Please note: All research in progress seminars are off-the-record by default. Any information about methodology and/or results are embargoed until publication.
Over the last several decades, Emergency Medical Service (EMS) has become an important component of health care service. The main performance indicator in the EMS setting is the response time, i.e. the time to reach the patient once an ambulance is requested. Policy makers adopt a response time criteria to set the standards of this service, and the push to reduce it is justified by the assumed link between longer response time and worse health outcomes. However, current literature finds weak to no relation and this knowledge gap has been recently attributed to the endogeneity of response time. Indeed, the ambulance driver may take actions that result in shorter responses for most critical cases, and this unobserved behavior creates a downward bias in the results up to the point of finding zero effect. In line with previous literature, my analysis is performed on patients affected by cardiovascular disease, i.e. the time sensitive pathology adopted by policy makers to set the EMS standards in terms of response times and the main cause of death in developed countries. In my work I exploit changes in the amount of hourly rainfall and rationalization of emergency personnel during night shifts (i.e. 8pm to 7am) as instruments for response time. I document that a minute increase in response time results in a 2% increase in the probability of highly severe health conditions at the ambulance arrival on the scene and by 0.4% rise in the probability of death by the arrival at the hospital. Finally, I discuss and rank alternative solutions that may be implemented by policy makers to improve EMS performances.