Emergency responders’ experiences of multi-agency working during the COVID-19 response in the UK: Social identity as part of the problem and part of the solution

Recent research has shown that multi-agency emergency response is beset by a range of problems, calling for a greater understanding of the way in which these teams work together to improve future multi-agency working. Social psychological research shows that a shared identity within a group can improve the way in which that group works together and can facilitate effective outcomes. Thus, seventeen semi-structured interviews were conducted with strategic and tactical responders during the COVID19 pandemic to explore whether there was any evidence that a shared identity was part of the solution to challenges faced, and if so, how and when shared identity arose. Findings suggest that two forms of group relations were particularly relevant: horizontal intergroup relations – the relationships among responders at the local level; and vertical intergroup relations – the relationship between responders at the local level and national level. Three key factors appeared to contribute to a shared identity amongst responders. First, pre-existing relationships with other responders facilitated the ease with which responders were able to work together initially. Second, a sense of ‘common fate’ helped bring responders together, and finally, Chairs of groups were able to strategically reinforce a sense of shared identity.

involved in the COVID-19 response at a strategic or tactical level. Specifically, we sought to address a gap in current research by exploring the role of shared identity in multiagency working in this context: was this relevant, how did it arise, how did it function? However, before describing this research, we first provide an overview of the emergency response context in the UK, a summary of research into interoperability, and introduce the Social Identity Approach.

The emergency response context
Within the UK, the multi-agency response to major incidents is managed through a three-tiered structure set out in the Civil Contingencies Act (CCA) 2004 (see Table 2). This structure brings together partners from Category 1 (e.g., the emergency services, local authorities, the NHS) and Category 2 responding agencies (e.g. the Highway Agency and public utility companies; Cabinet Office, 2013). Multi-agency coordinating groups are often established at the strategic (Strategic Coordinating Group; SCG) and tactical (Tactical Coordinating Group, TCG) levels in a multiagency response (Cabinet Office, 2013). Here representatives from relevant agencies come together to provide a vital coordination role in an incident response (CCA, 2004).
An effective response at the operational level relies on a coordinated response at a strategic and tactical level -that is, it requires good multi-agency working (Cabinet Office, 2013). Accordingly, it is important to ensure the factors impacting effective multi-agency working at a strategic and tactical level are properly understood.

Understanding interoperability
Previous research looking at the strategic and tactical levels of emergency response has been focussed on factors such as decision-making (e.g. Power & Alison, 2017;Waring et al., 2020;Wilkinson et al., 2019; and information sharing (e.g. Waring et al., 2018). Amongst other things, one thing that this research shows is that the composition of the group can be an important determinant of effective group decision-making (see Bang & Frith, 2017, for a review). For example, in SCG groups, it has been found that decision-making activities were dependent on group characteristics -groups that were more likely to explore, rather than exploit information were more likely to make critical decisions (Wilkinson et al., 2021). This supports the idea that group composition is an important element of emergency response and reinforces the need to further understand group working in these contexts.
To understand the experiences of local and national strategic decision makers in the COVID-19 response, three reviews were conducted between April and October 2020, and multi-agency working was identified as a prevalent theme (Hill et al., 2020a(Hill et al., , 2020b(Hill et al., , 2020c; see Hill et al., 2021 for a summary). Factors that were identified as facilitating effective multi-agency working included pre-existing relationships between responders and effective leadership. On the other hand, factors that challenged multi-agency working included responders having no prior knowledge of structures and procedures, and a lack of information sharing from a national to local level.
Whilst previous research has sought to better understand interoperability in multi-agency response (e.g. Waring et al., 2020;Wilkinson et al., 2019;, further research is required to better understand psychological group processes in these unique contexts. More specifically, in order to make interoperable working as effective as possible, we first need to understand how individuals from separate organisations come to work together interdependently as a group. To our knowledge, research has not yet applied social identity principles which underpin group behaviour and explore how these might impact interoperability. Below we introduce the Social Identity Approach and discuss how shared identity and organizational behaviour might help understand multi-agency working better.

