The devastating COVID-19 virus has resulted in high morbidity and mortality rates, with the lack of integrated data systems affecting governments’ abilities to design responsive programming. In Rwanda, where the first confirmed case was on March 14, 2020, the count is nearing 100,000 cases with around 1,300 deaths at the time of writing. In the early phase of the outbreak, testing was limited to high-risk groups (individuals known to have contact with positive cases) that likely resulted in undercounting across communities. With local authorities starting to strengthen COVID-19 preventive and management measures, there is an emerging need for accurate and timely data on the prevalence, incidence and evolution of the outbreak to better outline an epidemiological profile. This data however is currently fragmented, incomplete and scattered across multiple institutions such as hospitals, clinics and testing sites and the challenge at hand is how to integrate these datasets to inform evidence-based decisions.
Rwandan scientists from different institutions have come together to respond to this challenge. Ranging from local, public and international spheres (namely the University of Rwanda, Ministry of Health/ Rwanda biomedical centre, Regional Alliance Sustainable Development and Gent University, Belgium), the consortium gathers experts from the data science and e-Health fields; academic centres of excellence; government agency; healthcare industry; and the private sector that are experienced in data management and community outreach. Premised on using artificial intelligence and data science techniques, the main goal of the consortium is to access, harmonize and analyse available COVID-19 data in Rwanda towards building a centralised data system that can be used for monitoring and predicting outbreaks.
The goal of our team is to identify and gain access to existing datasets, pertinent to the first 17 months of the outbreak in Rwanda. This will be complemented by enriched longitudinal data that will be collected over 24 weeks, through mobile application surveys, telephone calls and in-person surveys). Using technical and standardized framed models, in the first stage of the project, data will be pooled from multiple existing sources. These will include data in different file formats (e.g. Excel, Registers and Electronic Medical Records) on people who have tested both positive and negative for COVID-19 from 15 health facilities in regions with a high number of COVID-19 cases.
For the longitudinal data collection, a minimum of 214 respondents per administrative district is expected to be reached. We envisage that once this stage is concluded, the harmonized datasets can be updated to include the new data from the longitudinal survey. This is with the view of creating an integrated COVID-19 data system, which can be leveraged for future use, such as applying mathematical modelling techniques, statistical methods and machine learning methods to predict COVID-19 cases across communities. More importantly, the data insights can be used to help the Government of Rwanda to better allocate resources to ease the burden in facilities that are under pressure.
Over the past year, the team has invested in getting the project off the ground, including arranging a press conference in February 2021 to officially launch the project. The first stage of the project has been initiated, with significant progress made in identifying different datasets available; while for the second stage involving the survey, respondents have been sampled and the survey instrument is being finalized. The project team has also assembled a team of data collectors to support the phone and in person surveys.
We are encouraged by the attention our work has garnered across various media platforms, and motivated to move ahead towards achieving a meaningful impact. To promote dialogue around the work we’re doing so, we’ll be facilitating a series of events and making our knowledge products available. To that end, we are also pleased to share our poster presentation, titled “A federated data network to support COVID-19 research in Rwanda”, which was accepted at the 2021 Observational Health Data Sciences and Informatics (OHDSI) Conference this year.
We would like to express thanks to Canada's International Development Research Centre (IDRC) and the Swedish International Development Cooperation Agency (Sida) for supporting this initiative through the Global South AI4COVID Program. In addition, our work would not be possible with the support of the Government of Rwanda and all the institutions who are part of this consortium.
To know more about project, please visit our website http://laisdar.rbc.gov.rw and follow our work on Twitter: @laisdar and Facebook: Laisdar LaisdarTags: #COVID-19Data, #DataSystem, #PredictingOutbreaks