With everything that has happened lately, a new pandemic is the least desired thing in the world. However, the lessons that recent years have taught us invite us to better prepare for that eventual scenario, with preventive techniques and resources that allow early detection and intervention.
Hand in hand with artificial intelligence and machine learning, scientists linked to the University of Waterloo, Canada, through the company GoodLabs Studio, presented a proposal to alert medical authorities with data analyzed in real time, to make decisions before a possible new pandemic.
Pandemic detection using AI
In early 2021, Canada’s Department of National Defense (DND) launched a call for innovative proposals to respond to future pandemics. GoodLabs, co-founded by Thomas Lo, won its second consecutive grant to develop the Syndrome Anomaly Detection System (SADS).
SADS is a platform that monitors diseases at a general level, to detect atypical patterns in the spread of diseases and timely inform medical authorities for decision-making. “We have learned from COVID-19 how fast pandemics are, and therefore how valuable reliable real-time data is to understanding risk”says Dr. Jean-Paul Lam of the Department of Economics, special adviser and team leader for AI outbreak detection on the project.
The use of SADS is based on a mobile application in a medical consultation. The SADS app uses natural language processing to anonymously capture the symptoms described during the doctor-patient conversation. That data is then aggregated and classified using deep language machine learning, with the goal of detecting increases in atypical symptoms within the population and assessing the risk of spread.
SADS uses machine learning analytics to code symptoms according to the International Classification of Diseases (ICD-10) and classify how typical or atypical they are. By tracking atypical symptoms over time, SADS creates a statistical visualization that represents how a new disease might spread in a community. The system generates an alert with key information about a possible outbreak and shares it in real time with health and government authorities.
When the team ran a simulation of the 2020 COVID-19 outbreak in various Canadian cities, they found that the city of Toronto, for example, already had a detectable outbreak a week before the city declared a lockdown. The simulation suggests that if SADS had been available at the time, a more proactive response would have been possible.
To protect the confidentiality of those who participate, the team has implemented natural language processing AI technology within the application, so as not to upload the conversation data to the cloud. Personal patient information is protected and only details relevant to analysis (symptoms, age, gender, and location) are collected and aggregated.
With the potential to aggregate health data generated around the world, SADS could be used at the local, national and global levels, depending on the vision of its creators. “We fundamentally believe there is unlimited opportunity for positive impact”Lo told the University of Waterloo. “Our goal is to implement the Syndrome Abnormality Detection System in hospital triage, clinics, telehealth, and virtual health forums, a system that can provide government and authorized health entities with early warning of the upcoming pandemic and its pattern. propagation”, he added.
From GoodLabs they point out that their objective is not only to implement a real-time surveillance system and prepare us for the impact of the next possible pandemic, but they also seek to explore the possibility of developing a mobile personal healthcare virtual assistant with conversational AI, to provide health care where it is needed most.