Drones delivering automated external defibrillators for out-of-hospital cardiac arrest: A scoping review

Aim

Out-of-hospital cardiac arrest (OHCA) remains a critical health concern, where prompt access to automated external defibrillators (AEDs) significantly improves survival. This scoping review broadly investigates the feasibility and impact of dronedelivered AEDs for OHCA response. Methods: PubMed, Cochrane, and Web of Science were searched from inception to August 6, 2024, with eligibility broadly including empirical data. The charting process involved iterative data extraction for thematic analysis. Results: We identified 306 titles and, after duplicate removal, title/abstract screening, and full text review, included 39 studies. These were divided into three categories: 1) Real-world observational studies (n = 3), 2) Test flights/simulation studies and qualitative analyses (n = 15), and 3) Computer/prediction models (n = 21). Real-world studies demonstrated the feasibility of drone AED delivery, with a time advantage of 01:52 – 03:14 min over ambulances observed in 64–67 % of cases. Test flight/simulation and qualitative studies consistently reported feasibility and positive bystander experiences. Computer/prediction models exhibited considerable heterogeneity, yet all indicated significant time savings for AED delivery compared to traditional EMS methods. Moreover, seven studies estimated improved survival rates, with five assessing cost-effectiveness and favouring drone systems. Regional factors such as EMS response times, volunteer responder programmes, terrain, weather, and budget constraints influenced the system’s effectiveness. Conclusion: Across all categories, studies confirmed the feasibility of drone-delivered AED systems, with significant potential for reducing time to AED arrival compared to EMS arrival. Prediction models suggested enhanced survival alongside costeffectiveness. Further research, including more extensive real-world studies and regulatory advancements, is imperative to integrate drones effectively into OHCA response systems.

Introduction

Out-of-hospital cardiac arrest (OHCA) is a pervasive challenge for healthcare systems and communities worldwide. Despite variations in survival outcomes across different regions, the overall prognosis for OHCA remains poor with survival rates typically ranging from 8 % to 10 %, emphasizing an urgent need for targeted interventions and innovative strategies to enhance survival rates.1, 2, 3 Efforts to improve OHCA survival should be based on the internationally widely recognized model “chain of survival”.4, 5, 6 The first three and most important links of the chain comprise 1) OHCA recognition and activation of the emergency response system, 2) immediate initiation of bystander cardiopulmonary resuscitation (CPR), and 3) rapid defibrillation before arrival of the Emergency Medical Services (EMS).7, 8, 9, 10, 11 Bystander defibrillation has the greatest impact on survival, and an array of different interventions have been introduced focusing on improving defibrillation by laypersons.12, 13 However, bystander defibrillation remains low at less than 3 % globally.¹⁴

Drones have emerged as a promising solution for delivering automated external defibrillators (AEDs) before EMS arrival.15, 16, 17, 18, 19, 20 They offer the potential to significantly reduce response times and, improve survival rates. This exploration is particularly relevant in contexts involving private and residential locations, which have historically presented challenges for timely defibrillation.21, 22, 23, 24

Given the lack of a comprehensive review on this approach, this topic was prioritized by the Basic Life Support (BLS) Task Force of the International Liaison Committee on Resuscitation (ILCOR) for investigation. This scoping review aims to map the type and volume of existing research and identify gaps, rather than conduct a quantitative synthesis of the current literature on the use of drones for delivering AEDs. By focusing on feasibility, safety, time gain, survival rates, cost-effectiveness, and human-drone interaction across various settings, we seek to provide a broad understanding of the current state of research in this field.

Methods

This scoping review followed the ILCOR process for evidence review, including a prespecified plan approved by the ILCOR BLS Task Force prior to releasing a Consensus on Science with Treatment Recommendations (CoSTR) (Supplemental 1).²⁵ Inclusion criteria and search strategy were subsequently updated and approved to better align with scoping review guidelines.

Inclusion criteria

Inclusion criteria were broad, guided by the Population-Concept-Context (PCC) elements for conducting scoping reviews.²⁶ The research question was formulated “What evidence exists regarding the use of drone-delivered AEDs for OHCAs?”. The population was defined as individuals experiencing OHCA, including suspected OHCA. The concept focused on AEDs delivered by drones. The context was open to include various study settings. Grey literature was included only to the extent available in the chosen databases and only if it provided empirical data on the topic.

Exclusion criteria

Publications concerning the use of medical drones for other purposes than AED delivery were excluded. Text and opinion literature, with no empirical data, e.g., previous reviews or comments on publications, were included as background knowledge, but not included in the final review.

Information sources and search strategy

With the assistance of a professional health science librarian, we conducted a comprehensive search strategy for PubMed, the free online database managed by the National Library of Medicine, and Web of Science, a broader multidisciplinary database. The search combined the words “DRONE” and “OHCA” or “AED” including synonyms for all three words and Medical Subject Headings (MeSH) terms when available (Supplemental 2). The search strategy was not peer reviewed.

