3.1. Bibliometric Analysis:
It is necessary to capture and present the distribution of research to help identify any gaps within the evidence-base and frame any narratives toward the distribution of research, and the potential sparsity of research. In total, less than 1% (0.02%, N = 56) of the studies identified during the initial systematic identification of relevant literature was considered eligible for analysis. A majority of all eligible studies represented UK perspectives (58.9%; n = 33), with 28.5% (n = 16) representing perspectives from the USA, and 12.5% of studies (n = 7) having an undefined country of origin. Analysis of publication count revealed an increase in publications over the last 21 years. However, by year of publication, 2018 to 2020 collectively accounted for 33.9% (n = 19) of all studies published, with the remaining years accounting for 76.1% (n = 37; M = 2.1 studies per year) of all studies published. An increase in critical incident events – specifically terrorism - appeared to have accounted for 57% of the variance in the increase in publication count between 2000 and 2017. Further, analysis of open access data relating to terror-related incidents (obtained via the Global Terrorism Database) revealed a significant correlation between critical incident events and publication count, rs (16) = .755, p < .001 (see Fig. 3, for a visualisation and forecast plot).
Across the 56 studies, 49 unique first authors were identified, with only 7 authors publishing more than one study (Alison, Brown, Cohen-Hatton, Kapacu, Power, van den Heuvel, Waring). Cohen-Hatton and van den Heuvel were the only authors to examine singular emergency response agencies (fire and rescue services, and police services, respectively), whilst the remaining authors publishing more than 1 study examined decision-making across multi-agency systems.
To assess the heterogeneity of evidence, and overall distribution of evidence across macro-, meso-, and micro- levels, literature was coded to identify first order and sub-level narratives. The cells in Table 3, represent a count of the papers by category with emphasis on the primary, secondary and tertiary narratives. Note, papers often featured more than one narrative. However, this coding did not attempt to capture or code peripheral narrative undertones (i.e., a sub-surface narrative, such as a socio-political factor, that may have existed within the research).
Inspection of the first order narratives revealed moderate heterogeneity, with 47.4% (N = 27) of papers as fitting within the definition of macro-centric research, 39.29%, (N = 22 as fitting within the definition of meso-centric research, and 12.5% (N = 7) as fitting within the definition of micro-centric research. Coding of the secondary narratives revealed that meso-centric and micro-centric research was the most coded categorisation (12.3%; N = 7), followed by macro-centric (7%; N = 4). Thirty-eight papers did not have an identifiable secondary narrative. Coding at the tertiary level, revealed that 1.8% (N = 1) of the studies had macro-centric characteristics, with micro-centric characteristics accounting for 3.6% (n = 2).
At a more fine-grained level, interoperability, communication, and coordination was the most identified factor (28.1%; N = 16), followed by political, policy and procedural factors (22.8%; N = 13). A full breakdown can be found in Table 3.
Table 3
Narrative bibliometric coding, showing the identification of first-order narratives categorised against sub-level narratives. Note. Cells in bold represent areas of research where factors related to emergency response decision-making were lower than expected, if all narratives were evenly distributed across macro-, meso-, micro- levels.
Sub-Level Narratives | First-Order Narratives (%) |
Primary | Secondary | Tertiary |
Macro-Centric | |
Politics, Policy, and Reform | 13 (22.8) | 2 (3.5) | 0 (0.0) |
Emergency Management Doctrine | 6 (10.5) | 2(3.5) | 0 (0.0) |
Ethical Issues | 7(12.3) | 0 (0.0) | 0 (0) |
Economic Issues | 1(1.8) | 0 (0.0) | 1 (1.8) |
Meso-Centric | |
Interoperability, Communication & Coordination | 16 (28.1) | 4 (7.0) | 0 (0.0) |
Technology | 1 (1.8) | 0 (0.0) | 0 (0.0) |
Policy, Procedure & Process | 0 (0.0) | 2 (3.5) | 0 (0.0) |
Organisational Structure | 1(1.8) | 0 (0.0) | 0 (0.0) |
Risk, Uncertainty & Avoidance | 4 (7.0) | 1 (1.8) | 0 (0.0) |
Micro-Centric | |
Decision Inertia | 3 (5.3) | 3 (5.3) | 0 (0.0) |
Risk Aversion | 0 (0.0) | 0 (0.0) | 0 (0.0) |
Other Cognitive Constraints | 2 (3.5) | 1 (1.8) | 1 (1.8) |
Therapeutic Jurisprudence | 2 (3.5) | 3 (5.3) | 1 (1.8) |
3.2. Holistic Narrative Analysis and Discussion:
Six narratives that were thought to comprise decision-making for critical incident response emerged from the analysis: (1) political reform and modernisation of emergency management doctrine; (2) difficulties of operating under austerity; (3) uncertainty and accountability; (4) inter-intra government and organisational ethics; (5) failures in collaborative information networks; and (6) limited research-focused horizon scanning.
Each narrative was considered at the primary and sub-level and included consideration for where states-of-emergencies revolve around a state of preparing and responding. For example, macro-centric analysis identified factors that were influenced by the interactions between geo-political, legal and ethics dimensions; meso-centric concerned the identification of communication networks and the stratification of organisations; and micro-centric analysis identified localised cognitive-driven factors that impacted decision-making (e.g. decision inertia).
3.2.1. Narrative One: Political Reform and Modernisation of Emergency Management Doctrine
Political reform and modernisation were found to be integral to enhancing emergency response resilience and preparedness to emerging and increasing threats (Kapucu, 2009). This was identified in several papers that discussed the need for political, policy and procedural reform (e.g., Banipal, 2006; Birkland et al., 2011; Gerber, 2007; Kapucu, 2009; Kapucu, and Garayev, 2011; May, 2004; Scavo, Kearney, and Kilroy Jr, 2008; Stevens, 2020). For example, a 2001 review of emergency response and preparedness procedures highlighted the complexity and lack of capacity for information provision, action, and management of emergency response at local levels (Home Office, 2001). The review also highlighted limited centralised focus, leading to politically driven reformation (e.g., when the UK adjusted its emergency management systems and frameworks to a centralised governing body - the Civil Contingencies Secretariat (CCS)). However, legislative reforms of emergency management structures and frameworks are not uncommon (see Table 4, for a full breakdown of the evolution of key legislative reforms).
Table 4
Key legislative reforms and reviews of emergency management structure and frameworks
Key Legislation and Reviews |
The Civil Defence Act 1948 |
The Local Government Act 1972, Section 138 |
The Control of Industrial Major Accidents Hazard (CIMAH) Regulations 1984 |
The Civil Protection in Peacetime Act 1986 |
The Local Government and Housing Act 1989 |
The 1989, 1991, and 1997 Peacetime Emergency Planning Reviews |
The Civil Defence (General Local Authority Functions) Regulations 1993 |
The 2001 to 2002 Emergency Review |
The Civil Contingencies Act 2004 |
Politically driven reforms have been found to be associated with centralising a framework for responding and preparing for emergencies at national, regional, and local levels (O’Brien and Read, 2005). Further, such reformation supports resilience through identifying potential challenges, assessing, and managing contingencies, and planning for future risk (Civil Contingencies Secretariat 2009b). However, it is postulated that politically driven fiscal measures (e.g., austerity) undermine centralised frameworks, reducing resilience and increasing cognitive vulnerabilities (e.g.,Wright, 2016). Policy that has attempted to direct and govern under-resourced multi-agency response have been found to not to be fit for purpose (Home Office, 2020; Laufs and Waseem, 2020). For instance, Stevens (2020) highlighted the potential risk of policy-related bias when scientific evidence entered politically driven policy development (e.g., COVID-19). In other words, ministers often trawl through evidence to suit a specific agenda (e.g., see Stevens, 2007), manipulating scientific guidance, rather than following the scientific evidence (Stevens, 2020).
The CCS also hold responsibility for bringing together those lessons learnt from previous emergency response outcomes, at all levels, whereby a multi-agency response was needed. Whilst it was found that these well-established information networks informed cross-governmental capabilities (e.g., Chen et al., 2008; Gil-Garcia, Pardo, and Sayogo, 2016; Waring et al. 2018), they seldom delivered operational emergency response change in practice to prevent repetition of faulty decisions (Pollock, 2013). This realisation appears to be in direct conflict with the core purpose of the CCS, as the dissemination of those lessons learnt only serve to remind multi-agency systems and individual response agencies of their potentially transgressive actions. In other words, lesson learned do not necessarily translate to effective future decision-making, as decisions are often found to be made in context to an incident (Rebera and Rafalowski, 2014), as agencies focus toward least-worst decisions, rather than optimal outcomes (Alison et al., 2018).
The reformation of centralised governing bodies highlights insufficiencies of UK government emergency response management structures (Kapucu, 2009). For example, the Civil Contingencies Act (CCA, 2004) introduced singular legislative and operational frameworks (e.g., JESIP) that provide emergency response decision-makers with civil protection legislation. This enables key decision-makers to define an emergency; identify clear boundaries, responsibilities, and roles for multi-agency systems; allowmsenior decision-makers to explore duties of local and governmental agencies; and offers governmental bodies legislative powers (O’Brien and Read, 2005). However, it was found that the CCA was not always considered by political and policy decision-makers (Home Office, 2020; Stevens, 2020). For instance, in response to the COVID-19 pandemic, it was identified that the diversity in policing roles and responsibilities given to operational policing teams hindered effective decision-making. It was believed that operational priorities and duties often evolved beyond the operational capabilities of policing teams. Policing teams were required to enforce restrictive community measures using militarised approaches, which reduced the overall professional standards of community-focus and collaborative policing (Laufs and Waseem, 2020). Further, government legislation, which purports to minimise operational harms and enhance decision-making (e.g., Coronavirus Act 2020) potentially hinders emergency response decision-making, as there has been minimal public and parliamentary scrutiny (Home Office, 2021). Notwithstanding this, there is also little guidance and support to effectively respond and resolve complex taskings (e.g., Ghaemmaghami et al., 2021; Newiss et al., 20201; Stanier and Nunan, 2020), hindering effective decision-making capabilities, as the operational landscapes remains uncertain. Whilst it was recommended in the CCA that emergency response organisations amend their strategic response styles (e.g., Bonkiewicz and Ruback, 2012), it is unclear to what extent multi-agency systems and emergency response personnel can adapt their style of response. It is well cited that multi-agency systems structure their response style toward intra-inter agency collaboration to enhance decision-making (Alison et al., 2013; Kapucu and Demiroz, 2010; Laufs and Waseem, 2020). However, whilst there is extensive research that demonstrates the need to improve training and interoperability-related skills (Brown et al., 2021; Cohen-Hatton et al., 2015; Hine et al., 2021; Kapacu et al, 2011; Power et al., 2018; Waring et al., 2018, 2020), there is limited research that identifies how under-resourced response teams effectively make collaborative decisions. In particular, how collaborative decisions are made in context to leadership and role-taking, and plan enactment in the preparation and response to critical incident events (see, Steigenberger, 2018, for a detail review of case studies and research agendas for multi-agency disaster response).
Ambiguity in what constitutes a critical incident under the CCA (see, JESIP, 2021, for a detailed definition) is also identified as a factor that may limit decision-making. For instance, key decision-makers in the strategic and tactical coordinating groups are not always able to identify critical incidents. It is thought that communication and coordination between emergency response structures lack prior knowledge of CCA legislation and procedures (e.g., JESIP) that assist responding agencies in identifying a critical incident under the CCA (Power and Alison, 2017; Waring et al., 2020). Worryingly, such challenges in identifying a critical incident in the early stages have been found to impede response decision-making (Waring et al., 2020), as decision-makers are unable to rapidly deploy appropriate multi-agency partners from Category 1 (e.g., Police, Fire, Ambulance, etc.) and Category 2 response agencies (e.g., Health and Safety Executive, Department for Transport, Utility Companies, etc; see, Cabinet Office, 2013 for more details).
3.2.2. Narrative Two: Difficulties of Operating Under Austerity
Over the last decade, fiscal austerity has forced governments to cut response agency budgets, resulting in increased operational pressures, reduced capacity for response, decrease in Category 1 resources, and increased cognitive and macro cognitive stresses (Power and Alison, 2017; Wright, 2018). In other words, austerity has been found to impede effective decision-making. For instance, response commanders are often required to resolve incidents quickly, rather than effectively, to free up resources (Power and Alison, 2017). Moreover, commanders are more likely to triage their decision based on responder wellbeing and capability (e.g., Power and Alison, 2017), as burnout prevalence within the emergency services has become more prevalent with exposure to long-term response efforts (e.g., long-term exposure to response efforts in wildfires; Miller, 2021). It seems apparent then, that decision-making under these pressures potentially derails operational priorities, as organisational priorities shift toward self-preservation and therapeutic jurisprudence (e.g., Milne and Clarke, 2018; Rutkow, Gable, and Links, 2011). Larkin et al (2003) highlights the challenge of resource allocation, as decisions are triaged based on resource allocation. Operating under these contexts has implications for effective response management, as austerity has been widely reported to undermine response resilience, contributing to increased political, organisational, and individual vulnerability for current and future threats (e.g., Wright, 2016). Whilst austerity measures have consequences for emergency response decision-making, it is important to highlight that there is a general lack of research that explores the extent to which austerity has comprised effective response decision-making.
3.2.3. Narrative Three: Uncertainty and Accountability
The National Security Strategy (NSS) sets out how multi-agency systems address and manage diverse challenges to security-related critical incidents (e.g., terrorism; HM Government, 2015). The NSS also guides senior decision-makers assessments of security threats, as the threats faced by the UK increase in ‘scale, diversity and complexity’ (HM Government, 2015, pp.15). However, whilst the aim and purpose of this strategy has provided direct guidance to strategic, and has driven UK security priorities (e.g., enhancing local and national security capabilities), these strategies do not always mitigate against faulty decision-making (Alison et al., 2010). This includes acknowledging factors of accountability, uncertainty, temporality, and cultural and historical concepts (Alison et al. 2010). For instance, the NSS and the national security risk assessment (NSRA) lack principles of accountability – a construct that refers to the expectation that an individual or organisation may be evaluated by a salient audience (e.g., CCS). In part, this may be because operational response agencies and frontline responders favour self-preservation (Alison et al., 2010). Further, anticipating future accountability is also thought to impede situational assessments (Alison et al., 2013), limit pro-social behaviours (Frink and Klimoski, 2004), limit effective plan execution in aid of common strategic and operational goals (Gollwitzer and Moskowitz, 1996), and limit the ability to discern relevant versus irrelevant information (Waring, Alison, Cunningham, and Whitfield, 2013).
High levels of uncertainty – a crucial component in governing strategic decision-making (Van de Heuvel et al., 2012) – has also been found to potentially influence decision-making, as multi-agency systems address higher levels of uncertainty through risk avoidance strategies (Bernstein, 1998; Van de Heuvel et al. 2012). Where future threats are uncertain, diverse in nature, and complex, multi-agency systems deploy risk avoidance strategies to avoid actions that might lead to potentially negative consequences (e.g., decision inertia; Alison et al., 2015; Shortland et al., 2020). For example, when perceptions of uncertainty are high, response personnel are more likely to narrow their focus toward negative consequences, and thus become more risk avoidant to preserve self-serving issues, in lieu of community protection; defer choice; or favour an act of omissions over commission (Van de Heuvel et al. 2012a, b). In summary, the results suggest that the characteristics of critical incident events often exacerbate perceptions of uncertainty. As emergency response organisations adopt a more risk adverse strategy to help promote self-serving bias, and reduce the negative consequences associated to accountability (i.e., future inquiries into actioned decisions) current decision-making models are not considered to be robust. In other words, the prescriptive nature of current decision-models appears to promote triaging systems that focus on community confidence and not community protection.
