Logo image
Back
Global epidemiology of botulism outbreaks from open-source intelligence, 2016-2025
Poster

Global epidemiology of botulism outbreaks from open-source intelligence, 2016-2025

Damian Honeyman, David James Heslop A/Prof and Raina MacIntyre Professor
Communicable Diseases & Immunisation Conference 2025 (CDIC) (Adelaide, Australia, 10/06/2025–12/06/2025)
11/06/2025

Metrics

1 Record Views

Abstract

Public health Epidemiology Artificial intelligence
Introduction: Botulism is a rare, potentially fatal illness resulting from intoxication with a botulinum neurotoxin. Outbreaks are reported worldwide in humans, animals, and the environment. With toxin identification reliant on traditional surveillance methods conducted in laboratories. Open-source epidemic intelligence (OSINT) provides an alternative disease surveillance method in which outbreaks are reported from open sources such as news media, websites, or social media. Methods: Using the AI-driven OSINT early warning system EPIWATCH, we extracted a dataset published between 31 December 2016 and 1 January 2025 and filtered by our search terms to identify botulism outbreaks. Data collected included country, event date, location, symptoms, outbreak type (human/animal), case numbers, hospitalisations, antitoxin treatment, deaths, source and origin of outbreaks, botulism type, toxin strain, and time from exposure to symptom onset. Analyses were conducted using STATA/BE 17.0 and hotspot mapping in ArcGIS Pro v.3.1. Results: We identified 296 botulism outbreaks, affecting 28,634 individuals, with 27,334 deaths. Human outbreaks (87.5%, n = 259) resulted in 1,097 cases and 47 deaths, while animal outbreaks (12.5%, n = 37) accounted for 27,537 cases and 27,287 deaths. The highest number of human outbreaks occurred in Ukraine (56.0%, n = 145), the Russian Federation (10.8%, n = 28), and the United States of America (7.0%, n = 18). Common symptoms included dysphagia (11.0%), nausea (9.0%), and generalised weakness (8.8%). The primary transmission routes were foodborne (79.5%), unknown (13.5%), and iatrogenic (3.1%). The most frequently implicated sources were dried fish (33.2%), unknown (15.4%), and canned fish (5.4%). Toxin strain was rarely reported (97.7%), with strain B (1.2%), strain E (0.8%), and strain A (0.4%) identified. Symptom onset most often occurred within one day (8.1%) or the same day (5.0%) after exposure. Conclusions: There is no global surveillance system for botulism. OSINT-based surveillance provides a unique way to understand the global epidemiology of outbreaks, especially for diseases for which global surveillance systems do not exist, and to guide outbreak response.

Details

Logo image