Strategic PAD positioning: a scoping review Authors: Luka Petravić, Kaja Gril Rogina, Rok Miklič, Mitja Lainscak, Tinkara Langus, Anže Žnidaršič, Andreja Kukec, Matej Strnad Published: 2025 ABSTRACT BackgroundOut-of-hospital cardiac arrest (OHCA) is a major contributor to overall morbidity and mortality. Survival depends on several factors, including rhythm shockability and timely defibrillation; for the latter, the availability of Public Access Defibrillators (PADs) is crucial. Across countries, PADs can be a limited resource, necessitating strategic placement to ensure optimal accessibility and effective utilization. The objective of this scoping review was to analyse existing experience with PAD positioning and to identify potential for the geospatial optimisation of PAD positioning.MethodsAn electronic search of PubMed and Scopus databases was performed to identify studies on PAD positioning. We included peer-reviewed articles as well as graduate and undergraduate theses, and excluded conference abstracts, books, and book chapters. Only English-language publications were considered. A Risk of Bias analysis was conducted and reported using Mixed Method Appraisal Tool. We adhered to PRISMA-ScR guidelines. Using the PCC framework, the Population was the general public, the Concept was PAD optimization, and the Context ranged from buildings to nations. Data charting was done manually. We quantitatively synthesized findings from included studies based on extracted data and Synthesis Without Meta-Analysis guidelines.ResultsWe included 43 works, mostly from North America. We found substantial diversity in study methodologies, outcome measures and results reporting. Within buildings, PADs are recommended to be positioned centrally or in elevators. Outside, places with high-footfall were preferred. Optimisation models most often employed the Maximal Covering Location Problem. Most studies based their models on the locations of existing and candidate PADs, alongside historical or simulated OHCA occurrences, and tried to minimise that distance.ConclusionsIn this scoping review of studies optimising PAD network deployment, we found that PAD network optimisation is feasible and improves coverage of historical OHCA locations. Data-driven positioning strategies should incorporate historical OHCA locations, population health indicators, and socioeconomic variables to identify high-risk zones effectively and inform appropriate PAD density. Future studies should adopt prospective and real-world design, while adhering to guidelines for reporting results.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s13049-025-01527-8. Background Out-of-hospital cardiac arrest (OHCA) affects 82.3 patients per 100,000 inhabitants per year in Europe, with overall survival of 4.5 per 100,000 [1]. Resuscitation can be initiated by non-EMS-first responders (e.g., trained laypersons, police, fire fighters), EMS or laypersons [2]. After basic life support and calling emergency services, Public Access Defibrillators (PADs) should be used. PADs detect shockable rhythms and deliver shocks, which must be administered quickly to prevent deterioration into non-shockable rhythms [3]. PADs designed for layperson use reduce time to first defibrillation [4], improve OHCA survival 1.36–2.18-fold, and increase favourable neurological outcomes 1.36–3.29-fold [5, 6]. Despite rising PAD availability, layperson defibrillation before EMS arrival remains low [7–10], with PAD use rates of 12–59% [11]. Contributing factors include low bystander response rates [8], unregistered nearby PADs [12], reluctance to use them [13] and limited access [14]. The latter is not surprising as most OHCAs occur in residential areas, usually homes [8, 9, 15], while most PADs are in public spaces [9, 16]. Spatial optimisation of resources helps determine best coverage, reduce costs, and maximize resource use [17, 18], and could similarly improve PAD positioning [19]. The International Liaison Committee on Resuscitation highlights strategic geospatial placement as key to increasing bystander PAD use [20]. Cheng-Heng et al. [19] found no universal approach to PAD positioning. Tools like PadMap can assist in translating theory into practice [21]. Guidelines discuss strategic placement but remain open to interpretation [9, 10, 22]. Evidence on optimal PAD placement and evaluation methods is limited. To map existing research and identify gaps, we conducted a scoping review asking: What strategies, evaluation methods, and outcomes have been used to optimize PAD placement, and what are priority directions for future research? Our primary objective was to identify optimal PAD placement strategies; secondary objectives examined optimization methods, contextual variables, and effects of different approaches. Methods We report our scoping review according to PRISMA-ScR [23] (Supplement 1). The study protocol was preregistered [24]. During the charting step the review type was changed from systematic to scoping, due to the heterogeneity of included studies both methodologically and in reported results. Using the PCC framework, the Population was the general public, the Concept was PAD placement optimization, and the Context ranged from buildings to nations. Eligibility criteria We restricted our search to exclude conference abstracts and books or book chapters. Only results published in English language were included. Graduate and undergraduate theses were included, on the grounds that these works provide a valuable contribution to the evidentiary base for recommendations presented in this paper. Information sources and search strategy We performed electronic searches of online databases PubMed [25] and Scopus [26]. Search fields were restricted to title/abstract. Our search strategy was informed by a previous systematic review [19]. The final search strategy was: (Defibrillators OR Defibrillators OR Electric Shock Cardiac Stimulator OR Automated External Defibrillator OR External Defibrillator OR AED) AND (Location OR Placement OR Distribution OR Site OR Network OR Deployment OR Deposit) AND (Optimisation OR Improvement OR Strategy OR Access). Dates searched were between 1/1/2010–2/1/2025. Selection process and data extraction Two authors (L.P. and K.G.R.) screened the extracted studies based on title and abstract. To increase consistency among reviewers, all reviewers screened ten randomly selected publications, discussed the results and amended the screening and data extraction manual before beginning screening for this review. Disagreements between the reviewers were resolved through consensus by a third reviewer (R.M.). Full texts were obtained for selected articles. Two authors (L.P. and K.G.R.) reviewed each article’s eligibility for inclusion. A data-charting spreadsheet was jointly developed by two reviewers (L.P. and K.G.R.) to determine which variables to extract. Three authors (L.P., K.G.R and R.M.) performed data extraction (see supplement 2). One author performed the data extraction for each article, and two authors double checked the extracted data. No subgroup analysis was carried out. For each included work we extracted study title, country of study, area covered, number of PADs in network, methods used for optimisation, list of all variables used for optimising network, population in which the network was optimised, incidence of OHCA in the area where network is optimised, locations where PADs were placed with highest priority, temporal and spatial components used, PAD density per km2, effect measurement methods, optimisation effects, outcomes and future directions. Critical appraisal of individual sources of evidence Risk of Bias was assessed using the Mixed Method Appraisal Tool (MMAT) [27] and three additional dimensions from the authors: “study environment” (real world vs. in-silico), “model code published” (yes/no) and “type of model validation” (internal/external/mixed). Synthesis methods and outcomes Due to substantial heterogeneity in study designs, populations, outcomes, and reporting formats, a meta-analysis was not feasible. Consequently, findings were synthesized narratively, following the Synthesis Without Meta-Analysis guidelines [28]. We employed the SR-Accelerator [29] and Zotero (RRID:SCR_013784) to support the writing of this scoping review. Results Search and selection Our scoping search resulted in 1329 eligible articles. After the selection process, 42 articles [30–70] and one graduate thesis [71] were included in the review. Two articles met the criteria but were written in a foreign language and excluded [72, 73]. Figure 1 shows the PRISMA 2020 flow diagram of the study [74].Fig. 1Prisma 2020 flow diagram of manuscript selection Prisma 2020 flow diagram of manuscript selection Basic characteristics of included studies Baseline characteristics of included studies are summarized in Supplement 3, Table 1. As shown in Fig. 2, most studies were conducted in North America (n = 15). Positioning optimisation was primarily evaluated in outdoor settings, though several studies investigated in-building positioning [30, 49, 57, 67]. Population coverage varied from as few as 20 individuals [49] to over seven million [44]. Area coverage was inconsistently reported and ranged from single buildings [36, 49] to cities [33] and whole regions [39]. The number of PADs considered in optimisation ranged from one [46, 49] to 9,880 [40], with several studies omitting this detail [34, 37, 39, 