Improving AED availability:
A framework to support decision-making
on where to develop AED drone networks
Thesis
University of Groningen
Faculty of Economics and Business
MSc Technology and Operations Management
Christian Westendorp C.F.Westendorp@student.rug.nl
S3224066
Supervisor/ university dr. ir. D.J. van der Zee
Co-assessor/ university dr. N. Szirbik
Supervisor/ field of study ir. J. Hatenboer UMCG Ambulancezorg
Abstract
Background – Out-of-Hospital Cardiac Arrests (OHCA) are a leading cause of death in Europe. Time matters in the treatment of an OHCA, i.e., the sooner care is provided, the better patient outcomes will be. Especially, early defibrillation improves the survival rate of a patient. Defibrillation can be achieved by using a so-called Automated External Defibrillator (AED) to restore the heart rhythm. AEDs can be found in public locations such as train stations and stores. However, most of the OHCAs occur in private settings, where there is no nearby AED available. In addition, availability of AED is often restricted to the opening hours of the building. Especially in rural areas, AED coverage is difficult to achieve. A possible solution is the use of AED drones. However, there is a gap in the literature on where to use AED drones.
Objective – The aim of this study is to develop and evaluate a framework that supports efficient decision-making on where to use AED drones.
Methods – A design science research method has been applied to develop this framework. First, the literature is reviewed. Second, static AED and AED drone networks are compared in terms of effectiveness and costs. In addition, determinants for successful operation of AED drone networks are identified. Next, the framework is designed based on information from domain experts and the previously identified determinants. Finally, the framework has been applied and evaluated based on a pilot area.
Results – The framework is fed with regional characteristics and decisions. The framework consists of five different steps in order to arrive at the regions with the highest potential for the implementation of an AED drone network. The ground step is to select the starting candidate regions by setting the boundaries and the level of detail. The first step of the framework is to identify regions that are not being covered by the current AED infrastructure, or where the current AED infrastructure performs bad. The second step is to assess the regions based on network characteristics in order research if AED drone operations are possible. Thirdly, an initial cost comparison is made for static AEDs vs. AED drones. Fourthly, the network is optimized based on patient demand and high-risk areas. Finally, the solution is assessed and tested. The output of the framework are regions that are have the high-potential for the implantation of an AED drone network. The framework provides key elements and help supporting the decision-making process.
Conclusion – The provided framework helps the client to select regions that have the highest potential for the implementation of an AED drone network. The framework guides the process and decision-making.
Preface
Before you lies the master thesis “Improving AED availability: A framework to support decision-making on where to develop AED drone networks”. This thesis concludes the final step in completing the master program: Technology and Operations Management.
The world of drones is relatively new, and therefore, challenging. A large amount of information is still missing and often the available data consists of working documents. However, the experience in researching the design of an AED drone network was very rewarding. The people researching this field are very warm and enthusiastic.
I would like to thank ir. J. Hatenboer from UMCG Ambulancezorg for his supervision and guidance throughout this process. His enthusiasm about (AED) drones worked contagious. Furthermore, I would like to thank Mr. T. Schrijnemaekers from HartslagNu & Ambulancezorg Limburg-Noord for providing me with the quantitative data. Next, I would like to thank Mr. R. Heidekamp from ANWB Medical Air Assistance for his time explaining the airspace structure and the regulations. Also, I would like to express my gratitude to all other people whom were involved in this research.
I would also like to thank my supervisor from the university, dr. ir. D.J. Van der Zee, for his constructive feedback during this semester. The feedback sessions always provided me with new insights and feedback to improve my work. In addition, I would like to thank my second assessor, dr. N. Szirbik, for his time and his feedback.
Finally, I would like to thank my parents, brother, and friends for their support during the master thesis project and throughout my whole study career.
Groningen, January 2019
Table of content
Abstract ... 3 Preface ... 5 List of abbreviations ... 9 1 Introduction ... 11 2 Theoretical background ... 132.1 Out-of-Hospital Cardiac Arrest (OHCA) ... 13
2.2 Factors that influence the risk of an OHCA ... 13
2.3 OHCA treatment- chain of survival ... 14
2.4 OHCA pathway ... 15
2.5 Improving AED availability: AED drone networks ... 16
2.5.1 Static AEDs ... 16 2.5.2 Drones ... 16 2.5.3 AED drones ... 16 2.6 Summary of findings ... 17 3 Methodology ... 18 3.1 Problem background ... 18 3.2 Research objective ... 18 3.3 Conceptual model ... 19 3.4 Research outline ... 20 3.5 Data sources ... 21
4 AED drone networks ... 22
4.1 AED networks ... 22
4.1.1 Static AED networks ... 22
4.1.2 AED drone networks ... 23
4.2 Assessing AED networks ... 25
4.2.1 Effectiveness ... 25
4.2.2 Costs ... 26
4.2.3 Summary of findings ... 27
5 Determinants for successful operation of AED drone networks ... 28
5.1 Determinants identified in related fields ... 28
5.1.1 Cell tower placement ... 28
5.1.2 Helicopter Emergency Medical Services ... 29
5.1.3 Ambulance care ... 30
5.1.4 Summary of findings ... 30
5.2 Determinants for selecting AED drone regions ... 31
5.2.1 Laws and regulation ... 31
5.2.2 Geography ... 31
5.2.3 Operational considerations ... 31
5.2.4 Population: patient demand ... 32
6 Framework design ... 33
6.1 Framework design approach ... 33
6.2 Framework input ... 35
6.3 Framework ... 36
6.3.1 Use of the framework ... 36
6.3.2 Detailed description of the framework ... 36
6.4 Framework output ... 40
7 Application and evaluation of framework ... 41
7.1 Evaluation approach ... 41
7.2 Framework input ... 41
7.3 Framework ... 42
7.4 Framework output ... 50
7.5 Summary of findings and evaluation ... 51
8 Discussion ... 52 8.1 Main findings ... 52 8.2 Theoretical contribution ... 53 8.3 Managerial contribution ... 53 8.4 Limitations ... 53 9 Conclusion ... 55 9.1 Conclusion ... 55 9.2 Future research ... 55 References ... 56
Appendix A: Causes of an OHCA ... 64
Appendix B: Survival probability rate ... 65
Appendix C: AED drone radius ... 66
Appendix D: Cost of ownership for an AED ... 67
Appendix E: Investment and maintenance costs for a static AED ... 68
Appendix F: AED drone costs ... 69
Appendix G: Costs comparison between static AEDs and AED drones ... 71
Appendix H: Linear programming model ... 72
Appendix I: Number of inhabitants per zip code ... 73
Appendix J: Detailed application framework step 2 ... 74
Appendix K: Overview of drone flying map ... 76
Appendix L: Number of AEDs per zip code ... 77
List of abbreviations
AED Automated External DefibrillationALS Advanced Life Support
BLS Basic Life Support
BVLOS Beyond Visual Line-of-Sight CPR Cardiopulmonary resuscitation
EMS Emergency Medical Services
GIS Geographic Information System
HEMS Helicopter Emergency Medical Services OHCA Out-of-Hospital Cardiac Arrest
RIVM Dutch National Institute for Public Health and the Environment
SES Socio-Economic Status
UAV Unmanned Aerial Vehicle
VLL Very Low Level
1 Introduction
In 2016 around 34,000 people were hospitalized in the Netherlands after an Out-of-Hospital Cardiac Arrest (OHCA). In addition, every day around 14 people die from the consequences of a cardiac arrest (Hartstichting, 2018b). Time matters in the treatment of a cardiac arrest. The health outcomes for a patient rapidly decrease as time progresses. With every minute that passes the survival rate decreases with 10 percent (American Heart Association, 2000). Early recognition and treatment improves the outcomes of the chain of survival (Helsen, Vanbrabant, & Dewolf, 2016).
OHCA treatment entails four sequential steps, the so-called chain of survival: early recognition, cardiopulmonary resuscitation (CPR), early Automatic External Defibrillation using AEDs that administer a shock to the patient, and Advanced Life Support (ALS) provided by an emergency medical service (EMS) (Bruining, Lauwers, & Thijs, 2017; Helsen et al., 2016). A study of Ringh et al. (2015) showed that especially early defibrillation can improve the health outcomes of a patient. Hartslagnu (2018) has started several campaigns to improve availability of AEDs by increasing the number of AEDs in public locations and companies to improve the chain of survival in terms of early CPR and early defibrillation.
However, early defibrillation is difficult to achieve because most of the OHCAs occur in private locations. Private locations may be difficult to reach by ambulances in time. This is especially true for rural areas, where in contrast to urban areas, public AEDs are less available (Cooper, Swor, Jackson, & Chu, 1998). Eisenburger et al. (2006) states that patients having a cardiac arrest in a non-public location are more significantly more prone to die or have severe brain damage. To overcome the problems in rural areas, Momont (2014) proposes the idea that AED drones can be used to save time and improve patient outcomes. The AED will be transported to the patient by a drone. There are several advantages in the use of AED drones: the time to defibrillation is decreased and the utilization of drone AEDs is higher compared to the static, public, AEDs (Boutilier et al., 2017; Claesson et al., 2017). Current dilemmas hindering their implementation are: where to deploy drones most efficiently and where not, and how to integrate AED drones in current networks. For example, are AED drones providing an alternative for public AEDs or do they extend the network of public AEDs?
efficient to use drones in addition to current static AED networks. It will help in selecting candidate regions and testing them for the potential of AED drones, based on determinants provided in the framework.
The four phases of Wieringa (2014) will be applied to design and evaluate the framework. The phases concern: literature review, characterization of regional factors that influences the implementation of AED drones, design and evaluation of the framework, and generalizing the findings in terms of theoretical contributions.
2 Theoretical background
In this chapter, the current literature is reviewed. Chapter 2 consists of four different sections. First, the out-of-hospital cardiac arrest (OHCA) is explained. Second, factors that influence the risk of an OHCA are discussed. Third, essentials steps in OHCA treatment, i.e., the so-called chain of survival are examined. Fourth, the OHCA care path is briefly introduced. Finally, the possibilities and usages of (AED) drones for supporting the chain-of-survival are explained.
2.1 Out-of-Hospital Cardiac Arrest (OHCA)
A cardiac arrest could be defined as the stopping of any cardiac activity. It can be proven by the absence of signs of blood circulation (Jacobs et al., 2004). Consequences of a cardiac arrest are: loss of heart function, loss of breathing, and loss of consciousness (Mayo Clinic, 2018). Therefore, an OHCA is a cardiac arrest outside of a hospital setting. A study of Sasson, Rogers, Dahl, & Kellermann (2010) showed that the survival rate of an OHCA has not significantly improved over 30 years. The survival rate is between 6.7% - 8.4% measured across different populations over a time span of 30 years (Sasson et al., 2010).
A cardiac arrest can have different causes (see Appendix A). The most common type of sudden cardiac arrest is an acute myocardial ischemia, where the flow of blood to the heart is reduced. This will result in a cycle of fast and chaotic contraction of the ventricles and this impacts the blood circulation (Bruining et al., 2017; Hartstichting, 2018c). The best way to treat a cardiac arrest is defibrillation, nonetheless, not all heart rhythms are shockable. This is depending on the electrical activity of the ventricles (Sayre et al., 2010). Defibrillation uses electric impulses to bring the ventricles back to a normal rhythm, so the heart can pump blood (Jones & Lodé, 2007).
2.2 Factors that influence the risk of an OHCA
There are several factors that influence the risk of an OHCA. The Socio-Economic Status (SES), ethnicity, and age, influence the risk of an OHCA.
Socio-economic factors
The Social Economic Status (SES) is a construct that combines the social and economic position of an individual within a certain society. It consists of indicators such as: education level, income, and area of residence (Baker, 2014). Studies have shown that the individual SES level is related to the health and mortality risk (Mackenbach et al., 2008; Stringhini et al., 2017).
Ethnicity
and inhabitants with a Moroccan background have a lower risk for a heart infarction compared to inhabitants with a Dutch background (Knol, 2012).
Age
In addition to socio-economic factors and ethnicity, age is also a determinant for the OHCA risk in an area. A study of McNally et al. (2011) showed that the average age of OHCA patients is 64 years old, based on an analysis of 40,000 patients. The survival rate of a patient is also related to the age. The chance of survival decreases with an increasing age of the patient (Herlitz, Eek, Engdahl, Holmberg, & Holmberg, 2003).
2.3 OHCA treatment- chain of survival
The treatment of an OHCA is linked to the chain of survival that consists of four sequential steps: early recognition, early CPR, early defibrillation, and early advanced care (figure 2.1). The survival rate of a patient is dependent on circumstances (place and age) and factors such as the timeliness of the treatment (Bruining et al., 2017; Helsen et al., 2016) (See Appendix B for the survival probability rate).
Figure 2.1 Chain of survival based on Perkins et al. (2015)
i. Early recognition and call of help
The first step emphasises the prevention and recognition of a cardiac arrest. The early recognition can trigger adequate medical treatment (Helsen et al., 2016), see ii-iv. Important symptoms of a cardiac arrest are: no heart rhythm, no breathing, and unconsciousness (Mayo Clinic, 2018).
ii. Early CPR
The second step is focused on the early start of CPR. Starting CPR straight away can double or triple the survival rate of a patient suffering from a cardiac arrest (Helsen et al., 2016). CPR is part of the Basic Life Support (BLS) and includes keeping the airways unobstructed and support the breathing and circulation without technical devices (Bruining et al., 2017).
Prevention
Early recognition and call for help
Buying time
Early CPR Restart the
heart
Early
defbriliation Quality of life Early
CPR consists of a series of chest compressions and ventilations to continue the blood circulation in the body. In addition, CPR will create time for the defibrillation and arrival of the EMS (Perkins et al., 2015).
iii. Early defibrillation
The third step emphasizes the significance of early defibrillation in case of a shockable heart rhythm. If CPR and defibrillation are performed within 3-5 minutes after the occurrence of a cardiac arrest, the survival rate can increase to 49-75% (Helsen et al., 2016). An AED consists of two pads which can be placed on the chest of the patient and gives a shock to the patient in order to restore the heart rhythm. Furthermore, an AED can indicate whether a heart rhythm is shockable or not shockable (Perkins et al., 2015).
iv. Early Advanced Life Support (ALS)
The final step is early advanced life support (ALS) and standardised post-cardiac arrest care. Often ALS is performed by an EMS service and also includes the administering of medication and transportation of the patient (Helsen et al., 2016).
2.4 OHCA pathway
The OHCA pathway consists of three major steps: telephonic care, pre-hospital care, and hospital care
The telephonic are consists of the dispatch of EMS, the activation of the volunteer/ lay rescuer network, and giving instructions to the caller. After the telephonic care, the pre-hospital care is provided. There are four different scenarios in providing the pre-hospital care (table 2.1).
Scenario Parties involved
1 Volunteer CPR + volunteer AED + EMS 2 Volunteer CPR + EMS
3 Volunteer AED + EMS
4 EMS
Table 2.1 Different scenarios in the pre-hospital care
Hospital care consists the admission of a patient. The patient will receive treatment in the hospital and will be discharged later.
2.5 Improving AED availability: AED drone networks
The concept for volunteers and static AEDs is designed for residential areas, and is therefore, less applicable for rural areas (Zijlstra et al., 2014). To overcome the challenges of static AEDs, in terms of availability and accessibility, Momont (2014) proposes to use a drone to rapidly deliver an AED to the OHCA location. This section is structured in the following way. First, the current AED infrastructure is discussed. Next, the dilemmas faced concerning the current AED infrastructure are sketched. Thirdly, the opportunities of AED drones are explained. Finally, it is explained where AED drones could be used.
2.5.1 Static AEDs
As clarified in chain of survival (section 2.3), early defibrillation can improve the survival rate of a patient with an OHCA. To improve the early defibrillation several public access defibrillation programs have been started (Zijlstra et al., 2014). The goal of these programs was to improve availability of AEDs by placing them at densely populated areas and high-risk sites. An example is the concept of the Dutch Heart Foundation, where volunteers receive a notification when somebody has an OHCA in their proximity (Hartstichting, 2018d). However, most OHCAs occur in non-public areas, where onsite AEDs are unavailable (Eisenburger et al., 2006).
The challenges faced concerning the availability of AEDs is the access and location. A news article of the NOS (2018) showed that one third of the Dutch people does not know where the nearest AED is in their area. In addition, most AEDs are not accessible 24 hours per day. For example, in the province of Zuid-Holland only half of the AEDs is accessible 24/7 (NOS, 2018).
2.5.2 Drones
Drones or Unmanned Aerial Vehicles (UAVs) are described as an aerial vehicle that does not carry a human and flies remotely controlled or autonomously (Gupta, Ghonge, & Jawandhiya, 2013). Drones are often controlled from a ground station. Drones need to be monitored, controlled, and the surroundings need to be controlled. There are two main types of control for an AED drone: Visual Line of Sight (VLOS) and Beyond Visual Line of Sight (BVLOS). With VLOS the remote pilot must always see the UAV. This means that the UAV cannot fly, for example, behind buildings or clouds (Gupta et al., 2013; Kopardekar et al., 2016).
2.5.3 AED drones
et al., 2017). The paper of Van de Voorde et al. (2017) extends the concept of Momont (2014) in terms of practical issues and proposes areas for further research. The paper of Claesson et al. (2017) compares drone flight times versus EMS response times. In all simulated cases the drone is faster than ambulances.
Boutilier et al. (2017); Claesson et al. (2016); Pulver et al. (2016) focus on the design of (AED) drone networks. The provided mathematical models are simplified by making assumptions. In the paper of Boutilier et al. (2017) one of the assumptions is that drones are only placed on EMS stations. In contrast, Pulver et al. (2016) proposes a model where the launch sites are placed on existing buildings (new locations), EMS locations, and a combination of both.
Finally, the model of Claesson et al. (2016) uses Geographical Information Systems (GIS) to place drones stations in Sweden. In this model, both rural and urban areas are covered by drones.
The literature on AED drones and the design of a network is lacking support in the decision-making of the selection of regions. There is no framework to guide the decision-decision-making to find the optimal regions to use AED drones. In addition, the integration of the current static AED network and drone-based networks is ignored in all the papers. A hybrid model between static AEDs and AED drones could be beneficial because it is difficult to operate a drone in densely populated areas.
2.6 Summary of findings
For the treatment of an OHCA time matters. The sooner care is provided, the better the patient outcomes will be. Especially early defibrillation is important in the survival chances of a patient. However, the low availability of static AEDs and the distances in rural areas make it difficult to start defibrillation within six minutes. In addition, static AEDs are often restricted to the opening hours of the buildings they are placed in.
3 Methodology
This chapter provides an overview of the research design of this study. Section 3.1 provides a description of the background of this research. Next, the research objective and deliverables are stated in section 3.2. Thereafter, section 3.3 provides the conceptual model that gives a graphical overview of the concept of AED drones. Furthermore, the research set-up is provided in section 3.4. Finally, the data sources are discussed in section 3.5.
3.1 Problem background
Timely defibrillation of an Out-of-Hospital Cardiac Arrest is critical. The sooner the patient is treated, the better the patient outcomes will be. EMS providers have difficulties reaching OHCA patients in rural areas in time. This results in a lower survival rate for OHCA patients in rural areas compared to urban areas (Eisenburger et al., 2006; Fredman et al., 2016). In addition, the major difficulty in the early defibrillation is the accessibility of AEDs. Most publicly accessible AEDs are subjected to the opening hours of buildings or require long walks. This is especially true for rural areas, where there are less publicly accessible AEDs and distances are longer (Van de Voorde et al., 2017).
To overcome these problems with the delivery of AEDs, Momont (2014) proposed to use drones as an AED carrier. Several studies have shown that the response time can be decreased, when using AED drones.
The National Institute for Public Health and the Environment (RIVM) sees that AED drones are an opportunity to improve the health outcomes. However, there should be a solid business case before regulations can be changed to make AED drone operations possible. Therefore, UMCG Ambulancezorg has put forward the question on: “where to use AED drones?”
3.2 Research objective This research has the following objective:
To develop and evaluate a framework that supports efficient decision-making on where to use AED drones.
Research deliverables:
• A characterization of factors that influence the decision-making to develop an AED drone network in a region;
• A validated framework to determine suitable areas for AED drones; • A pilot to illustrate and evaluate the framework
3.3 Conceptual model
Figure 3.1 depicts the conceptual model of the OHCA pathway, chain of survival, and the performance indicators. The chain of survival refers to the series of care services (section 2.3). The OHCA pathway refers to the care activities (section 2.4). The input consists of a patient with an OHCA. And the process is initiated by a phone call to the EMS dispatch center. In the pre-hospital care, the chain of survival is initiated. Typically, an ambulance is dispatched to provide ALS. In addition, volunteers in the proximity receive a text message and can provide CPR and defibrillation by means of a static AED. Alternatively, an AED drone can be dispatched to the location of the OHCA patient. The use of static AEDs or AED drones is influenced by regional characteristics such as: geography, population, distances, locations, and the coverage of EMS services.
3.4 Research outline
The research design is based on the design science method of Wieringa (2014). The method of Wierenga (2014) ensures the generalisation of theory and develops an artefact (framework) in order to design a future system. In contrast to surveys, analytical quantitative research, and a case study, with design science an artefact is created. The design method proposed by Wierenga is implemented as a four-phase approach: reviewing current literature, regional AED drone characterisation, framework design and evaluation, and the generalization of the findings (table 3.1).
Research phase Thesis chapter
1. Literature study 2. Theoretical background
2. Regional AED drone network characterisation
4. AED drone networks
5. Determinants for successful operation of AED drone networks
3. Framework design and evaluation 6. Framework design
7. Application and evaluation of framework
4. Theoretical contribution 8. Discussion
9. Conclusion
Table 3.1 Research outline
Phase 1: Literature study
In the first phase, the current literature is reviewed. The goal is to introduce subject of OHCAs, the treatment, the pathway, and the dilemma’s faced concerning AED availability. Next, the topic of AED carriers is elaborated and, finally, the AED drone network design is explained. Since AED drones is a new research subject, literature about determinants in the decision-making in related fields is reviewed. The goal is to find similarities in determinants and decision-making that could also be used for the AED drone placement. For example, determinants in the placement of ground ambulances, Helicopter Emergency Medical Services (HEMS), and designing cellular coverage networks. Those sectors have similarities with the placement of AED drones in terms of: coverage of networks, health decisions (for example, cost-effectiveness), and providing care.
Phase 2: Regional AED drone network characterization
In the second phase, the characteristics of the hybrid AED drone network are further elaborated and described in detail. The actual decision-making framework will be based on the regional characteristics. Factors that will influence the decision-making are included in the final framework. Sources in this phase include interviews with domain experts:
• In drone operations; • In healthcare operations;
• In Emergency Medical Services;
The domain experts offer support in validating the characteristics for AED drone networks and offer support to build the framework.
Phase 3: Framework design and evaluation
The framework will be designed based on the input of the previous phases. It will consist of different steps where the most potential locations are selected, adjustments to the network are made, and decision-moments. The framework will be build based on literature, interviews, and documents. In addition, there will be interviews and feedback moments with domain experts to evaluate and validate the framework.
The outcome of the framework can support a decision to place an AED drone station or drone network in a certain region. The framework will be illustrated and evaluated by a small pilot in a Dutch region.
Phase 4: Theoretical contribution
The final phases include generalizing the findings and elaborating the theoretical contribution of the thesis. The framework will be evaluated from a theoretical perspective and this will yield new knowledge that can be added to the theory about AED drones. The discussion will show the theoretical relevance, generalize findings, and shows the theoretical contribution of this research.
3.5 Data sources
4 AED drone networks
This chapter will discuss the factors that will influence decision making on the implementation of regional AED drone networks. First, AED networks are explained in terms of their key elements, i.e. procedures, resources and operations. Finally, networks based on static AEDs and AED drones are compared for effectiveness and costs.
4.1 AED networks
Two types of AED carriers can be distinguished: volunteers and AED drones. The former refers to “static” AEDs that are placed around strategic locations in, for example, gyms or universities. However, these AED needs to be picked up and brought to the patient by a care provider or volunteer. AED drones can bring the AED to the patient scene, without needing such support.
4.1.1 Static AED networks
Static AEDs should be placed on strategic locations to ensure a high coverage. One of the earliest techniques to place AEDs strategically was to analyse locations with a high OHCA- risk (Becker, Eisenberg, Fahrenbruch, & Cobb, 1998). Locations with a high OHCA risk are typically: airports, shopping malls, universities, and banks. Therefore, AEDs are placed in these areas (Brooks, Hsu, Tang, Jeyakumar, & Chan, 2013; Sun, Brooks, Morrison, & Chan, 2017). Furthermore, EMS providers also carry AEDs in their vehicles, for example, ambulances.
However, most OHCAs occur outside of buildings (Fredman et al., 2016). Static AEDs are often placed in urban areas where the population is clustered. It is more difficult to place AEDs in rural areas, since, the population is scattered over a large area. Another method to choose locations is to use GIS to analyse the distributions of OHCAs and then identify high risk areas based on the historical data (Chrisinger et al., 2016; Mao & Ong, 2016). Cooper et al. (1998) investigated certain high-risk areas and found that only three locations had more than one OHCA in six and half years out of 284 total OHCA cases. The high-risk areas were a golf course, a young men’s Christian association, and a health club.
4.1.2 AED drone networks
The framework of Ruijfrok (2018) divided the AED drone network into three different elements: operational processes, network resources, and the control function. The different elements are explained below and applied to the regional selection when possible. For example, the AED drone operations are still the same, but in the resources and control function some additions have been made.
4.1.2.1 AED drone resources
AED drone resources are physical assets that support the operational processes. The different resources are: the drone, the drone stations, and AED.
Drone
The area will eventually determine the type of drone that is needed to carry out the drone operations. For AED drones there are three different options: a quadcopter, a fixed wing drone, and a hybridcopter (which is a combination of both). The drone needs to be as safe as possible, for example, there needs to be a good functioning autopilot and the kinetic energy needs to be minimized.
Drone stations
The stations are used to store and launch the drone and the AED attached to the drone. There are several forms of stations. For example, the drone can be stored in a box that can function as a launch pad.
AEDs
An AED is a device that can be carried and can administer a shock to a patient with an OHCA in order to restore the heart rhythm. In the case of AED drones, the AED will be attached to the drone and be flown to the location of the OHCA patient.
4.1.2.2 AED drone operations
Figure 4.1 Operation process of an AED drone (Ruijfrok, 2018)
1. Recognition of OHCA: In this phase a bystander witnesses an OHCA and contacts the dispatch and control centre.
2. Dispatch of drone: The second step the dispatch and control centres assess whether an AED drone is required.
3. Prepare the flight: Based on the selected AED drone station a flight plan should be made.
4. Activating AED drone: The drone is cleared for take-off and can fly to the OHCA location.
5. Flying: The flight is BVLOS and at Very Low Level (VLL) altitudes in order to minimize risks.
6. Delivery of AED by drone: The AED is delivered at the OHCA location.
7. Returning AED drone: The drone and AED will return to the location of origin. 8. Preparation for next dispatch: After the return of the AED and drone, the drone is
prepared for the next flight.