The search encompassed all years from database inception (PubMed from 1966 and Web of Science from 1900) and was limited to publications in the English language. Searches were run on October 13, 2023, and updated February 4, 2024, and once again on August 5, 2024. Additionally searches in the Cochrane Library were performed on these same dates. Reference lists of included studies were also examined to identify additional publications not captured by the database searches.

Evidence screening and selection

Two reviewers screened the publications independently; duplicate removal and title and abstract screening were performed assisted by the online tool Covidence for systematic reviews.²⁷ Disagreements were resolved by discussion. No detailed assessment of the quality or stringency of individual sources was planned or performed.

Data charting process, items, and synthesis

The data charting process was iterative. Initial key information (aims/purpose, methodology, and main findings) was extracted and charted in a template data table for thematic analysis. Through this process, additional specific details of interest (e.g., drone flight reliability, drone specifications, degree of urbanization, etc.) were identified, discussed iteratively, and included after mutual agreement.

Ethical approval

None.

Results

The searches in PubMed and Web of Science identified 302 publications in total. Manual searches of reference lists added 4 more publications. The Cochrane Library searches did not yield any additional publications. After removing duplicates, 148 publications were screened by title and abstract. This process left 59 publications for full-text review, out of which 39 met the inclusion criteria (Fig. 1).

Fig. 1.

Flowchart illustrating the screening process and inclusion of publications.

The included studies originated from various regions and countries, including Sweden (n = 8), Germany (n = 4), Austria (n = 2), France (n = 1), UK (Wales, Northern Ireland) (n = 2), Canada (Ontario, British Columbia) (n = 5), USA (Michigan, North Carolina, Washington, Virginia, Utah) (n = 12), South Korea (n = 2), China (Tianjin) (n = 2), and Thailand (n = 1).

The thematic data analysis enabled categorization of the content into three groups:

• Real-world drone AED delivery for OHCA: Three publications, all from Sweden, explored real-life AED-delivery by drone to OHCA patients.18, 28, 29 (Table 1).

• Test flights/simulation studies & qualitative analyses: 15 studies in this category evaluated the feasibility of real-world AED delivery by drones through simulated OHCA scenarios and analysed laypeople's perceptions of the concept.15, 19, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42 (Table 2).

• Computer/prediction models for drone AED feasibility and effectiveness: 21 studies employed various strategies to determine optimal sites for the placement of AED-drone bases and estimate time gains compared to EMS response times, cost-effectiveness, and more.16, 17, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61 (Table 3).

Table 1.

Real-world drone AED delivery for OHCA.

Author, year, and country:	Study design / methodology:	Main findings:
Schierbeck et al., 2023, Sweden.18	Prospective observational study, AED delivery frequency and time gain by five drones for real OHCAs.	72 drones dispatched for 211 suspected OHCAs; AEDs successfully delivered in 81 %. Drones arrived before EMS in 67 % of cases with a median time gain of 3:14.
Schierbeck et al., 2022, Sweden.29	Case report, OHCA involving the use of a drone-delivered AED.	71-year-old man with an OHCA, received defibrillation with a drone delivered AED following prompt CPR. The man fully recovered.
Schierbeck et al., 2022, Sweden.28	Feasibility study, three drones delivering AEDS for real OHCAs. Additionally, 61 BVLOS test flights were conducted.	Successful AED delivery in 11 out of 12 cases. Drones arrived before EMS in 64 % of cases.

Abbreviations: AED (automated external defibrillator), OHCA (out-of-hospital cardiac arrest), CPR (cardiopulmonary resuscitation), EMS (Emergency Medical Services, BVLOS (beyond-visual-line-of-sight). Detailed table 1, Supplemental 3.

Table 2.

Test flights/simulation studies & qualitative analyses.

