top of page

Developing AI Algorithms for EMS

Dream Journal

Madison Melton - Department of Emergency Medicine



A crucial component of emergency medical services (EMS) is assessing patient severity to provide appropriate prehospital management and transfer to an emergency department (ED). In this study, an artificial intelligence (AI) algorithm based on deep learning was created to predict the necessity for critical care during EMS. The algorithm performs “automatic calculations based on basic information,” including “age, sex, chief complaint, symptom onset to arrival time, trauma, and initial vital signs” [1]. It does not require expert judgment or medical experience. 

Using this algorithm, patients’ need for critical care and the optimal destination hospital, based on data like ICU and ED capacity, could be accurately predicted amid EMS. Further, the variables used by the algorithm “were simple and could be used via a wearable device and information from a patient or their family” [1]. As a result, patients could also receive daily monitoring from home regarding their critical care needs and be referred to the hospital, if necessary. 

This study showed that the AI algorithm could accurately predict critical care needs, outperforming the three ED medical staff members in deciding the triage level with EMS run sheets. However, it has several limitations. It is difficult to completely understand how the AI model predicts critical care, and “contrary to traditional methods, such as XGboost or CatBoost that can present uncertainty measures (e.g., 95% confidence interval), deep learning has greater difficulty in quantifying” this [1]. Additionally, the study was performed on only 2 hospitals in Korea, so it is important to ensure the model is not overfitted to a single environment, which could lead to biased results when utilized in larger populations or different countries. 

Overall, this new algorithm shows great promise in improving the field of EMS, especially if further research and development were to be done.


References:

  1. Kang DY, Cho KJ, Kwon O, et al. Artificial intelligence algorithm to predict the need for critical care in prehospital emergency medical services. Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine. 2020;28(1). https://doi.org/10.1186/s13049-020-0713-4 


Edited By: Firas Batrash, Editor-in-Chief

6 views0 comments

Comments


Subscribe to Our Journal

Thank you for subscribing! We can't wait to share what's coming next with you!

  • Youtube
  • Instagram

Learn more to dream more!

​​

We are a registered 501C3 non-profit organization and all donations are tax- deductible.

Please note that the information on this site is not medical advice. It is created by students with an interest in medical literature and evidence-based medicine. For the most accurate and contextual information, please refer to the original sources cited in each post.

Dr.eam Journal Company

bottom of page