Establishment of machine learning datasets for better healthcare outcomes: A focus on rabies

The project aims to create machine learning-ready datasets for rabies diagnosis and outbreak prediction, and for the assessment of case detection and viral spread with additional genomic surveillance.

The datasets will be created from existing systems for rabies surveillance that includes Integrated Bite Case Management (IBCM) and contact tracing. The application of machine learning to IBCM data can be used to develop diagnostic algorithms to distinguish between high-risk and low-risk bites, and for field diagnosis of potential rabies animals.

The creation of these machine learning-ready datasets will help guide interventions in rabies control. Machine learning datasets for rabies will support the development of improved practices for controlling rabies in animals and preventing disease in people, including guiding treatment decisions, efficient and accurate diagnoses, and policy decisions including outbreak response.

Datasets will be obtained from the existing systems for rabies surveillance containing extensive data from 2010 onwards with over 15,000 bite patient records, and over 3,000 animal rabies cases. The datasets will be annotated and labeled, ready for machine learning in the diagnosis of rabies from reported clinical signs, prediction of the timing and spread of outbreaks and demand for costly human rabies vaccines. The proposed dataset will be made publicly accessible for researchers, students, and practitioners for ethical and responsible use to address the burden of rabies.

The project will be conducted by Ifakara Health Institute in collaboration with Muhimbili University of Health and Allied Sciences (MUHAS) and the University of Glasgow, UK.

The study will involve the following researchers:
1.    Kennedy Lushasi - Principal Investigator
2.    Rebecca Beldon Chaula - Co-principal investigator
3.    Prof Katie Hampson – Project Collaborators
4.    Joel Changalucha – Project collaborators
5.    Dr Deogratias Mzurikwao – Project collaborator
6.    Dr Kirstyn Brunker – Project collaborator
7.    Dr Elaine Ferguson – Project collaborator

Study duration: 11 months from July 2022 to May 2023
Support and Funding: Meridian Institute, through Lacuna fund.