AI to stop the Silent Pandemic
It makes no noise and is barely heard of yet. Yet in 2019 globally it claimed the lives of 4.9 million people and it is estimated that the number could rise to 10 million per year by 2050, making it the leading cause of death in the West, ahead of heart attacks and strokes. All the numbers are beginning to be in place for public attention to focus on antimicrobial resistance (AMR), a.k.a. The Silent Pandemic, recently listed by the World Health Organization as one of the ten most dangerous threats to public health.
Antimicrobial resistance refers to the ability developed by certain bacteria-but also by viruses or parasites-to resist the drugs used to treat the diseases they cause. Antimicrobial resistance is not a recent phenomenon: for millions of years, in fact, the purpose of all pathogens has been to multiply and evade the strategies used to treat them. If we focus on bacteria, one of the direct causes of the development of resistance is the massive use of antibiotics, both in human and veterinary medicine. With often disastrous consequences for the environment as well. Once resistance develops, these bacteria can spread from person to person, through the environment, travelers and more.
At the root of the ineffective treatment of an increasing number of infections is precisely the overuse and misuse of antibiotics, which nearly a hundred years ago opened the door to modern medicine and marked a quantum leap for human life expectancy. Alexander Fleming himself, the man who discovered penicillin, warned in his Nobel Prize acceptance speech of the potential risk of resistance development. And he was right: for every antibiotic created over time, resistance has always emerged.
To counter this trend with increasingly alarming contours, a research project has been launched in Switzerland that aims to have an impact on a global scale. It is SPEARHEAD (an acronym for Swiss Pandemic & AMR - Health Econonomy Awareness Detect), and the Dalle Molle Institute for Artificial Intelligence Studies USI-SUPSI, which is in charge of developing artificial intelligence to help physicians optimally prescribe antibiotics to their patients, and SUPSI's Design Institute, which is instead in charge of digital communication aspects and the organization of workshops to raise awareness on the issue among citizens and patients.
Laura Azzimonti is a senior lecturer-researcher at IDSIA where she closely follows a strand of research related to machine learning from data to support medical decisions. About SPEARHEAD, she explains how "cutting-edge artificial intelligence models are employed within the project that can predict antibiotic resistance and help doctors select the best therapy for their patients. Or in some cases to provide them with the information they need to avoid prescribing antibiotics altogether when it is inappropriate. Other working groups at IDSIA focus instead on identifying more effective molecules for transporting drugs across cell membranes, based on simulations and available data."
The large amounts of data available are the basis of how well artificial intelligence works and thus the quality of the solutions it can generate. This is true for all fields in which it is employed, but in the medical field it poses a variety of problems.
In this regard, Dr. Azzimonti explains how "it is often necessary to integrate data collected in different hospital facilities. This obviously brings challenges because the information is heterogeneous with each other, and it is necessary to treat it in a way that protects patients' privacy. Information from each hospital is collected and combined with information from other hospitals to create a global predictive model, which can be used by all healthcare facilities. These methods are particularly relevant in the medical field because, by ensuring privacy, they facilitate collaboration among different hospitals. As a result, by having more data available, they enable the development of more robust and accurate prediction methods.
We also work closely with hospitals in raising awareness, aimed at removing some of the resistance about the use of artificial intelligence, which is in no way intended to replace the work of medical staff, but rather to enhance it and provide society with an increasingly high-quality service."
The large amounts of data available are the basis of how well artificial intelligence works and thus the quality of the solutions it can generate. This is true for all fields in which it is employed, but in the medical field it poses a variety of problems.
In this regard, Dr. Azzimonti explains how "it is often necessary to integrate data collected in different hospital facilities. This obviously brings challenges because the information is heterogeneous with each other, and it is necessary to treat it in a way that protects patients' privacy. Information from each hospital is collected and combined with information from other hospitals to create a global predictive model, which can be used by all healthcare facilities. These methods are particularly relevant in the medical field because, by ensuring privacy, they facilitate collaboration among different hospitals. As a result, by having more data available, they enable the development of more robust and accurate prediction methods.
We also work closely with hospitals in raising awareness, aimed at removing some of the resistance about the use of artificial intelligence, which is in no way intended to replace the work of medical staff, but rather to enhance it and provide society with an increasingly high-quality service."
The project is funded by Innosuisse's Flagship program, and is being carried out by a Swiss multidisciplinary consortium of research institutions headed by the University of Basel. It includes university hospitals and industrial partners with expertise ranging from medical, biological and pharmacological fields to artificial intelligence and digital communication. In this regard, we highlight the MAKEAWARE population engagement project curated by SUPSI's Design Institute.