TL;DR:
- Emory University and Georgia Tech leverage artificial intelligence (AI) to enhance COVID-19 treatment access.
- Natural Language Processing (NLP) AI model classifies patient-initiated messages with 94% accuracy.
- Faster response times to patient messages improve the likelihood of receiving antiviral prescriptions within a critical treatment window.
- Collaboration with Switchboard, MD facilitates the development and deployment of the NLP model.
- AI integration in healthcare holds the potential to reshape medical practices by prioritizing and enhancing human interactions.
Main AI News:
In the ever-evolving landscape of healthcare delivery, Emory University’s School of Medicine and the Georgia Institute of Technology are pioneering the use of artificial intelligence (AI) to enhance the efficiency of diagnoses and treatment. Their groundbreaking research aims to address the pressing challenges faced during the COVID-19 pandemic, particularly the need for timely access to medical care and treatment.
With the rapid rise of telemedicine and electronic health record (EHR) messaging, virtual visits have become increasingly popular. Patients now have the convenience of reporting positive COVID-19 test results from the comfort of their homes, eliminating the need for in-person visits. While this shift has its advantages, the sheer volume of messages without an organized triage system has resulted in delays in response time and hindered access to critical treatment.
To tackle this issue head-on, a recent study published in JAMA Open Network investigated the effectiveness of natural language processing (NLP), a specific type of AI, in expediting patient-initiated messages, physician responses, and access to COVID-19 antiviral treatment. Leveraging deep learning predictive models and building upon their previous research, the team developed an innovative NLP model to classify patient-initiated EHR messages. This model’s accuracy was evaluated across five hospitals in the Atlanta area between March 30 and September 1, 2022, with an impressive 94 percent accuracy rate.
The findings revealed a compelling correlation between the speed of responses to patient messages and the likelihood of patients receiving antiviral medical prescriptions within a crucial five-day treatment window. Nell Mermin-Bunnell, lead author of the study and a third-year student at Emory School of Medicine, expressed excitement about the potential impact of natural language processing, stating, “We were thrilled to witness how this advanced technology accurately and instantaneously triaged patient messages, ensuring improved access to treatment. While our study focused on COVID-19 diagnoses, the scope of this model could be expanded to other medical conditions.”
May Wang, PhD, a co-author and professor at Georgia Tech, highlighted the power of advanced NLP models in real-time identification of patients at risk for specific diseases. Wang stated, “These results underscore how leveraging NLP technology can significantly expedite patient access to healthcare, emphasizing the pivotal role of speed in providing prompt medical attention.”
The study’s success was made possible through a collaboration between Emory University, Georgia Tech, and Switchboard, MD, a pioneering data science and artificial intelligence company founded by physicians from Emory Healthcare. The NLP model utilized during the study, named eCOV, was initially developed by Dr. Blake Anderson, CEO of Switchboard, MD, and a primary care physician at Emory. As the adoption of EHR messaging surged, Dr. Anderson recognized the need for efficient message organization to alleviate the burden on clinical staff and prevent burnout. His team conducted extensive experiments to refine the model’s performance and focused on an algorithm that considered the context of the message, going beyond simple keywords.
Dr. Anderson explained, “Our goal is to sift through a massive influx of data and extract what is most relevant for those who urgently need it, enabling patients to receive care in a more expedient manner.” Once the model was fine-tuned, he collaborated with Georgia Tech to ensure its reproducibility and initiated its deployment to evaluate its efficacy in facilitating physician-patient communication.
While further analysis is required to determine the model’s impact on clinical outcomes, one thing is certain: as AI becomes increasingly integrated into mainstream healthcare, it possesses the potential to reshape medical practices. Dr. Anderson reassured those concerned about the role of AI in medicine, emphasizing that “this type of NLP empowers AI to prioritize and enhance human interactions rather than replacing them.”
Conclusion:
The utilization of artificial intelligence, particularly through advanced NLP models, has the potential to revolutionize healthcare delivery. By accurately classifying patient-initiated messages and expediting response times, AI technology enhances access to COVID-19 treatment. This innovative approach not only demonstrates the power of AI in improving patient outcomes but also signifies a significant market opportunity. As AI continues to be integrated into mainstream healthcare, companies specializing in AI solutions are poised to play a pivotal role in transforming the industry and improving overall patient care.
