structural and geographic inequalities rather than fuel type alone, necessitating equity-focused subnational policies to effectively reduce the disease burden.

The Gist

🏠 The Big Picture: Why Do Nigerian Children Get Sick?

Imagine a child's lungs as a delicate garden. Sometimes, this garden gets sick (Acute Respiratory Infection, or ARI). For a long time, scientists have pointed at one specific weed in the garden: dirty cooking fuel (like wood, charcoal, or kerosene). They thought, "If we just pull out the dirty fuel and plant clean gas, the garden will be healthy."

But this new study, looking at data from nearly 29,000 children in Nigeria, suggests the problem is much more complicated. It's not just about the weed; it's about the entire soil, the weather, and the fence around the garden.

The researchers found that while dirty fuel is bad, the real reason children get sick is a mix of where they live, how much money their family has, and their age. If you only fix the fuel but ignore the rest, the garden might still struggle.

---

🔍 How They Did the Study (The Detective Work)

The researchers acted like super-detectives using a massive map of Nigeria (the 2024 Demographic and Health Survey).

1. The Standard Detective Work: First, they looked at individual clues. They asked: "Does using wood for cooking make kids sick?"
   • Initial finding: Yes, it looked like it did.
2. The "Intersectional" Detective Work (The Magic Lens): Then, they used a special tool called MAIHDA. Think of this as a 3D kaleidoscope. Instead of looking at fuel, money, and location separately, they spun the kaleidoscope to see what happens when you combine them all at once.
   • Example: They didn't just look at "Poor people." They looked at "Poor people living in rural villages in the North East who use wood."
   • This helped them see that these specific combinations create "hotspots" of sickness that you can't see if you just look at one factor alone.

---

📉 The Big Surprises (What They Found)

1. The "Fuel Myth" (It's Not Just the Smoke)

When they looked at the data without adjusting for other factors, it seemed like dirty fuel was the main villain.

• The Analogy: Imagine you see a car crash and blame the red paint on the car. But when you look closer, you realize the red cars were mostly driven by people speeding in the rain on icy roads. The ice and speed were the real problem, not the paint.
• The Reality: Once the researchers accounted for poverty, location, and education, the link between dirty fuel and sickness disappeared. The fuel wasn't the main cause; it was just a sign that the family was living in a difficult situation (poor, rural, remote). The conditions of poverty were the real killers.

2. The "Age Shield" (The Strongest Factor)

The most consistent finding was about age.

• The Analogy: Think of a baby's immune system as a new, untested umbrella. It's small and fragile. As a child grows, their immune system becomes a big, sturdy umbrella that can handle the rain.
• The Reality: Babies (under 1 year) were the most likely to get sick. Children aged 2 to 5 years were much healthier. Their bodies had built up defenses, just like that sturdy umbrella.

3. The "Geography Map" (Where You Live Matters Most)

Where a child lived in Nigeria was the biggest predictor of sickness.

• The Analogy: Imagine Nigeria is a giant pizza. Some slices are "safe zones" (like the South West), and some slices are "danger zones" (like the North East).
• The Reality: Children in the North East were much more likely to be sick, while those in the South West were much less likely. This wasn't just about fuel; it was about the whole environment—healthcare access, climate, and local infrastructure.

4. The "Rich Kid Paradox"

Here is a weird twist: In the adjusted models, children from the richest families actually had higher odds of being reported as sick than the poorest.

• The Analogy: Imagine a rich person and a poor person both have a headache. The rich person goes to a fancy doctor who diagnoses it immediately. The poor person stays home and ignores it.
• The Reality: Wealthier families likely have better access to doctors and are more likely to report symptoms. Poorer families might not seek help or report the illness, so the data makes it look like they are "healthier," even if they are suffering.

---

💡 What Does This Mean for the Future?

The study concludes that you can't fix child sickness in Nigeria just by handing out clean gas stoves.

• The Old Way: "Here is a clean stove. You are now safe."
• The New Way (The Study's Advice): "We need to fix the whole neighborhood."

To truly protect children, we need intersectional policies. This means:

1. Targeting the youngest: Focus extra protection on babies under 1 year old (their "umbrellas" are too small).
2. Fixing the "Danger Zones": Pour resources into the North East and other high-risk areas to improve healthcare and living conditions.
3. Tackling Poverty: Since poverty is the root cause that links dirty fuel, bad housing, and lack of healthcare, fighting poverty is the best way to fight sickness.

🏁 The Bottom Line

Think of a child's health as a house. Dirty fuel is a leaky window. But if the roof is caving in (poverty), the walls are crumbling (bad healthcare access), and the foundation is shaky (geographic inequality), fixing just one window won't save the house. You have to rebuild the whole structure to keep the children safe.

Technical Summary

1. Problem Statement

Acute Respiratory Infections (ARI) remain a leading cause of morbidity and mortality for children under five in Nigeria, which bears the highest absolute number of child pneumonia deaths globally. While household air pollution (HAP) from polluting cooking fuels (solid biomass, kerosene) is widely recognized as a risk factor, existing literature often treats determinants like fuel type, poverty, and geography as independent variables. This approach fails to capture how these factors intersect to create concentrated disadvantage. Furthermore, robust evidence linking HAP to child health in Sub-Saharan Africa is limited, and standard regression models struggle to analyze the complex, sparse data generated by multiple intersecting social categories.