Author, year, country:	Study design / methodology:	Main findings:
Davidson et al., 2024, Michigan, USA.*42	Randomized simulation of 24 emergencies. Assess safety and efficacy of bystander's interaction with a grounded drone under varying 9–1-1 dispatcher guidance.	Drone-specific dispatch instructions positively impacted bystander safety.
Leith et al.,2024 Michigan, USA.*41	Randomized simulation of 24 emergencies. Qualitative assessment of bystander's interaction with a grounded drone with a medical kit.	Lay bystanders can effectively and comfortably interact with a medical drone.
Starks et al, 2024, North Caolina, USA.40	Simulated 51 public OHCAs with single bystanders of varying CPR training levels; timed AED retrieval/application and assessed CPR quality.	Median time to retrieve and use AED was 1:59, suggesting drones must arrive 2 min before EMS for effective intervention.
Scholz et al., 2023, Germany.30	Assessed feasibility of BVLOS drones for nighttime AED delivery with 10 night and 10-day flights to simulated OHCA sites in rural Germany.	Nighttime missions were feasible with no safety incidents and no major operational differences from daytime flights.
Fischer et al., 2023, Austria.31	Examined BVLOS drone AED delivery feasibility in mountainous regions through 29 simulated OHCA scenarios.	BVLOS drone delivery of AEDs in mountainous regions was feasible.
Purahong et al., 2022, Thailand.32	Developed and tested a mobile phone application for bystanders to request a medical drone for AED delivery in OHCA cases.	The operational field test produced successful results.
Baumgarten et al., 2021, Germany.19	BVLOS drone AED delivery tested in 46 rural OHCA simulations with timing and interviews of lay bystanders and community first responders.	Integrating airborne AED delivery with EMS and community first responders was feasible and safe, though still experimental.
Rees et al., 2021, Wales, UK.33	Tested a fixed-wing drone for AED delivery via parachute in 6 BVLOS flights.	4/6 flights achieved successful AED drops. A full-scale OHCA simulation confirmed proof of concept.
Kim et al., 2021, South Korea.34	24 simulated OHCA-scenarios with drone-delivered AEDs and lay bystanders; compared audio vs. video instructions in dispatcher-assisted CPR.	Participants valued drone AED delivery in areas lacking AEDs. Video instructions improved CPR quality over audio-only
Zègre-Hemsey et al., 2020, North Carolina, USA.**35	Qualitative assessment of lay perceptions of AED-equipped drones during 35 simulated OHCA events.	17 of 35 participants were interviewed, reporting positive experiences with AED-equipped drones.
Cheskes et al., 2020, Ontario, Canada.15	Examined feasibility of AED drone delivery, comparing response times of simultaneous drone and ambulance dispatch in 6 OHCA simulations.	The AED drone arrived 1.8 to 8.0 min before the ambulance, demonstrating feasibility in simulated OHCA events.
Rosamond et al., 2020, North Carolina, USA** (Author Manuscript).36	Quantitative assessment of lay performance in 35 simulated OHCA events comparing ground AED search to drone-delivered AED times.	In 4 out of 5 zones, the drone AED arrived before the ground AED (0:17–02:56 min). In the remaining zone, the ground AED, located 77 m away, arrived faster.
Sedig et al., 2020, Ontario, Canada.39	Conducted interviews and focus groups with 67 participants to explore public perception of drone AED delivery in OHCA.	Evaluated lay use of drone-delivered AEDs to be feasible, highlighting the need to address community cardiac arrest literacy before implementation.
Sanfridsson et al., 2019, Sweden.37	Explored lay bystander experiences with drone-delivered AEDs in 8 simulated OHCAs. Evaluated time delays and qualitative analysis.	Bystanders found drone-delivered AEDs safe and feasible.
Claesson et al., 2017, Sweden. (Research Letter).38	Conducted 18 BVLOS test flights to historical OHCA locations within a 10 km radius, comparing drone AED delivery times to historical EMS response times.	Arrived faster than EMS in all 18 cases, with a median response time reduction of 16:39 min.

Abbreviations: EMS (Emergency Medical Services), OHCA (out-of-hospital cardiac arrest), AED (automated external defibrillator), BVLOS (beyond-visual-line-of-sight). Detailed Table 2, Supplemental 3.

^*Report additional findings from the same 24 emergency simulation.

^**Report additional findings from the same 35 testflights.

Table 3.

Computer/prediction models for drone AED feasibility and effectiveness.