43, 52, 57, 64, 67, 68, 70, 75]. Common outcome metrics included spatial coverage, distance to the nearest PAD, coverage efficiency ratios, retrieval times, percentage of population or historic OHCAs covered and deployment cost-effectiveness. Five studies were endorsed by societies related to resuscitation or cardiovascular health [38, 43, 63, 68, 69].Fig. 2Geographic distribution of included studies Geographic distribution of included studies Primary outcome: where should PADs be placed? Included studies analysed PAD positioning across varied settings and geographic contexts, making comparison difficult. Recommended positioning locations are summarized in Fig. 3. Limited research exists on indoor positioning (n = 4); PADs are generally recommended to be positioned centrally within buildings or in elevators [30, 36, 49, 57]. Most included studies (n = 39) explored how to place PADs outside. In public spaces, high-footfall venues such as restaurants, cafes, bars, and pubs [38, 43, 58, 65, 66] are commonly identified as suitable locations. For outdoor positioning, leveraging existing infrastructure such as bike-sharing stations [38], ATMs [58], and postal boxes [43, 59] is suggested. Sports facilities represent a distinct category and are consistently recommended to have PADs on site [68]. Areas with high incidence of OHCA [45, 48, 55], OHCA risk [31, 45, 50, 64] and socioeconomically deprived areas [41, 55, 57] are recommended to be equipped with more PADs. A denser PAD network is advised in urban areas compared to rural ones [42, 51, 55].Fig. 3Primary outcomes: where should PADs be placed with higher priority Primary outcomes: where should PADs be placed with higher priority Secondary outcomes Methods employed for strategic PAD positioning We found seven method families, which range from basic to more complex, like machine learning models (Fig. 4).Fig. 4Employed methods for optimising the PAD network, grouped by method family and showing the number of times the method was used in included studies Employed methods for optimising the PAD network, grouped by method family and showing the number of times the method was used in included studies Optimisation models most often employed the Maximal Covering Location Problem (MCLP) [30, 31, 35, 37, 47, 52, 55, 56, 62, 65, 68, 69] to identify PAD positioning that maximizes historical OHCA coverage within spatial or temporal constraints. Simulation and modelling techniques further refined these strategies and enabled the dynamic testing of PAD deployment scenarios, incorporating spatial and temporal accessibility [61]. Statistical and machine learning methods provided predictive capabilities, enhancing the capacity to model OHCA risk and potential PAD configurations [33]. Heuristic and metaheuristic algorithms were used to address the computational complexity of large-scale location problems and offered adaptive solutions to optimisation models [31]. GIS (geographic information system) and spatial methods ensured precise geospatial planning. Some studies leveraged tools such as QGIS (Quantum Geographic Information System) [32, 39] and real-world walking routes [32] models to ensure geospatial precision and to more accurately estimate walking distances and travel times, enhancing accuracy in urban settings. Empirical and survey-based studies uncovered OHCA hotspots [42], sociodemographic barriers [75] and which businesses are best placed to improve PAD coverage [38]. Additional models, such as the Maximum Expected Survival Rate Model/Maximum Coverage Model were integrated to directly maximise survival rates [40] or efficacy in specific environments such as golf courses [46], high rise buildings [57] or 3D indoor environments [30]. Variables adopted for positioning optimisation A broad range of variables was used to inform PAD positioning optimisation, reflecting the complex determinants of OHCA risk, accessibility, and response dynamics (Fig. 5). Most studies based their models on the location of existing and candidate PADs [31, 38, 40–42, 45, 47–50, 53–55, 58, 61, 63, 68, 69] alongside the historical or generated OHCA occurrences and locations [31–33, 36, 38–45, 47, 48, 50, 51, 53–55, 61, 63, 65–70, 75].Fig. 5Employed variables for optimising the PAD network, grouped by variable family and showing the number of times the variable was used in included studies Employed variables for optimising the PAD network, grouped by variable family and showing the number of times the variable was used in included studies PAD-specific data such as the type of device (fixed, summonable, patrolling) [36, 42], and temporal availability of devices [34, 50, 53, 55, 58, 61, 65], alongside optimised/potential PAD sites [35, 37, 45, 52–54, 65, 66] were