A groundbreaking study has revealed a new approach using machine learning to efficiently allocate medical treatments during a pandemic or when there is a shortage of therapeutics. The research, published in JAMA Health Forum, showed a significant decrease in expected hospitalizations by about 27% when machine learning was employed to distribute medication during the COVID-19 pandemic.
Traditional methods of treatment allocation often overlook the complex interactions that can occur in patients when determining the expected clinical effectiveness of medications. Dr. Adit Ginde, the senior author of the study and a professor of emergency medicine at the University of Colorado Anschutz Medical Campus, emphasized the importance of using real-time, real-world evidence to inform public health decision-making.
The study focused on developing a machine learning model that considers how individual patients benefit differently from treatment. This approach provides healthcare professionals with more accurate information in real-time compared to traditional allocation score models. Dr. Mengli Xiao, an assistant professor in Biostatistics and Informatics, played a key role in developing the mAb allocation system based on machine learning.
The researchers introduced a novel Policy Learning Trees (PLTs)-based method to optimize the allocation of COVID-19 neutralizing monoclonal antibodies (mAbs) during periods of limited resources. The PLT approach aims to maximize the overall benefits for the population by prioritizing patients with a higher risk of hospitalization and considering various factors that affect treatment effectiveness.
Comparing the machine learning approach with real-world decisions and a standard point allocation system used during the pandemic, the researchers found that the PLTs-based model significantly reduced expected hospitalizations. This innovative approach could have a lasting impact beyond the current pandemic, providing personalized public health decisions even when resources are limited.
The study, part of the Monoclonal Antibody (mAB) Colorado project, underscores the importance of data-driven decisions during public health crises. The researchers hope that their findings will encourage public health entities, policymakers, and disaster management agencies to consider implementing machine learning methods in future public health emergencies.
For more information, you can access the study titled “A Machine Learning Method for Allocating Scarce COVID-19 Monoclonal Antibodies” in JAMA Health Forum. This research highlights the potential of machine learning to improve healthcare resource allocation and reduce hospitalizations during pandemics.