2. Methodology

The study utilized a secondary analysis of the 2024 Nigeria Demographic and Health Survey (NDHS), a nationally representative cross-sectional dataset.

• Sample: 28,728 children aged 0–59 months.
• Outcome Measures:
  • Strict ARI: Maternal report of cough, rapid/difficult breathing, and fever within the last two weeks (Prevalence: 1.9%).
  • Broad Respiratory Symptoms: Cough combined with either rapid breathing or fever (Prevalence: 8.3%).
• Primary Exposure: Household cooking fuel type (Clean vs. Polluting: solid biomass, kerosene/other petroleum).
• Covariates: Child age/sex, household wealth quintile, residence (urban/rural), geopolitical zone, and parental education.
• Analytical Approaches:
  1. Survey-Weighted Quasibinomial Logistic Regression: To estimate associations between fuel type and ARI, adjusting for confounders. Population-attributable fractions (PAF) were calculated.
  2. Intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA): An innovative approach to address sparse data in interaction terms.
     • Strata Construction: Created 164 empirically observed intersectional strata based on combinations of fuel type (4), wealth (5), residence (2), and geopolitical zone (6).
     • Modeling: Generalized linear mixed models with random intercepts for strata were used to quantify between-strata variance (Intraclass Correlation Coefficient - ICC) and assess whether child age/sex explained this clustering.

3. Key Contributions

• Intersectional Framework: This is the first study in Nigeria to apply MAIHDA to child ARI, moving beyond additive models to examine how the combination of energy poverty, wealth, and geography creates unique risk profiles.
• Deconstruction of Fuel-ARI Link: The study provides evidence that the apparent link between polluting fuels and ARI is largely a proxy for broader structural inequalities (geography and wealth) rather than a direct causal effect of fuel exposure alone.
• Outcome Definition Sensitivity: It highlights that the association between fuel and health varies significantly depending on whether "strict" (febrile) or "broad" (symptomatic) definitions of ARI are used.
• Methodological Rigor: It demonstrates the utility of MAIHDA in handling sparse data and low-prevalence outcomes where traditional interaction terms fail.

4. Key Results

Regression Analysis

• Fuel Association: In unadjusted models, polluting fuel use was associated with higher odds of broad respiratory symptoms (OR 1.85, 95% CI 1.43–2.39). However, after adjustment for structural factors (wealth, zone, residence), this association became non-significant (aOR 1.17, p=0.4).
• Structural Determinants:
  • Geopolitical Zone: The strongest predictor. North West and South West zones had significantly lower odds of ARI compared to North Central, while North East showed elevated risks for broad symptoms.
  • Child Age: The most consistent individual predictor. Children aged 24–59 months had roughly half the odds of strict ARI compared to infants <12 months (aOR 0.53).
  • Wealth Paradox: In adjusted models, children in the Richest quintile had higher odds of strict ARI than the poorest (aOR 1.86), likely due to negative confounding by zone or better healthcare-seeking behavior (surveillance bias).
• Population-Attributable Fraction (PAF): The PAF for polluting fuel dropped dramatically from 41.4% (unadjusted) to 12.4% (adjusted), indicating that the majority of the fuel-ARI burden is attributable to the structural conditions determining fuel access, not the fuel itself.

MAIHDA Analysis

• Variance Clustering: The Intraclass Correlation Coefficient (ICC) was 0.09, meaning 9% of the total variance in ARI risk was attributable to differences between intersectional strata.
• Independence from Demographics: Adding child age and sex to the model did not reduce the between-strata variance (ICC remained 0.09). This confirms that the clustering of risk is driven by the intersection of social positions (fuel, wealth, zone, residence) rather than just the demographic composition of those groups.
• Strata Specificity: The analysis identified specific combinations of disadvantage (e.g., solid fuel + poorest wealth + rural + specific zones) that carried amplified risk beyond the sum of individual factors.

5. Significance and Implications

• Policy Shift: The findings suggest that interventions focusing solely on transitioning households to clean fuels may be insufficient if they do not address the underlying structural inequalities (regional disparities, poverty, and healthcare access) that drive both fuel choice and disease risk.
• Equity-Focused Interventions: Policies must be subnational and intersectional, targeting specific "hotspots" of compounded disadvantage (e.g., poor rural households in the North East) rather than applying uniform national fuel subsidies.
• Surveillance Bias: The study highlights how wealth and education can paradoxically appear as risk factors in cross-sectional data due to increased healthcare utilization and symptom reporting, necessitating careful interpretation of socioeconomic gradients.
• Methodological Advancement: The successful application of MAIHDA in this context validates its use as a "gold standard" for investigating health disparities in social epidemiology, particularly in low-resource settings with complex social stratification.

Conclusion: The study concludes that ARI risk among Nigerian children is shaped more by structural and geographic inequalities than by household fuel use alone. Reducing the ARI burden requires equity-focused policies that address intersecting socioeconomic and regional disadvantages.