Sep 15, 2021

Cooperation in Tertiary Prevention of Islamist Extremism

Issue Paper from InFoEx Workshop, September 30 – October 1, 2020

In tertiary prevention of Islamist extremism, civil society and governmental exit programs support individuals (and their families) who wish to disengage from violent extremist groups and distance themselves from extremist ideologies. Exit work and successful reintegration into society involves security agencies as well es very practical elements provided by municipal actors, public services, and civil society organizations. Effective cooperation between civil society and governmental actors including statutory bodies is crucial but can be challenging.



Please note: The online version of the text does not contain footnotes. To see the citations, please download the PDF version here.



Dimensions of Cooperation in Tertiary Prevention 

Clarity of Objectives and Roles 

Cooperation and Information Exchange

Individual Versus Structural Cooperation

Trust Issues

Cooperation with Religious Organizations   

Challenging Target Groups       

Research and Evaluation   

How to Transfer Good Practices   

Multi Agency Cooperation in Practice 

VVSG: a trusted umbrella organization in Belgium

NeDiS: a network in the penitentiary system in Germany 

The Safety House: a holistic approach in the Netherlands


Executive Summary

In tertiary prevention of Islamist extremism, civil society or governmental exit programs support individuals who wish to leave extremist groups and ideologies behind, while security agencies aim to detect and protect against possible security risks. A person’s successful reintegration into society also depends on more practical elements like finding a place to stay and a job as well as receiving access to the health services. Effective cooperation between civil society and governmental actors including statutory bodies that provide these services is crucial but can be challenging.

Against this background, the German Council on Foreign Relations (DGAP) in cooperation with the Federal Office for Migration and Refugees (BAMF) organized an international virtual workshop in autumn 2020 on the role of statutory bodies and multi-agency cooperation in tertiary prevention of Islamist (violent) extremism. This workshop took place as part of the International Forum for Expert Exchange on Countering Islamist Extremism (InFoEx), providing the BAMF’s governmental and civil society network partners with a platform to discuss challenges and good practices with several external experts. Among the topics discussed were the different dimensions of multi-stakeholder cooperation (such as clarity of roles, information exchange, and trust building), the transferability of good practices as well as several concrete multi-agency structures in Belgium, Germany, and the Netherlands.

Based on the discussions at the workshop, the following key recommendations have been formulated for practitioners, governmental actors working in a multi-agency setting, and policy makers:

Key Recommendations
1. Bring together stakeholders who work on the same cases for regular exchanges on needs and expectations to improve ad-hoc cooperation, establish new multi-agency structures, and strengthen existing ones.
2. Develop common guidelines for establishing working definitions, objectives, and responsibilities in order to regulate cooperation and information sharing (including aspects regarding data protection) between institutions.
3. Promote existing and well-functioning approaches to serve as role models for actors wanting to establish new cooperation structures.
4. Foster trust and the exchange of good practices by strengthening (inter)national peer-to-peer knowledge exchanges and establishing local multi-professional networks or work shadowing offers.
5. Strengthen expertise and capacity of staff working in youth welfare or public health services by providing needs-based and targeted trainings, for example on (de)radicalization processes. Allocate additional resources to ease their workload.
6. Promote a culture of learning from mistakes, for example by actively encouraging both internal and external monitoring and evaluation and by promoting the understanding of multi-agency cooperation as a continuous process.


In tertiary prevention of Islamist extremism,  civil society and governmental exit programs support individuals (and their families) who wish to disengage from violent extremist groups and distance themselves from extremist ideologies. In some cases, it is important to involve security agencies, for example, to prevent people from hurting themselves and others. But exit work and successful reintegration into society also involves very practical elements provided by municipal actors, public services, and civil society organizations: help in finding employment, especially after being released from prison, access to the health services, or support for one’s children provided by youth welfare services. Effective cooperation between civil society and governmental actors including statutory bodies is crucial but can be challenging.

Countries across Europe have established different approaches to cooperation in tertiary prevention, but it is helpful to differentiate between three main approaches: Some countries have established nationwide “multi-agency cooperation” or “multi-agency workings,” where permanent structures have been set up to enable mostly government actors to work on individual cases in each municipality or region. Examples are the Safety Houses in the Netherlands (p.15) or the so called LIVC-R units in Belgium (p.14). In those cases, tertiary prevention does not need to rely quite so much on civil society. Another approach are the rather unique programs that only exist in one or several cities but are also based on multi-agency cooperation. Examples are the Aarhus Model in Denmark (which relies on police and municipal actors) or PAIRS in France (which is funded by the government, but with services provided by a civil society organization (Koller 2019, p.5)). Finally, some countries have exit programs led by civil society or governmental actors which cooperate with other relevant actors on a case-by-case basis. In Germany, civil society and government actors work together on individual cases by means of case conferences but otherwise remain separate working entities. Another example from Germany is the model project “Returnee Coordination,” which is based on multi-agency cooperation and deals with foreign fighters and their families returning from Syria and Iraq (p.10). But independently of the specific approach, cases usually also involve statutory bodies like youth and child welfare offices, prison and probation, or public health care institutions that mostly serve the general public.

While the various models of cooperation between different actors and institutions have their specific advantages and disadvantages, they do share some common challenges. It takes time, for example, to share a belief in a common objective and build up trust in the relationship, yet both are essential to establishing an effective cooperation. Also, ethical and data protection issues make information sharing on relevant cases between involved actors difficult. In addition, a public health official can lack relevant knowledge on (de)radicalization processes, or an exit counselor may not be able to diagnose a mental health issue. Still, a growing body of experience from working together on cases suggests that a multi-professional approach can help make tertiary prevention more effective. Some researchers also point out the advantages of multi-agency work because “when such actors manage to come together in trust-based networks, they are able to mobilize a range of useful skills and resources ranging from providing someone with a job to offering religious counselling” (Dalgaard-Nielsen 2016). Relevant governmental authorities and agencies can thus help address clients’ needs in the long-term.

This issue paper presents the results of an InFoEx workshop in the fall of 2020. Practitioners, researchers, and public officials from Belgium, France, Germany, the Netherlands, and Norway came together to discuss their experiences regarding cooperation in tertiary prevention. The first chapter provides insights into the different dimensions of this cooperation including challenges and good practices. The second chapter focuses on the conditions for successfully transferring good practices to other countries and contexts. The final chapter presents three examples of concrete cooperation models in Belgium, the Netherlands, and Germany.

Dimensions of Cooperation in Tertiary Prevention

Cooperation in tertiary prevention can range from discussions on specific cases as required, such as case conferences in Germany, to full-fledged multi-agency workings like those taking place in almost all municipalities in Belgium, at the Safety Houses in the Netherlands, and the Infohouse in Aarhus, Denmark. This chapter provides insights into the various dimensions of cooperation in tertiary prevention of Islamist extremism. Each of the following sub-chapters includes a short description of the main challenges that participants are facing and then lists good practices that they have identified in their daily work to address those challenges and needs.

Clarity of Objectives and Roles

One of the most important challenges according to participants is to build a common understanding of the issues and agree on which problem should be solved. Several participants pointed out that the perspectives of actors could differ (for example between security service and civil society actors) and that this could make effective cooperation difficult. Despite different missions, a common interest must be established. Participants also reported that evaluating prevention work within a multi-agency setting was difficult if no clear objective was agreed. In addition, it was necessary to develop a similar understanding of the case – for example between probation officers and the counseling center – to make tertiary prevention work more sustainable.

Another difficult aspect was to establish a common understanding of what cooperation and partnership entails. One participant argued that while everybody could easily agree to the idea of partnerships, putting this partnership into practice remained difficult. The important question was whether and how to establish long-term multi-agency partnerships (beyond for example cooperation on cases as required). A German researcher with a civil society organization also mentioned that it was challenging to translate an existing good cooperation into reliable and mandatory structures. Another German researcher, also with a civil society organization, asked how cooperation models could provide clients with “control-experiences”, meaning for example experiences of integration and meaning. In her opinion, some of the governmental actors were unable to help as clients were likely to experience a high degree of external control, if governmental actors imposed measures or took decision without the client.

Organizational changes in government agencies could also cause problems. Participants gave the example of a government office which had decided to stop sharing certain information and sending certain representatives to take part in the case discussions. In addition, the high workload in child and youth welfare services meant that case files were piling up and that staff was not given enough time to attend training. Finally, a German ministry official said that the communication on funding between the local and the federal level remained challenging, since no format existed to bring together municipalities and the federal level.

The following good practices were shared to address these challenges:

  • When establishing a new multi-agency structure or ­c­ooperation, consult all stakeholder regarding their needs and expectations so that structures and rules can be established accordingly.
  • Encourage all actors to work toward a clear definition and understanding of different roles: Actors need to be aware that perspectives on radicalization and the goals of an intervention can differ. They should find out “who has which understanding of the objective and who expects what from whom,” as a German researcher with a civil society organization pointed out. Being transparent about goals and responsibilities also helps to establish trust between actors from different institutions, for instance because it becomes clear that a participating official does not represent the entire government but only the specific task of prevention.
  • Develop common working definitions, including for the broad term “cooperation.” For some, it can mean simply talking to other actors, while others understand cooperation as a relationship which includes for example financial expectations.
  • Follow a needs-based approach: A German ministry official said that to draw up the objectives for their work, he and his colleagues ask questions like “What are the needs on the local level? What works well and what does not? How can governments actors such as we help?”
  • Provide additional resources: Several participants pointed to staff discontinuity as a difficulty. Civil society organizations mostly depended on project-oriented funding which led to a relatively high turnover. To avoid friction, tertiary prevention projects should be given a longer-term perspective.

Cooperation and Information Exchange

Exchanging and managing information represents a particularly important challenge. Different actors with different objectives need to find ways to build up trust and to make it attractive for everybody to share information. One example mentioned was that security agencies had a professional interest to maintain secrecy on individual cases, while youth welfare offices were under legal obligation to respect confidentiality but also to report possible security risks. Hence, both actors find it difficult to obtain information from each other. These issues were considered especially relevant if a network was newly established because of the time needed to develop enough personal trust to speak openly.

Also, different levels of cooperation might require different levels of communication. For example, it was mentioned that the Dutch Safety Houses (see p. 15) had a clearly defined responsibility to share information within a Safety House, but that the municipalities struggled to decide what kind of information they should demand from external actors. In this context, a governmental exit counselor considered the cooperation between different levels (local, municipal, federal, and civil society) as the biggest challenge.

Another German participant said the rules for exchanging information between governmental agencies were not clear enough. In this context, the issue of security clearance was also discussed: State actors insisted that security clearance was just an administrative process to allow classified information to be accessed, and that it was not about judging character. Yet according to another participant, a person would only receive security clearance if he or she was credibly committed to Germany’s democratic order. Both government and civil society actors agreed that while a security clearance was the formal prerequisite for exchanging information, it did not necessarily mean that personal trust was established as well.

Finally, having as much information as possible may not always be helpful: One civil society actor said he found it difficult to approach the client without a preconceived picture if he already knew certain things about him or her. Another said that every counselor needed to choose whether to consult the file provided by the police. Another difficulty consisted of adequately recording and managing (personal) data and knowledge. It was mentioned that for example the health care sector had not been sufficiently digitalized.

The following good practices were shared to address these challenges:

  • Develop joint guidelines to regulate cooperation between institutions: In the German state of Hesse, for example, the laws on data protection and the penitentiary permit an exchange of data. Nevertheless, according to one participant, counselors often ask for the client’s consent for passing on information as this is considered important for building trust. Joint guidelines of the Ministry of the Interior, the Justice Ministry, and the Social Affairs Ministry which detail which information can be exchanged, and how, as well as the rights and obligations of the actors involved exist since 2015. In one participant’s opinion, cooperation had continuously improved since the guidelines had been introduced. There was now a high degree of information flow and trust between actors as well as an increased understanding for the different perspectives. Another example was shared by a German practitioner employed at a civil society advice center who works with individuals assessed as “Gefährder,” meaning that they are considered to pose a security risk. He stated that having role clarity made it possible to take different perspectives into account and created synergies. For him, it was important to understand that while security actors needed “objective facts” that are verifiable and could be used in court, there were only “subjective facts” when speaking with a client. Therefore, the advice center’s data protection agreement ruled that no information stemming from the counseling process could be shared unless there was an acute threat. The counselor added that he found case conferences useful that either included the client or were held after he had spoken with the client. In these settings, trust had grown over time.
  • Agree on clear, written rules regarding data management, information exchange, and cooperation with third actors: While written rules may hinder creative solutions, a lack of rules can lead to problematic semi-legal situations, especially regarding data exchange.
  • Consider building structures that facilitate information exchange: One researcher said that the advice center where he was working shared offices with a government agency, making it easier to exchange information between security agencies and the advice center. Both sides held a monthly meeting to discuss specific cases as well as the general situation.
  • Anonymize information on specific cases: To avoid passing on sensitive personal data, actors could opt to discuss cases in anonymized form. However, this might not be possible if the case was already publicly known and highlighted in the media.
  • Tell the client why it is important to share information: A German researcher with a civil society advice center said he found it helpful to clearly communicate to the client what the objective of data protection was and why being able to share information with others could still be very important.
  • Acknowledge the tensions between police and non-police partners and defuse this tension through clear agreements: One participant said that it was important to provide a place where meetings could be held without having security actors take part. At the same time, these should not be branded negatively to make it possible to integrate them easily into the discussion at  a later stage.