Shared identity and organizational behaviour
The Social Identity Approach is a social psychological framework that seeks to understand the distinct contribution that group life makes to people's psychology and behaviour. The approach is comprised of two inter-related theories -social identity theory (Tajfel & Turner, 1979) and self-categorization theory (Turner et al., 1987(Turner et al., , 1994)that are built upon a foundational insight that as well as defining themselves, and behaving, in terms of their personal identity as individuals (Turner, 1982), people can, and often do, also define themselves, and act, in terms of their social identity as members of social groups (Tajfel & Turner, 1979). So, whereas personal identity defines a sense of 'I' and 'me' that describes a person in contrast to others, social identity defines the self in terms of 'we' and 'us' in ways that psychologically connect people to other members of their ingroup.
In a range of social and organizational contexts, this sense of social identity is observed to be the primary driver of people's behaviour primarily because, as Turner (1982, p.21) argues, it is what "makes group behaviour possible" (cf. Haslam et al., 2003). In particular, a shared identity within a group is a basis for coordination and cooperation between group members because it increases their psychological sense of inter-connection and common purpose . At the same time, it can provide group members with shared definitions of situations, as well as common norms for behaving in those situations (Reicher et al., 2010). Consequently, these shared definitions and common norms can increase the ability of those who perceive themselves to share social identity (i.e., who are bound by a common sense of 'us') to work effectively together (Drury et al., 2009;Haslam et al., 2009) while also fostering trust and respect among group members (Haslam et al., 2012;Turner et al., 1987).
Demonstrating these positive effects, Haslam et al. (2009) showed that individuals who had high group identification were more willing to display organization citizenship, Qeios, CC-BY 4.0 · Article, October 12, 2021 Qeios ID: MHJTNY · https://doi.org/10.32388/MHJTNY 2/11 had greater work satisfaction, and had more pride in their work than those with lower levels of identification. More generally, a large body of research demonstrates that when group members perceive themselves to share social identity this increases their motivation to contribute to the groups' success as well as their ability to do so (as reviewed by Ellemers et al., 2004;Haslam, 2004).
At the same time though, in organizational contexts, it is apparent people have multiple social identities which can each become salient in different contexts (e.g., us women, us Londoners, us paramedics; Millward & Haslam, 2013;Turner et al., 1987). According to self-categorization theory, these can also be defined at multiple levels of abstraction (Turner, 1985). For example, a paramedic, Anne, can define herself, as a member of a particular team, as a member of a particular profession, or as an emergency worker. It follows too that this is likely to have a significant bearing on her behaviour. For example, when (and to the extent that) she defines herself as a member of a particular team Anne should be motivated to advance the interests of that team but when (and to the extent that) she defines herself as an emergency service worker Anne should be motivated to advance the interests of emergency service workers.
This analysis is clearly relevant to the analysis of interoperability in the emergency services because the organizational entities here (e.g. SCG and TCG's) typically involve individuals from usually independent organizations who are required to work collectively and collaboratively with each other. In particular, responders are required to jointly provide the overall multi-agency management of the incident, yet they still have agency-specific responsibilities (Cabinet Office, 2013) which can conflict with the overall superordinate goal of the response (Mathieu et al., 2001). As a result, there is a need to understand what factors impact interoperability in a multi-agency response to improve the effectiveness of these unique groups in future incidents. Hitherto, though, despite its obvious relevance, social identity theorising has not been used to gain traction on this issue. This is therefore the goal of the present research.

The present study
Existing research highlights the recurring problems that arise in multi-agency response and calls for a greater understanding of the way in which these teams work together, in order to improve multi-agency working in the future. Potentially exacerbating existing challenges, the COVID-19 pandemic has presented a unique set of challenges to emergency responders, in terms of the scale, longevity and complexity of the response required. For example, most emergencies in the UK are handled at the local level with no direct involvement from a national level (Cabinet Office, 2013). However, in the COVID-19 response, central Government played a leading role in the response, adding an additional consideration for responders to manage.
With this in mind, we conducted regular, semi-structured interviews with responders at the strategic and tactical levels of response from across the UK to identify good working practices that engendered effective multi-agency working during the COVID-19 response, and to understand the possible role of shared identity in facilitating effective working in these contexts.
More specifically, the aims of this research were to identify: 1. The key challenges that emergency responders faced in developing and delivering a multi-agency response to COVID-19 at the strategic and tactical level.
2. Whether there was any evidence that emergency responders' shared identities were bound up with these challenges and/or were part of the solution to them.
3. If so, how and when a shared identity arose in multi-agency groups.

Participants
Fifty-two semi-structured interviews were conducted with 17 responders from the Police (N=8), Fire and Rescue (N=7) and Ambulance Service (N=2) who were involved in the COVID-19 response at a strategic and/or tactical level -all responders were involved in the SCG and/or TCG within their local area. One responder was from Scotland, one was from Northern Ireland, three were from Wales, and 12 were from England (see Table 3 for a full list of participant details). Repeated Interviews took place between 13 th April 2020 and 27 th July 2020. Due to availability challenges because of commitments in the ongoing COVID-19 response, responders took part in an unequal number of interviews, ranging between 1 and 12 each (M = 4, SD = 3.15). Note: Participant's age was not recorded.

Procedure
Interviews took place either over the telephone or via the online platform, Microsoft Teams, and were recorded with a dictaphone. Responders were provided with an information sheet electronically before their first interview and a verbal consent protocol was read out to responders before their first interview and they were asked to verbally consent to take part.
Subsequent interviews were carried out between 6 and 56 days after the previous interview (M = 17 days, SD = 13.2). The first interview for each responder lasted on average 41 minutes and 28 seconds (max = 56 minutes 59 seconds, min = 26 minutes 26 seconds). Subsequent interviews lasted on average 23 minutes 37 seconds (max = 42 minutes 8 seconds, min = 10 minutes 42 seconds).

Interview schedule
The interview schedule was developed following discussions between members of the research team. For the first interview, questions focussed around roles and responsibilities (e.g. "What is your current role within the COVID-19 response?"), multi-agency working (e.g. "Can you tell me about the range of partners that you are involved with in this response?"); strengths and weaknesses of the response (e.g. "Can you tell me about any challenges you have faced?"); adaptation (e.g. "are there any specific areas of improvement that you have recognised in this response?"); and training and guidance ("Is there any specific training or guidance you are following in your response?"). Subsequent interviews focussed on any changes or developments in the response since the previous interview. The full interview schedules can be found in the Supplementary Materials 1 & 2.

Context
A summary of key dates, events and response activities provided by participants during the interviews is presented in Table 4. Local authorities given power to enforce local lockdown * The summary of response activities was provided by participants during the interviews within the time period specified.

Data analysis
Familiarisation with the interviews took place by the lead author through listening to the recordings of all interviews. Sections of interviews relevant to the research question Qeios, CC-BY 4.0 · Article, October 12, 2021 Qeios ID: MHJTNY · https://doi.org/10.32388/MHJTNY 4/11 were then transcribed. For example, sections of the interview where the responders discussed more general response activities that were not specific to the multi-agency response to COVID-19 were not included. Transcribed interviews were analysed using thematic analysis -a method for identifying, analysing and reporting patterns (themes) in data (Braun & Clarke, 2006). Analysis involved six key stages: (i) familiarization with data; (ii) generating initial codes; (iii) searching for themes; (iv) reviewing themes; (v) defining and naming themes; and (vi) producing the report . The research team met on a fortnightly basis throughout the analytical process to discuss the analysis.

