The HarmonAIze Project: Pioneering Automated Climate and Health Data Harmonization and Integration

Wits Health Consortium (Pty), Ltd, Johannesburg

Investigators

Linked publications & trials

Abstract

Project Summary An effective and efficient response to climate change and other major public health threats in Africa requires the optimal use of existing and routinely collected data, in accordance with the WHO FAIR principles. The DS-I Africa initiative, and other projects and programs, have made significant progress in addressing barriers to data sharing. Progress in upscaling data harmonization has, however, lagged considerably. The HarmonAIze project addresses this key obstacle to realizing the potential of data science and Artificial Intelligence to advance climate change and health research, and public health more broadly, in Africa and, indeed, globally. The overall objective is to render data harmonization and climate integration more feasible, while optimizing the inputs of health, data science and climate science experts. Firstly, we aim to develop an open-source toolkit that leverages Artificial Intelligence to facilitate semi- automated harmonization of health data, using large-language models to perform repetitive tasks that currently require significant time and specialist expertise. Advances in the underlying methods will be supported by a human-in-the-loop, judiciously utilizing domain-specific experts, while reducing common barriers to their involvement in data science approaches. Achieving this will accelerate progress in the HE 2AT Center, but importantly will also contribute to similar global projects, and, in fact to research on any health data reanalysis. Secondly, the toolkit will perform spatiotemporal integration of climate and health data, leveraging advances in African-optimized, semi-automated, geocoding pipelines and curated, downscaled, remote sensing climate data. This approach allows for a federated approach to preserve data privacy. Additionally, this harnesses the experience and expertise developed within the HE2AT Center to catalyze climate and health research on the continent and beyond. We aim to pilot the approach through the DS-I Africa Research Hubs, initially UZIMA- DS and MADIVA, contributing to a critical mass of climate-health researchers in Africa. Thirdly, to achieve these ambitious goals, we will place an emphasis on capacity building within the HE2AT Center and partners, focusing on early-career investigators and transdisciplinary collaboration across the continent within the supportive scaffold of the Research Hubs, and Training Programs. Lastly, we will collaborate with the eLwazi Open Data Science Platform, in the pursuit of establishing a Center of Excellence for data harmonization on the continent. This platform will foster sustainability beyond the lifespan of the HarmonAIze project and facilitate ongoing improvements and implementation in future projects. The HarmonAize project has the potential to present a step-change in climate and health research within the HE2AT Center, across the DS-I Africa initiative, and other projects, through developing and sharing innovative data science tools to facilitate climate-health analyses. The project has the potential to accelerate the broad transformation of research into public health policy and interventions, on the African continent and beyond.