Modelling the impact of climatic and environmental variables on malaria incidence in Tanzania: Implications for achieving the WHO's 2030 Targets

Abstract

Malaria remains a significant public health challenge, particularly among vulnerable populations in high-burden countries like Tanzania. Despite progress in reducing malaria incidence, climatic and environmental condition variability has led to uneven reductions, hindering the achievement of the WHO 2030 targets. We assessed the impact of climatic and environmental variables on malaria incidence to better understand spatial and temporal trends and their implications for the WHO targets. We utilized geo-covariate data from the Demographic and Health surveys (DHS) program, applying a Moran's I test for spatial autocorrelation, a geostatistical Bayesian-based model to predict malaria incidence at an unsampled locations, and calculated the percentage change in predicted incidence over a ten-year interval. The results showed that malaria incidence decreased with greater variance across Tanzania. Mean malaria incidence decreased from 0.347 (95% CI: 0.336, 0.357) in 2000 to 0.118 (95% CI: 0.114, 0.122) in 2020, relative to the increasing insecticide-treated bednets (ITNs) coverage (0.037; 95% CI: 0.036, 0.039 in 2000 to 0.496; 95% CI: 0.476, 0.517 in 2020). Malaria incidence was higher in the Lake, western, eastern and southern zones compared to others, with spatial clustering observed (Moran's I of 0.93 in 2000, 0.87 in 2010, and 0.74 in 2020). Higher temperatures increased malaria incidence (Odds ratio (OR): 1.06; 95% credible interval (CI):1.04,1.08 and 1.13;95% CI:1.10,1.16) in 2000 and 2010, respectively). Enhanced vegetation index increased the likelihood of malaria incidence (ORs ranging from 5.28; 95% CI: 4.96,5.61) in 2000 to 6.22; 95% CI: 5.91,6.55) in 2020 and higher aridity was associated with higher malaria incidence (ORs: 1.11; 95% CI: 1.10,1.13) in 2010 and 1.07; 95% CI: 1.06,1.07) in 2020). To achieve the WHO 2030 malaria reduction targets, fine-scale and region-specific interventions are essential to mitigate the impact of climate and environmental factors on malaria incidence.

Fig 1. Trends in malaria incidence, climatic and environmental factors, and ITN coverage in Tanzania.

Mean values and their confidence intervals are indicated as dots with error bars.

Fig 2. Spatial distribution of malaria incidence.

Each circle represents a sampled cluster location. Column 1, 2 and 3 represent years 2000, 2010, and 2020 respectively. These maps were generated in R using a basemap shapefile obtained Natural Earth Data (https://www.naturalearthdata.com/) via the rnaturalearth package under a public domain license (https://www.naturalearthdata.com/about/terms-of-use/).

Fig 3. Moran’s I scatterplot of the relationship between malaria incidence in cluster and average values of neighbourhood clusters.

Fig 4. Comparison of predicted malaria incidence between model without covariates (Row A) and model with covariates-best model (Row B) and by years.

These maps were generated in R using a basemap shapefile obtained Natural Earth Data (https://www.naturalearthdata.com/) via the rnaturalearth package.

Fig 5. Percentage change in predicted malaria incidence.

Positive % change indicates a decrease in predicted malaria incidence and negative % change indicates an increase in predicted malaria incidence. (A) Overall country mean percentage change in predicted malaria incidence; (B) Spatial distribution of percentage change in predicted malaria incidence across the country, and (C) Histograms show the distribution of the percentage change. The maps presented here were generated in R using a basemap shapefile obtained Natural Earth Data (https://www.naturalearthdata.com/) via the rnaturalearth package.