RESULTS
The results are presented under two key topic areas: horizontal intergroup relations and vertical intergroup relations (see Table 5). Table 5. Overview of topic areas and themes.

Topic Area Theme Definition
Horizontal intergroup relations: the relationship between different organizations within a local area Pre-existing relationships The impact of relationships in place between responders before the COVID-19 pandemic.

Understanding roles of partners
Responders having a clear understanding of what the roles and goals of partners from different agencies are in the response to COVID-19.

Shared understanding of the response
Responders from different agencies all having the same awareness and understanding of the COVID-19 response.
Vertical intergroup relations: the relationship between responders at a local level and national government agencies Communication Challenges in communication and information sharing between national government organizations and responders at the local level.
Cross-area lesson sharing Local level responders sharing information with regional, national and international partners.
The themes are presented alongside representative extracts from the interviews. To ensure anonymity, each responder was given a unique participant number (1-17; see Table 3) which is presented alongside extracts. Non-typical quotes are highlighted in the results.

Horizontal intergroup relations
3.1.1 Pre-existing relationships. Nearly all responders said a key strength of the response was how well everyone came together as a group to put a response in place in such uncertain terms, and credited pre-existing relationships between partners from different organizations in facilitating this: In a long-playing incident like this you can utilise those relationships you have already got and work quite effectively as opposed to coming together at the point the incident started and develop the relationships from there. So, having those pre-existing relationships is what has made this response so effective (P7).
One responder said that they already trusted each other and saw each other as a "unit" (P5). Furthermore, one responder from London described their partnership as "welloiled" due to the regular recent incidents they had attended, such as the London Bridge and Finsbury Park attacks, and Grenfell Tower fire (P3). Another responder from London expanded on this saying they have "worked as one" for so long (P2).
Yet, it is not just about the presence of a prior relationship that helped the group come together, some responders suggested the quality of that prior relationship was also important in group formation. For example, knowing people on first name bases and having their phone numbers already saved meant that when challenges arise, they can be quickly resolved by a phone call to the right person. One responder said that friendly relationships can help "lighten the mood" when the pressure on them is high (P12).
In addition, another responder said friendly relationships are beneficial not just for the current response, but also for future group working: The Some responders commented on the unique nature of COVID-19 and the requirement for people to work remotely where possible. This created challenges in getting to know others when pre-existing relationships were not formed because they were not able to have the face-to-face interaction they would usually have. Responders from the South East discussed setting up a Strategic Coordination Centre (SCC) where they were able to co-locate and have meetings in person. One responder from this area said that being able to attend the SCC (as opposed to attending virtually) made it much easier to know who people were and who they needed to speak to. Another responder commented that virtual SCG meetings made it much harder to establish a relationship when they did not previously know an individual: What is harder with COVID-19 than in a usual incident is that if you don't know another responder at all and if they are now ringing into these meetings rather than attending them, due to social distancing, it is much harder to get that relationship (P2).

Understanding the roles of partners.
Several responders said that the biggest difference of the COVID-19 response compared to other incident responses is that it is a health-led initiative and they have spent a lot of time trying to understand the nature of the health service. Exacerbating this challenge, one responder (P4) said they had a health Chair of the SCG who had not previously chaired an SCG before and it took them a couple of weeks to fully understand the purpose of the group.
Yet, it was not just the health service who presented challenges; one responder (P12) said that the COVID-19 response involved a number of guest agencies who would not normally be involved in a response, such as the prison and probation service. This created challenges with new partners being initially hesitant to share problems they were facing, leading to delays in resolving them. Some responders said that this challenge eased as time went by: Further, one responder (P4) highlighted the important role of the Chair of the group in facilitating an understanding of roles amongst partners. The Chair going over the roles and responsibilities of each partner at the beginning of the response made logistics easier. Another responder said that once they had overcome initial differences between partners, they were able to collectively deal with any challenges, rather than working independently of each other: 3.1.3 Shared understanding of the response. Initially, some responders talked about how COVID-19 brought them together and facilitated their group working. One responder said the external threat of COVID-19 allowed partners to work collectively in the response because they had a "unity", "one aim" and were all "working to the same goal" (P5). Other responders said they were all working together for "a common purpose" (P6) and convening around "a common enemy" (P11).
However, in later interviews some responders said that different organizations developed a different understanding of where they were at in the response, and that this reduced the shared sense of common fate. One responder who initially talked about the unity of the group said that later in the response the group cohesiveness that was originally formed started to weaken because there was no longer a clear common purpose for why they were convening: It takes an external threat for everyone to come together to work for the greater good and taking one for the team [...] but as soon as that external threat slightly dissipates, even if it is just that we are over the initial peak [...]

