Filip Gnesin - PhD Scholarship 2022
Project summary:
Detecting acute heart disease outside the hospital
Although chest pain and breathing problems are common symptoms reported in emergency calls, many still die from acute heart disease. This project will use health data from the emergency calls, the ambulances and Danish national registries to predict whether a patient calling with these symptoms suffers from acute heart disease. This could help dispatchers determine over the phone which patients should be prioritized and the ambulance crew to start early treatment.
Project title
Predicting acute cardiovascular disease in a pre-hospital setting
Background
In Denmark, the potential to increase and improve pre-hospital cardiovascular care is high. In a recent study of patients receiving an ambulance by the Emergency Medical Services (EMS) in the North Region of Denmark, we showed that more than 50% of patients reporting symptoms of a presumed cardiac condition as evaluated by EMS and more than 25% of those reporting dyspnoea later received a cardiovascular hospital diagnosis, but only a small proportion of these patients had received specialised care prior to hospitalisation.
Aim
The aim of the current project is to examine whether data collected by ambulance operators can be used to predict the cardiovascular origin of chest pain and dyspnoea in a pre-hospital setting. This observational study is based on EMSs across Danish regions. Data comes from electronic records in relation to emergency calls, ambulance electronic patient files and national registries.
Methods
Based on data collected in ambulances including electrocardiograms (ECG), we will focus on four main studies: predicting myocardial infarction, acute heart failure, cardiac arrhythmia and pulmonary differential diagnoses. We expect to have a range of parameters from ambulance personnel covering >4 million calls and >400,000 ECG recordings covering the whole nation.
The main statistical methods are based on machine learning prediction using the highly flexible methods from ’Super learner’. A clinical risk score system may focus current specialised pre-hospital treatment to patients with poor prognosis and guide future interventions.
Filip Gnesin
- MD
- University of Copenhagen and Nordsjællands Hospital Hillerød, Department of Cardiology
Main supervisor:
Professor Christian Torp-Pedersen, MD Department of Cardiology, Nordsjællands Hospital Hillerød
Co-supervisor:
Professor Fredrik Folke, Copenhagen University and Senior Consultant at Department of Cardiology, Gentofte Hospital