Author, year, country, and type of area:	Prediction-model target:	Main findings:
Starks et al, 2024, USA, North Carolina. 61	Combined AED-equipped drones with providing First Responders with AEDs to ensure AED arrival within 5 min.	First responder AEDs and 326 drones increased 5-minute coverage from 16.5 % to 56.3 %. Estimated survival increased from 14.5 % to 19.4 % for witnessed OHCA.
Ren and Li, 2023, China, urban area.43	Ensured an AED on site within 4 min, based on 62 simulated OHCAs.	24 drone launch sites for area coverage, estimated survival increased from 22 % to 64 %.
Frigstad et al.,2023, Sweden, mixed area.44	Combined dispatch of volunteer responders and AED-equipped drones, estimated survival impact.	AED arrival times were reduced by 05:29 min using 20 drone launch sites. Estimated survival increased from 9.6 % to 15.8 %.
Yukun et al., 2023, China, urban area.45	Compared a heuristic drone model with an improved algorithm, based on 300 simulated OHCA locations.	The improved algorithm increased system stability and reduced required bases by 29 %. Estimated response time was 2:07 min with 25 launch sites.
Röper et al., 2022, Germany, rural area.46	Estimated cost differences between a stationary AED network and a drone based.	Drone-based network was significantly more cost-effective, costing €1,000/km2 vs. €2,500/km2 for a stationary network.
Leung et al., 2022, Canada, mixed area.47	Compared five drone configurations to EMS response times for historical OHCAs.	All drone systems reduced median AED arrival times from 6.4 to 4.2–5.4 min, arriving before first responders in 53–77 % of cases.
Choi et al.,2021, urban area, South Korea.48	Impact of topography, drone type, and weather on AED delivery time.	Drone pre-arrival in 12 % − 27 % of OHCAs (compared to EMS), depending on drone capabilities.
Ryan, 2021, suburban area USA, Virginia.49	Two drone configurations, compared them to EMS response times for simulated OHCAs.	Drones covered 98 % of the study area within 10 min and 70 % within 5, consistently outperforming EMS response times.
Lancaster and Herrmann, 2021, urban/suburban area, USA, Washington.50	Costs and survival based on historical witnessed, shockable OHCAs, and mobile responder density.	Showed a cost-survival trade-off over 10 years. A five-base drone system increased survival from 20 % − 30 % and was cost-effective.
Schierbeck et al., 2021, Sweden mixed area.51	Drone networks covering 50 %-100 % of areas with EMS response times over 8 min, including a focus on high-incidence areas.	The number of drones needed varied from 21 to 2,408, with 61 drones required to cover all high-incidence OHCAs within 8 min.
Derkenne et al., 2021, France, urban/suburban area.52	Drone vs. historical OHCA response times, including OHCA recognition, bystander presence, time of day, and weather.	Estimated AED delivery by drone before EMS in 26 % of OHCAs (n = 777) in Greater Paris, arriving on average 190 ± 7 s faster in 93 % of cases.
Chu et al., 2021, mixed area, Canada.53	A machine-learning drone dispatch ruleset to minimize unnecessary flights.	The machine learning model maintained reduced median response time of 33 % and decreased dispatches by 30 %, with minimal sensitivity loss.
Bauer et al., 2021, Germany, rural area.54	Drones covering 80 %-100 % of areas with EMS response times over 7 min, comparing cost-effectiveness and life years gained.	100 % coverage required 1,933 drones, yielding a 150 % increase in life years over current EMS. All configurations were cost-effective.
Lancaster and Herrmann, 2020, urban/suburban area, USA, Washington.55	Drone network AED response times compared with EMS; include mobile responder density and survival estimation.	Dispatched mobile responders and 5 drone AED launch sites reduced time to defibrillation, survival increased from 20 % to 30 %.
Mackle et al., 2020, mixed area, Northern Ireland.56	Drone network AED response times compared with EMS and bystander AED retrieval times.	A 78-site drone network, reduced mean response times by up to 50 % in some regions.
Wankmüller et al., 2020, rural (mountain) area, Austria.57	Compared two drone types, varying in speed, to reach simulated OHCAs in a mountainous region.	Fast-flying drones from 36 launch sites reached 95 % of OHCAs within 5 min, outperforming air ambulance times.
Bogle et al., 2019, mixed area, USA, North Carolina.58	Drone AED networks to estimating the incremental cost-effectiveness ratio per quality-adjusted life years (QALYs).	A 500-drone network reduced median AED arrival time from 7.7 to 2.7 min, increased survival from 12.3 % to 24.5 % and gained 30,267 QALYs.
Pulver and Wei, 2018, mixed area, USA, Utah.59	Number of drones and launch sites needed to cover historical OHCA locations.	68 launch sites and 71 drones to reach 90 % of historical OHCAs within 1 min. 59 % of cases were backed up by two or more drones.
Boutilier et al., 2017, mixed area, Canada.60	Number of drones and launch sites needed to cover historical OHCA sites 3 min before median EMS response.	81 launch sites and 100 drones reduced AED arrival by 6:43 min in urban and 10:34 min in rural areas.
Claesson et al., 2016, mixed area, Sweden.16	Compared a drone model balancing OHCA incidence and EMS delay with one prioritizing EMS delay.	Balancing OHCA incidence and EMS delay, drones arrived 1.5 min before EMS in 32 % of cases; prioritizing EMS delay, they arrived 19 min earlier in 93 % of cases.
Pulver et al., 2016, mixed area, USA, Utah.17	A drone network compared to estimated EMS response times, aiming for arrival within one minute in 90 % of incidents.	Existing EMS infrastructure delivered an AED within 1 min in 4 % of cases; a drone network increased this to up to 90 %.

Abbreviations: OHCA (out-of-hospital cardiac arrest), AED (automated external defibrillator), EMS (Emergency Medical Services). Detailed Table 3, Supplemental 3.