everyone starts petty squabbling and it just unravels from the top (P5).
To try to overcome this challenge and maintain group cohesiveness, this responder (P5) said they laid out eight strategic goals for the SCG at the beginning of each meeting.
Further, another responder said that they began each of their SCG meetings with an overview of the common picture of the incident, so that each partner knew exactly what was happening, what the challenges were, and what actions needed to be taken, suggesting that leadership is important in maintaining group cohesiveness:

At the start of the meeting you start with this is where we are and these are the previously identified risks […] everyone needs to leave the room with a clear line of sight of everyone else's position […] no matter where they are from, they have a clear line of sight of what is happening [...] where the pinch points are and what mitigating action needs to be taken (P12).
In response to the changing situation, one responder said that their SCG introduced a new phase called 'stabilisation' which occurred after the initial response phase, but before the recovery phase. Within this phase, the SCG members were not meeting regularly as they had done in the initial response phase, but partners were still working together and ready to meet again if or when it was necessary. This was so that partners were aware they still had access to the resources and support the SCG could provide:

The other alternative is to close [the SCG] down and the message that sends is message complete [their response to COVID-19 is over] […] that sends all kinds of dangerous signals so the other alternative is to leave it running in the background so it is technically in existence but there is nothing happening in it […] the SCG is a leadership group of senior people across the partnership saying this is still very important to us. If we all walk away […] what we're saying is it's not very important
anymore (P11).

Communication.
Initially, a key challenge consistently highlighted by nearly all responders centred on communication between national and local level. In particular, responders reported that challenges were created by receiving key announcements from central government at the same time that the public received them. One responder said that not having a clear idea of changes in policies made it very difficult to prepare and to provide a timely response:

With other incidents it is much more clear about what is going to happen next, the big thing with this one is [...] we are finding out at the same time as everyone else, then having to respond to this as a strategic and tactical body, which is unusual in responses to emergencies because usually we know what is coming next and can start to plan and map out what it looks like. Finding out at the same time as everyone else makes it a little more interesting (P4).
This communication challenge seemed to be exacerbated in the devolved nations of Northern Ireland and Wales. A responder from Northern Ireland said it was challenging to understand whether to follow a UK approach or a devolved approach and often there was mixed messaging between the two administrations which created confusion in the response (P8). A responder from Wales said additional measures had to be put in place before any changes in rules surrounding COVID-19 could be enforced. This led to further delays in responders in Wales receiving new information. One responder said this created a "false start" in their response because when the new rules were published by the media it was not made clear that this was the law for England not Wales: Yet, it was not just delayed information sharing between a local and national level which made the response challenging, some responders also said conflicting information created challenges. One responder said this caused "confidence and reputational issues" with responders towards the government:

