Logo image
Enhanced surveillance of a pneumonia outbreak of unknown cause in Argentina in 2022 using the AI system, EPIWATCH
Abstract   Open access   Peer reviewed

Enhanced surveillance of a pneumonia outbreak of unknown cause in Argentina in 2022 using the AI system, EPIWATCH

Dr Ashley Quigley, Associate Professor Samsung Lim, Professor C Raina MacIntyre and Damian Honeyman
International journal of infectious diseases, Vol.152(Supplement), pp.53-54
International Congress on Infectious Diseases 2024 (ICID CT 2024), 20th (Cape Town, South Africa, 03/12/2024–06/12/2024)
03/2025
url
Enhanced surveillance of a pneumonia outbreak of unknown cause in ArgentinaView
Published (Version of record) Open CC BY-NC-ND V4.0

Related links

Metrics

1 Record Views

UN Sustainable Development Goals (SDGs)

This output has contributed to the advancement of the following goals:

#3 Good Health and Well-Being

Source: SDGs from InCites

Abstract

Introduction: In 2022, an outbreak of pneumonia of unknown origin occurred in San Miguel de Tucumán, Argentina, later confirmed to be Legionnaires’ disease. Additionally, reports of similar outbreaks emerging in a hospital in Mendoza (approximately 960km away from Tucumán) and in a college in Tucumán where 60 cases of gastroenteritis were reported during the same period, complicated response efforts. Confirmation of a Legionnaires’ outbreak is complex and requires matching both patient with environmental source samples. Artificial intelligence (AI) using open-source data can provide epidemic intelligence to generate early warning outbreak signals to assist with response efforts. Methods & Materials: Open-source intelligence reports of ‘pneumonia of unknown origin’ in San Miguel de Tucumán, Argentina were sourced from EPIWATCH, an AI-driven system harnessing vast open-source data to generate automated early warnings for epidemics. We extracted data between July 1st, 2022 and January 1st, 2023 (before and after the official diagnosis on September 5th, 2022) to analyse early outbreak signals. These included reports of diseases reported as unknown, and syndromes such as severe acute respiratory syndrome, influenza-like illness, pneumonia, acute gastroenteritis and febrile syndromes. To analyse the distribution of cases, we extracted the geographic information from the data, accumulated the number of reports for each state monthly, and then applied advanced hotspot analysis to the aggregated data to understand and characterize the unknown pneumonia outbreak. Results: A localised outbreak of pneumonia was detected in a health facility in San Miguel de Tucumán, Argentina, between 18-22 August 2022. Hotspot analysis on the EPIWATCH data points shows that Tucumán became a hotspot by the end of August 2022, which peaked in the following month. Nine of the total 24 states (37.5%) had pneumonia-related signals, with Tucumán (14 reports) having the highest signal. In September 2022, the number of reports in Tucumán (192 reports) increased by 13.7 times whereas other states had no abrupt changes in numbers. EPIWATCH detected an outbreak cluster well before the August 30th, and before WHO first reported the outbreak with confirmation of Legionnaires’ disease. Discussion: Harnessing AI for outbreak detection holds immense potential for early warning and assessment of unknown epidemics and is particularly useful to identify geographical areas for targeted outbreak investigation efforts. Pinpointing locations of geospatial clusters can identify disease outbreaks and can often provide clues about their cause and spread. In this study, EPIWATCH identified early signals of the Legionnaires’ outbreak and mapped the distribution of pneumonia clusters. Conclusion: EPIWATCH can provide enhanced epidemic intelligence and risk mapping to inform infectious disease control and trigger timely resource mobilization in the face of an emerging epidemic. This research highlights the capability of EPIWATCH as a valuable tool for non-traditional surveillance enabling resource allocation and response efforts to be optimized.

Details

Logo image