|
|
 |
| |
|
|
|
Climate-Informed Malaria Prediction Models: A Bayesian Approach For South African Endemic Provinces |
|
|
|
PP: 285-306 |
|
|
|
doi:10.18576/jsap/140210
|
|
|
|
Author(s) |
|
|
|
Makwelantle Asnath Sehlabana,
Daniel Maposa,
Alexander Boateng,
Sonali Das,
|
|
|
|
Abstract |
|
|
| In this study, we predict malaria cases using climate factors and Bayesian methods. Climate change plays a pivotal role in determining both the geographic spread and severity of malaria outbreaks. Recent research underscores that climate-related factors outweigh other contributors, such as epidemiological, socio-economic, and environmental factors, in the resurgence of malaria cases. The coronavirus disease of 2019 (COVID-19) pandemic have caused a setback in the global strides made towards malaria control and elimination. South Africa has not met its malaria elimination targets, despite strategic plans like the National Malaria Elimination Strategic Plan (NMESP), which emphasises strengthening surveillance systems. Researchers are developing malaria forecasting and prediction models incorporating climate factors, primarily using time series and machine learning techniques. While time series models exhibit shortcomings in long-term forecasting, machine learning models have shown promise in prediction but did not prove granularity in delineating critical malaria seasons or providing climate-specific predictions. This study seeks to edify these models using a Bayesian framework to predict malaria in South Africa’s endemic provinces based on climate and environmental factors. The study found that malaria transmission is high in regions with temperatures of 20-30◦C, rainfall of 0-200 mm, and normalised difference vegetation index (NDVI) levels of 0.5-0.8, predicting 200 to 1000 malaria cases in these conditions. The Ehlanzeni district in Mpumalanga and the uMkhanyakude district in KwaZulu-Natal are identified as high-risk areas with elevated malaria counts. Targeted prevention and control measures are recommended for these districts. Future research should explore malaria prediction using subjective informative prior distributions for deeper insights.
|
|
|
|
|
|
|
|