…] Consideration as to what comes out and even if it is delayed by a day it can avoid a lot of the contradiction and confusion (P6).
Further exacerbating this disconnect between the national and local level was the inconsistent presence in some areas of a representative from government (Government Liaison Officer [GLO]) in the SCG meetings to act as a link between the local responders and central government. One responder said they were given a different GLO each week which was difficult as they were unable to form any consistent relationships (P4). They said this presented challenges as there was no dependable channel of communication into central government, meaning the national and local level groups were working independently of each other in the initial response to COVID-19. For example, one responder described how national agencies were planning on building a testing site at the same location they had locally planned on building a temporary mortuary and this was only uncovered by a chance encounter between a local police officer and members from the national government agency at the site (P16).
However, the inconsistent presence of a GLO was not reported by all responders, with responders from London saying they had a GLO that attended meetings from the beginning. Yet, there was still a clear disconnect between national and local level in this area: We are not getting information down from government as the government are keeping everything very close to their chest. We have got a number of GLOs in the SCGs that have been involved right from the start, they help with the information flow downwards and they can help facilitate a workaround to ensure that supply chains and what needs to be done can be done to ensure that communication channels can open up again (P3).

Cross-area lesson sharing.
In later interviews, whilst communication challenges between national and local levels were still prevalent, some responders reported being less reliant on information coming down from a national level to guide their response, instead taking a more local-level approach. Responders in the South East credited the development of connections with regional partners which facilitated a common regional approach. They said this allowed responders from different areas to compare and discuss what actions were being taken in each region, share relevant information, provide updates on what was working well and provide a vital coordination role between regions: We are now solving the same problem but in a different way […] we have focussed on regional colleagues and partners, […] we have agreed a common approach across the South East. We have a workshop on Thursday to compare approaches to the modelling cells, this has enabled us to start comparing (P11).
Further, one responder (P2) in London talked about lesson sharing nationally to enable other areas to learn lessons from London, who in the early months of the pandemic seemed to be worse affected than other areas of the country. In addition, one responder (P17) in Scotland discussed sharing lessons internationally, which came about due to the strong international links they have.
As well as providing practical support, this cross-area connection also provided emotional support through "providing the opportunity to vent and also assurance" (P16). This was also echoed in Scotland where one responder said they recognised that several lives had been lost, and that this was causing an emotional strain for responders: When London were getting hit about three to four weeks ahead of us, I had a number of one-to-one strategic meeting calls with other strategic commanders in London about how does this feel, not the numbers or sterile meeting room environment but the 'phone a friend' item, how's it going? What does it feel like? (P17).
However, pre-existing relationships with responders from other areas was credited as a key driver behind this cross-area lesson sharing. The responder from Scotland said they needed to depend on their network and relationships that they had built up with others before COVID-19 (P17). Furthermore, the responder from the South East said that there was nothing in guidance about talking to people from other regions:

DISCUSSION
Based on interviews with Police, FRS and Ambulance Service responders from across the UK who were involved in the COVID-19 response at a strategic or tactical level, there were two key areas where inter-group relations were discussed: (i) horizontal -the relationship between different organizations within a local area, and (ii) vertical -the relationship between responders a local level and national level. Horizontal intergroup relations were strengthened by pre-existing relationships, understanding the roles of partners, and having a shared understanding of the response. Vertical intergroup relations were challenged by poor communication between national and local level, with this challenge eased in some areas by cross-area lesson sharing.
Key challenges responders faced in the response are discussed below alongside potential solutions and how shared identity arose in the multi-agency groups. Discussion is separated by three potential solutions: relationships, common fate, and leadership. Because of these recent shared experiences, and relationships already being established, a sense of shared identity was likely to already embedded in these responders.