35, 36 Zègre-Hemsey et al. focused on qualitative data analysis, while Rosamond et al. presented quantitative findings. Both studies were incorporated into the review in category 2. Similarly, Leith et al. and Davidson et al. reported additional findings from the same simulated OHCAs, and these studies were also included in category 2.41, 42 Two studies contributed data to multiple categories but were included only in the category of their predominant focus.16, 18

Real-world drone AED delivery for OHCA

Three real-world studies (Table 1) were reported by the same group in Sweden.18, 28, 29 Adverse events were monitored in two of the studies, with none reported in either case. 18, 28.

A feasibility study assessed using drones for delivering AEDs in real-life suspected OHCA cases.¹⁸ Over 4 months, 14 cases (out of 53 suspected OHCAs) were eligible for drone dispatch. Drones were deployed in 12 of these cases, successfully delivering an AED in 11 cases. In 64 % of these instances, drones arrived before ambulances, achieving a median time benefit of 01:52 min (IQR 01:35–04:54). Additionally, the study described 61 beyond-visual-line-of-sight test flights, achieving a 90 % operation reliability.

A prospective observational study examined drone AED delivery compared with EMS in real-life suspected OHCAs. ²⁸ In 72 of 211 suspected OHCAs a drone was deployed, with 58 AEDs successfully delivered. Drones arrived before ambulances in 37 of cases, with a median time gain of 3 min and 14 s. In six cases the AED was attached, and in two of the cases the AED shocked the patient before ambulance arrival. One person achieved 30-day survival.

The third study was a case report, detailing successful resuscitation after defibrillation by a drone-delivered AED. ²⁹ A 71-year-old male had a cardiac arrest outside his house and promptly received CPR. The drone reached the patient before EMS, and the drone AED was attached and delivered one shock before the EMS arrival. Subsequently, the patient received three more defibrillations by the EMS crew and was discharged from the hospital with full neurological recovery after seven days.

Test flights and qualitative analyses

Twelve studies conducted drone test flights to assess AED delivery to simulated OHCA sites. Nine of the studies flew their drones beyond-visual-line-of-sight, totaling 151 flights.15, 19, 30, 31, 32, 33, 35, 36, 38 Three studies conducted a total of 83 flights within line of sight, with their primary outcome being bystander interventions and experiences with the drone.34, 37, 40(Table 2).

Three studies did not conduct drone flights.39, 41, 42 Sedig et al. qualitatively investigated society’s readiness for the use of AED-equipped drones as an EMS supplement,³⁹ while Leith et al. and Davidson et al. investigated different perspectives of bystander interaction with a non-flying drone in 24 simulated emergencies (including 12 OHCA simulations).41, 42

Beyond-visual-line-of-sight and flight reliability

Baumgarten et al. reported a 93.8 % operational reliability in beyond-visual-line-of-sight flights, with four flights out of 50 cancelled due to wind and technical issues.¹⁹ Fischer et al. cancelled one flight out of the planned 30 due to adverse weather conditions.³¹ The rest of the studies conducted their flights without any reported cancellations.

Time gain for AED drone delivery

Cheskes et al. compared drone flights (n = 6) to actual ambulance response times and found that drones consistently delivered the AED before ambulance arrival, regardless of different startup locations for the drone.¹⁵ The drone had a travel-distance ranging from 6.6 km to 8.8 km and the ambulance travelled a distance ranging from 6.6 km to 20 km. The drone-delivered AED was applied on a mannequin from 1.7 to 8.0 min before EMS arrival in the simulated OHCA events.

Rosamond et al. conducted 35 test flights in a community setting, comparing autonomously drone delivered AEDs with bystander-retrieved fixed-location AEDs.³⁶ In four out of five different OHCA locations the drone was faster than ground search and retrieval of an AED. The ground AED was placed between 77 – 163 m from the OHCA location and the drone was launched between 238 – 393 m from OHCA-location.

Claesson et al. compared 18 drone flights to EMS response times from historical OHCA.³⁸ The drone was based in an area with known extensive EMS delay and ranged a 10 km radius. The median flight distance of the drone was 3.2 km with arrival within 5 min 18 sec (IQR 3:00 – 8:30 min). In contrast, ambulances took a median of 22:00 min (IQR, 17:48 – 29:00) from the EMS station.

Human-drone interaction

Four studies primarily assessed bystander experiences and performance in simulated OHCA scenarios with AEDs delivered by drones.34, 37, 35, 41 Lay individuals, both individually and in pairs, initiated CPR on a mannequin and placed a mock emergency call. The AED was delivered by a drone either beyond-visual-line-of-sight,34, 35 within visual line of sight ³⁷ or from a grounded drone with rotors initially turning.⁴¹ Post-interaction interviews were either unstructured ³⁷ or semi-structured.34, 35, 41

Overall, bystanders' interactions with the drone were mainly positive. All four studies reported that bystanders felt relieved when the drone arrived and saw it as positive and helpful. Kim et al. and Zègre-Hemsey et al. reported that despite some bystanders' hesitancy regarding approaching the drone, they still found it beneficial and would use it in a real-life scenario.34, 35.