frequently utilised. Historical OHCA data included details such as the cause [42], location type [32, 42], incident details and risk metrics [32, 39, 45, 64], and clinical outcome predictors [33, 40], which were used to establish high-risk areas. Temporal dimensions were captured through variables related to emergency service response times [32, 33, 70] and PAD retrieval times [30–33, 36, 39, 40, 46, 57]. Spatial coverage and accessibility measures [34, 44, 44, 47, 49, 65, 69], such as demand nodes and hosting facilities data [37, 52, 56, 65, 66], as well as spatial/demand considerations [35, 42, 43, 45], were widely used to estimate effective coverage. Many countries have successful volunteer programs, whose data have been used [33]. Environmental and infrastructural variables included building characteristics [49, 55, 57], operational hours, pedestrian traffic indicators [47, 49, 50, 59, 65], potential location category data [58], road network distances [32, 46, 54, 56], and other geospatial measures [39, 64, 67, 75], which further enhanced the precision of position modelling. Demographic and socioeconomic variables, including population density [50, 55, 56, 60, 75], demographic structure [45, 48, 51, 60], socioeconomic and land-use data [41, 43, 48, 51, 60, 64, 71], and cost parameters [33], were employed to identify high-risk or underserved areas. Finally, several studies incorporated additional factors such as travel modes [31], emergency call data [37, 43], timing and routing details [31], planning and regulatory constraints [35, 37, 75], local context and demographics [34, 75], and operational safety and readiness [34, 75] to increase real-world applicability. Optimisation effects PAD optimisation strategies demonstrated consistent improvements across multiple domains (Supplement 3, Table 2). Coverage increased substantially across studies, ranging from 1 to 175% depending on baseline infrastructure and intervention scope [31, 35, 47, 48, 50, 53]. Strategies combining device reallocation with new placements proved particularly effective, with reallocation alone improving coverage by 23%, while combined approaches yielded greater gains [42, 48]. These accessibility improvements translated into measurable clinical benefits. Sun et al. [47] reported increased bystander defibrillation (14.6% to 26.9%) and improved 30-day survival (31.3% to 35.4%). Advanced optimisation algorithms further enhanced performance: robust models improved retrieval distances by 9–20% [62], while other approaches reduced retrieval time by 98 s—suggesting an 11–16% survival increase [66]. From an implementation perspective, optimisation proved both cost-effective and equitable. Despite requiring up to a 15-fold increase in device numbers in some scenarios, optimisation remained cost-effective at €80,000/QALY [31]. Additionally, optimisation naturally directed resources toward underserved populations with higher OHCA incidence [41]. Critical appraisal of individual sources of evidence Full critical appraisal of individual sources of evidence is reported in Supplement 3, Table 3. MMAT quality ratings were generally high (grade 4), though some scored lower (grade 3). We expanded on the MMAT by using our own dimensions. Specifically whether the validation was internal or external, whether the model was made public and whether the study was a real world or an in-silico study. One study used external validation [53], others were internally validated or did not perform result validation at all. A minority of studies were based on real-world scenarios (n = 19). Only one study published the model code [37]. Six studies did not specify if the model code is publicly available [42, 46, 47, 49, 62, 76]. Most studies didn’t confirm optimisation results in real-life settings. We observed a lack of uniform reporting among different research groups. Discussion Cheng-Heng et al. [19] conducted the first systematic review on public PAD location strategies. They identified three main positioning strategies: guideline-based, landmark-based, and grid-based. Their findings emphasized the need for both spatial and temporal optimisation [19]. Different guidelines that touch on the issue of strategic PAD positioning recommend that PADs should be placed in areas with a high risk of OHCA [9, 10], inside sports facilities [22]. A best practice example is set by the Japan Foundation for Emergency Medicine. They offer a list of 16 location types where PADs should be placed, including residential areas [76]. Our review builds on this foundation and offers more specific suggestions as to PAD network design and research. Domains of PAD positioning strategies We found that PAD positioning strategies can be broadly