Region Specific Data-Driven Malaria PrevalencePrediction Systems


Contact Info

UCL - London

Plasmodium falciparum malaria still poses one of the greatest threats to human life with over 200 million cases globally leading to half-million deaths annually. Of these, 90% of cases and of the mortality occurs in sub-Saharan Africa, mostly among children. Although malaria prediction systems are central to the 2016-2030 malaria Global Technical Strategy [1], currently these are inadequate at capturing and estimating the burden of disease in highly endemic countries.

Historically the tendency has been to build monolithic explicit mathematical models that have been unable to provide accurate performance across different malaria regions. As a change in paradigm, we propose data-driven healthcare region-specific systems, see Region-specific Elastic-Net based Malaria Prediction System (REMPS) [2], as a simple and deployable malaria prediction system suitable for high-transmission sub-Saharan all-year-round settings [2]. These regional data-driven systems can be fine-tuned to support regionally dependent adaptability and readiness of healthcare pathways, each with their own critical bottlenecks.

We propose a deployment scenario where many regional centers (figure), each a regionally trained REMPS node (harnessing such local data at its best), push their data and predictions into a distributed ledger that ensures consensus, consistency and immutability of information across participating nodes. As the network of locally specialized predictors grows, it opens the possibility of meta-learning and novelty detection algorithms to be applied for tasks such as early epidemic prediction. In this context, our work provides a realisable example towards achieving truly data-driven open and distributed digital global health.

Research Funded by: