A new study sheds light on a promising approach using machine learning to more effectively allocate medical treatments during a pandemic or any time there's a shortage of therapeutics.

The findings, published in JAMA Health Forum, found a significant reduction in expected hospitalizations when using machine learning to help distribute medication during the COVID-19 pandemic to test the model. The model proves to reduce hospitalizations relatively by about 27% compared to actual and observed care.

"During the pandemic, the health care system was at a breaking point and many health care facilities relied on a first-come, first-serve or a patient's health history to implement who received treatments," said the paper's senior author Adit Ginde, MD, professor of emergency medicine at the University of Colorado Anschutz Medical Campus.

"However, these methods often don't address the complex interactions that can occur in patients when taking medications to determine expected clinical effectiveness and may overlook patients who would benefit the most from treatment. We show that machine learning in these scenarios is a way to use real-time, real-world evidence to inform public health decision-making," Ginde adds.

In the study, the researchers showed that using machine learning that looks at how individual patients benefit differently from treatment can provide doctors, health systems and public health officials with more accurate information in real-time than traditional allocation score models. Mengli Xiao, Ph.D., assistant professor in Biostatistics and Informatics, developed the mAb allocation system based on machine learning.

"Existing allocation methods primarily target patients who have a high-risk profile for hospitalizations without treatments. They could overlook patients who benefit most from treatments. We developed a mAb allocation point system based on treatment effect heterogeneity estimates from machine learning. Our allocation prioritizes patient characteristics associated with large causal treatment effects, seeking to optimize overall treatment benefits when resources are limited," said Xiao.

Specifically, the researchers looked at the effectiveness of adding a novel Policy Learning Trees (PLTs)-based method for optimizing the allocation of COVID-19 neutralizing monoclonal antibodies (mAbs) during periods of resource constraint.

The PLT approach was designed to decide which treatments to assign to individuals in a way that maximizes the overall benefits for the population (ensuring those who are at the highest risk of hospitalization are sure to receive treatments, especially when treatment is scarce). This is done by taking into account how different factors affect the effectiveness of the treatment.

The researchers compared the machine learning approach with real-world decisions and a standard point allocation system used during the pandemic. They found the PLTs-based model demonstrated a significant reduction in expected hospitalizations compared to the observed allocation. This improvement also surpassed the performance of the Monoclonal Antibody Screening Score, which observes antibodies for diagnosis.

"Using an innovative approach like machine learning expands beyond crises like the COVID-19 pandemic and shows we can provide personalized public health decisions even when resources are limited in any scenario. To do so, though, it's important that robust, real-time data platforms, like what we developed for this project, are implemented to provide data-driven decisions," adds Ginde, a leader in the Colorado Clinical and Translational Sciences Institute at CU Anschutz.

The paper in JAMA Health Forum will be the 15th publication to come out of a project called Monoclonal Antibody (mAB) Colorado. The project focused on doing the most good for the most people, using real-world evidence for data-driven decisions during the COVID-19 pandemic.

The researchers hope this paper will encourage public health entities, policymakers and disaster management agencies to look into methods like machine learning to implement in case of a future public health crisis.