Improving Ambulance Efficiency Using AI
ICT & Artificial Intelligence
Client company:Safety & Security Campus
Aleksandar Nechev
Danial Zarei
James Zhou
Aleksandar Nasev
Project description
The primary research question was to investigate whether ambulance response times could be lowered by adjusting the placement of their dispatch locations.
Context
The project is firmly in the domain of healthcare, though many of the principles can also be applied to other public safety programs, such as the firefighting department.
Given the steady rise of ambulance calls in the recent years, as well as difficulties in keeping up with the response time demands, we were asked to investigate the situation and see whether it is possible to pin-point issues and optimize the placement of vehicles in the city of Eindhoven.
We worked together with SSC to look into the current way of dispatching vehicles as well as determining where the dispatch locations are located and compared it with attempts to modernize and optimize that placement in other countries. Then, we created a model to determine where these new centers should be, as well as a comprehensive validation to show whether dispatching from the new centers would be more efficient or not.
Results
Our project bore fruit to a dashboard that allows the user to enter a time period and then generate a map with the most efficient dispatch points for that specific time period, based on historical data from the previous two years.
By generating the dispatch centers with data from 2019 and 2020 and validating using ambulance calls strictly from 2021, we managed to achieve an estimated two minute reduction in the time it takes for the ambulance vehicle to arrive to the place of the medical emergency.
About the project group
We're a group with a software and business background. For this project, we dedicated two days a week and adapted a more research oriented working style, focusing on exploring different options and substantiating our choices based on previously reviewed literature. Given the nature of the assignment, we spent a lot of time validating the results to ensure as much precision as possible given the circumstances, which was a fun challenge.