Key challenges, potential solutions
Qeios, CC-BY 4.0 · Article, October 12, 2021 Qeios ID: MHJTNY · https://doi.org/10.32388/MHJTNY 7/11 It is well documented within the social identity literature that when individuals identify with members of their group this can help the group work more effectively together (Drury et al., 2009;Haslam et al., 2009) and can foster trust in other group members (Turner et al., 1987). In other words, when these relationships are already formed and individuals have had recent experience of a shared identity with each other in the past, it makes it easier for them to act as a group in the present. Based on this, responders in the COVID-19 response were likely able to benefit from the positive effects of their existing shared identity immediately.
Despite this, a key challenge several responders highlighted was that delayed or conflicting communication from a national to local level created difficulties for responders in preparing for and providing a timely response. This seemed to be exacerbated in the devolved nations of Northern Ireland and Wales where at times it was unclear whether to follow a national approach, or a devolved approach. Yet, to manage this challenge some responders discussed sharing lessons regionally (South East), nationally (London) and between-nations (Scotland). Thus, responders utilised relationships with partners outside of their local area as a potential solution to a challenging relationship with national partners. In addition, some responders commented that this cross-area lesson sharing was facilitated by pre-existing relationships with responders from different areas, suggesting that these relationships can facilitate the response both early on, and throughout the response.

Common fate.
Several responders used collective terminology when describing the response, particularly in the initial interviews (e.g. "common enemy", "unity", "one aim", "common purpose"). As such, it seems that the nature of the response also facilitated the group coming together, in other words the responders experienced a sense of common fate -"a coincidence of outcomes among two or more persons that arises because they have been subjected to the same external forces or decision rules" (Brewer, 2000, p. 118). Common fate between individuals can facilitate a shared identity between members (Drury, 2018). Subsequently, this shared identity can encourage helpful and empathetic behaviour between group members (e.g. Levine et al., 2005), enhance people's trust with other group members (Cruwys et al., 2020) and increase their willingness to co-operate in working towards group goals (Haslam, 2004). Taken in the context of the current research, a shared sense of purpose in the response to COVID-19, as evidenced by the collective terminology used, is likely to have facilitated responders coming together psychologically and increased their ability to work collaboratively on the response.

4.1.3
Leadership. An early challenge discussed by responders was the wide range of partners involved in the response who would not typically be involved in an incident response. As such, pre-existing relationships were not present with these partners. Furthermore, this was exacerbated by the virtual nature of the meetings. To overcome this challenge, in some areas the Chair of the meetings went over roles and responsibilities at the beginning of the response, or when new partners joined, facilitating their ability to work interdependently with, as opposed to independently of each other.
Furthermore, when the initial wave of COVID-19 came to an end, the initial sense of shared purpose seemed to reduce in some areas. A potential solution for this was for the Chair to go over where they were at with the response and outline any outstanding issues to help maintain a common operating picture amongst responders. Thus, maintaining a shared awareness of the situation in this way helped facilitate group cohesiveness and was achieved strategically by the Chair of the group through ensuring common goals were communicated to all. Taken together with the role of the Chair group formation, leadership played an important role during the multi-agency groups in this context.
According to Zehnder et al. (2017), effective leadership can help organizations foster a sense of shared identity among members -i.e. helping group members see themselves as 'we' as opposed to 'I' (Steffens et al., 2014). Effective leaders can facilitate collaboration between group members (Ellemers et al., 2004), making the group more likely to succeed in their goals (Carton et al., 2014). In a recent study, Fladerer et al. (2020) showed that leaders were able to reinforce a sense of shared identity amongst group members by using collective language such as "we" as opposed to "I". In turn, this reinforced sense of shared identity within the group was subsequently associated with improved organizational performance, emphasising the importance of effective leadership.
Further, in the context of emergency response, recent research has shown that individual characteristics of the Chair of SCG groups can influence decision-making processes within the group (e.g. Waring et al., 2020;Wilkinson et al., 2019). The evidence presented in the current paper supports this and suggests that the Chairs (i.e. leaders) of the multi-agency coordination groups were able to strategically embed a shared identity among the group through specific actions throughout the duration of the response. In addition, the research in the current paper showed that leaders played a particularly important role when pre-existing relationships were not already present, or when the sense of common fate began reducing, or in other words, when there did not seem to be a strong sense of shared identity between responders. This echoes further recent research by Fladerer and colleagues who found that identity leadership was particularly relevant in situations where co-workers group identification was low (Fladerer et al., 2021). Thus, this highlights the importance of effective leadership, particularly in new or less well-established groups.

