The Dalle Molle Institute for Artificial Intelligence (IDSIA USI-SUPSI) and the SUPSI Institute of Design (IDe) joined SPEARHEAD, an Innosuisse-funded consortium comprising 12 Swiss public and private partners, to fight antimicrobial resistance (AMR).

SPEARHEAD will contribute to optimise AMR management practices that may easily be harnessed to address future emerging infections, ultimately resulting in a more resilient healthcare system and society.

AMR is the ability of microorganisms (like bacteria, viruses, and some parasites) to prevent antimicrobials (such as antibiotics, antivirals and antimalarials) from working against them. As a result, many treatments are becoming ineffective against infections that were previously easily treatable.

The partnership, funded by the Swiss Innovation Agency (Innosuisse), is meant to built a first-in-class, globally scalable, modular digital platform to improve antibiotic stewardship, with advanced patient stratification techniques and direct access to results from faster diagnostics. Direct citizen engagement will be used to raise awareness about the issue of AMR and pandemic preparedness.

The project has five main goals: real-time data flow to and from decision-makers; better use of big data for risk stratification; timely diagnostics that can leapfrog a centralized laboratory infrastructure; better and broader citizen and community engagement and early attention to the financial implications of the proposed innovations.

«Starting from January 2022, IDSIA will join the Innosuisse project SPEARHEAD, providing the consortium with its expertise in artificial intelligence and machine learning» stated Laura Azzimonti, Senior Lecturer and Researcher, leading the IDSIA research team involved in the project. «During the 4-year project, IDSIA will be involved in the development of a clinical decision support system based on machine learning models to provide data-driven insights for antibiotic prescriptions and to predict patient-specific clinical risk of antibiotics resistance. This would allow to direct the use of last-resort antibiotics towards patients at risk of resistant infections. The machine learning models will be trained by pooling together the information collected in different Swiss hospitals and properly taking into account the expected variability between the available datasets. Moreover, continuous updates of the models based on new available data will provide near real-time improvements in prediction.»