Data from basic blood work and other vitals can accurately project a COVID-19 patient’s outcomes.
So revealed the results of a nearly three-year-long study led by Northwell Health's research arm, Feinstein Institutes for Medical Research, published in Nature Communications this week. The scientists relied on a data set of 35,000 patients from Northwell Health and built an AI tool to help inform front-line staff on a patient’s prognosis and severity of the disease.
The research accounted for rapid changes in a patient’s condition and outcomes across different COVID-19 waves and the variants that caused them, along with the introduction of vaccines and other treatments. The five data points that were tracked related to a patient’s condition were: age, serum urea nitrogen, lactate, serum albumin and red cell distribution width. EHR data from inpatients hospitalized across 13 Northwell hospitals between April 2020 and May 2022 were analyzed.
“COVID-19 was one of the most dynamic diseases we’ve witnessed in modern history and information about how to care for patients was constantly evolving,” Theo Zanos, Ph.D., senior author of the study, said in a press release. “By harnessing data and developing a real-time auto-updating clinical tool, we set out to create a tool that accounts for these developments and helps clinicians make the decisions they need to deliver better care.”
The model was accurate throughout the research period with changing waves and variants and performs equally regardless of race, ethnicity or gender. It can also be applied to other dynamic diseases.
When COVID-19 first hit, providers scrambled to understand patient conditions amid rapidly changing symptoms and science. Predictive models were developed throughout the pandemic, but most existing ones don’t account for real-time shifts in patient characteristics and outcomes, the announcement highlighted. The latest AI tool monitors nonstop the predictive performance and is automatically updated when it detects changes. This, the researchers argue, is critical to providing clinicians with accurate 28-day survival predictions.
“This important study harnessed data analytics and technology to develop novel insights into a new illness,” Kevin J. Tracey, M.D., president and CEO of the Feinstein Institutes, said in the press release. “Dr. Zanos’ strategy provides a new model to study COVID-19 as a guide to clinical decision-making and better outcomes.”
The researchers published the validation methods and techniques used to build the tool and are also providing the medical community with the open-source code of the framework, the clinical data used in the models and an online calculator that demonstrates the tool’s functionality.