Dynamic Deployment

Exploring the predictability of Emergency calls

Historically, the positioning of resources in any emergency system is a highly manual decision by dispatchers aided by static post plans. With Dynamic Deployment predictive AI is used to send units into the vicinity of the next incident before it occurs.

This can be achieved by using demand forecast for the next 24-48 hours utilizing a combination of historical data and demand predictions, that are automatically adjusted based on changes in demand over time compared to the actual history. This means that seasonal variations, new settlements and other changes in data over time are considered automatically.

Using AI for demand prediction to narrow down what will happen in the next 40 minutes has shown, that it is possible to pinpoint future incidents in more than 80% of the time to a location within a quarter mile radius and within 3.5 minutes of those calls happening. Applying unit coverage and business rules around how units can be deployed allows for automated posting of units close to future incidents, which in turn decreases response time and ultimately improves patient outcomes.

Increase coverage to avoid level 0 system status

Another area in which predictive AI provides big advantages is in optimizing the utilization of avaiable units. An EMS organization may often have periods of time where the total volume of calls exceeds the number of available units. In those cases there really are not many options, but what if you could use data to reduce the frequency and length of those level 0 situations?

Since implementing Logis’ Dynamic Deployment innovative resource posting feature, we have seen our Med 1 response times improve by an average of 6-8%.

Katie Arens,

Vice President of Customer Access,

Life EMS Ambulance

Using predictive AI to both gain insight into where those near future calls will happen and business rules around unit posting has shown that this is in fact possible. Placing units geographically close to the future calls at the right time will reduce the response time to those calls and hence also reduce the time on task – in effect resulting in more available unit hours.

As shown in the illustration there is a peak demand for emergency services during the morning and afternoon hours, which is illustrated by a decrease in the number of available units. When availability drops, there is normally a decreased coverage of the geographical area serviced, which in turn leads to longer response times.
In the graph the expected average response time using Dynamic Deployment is shown against the average response time without Dynamic Deployment.

Naturally, Dynamic Deployment has the most impact when resource availability is high, which is shown in the graph as a relatively larger difference between normal and Dynamic Deployment enabled response times. When there are very few or no available resources, then Dynamic Deployment does not impact response times, as there are no units to move.

Note, that while Dynamic Deployment does not have a significant impact when there is a very low level of available resources, the benefit is still that the time period where the system experiences low availability is shortened. The reason for this is that when response times decrease then the total time on task also decreases which again increases available unit hours as previously mentioned.

To learn more, download our free whitepaper on Dynamic Deployment