Title of talk
Where to place automated external defibrillators?: Using data analytics and mathematical optimization to guide location decisions
Out-of-hospital cardiac arrest is a significant public health issue, and treatment, namely, cardiopulmonary resuscitation and defibrillation, is very time sensitive. Public access defibrillation programs reduce the time to defibrillation and improve cardiac arrest survival rates. However deciding where to put automated external defibrillators (AEDs) is a complicated problem. Research often looks at where the cardiac arrests happened before and determines AED locations based on that. We claim that historical data should not be used blindly to guide AED deployment, but should be used to estimate the cardiac arrest risk. In this talk, we explore a statistical smoothing technique, called kernel density estimation, to derive a risk map of cardiac arrests in Toronto. We suggest ways to incorporate this tool to simulate future cardiac arrests, and develop mathematical optimization models to guide the deployment of public AEDs. Lastly, we evaluate deployment policies using simulated cardiac arrest data sets and show that the optimization outperforms the existing approach.