Three test flight studies reported bystander experiences as secondary outcomes.19, 31, 36 Fischer et al. found that 93 % of participants (including 10 paramedics and 19 laypeople) felt safe as the drone approached.³¹ Baumgarten et al. conducted structured interviews with 22 Community First Responders and 45 laypeople after each of 46 simulated OHCAs.¹⁹ Among lay responders, 8.9 % hesitated to collect the AED, and 2.2 % found it cumbersome. None of the Community First Responders reported any problems. Observers deemed the interaction between bystanders and the drone as safe.

AED delivery methods

Eight studies landed the drone on-site for a person to fetch the AED directly from the drone.15, 19, 30, 34, 36, 37, 38, 40 Fischer et al. attached the AED to the drone using a long line. Upon arrival, the drone was manually descended until the AED reached approximately 1 m in height, at which point the line was released.³¹ Rees et al. conducted the sole study using a fixed-wing drone with parachute landings for the AED, finding it feasible, albeit with room for improvement in landing precision (about 50 m from the OHCA location).³³

In a 2016 study (included in the category of computer/prediction models), Claesson et al. conducted 13 test flights for AED delivery.¹⁶ They tested three methods: latch release from a 3–4 m height (n = 6), landing the drone (n = 6), and parachute landing of the AED (n = 1). The parachute landing method was deemed too uncertain for precise AED placement, emphasizing the superiority of landing the drone or using a latch release method.

The real-world studies employed a system where the AED was attached underneath the drone using a wire and winching mechanism.18, 28, 29 Upon reaching the designated location, the AED was winched to the ground before the wire was released from the drone. In the first study, the AED was delivered within a median distance of 9 m (IQR 7.5–10.5) from the patient or building.¹⁸ In the second study, the AED was delivered within 15 m of the building or the patient in 91 % of cases.²⁸

Safety of drone AED delivery

Overall, none of the studies in the test flight category reported serious adverse events, including ground risks (e.g., crash, loss of AED, dangerous landings) and air risks (conflicts with other aircraft). Zègre-Hemsey et al. highlighted bystander hesitancy due to propeller concerns,³⁵ while Kim et al. emphasized the heightened risk of drone crashes during take-off and landing on various surfaces across diverse geographic areas, suggesting that landing might not always be feasible.³⁴ Schierbeck et al. reported one instance of emergency parachute deployment during 1 of 61 test flights.¹⁸ Baumgarten et al, noted a minor rotor damage in one out 46 flights during landing.¹⁹ However, no safety concerns were identified, and the remaining studies reported no adverse events.

Society readiness level

A Canadian study by Sedig et al. surveyed a community's readiness for AED-delivering drones, interviewing 65 residents.³⁹ Residents found the concept generally acceptable, but concerns revolved around CPR and AED use. The study emphasized the importance of addressing community OHCA literacy level and information needs for successful drone-delivered AED programmes. In a study by Zègre-Hemsey et al. from North Carolina, 17 lay people in simulated OHCA events expressed concerns mainly about CPR and defibrillation procedures rather than about drone delivery of the AED.³⁵

Computer/Prediction models for drone feasibility and effectiveness

The computer/prediction model category comprised 21 studies investigating the theoretical feasibility and effectiveness of drone-delivered AEDs for OHCAs. These studies utilized sophisticated algorithms and statistical analyses to estimate drone AED coverage and response time reduction compared to standard EMS times. Given the heterogeneity among the studies, we have outlined key themes rather than providing a comprehensive coverage of all topics (Table 3).

AED drone delivery in urban setting

Three studies focused on densely populated urban areas.48, 52, 55 In the greater Paris area, Derkenne et al. found that AED-drone arrival preceded basic life support in 26 % of cases, resulting in a time gain of 190 s (± 7 s) in 93 % of instances.⁵² Choi et al. modelled a drone AED system for the city of Seoul, which is characterized by population density, high rise buildings, and hilly terrain.⁴⁸ They simulated various drone capacities and compared topographic flight pathway with a standard Euclidean flight pathway. Standard Euclidean flight pathways were estimated to arrive before the ambulance in 38 % of cases and topographic pathway 27.0 % of cases. In Bellevue, Washington, Lancaster et al. estimated a 2.3-minute reduction in average response time compared to standard EMS.

AED drone delivery in mixed urbanized settings

Boutilier et al. focused their drone network study on eight regions in Southern Ontario, Canada.⁶⁰ When aiming to reduce the median AED arrival time by 3 min compared to traditional 911 response, the 90th percentile in urban areas experienced a response time reduction of 6 min and 43 s, whereas in rural areas, it was 10 min and 34 s.

