Monday, May 4, 2015

New tool optimises positioning of emergency vehicles

Bochum: Where and when should emergency ambulances be positioned within a city to ensure that all locations across the urban area will be reached in good time? Economists are helping to solve this complicated question with IT-based decision support. Once the fire department’s control room receives an emergency call, every second counts: only eight to ten minutes are permitted to pass before the emergency services have to arrive on scene and launch rescue operations, no matter what the day of the week, what time of day or night, no matter if there is heavy traffic, no matter if it’s in the middle of the city or in remote outskirts. Keeping within this so-called response time is a challenge for emergency service planners, especially since monies are scarce and the number of rescue operations is on the increase. Approximately 23,000 rescue operations per annum have been served by Bochum’s 13 emergency vehicles in the recent years. Due to the rise in numbers of older people, the frequency of emergency calls is growing.

In order to develop better strategies and tactics to deal with these challenges, the Chair of Management, Operation Research and Accounting in Bochum headed by Prof Dr Brigitte Werners (fig. 1) has been collaborating with the fire brigade. “We have been working with them for the purpose of case-study seminars for years,” as Werners elaborates the background. “Dr Dirk Hagebölling, head of Bochum’s fire brigade, is very open-minded when it comes to new developments.” When the plan was hatched to use scientific research for designing an IT-based optimisation tool for emergency response planning, it exceeded the capacity of a student project: the institute got the “Stiftung Zukunft NRW” on board for its undertaking and was granted financial support for two years. Four members of staff were looking into the complex questions surrounding Bochum’s emergency response services.
“Essentially, it’s all about how many emergency vehicles have to be placed where at what time to meet the legal requirements as well as possible,” explains Prof Werners. In order to compile the relevant information, the project team used existing data of emergency responses in the city. They found out that the peak time for emergency responses is between 10am and 2pm. At night, far fewer emergency calls come in. Emergency call figures vary on different days of the week, too. They are much more frequent on weekends and on Wednesdays than on any other day of the week. Researchers expect that this phenomenon can be explained by the office hours of doctor’s surgeries. More emergency responses take place in the city centre than in the outskirts or in industrial estates. The time it takes an emergency vehicle to get to an injured or ill person very much depends on how fast it can go. Its speed, in turn, is determined by current traffic conditions, which vary strongly in the course of the day. Before six in the morning, they make good headway; during the day, the average travel speeds drop. All these data were fed into the researchers’ analysis, as well as the current and possible locations of emergency vehicles and information regarding the overall conditions in the city. In order to render the latter manageable, the researchers used the zoning plan of the entire urban area, which is subdivided into 1 x 1 km grid squares. The data analysis has shown that the inner city of Bochum is currently very well covered, excessively covered even: emergency vehicles reach any location within a short period of time; simultaneous emergency responses do not pose any difficulty, either. This is not the case in some of the outskirts, however (fig. 2).
“A crucial factor in the optimisation of emergency response services is the positioning of emergency vehicles,” explains Brigitte Werners. At certain times, it may make sense to position the emergency vehicles at so-called flexible stations, rather than at one of the three headquarters. “Flexible stations could be, for example, offices of the volunteer fire brigade, as well as hospitals and other municipal facilities, such as schools,” says Prof Werners. “When choosing a location, it’s important that it provides the necessary conditions for the emergency response team to disinfect the emergency vehicle after a rescue operation and that there are staffrooms and sanitary facilities.”
Using complex algorithms, the optimisation tool SPR2 (Strategic Planning of Resources for Emergency Medical Services) uses the input data to calculate the optimal locations for emergency vehicles in order to provide best possible coverage for the entire urban area, measured by the reachability rate. Specific local features, such as locations of headquarters, must be considered. The result is a detailed map outlining recommended potential locations for the emergency vehicles (fig. 2).
The tool is made up of several units and features a component which the respective planner can use to examine certain situations or changes in more detail. “Through simulation, we perform a countercheck of sorts on the optimisation,” describes Prof Werners. “It is also useful for monitoring the effects of certain changes, for example the relocation of a recommended flexible station to a different location, which may perhaps seem more appropriate for some reason.” (fig. 3)
Bochum’s fire brigade is now considering if and to what extent the changes recommended by the optimisation tool can be implemented across the city. “An undertaking like this naturally takes some time. We can integrate many factors into our tool, but not all of them,” says Prof Werners. “For example, we can’t include any internal aspects, such as specific additional duties of the employees.” Today, she travels a lot to present the tool in other cities across NRW as well. It meets with considerable interest: several NRW cities have forwarded their data from the recent years to the researchers in Bochum, in order to have their processes optimised, too. “It will take some time to complete the task,” presumes Brigitte Werners. “This is because, in order to make the data compatible with our system, we have to process them first. In addition, we have to clean them up if we find any errors, i.e. data which are just not plausible.”
In order to enable users to apply the system independently, additional programming has still to be conducted, for the purpose of which interested parties have to come together to finance the undertaking. “As scientists, this is not something we can do ourselves; our task is a different one,” explains the researcher. Her enthusiasm for the project is unabated: “The methods deployed in mandatory emergency services requirements planning are governed by specific rules”, she explains. “If it turns out that the new methods are better, it could result in these rules being changed – but that’s a long way away.”