Natalia Pérez de la Ossa
Artificial intelligence to quickly categorise stroke patients and improve their chances of recovery
Natalia Pérez de la Ossa
Fundació Institut d’Investigació en Ciències de la Salut Germans Trias i Pujol (IGTP), Badalona, Spain
Stroke is a highly prevalent disease that affects 15 million people worldwide each year. 40 % of cases are caused by an occlusion in one of the major arteries of the brain and can be effectively treated with endovascular therapy. However, this therapy is only available in a few highly specialised hospitals with experienced staff. It is estimated that half of the patients are initially transferred to centres that lack the skills to carry out this therapy, which can result in significant delays in treatment, unnecessary health care costs and worse patient outcomes. On the other hand, 25 % of strokes are haemorrhagic in nature, and in these cases the earliest possible care at the nearest hospital is needed, where measures to reduce the bleeding can be initiated.
In all cases of stroke, response time is crucial and the chances of neural tissue recovery depend on an agile and precise response in its characterisation and treatment. Artificial intelligence applied in the pre-hospital ambulance setting can be a key tool to identify the type of stroke and the treatment needed by each patient. It can thus facilitate individualised decisions that enhance and optimise the process of transfer management.
The stroke severity assessment tools currently available in the pre-hospital setting, such as the RACE scale, previously developed by the research group and implemented globally, are based on clinical scales that grade the severity of symptoms, but have limited diagnostic capabilities. This project proposes the development of RACE-Plus, a new application aimed at the pre-hospital environment (ambulances) that will use artificial intelligence algorithms to predict the type of stroke a patient is experiencing in order to direct them immediately to the health centre that will ensure the best clinical recovery. The researchers estimate that the correct categorisation of patients and their consequent transfer to the appropriate hospital could increase access to endovascular therapy by up to 20 % and advance treatment by two hours, leading to a 10 % increase in chances of recovery.
RACE Plus: AI-driven application to allow pre-hospital diagnosis and management of stroke patients