How integrating AI and clinical decision support systems can help in the emergency room
The use of artificial intelligence for clinical decision support at the point of care is still in its infancy. Despite media attention and the increase in AI research, translation into clinical practice is rare. Little evidence exists on best practices for deployment, particularly in emergency medicine.
Emergency medicine serves as the front line of healthcare and the integration of AI and Clinical decision support at this critical point of care has the potential to revolutionize the way care is delivered, impacting numerous downstream processes, says Andrew Taylor, associate professor of emergency medicine, director of clinical in informatics in the emergency department and associate director of informatics and data science research at Yale University School of Medicine.
Combine AI and CDS
Taylor will speak on this topic at the HIMSS24 Global Conference and Exhibition in an educational session titled “Deploying Artificial Intelligence for Clinical Decision Support in Emergency Medicine.”
“In the emergency department, where fast and accurate decision-making is critical, AI-CDS tools can significantly streamline processes, improve patient outcomes and optimize resource use,” he explains. “However, this is a complex environment with many variables – from patient demographics to symptom presentation. Therefore, the deployment of AI tools must be done with careful planning and sensitivity to the unique stressors and workflow of the emergency department.
“During this session at HIMSS24, we will explore various applications of AI-CDS in the emergency department, including triage, patient disposition, diagnosis and risk assessment,” he continued. “We will also continue to focus on a guiding philosophy: AI in medicine must grow as an organic extension of human empathy and care, not as a distant technological force.”
The human elements of healthcare
Taylor’s approach emphasizes the creation of AI systems that are technically advanced yet seamlessly integrated with the human elements of healthcare.
“It’s about cultivating AI tools that support rather than replace doctors, and ensuring that technology is a means to improve the human-centered care that is at the heart of medicine,” he said.
Session participants should leave with a deep understanding of AI applications and workflow integration and stakeholder engagement, Taylor said.
“In terms of AI applications, our discussion will delve into the ways in which AI-CDS enables rapid and accurate triage, which is just one facet of its broader capabilities,” he explained. “By quickly analyzing complex patient data, AI algorithms can accurately assess the severity of a patient’s condition, ensuring timely and appropriate medical attention.
“This advanced triage process is not the only benefit; AI-CDS extends its usefulness to risk assessment, helping predict patient outcomes and contributing to more informed decision-making about patient disposition – whether admission to intensive care, a hospital bed or discharge to home care” , he continued.
Furthermore, AI-CDS systems play an important role in improving diagnostic accuracy, which is crucial in the high-stakes environment of the emergency department, he added.
“By integrating these varied functions, AI-CDS supports more nuanced and efficient allocation of emergency department resources and drives better patient outcomes through multifaceted clinical decision support,” said Taylor.
Acceptance and integration
In terms of workflow integration and stakeholder engagement, the success of AI-CDS depends not only on the sophistication of the technology, but also on the acceptance and integration of these systems by those directly affected by their use of it, he said.
“Physicians, healthcare professionals and patients are key stakeholders whose insights, expertise and experiences drive the development of AI solutions that are ethically conscious, transparent and free from bias,” he said. “By actively involving these stakeholders, AI tools can be developed to meet the nuanced demands of healthcare, ensuring that such innovations serve as a supportive extension of human care.
“This engagement process is critical in advancing AI systems that are not only technologically advanced, but also resonate with the core values of healthcare – compassion, privacy and equality,” he continued. “This collaborative approach will allow us to develop AI-CDS tools that respect the delicate human aspects of healthcare, and ensure that these systems are seen as allies in clinical decision-making rather than as impersonal or disruptive forces.”
A robust infrastructure
On another front, an additional lesson from this HIMSS24 session will be the importance of establishing a robust infrastructure for the long-term deployment and use of AI-CDS, Taylor revealed.
“The effectiveness of these systems depends on their ability to blend into existing clinical processes, improving rather than complicating the decision-making process,” he explains. “AI-CDS tools must therefore be designed with the user experience at the forefront, so that they are easy to use, intuitive and provide actionable insights that align with clinicians’ thought processes.
“Additionally, the infrastructure supporting AI-CDS deployment must be robust, adaptable, and able to evolve with the changing landscape of clinical data and healthcare practices,” he continued. “The deployment strategy should include the implementation of machine learning operations, known as MLOps, which are critical to the monitoring, maintenance, and continuous improvement of AI applications.”
This framework ensures that AI-CDS tools remain effective, secure and relevant in the long term, maintaining compliance with strict data security standards and adapting to the dynamic environment of emergency medicine, he added.
Improving patient care
“By building a resilient infrastructure that takes into account the lifecycle management of AI tools, we ensure that these systems can become sustainable assets in medicine, continuously improving patient care while dealing with complexity and ever-changing demands of the healthcare landscape,” he said.
“It is this meticulous attention to operational infrastructure and cultivating a symbiotic relationship between AI-CDS tools and clinical workflows that will drive the success and sustainability of AI in emergency medicine,” he concluded.
The session, “Deploying Artificial Intelligence for Clinical Decision Support in Emergency Medicine,” will take place on HIMSS24 in OrlandoMarch 12 from 1:15 PM to 1:45 PM in room W307A.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.