Gamal Abdel Nasser University of Conakry, Guinea
This application seeks to characterise the evolution of the signs and symptoms of EVD using individual patient data (IPD) hosted by the Ebola Data Platform. In an emergency outbreak situation, it is important to have a reliable indicator for medical screening of patients and monitoring for evolution of signs and symptoms associated with EVD to identify those at increased risk of death, helping to optimise health care delivery. The study uses IPD meta-analysis, considered the gold-standard approach for synthesis of results across studies, where the volume of available data provide a unique opportunity to characterise the evolution of EVD symptoms. Clinical signs and symptoms at baseline and during hospitalisation period in Ebola treatment centre (ETC) patients with confirmed EVD are described, with longitudinal mixed effects modelling to characterise the population-averaged trajectory of the clinical signs and symptoms, and competing risk survival analysis to estimate the cumulative incidence of fever clearance, haematological recovery, and viral clearance.
Read full application
École Polytechnique Fédérale de Lausanne (EPFL), Switzerland
The study will create an open-source collaborative data analysis platform that allows users to crowdsource diverse clinical insights on EVD from rapidly accumulating data across fragmented datasets, whilst preserving patient privacy and local ownership. Using the Ebola Data Platform to offer evidence on the applicability and performance of new, decentralized machine-learning methods, the study will create a platform for robust real-time data analysis during an outbreak. These methods will enable clinical researchers to generate reliable models using patient data that is distributed across sites, without the need for data owners to share any original data or provide their patient records to a centralised repository. This will strengthen patient privacy and incentivise collaboration and interoperability between competing research parties, which is especially relevant during health emergencies and in rural settings without reliable centralised medical infrastructure.
Read full application