Looking Ahead: The Future of Artificial Intelligence in Emergency Medicine

Zoee Castro, MSIII, University of Nevada, Reno School of Medicine
EMRA MSC Student Advising Coordinator, 2024 - 2025

Between social media outlets, news articles, and general conversation, Artificial Intelligence (AI) has been a highly discussed topic in the past few years. It is quickly gaining popularity and becoming integrated into our society. Individuals are beginning to understand how AI can help with efficiency and work production, but there is also the counter debate that AI is affecting us negatively. Regardless of the ongoing debate around AI, advancements have been made to incorporate it into medical practice. Artificial Intelligence will be in our future as practicing physicians, but what exactly does this mean for the field of emergency medicine?

Some argue that Artificial Intelligence will improve efficiency, especially in busy emergency departments. As of now, physicians are tasked with completing multiple forms of documentation regarding patient encounters. Spending time requesting labs/imaging, orders for medications, and completing discharge summaries all take time away from bedside interactions. Dr. Jace C. Bradshaw has offered insights suggesting the ability to employ AI to quickly and accurately generate personalized discharge summaries for patients, specifically targeted to a patient’s level of understanding.1 He states that emergency physicians should be one of the first to fully adopt artificial intelligence as it can provide us with new capabilities to improve patient care and spend more time with them at the bedside.

Artificial Intelligence can also provide the opportunity for rapid diagnostics. AI algorithms can be used to analyze chest x-rays, CT scans, MRIs, etc., all in a prompt manner. This is especially useful in the emergency department as patients can present with symptoms concerning for acute and life-threatening conditions like a stroke or aortic dissection. By utilizing the rapid diagnostics of imaging that AI can provide, this can lead to quicker interventions in these critical situations where every minute matters. Additionally, these rapid diagnostics capabilities of AI could be beneficial to ED departments that are in rural areas with limited radiology support.

As great as AI sounds, there are also a few limitations that must be considered when integrating it into medical care and emergency medicine. One of those limitations that we must address is the bias that AI systems interpret data through. AI algorithms can only be as good as the data they are trained on, meaning that the current AI systems have been fed historical data with underrepresentation of minorities.2 Allowing AI systems to utilize this data could only further perpetuate under-treatment of minorities that already struggle to find equitable health care. Additionally, AI systems have the potential to disregard significant patient history that it deems as “unimportant” since it does not fit into the algorithm box for a diagnosis or treatment.

Another argument against AI is that people fear EM physicians may become too reliant on these AI systems. AI is meant to be used as a tool and should not become a crutch that guides every single patient encounter or treatment plan. Many fear that by incorporating AI into emergency departments, it also encourages physicians to disengage from patient care at the bedside and allow algorithms to determine plans of care. It is essential that as future EM physicians, we continue to utilize our time at the patient's bedside and critical thinking skills with patient presentations.

Overall, Artificial Intelligence has the potential to become a great tool in the emergency department. If used within its limitations, it can help increase efficiency and patient care. As future EM physicians, though, we must remain aware of the bias that these AI systems can interpret data through and ensure that it is not placing our patients in a box they do not belong. It is essential that we continue to look at patients individually and not as a cluster of complaints that can be sorted. As we continue to navigate this new world of Artificial Intelligence in the emergency department, we must remember that the one thing AI can’t do is spend time at a patient’s bedside and establish a genuine human connection.

References:

  1. Bradshaw M.D., J. (2023, June). The ChatGPT Era: artificial intelligence in emergency medicine. Annals of Emergency Medicine. Retrieved February 8, 2024 from https://www.annemergmed.com/article/S0196-0644(23)00035-5/fulltext
  2. Chenais, G., Lagarde, E., Gil-Jardine, C. (2023, May 23). Artificial Intelligence in emergency medicine: viewpoint of current applications and foreseeable opportunities and challenges. Journal of Medical Internet Research. Retrieved February 8, 2024 from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10245226/

Related Content

Nov 29, 2021

Program Director Interview Series: Laura Smylie, MD | EM Residency Assistant Program Director at Wayne State University Detroit Receiving Hospital

In this month’s addition to the Program Director Interview Series, we got to chat with Laura Smylie, MD, to learn about the Wayne State University Detroit Receiving Hospital. Dr. Smylie tells us more about residency in the amazing Motor City.

Jun 24, 2021

Program Director Interview Series: Gillian Baty, MD | EM Residency Program Director at University of New Mexico, Albuquerque, NM

In this month’s addition to the Program Director Interview Series we got to chat with Dr. Gillian Baty to learn about the UNM EM Residency Program. If residency in the Land of Enchantment sounds appealing to you, then Dr. Baty has a lot to share below.

Jun 23, 2021

Program Director Interview Series: Shannon Toohey, MD | EM Residency Program Director at University of California, Irvine, CA

In this month’s addition to the Program Director Interview Series we got to chat with Shannon Toohey, MD to learn about the UCI Residency Program. Dr. Toohey tells us more about residency in the Golden State.