Skip to content

🌟 JunctionX Budapest Finalist Project: Varian's Challenge! πŸš€ Experience the thrill of innovation unleashed in just 48 hours. Feel free to check it out! πŸ‘€ πŸ’‘

otahina/ExcellaCare-Planner-Hackathon-JunctionX

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

11 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Header Image

πŸ₯ ExcellaCare can both save patients and doctors. πŸ§‘β€βš•οΈπŸ™†β€β™€οΈβœ¨

Reduce the amount of stress on doctors, save more patients' lives, and increase the quality of care with ExcellaCare.

Demo Highlights 🎬

record_github.mov

Frontend UI Features 🎨

React

All you need is button clicks! β˜‘οΈ

  • Time Table for each machine, doctors can see appointments at a glance. πŸ‘€
  • Recommended date and time are shown, just need to click and choose! Confirm and all done!
  • Emergency handling: Remove patient and make a space for an emergency patient.

Technical Details 🧰

Open Sesame πŸͺ„ For simplicity, we made few assumptions:
  • timetable only shows 5 days.
  • Treatment should be done every day constantly unless there is an emergency.

Backend Features πŸ€–

Python Flask

The power of prompt engineering βš™οΈ

  • Used GPT 3.5 turbo to process the data and show available time slots.

Technical Details 🧰

Open Sesame πŸͺ„

Prompt Example

data_to_be_known = ("TrueBeam, VitalBeam, Unique are the machine names for radiotherapy. Overweight patient can't "
                        "use VitalBeam. Underweight patient should prioritize VitalBeam. If the treatment has to be "
                        "done today, and there is no available date and time today, then return only machine name "
                        "which should be used.")

patient_condition = (f" Patient is {weight_type}, who has {cancer_type} cancer at {cancer_stage}. "
                         f"The treatment time per session is {treatment_session_time}.")

available_time_and_date_data = (f" Here is the available date and time for each machine: "
                                    f"{available_time_and_date_for_each_machine}")

instructions = (f"/Calendar_data finished/ Please find the 5 continuous days the machine can be used in the range, "
                    f"{the_range_of_date_the_patient_should_start}. Based on the condition, {machine_priority}, return "
                    f"the best machine to be used, and as many available time as possible per date. Make available date"
                    f"for all machines which can be used. The date should be continuous "
                    f"5 days, and each day, the treatment starting time can be +/-2 hours.")
                    
summary = data_to_be_known + patient_condition + available_time_and_date_data + instructions

Finally, summary is fed into GPT 3.5 turbo to generate the available time slots.

GPT response Example

Based on the condition given, the available machines for the treatment are VitalBeam and TrueBeam. Here are the available dates and times for each machine, with a window of +/- 2 hours for the treatment starting time:

VitalBeam:
2023-11-01: 11:00-11:15, 12:00-12:15, 13:30-13:45
2023-11-02: 11:00-11:15, 12:00-12:15, 12:30-12:45
2023-11-03: 10:00-10:15, 15:30-15:45
2023-11-04: 9:00-9:15, 9:30-9:45, 10:00-10:15
2023-11-05: 9:00-9:15, 11:00-11:15, 12:30-12:45

TrueBeam:
2023-11-01: 12:30-12:45, 13:30-13:45, 16:00-16:15
2023-11-02: 14:30-14:45, 13:30-13:45, 14:00-14:15
2023-11-03: 14:00-14:15, 11:00-11:15, 15:00-15:15
2023-11-04: 12:00-12:15, 12:30-12:45, 16:30-16:45, 16:00-16:15
2023-11-05: 13:00-13:15, 9:00-9:15, 15:30-15:45, 14:00-14:15

That is how we can retrieve the available time slots for each machine, and pass it to the frontend.

Next Steps πŸƒβ€β™€οΈ

It is better if the system can also recommend which patient should be postponed or cancelled.

Team Members πŸ‘©β€πŸ’»πŸ‘¨β€πŸ’»

Hina Ota

ZoltΓ‘n SinglΓ‘r

About

🌟 JunctionX Budapest Finalist Project: Varian's Challenge! πŸš€ Experience the thrill of innovation unleashed in just 48 hours. Feel free to check it out! πŸ‘€ πŸ’‘

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published