This document was created during the 10th Annual NYUAD Hackathon as a Quantum Computer Solution for Social Good. The solution is to assign patients to hospitals based on their characteristics. The decision is influenced by many factors, including the distance between the patient's location and each hospital, the number of beds available at that hospital and the severity of the patient's condition.
The values in the list of equipment can be calculated based on the following levels medical need and severity:
- Primary care hospital:
- Basic Live support (BLS)
- Basic Live support (BLS) emergency
- Secondary referral hospital
- Advanced life support (ALS) level one
- Advanced life support (ALS) level one emergency
- Advanced life support (ALS) level two
- Tertiary referral hospital
- Specialty care transport
- Paramedic Intercept The different levels have to be normalized and added to the list of alpha values.
The world is not prepared when an emergency calamity hits. Data at hand shows that in every 15 years, there is a new pandemic. When covid-19 broke out, in Tokyo alone, 160,000 people were harmed because they were unable to be offloaded to A&E. In a very busy hospital, only 5 calls were received out of 200 calls made in one day because of hospital's unavailability. 10,000 calls were received by all hospital on average in Tokyo in one day. There was a need to map people to availble emergency health care center while also providing an optimal path. Our team looked into the problem to map emergency health care seekers to respective hospitals based on availability and necessary conditions for the patient.
Under the hood, the program is minimzing the cost function below so that an optimal hosptital can be matched for the patient in an emergency case:
Using classical computers, it is not feasible to minimize the cost function for a large number of patients given multiple different parameters. Our approach uses quantum annealing algorithm to minimize the cost function to find betas (matching) given the alphas: (calculated from various paramters like distance to hospital, equipment availability score and so on). A typical runflow interface looks like the following:
Image 1: Selecting number of patients and number of hospitals
After the program finishes running, the assignments are shown as below:
Image 2: Final output after algorithm runs
- Farai Mazhandu (Mentor)
- Ken Tanaka (Mentor)
- Cyril Karam
- Saimon Tsegai
- Nadeen Tarek
- Qasem Ahmad
- Omar Rayyan
- Edgar Palomino
- Dhurba Tripathi
This project was created at the 2022 NYUAD Hackathon for Social Good in the Arab World: Focusing on Quantum Computing (QC).