Sign up to save your podcasts
Or
Share 54. Improving EMS Response to Mental Illness with AI
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
54. Improving EMS Response to Mental Illness with AI
This week is the first of three in a row in which we feature students who have done amazing work to improve the lives of others.
This presentation is an overview of a project our speaker, Pierce Wright, developed aiming to predict the nature of incoming Emergency Medical Service (EMS) calls by training an algorithm.
A volunteer EMR, Wright found most ambulance calls he fielded related to mental health and substance abuse—which most EMS responders are not trained to treat—resulting in the patient being transported to an emergency department (ED). Utilizing EMS to “transport” the increasing number of mental health calls creates two public health risks: (1) an ambulance becomes unavailable for a trauma call; and (2) the patient may be admitted to an ED, often at public expense, when mental health treatment is more appropriate.
A policy solution is to create specific EMS crews staffed with mental health professionals, using the predictive algorithm to identify such calls. After a retrospective study of AI projects in this space, Pierce established an algorithm using 24 million interactions from New York City’s (NYC) EMS Incident Dispatch Data, from January 2005 - March 2022. The algorithm was validated with additional data from April - December 2022. The algorithm achieved 94.5% predictability, compared to 92.3% accuracy for EMS operators. Across 9 million NYC 911 calls annually, the algorithm could more effectively allocate resources for 198,000 cases.
Extrapolated nationally based on data found in Karaca & Moore (2020), where $5.6 billion is spent annually on mental and substance use disorder ED visits, savings of $123 million may be possible—more if the algorithm can prevent misguided hospital admissions. Additional data, such as new call types and customization for regional municipalities, would make the model more robust and scalable.
Pierce Wright is a rising high school senior at The Browning School in New York City. He is a nationally certified Emergency Medical Responder and has spent more than 500 hours on shift with the Westport (CT) Volunteer EMS.
His paper on his working with EMS incident data has been accepted for presentation at conferences around the world, including as the keynote speaker at the International Experts Summit on Public Health and Preventative Medicine in 2025. Pierce’s work has received numerous awards, including first place at the TerraNYC STEM Fair and second place at the New York State Science & Engineering Fair. He has also received the President’s Volunteer Service Award (Silver Medal).
In his free time, Pierce enjoys listening to music, fly fishing, computer builds, and walking his Samoyed, Cloud, in Central Park.
Youtube video of this presentation: https://youtu.be/nil401MZoGg
To read the paper about this project, please visit: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4782065
View all episodes
By The Rotary eClub of Silicon ValleyThis week is the first of three in a row in which we feature students who have done amazing work to improve the lives of others.
This presentation is an overview of a project our speaker, Pierce Wright, developed aiming to predict the nature of incoming Emergency Medical Service (EMS) calls by training an algorithm.
A volunteer EMR, Wright found most ambulance calls he fielded related to mental health and substance abuse—which most EMS responders are not trained to treat—resulting in the patient being transported to an emergency department (ED). Utilizing EMS to “transport” the increasing number of mental health calls creates two public health risks: (1) an ambulance becomes unavailable for a trauma call; and (2) the patient may be admitted to an ED, often at public expense, when mental health treatment is more appropriate.
A policy solution is to create specific EMS crews staffed with mental health professionals, using the predictive algorithm to identify such calls. After a retrospective study of AI projects in this space, Pierce established an algorithm using 24 million interactions from New York City’s (NYC) EMS Incident Dispatch Data, from January 2005 - March 2022. The algorithm was validated with additional data from April - December 2022. The algorithm achieved 94.5% predictability, compared to 92.3% accuracy for EMS operators. Across 9 million NYC 911 calls annually, the algorithm could more effectively allocate resources for 198,000 cases.
Extrapolated nationally based on data found in Karaca & Moore (2020), where $5.6 billion is spent annually on mental and substance use disorder ED visits, savings of $123 million may be possible—more if the algorithm can prevent misguided hospital admissions. Additional data, such as new call types and customization for regional municipalities, would make the model more robust and scalable.
Pierce Wright is a rising high school senior at The Browning School in New York City. He is a nationally certified Emergency Medical Responder and has spent more than 500 hours on shift with the Westport (CT) Volunteer EMS.
His paper on his working with EMS incident data has been accepted for presentation at conferences around the world, including as the keynote speaker at the International Experts Summit on Public Health and Preventative Medicine in 2025. Pierce’s work has received numerous awards, including first place at the TerraNYC STEM Fair and second place at the New York State Science & Engineering Fair. He has also received the President’s Volunteer Service Award (Silver Medal).
In his free time, Pierce enjoys listening to music, fly fishing, computer builds, and walking his Samoyed, Cloud, in Central Park.
Youtube video of this presentation: https://youtu.be/nil401MZoGg
To read the paper about this project, please visit: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4782065