Implementing Focused Deterrence for Intimate Partner Violence: The SRFV Project in Malmö

5.1.1 Organization

SRFV is a collaborative project involving the social services, the probation service, and the police. The project is overseen by a cross-agency steering group with representatives from Malmö municipality, the police, the probation services, and the County Administrative Board. Malmö university, responsible for the evaluation, is also represented in the steering group to update project owners about the progress of the evaluation. SRFV is operationally overseen by two project leaders representing the social services and the police, respectively, who are also members of the steering group. Despite being an inter-agency collaboration, the project is primarily anchored around the police. This is largely because cases are initiated through police reports and because the police have broader legal possibilities to share and access information compared to other agencies.

To organize the project and the operational work, a project group was established. The project group consists of approximately 12 representatives from the police, the social services, the probation service and the social administration of Malmö municipality. The project group members have different roles within their respective organization: coordinators, secretaries, probation officer, and analyst. The group has remained relatively stable over time, although some members have been absent for short periods or have left entirely, while others have joined over the course of the project. The project is implemented under the guidance of NNSC.

5.1.2 Description of the strategy

Drawing from the IPVI framework, the SRFV strategy applies focused deterrence principles with the aim of reducing IPV within Malmö municipality. By focusing on interventions as a form of secondary and tertiary prevention, the explicit goal of SRFV is to reduce intimate partner violence, disrupt the cycle of abuse, and prevent its escalation into more serious crimes (malmo.se).

A key feature of the SRFV strategy is inter-agency collaboration, with a primary focus on information sharing to enable a more detailed and comprehensive understanding of IPV cases. Since each agency holds different pieces of relevant information, collaboration facilitates more informed assessments of which cases to prioritize and how to tailor interventions to the specific circumstances at hand. In addition to its practical utility, collaboration is also thought to serve a symbolic function by signaling that society stands united against violence. To support this work, representatives from the police and social services hold weekly briefings to exchange information about ongoing cases.

In the strategy, suspected offenders are classified into four categories, labeled A through D. Each category is based on the following criteria:

In the Swedish system, “reasonable suspicion” is based on observed facts, but requires a lower standard of proof than “probable cause” and is used to justify police interventions such as temporary detentions and limited searches. Probable cause, on the other hand, requires a higher likelihood of involvement in a crime and is necessary for more invasive actions, such as arrests or search warrants.

In each category, both the offender and the victim receive basic information about the project, either verbally or in writing. Beyond this, each offender level has a set of interventions, referred to as the response chain (see Table 2). Some project interventions are newly introduced, while others are the same as prior to the release of the project but are now implemented within a newly structured framework.

A new aspect of the strategy applies to individuals at the D-level, where no specific suspicion of a crime exists. In these cases, the residence is flagged in an incident report. Based on this report, the SRFV group can submit a concern report to social services—a step that was previously reserved for cases involving children in the household. Additionally, social services now proactively reach out to every victim at the C-level, offering guidance on how to access support.

At the B-level, a key intervention in the strategy is custom notification. During this intervention, police and social services hold a face-to-face meeting with the alleged offender to deliver a clear message: violence will not be tolerated, help is available, and any further violence will have consequences. This meeting occurs only with the offenders’ consent, and it is not related to any ongoing criminal investigation.

While some measures are not entirely new, they are now implemented in a more structured and systematic manner. For instance, at the A-level, focused investigation entails prioritizing cases involving persistent offenders, who may also be suspects in other crimes. Although the investigative procedures themselves remain largely unchanged, the added structure ensures that such cases receive heightened attention.

It should be noted that this is the protocol for SRFV. When put into practice, unforeseen challenges may occur. Therefore, during 2024, the strategy has been tested through seven scheduled flow tests: limited time periods where a certain amount of personnel works according to the strategy, followed by discussions on its implementation. With each flow test, the number of personnel managing IPV cases according to the strategy has increased, providing opportunities to observe case management in practice as the workload increases. During the flow tests, more than 70 cases were handled.

The flow tests were for the most part described in positive terms. During the time period, positive relations between agency representatives emerged, information was being shared on a set schedule, and the strategy was well received outside of the project. Meanwhile, police analysts created a registry where each client is assigned a unique code name linked to various key variables, such as interventions and re-offenses. This data will be crucial for evaluating the strategy’s impact. Beginning in 2025, the goal is to scale up the project to involve all personnel working with IPV cases in Malmö. At this stage, the program can be viewed as a “proof of concept.” An impact evaluation is planned, and if the results prove successful, the ambition is to further develop the strategy (Mellgren et al., 2024).

