The Argentinean Public Health Research on Data Science and Artificial Intelligence for Epidemic Prevention (ARPHAI) is an interdisciplinary research consortium, whose mission is to develop technological tools and recommendations to anticipate and manage epidemiological events. Its key focus areas include: i) researching and piloting data-driven tools using artificial intelligence and data science towards upgrading the national electronic health record (EHR) system; as well as ii) monitoring the implementation of the EHR system and providing support for its national scaling-up strategy. With the COVID-19 crisis impacting the health system in Argentina, accelerating these efforts become all the more important to promote evidence-based decision making on public health.
ARPHAI is led by the Interdisciplinary Center for Science, Technology, and Innovation Studies, known as CIECTI. Other members of the consortium include the Secretariat of Planning and Policies of the Ministry of Science, Technology, and Innovation and the National Directorate of Information Systems at the Secretariat of Access to Health, Ministry of Health in Argentina. As part of its mission to expand capacities across diverse disciplines throughout Argentina, ARPHAI also convenes scientists, technical staff and policy makers from 19 institutions, including ministries, health institutions, universities and research centers located in six Argentine provinces and the Autonomous City of Buenos Aires.
The national EHR system is called Historia de Salud Integrada in Spanish, or HSI in short. ARPHAI is piloting three HSI-based components in parallel to anticipate and detect potential epidemic outbreaks. Their modular design allows for integration to better facilitate public health decision-making on COVID-19 and other infectious diseases. In particular, these components include:
1) the extraction of computable phenotypes of diseases, symptoms, and syndromes using natural language processing to analyse EHR structured and free-text;
2) models for understanding and prediction of relevant epidemiological variables (e.g., intensive care unit bed occupation, number of disease cases, number of deaths, etc.) using computable phenotypes and open data information as input; and
3) dashboard visualization of the results from both points above, along with additional open data sources to inform decisions made by public sector epidemiological meso-management authorities (e.g., public servants in provinces or other large jurisdictions).
There are two additional lines of work ARPHAI undertakes that are transversal to these three research developments, which include a) diversity, equity, and inclusion (DEI) with a focus on gender and b) responsible use of health data.
To implement DEI in the development of the data-based tool, our research team is working to detect and, when feasible, mitigate biases in the phenotypes, models, and dashboards developed. We are also developing recommendations that lend a gender-inclusive perspective to the HSI system, with particular attention to the application of Argentina’s Gender Identity Law (Law 26,743, 2012).
To advance the responsible use of health data, ARPHAI is developing key guidelines to ensure an ethical secondary use of EHR for research and socially responsible decision making, including ethical data management, ethical approval and monitoring, and consented data use. ARPHAI is also carefully documenting the anonymization and security processes to protect EHR privacy. These processes include tuning to Spanish existing natural language processing algorithms used for free-text entity detection.
Regarding the second component of facilitating HSI implementation and its national scalability we have discussed above, ARPHAI is currently supporting the implementation of two pilots in Quilmes and Almirante Brown districts in Argentina. ARPHAI is providing support around systematizing vital information for the HSI national scaling-up and deployment proposal.
We expect that ARPHAI’s work will have an impact on provincial and municipal government levels, by activating epidemiological early alerts and providing evidence-based information for the management of the public health system. ARPHAI also seeks to strengthen the national health system and lay the foundations for addressing practical ways of managing and safeguarding sensitive data being harnessed for decision making. All of ARPHAI resources will be open-source and reusable, with the potential of being useful in other Latin American countries.
ARPHAI is pleased to be part of the Global South AI4COVID Program which is funded by Canada’s International Development Research Centre (IDCR) and the Swedish International Development Cooperation Agency (Sida). We would like to thank IDRC and Sida for this unique opportunity that allows policymakers and researchers to engage in this indispensable dialogue to solve problems that are greatly impacting society.