categorized into three domains shown below. Research shows that optimal positioning of community first responders and PADs is up to 500 m from any address, but this might only be useful in cities with perpendicular street designs [77]. Building based Building-based positioning refers to the internal positioning of PADs within specific structures [30, 49, 57, 67]. Optimal positioning is either centrally aligned with the building’s geometric layout or in elevators [30, 49, 57, 67]. The latter approach is particularly practical, as elevators are typically located to provide equitable access across all areas of the floor. Moreover, elevators are universally used and well-recognized landmarks within buildings. Location based Location-based positioning can be subdivided into four categories. (I) Public spaces include high-traffic venues such as restaurants [38, 43, 58, 65, 66], government buildings [32, 35, 65, 66], train stations [9] and schools [39, 68]. This approach aligns with existing national and international guidelines [9, 22, 76, 78, 79]. Japanese guidelines specifically recommend PADs in schools, despite a relatively low incidence of OHCA, on the basis that younger individuals tend to have higher survival probabilities [76]. (II) Outdoor urban infrastructure includes postal collection boxes [43, 59], ATMs [58], public bike-sharing stations [38], and notable landmarks [38]. While the original design of these infrastructures was not intended for PAD deployment, they offer an existing grid network that can be repurposed. Landmarks and similar outdoor locations attract significant pedestrian traffic, increasing the probability of a bystander-witnessed cardiac event [35]. (III) Sports facilities are associated with a higher risk of OHCA due to physical exertion and the potential presence of undiagnosed cardiac conditions. (IV) Residential deployment strategies should account for population density, with a denser PAD network in urban areas compared to suburban zones. High-rise residential buildings, which accommodate many individuals in a confined vertical space, are particularly suitable for PAD positioning. Given that the majority of OHCAs occur in the home, this category remains critically underrepresented, which was noted by European resuscitation council (ERC) resuscitation guidelines [9]. Japanese guidelines argue that due to the higher likelihood of shockable rhythms and bystander presence, public spaces are more suitable for PAD positioning than residential environments [76]. While we acknowledge the rationale provided, we partially disagree, proposing the following hypothesis: in many residential cases, the initial shockable rhythm deteriorates into a non-shockable rhythm due to delayed defibrillation. This assertion is supported by recent research [80] and emphasizes the need for more strategic PAD positioning. We agree with the Japanese workgroup on the limited utility of PADs in residential areas without witnesses [76]. However, technological advances, such as smart wearables that detect the loss of pulse [81], volunteer responders dispatched via mobile apps [82], and rapid PAD delivery by drones could help overcome the barrier of absent witnesses, making residential PAD deployment more effective [9]. Risk based In risk-based deployment, PADs are placed in areas with a high risk of OHCA [31, 45, 50, 64], mostly in areas with an elderly population [48], historical high incidence [45, 48, 55] and in socioeconomically disadvantaged areas [41, 55, 57]. Other considerations Cost-effectiveness analyses indicate that the strategic expansion of PAD networks remains within acceptable willingness-to-pay thresholds while reducing the necessary number of PADs by up to 15 times to achieve equivalent coverage compared to non-optimised approaches [47]. There is no data regarding the average use of a PAD throughout its lifespan. This makes calculating QALYs or price per QALY difficult, although some researchers have studied the cost-effectiveness of PAD networks in relation to QALYs. Folke et al. found that unguided PAD positioning in Copenhagen, assuming an 80% probability of PAD use, costs $135,900 per QALY [83]. Methodologies and variables The secondary aim of our study was to identify specific methods used for the optimisation of PAD positioning, list the variables employed, and summarize the effects of optimisation. The MLCP was the most used method [30, 31, 35, 37, 47, 52, 55, 56, 62, 65, 68, 69]. We identified seven method families and eight variable families [30–70]. Most often existing and candidate PAD locations [31, 38, 40–42, 45, 47–50, 53–55, 58, 61, 63, 68, 69], PAD hours of operation [34, 50, 53, 55, 58, 61, 65] and historical OHCA occurrences [31–33, 36, 38–45, 