Artificial Intelligens in Healthcare
The technical development of various branches of AI has been extremely strong in recent years. The area in focus here is population-level analyses using modern machine learning (e g deep learning) – which we refer to as Big Data Analytics.
Modern AI is a toolbox that fits perfectly into the healthcare vision of “precision medicine”, the fully tailored treatment for each patient.
Big Data Analytics
Big data analytics is a complement to traditional evidence creation (clinical studies) that instead directly utilizes knowledge that is recorded in the information routinely stored for all previous patients. This means that it is possible to provide decision support for each new patient through large-scale analyses on comparable previous patients in a routine setting.
Today the challenge lies in creating complete solutions that can be put to actual use in healthcare. This is a demanding systemic challenge that requires both interdisciplinary and cross-sectoral collaboration. Unlike the lab environment, the “clinical wilderness” requires robust handling of all situations, even where incoming data is of poor quality, contradictory or incomplete, as well as managing constant change of, for example, the set of data sources and operators. Even the most powerful algorithms need to be carefully put in context with innovations in workflows and interaction schemes to be beneficial.
AIDA aims to lead Big Data Analytics to real benefit in healthcare. This is done by creating a large-scale demonstrator, that is, an IT system launched in an environment as similar to a diagnostic department as possible with patients, equipment, personnel, etc.
The diagnostic tools developed within AIDA consist of decision support based on AI methods, emphasizing man-machine interaction. The scale in terms of number of patients should be so large that further scaling up to the entire region or even a whole country should be straightforward.
AIDA essentially consists of two parts, the inner core and the outer subprojects. The core contains the technical infrastructure and human expertise to underpin the AIDA efforts. The core is provided by three Linköping-based partners: Center for Medical Image Science and Visualization (CMIV, Linköping University), Sectra AB, and Region Östergötland. The physical location of the AIDA arena is CMIV. For the outer subprojects, AIDA welcomes proposals from any organization working in the area in Sweden, such as those engaged in the Medtech4Health program, both from academia, industry and healthcare.
A large part of AIDA consists of subprojects run by research & innovation groups throughout the country. These projects will be focused on decision support development in large-scale image data annotation, analysis development or pseudo-clinical evaluation.
There will also be opportunities for clinical fellowships where an individual working clinically with Diagnostics works with a project in AIDA (including prolonged stay on site in Linköping) that contribute to the clinical use of AI tools. The clinical fellowships are especially intended for radiologists and pathologists.
You can read more about how to apply for projects and fellowships on the Medtech4health site http://medtech4health.se/aida/