Proposal for master thesis at Norwegian Open AI-Lab, NTNU, from Norwegian National Advisory Unit for Prehospital Emergency Medicine (NAKOS), and Department of Emergency Medial Communication Centre (EMCC), Division of Prehospital Services, Oslo University Hospital.
External contact persons: Professor Jo Kramer-Johansen (NAKOS), Håvard Wahl Kongsgård (NAKOS)
In emergency medicine, time is of the essence. Acute conditions like stroke, myocardial infarction, major trauma, and cardiac arrest, result in loss of vital functions and life within minutes. EMCC is the key to rapid first aid when the public calls them for help by providing first aid instructions, but definitive treatment requires arrival of qualified personnel and often, rapid transport to hospital.
However, most cases handled by the EMCC, are of less acuity, and to aid call takers at the EMCC to decide and prioritize, they use a paper-based triage-tool. Both acute and less acute cases may need ambulance transport, and usage of available resources are high and growing. When an ambulance
transports a patient, it is not available for other missions, so usage of ambulance resources must balance the current load of missions and availability for the next (unknown) acute case. The current solution involves specially trained and experienced dispatchers (Resource Coordinators – RC) in the EMCC that manually assigns each mission to an ambulance, maintaining the delicate balance of workload and contingency within geographic regions.
EMCC in Oslo, is the largest in Norway and answers calls to the medical emergency phone number 113 from a population of 1.6 million (Oslo, Akershus, and Østfold). The available resources includes 29 (night) to 45 (day) ambulances based at 15 ambulance stations. Each year EMCC handles 500 000 telephone calls, and the Ambulance department executes more than 150 000 missions each year.
The existing data set is an extract from the clinical data system used in EMCC to record each mission – (Akuttmedisinsk informasjonssystem, AMIS). Each mission has a position that has been mapped to a standardized 1000m by 1000m grid from SSB (Statistics Norway). We will anonymize the dataset, but
the SSB-grid ties each mission to more sociodemographic information, as well as historical information such as traffic, weather and climate, public events, and moveable public holidays. Each observation is per grid per hour and includes; number of events, response intervals for both acute and non-acute missions, and idle-time.
Example data structure:
Use of data – dynamic resource placement and prediction
This is a project in collaboration with an external partner. If you choose this project, then I will serve as the responsible from NTNUs side, but the actual work will also be in tight collaboration with personell from the external partner as listed above.
If you consider picking a project with me as the supervisor, then please look at www.idi.ntnu.no/~helgel/teaching/proposals/.