A new study reveals a promising way to use machine learning to more effectively allocate medical treatment during a pandemic or in the event of a shortage of any therapeutic drug.
Research published in the JAMA Health Forum found that significant reductions in hospitalizations were expected when machine learning was used to help distribute drugs and the COVID-19 pandemic was used to test models. This model has been shown to reduce hospitalizations by approximately 27% compared to actual and observed care.
“During the pandemic, the health care system is on the verge of collapse, and many health care institutions rely on a first-come, first-served or a patient’s health history to determine who is receiving treatment,” said Adit Ginde, M.D., senior author of the paper. Professor of emergency medicine at the University of Colorado Anschutz School of Medicine.
“However, these methods often fail to address the complex interactions that patients may experience while taking drugs to determine the expected clinical effects, and may overlook those patients who will benefit most from treatment.” We show that in these cases, machine learning is a way to use real-time, real-world evidence to inform public health decisions,”Kinder added.
In this study, researchers showed that using machine learning to study how individual patients derive different benefits from treatment could provide doctors, health systems and public health officials with more accurate real-time information than traditional allocation scoring models. Dr. Mengli Shaw, assistant professor of biostatistics and informatics, has developed a machine learning-based mAb allocation system.
“Existing allocation methods mainly target high-risk patients who are hospitalized without treatment. They may ignore the patients who benefit most from treatment. We developed the mAb allocation point system based on machine-learning estimates of heterogeneity of treatment effects. Our allocation prioritizes patient characteristics related to large causal treatment effects and seeks to optimize overall treatment benefits with limited resources,”Shaw said.
Specifically, the researchers investigated the effectiveness of adding a new method based on a strategy learning tree (PLT) to optimize the allocation of COVID-19 neutralizing monoclonal antibodies (mAbs) during periods of limited resources.
The PLT approach is designed to determine which treatments to allocate to individuals to maximize the overall benefit of the population (ensuring that those at the highest risk of hospitalization will always receive treatment, especially if treatment is scarce). This is done by considering how different factors affect treatment outcomes.
The researchers compared machine learning methods to real-world decision-making and standard point allocation systems used during the pandemic. They found that the PLT-based model showed a significant reduction in expected hospitalizations compared to the observed allocation. This improvement also exceeds the performance of the monoclonal antibody screening score, which makes diagnosis by observing antibodies.
“Using innovative methods such as machine learning can transcend crises like the COVID-19 pandemic and show that we can provide personalized public health decisions even in any case with limited resources. However, to do this, what’s important is a reliable real-time data platform, like the platform we developed for this project, implemented to provide data-driven decision-making,”added Ginde, leader of the Colorado Clinical and Translational Science Institute at the University of Colorado at Anschutz.
The paper at the JAMA Health Forum will be the 15th published by the Colorado Monoclonal Antibody (mAB) Project. The focus of the project is to provide the greatest benefit for most people, leveraging real-world evidence for data-driven decision-making during the COVID-19 pandemic.
The researchers hope the paper will encourage public health entities, policymakers and disaster management agencies to study methods such as machine learning to implement them in future public health crises.
There are links in the video below this video. If you are interested, you can open it and have a look.
Thank you for watching this video. If you like it, please subscribe and like it. thank
Original text:https://medicalxpress.com/news/2024-09-machine-hospitalizations-pandemic.html
More information: Machine learning methods for allocating scarce COVID-19 monoclonal antibodies, JAMA Health Forum (2024). DOI：10.1001/jamahealthforum.2024.2884
Journal Information: JAMA Health Forum
Oil tubing: