Balancing in the pandemic: how social workers respond to new risks when supporting clients who experience domestic violence

ABSTRACT Since the winter of 2020, COVID-19 has significantly changed the lives of many people. The aim of the present study is to explore how social workers specialised in the field of domestic violence (DV) balance the expectation and need to protect themselves from infection with the expectation and need to protect their clients and how this balancing affects their discretion. The method was qualitative interviews with social workers, team leaders, and managers in two Swedish municipalities. The results offer insights in how professionals experience and respond to changing conditions of social work and the potential consequences for victimised women. Three street-level responses to pandemic restrictions are presented: professionals who complied with new restrictions, professionals who negotiated some far-reaching restrictions, and professionals who resisted restrictions. Consequences for the social services, women experiencing DV, and civil society are discussed. The authors suggest that the results provide learning opportunities for managers and social workers to better understand the complex everyday life that surrounds their mission.


Introduction
Since the winter of 2020, COVID-19 has posed major challenges for anyone who has had to face the risk of becoming seriously ill or face restrictions on personal freedom due to lockdowns. Such restrictions have limited people's freedom of movement, forcing them to stay at home. This has particularly affected people with failing health, difficult home conditions, homelessness, poor finances, or family tensions that may have escalated during the pandemic. Challenges have also arisen for people at the forefront of social work and health care as they have had to deal with new situations in their work. These changes have led to cutting and re-prioritising previous standards of services, which may have affected social workers' ability to carry out their work. Social workers have had to balance contradicting demands (e.g. protecting themselves, their co-workers, and clients) and providing quality support and help to their clients. The pandemic has disrupted ordinary practices and created significant uncertainties. No matter how one responds to these challenges, there is a risk that things will go wrong (Banks & Rutter, 2022), meaning that professional judgment is needed. COVID-19 resembles an exogenous shock that might create radical institutional change with new ways of working and new professional norms (Mahoney & Thelen, 2010).
This study heeds the call for more empirical research that investigates how professionals perform their work during a crisis (Brodkin, 2021) and provides opportunities for reflections and learning to improve practice. We explore how social workers specialised in the field of domestic violence (DV) balance the expectation and need to protect themselves with the need and expectation to protect their clients and how this balancing affects their discretion. DV is a particularly suitable case to study since women run a greater risk of being exposed to violence during the pandemic due to isolation at home. This isolation may create a need for urgent social interventions and protection. Specifically, this study will answer two questions: RQ1: How do social workers within DV handle the risk of becoming infected with COVID-19? RQ2: How do social workers within DV support their clients during the pandemic?
During the last decades, DV has attracted increased political attention in Sweden with stricter legislation, training of social workers, public funding for women's shelters, and the implementation of evidence-based treatments and risk-assessment instruments (Skillmark et al., 2019). According to the Social Service Act (2001:453), municipalities have the ultimate responsibility to provide help and support to those in need and women (as well as children) who are subjected to violence are considered a target group. New binding regulations and advice were issued by the National Board of Health and Welfare in 2014 (SOSFS 2014, p. 4). They mainly concern the social services obligation to offer victims of DV urgent support based on the need for protection as well as formalisation of investigation and assessment work. General advice suggests that the social services should offer counselling, permanent accommodation, parenting support, and support in navigating the welfare system. Now specialised social workers in Swedish municipalities provide such support, including housing and long-term treatment, for women exposed to DV. We have followed two municipalities' work with DV since 2013 and have also collected data from vulnerable women exposed to DV and collaboration partners, data that will be reported in subsequent articles.

The problem as highlighted by previous research
This review of current research related to COVID-19 start with how the pandemic has affected social work in general and DV in particular. Then we describe what is known about the pandemic's consequences for abused women. The pandemic is an exceptional situation that burdens health services with many intensive care admissions and increased pressure on staff. At the same time, there was insufficient protective equipment and a lack of preparedness. Vulnerable groups were hit harder than others, and socio-economic and medical factors influenced how they experienced the pandemic. In Sweden, voluntary infection control measures, recommendations to work at home, and personal responsibility were emphasised rather than more intrusive measures such as the stringent lockdowns implemented in many other countries. A national commission concluded that the Swedish pandemic preparedness was substandard and tardy, and these shortcomings affected staff in several fields (SoU, 2021).

Challenges for social work
A Finnish study that followed 33 social workers for a few months showed that the pandemic has challenged workers at all levels, including the lack of face-to-face meetings and deceased relationships with collaboration partners . The pandemic exposed other difficulties related to administrative, managerial, and technical weaknesses as well as strains within established working groups. It changed the working life of social workers now using protective equipment, gloves, and visors and reduced home visits and travel. Therefore, some of the work was conducted via video and telephone. Other studies have shown how social workers need to improvise and be adaptive and creative as well as how the pandemic has made relationship-based social work more difficult (Ashcroft et al., 2022;Ferguson et al., 2022;Kong et al., 2022;Manthorpe et al., 2021). Banks and colleagues reported on ethical challenges creating and maintaining client trust, especially when the work was conducted at a distance (2020). Consequently, Manthorpe et al. (2021) report how restrictions on discretion in a time of crisis might undermine good practice (see also Banks & Rutter, 2022). Research indicates negative effects due to changing working conditions for frontline staff in social work.

Challenges for abused women
There is no clear picture of how the pandemic has affected women exposed to DV according to the Swedish Commission. On the one hand, there are reports of an increased influx to non-profit organisations such as women's shelters; on the other hand, there is a significant reduction in women's contacts with authorities (see also Petersson & Hansson, 2022;SoU, 2021, p. 63). According to some international research, women are increasingly exposed to DV: 'Domestic violence rate are rising, and they are rising fast' (see also Ashcroft et al., 2022;Bradbury-Jones & Isham, 2020, p. 2047). Because of COVID-19 restrictions, abusive partners can increase their power and ability to enforce control (Kaukinen, 2020). Sharma and Borah (2022) call the COVID-19 pandemic an exogenous shock that significantly impacts incidence of DV around the globe. Helpline inquiries (Bradbury-Jones & Isham, 2020;Kofman & Garfin, 2020) and Google searches about DV have increased (Usher et al., 2020). When police are perceived as hesitant to intervene and sheltered housing becomes more inaccessible, women's vulnerabilities will increase (Bennett et al., 2021;Buttell & Ferreira, 2020;Wood et al., 2020). At the same time, there are hidden numbers internationally as only between 2,5% and 15% of women exposed to partner violence report the violence (Akel et al., 2022). According to a systematic review of international trends, DV increases but reports of DV decrease when victims and perpetrators are isolated together (Kourti et al., 2021). In summary, although empirical research is not clear on the actual increase in DV during the pandemic, there is an unequivocal risk of growing partner violence because of increased isolation and its strain on family life and family relationships.

Consequences for practice
Kingstone et al. use the term 'managing uncertainty' in an analysis of social workers' work during the pandemic based on interviews with 13 social workers from the UK (Kingstone et al., 2022). They identify difficulties providing social care at a distance, the challenge of working from home, the management of risks in the face-to-face meetings, as well as the potential long-term implications (see also Kong et al., 2022). Although the latter means future cuts, it also opens new ways of working that can benefit clients. As in other social work fields, DV practitioners have experienced significant changes in the way they deliver services (Cortis et al., 2021;Emezue, 2020;Wood et al., 2020). Some changes seem to appeal to practitioners such as the possibility to improve accessibility and efficiency through digital solutions as well as reaching new (younger) women (Petersson & Hansson, 2022). However, such changes might negatively impact work as telephone calls make it impossible to assess non-verbal cues and determine whether the perpetrator is monitoring phone calls (Cortis et al., 2021; see also Kong et al., 2022).

Addressing a knowledge gap
With unclear and constantly changing COVID-19 guidelines , increased demands, uncertainty, and insufficient resources (Gofen & Lotta, 2021), much faith is put on the local social service organisation and its personnel to make sense of and navigate their work during the crisis. The above-cited research shows the need for professional actors to use their judgment and creativity when working to defend social justice, combat DV, support clients, and protect themselves from illness. However, how professionals handle this challenging and stressful situation is not well researched (Casquilho-Martins et al., 2022). Most of the research we have taken part of concerns social work at the beginning of the pandemic. At the start of our research, the pandemic was in its second year, so street-level responses had moved beyond the more acute phase. This provided us an opportunity to explore how professional actors managed interruptions and how they worked to restore organisational stability over a longer period.

Methods and analytical venue
The qualitative study was conducted in two cities in Sweden during the summer of 2021. These cities had been developing specialised units for two decades and now serve a population of about 170,000 people. Since 2013, the research team has conducted several studies focusing on how they organise and implement interventions within DV as well as on how their clients experience the provided support. Following the outbreak of the COVID-19 pandemic, we extended the ongoing research to explore how the two organisations managed the pandemic and how victims of violence were affected. The analytical focus in this article is on the practices of discretion in times of crisisi.e. how actors balance autonomy/external restrictions and the judgments they need to exercise (Hupe & Evans, 2020). In this case, concerning their own, their colleagues, and their client's safety and wellbeing.
The empirical material draws on interviews with seven social workers (SW 1-7) specialised in DV and working directly with victims in the two cities, two coordinators/team leaders (TM 1, 2) (who share their duties between clinical and administrative work), and two managers (M 1, 2). These professionals were selected due to their roles as key actors in the daily task of organising and providing support (such as assessment, decisions, and counselling) to victims in the organisations. One manager refrained from participating because of time constraints.
Because of governmental restrictions and injunctions, we conducted the interviews using Zoom. An interview guide was developed, with slight adjustment to fit the respondents' positions in the organisations. The construction of the guide was based on our interest in the various ways the organisations, and particularly those on the frontline, responded to the extraordinary conditions due to the pandemic. They were asked about their work experiences since the COVID-19 outbreak, the consequences for victims, how they interpreted information about the pandemic, organisational adaptability, and their view of work post-pandemic. We frequently made use of follow-up questions and explicitly asked for concrete examples from their daily work.
The interviews were conducted by the first and second authors and lasted between 33 and 57 min. The interviews were conducted in Swedish, recorded, and transcribed verbatim. A thematic-analytic approach was used to analyse the transcripts and recordings (Braun & Clarke, 2006). We started by listening to the recordings and reading the transcripts. With the aim and research questions as a backdrop, each data item was then coded manually and collated together. Terms used in this stage of the analysis were close to the raw data. Examples of codes were for instance 'avoids physical contact' and 'positive about remote work'. The succeeding interpretations and the development of themes were made by moving back and forth between the data, previous research (see introduction above), and literature on the use of discretion in human service organisations (Evans, 2020;Molander, 2016). Following Brodkin (2021), who suggests that frontline workers can respond to crisis by adapting practices, resisting management demands, or even changing the directions of work, we paid particular attention to accounts signalling how the participants positioned themselves in relation to changes implemented in the organisation (for instance, why they were positive about remote work and complied with restrictions in their practice). Drawing on Molander's (2016) description of discretionary reasoning, we interpreted responses as conclusions about what actions are called for under conditions of uncertaintyi.e. when balancing protecting women from violence and decreasing the risks associated with the virus. The norms used to justify different responses were further scrutinised and interpreted. For example, when participants talked about how they negotiated restrictions due to the importance of proximity to colleagues when making decisions, we turned to research on street-level decision making and the significance of deliberation (Møller, 2021;Zacka, 2017) to understand their accounts. This abductive analysis led us to construct three empirically overlapping themes, which structure the results section below.
This research has been approved by the Swedish Ethical Review Agency, 2021 (2021-02442). Informed consent was provided by all participants before the interview.

Results
The interviews revealed three street-level responses to restrictions posed by the pandemic. First, professionals spoke about how they complied with pandemic restrictions. However, there are two sides to this story: maintained quality of work that justifies compliance and enforced compliance that deteriorates practice and the life situation of the women. The latter calls for new and creative solutions. Second, professionals negotiated practices to counter the negative effects restrictions would have on the services they provided to clients. However, as will be evident below, because some professionals were also fearful due to the apparent lack of restrictions, they negotiated for the implementation of further protective actions for themselves and their clients. Third, some professionals resisted restrictions altogetheri.e. they claimed to break the rules for a higher moral purpose, serving women in need of protection.