4.1.2.3 AED drone control
For the operational control of the AED drones there should be a control centre. The control centre oversees operating and controlling the drones. In addition, there could be one or multiple control centres. There should be a pilot to control and monitor the drone and intervene when necessary. Furthermore, a support team should be present to maintenance the drones and monitor the health.
AED network
4.2 Assessing AED networks
Both networks can be assessed in terms of effectiveness and the involving costs. Section 4.2.1 assess both static AEDs as AED drones based on the effectiveness and the area than they can cover. Section 4.2.2 assess both networks based on the costs that are involved.
4.2.1 Effectiveness
Time to patient is one of the most important factors in the treatment of an OHCA. The Dutch Heart Foundation has introduced a six-minute window between the incidence and the time help should be on site. There are three things needed for an optimal working six-minute system: an active notification system, volunteers, and sufficient AEDs (Hartstichting, 2018a).
Static AED effectiveness
There is not a standard radius for the placement of AEDs. Zijlstra et al. (2014) assumes a radius of 500 meters between the patient and the nearest AED. However, in the study of Winkle (2010) it is described that the AEDs on, for example, an airport were placed in a 60 to 90 second walk from the patient.
It is possible to calculate a radius, based on some characteristics and assumptions. The first two minutes are needed for the recognition, calling, and locating the nearest AED. There are two different scenarios. In the first scenario, two minutes is the maximum radius, since, the bystander needs to return to the patient with the AED. In the second scenario, the bystander travels from the AED to the patient, therefore, four minutes are available.
The assumption is that the speed of a bystander is between a walking speed of 6 km/h and a running speed 15 km/h. Based on this information, the radius of both walking and running is calculated (table 4.1). The radius of a static AED is between 200 and 500 meters.
Speed (km/h) Radius (R)
Walking (retrieving) 6 km/h 200 meters
Running (retrieving) 15 km/h 500 meters
Walking (one way) 6 km/h 400 meters
Running (one way) 15 km/h 1000 meters
Table 4.1 Radius of a static AED
AED drone effectiveness
A hybridcopter is the best option for as an AED drone, since, it has the largest radius and is suitable for these operations. The speed of a hybridcopter is 108 km/h, therefore, the radius is 7.2 km (see Appendix C for comparison between drones and the radiuses).
In conclusion, the AED drone is more efficient than the static AED in terms of time to patient and radius. Another advantage is that AED drones can fly Euclidian distances and are not hindered by traffic, as is the case for static AEDs.
4.2.2 Costs
Static AED networks and AED drone networks can also be assessed in terms of costs.
Static AED costs
In the costs of an AED there are two important factors: investment costs and maintenance costs. Especially cheaper AEDs tend to be more expensive as time progresses. The average lifecycle time of an AED is 10 years (Medisol, 2018).
The average investment cost is €1,780 and the average maintenance cost is €493 (based on a life expectancy of 10 years). Therefore, the average total cost of ownership is €2,273 depending on the type of AED (Appendix D & Appendix E). The actual investment cost might be lower due to tax arrangements and insurance funding.
AED drone costs
There is not yet a functioning or developed AED drone, therefore, it is difficult to estimate the exact costs. However, based on some calculations and assumptions, the costs of an AED drone network with one AED drone is €966k per year (see Appendix F for the calculations).
An increasing number of drones will have a decreasing effect on the price, since, the fixed costs can be divided over more drones. AED drones will only be affordable if they are used in a network of multiple drones. For example, the cost of having one AED drone is around €966k per drone per year, in comparison, having twenty drones will result in a cost of €56k per drone per year (Appendix F).
Radius Costs Costs per km2
0.13 km2 €2,273 (static AED) €17,485
0.79 km2 €2,273 €2,877
18 km2 €966,000 (AED drone) €53,667
163 km2 €966,000 €5,926
Table 4.2 Cost per km2 for static AEDs and AED drones
Table 4.2 shows the comparison in costs per km2 for both static AEDs and AED drone networks. In the scenario that one AED drone is used in the whole network, it is cheaper to use static AEDs (in terms of costs per km2). However, if the AED drone network consists of more than three drones, it is cheaper to use AED drones to cover the area (Appendix G).
4.2.3 Summary of findings
5 Determinants for successful operation of AED drone networks
In this chapter establishes the determinants for successful operations of AED drone networks. First, the literature in related fields is reviewed to find determinants that could also be applied in the field of AED drones. A determinant is a critical factor that decides that will decide the outcome. In this case, determinants are identified for the successful operation of AED drone networks.5.1 Determinants identified in related fields
Since the literature about the regional factors influencing the adoption of AED drone networks is scarce, the literature in related fields will be used as guidance in determining which regional factors may influence the decision-making to place an AED drone network in a region. Related fields are: cell tower placement, helicopter emergency medical services, and ambulance care.
5.1.1 Cell tower placement
There are many similarities between the placement of an AED drone network and the placement of cell towers in a network. For example, both networks consist of multiple locations to require a certain coverage of an area. The goal of cell tower placement algorithms is to minimize the cost of placement with the maximum coverage of an area (Kashyap et al., 2015).
There are several methods to obtain the optimal cell tower locations. The method of Kashyap et al. (2015) consists of three sequential steps: identification of locations, determination of coverage, and optimization of the cell tower distribution. Firstly, the potential locations are identified based on information from typical sources such as, satellite images, population density, and topographical information from GIS. Secondly, the satellite images and the information about population density are overlaid. Based on this information, the coverage and grids of potential locations can be determined. Finally, the potential locations are numbered and sorted in a decreasing order of heights (in order to minimize costs).
5.1.2 Helicopter Emergency Medical Services
In Helicopter Emergency Medical Services (HEMS) there are several considerations in the placement of launch pads and helicopters. One of the most important considerations is the cost-effectiveness (Heggestad & Børsheim, 2002). In addition, criteria for use, cost containment, and safety are other challenges in the placement of HEMS.
The models of use predictors for the use of HEMS can be divided into three different categories: supply factors, patients’ needs, and practice variations. The different determinants for HEMS use are visualized in figure 5.1.
Figure 5.1 Theoretical model of HEMS use (Heggestad & Børsheim, 2002)
The study of Porton (2015) shows that there are four regional characteristics that have an influence on the effectiveness of HEMS operations:
• The HEMS can meet the accessibility targets,
• And/or shows an advantage compared to EMS alternatives • The demand is low and distributed over an area,
5.1.3 Ambulance care
The goal of ambulance station placement is to minimize the maximum travel distances to accidents and maximize the coverage (Maleki, Majlesinasab, & Mehdi Sepehri, 2014). The maximum travel distance is influenced by traffic, which can shorten the maximum distance (McAleer & Naqvi, 1994). The concept of ambulance coverage for response is that a demand location needs to be covered if an ambulance can reach it within a standard (for example, time or distance) (Knight, Harper, & Smith, 2012).
McAleer & Naqvi (1994) have identified several constraints to the (re)location of ambulance stations, such as, accessibility of sites, operational costs, and the construction costs. The allocation of ambulance stations differs per country based on the responsible authority.
To conclude, in the placement of ambulance stations in a region, demand and operational considerations play an important role.
5.1.4 Summary of findings
The different sectors have similarities in the determinants that are related to the determinants in AED drone networks (table 5.1). The different determinants are considered from a cost-perspective. In cell tower and ambulance networks (patient) demand is important. For HEMS patient demand is less important, since, areas need to be covered. Next, regulation is a factor for the cell tower networks, but less for medical services. Medical services often receive exceptions in regulations. Geography is most important in the placement of cell towers and for HEMS. In ambulance care (in the Netherlands) geography is less important, since, it is possible to reach all regions. Finally, operational considerations, such as, maintenance are important in all the described fields.
Cell tower HEMS Ambulance Determinants Population (demand) + +/- + Regulations + +/- +/- Geography + + - Operational considerations (e.g. maintenance) + + +
5.2 Determinants for selecting AED drone regions
Based on the previous sections regional characteristics and considerations are identified. The regional characteristics influence the decision-making process.
5.2.1 Laws and regulation
Laws and regulations will affect the operations of drones (Van de Voorde et al., 2017). Current regulations prohibit drones from flying BVLOS and focus on minimizing risks. Future regulation has a focus on mitigating risks to an acceptable level (JARUS, 2017; Rijksoverheid, 2018). In this thesis, the assumption is that the new laws will replace the old laws, making AED drone use possible.
Laws and regulations in a region, can give an indication whether a region has potential for the operations of AED drones. By looking at laws and regulations regarding the operation of drones, it can be assessed whether drone operations are possible or that there are adjustments needed.
5.2.2 Geography
Geography influences the feasibility of a region to implement an AED drone network. Not all regions in the Netherlands might be suitable for AED drones. There are for example: buildings, traffic, and landscape conditions in areas that could influence the location of a drone station (Pulver et al., 2016).
Landscape
The landscape of an area determines the suitability for drone operations. Landscape is seen as natural characteristics, such as, mountains or volcanoes, that could impact the drone operations, and might require special adjustments.
Environment
Environmental factors could influence the drone operations in an area. For example, the presence of agriculture or national park influences whether drones can be operated. In addition, drone operations are prohibited in certain Natura2000 areas, only with special flying permits drone operations are allowed (Ministry of Agriculture, Nature and Food Quality, 2018). Some environmental characteristics, such as, windmills or high buildings might impact the drone operations.
5.2.3 Operational considerations
after each dispatch the batteries should be checked, and the electrodes should be replaced. In other words, the AED drone must be prepared for its next dispatch.
5.2.4 Population: patient demand
The patient demand for a drone could be different between certain regions (Van de Voorde et al., 2017). Demographics such as, socio-economic factors, ethnicity, and age, influence the risk of an OHCA. The coverage of a network should be optimized based on these regional characteristics.
5.2.5 Social considerations
6 Framework design
In this chapter, the focus is on the development of a framework to help guide the efficient decision-making on where to use AED drones. The framework concerns a stepwise approach to find the regions that have the highest potential for the implementation of an AED drone network.
6.1 Framework design approach
The aim of the framework is to guide the efficient decision-making on where to use AED drones. The healthcare sector has a focus on minimizing costs against the best patient outcomes. The decision problem is approached in an efficient way by using already available data and means to develop and provide input to the framework. The framework consists of five different steps. Every step will require a higher investment from a research perspective (efforts, costs, and time for doing research).
The framework consists of desk-research and field-research. The ground step and the first step are desk-research. The second step is both desk-research and field-research. The network characteristics are first identified based on desk-research; however, some characteristics require field research. For example, the impact of airflows around mountains on an AED drone needs to be assessed in the field. The third step consists of desk-research. The fourth step consists of desk-research; based on available OHCA data and risk factors, the network is optimized. The fifth step of the framework consists of both desk-research and field-research. The final solution can be assessed by means of simulation or a real-time pilot. Data that is already available (volunteer, EMS, AED, OHCA demand) is used as much as possible in the desk-research steps.
In the first step, the goal is to narrow down the number of possible candidate regions to cover. The final step of the framework concerns an investigation in terms of solution assessment and testing. The framework is a funnel model where the number of potential locations is reduced with every step, in order to arrive at those regions with a high potential.
Framework input
The framework is fed by information about the regions and their characteristics. All different inputs are further explained in section 6.2.
Framework
whether a region has potential for AED drone operations. In the third step, an initial cost comparison is constructed for static AEDs and AED drones in order to determine whether it is cost-effective to cover the selected area. Next, the number of drone stations is optimized based on patient demand and social considerations. Finally, the coverage regions are selected based on testing and initial solution assessment.
Framework output
The output of the framework will consist of different coverage regions that are suitable for the implementation of AED drones. The final step gives an indication which coverage regions are most suitable, and which regions are less suitable for AED drone networks.
6.2 Framework input
The framework is fed by different regional characteristics and decisions. The different characteristics and decisions are explained below.
Input step 1: Selecting regions that are not being covered by the current AED infrastructure
• Volunteer data
The current AED infrastructure relies on the use of volunteers. Without volunteers the AED infrastructure would be impossible to accommodate in a speedy manner.
• Public AED locations
There are several maps to see where AEDs are registered and to see which AEDs are available 24/7.
• EMS data
Based on the National acute care map it is possible to identify regions that cannot be covered by an ambulance within 10 minutes (Landelijk Netwerk Acute Zorg, 2018).
Input step 2: Make an assessment based on network characteristics • Geography
It might not be feasible to use drones in a certain region. There could be environment factors, such as, national parks or landscape characteristics, that restrict the drone operations.
• Laws and regulation
Based on the type of drone operations, area, and drone, a risk profile is composed. The risk profile consists of the ground risk and the air risk of the drone operation.
• Operational considerations
The area should be reachable by maintenance crew in order to prepare the drone for its next dispatch. This also include the periodic maintenance of an AED drone launch station.
Input step 3: Initial cost comparison: static AEDs vs. AED drones • Costs static AEDS
The different costs of static AEDs are identified in section 4.2.2. These costs will be used to make calculations in the second step of the framework.
• Costs AED drones
Input step 4: Optimization of the network • OHCA patient demand
OHCA patient demand can be determined by the historical treatment records. Based on data it might be possible to indicate “hot spots” or higher risk areas where there are more OHCA incidences compared to other regions.
• Social considerations
There might be social considerations, in a region, that influence the placement of an AED drone.
6.3 Framework
This section provides the developed framework to guide the decision-making on where to use AED drones. The framework intends to first select candidate regions. The framework consists of five steps, and each step is followed by a decision-moment. The decision-moments help the stakeholders to enter the decision process. The goal is to find coverage regions that are most suitable and efficient to implement AED drones.
6.3.1 Use of the framework
This section will clarify the use of the framework. The framework is intended to find regions with the highest potential for the use of AED drones. A region can be defined in several ways, depending on who the client initiating the study is, and the influence of the client. In addition, the way of facilitating the AED infrastructure influences the use of the framework. The framework can be applied to several layers of detail. For example, it can be applied to a country, provinces, municipalities, towns, and zip code areas.
In some cases, step 2 follows step 3 in the framework if the field-research requires elaborate means. For example, the influence of airflows around mountains on the fly-path of AED drones. In this case, making an initial cost comparison before the assessment based on network characteristics is more efficient in terms of resources for doing research.
6.3.2 Detailed description of the framework
Step 0: Choice of start regions – boundary and detail
Step 0 concerns the selection of start regions for the analysis on where AED drones can be used most efficiently. The start region sets the boundary of the research. The choice of the set of regions is dependent on the client. For example, an EMS provider which serves certain regions or a regional authority being responsible for healthcare. This step also determines the level of detail in the further application of the framework. The more detailed the start regions are, the more detailed the outcomes will be. However, more detail can also lead to regions that are too small to cost-efficiently cover with AED drones (see step 3).
Regions can be provinces, municipalities, towns, EMS areas, safety regions, care regions, air space classes, and even zip code areas.
Step 1: Selecting regions that are not being covered by the current AED infrastructure The first step of the framework is to select regions that are not being covered by the current AED infrastructure. The AED infrastructure consists of the chain-of-survival (early recognition, early CPR, early defibrillation, and advanced life support). Early recognition is beyond the scope of the research, since, it relies on bystanders that witness an OHCA, and cannot be influenced using AED drones.
In current AED networks, EMS acts as an AED carrier. In addition, there could be other parties that carry AEDs, for example, police or the fire brigade.
The areas that are not being covered by the current AED infrastructure are identified by using volunteer, AED, and EMS data. The sequential steps of the chain-of-survival are followed in this process. First, the volunteer coverage is considered. As described before, CPR and early defibrillation should be performed within a six-minute window for the best patient outcomes (Hartstichting, 2018a). Thereafter, the AED coverage is determined to indicate which areas are not covered. Finally, the current EMS coverage is analysed
Step 2: Make an assessment based on network characteristics
In the previous step, regions that are not being covered by the current AED infrastructure are identified. In this step the regions will be assessed on network characteristics. Based this step, regions are selected with the most potential for AED drone operations. For example, the risk profile is low or medium and there are little geographical characteristics that hinder the drone operations.
Regulations
and regulations. For example, current or future regulations will influence the operations. In addition, the risk profile will influence the potential of a regions. There might be regions with a high risk that could limit the drone operations, since, mitigations are needed to lower the risk.
Laws and regulations Impact (yes/no) What is the impact Current or future
regulation
Risk profile (ground & air risk)
Table 6.1 Laws and regulations
Geography
Geography is divided into two different categories: landscape and environmental characteristics (table 6.2 and table 6.3). Landscape are natural characteristics and environmental are human characteristics. The different characteristics might impact the drone operations. Some impacts can be mitigated or controlled in order to ensure drone operations.
Landscape characteristics Impact (yes/ no) What is the impact Climate Mountain Hill Desert Sea/ lake Urban Rural Forest
Table 6.2 Landscape characteristics
Environmental characteristics
Impact (yes/no) What is the impact
National park Animals
Infrastructure
Attractions (theme parks etc.)
Table 6.3 Environmental characteristics Operational considerations
and the possibility for maintenance. The maintenance time should be minimized, since it is likely that the drones are scattered out over a larger area. In addition, the locations should be easy to reach in order to change batteries and prepare the drone for the next dispatch (table 6.4).
Operational considerations Yes/No What is the impact
Is the area reachable for maintenance
Table 6.4 Operational considerations
Step 3: Initial cost comparison: static AEDs vs. AED drones
The third step is to determine the minimal area that needs to be covered. The size of the area indicates whether it will be suitable and efficient to use AED drones or static AEDs are a better option. Based on the AED data, the minimal number of AEDs can be calculated. Thereafter, the minimal number of AED drone stations are calculated (see model Appendix H). The costs of both static AEDs and AED drones are identified in section 4.2.2. Based on the OHCA incidence rate an assessment can be made whether a region may require multiple drones at a single launch site. However, in most rural areas a single drone at a single launch site will be enough.
After this step, a decision is made if AED drones are a cost-effective option, the area needs to be bigger, or if it is better and cheaper to cover the area with static AEDs.
Step 4: Optimization of the network
The fourth step of the framework is to optimize the number of AED drones based on patient demand, OHCA risk indicators, and social considerations.
Patient demand
The frequency and distribution of OHCAs over a different region can indicate whether there are so-called hotspots. Hotspots are areas with a high number of OHCAs compared to surrounding areas. It might be beneficial to use more AED drone stations of more AED drones per station to cover those areas.
OHCA risk indicators
Social considerations
There could be social characteristics that influence the drone operations (table 6.5). For example, the placement of an AED drone nearby a sports complex or university campus, since, the social impact of an OHCA is high.
Social considerations Impact (yes/no) What is the impact University campus
Sports complex Tourism
Table 6.5 Social considerations
Step 5: Testing and assessment
The final step in de framework consists of testing and the assessment of the regions. The decisions and selections made in the previous sections will be tested and further analysed. The output of the previous steps will lead to several regions that are most suitable for AED drone operations. It is important the selected regions are further tested, and the network is further designed.
6.4 Framework output
7 Application and evaluation of framework
In this chapter, the proposed framework in chapter 6 will be evaluated. The framework will be applied to the province of Drenthe. The framework will be checked for completeness and efficiency. Section 7.1 explains the evaluation approach. Next, the framework input is presented in section 7.2. The framework is evaluated in paragraph 7.3 by following all the steps and input, in order to discuss the output in section 7.4. Finally, the findings are summarized in paragraph 7.5.
7.1 Evaluation approach
A full evaluation is not possible due to time and data constraints. For the patient demand information, AED data, and volunteer information, it was only possible to receive them per zip code area.
The focus of the evaluation is to demonstrate how the framework performs efficiently when it is applied to a region. In addition, the evaluation will give a check of completeness. The first step in the evaluation is to describe the input of the framework. Real data will be used as an input to determine the output. Next, the framework itself will be evaluated by means of finding areas with high-potential for AED drone operations. Firstly, the regions will be selected that are not being covered by the current AED infrastructure. Secondly, the regions will be assessed based on AED drone network characteristics. Thirdly, an initial cost comparison will be made for static AEDs vs. AED drones. Fourthly, the network is optimized based on patient demand and OHCA risk-factors.