47, 48, 50, 51, 53–55, 61, 63, 65–70, 75] were used. Optimisation effects We found consistent evidence that strategic PAD positioning significantly improves OHCA coverage (a 1-175% increase) [31, 35, 47, 48, 50, 53], with optimised networks achieving 38–74% coverage [35, 55] compared to 15–30% in conventional deployments [47, 49]. Relocating existing PADs often matched the benefits of doubling PAD numbers. Optimised networks reduced retrieval times, which could potentially improve patient outcomes. Even with an optimal PAD positioning strategy, the critical question of PAD maintenance and accountability remains unanswered [11]. There have been reports of PADs malfunctioning due to inadequate maintenance, failing to deliver shocks in critical moments [84, 85]. In addition to strategies for optimal PAD positioning, it’s crucial to have a well-considered policy in place for network maintenance and upgrades [86]. Critical appraisal of individual sources of evidence We found that all studies fell between grades 3 and 4, indicating a low risk of bias in the reported results. This aligns with the findings of a previous review [19]. Most studies used internal validation. External validation is necessary to make the reported results more robust and applicable beyond the training data. Most studies were in-silico. We also observed heterogeneity in reporting across research groups. To address this, we recommend international harmonization efforts, such as Delphi-based expert consensus, to develop shared definitions (e.g., coverage metrics, retrieval time, and distance), and validation protocols. Adoption of such consensus-based standards by journals and professional societies would enhance comparability and implementation readiness. Future directions Authors of included studies mentioned that multiple directions should be explored to improve coverage and outcomes. Priority areas include incorporating volunteer first responders and transportation modes in emergency response models, transitioning from straight-line to actual walking route calculations, and developing three-dimensional frameworks that account for vertical distances in buildings. Optimisation approaches should target health outcomes like QALYs rather than simple coverage metrics and address temporal variations. Technological innovations should focus on smartphone applications for PAD location and drone or vehicle-based delivery systems. Implementation strategies should emphasize 24/7 accessibility, relocation of poorly positioned devices, and centralised coordination of PAD networks integrated with emergency services. Future studies would benefit from machine learning to predict OHCA locations, real-world validation of theoretical models, and rigorous assessment of how optimised positioning impacts actual survival rates beyond theoretical coverage improvements. We believe that Australia could be a good study environment, with strategic PAD positioning legislation going into effect in 2026 [79], making it possible to compare OHCA responses and outcomes before and after. Another area is AED drone delivery, which could potentially impact response times, cost-effectiveness, and possibly even survival rates [87]. Currently, the body of evidence is limited, lacking real-world studies showing these effects, although published studies demonstrate the feasibility of such technology use [87]. In line with guidelines’ call for additional research, future studies should evaluate the cost-effectiveness of optimized PAD networks against traditional willingness-to-pay thresholds, investigate optimal dispatcher instruction protocols for AED retrieval using real-time registry data, assess the impact of comprehensive risk mapping on patient outcomes, examine the balance between CPR continuation and AED retrieval in lone-rescuer scenarios, and develop standardized procedures for device redeployment after use as recommended by the ERC [9]. Recommendations Our findings align with and expand on the 2025 ERC and American Heart Association (AHA) guidelines on public access defibrillation systems [9, 10], which emphasize deploying AEDs in high-traffic locations like airports, shopping centres, train stations, and in communities with higher incidences of cardiac arrest [9, 10]. Our review adds granularity for PAD network optimization. Specifically, our analysis strongly supports current ERC recommendations for AED places that have longer working hours, are open 24/7 or allow for the outside placement of PADs [55, 65] and transportation hubs [34, 68]. Building on the ERC recommendation, our analysis identified specific location types, such as restaurants, cafes, bars, and pubs [38, 43, 58, 65, 66], sport facilities [68], pharmacies and local