Housing Improvements and Malaria Risk in Sub-Saharan Africa: A Multi-Country Analysis of Survey Data

Abstract

Background: Improvements to housing may contribute to malaria control and elimination by reducing house entry by malaria vectors and thus exposure to biting. We tested the hypothesis that the odds of malaria infection are lower in modern, improved housing compared to traditional housing in sub-Saharan Africa (SSA).

Methods and findings: We analysed 15 Demographic and Health Surveys (DHS) and 14 Malaria Indicator Surveys (MIS) conducted in 21 countries in SSA between 2008 and 2015 that measured malaria infection by microscopy or rapid diagnostic test (RDT). DHS/MIS surveys record whether houses are built with finished materials (e.g., metal) or rudimentary materials (e.g., thatch). This information was used to develop a binary housing quality variable where houses built using finished wall, roof, and floor materials were classified as "modern", and all other houses were classified as "traditional". Conditional logistic regression was used to determine the association between housing quality and prevalence of malaria infection in children aged 0-5 y, adjusting for age, gender, insecticide-treated net (ITN) use, indoor residual spraying, household wealth, and geographic cluster. Individual survey odds ratios (ORs) were combined to determine a summary OR using a random effects meta-analysis. Of 284,532 total children surveyed, 139,318 were tested for malaria infection using microscopy (n = 131,652) or RDT (n = 138,540). Within individual surveys, malaria prevalence measured by microscopy ranged from 0.4% (Madagascar 2011) to 45.5% (Burkina Faso 2010) among children living in modern houses and from 0.4% (The Gambia 2013) to 70.6% (Burkina Faso 2010) in traditional houses, and malaria prevalence measured by RDT ranged from 0.3% (Senegal 2013-2014) to 61.2% (Burkina Faso 2010) in modern houses and from 1.5% (The Gambia 2013) to 79.8% (Burkina Faso 2010) in traditional houses. Across all surveys, modern housing was associated with a 9% to 14% reduction in the odds of malaria infection (microscopy: adjusted OR 0.91, 95% CI 0.85-0.97, p = 0.003; RDT: adjusted OR 0.86, 95% CI 0.80-0.92, p < 0.001). This association was consistent regardless of ITN usage. As a comparison, the odds of malaria infection were 15% to 16% lower among ITN users versus non-users (microscopy: adjusted OR 0.84, 95% CI 0.79-0.90, p < 0.001; RDT: adjusted OR 0.85, 95% CI 0.80-0.90, p < 0.001). The main limitation of this study is that residual confounding by household wealth of the observed association between housing quality and malaria prevalence is possible, since the wealth index may not have fully captured differences in socioeconomic position; however, the use of multiple national surveys offers the advantage of a large sample size and the elimination of many biases typically associated with pooling observational data.

Conclusions: Housing quality is an important risk factor for malaria infection across the spectrum of malaria endemicity in SSA, with a strength of association between housing quality and malaria similar to that observed between ITN use and malaria. Improved housing should be considered a promising intervention for malaria control and elimination and long-term prevention of reintroduction.

Fig 1. Reduction in the odds of malaria infection in children aged 0–5 y living in modern houses in sub-Saharan Africa.

Values to the left of the vertical line representing the null value indicate a reduction in the odds of malaria infection in modern housing compared to traditional housing. Data are taken from 15 Demographic and Health Surveys and 14 Malaria Indicator Surveys conducted between 2008 and 2015. Houses built with a finished wall, finished roof, and finished floor material were classified as modern, and all other houses were classified as traditional (S2 Appendix). ORs are adjusted for age, gender, insecticide-treated net use, indoor residual spraying in the past 12 mo (where measured), household wealth, and geographic cluster. Summary effects are from random effects analysis. Sub-groups show diagnostic test. Error bars show 95% confidence intervals. DRC, Democratic Republic of the Congo; OR, odds ratio; RDT, rapid diagnostic test.

Fig 2. Reduction in the odds of malaria infection in children aged 0–5 y sleeping under insecticide-treated nets in sub-Saharan Africa.

Values to the left of the vertical line representing the null value indicate a reduction in the odds of malaria infection in users of insecticide-treated nets compared to non-users. Data are taken from 15 Demographic and Health Surveys and 14 Malaria Indicator Surveys conducted between 2008 and 2015. ORs are adjusted for age, gender, indoor residual spraying in the past 12 mo (where measured), household wealth, house type, and geographic cluster. Summary effects are from random effects analysis. Sub-groups show diagnostic test. Error bars show 95% confidence intervals. DRC, Democratic Republic of the Congo; OR, odds ratio; RDT, rapid diagnostic test.