Complying with restrictions
Unsurprisingly, participants reported disruptions in their day-to-day practice because of the pandemic. The most common restriction designed to avoid infection was the expectation to minimise physical interactions with clients and colleagues by shifting to remote service delivery. There is somewhat contradicting evidence concerning consequences of this shift, from inhibiting relationship building (see Kingstone et al., 2022;Pascoe, 2022) to enhancing accessibility (see Cortis et al., 2021). Different voices are also present in our material. Consider the following statement from a social worker who switched to remote practice early in the pandemic and primarily provided counselling to women who had left their abusive partners: Many [women] connect [to online-counselling] even though they may be little sick. Even though they do not have the strength. Its less driving, less time away from work, many have been able to solve it, even if they work, they can connect to a meeting. (SW 1) According to this social worker, to meet clients remotely meant fewer dropouts, while at the same time experiencing a possibility to 'get close to the client' and 'see if they have tears in their eyes'.
Others were initially highly sceptical of the organisational restrictions, but after switching to remote practice they as well as their clients were positively surprised: And I also asked my clients, how does it feel and how is it [to use digital solutions]? And to my surprise, and to their surprise, it also turned out to be, no, it did not make a huge difference. It works great. And then we did it, looked at each other, and it went really well. (SW 5) However, such positive statements from the interviews should be further scrutinised. First, according to our data, for remote service delivery to work well, the women's life situation should be stable without acute risk for exposure to violence. For example, when SW1 reflected on why she thinks the transformation to digital solutions has worked so well, she noted that 'it is probably mostly based on the fact that I have had cases where they have left [their abusive partners]' (SW 1). As other research has noted, supporting women remotely when a violent partner may be at home poses significant challenges for social workers (Cortis et al., 2021). Second, as evidenced in the second quotation above, it is easier to switch work procedures after relationship and trust have been created face-to-face, for example, when 'we [the social worker and the client] sort of talk the same language' (SW 4) and 'you [social workers] can continue with remote counselling when you [social workers] have established a contact, and you know, have had an optimal chance to create an alliance' (M 2).
Other examples illustrate the tension between organisational COVID 19 restrictions and considering what is best for the clients. This was especially the case in acute situations and among those with experiences of working with assessments during the pandemic. The following example, described by a social worker who switched to counselling duties during the pandemic, illustrates this: SW 4: Working with assessments, it was a big difference. If you compare with how we worked before [pre-COVID]. Because suddenly, we were not allowed to go and pick up women as we had done before. We were not allowed to sit in the same car with them.
I: Who decided that? SW 4: The managers, the organization, you know, we were not allowed to. We had a car that was ours, which we used to pick up [women], we had to like go urgently if needed, but based on the restrictions we were not allowed to sit in the same car, so a taxi had to pick them [the women] up, or they had to get to sheltered housing on their own. But then, then there were many shelters who changed their routines so that they could instead pick them up, so that solution was made that way.
The usual way of supporting women in crisis, driving women to shelters, was now no longer an option. The social worker also described how women were instantly sent away when coming to the social services waiting room and that risk assessment was impossible to do over the telephone or using digital tools, which meant that 'we did not do it, that assignment was placed on the women's shelters' (SW 4). Our interpretation is that this shift in responsibility primarily seems to have been implemented to protect the professionals from infection, acknowledging their, rather than the women's, vulnerability. The fact that the victimised women needed to sit in a car with other people was not the issue; the issue was the protection of the social worker.
However, rigidly following local organisational imperatives does not fully explain compliance with restrictions in cases where social workers identified a possible reduction of work quality. Another social worker, recounting an interaction with a client with cold symptoms, noted that 'neither of us [the professionals] felt like sitting in the same car, not even with protective masks' (SW 6). Since women need and indeed are entitled to protection, other solutions were called for: So, we found a person in her network who could, since this person had met her and potentially already were infected, it was not an optimal solution based on protection maybe, but it was, we did a safety plan around it, so that it became someone in the network who drove. (SW 6) Rather than turning to other actors in the DV field (as in the example above), this social worker recruited help from a person from the woman's network to manage her feelings of discomfort.
Our data illustrate how social workers, when they comply with organisational demands, act under conditions of uncertainty even when such actions are not optimal. To support victims of DV and remain safe, social workers made use of their creative discretion to solve problem and craft new work proceduresto think outside the box (see Evans, 2020). As this case reveals, solutions outside the box are contingent on other actors who put their own risk aside to support women.

Negotiating workplace restrictions
Our informants told us that the lack of proximity to colleagues and a wish to discuss assessments and decisions collectively compelled work-place negotiations about COVID-19 restrictions. The following excerpt comes from an interview with a social worker who started working in her current position (assessments in DV cases) during the second wave of the pandemic in Sweden: SW 2: Yes, it is difficult, and we raised it with our manager quite early, that we cannot see how our tasks could be done from home. We said that we want an exception from this demand. Because it was a demand that we should work from home. I: mm SW 2: And then we finally got it approved, we are not so many who work with violence, and we have our own corridor, so really, there were no obstacles really for us to be able to work at the office. But yeah, before we understood, because both me and my colleague were new on this position, and it took some time for us to understand that this . . . this won't work.
She also criticised the idea of work becoming more efficient due to limited travel time to and from different collaborating organisations since 'you sort of prepare for the meeting during the ride in the car, and afterwards you can sort of land in what you have gained and reflect on what you had and maybe talk to a colleague in the car' (SW 2). The uncertainty and complexity inherent in DV work, while also dealing with the uncertainty of being new on the job, called for negotiations about the requirement to comply with organisational restrictions and working remotely. The 'demand to work from home' was not acceptable according to this social worker since it inhibited proximity to colleagues and, by extension, proper decision making. This reasoning is in line with Zacka (2017) proposition that street-level bureaucrats often look to their colleagues for cues about the proper action to take in their daily doings. The importance of colleague's rests in the ability to share the burden of discretion (Molander, 2016)i.e. to reason about the appropriate path to take in individual cases (Pascoe, 2022). In relation to this argument, the team leader of the social worker cited above developed reflections about limiting restrictions that interfere with providing services: TM 1: You need help in these cases as well and it is also, I think it has also been very difficult. Somewhere around Christmas these two social workers would then rotate so they did not meet, one at home and one in the office. But when they are only two and it is a case that becomes urgent or needs to be placed in protection, then they must work with the case both of them. Then it is very difficult if they do not meet or are working on distance. So, they actually asked to work in-house, they sort of saw that it was impossible to work like that.
These examples illustrate how the social workers stepped forward and questioned institutional constraints by voicing their concerns about limited spaces and places to reflect on cases. That is, they expanded their discretion with the consent of management. We interpret these actions as a way to safeguard what Møller terms deliberative organisational routines with low degree of formalisation (e.g. ad hoc discussions in a car): 'organisational routines that enable deliberation, during which one or more cases are discussed and potentially decided upon ' (2021, p. 478).
According to our participants, those who work with assessment and subsequently granting (or rejecting) interventions such as safe accommodation or in-house treatment experienced strict pandemic rules imposed by management. One social worker reflected about such differences: Especially those who work with assessments and decisions, they have very, at least in this municipality, have very strict rules to follow, they do not get, they don't really have the same freedom to decide in each individual case if they are allowed to meet physically or not, we have a little more flexibility there, I think. (SW 7) Another participant reflected about her experiences of differences between working with assessments and working with counselling during the pandemic: There were meetings and the managers got together and they even had a meeting where they informed about how we should wash our hands correctly, and like showed us how to do it. So, there was a lot of information, and they were very scared, the managers. When we started to work here [at the counselling unit] it was nothing at all. It was more like we had to ask, how do we think about Corona, how will it be safe, what do we answer when clients ask, what do I do when I meet a client that is worried? And then the manager had to think a little. And, not much happened. It was not until I got sick that the manager has an obligation to tell senior managers and then they must do some kind of analysis of where it comes from and then it became a little more, but barely. (SW 4) The above excerpt shows that social workers not only negotiated to limit some restrictions as illustrated previously but also raised their voices when their own and their clients' protection were not taken seriously enough by management.
Our previous research from these municipalities have shown that the counselling units have had significant autonomy within their own jurisdiction and trust from the local management and the political domain (Skillmark et al., 2019). This seems to apply also during the pandemic although now professionals are requesting more guidance. Soft Swedish public governance through recommendations has enforced local municipalities to devise their own restrictions. Negotiations of restrictions could be interpreted as local sensemaking activities, spurred by an ambiguous event, inconclusive directions, and with the aim to restore stability (Sandberg & Tsoukas, 2020). In play here are different norms (protection and deliberation) that work to justify negotiations concerning how to work. However, discrepancies between different norms are not always made visible and negotiated, as will be illustrated below.

Resisting workplace restrictions
In comparison with negotiating restrictions above, the interviews with the social workers revealed a few instances of more informal modes of organisational 'misbehaviour' or resistance to restrictions: 'I guess I am a bit to rebellious [. . .] I don't ask if its ok' and 'what are they [management] going to do [when I meet clients face-to-face]?' (SW 3). Here, women's safety and vulnerability trump the risk of infection and therefore justifies resistance. The following example further illustrates this dilemma: SW 3: And it is also strange, we would go, for example, me and my colleague would go and place a woman up in the country, and they [management] wanted us to take one of those minivans, and then you should go with one of those minivans that is covered in plastic, it just felt, it did not feel like, you know 'alien', when they come in and it almost feels absurd.

I: Mm
SW 3: This is about vulnerable people.

I: Yes
SW 3: It feels really weird.
I: And what did you do then? SW 3: No, but we . . . we were going out to do this placement and it was a young girl who was particularly vulnerable, a particularly vulnerable group, and then she got to sit in the back seat.
The social worker described how her organisation had prepared a minivan with different safety devices to minimise infection risk. This was done in an 'absurd' way, making the minibus resemble something 'alien' and 'really strange', seemingly unfit with the needs of vulnerable persons. When supporting a young, victimised woman, she and her colleague abandoned the idea of using the minivan according to local regulations. Instead, they travelled by car with the client 'sitting in the back seat' (SW 3). Here, we see resemblance with research from other contexts. Banks and Rutter, for example, observed the use of ethical creativity that entails 'questioning, bending or breaking institutional rules and norms ' (2022, p. 3471).
In this case, there was no lack of or confusing guidance on how to act but rather a clash between what was considered right according to their professional judgment and what was expected by the organisation. That is, they acted 'just as the Public Health Agency does, you know, you put it [risks] against each other' (SW 3). Breaking rules does not seem to trouble the social workers or make them fear repercussions: 'I really think the manager appreciates us not talking about it [how we broke the rules]' (SW 2).

Discussion
This study explores how (1) social workers specialised in the field of DV balance the need to protect themselves and their colleagues and (2) the need to protect their clients and how this balance affects their discretion. The answers to the research questions are found in the three street-level responses complying, negotiating, and resisting. We will now deepen and problematise these results before reflecting on the study's limitations and highlighting implications for social work practice.
As our results illustrate, although there was fear and scepticism associated with switching to remote working, under some conditions and circumstances it could benefit clients (see also Cortis et al., 2021) if the clients have access to digital solutions and if relationship and trust have been created. Experiencing client benefits facilitated further compliance with restrictions. Nevertheless, we concur with others (Banks et al., 2020) and raise a warning against what many might perceive as a practical possibility rather than an 'unchecked digital erosion' (see Pentini & Lorenz, 2020, p. 548). Problems of confidentiality may arise when social work becomes dependent on services from large commercial tech companies such as Metaverse and Alfa (see Zuboff, 2019). The meeting between a social worker and a client, usually understood as a somewhat protected zone, might be abolished in the digital era. Clicking down one meeting and clicking up another (as one social worker said during the interview) risks hampering thorough judgments and therefore overriding what Banks and Rutter (2022) call slow ethics (i.e. taking time to reflect about ethical implications) as well as hinders possibilities for deliberation to support learning (Møller, 2021;Zacka, 2017). Our results show how the social workers faced these challenges by negotiating or resisting workplace restrictions.
In social work in general and in DV work in particular, individual and often acute interventions require general knowledge of social problems. As there are no simple and clear paths, social workers must rely on their ability, experience, and capacity to assess situations and support victims of DV. This is where the room for creative discretion becomes central, as legislation and regulations cannot provide sufficient support in all client-centred work (Evans, 2020). However, as shown in our results, even creativity has its limits as women exposed to men's violence need timely help. The restrictions organisations have put on social workers, causing some abused women to search for other solutions, are in line with other research from civil society that reports an increase in people seeking support during the pandemic (Petersson & Hansson, 2022).
What does our case say about the exogenous crisis that has affected social work? Our material shows the change to use distant methods that have become possible in our digital age. Social work is strongly institutionalised through education, legislation, and professional experience. Normative assumptions about how social work should be conducted through relationships and faceto-face meetings are strong and stable. This has been put to the test through lots of new restrictions during the pandemic. Distance work entails rationalisation and increased accessibility for some people, but also that the tradition-based social work with closeness and alliance building are changing. Covid-19 may be the push into a new way of working that will be difficult to get out of and people without technical resources might be excluded.

