Stroke prediction in emergency department aided by machine learning

Machine learning models can help predict those patients in an emergency department who have had an ischemic stroke, researchers reported at the ISC 2022 conference.

About 9% of patients with stroke are misdiagnosed, which is associated with poor outcomes. The use of machine learning in the emergency department could help detect ischemic stroke and avoid misdiagnosis.

The researchers from the Geisinger Health System in Danville, Pennsylvania, USA, collected data on patients treated in emergency departments at 13 hospitals between September 2003 and May 2019. Subsequent data (from June 2019 to December 2020) were divided into pre- and post-COVID cohorts for validation.