Inspiration
About 25% of Americans live in rural areas, yet more than 60% of U.S. trauma deaths occur there (Vanderbilt University Medical Center). Blood loss is also a huge concern with trauma cases. Standard responding ambulances (BLS and most ALS) cannot carry or administer blood, while the nearest trauma center or blood-carrying supervisor is often miles away. According to ems.gov, around 37% of severely bleeding trauma patients can be saved with prehospital blood. Researchers at Pitt state that every 1-minute delay before prehospital blood transfusion increases 30-day mortality odds by 2%, and every 5 miles farther from a trauma center raises the odds of death by 8%-23%.
Our solution is a web application that, based on areas of frequent trauma incidents, finds optimal places to put blood vehicles on standby, in order to reach as many people as possible in a certain amount of time. It also finds the best place for the blood vehicles and patient to rendezvous on the way to the hospital.
This can be used by first responders and people in the chain of command in order to reduce the time before severe trauma patients receive blood, saving lives.
How we built it
User interface: Next.js + React
Maps and routing: Mapbox API
Backend:Python FastAPI
OR-Tools: Mathematical optimization python library. Decides where blood response units should be placed / best rendezvous plan
Challenges building BloodGrid
Defining when it is not worth rendezvousing: Getting blood tranfusions as quick as possible is important, but quickly getting to the hospital is even more important. We had difficulty defining the cutoff between when it was worth rendezvousing to a blood vehicle (lengthening total travel time to hospital) or to just not rendezvous and go straight to the hospital (prolonging time without transfusion).
Finding a route for which the ambulance and blood vehicle meet where the ambulance goes the route with least resistance towards the hospital whilst also meeting with the blood vehicle as soon as possible: If the hospital is 20 minutes away whilst the blood vehicle rendezvous point is 12 minutes away, its a no-brainer to rendezvous, right? This would be a horrible thing to generalize because what if the rendezvous point is in a opposing direction to the hospital. Defining what blood vehicle and rendezvous point to use to minimize both the time to get a transfusion and to get to the hospital was a huge hurdle to overcome. Ideally, the rendezvous point should be in the general direction of the hospital where the ambulance is at no point driving in a different direction.
Creating an intuitive smooth UI experience: Organizing the front end to be as easy as possible for a distracted first responder to use during a time of crisis was very important. We made the UI as user-friendly and efficient as possible in the time we had: minimal prompting and simple graphics.
Accomplishments that we're proud of
1. Creating a working proof-of-concept: With this being our first major hackathon, at first we were worried if we could even could have a viable prototype with an ambitious project like this. We had to learn many new things quickly, and work efficiently as a two-person team to make quick progress. We're proud of being able to prove our concept.
2. Enjoying the creation of our product: We are glad that we had a lot of fun creating this project and in the process of creating this product, and that the meaning behind it kept us motivated.
What we learned
Having a problem-centered approach to brainstorming. Half the solution is understanding your problem.
What's next for BloodGrid
Real-world field testing would help us translate the benefits to real life and find gaps to address that we previously haven't.
Log in or sign up for Devpost to join the conversation.