5.2.1 Sharing information?

Going into the project, the project members had a very positive mindset about collaboration as a way to reduce IPV. Typically, when it comes to collaboration between agencies, the possibility to share information is a key aspect. Uncertainty about the right to share information is a common challenge in collaborative projects and has also been documented in similar projects in Sweden (Ekström & Olin Dahl, 2023). Hence, an important early step was to assess to what extent the implementation of IPVI would be legally feasible in a Swedish context. To this end, a “legal framework” was developed for the project by lawyers from the different agencies, examining the legal prerequisites for the strategy. Put simply, it was concluded that a Swedish version of IPVI could indeed be implemented in Malmö. Although, the answer is not entirely straightforward—some pieces of information are easier to share than others, and some information can be shared by some agencies, but not with others.

The Swedish Secrecy Act (OSL) is a complicated legislation to work with, an issue that does not get resolved by one project. In addition to assessing the legal admissibility of the strategy, the legal framework was also designed to support practitioners in their operational decision-making. Members of the project group expressed strong appreciation for this document. Since different agencies have access to different types of information relevant to the project, the framework plays an important role in guiding information sharing. For example, the probation service may hold information about a client’s mental health, difficult separations, or substance abuse—details that can inform the categorization of cases and the tailoring of interventions. Similarly, social services may have received multiple reports of concern in cases that have not been reported to the police. The general perception was that this document would provide a solid ground for practitioners in their daily work. Although this might have been the case, difficulties related to confidentiality would inevitably still arise. As Johanna, one of the lawyers behind the legal report, expressed during an interview:

Yes, it’s a challenge, as I see it anyway [….] We cannot give general answers that you can always share certain information, that assessment must always be made in the individual case. And that can be time-consuming, and it requires administration and that you actually sit down and look at the circumstances.

This issue is central to the project. For instance, agencies are expected to respond swiftly to reoffending, but doing so requires timely information sharing about new offenses. The extent to which this is legally permissible depends on the specifics of the case. If a child is present in the household, information can be shared more easily to assess potential child protection needs. Additionally, the legal basis for sharing information depends on the nature of the crime. If sharing information can be assumed to contribute to preventing a crime with at least one year of prison in the penal code, confidentiality can be breached (OSL 10:18 c). The citation sheds light on the difficulties that follow with the adoption of new strategies: the legal ambiguity in procedures of sharing information makes the thought of institutionalizing or making such sharing “common practice” challenging. In handling these complexities, the project group has occasionally looked to Sluta Skjut for guidance, as it is also based on focused deterrence principles and collaboration between agencies. However, compared to gang violence, IPV was often described as more complex, particularly in terms of information sharing. A probation service lawyer noted that, unlike GVI, consent would play a central role in the SRFV project:

This type of crime is not considered to be organized crime or other crime that causes us to have a breach of confidentiality provision. We don’t have it when it comes to violence in a close relationship, but it [sharing information] is perhaps about the fact that we have to get consent.

The fact that obtaining consent, whether from the offender or the victim, is a key factor that will probably have significant implications. Unlike GVI, this requirement is likely to highlight the importance of building good relationships with both offenders and victims. This will, of course, demand more attention to engagement and communication, which reflects the unique dynamics and sensitivities involved in addressing IPV compared to when dealing with gang violence.

For better or worse, secrecy laws are not only legal boundaries, but symbolic ones. In many ways, representatives from these agencies have formed a new agency unit. They reason together about cases, work together directly with clients, deliver public presentations about the project, and have shared experience of working with colleagues within their respective organizations who are not part of the project. These shared experiences may foster a sense of unity, a “we” that exceeds organizational boundaries. In this context, confidentiality underscores the lines between the agencies that still exist. Importantly, these boundaries are not necessarily problematic. Different agencies bring different tools, mandates, and professional identities to the table. These differences can be a strength by offering complementary perspectives, even as they sometimes complicate communication or decision-making. The project group is, in this sense, engaged in a balancing act: working to establish a shared identity and collaborative practice, while also navigating the tensions that arise from organizational diversity. This ongoing negotiation of unity and difference is central to the work of inter-agency collaboration.

Confidentiality rules may also give rise to peculiar and time-consuming situations. While the police can share some pieces of information with both social services and the probation service, these two agencies are not permitted to share the same information directly with each other. This creates a scenario where the police act as a messenger between the two, which ultimately receive the same information. For much of the project’s duration, information was exchanged in an analog form. This required practitioners to agree on when to meet, and to physically deliver documents. A police employee later suggested switching to digital information sharing, using personal codes assigned to clients within the project. This would enable direct contact between agencies in a faster and more flexible way. As the time gone between the incident and the response was viewed as a crucial aspect when dealing with IPV cases, digital contact could potentially have positive effects. This proposal, it turns out, was permissible, as it does not reveal any personal information about the client. This was not the original intention with the project, but it became a positive side effect. Although not part of the project’s original design, this proposal was legally permissible, as it did not involve disclosing personal data, and it emerged as an unintended but valuable outcome.

Despite extensive preparatory work, confidentiality issues have repeatedly arisen during meetings and have arguably been one of the most impactful aspects of the implementation process. These issues can be relatively minor, such as how the project group members are allowed to talk about the cases depending on who is present in the room. However, confidentiality issues may also become an obstacle when it comes to circumstances that are central to the strategy of the SRFV. For instance, information about whether or not a client has received and accepted support measures is crucial to authorities that are expected to follow up on the individual in question. Nevertheless, the question of whether one could share such information did not arise until a fairly late stage of the project implementation. This is particularly problematic given that a central aspect of the strategy is to communicate—and recommunicate—with the perpetrator.

A great deal is happening in the crime policy area regarding these issues. It is possible that the confidentiality legislation’s form might change, although the outcome remains uncertain, as does its potential impact on the project’s legal and operational framework. In either case, the project takes place in the context of a legal culture, and the notion that paying close attention to the legal framework is essential when implementing a new crime prevention program (Bayley, 2006) appears equally relevant to this project.

5.2.2 Cultural aspects and reaching a “shared understanding”

Most participants expressed a positive attitude going into the project. In fact, the project was commonly viewed as inherently positive. This was rooted in a shared view that intimate partner violence is often not given priority, especially by the police, and is not taken as seriously as issues like gang violence, despite its often severe consequences. As Daniel, a crime investigator in the domestic violence unit, expressed in an interview: “We are a category that is far down the food chain. Even if it is a very serious crime, it can still go far down the food chain, unfortunately.”

On the same note, a member of the steering group compared the mentality toward IPV with the mentality toward other crimes:

We asked ourselves the question: Would it be possible to work with reconnaissance missions when it comes to serious violations of women’s privacy? Or a serious rape or whatever, something that seems to happen repeatedly. “No, we can’t sit and look in on families”. But we kind of do that when it comes to drugs, so what’s the difference? But there is something psychological that stops it, that we cannot intervene in a family when domestic violence is committed, so to speak, but we can do it with drugs. I think it’s mental and cultural barriers, where intimacy and privacy create that for us.

The project, it is believed, can elevate the status of working with IPV within law enforcement. Early signs of this was expressed by one of the project leaders, who hoped that the project could spark an interest in working at the IPV unit:

If we do it right, it will be more interesting to work with domestic violence. We can offer something new and fresh […] when we start talking about an operational group that has the ability to do more than just investigations… Then it is clear and obvious that it becomes more interesting for those who are police officers. Maybe they won’t care about applying to patrol, but you stay, because here you get opportunities to do police work to the full anyway.

This quote expresses not only a desire to enhance the status of IPV work but also a vision for how that might be achieved. The suggestion is that by developing a more operational and proactive approach—moving beyond standard investigations—IPV work might come to be seen as more attractive within the police organization. Interestingly, this framing does not challenge existing ideas of what constitutes “real police work.” Rather, it seems to reflect an attempt to reshape IPV work so that it fits within current occupational ideals, often associated with action, autonomy, and enforcement.

The same observation was made in an evaluation of how the Swedish police work with violence risk assessment, primarily in cases of IPV (Mellgren et al., 2012). Interviewed police officers expressed that it was difficult to get recognition for working with “soft issues,” questions relating to crime victims and how victim-centered work was not central to policing. Police culture has historically been shaped by a pressure to appear “efficient” (Reiner, 2010), which often leads to a focus on immediate, tangible outcomes—such as making arrests or responding to emergencies—rather than on tasks viewed as more administrative or preventive.

At times, this cultural tendency became a point of friction during group meetings. Police representatives wanted to achieve clear, concrete results, whereas representatives from other agencies appeared more comfortable engaging in lengthy discussions that did not necessarily lead to definitive and practical conclusions. Perhaps, this difference in need to appear efficient may also have influenced participants’ readiness to “scale up” the project. While the police favored a faster pace of implementation, social services expressed concerns about the potential impact on their workload. In part, this may have practical reasons, as social services handle cases involving not only victims and perpetrators, but also children. However, since none of the actors knew at the time how resource-intensive an SRFV case would be compared to handling cases through traditional methods, the various approaches clearly reflect a culturally conditioned difference in mindset.

Cultural aspects can also result in the undervaluation of work that requires patience, long-term planning, or collaboration with external agencies, which is often necessary when addressing IPV. Therefore, in the context of SRFV, it’s important to rethink what constitutes “real police work.” Building trust with crime victims, enhancing communication strategies to encourage offenders to seek help, and gaining a deeper understanding of the situation, beyond the scope of an investigation, are all critical for helping other agencies prioritize cases and tailor interventions. These efforts are essential in addressing IPV and are vital for the success of the project.

While the police have already taken some strides in this direction, it is not yet the norm. And cultural norms can have real-world consequences. As the project moved into its operational phase, a major challenge within the police organization has been motivating officers to perform tasks outlined in the strategy, such as delivering the SRFV message on-site. This is a difficult issue to address, given the significant discretionary power police officers have, especially further down in the organization (Wilson, 1968). However, this challenge also highlights a potential shift in police culture. The tasks central to the project, such as engaging directly with perpetrators, do not fit neatly within traditional notions of “real police work,” which have historically emphasized action, visibility, and enforcement. Although the project was not explicitly designed to challenge police culture, the difficulties in motivating officers to engage with these tasks reveal a potential need to reframe how IPV-related work is perceived. Reframing this work could help position IPV efforts as integral to policing, expanding the notion of “real police work” to include prevention and community engagement. This shift would align with the evolving role of policing, where addressing complex social issues like IPV is increasingly seen as a core responsibility.

The cultural aspects that organizations bring into the project, and how those aspects influence its various stages, were not as prominent in the other agencies as they were within the police. For the social services, the project interventions appear to align rather well with existing views and practices and is thus viewed as “real social work.” Although the social services have been assigned some new tasks through the project, such as proactively contacting clients at the C-level, the measures implemented are largely seen as similar to those previously taken—the significant difference lies in the coordination with other agencies. Regarding what SRFV adds to existing work routines for the social services specifically, a social services representative explained it this way during an interview:

It is the coordinated dance, or what should I call it [….] Each party has done what they are supposed to, but it has not been the case that we [the agencies] have been coordinated with the one being victimized, with the one exposing someone [to violence]. So it feels like there is a big difference in how quickly we can work together.

[…] previously, it [social service work] wasn’t linked to the police investigation and when they see [IPV-related] things. So, I think it makes a big difference if we can work this closely together at the same time. Before, we were running around after each other, in a way.

The quotes illustrate how the new enhanced inter-agency collaboration is what is perceived as the “new element” introduced by the project. As such, collaboration is almost perceived as a solution in itself, a view that was commonly held. This perception stems from the recognition that IPV is a complex issue, requiring a multidimensional approach. In more concrete terms, there was a perceived need to address both the suspect and the victim simultaneously, which aligned well with the responsibilities of the different agencies.

However, the varying aims, languages, procedures, cultures, and perceptions of power among different actors often make collaboration challenging (Huxham, 1996). If the project did not introduce new tasks that challenged the authorities’ self-image within the other two agencies to the same extent as it did for the police, cultural aspects could emerge within the collaboration itself. This was evident in how representatives from different agencies perceived IPV and its perpetrators, and how these perceptions influenced their communication about the issue. As John from the probation service noted:

[…] in the beginning, before I understood the police officers who are in the group … [I reacted to] how they spread the message to patrols. Of course, you have to have a certain language so that they [patrols] are also on track … but it sounded a bit harsh. That, “now we are going out on their hunting grounds, these perpetrators.” […] And when they say certain things, use a certain language, of course I get ideas like, how do they really see this target group? And how will those on patrol view the target group? Because that’s the whole thing, the whole thing [working with IPV perpetrators] is based on the approach [toward the target group].

This illustrates a difference in mentalities between agencies, which may not be surprising given their different missions. A general difference is that the probation service is based more on communication and treatment, while the police’s starting point is more about enforcing the law. Similarly, discussions emerged during the meetings about the differences between the police and social services. During one meeting the social services were described—both by themselves and by the police—as more subjective and perhaps emotionally driven in their approach when handling IPV cases. In contrast, the police were portrayed as more objective and focused primarily on enforcing the law. These distinctions, while reflecting a common view of each agency’s role, may oversimplify the complexities of their work. In practice, police work involves much more than enforcement, as it requires communication and building trust and legitimacy, especially when addressing sensitive issues like IPV.