Limitations
The data in this study are limited in scope and focused on work in two of Sweden's municipalities, so generalising the findings to other contexts should be done with caution. However, our results are in line with previous research from other settings and contexts, which seem to resemble the challenges faced by social workers during the pandemic (see Banks et al., 2020;Manthorpe et al., 2021;Pascoe, 2022).
These similarities might open for more analytical generalisations and transferability of the results. Second, interviews focus on what participants say they do rather than what they actually do. An ethnography would probably provide more detailed and in situ considerations to analyse. The pandemic dynamics, including stay-at-home policies, were an obvious obstacle for us in this regard.

Implications
This research provides learning opportunities for managers and social workers. The presented typology (compliance, negotiation, and resistance) might support managers to better understand the complex everyday life that surrounds their mission, especially in times of crisis. Awareness of the three different positions that social workers occupy and their need to balance different demands provides insight into the negotiations that the work entails. Managers' ever so clear directives and wishes are not only complied or resisted but often transformed into negotiations that social workers carry out to protect their interests or those of their clients.
We also suggest managers in social work to pay attention to the extent of the potential shift from publicly run social work to voluntary efforts and efforts made at women's shelters and within civil society. According to our data, social workers' discretion has in some cases been circumscribed; they have not been given the opportunity to work fully in risky situations, and the consequences of this for victimised women remain unclear (see Manthorpe et al., 2021).
Finally, social workers should be attentive how they perform social work and how their discretion creates possibilities they might not be aware of. The social workers in this study implemented new ways of working through digital solutions, a development that can be perceived as a future threat or opportunity. They also used soft spots in their negotiations and bent the rules. Their managers closed their eyes and accepted a space for discretion. Admittedly, they have added new rules for how the work is to be carried out. They took risks and organised their work in their small contexts. We regard this as inspiring since they proved to have a great deal of room for their own application and their own interpretations of risks. However, this study identifies a future field of tensions that will affect the working conditions. This study provides opportunities to reflect on changes in social work during the pandemic. Institutional changes occur through severe crises, more often slowly and gradually. The pandemic shows that albeit a crisis, there are still opportunities for social workers to use their room for discretion. We look forward to further studies of how this room is affected by the opportunities and challenges of the ongoing digitalisation.

Notes to contributors
Mikael Skillmark (PhD in Social Work) is Senior Lecturer at the Department of Social Work at Jönköping University. He is particularly interested in the implementation and execution of assessment tools in the social services and what consequences standardisation might have for social work as a profession and a field of practice. He has also published books and articles on social work, masculinities, domestic violence, and victimisation.
Verner Denvall is Professor in Social Work at Lund University and was previously Professor at the Department of Social Work at Linnaeus University. His work is related to evaluation and planning with a focus on issues of implementation and de-implementation of interventions to combat inequality and domestic violence. He has published articles together with colleagues on meta-governing EBP and motivational interviewing and is currently co-editing a book on participation.
Cecilia Kjellgren is a clinical social worker and associate professor in social work, Linnaeus University. She is experienced from clinical work with children being victims of abuse and neglect. One field of research is consequences of physical abuse of children and the effects of interventions. Another field focus on young people with harmful sexual behaviour; prevalence, risk factors associated with the behaviour and outcome of the interventions offered. She runs research projects on the aftermath of institutional sexual abuse and is involved in research exploring interventions for victims of domestic violence and effects experienced by clients.
Johanna Thulin (PhD in Social Work) is Senior Lecturer at the Department of Social Work at Linnaeus University, Sweden. She is experienced from clinical work with abused and neglected children placed in out of home care. Her areas of research interest include child physical abuse and domestic violence, its consequences and effects of treatment, with special interest in how children and women experience their participation in social welfare processes.