The final step of the framework, the selection based on testing and initial solution assessment, and the output of the framework are beyond of the scope of the evaluation, due to time constraints. However, the evaluation is intended to generate a better understanding of the functioning of the framework.
7.2 Framework input
The province of Drenthe, where the EMS provider operates, is chosen as pilot area for the evaluation. The ambulance region is divided into zip code areas to create more detail. The assumption is that the air regulations will change, otherwise, drone operations are not allowed. The province of Drenthe consists of 310 zip codes and has 244 towns. The input is presented below.
• EMS data
• OHCA patient demand from 2012-2016, provided by HartslagNu • New laws and regulations
• Geographical data • Spatial data
7.3 Framework
The province of Drenthe will be evaluated and the high potential regions for AED drone networks will be selected based on the previously described input. The framework will follow the different steps that are described in section 6.3. The first step is to select regions that are not being covered by the current AED infrastructure. Second, an assessment, based on network characteristics, is made to determine the feasibility of a region for the use of AED drones. Third, an initial cost comparison is made of static AEDs vs. AED drones. Thereafter, the number of drones is optimized based on patient demand and social considerations. As described before, the final step, testing and assessment is beyond the scope of this research due to time constraints.
The detailed illustration and description of the framework applied to the province of Drenthe are provided in the textbox below. The different steps of the framework, the input, the key decisions, and the details are discussed. The different steps are described in detail in section 6.3. The steps below are placed in the main text because it shows the functioning of the framework. But this part can be skipped by the reader, therefore, the detailed evaluation is placed in a grey textbox.
Step 0: choice of start regions – boundary and detail
UMCG Ambulancezorg is the EMS operator in the province of Drenthe. Therefore, the province of Drenthe is chosen as the boundary of the start region. This area is further divided into zip codes in order to create a sufficient level of detail, since, there are differences in the number of AEDs, volunteers, OHCAs, and EMS travel times per zip code.
In Drenthe there are several Post Office (PO) boxes that are privately owned. PO boxes are not connected to streets, therefore, there will be no volunteers, OHCA incidences, and AEDs registered on these locations. In total, there are 310 zip codes in Drenthe, however the 35 PO boxes are excluded from the dataset. Therefore, the start region consists of 275 zip codes.
Step 1: Selecting regions that are not being covered by the current AED infrastructure
The goal of this step of the framework is to find regions that are not being covered by the current AED infrastructure. By combining the data below such regions can be identified.
Volunteer data
There is no minimal requirement for the number volunteers per zip code. The more volunteers the better the coverage will be. Figure 7.1 shows the distribution of the volunteers in Drenthe. There are 29 zip codes with 0-5 volunteers and 102 zip codes with 6-25 volunteers. Furthermore, 14 zip codes have more than 100 registered volunteers.
Figure 7.1 Number of registered volunteers per zip code
AED data
Figure 7.2 shows the number of AEDs per zip code. There are 69 zip codes without any AEDs. In total 229 zip codes have less than the minimal requirement of four AEDs.
EMS data
Figure 7.3 shows the province of Drenthe and the ambulance coverage. The map shows that the dark-red areas cannot be reached by an ambulance within 10 minutes. The light-yellow areas are next to ambulance posts, and can be reached within three minutes.
Figure 7.3 Ambulance coverage of Drenthe, adapted from: Landelijk Netwerk Acute Zorg (2018)
Figure 7.4 Overview of regions that are not covered
The selected region (figure 7.5) has ambulance travel time of more than 10 minutes and the average number of AEDs per zip code is one. The average number of OHCAs in this area is around 12 per year. The selected area has a total number 13,610 inhabitants (see Appendix I for inhabitants per zip code).
Step 2: Make an assessment based on network characteristics
The second step of the framework is to assess whether the region has potential for the use of AED drones. The selected region is assessed based on network characteristics. There are three different network characteristics: laws and regulations, geography, and operational considerations. The complete filled in tables from section 6.3.2 can be found in Appendix J.
Laws and regulations
AED drone operations are only possible if the new regulations are in place. Furthermore, the risk-profile impacts the selected region. Large cities are avoided; however, the drone still needs to cross highways. The ground and air risk are both medium.
Geographical considerations
There are several geographical characteristics that influence the selected region. For example, there are a few high was and in addition there are military exercise zones where drone operations are limited. In addition, drone operations in Natura2000 areas is only allowed with special permits (Appendix K). Both will hinder the operations of AED drones. However, a large part of Drenthe consists of military exercise zones, therefore, arrangements should be made between the military and the drone operator. The same is true for the Natura2000 areas.
Operational considerations
Operational considerations could influence whether a region is suitable for drone operations. The selected region can be reached for maintenance (see appendix J).
Conclusion
AED drone operations will be difficult in the selected area, since, it is in a military exercise zone. However, most of the airspace in Drenthe consists of military exercise zones, therefore, arrangements should be made with the government and the military for the operation of AED drones.
Step 3: Initial cost comparison: static AEDs vs. AED drones
In this step of the framework an initial cost comparison is made for static AEDs vs. AED drones.
Costs to cover the area with static AEDs
There should be at least four AEDs per zip code. The number of missing AEDs can be calculated based on the AED data from the first step and the requirement of at least four AEDs. In total there are at least 68 AEDs needed for a minimal coverage of the area. According to the AED data, there are already 22 AEDs (Appendix L). Thus, 46 static AEDs are still needed to cover this area. The costs of €2,273 per static AED are identified in section 4.2.2. Therefore, the total cost of a onetime investment is €104,558 to cover the selected area with static AEDs.
Costs to cover the area with AED drones
goal is to cover the whole area with the minimal number of AED drone stations. Based on the OHCA incidence of 12 per year, one AED drone per launch site is enough.
Figure 7.6 shows the coverage of the area. The minimum number of AED drone stations is two. The drone stations are placed on borders between zip codes to maximize the coverage.
Figure 7.6. Coverage of two AED drone stations
The costs per AED drone are based on section 4.2.2. The costs to implement a network of two drones is high (€483,651 per drone per year), since, the fixed costs can only be divided over two drones. In this case it is cheaper to use static AEDs to cover the area.
In order to continue the illustration of the framework, it is assumed that the other regions, that are not being covered in step 1, are also part of the implementation of an AED drone network. Therefore, it is assumed that the drones are implemented in a larger AED drone network in Drenthe consisting of a total of 25 AED drones, therefore, the total costs will be €50,000 per AED drone as a onetime investment. This number is based on an interview with a domain expert and the cost assessment in section 4.2.2. Thus, the total costs to use AED drones in the selected area is €100,000, based on two drone stations with each one AED drone.