3.2.4. Narrative Four: Inter-Intra Governmental and Organisational Ethics
Ethics are crucial for social, political, and economic practices, particularly when framed in context of emergency response (e.g., French and Raymond, 2009; Kalajtzidis, 2016; Leider et al., 2017; Rutkow et al., 2011). When multi-agency systems become stretched beyond their operational capabilities, allocation of resources, civil protection response strategies, and duties to vulnerable persons become an ethically and legally complex problem (e.g., Gostin et al., 2020; Leider et al., 2017). However, in practice decisions are often determined by the best possible, or least-worst outcome (Alison et al., 2018), utilising effective information sharing networks (Waring et al., 2018) and triaging processes (e.g., WISCI; Milne and Clarke, 2018) rather than focusing specifically on ethical and legal guidelines. Ethical guidance is also an increasingly common component of emergency management frameworks (e.g., see Leider, DeBruin, Reynolds, Koch and Seaberg, 2017), as triaging systems rely on technological solution to inform operational, tactical, and strategic decision-making (e.g., the ethical use of Big Data through social media and mobile data mining; Crawford and Finn, 2016; French, 2009). However, ethical principles are considered too ambiguous, and present potential operational and strategic risk (Crawford and Finn, 2016) as frontline responders and response agencies struggle to fully comprehend and apply ethical practice. That is, responders focus toward triaging why ethics was needed, rather than understanding the practical application and implementation of ethical principles (Leide et al., 2017).
Integrating ethical principles also required a collaborative and coordination multisystem command structure (Jennings, Arras, Barrett, and Ellis, 2016). Further, strong decision-making frameworks are necessary to instil confidence and promote inter-intra-agency communication and coordination through common goals for preparedness planning, emergency response, and post-event recovery (Waring et al., 2018). However, it is suggested that multi-agency systems seldom achieve synergetic coordinated collaboration (Alison et al., 2015; Chen et al., 2008; House et al. 2014; Waring et al., 2018) as operational end goals consistently differ between response organisations, and organisational structures.
3.2.5. Narrative Five: Failures in Collaborative Information Networks
Public inquiry into critical incident events have repeatedly highlighted a lack of interoperability of a key factor in response outcomes (Hine and Bragias, 2021; Waring et al., 2018). However, the extent to which multi-agency systems facilitate effective interoperability remains an area of key interest. Whilst it is widely agreed that interoperability enhances critical incident response (e.g., Chen et al., 2008; Hine and Bragias, 2021; House, Power, and Alison, 2012; Waring et al., 2018) it is largely dependent on an organisation’s ability to develop a shared characterisation of a critical incident event (e.g., do responding agencies have a shared operational picture; Salas and Cannon-Bowers, 2001). For example, whether responding agencies have sufficient capability to share information, both reliability and quickly (e.g., information networking technologies; Allen, Karanasios, and Norman., 2014; Waring et al., 2018). Further, whether agencies have the desire to collaborate between other responding agencies (e.g., Alison et al., 2015; Chen et al., 2008; Kapucu and Garayev, 2011).
Kapucu and Garayev (2011) highlighted that whilst decision-making performance was relatively satisfactory in response to Hurricanes Katrina and Rita in 2005, there were failures in interoperability networks, as communication network capabilities were not fit for purpose; information was not always disseminated due to a lack of trust between agencies; and inter-agency values differed. In addition, the phase one report into the 2019 Grenfell Disaster highlighted several challenges associated with interoperability, as response agencies were not able to fully develop a shared situational awareness. In part, this was due to communication technologies not being fit for purpose, responding agencies were not able to effectively utilise specialist communication technologies, and internal communication policies were not fully understood in context of the on-going events.
