Use of discrete event simulation and genetic algorithms to estimate the necessary resources to respond in a timely manner in the emergency medical services in Bogotá

Introduction Bogotá has public and private emergency medical services that respond to health incidents. However, its sufficiency in terms of quantity, type, and location of demanded resources is unknown. Objectives Using data from the emergency medical services of Bogotá, Colombia, we sought to char-acterize the pre-hospital response to cardiac arrest and determine the model with the least number of resources needed to respond within eight minutes, considering their location, frequency, and type. Methods A database of incidents reported in administrative records of the district health authority of Bogotá (from 2014 to 2017) was obtained. Built on this information, a hybrid model based on discrete event simulation and genetic algorithms was designed to establish the amount, type, and geographical location of resources, according to frequencies and type of events. Results Bogotá had 938 671 ambulance dispatches in the studied period. Of these, 47.4% were high priority, 18.9% medium priority, 33.74% low priority, and 92% corresponded to 15 of the 43 medical emergency codes. Response times were longer than expected, especially for out-of-hospital cardiac arrest (median 19 minutes). In the proposed model, the best scenar-io required at least 281 medical and basic ambulances in a 3:1 ratio to respond in adequate time. Conclusions The results suggest the need to increase the resources of the emergency medical services to bring response time closer to the needs of our population.