Despite these differences, the project seems to have fostered a “shared perspective”, or at least an enhanced understanding of each other’s work, responsibilities and limits. In this sense, the SRFV project provided a platform for conversations that might not have taken place otherwise. Despite differing views, there seemed to be a shared understanding in the group that reducing violence now takes precedence over making arrests. Although, how this shared understanding translates into better crime prevention has yet to be demonstrated, as it depends on its translation into concrete interventions. Neither does a shared perspective mean that all disagreements or challenges have been resolved. Instead, it reflects a better understanding of how collaboration can bridge gaps between different professional approaches. For instance, the police seem to have gained a better understanding of the legal constraints faced by the social services, as well as the rehabilitative role played by both the social services and the probation services. Conversely, other agencies have developed a clearer understanding of what type of information is most useful to the police.

5.2.3 Theoretical and practical challenges

Ideally, crime prevention initiatives are grounded in solid theoretical foundations and empirical evidence which, on average, should increase the likelihood of achieving the desired outcome (Lab, 2023). For practitioners, however, the underlying theory of an intervention may feel distant. For them, implementing a new project is more about how it will impact their day-to-day work in practical terms—managing workload, taking on new responsibilities, adapting to new systems, and navigating the complexities of sharing information with legal uncertainties. This is evident in the group meeting discussions, where the agenda often includes various practical problems, such as finding time to schedule meetings for information sharing between the agencies or transferring information in order to categorize offenders.

A core component of the strategy, custom notifications, is theoretically interesting as it is based on the idea that face-to-face interactions can disrupt the isolated context in which violence occurs, potentially reducing a perpetrator’s likelihood of reoffending. However, regardless of whether this theory holds true, and if so, how those meetings should be conducted to achieve that effect, it was not subject to comprehensive discussion. Instead, a more practical challenge was central: it was difficult to get in contact with the perpetrators. This raised the question of how much time and effort practitioners should put into reaching out to the individual. Such aspects of translating a theoretical concept into concrete actions highlight that access to resources can influence the implementation process. These aspects may impact the extent to which interventions are carried out as intended.

On a similar note of having a theoretical model to govern practical work, a recurring discussion was related to having the categorization model as a basis for prioritization of cases. When working with a model, the need to balance between actuarial methods—making assessments based on statistical data—and practitioners’ experience, judgment, and tacit knowledge is unavoidable. During meetings, concerns were raised that the categorization (A–D) was too rigid, failing to account for the diverse needs in each case. The complex reality of working with IPV offenders cannot be “reduced to ones and zeros,” as one project group member framed it during a meeting. Although categorization is a key part of the strategy with an intent of directing resources, practitioners seemingly felt that it limited their discretion in a negative way. Discussions began around how individuals should actually be categorized. One suggestion that came up during meetings was that the criteria in the model could work as an initial categorization, which can then be updated by a practitioner. Naturally, this leaves the function of the first categorization in an uncertain spot. Another suggested approach was to find subcategories within the model categories. Certain cases labeled as category A that by practitioners were perceived to be more urgent—referred to as “A+”—while other A-clients were not seen as equally urgent. Similarly, it was stated during a meeting that “Sure, this client is categorized at the B-level, but they’re a problematic case, so of course we’ll prioritize them.” This suggests that the current way of classifying offenders may not fully capture how practitioners assess the urgency of a case in practice. One possible approach could be to reclassify less urgent A-cases to category B, prompting a revision of the model and the criteria for assigning categories. Yet, instead of seeking to adjust the model, practitioners tended to find ways to retain their discretion within the existing framework.

Determining whether SRFV is implemented as intended requires understanding how much flexibility is built into the strategy. This is particularly complex because SRFV is based on focused deterrence, which relies on guiding principles rather than fixed procedures. A core principle is the classification of offenders and the delivery of tailored responses. However, while classification is central, the specific methods for identifying categories and designing responses are left open to local adaptation. This built-in flexibility makes it challenging to define and measure program fidelity, because the strategy is meant to vary depending on local conditions, crime patterns, and available resources. In the case of SRFV, the suggested approach when classifying offenders was to use the categorization as a starting point, but to allow reconsideration following a professional evaluation. If deemed necessary, this could then lead to an updated categorization, which would be registered as a “deviation” from the formal model. So far, registered deviations from the categorization have remained relatively few, but it is still early in the process and this discussion highlights a lack of clarity in the balance between tacit knowledge and actuarial methods in guiding resource allocation.