In line with current recommendations, agencies can facilitate interoperable decision-making through system monitoring (see, Healey et al., 2009). However, despite facilitating technologies purportedly aiming to enhance decision-making capabilities (House et al., 2012), developing interoperability remains a challenge, as responding agencies seldom have the operational experience of working collaboratively in extreme contexts (Brown et al., 2021; Waring et al., 2018). In addition, standard operating procedures, potentially hinder effective decision-making. This is because the response to a critical incident event often results in organisations adopting a consequentialist approach - where agencies undertake rapidly evaluating courses of action (i.e., on the spot decision-making; Rebera and Rafalowski, 2014) without consideration of the wider operational objectives and operating procedures (see, Shortland et al., 2020). Standard operating procedures were also found to assume an operational hierarchy, whereby commanding personnel (e.g., Gold Command) were assumed to know best how objectives should be achieved. As such, there is a general expectation amongst responding agencies that command strategies and tactics should be followed (e.g., Critical Incident Management Guidance; Home Office, 2021). Whilst hierarchal structures within response agencies reflect a centralised strategic and tactical decision-point, where command personnel operate a ‘top-down’ approach to decision-making, hierarchical structures in emergency response can be counterproductive. In practice, this is due to a lack of consistency in communication channels across teams (e.g., SSG and TSG). This impedes fluid communication (Brown et al., 2021), as decision priorities vary across responding agencies, limiting shared operational understanding and plan formulation (House et al., 2012; Waring et al., 2018) as decision-makers fear breaching operational norms (e.g., Robert and Lajtha, 2002).
A lack of coordination between response agencies (Smith and Dowell, 2000), multi-agency authorities and conflicts of interest (Chen, Sharman, Rao and Upadhyaya, 2008) have been widely as factors that limit effective interoperability. In part, this is due to inter-and-intra agency information communication networks dissemination incomplete information that aids decision-making (e.g., Banipal, 2006). For example, Salmon et al. (2011) found that multi-agency communication networks disseminate incomplete information between agencies, resulting in a misunderstanding of the information communicated, limiting interpretability of the information, and consequently resulting in poor information management. Further, a lack of understanding and trust between agencies limit awareness of operational capabilities. This often results in agencies developing independent operational priorities and failing to formulate common operational end-goals (Salmon et al., 2011). Whilst a multi-agency response is required to mitigate least-worst outcomes, the lack of trust, understanding, and failure in communication networks and strategies only serve to exacerbate response challenges.
3.2.6. Narrative Six: Future Research-Focused Horizon Scanning
Experimentally examining all the factors, at a holistic level (e.g., Macro-, Meso, and Micro- level), is crucial to advancing decision-making (Brown et al., 2020). Yet, research remains in its infancy, as decision-making challenges have not yet been captured through empirical formats (Brown et al., 2020; Power, 2017). For example, using technological innovation to empirically examine critical incident decision-making is still a developing area of scientific understanding (Alison et al., 2013; Brown et al., 2020; Crego and Harris, 2017; Eyre et al., 2012). Brown et al. (2020), whilst supportive of innovation and technology, suggests that high-fidelity simulation research lacks context validity, limiting the translation of theory to practice. Further, Bayesian statistical approaches (i.e., an approach to assess epistemological uncertainty in fields of study where there is a lack of knowledge about a fundamental phenomenon) seldom feature in research. For example, Brown (2020) articulated the necessity to utilise more robust and powerful statistical approaches to incorporate prior expectation, without the need to assess critical incident data independently for each study (see, Zyphur and Oswald, 2015 for a more complete understanding of Bayesian approaches). Whilst this is not a suggestion that research cannot provide crucial insights, it does suggest that the validity and impact of current research is limited.