In Northern Ireland, AED delivery time was reduced by 25 % in the dense populated region of Belfast and 50 % in the rest of the country compared to ambulance arrival or bystander retrieval of a publicly accessible AED.⁵⁶

Two Swedish studies underscore the importance of balancing placement and quantity of bases according to longer EMS response time and OHCA incidence. A model from Stockholm, placed 10 drones based on high OHCA incidence, reaching 69 % (n = 3041) of OHCAs before standard EMS, with a median time gain of 1.5 min.¹⁶ Alternatively, when OHCAs from areas with longer EMS response times were prioritized, only 3 % (n = 124) of OHCAs would be reached before EMS arrival, but with a mean time reduction of 19 min. A study by Schierbeck et al., highlighted the need for numerous drone bases to cover rural areas of Sweden.⁵¹ For OHCAs with ambulance response times exceeding 8 min, achieving 90 % coverage required 784 drones, escalating to 2408 for 100 % coverage.

While each study in this category employed different methodologies for location optimization to ascertain the ideal number of bases and drones for effective coverage, Chu et al. introduced a machine-learning model.⁵³ This model aimed to minimize drone dispatches for OHCA instances while preserving the initially estimated 2-minute reduction in response time compared to standard EMS. The machine learning dispatch rule programme achieved a 30 % reduction in dispatched drones without compromising time gain or AED-drone coverage.

Survival prediction and cost-effectiveness

Two studies assessed the incremental cost-effectiveness ratio (ICER) for drone AED-systems compared to standard EMS care. Bogle et al. reported a 5-minute response time reduction resulting in a doubled survival rate (12.3 %-24.5 %), adding to an estimated cost of €791 (US$858) per incremental quality-adjusted life year.⁵⁸ Bauer et al. achieved 90 % coverage (defibrillation within 10 min) in Germany with 1074 drones, obtaining an ICER of €14,548.⁵⁴ Röper et al. found drone networks to be more cost-efficient than establishing and maintaining traditional stationary AED networks in a rural German district.⁴⁶

Lancaster et al. compared different systems over a 10-year period, favouring a drone system delivering AEDs for bystanders or volunteer responders, in terms of cost-benefit.⁵⁰ By reducing time to defibrillation, they estimated a 10 % increase in survival over the EMS survival rate. Frigstad et al. modelled a system of combined dispatch of volunteer responders and a network of AED-equipped drones.⁴⁴ With 20 drone launch sites, they estimated an increase in survival from 9.6 % to 15.8 %. Starks et al. estimated the impact of equipping First Responders with AEDs and deploying AED-delivering drones to achieve 5-minute AED access for historical OHCAs across 48 counties in North Carolina, USA.⁶¹ Using a logistic regression model based on 28,292 historical OHCAs, the study predicted that these interventions could increase survival rates for witnessed OHCAs from 14.5 % to 19.4 %.

Discussion

Overall, we found that drone-delivered AED studies spanned three major categories: real-world studies, test flight studies with or without bystanders, and mathematical computer/prediction models for drone AED feasibility and effectiveness. Of the 39 studies included, 31 studies explored drone AED response times – from real-world systems, from test flights, or estimated through computer models. Collectively, they suggested a potential time gain compared to standard AED delivery methods. Despite increasing global interest in utilizing drones for AED delivery to OHCA incidents, only three studies provided data for real-world suspected OHCA. None of the studies involving real drone flights reported any safety issues or serious adverse events.

The identified studies exhibited significant methodological and outcome heterogeneity, leading to the decision against conducting a systematic review and meta-analysis.

The effectiveness of a drone AED system depends on the technical capabilities of the specific drone used, human/societal factors, as well as the geographical and meteorological conditions of the implementation area. In the computer prediction models, various technical capabilities were included, and differing data input were considered. Drone velocities ranged from 47 km/h to 100 km/h,17, 56, 59 with the majority falling within 60–80 km/h. Six studies included the possibility of drones being occupied due to multiple OHCA-incidents,17, 47, 52, 53, 60, 61 two studies included topography,48, 57 and some studies considered possible flight cancellation factors varyingly, e.g., weather, visibility, and time of day.31, 48, 50, 52, 55 While most studies evaluated drone AED effectiveness by comparing it to traditional EMS (ambulances and/or professional first responders), two studies included the probability of OHCA witnesses being able use the AED,52, 58 and three studies compared drone AED response time to estimated volunteer responder response times.50, 55, 56

While theoretical models provide valuable insights, real-world studies highlight significant practical challenges. In the latest and largest Swedish real-world study drones could be dispatched in 34 % of suspected OHCAs, with successful AED delivery achieved in 81 % of those cases.²⁸ Furthermore, as highlighted by two studies in this review 39, 35 and supported by Ringh et al., Dainty et al., and Brooks et al., the effectiveness of AED systems involving lay people highly depend on the societal readiness level.62, 63, 13 It is essential, however, to acknowledge that OHCA literacy is likely to increase over time, entailing a growing number of individuals capable of providing CPR and using AEDs. Additionally, anticipated technological advancements are expected to improve the systems and reduce the associated costs.

