New tool to predict ICU admission and discharge
August 5, 2020
TORONTO – Hospital Intensive Care Units (ICUs) cater to patients with severe illnesses requiring constant monitoring and specialized treatment, including ventilation. Pandemics like COVID-19 significantly increase the risk of ICU overcrowding given the surge in patients, which can negatively impact the quality of care. As a result, an early understanding of a patient’s unique treatment journey, including their ICU requirements, is critical to improved treatment and better management of ICUs.
To help achieve this, Bayer has partnered with Canada’s Digital Technology Supercluster and the Toronto-based clinical information company, Altis Labs, on a unique project to develop an innovative software tool to predict intensive care unit admission and expected date of discharge.
By using the tool, clinicians will be able to make better decisions, sooner, in regard to the capacity of beds, staffing, and ventilators during patient surges. This will enable hospitals to better manage and predict ICU capacity leading to better care and outcomes for patients.
Bayer Radiology will collaborate with Altis Labs in the development of machine learning models to extract features from image data for the prognostication of COVID-19 severity and other types of pulmonary infections. Specifically, Bayer will be developing the AI algorithms that will support the diagnosis of COVID-19 with X-ray and CT imaging.
“Throughout the COVID-19 pandemic, the Bayer Radiology team has continued to find innovative and new ways to support radiologists and hospitals,” says Jerry Orban, country head, radiology. “This partnership represents an extension of that commitment, to support our shared goal of supporting patients during these challenging times.”
“The requirements coming from the COVID-19 pandemic will significantly change healthcare systems and accelerate the role of AI across disciplines to clearly and immediately answer the relevant questions to treat patients,” said Guido Mathews (pictured), VP, head of digital diagnostics. “This very ambitious collaboration project currently starting with the research phase clearly shows from the beginning the potential of machine learning for comprehensive methods of ICU optimization based on radiology data.”
Bayer is providing support in the form of know-how and expertise in artificial intelligence algorithmic development, radiology and medical topics, as well as infrastructure for computing and deployment of AI solutions.
The deployment is expected to start across partner Canadian hospitals early next year.