Data-driven malaria prevalence prediction in large densely populated urban holoendemic sub-Saharan West Africa

Biobele J Brown,Samuel Omokhodion,Ikeoluwa Lagunju,Michael Shaw ,Rémy Claveau ,Olugbemiro Sodeinde,Gbeminiyi Oyinloye,Geraint Rees,K Osinusi ,Muna Elmi,Petru Manescu,Olayinka Kowobari,Alexander A Przybylski,Felix O Akinbami,Fabio Caccioli,Nathaniel K Afolabi,Wuraola A Shokunbi,Wasiu A Ajetunmobi,Adebola E Orimadegun,Francis Akinkunmi,Mandayam A Srinivasan,Vijay Pawar,John Shawe‐Taylor ,Delmiro Fernandez‐Reyes

doi:10.1038/s41598-020-72575-6

Abstract

Over 200 million malaria cases globally lead to half-million deaths annually. The development of malaria prevalence prediction systems to support malaria care pathways has been hindered by lack of data, a tendency towards universal “monolithic” models (one-size-fits-all-regions) and a focus on long lead time predictions. Current systems do not provide short-term local predictions at an accuracy suitable for deployment in clinical practice. Here we show a data-driven approach that reliably produces one-month-ahead prevalence prediction within a densely populated all-year-round malaria metropolis of over 3.5 million inhabitants situated in Nigeria which has one of the largest global burdens of P. falciparum malaria. We estimate one-month-ahead prevalence in a unique 22-years prospective regional dataset of > 9 × 104 participants attending our healthcare services. Our system agrees with both magnitude and direction of the prediction on validation data achieving MAE ≤ 6 × 10–2, MSE ≤ 7 × 10–3, PCC (median 0.63, IQR 0.3) and with more than 80% of estimates within a (+ 0.1 to − 0.05) error-tolerance range which is clinically relevant for decision-support in our holoendemic setting. Our data-driven approach could facilitate healthcare systems to harness their own data to support local malaria care pathways.

Highlights

Over 200 million malaria cases globally lead to half-million deaths annually
Using our Ibadan dataset Training set (DTRAS) dataset, we show that EN regularization-strength and L1-norm parametrization produce nextmonth prevalence estimates with low error and allows us to build a regionally adaptable Region-specific EN based Malaria Prediction System (REMPS)
Previous malaria studies in different world regions and our 60-years Ibadan academic healthcare system knowledge formed the basis to select the variables incorporated into region‐specific elastic‐net based malaria prediction system (REMPS)

Summary

Introduction

Over 200 million malaria cases globally lead to half-million deaths annually. The development of malaria prevalence prediction systems to support malaria care pathways has been hindered by lack of data, a tendency towards universal “monolithic” models (one-size-fits-all-regions) and a focus on long lead time predictions. Nigeria accounts for 29% of worldwide malaria cases and 26% of deaths in 2015 (mostly in children under five years of age), the largest proportion from any one c ountry[7] This global health challenge is striking in large urban densely populated cities such as Lagos (> 15 million inhabitants) and Ibadan (> 3.5 million inhabitants) both under large all-year-round malaria burden where stretched healthcare resources will benefit from advance knowledge of malaria prevalence to support their specific malaria clinical care pathways (Fig. 1). Model-based geo-statistics have provided important contributions to global estimates of the burden of malaria disease[8] These approaches have been less effective in short-lead prevalence prediction in the context of region-relevant (local-scale) clinical pathways

Methods

Results

Conclusion