clinics [32, 38], convenience stores [32, 67, 70] and schools [39, 68]. Consistent with the guidelines, we propose a comprehensive geographical analysis incorporating OHCA incidence mapping [45, 48, 55] to identify areas with documented higher cardiac arrest rates [9]. These can be affected by demographic risk factors including elderly population density [48], socioeconomic determinants to address health equity in deprived areas [41, 55, 57] and informing decisions by multifactorial risk modelling that combines age, socioeconomic status, and disease burden [31, 45, 50, 64]. Researchers should implement MCLP [34, 44, 47, 52, 55, 65, 66, 68] as a baseline to compare methodologies. When selecting variables for network optimisation, we recommend including at least the following: existing PAD locations [38, 40–42, 45, 47–49, 55, 58, 61, 63, 68], hours of operation for those location [53, 55, 58, 61, 65], OHCA locations [31–33, 36, 38, 40, 42, 44, 45, 47, 50, 51, 53–55, 61, 63, 65–68, 70, 75], EMS response times [33], volunteer response times [33], road networks and distance calculations [46, 54, 56], population density metrics [50, 55, 56, 60, 75], socioeconomic factors and land use data [41, 43, 48, 51, 60, 64, 71]. Beyond guidelines, we recommend developing national OHCA risk maps, which was already done by some research groups [88, 89], as certain diseases predispose individuals to OHCA [90]. At the European level, tools such as the European Community Health Indicators & Monitoring, where countries report their indicators (e.g., high blood pressure, high body mass index, smoking, social support, etc.), could be used locally to determine where the network should be more concentrated [91]. This data should inform future PAD positioning research and guideline updates. Limitations Improved PAD coverage is an intermediate outcome, not a direct measure of clinical benefit. Optimal PAD placement can increase survival potential but does not guarantee high OHCA survival, which depends on timely bystander intervention [91]. Cardiac arrests must be witnessed, and bystanders must locate and use PADs [91]. Community education and complementary strategies, such as 24/7 accessibility, maintenance programs, and integration with EMS/first-responder systems remain essential [9]. No prospective trials compared OHCA outcomes before and after optimisation. Future reviews should focus on a single aspect of PAD positioning, either indoor or outdoor. Comparisons were challenging due to heterogeneity. We used MMAT, which is a general framework and not specifically designed for network optimisation studies; TRIPOD might be better for machine learning approaches [92]. Due to most studies using only internal validation, the reported results might differ in different environments. Most studies were graded 4 [19] with MMAT and many were graded 3 [19], increasing the risk of bias. Suggested solutions should be reviewed in a regional context. Due to studies using diverse methodologies and pursuing different objectives, it was challenging to synthesize the findings. Consequently, the combined recommendations may be less applicable. Some variables were highly specific and not generalizable, so they were excluded. Chan et al. Studied the optimisation of high-rise building design with a particular focus on elevators [57]. Another study assessed the use of the 7-Eleven chain of convenience stores, which are not present in all countries [70]. In many instances, the methodologies were not explicitly described [59], but these studies were included to maintain the review’s broad scope. Limiting the database search to English-language publications, due to the team’s language proficiency, constitutes a potential source of language bias. This restriction may have resulted in the exclusion of relevant non-English studies, thereby limiting the comprehensiveness of the review. Conclusions and clinical implications In this scoping review of studies optimising PADs network deployment, we found great diversity in study methodologies, reported results formatting and measured outcomes. Findings consistently indicated that existing PAD positioning is often suboptimal; strategic positioning of even a few units can significantly enhance coverage. Most OHCAs occur at home and are inadequately covered, highlighting the importance of incorporating residential accessibility into PAD positioning models. Supplementary Information Supplementary Material 1. PRISMA-ScR ChecklistSupplementary Material 2. Additional Methodology descriptionSupplementary Material 3. Additional Tables Supplementary Material 1. PRISMA-ScR Checklist Supplementary Material 2. Additional Methodology description Supplementary Material 3. Additional Tables