Analytics in Action News

Predictive Analytics Detects Deterioration in COVID-19 Patients

The computer program uses predictive analytics and chest X-ray imaging data to predict which COVID-19 patients are likely to contract potentially fatal complications.

Predictive analytics detects deterioration in COVID-19 patients

Source: Getty Images

By Jill McKeon

- Using predictive analytics and X-ray imaging data, an artificial intelligence computer program predicted which COVID-19 patients would develop life-threatening medical issues within four days with up to 80 percent accuracy, according to a recent study

Conducted by researchers at NYU Grossman School of Medicine and published in npj Digital Medicine, the study stemmed from a need to understand and get ahead of sudden patient deterioration from COVID-19 patients.  

"Emergency room physicians and radiologists need effective tools like our program to quickly identify those COVID-19 patients whose condition is most likely to deteriorate quickly so that health care providers can monitor them more closely and intervene earlier," said study co-lead and assistant professor in computer engineering at NYU’s Abu Dhabi campus Farah Shamout, PhD, in a press release.  

Researchers input X-ray information, vital signs, age, gender, and race into the program, which then used the data to assess the patient’s risk level. The system learns using an imaging-based classifier based on the Globally-Aware Multiple Instance Classifier (GMIC), as well as collected clinical variables from a gradient boosting model (GBM). The program was trained with a dataset of 3661 patients, and the predictions of both models were combined to calculate an overall risk prediction, the study said.  

“Additionally, our results suggest that chest X-ray images and routinely collected clinical variables contain complementary information, and that it is best to use both to predict clinical deterioration. This builds upon existing prognostic research, which typically focuses on developing risk prediction models using non-imaging variables extracted from electronic health records,” the study stated.  

To ensure the predictive value of the AI program, researchers used 770 chest X-rays and accurately predicted which patients would go on to receive mechanical ventilation, intensive care, or would die within four days of hospital admission. The program was accurate for four out of five patients, according to these tests.  

"We believe that our COVID-19 classification test represents the largest application of artificial intelligence in radiology to address some of the most urgent needs of patients and caregivers during the pandemic," explained Yiqiu “Artie” Shen, MS, a doctoral student NYU’s Data Science Center. 

The study notes that the program only focused on chest X-ray images due to the nature of COVID-19, but cannot predict non-pulmonary-related complications. In addition, the dataset was collected only from NYU Langone Health in New York, and demographic considerations were not taken into account.  

Despite these limitations, AI programs thrive when given more data to work with, and researchers are continuing to improve and update the program as new data becomes available, according to study senior investigator and assistant professor in the Department of Radiology at NYU Langone Krzysztof Geras, PhD.  

The development of predictive analysis tools using AI and machine learning is on the rise. A recent study used machine learning to help clinicians pinpoint the most suitable imaging test for patients who may have coronary artery disease. While there are still significant hurdles and further research is needed in the AI space, many researchers feel that the future is promising.  

“Our findings show the promise of data-driven AI systems in predicting the risk of deterioration for COVID-19 patients, and highlights the importance of designing multi-modal AI systems capable of processing different types of data,” the study explained. 

“We anticipate that such tools will play an increasingly important role in supporting clinical decision-making in the future.”