Strengths and Limitations
One limitation of the research presented is that only responders from the blue light services were included but the COVID-19 response involved responders from several different organizations. Because of this, it is difficult to discern whether the challenges discussed by responders were common across responders from other organizations. In addition, the varying availability in responders may have biased the results to those who took part in the most interviews. However, several findings from the current research are echoed in the research by Hill et al. (2021). For example, collaborative working was facilitated by pre-existing relationships between partners, yet hindered by partners who had no prior knowledge of the structures or procedures of the SCG and TCG groups. This suggests that the findings of the current research are both common experiences and generalisable to the wider response partners. Additionally, Radburn et al. (2021) have conducted an in-depth case-study from one local area which will provide valuable insights into the wider local-level response to COVID-19.
The research discussed in this paper is focussed on UK emergency response, therefore the applicability to other countries may be limited. Yet, the principles of the Social Identity Approach have been used to explain group behaviour in various different contexts, including social solidarity (e.g. Reicher & Haslam, 2009), social influence (e.g. Postmes et al., 2005), mass emergencies (e.g. Carter et al., 2015) and health (e.g. Jetten et al., 2011). Therefore, it is likely the results of this research would be applicable to other countries. However, because different countries have different emergency response structures, further research would be beneficial to confirm this.
Qeios, CC-BY 4.0 · Article, October 12, 2021 Qeios ID: MHJTNY · https://doi.org/10.32388/MHJTNY 8/11 Of course, it is possible there are factors other than social identity principles that may have impacted multi-agency working in the COVID-19 response. For example, responders from the South East said they introduced a period of stabilisation in between the usual 'Response' and 'Recovery' phase -an interim control stage to mitigate the risk of secondary impacts occurring, as well as allowing multi-agency coordination groups to retain their overall focus on reducing the risk of the current threat (Deeming & Burgess, 2017;cf. Deeming, 2020). Recent research looking at multi-agency response to a simulated terrorist incident also found benefits of a three-phased approach (Brown et al., 2021). Brown and colleagues suggest that an additional phase in between response and recovery can increase opportunities for collaborative working across agencies and reduce demands on a single team. As such, it should not be ignored that factors other than shared identity can also facilitate effective multi-agency response.
Finally, to our knowledge, this is the first research that applies social identity principles to better understand multi-agency working in emergency response. This research clearly demonstrates that group relations played an important role in the multi-agency response to COVID-19. However, future research would benefit from further investigating this to allow for concrete practical recommendations to be drawn.

Conclusion
The evidence from interviews with strategic and tactical responders involved in the COVID-19 response show that a shared identity was important in helping to provide potential solutions to key challenges in the response. A shared identity was embedded within the multi-agency groups from prior relationships which facilitated the ease at which responders were able to come together initially and work together in the response. Further, a shared sense of common fate against COVID-19 also facilitated group working during the initial months of the pandemic. However, when pre-existing relationships were not present, or when the initial threat of COVID-19 started reducing, a shared identity was able to be reinforced strategically by the Chairs of the groups, through spending time going over roles and responsibilities of partners, or by going over the shared goals of the response.

Practitioner points
• Relationships between responders from different organizations should be nurtured to ensure that a shared identity is maintained between responders to facilitate the ease at which they are able to come together for future incident responses.
• When responders share difficult or challenging experiences with each other, this can help them feel connected to each other, regardless of their organization, and responders should be encouraged to use collective terminology, such as 'we' and 'us' to facilitate this.
• Leadership is important in facilitating a shared identity and Chairs of the multi-agency groups can help strategically embed a shared identity if relationships are not already present or if there is not a strong sense of common fate between responders through specific actions such making roles and responsibilities of partners clear and reiterating shared goals.