Comparing outcomes across prediction models was challenging due to variations in key outcomes, such as time gain, area coverage, cost-effectiveness, and survival estimates, each relying on differing underlying assumptions. For instance, some studies estimated the number of drones needed to reach all or varying proportions of OHCAs within a set response time,17, 43, 59, 60, 61 while others evaluated fixed drone numbers against EMS times. 16, 45, 50, 55 When comparing drone AED delivery to EMS response times in rural areas, the estimated response time reduction was significant, ranging from 10:34 min in Boutilier et al. to 19 min in Claesson et al.16, 60 Although the time gain in urban areas was smaller (6:43 and 1:30 min in these two studies, respectively) the overall number of OHCA cases benefiting from drone-delivered AEDs is considerably higher in urban settings due to the greater incidence of OHCAs. Overall, drone deployment showed promise in addressing health inequities by bridging geographical gaps and enhancing emergency medical services across diverse settings.

Estimating survival gain and cost-effectiveness is particularly challenging due to the numerous underlying assumptions involved.43, 44, 46, 50, 54, 55, 58, 61 Estimated survival ranged widely from 15.8 % – 64 %.43, 44 Six of the seven studies estimating survival gain,43, 44, 50, 54, 55, 58 based their predictions on previous literature.9, 64, 65, 66 In contrast, Starks et al. incorporated a logistic regression model based on their own data, accounting for potential changes in the first recorded shockable rhythm due to earlier AED arrival facilitated by a drone-network combined with First Responder AEDs.⁶¹ However, the lack of robust survival estimates tied to defibrillation time, combined with the reliance on numerous assumptions, adds complexity to the analysis and underscores the need for cautious interpretation of survival benefits. Despite great variation in estimated incremental cost-effectiveness (Bogle et al, €791 and Bauer et al. €14,548),54, 58 both drone systems were considered cost-effective based on thresholds for high-income countries.⁶⁷

The included studies collectively underscore regulatory constraints as a significant barrier to the broader implementation and testing of drone-delivered AED systems across regions and countries. The barrier revolves around the lack of existing Unmanned Air Traffic Management regulations specifically designed for beyond-visual-line-of-sight flights.

Future research should be focused on real-world studies looking at reliability and local system integration, as well as regulatory aspects and cost-effectiveness studies, which are all essential for further exploration of the potential of AED-delivery by drones. Addressing these factors, along with standardizing protocols for drone integration into OHCA response systems, will be crucial for the wider adoption of drone-delivered AEDs.

The limitations of this scoping review include the heterogeneity of study designs and methodologies, complicating direct comparisons between studies. Additionally, the lack of detailed assessment of the quality or rigor of individual sources and the reliance on simulated and theoretical models, with few real-world implementations, limits the generalizability of the findings.

Conclusion

This ILCOR scoping review provides a comprehensive overview of the type and volume of existing literature on drone-delivered AEDs for OHCA. Collectively, the included studies suggest that drone technology holds promise in reducing response times, increasing survival rates, and improving cost-effectiveness in OHCA scenarios. However, further research, full-scale real-world studies, and regulatory developments are needed to facilitate the integration of drones into OHCA response systems. Our findings consistently showed that drone-delivered AEDs are considered feasible in all included studies across the three categories, and many studies predicted time gains and increased survival rates, as well as positive human-drone interaction.

CRediT authorship contribution statement

Louise Kollander Jakobsen: Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Data curation. Victor Kjærulf: Writing – review & editing, Investigation, Formal analysis, Data curation. Janet Bray: Writing – review & editing, Supervision, Project administration, Methodology, Conceptualization. Theresa Mariero Olasveengen: Writing – review & editing, Conceptualization. Fredrik Folke: Writing – review & editing, Supervision, Project administration, Investigation, Conceptualization.

Funding

This scoping review was funded by the American Heart Association, on behalf of The International Liaison Committee on Resuscitation (ILCOR). The work has been supported by TrygFonden (ID 157642) and the Novo Nordisk Foundation (grant number NNF19OC0055142). TrygFonden and Novo Nordisk Foundation had no influence on the study planning or conduction. JB received a fellowship from the Heart Foundation of Australia.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Janet Bray is an Associated Editor and Theresa Mariero Olasveengen an Editorial Board member of Resuscitation Plus. JB and TMO are Editorial Board members of Resuscitation.

Appendix A. Supplementary data

The following are the Supplementary data to this article: