A Science Partner Journal published in Health Data Science has demonstrated the effectiveness of machine learning models as a decision support tool for emergency medical service personnel in Singapore. This model has resulted in an increase in the quality of triage and the utilization of ambulances, thereby providing significant breakthroughs in the field.
The research team collected data from 360,000 cases from the National Emergency Call Center in Singapore between 2018 and 2020 to develop a machine learning model that can predict the acuity of emergency cases.
This study differs from the previous one in that it presents a robust methodology for developing machine learning models to optimize ambulance triage, which enhances dispatchers’ ability to recognize cardiac arrest.
Machine Learning is an important innovation
Although protocols exist to guide decision making, the limited information obtained during short calls means emergency medical service operators often face challenges in determining the acuity of a case. This can compromise triage quality and become a major problem in cities with growing populations.
Machine learning models have the potential to capture complex and subtle relationships, and a well-trained data model can provide accurate predictions within seconds. This is why researchers at the National University of Singapore (NUS) are determined to harness this potential to improve ambulance dispatch triage.
Han Wang, a researcher at NUS said, “It is very possible to use machine learning to improve triage performance among call center specialists. We are determined to solve the triage problem of ambulance dispatches having too many redundant triages, which can result in overcrowding in the Emergency Department and waste of ambulance resources.”
Open protocol details will be provided
Mengling Feng, Assistant Professor at NUS said that this research is the first research to optimize ambulance triage with machine learning, and they are excited to see the potential of machine learning to save lives. The next step is to implement this system in the real world and compare its performance with the control group.
The research team is open to sharing protocol details and data upon request, with approval from Singapore’s emergency medical services system. The research team hopes that this research will inspire more research in this direction and serve as a roadmap for other emergency medical services systems around the world.