AI may be ready to “substantially improve” the clinical experiences and outcomes of COVID-19 patients, but available data “must be accessible, interpretable and actionable.”
That’s according to a study recently published in the journal BMJ Health Care & Informatics. The study, conducted by researchers at the federal Veterans Administration, has lead to the development of a tool, currently being piloted at 13 VA medical centers, that uses AI to calculate the mortality risk for COVID-19 patients.
According to a VA article, the tool creates a report providing 120-day mortality risk scores in both inpatient and outpatient settings that is based on two models.
The first model assesses conditions about a patient that are known before he or she enters a hospital and “relies heavily on age, body mass index (BMI), and co-existing health conditions that can be found in a Veteran’s electronic health record.”
As for the second set of models, the consider many of the same factors from the outpatient models, but add Veterans’ lab work and vital signs taken at admission. These extra data points “drastically improve” the accuracy of the model, according to Tim Strebel, the VA researcher leading the project.
“It’s no surprise that age and BMI are the most predictive factors for mortality in COVID-19 patients,” Strebel said of the first model. “An overwhelming amount of concurrent research confirms this. While a few comorbidities on their own are predictive of mortality, such as diabetes and dementia, we’ve found that the amount of and severity of comorbidities in a patient is the best way to use them to predict mortality.”
Concerning the second, more expansive models, Strebel said “we provide the outpatient models to try and provide clinicians with an additional perspective of who may be at risk based on the information we already know about the patient to promote early treatment.”
That said, Strebel cautioned, “One of the biggest challenges in any AI effort is bias. While age is no doubt one of the leading predictors of death, there are always exceptions. Some patients in their 90s survive COVID-19. Conversely, some really young patients die from COVID-19. . . To help reduce biased decision making, for each of our models we provide additional models that are mirror copies, except we take age out.”
As the BMJ paper puts it, “while there is a general understanding that demographics play a role in understanding differences in physical and mental health among individuals, including Veterans, there is also an increasing recognition that social determinants are potentially even more important. This comes at a time when new data is becoming available. For example, . . . (u)nstructured data can provide valuable information about Veteran experiences, allowing researchers to map qualitative information about experiences into comparable indices.”
According to Strebel, the tool could also serve as a model for AI initiatives that apply to conditions and viruses beyond COVID-19, such as suicide prevention.
Photo by Creativalmages/Getty Images