Individual versus Structural Cooperation

Workshop participants also discussed the advantages of structural multi-agency cooperation versus individual ad-hoc cooperation. While the first could bring together different perspectives on a daily base, the later could be sufficient to work on a case-by-case basis.

  • Establish cooperation depending on needs: One participant said that he had found it useful to discuss some cases with psychologists or employees of a youth welfare office. Another participant argued that a “system of levels” had worked well, meaning that security agencies were only included at the last level and in case of security issues.
  • Promote good examples and peer-to-peer exchanges: One German government official from the health sector argued that tertiary prevention work should also be discussed between health offices and youth welfare offices. More examples of good practices were needed to show what cooperation could be like. The BAMF, for instance, was funding a so-called Clearing Unit to encourage local advice centers and youth welfare offices to exchange expertise and information. Structures like InFoEx were also mentioned as a good practice as they bring together different actors and encourage them to be curious about working together more closely.
  • Consider separating strategic and operational roles: In established multi-agency structures, it is helpful to have some actors discuss strategic aspects, for example to gain an overview of local trends, while other actors carry out operational tasks. In Belgium, for example, most multi-agency workings were headed by the mayor. In practice, he or she was not involved in the discussion of the cases on an operational, but only on the strategic level (see p.14).
  • Encourage networking and peer-to-peer learning: Several participants mentioned the need to provide opportunities for networking and peer-to-peer learning at the local and municipal level, for example by bringing together youth welfare offices to exchange good practices. Participants said that more local networks were needed in general to identify the relevant cooperation partners in government structures and establish cooperation, for instance regarding returnees.

Trust Issues

Another challenge mentioned was lack of trust  between different actors as well as differences in institutional culture. One participant argued that trust depended on cultural characteristics, for example on perceptions of the police and other state representatives. Cooperation with officials from the penitentiary system might also present special challenges, as one official from a German state ministry of justice explained, adding that it was difficult for civil society actors to gain access to this closed system. A German researcher from a civil society advice center argued that rigid hierarchies and organizational identities could generally impede cooperation. For Germany, institutional challenges between the federal and the state level were also mentioned. The following good practices were shared to address these challenges:

  • Support trust building: Trust is best built by working together. For example, staff of the Dutch Safety Houses have been working together since around 2013. However, the trust building process can also be supported through other means. For example, as one participant explained, one German multi-agency structure asked all external staff to obtain security clearances. It also established detailed contracts which include descriptions of each actor’s needs.
  • Pick a neutral actor or mediator: According to one observation, the VVSG in Belgium was perceived as a trusted actor serving as an umbrella organization for civil administration and police. In Germany, the BAMF Advice Centre was also mentioned as being able to mediate between civil society and security actors; in this context, state actors which do not belong to the security agencies can have a good standing as intermediaries. The Safety Houses in the Netherlands were considered neutral spaces as well.
  • Ensure continuity: In France, some actors in a municipality had been working together on crime prevention for at least ten years before working on radicalization issues; hence they had a clear structure and set of rules, leading to a good partnership. Another good practice was to appoint deputies so that contact with the institution can still be maintained and continuity safeguarded if the person responsible is sick or leaves the job.
  • Strengthen the sense of responsibility: A German government participant said that in his experience, trust and cooperation on the local and subnational levels improve over time, if every actor feels as well as acts responsibly.
  • Provide multi-actor trust building: Several participants agreed that trust was crucial but that establishing it required working together for a longer period. A trust-building exercise or training with several actors could enhance understanding of each other’s needs and challenges, create interest as well as help manage expectations. Frontline practitioners from different institutions, who normally only speak to each other during a crisis, would benefit from a safe space where they can get to know each other without time pressure.
  • Offer job hopping or work shadowing: One governmental participant mentioned that many people working in governmental agencies have been doing their job for a long time. It would be beneficial to bring in staff from other sectors who can add fresh perspectives and improve cooperation between governmental agencies and civil society. One proposal was to establish job hopping formats: Governmental staff would spend a month working at an NGO, while civil society personnel (with security clearance) would experience governmental work for a limited time. Of course, not every institution is suited to such a format.

Cooperation with Religious Organizations

Participants from several countries said that cooperation with religious organizations could be difficult. State or non-state institutions needed to decide how to organize such cooperation and which role to assign it in tertiary prevention. In one German federal state, cooperation with religious organizations was mentioned as one of the most critical issues for governmental actors in tertiary prevention work. In the Netherlands, a governmental participant said no real cooperation had been established yet. One of the reasons was a dispute over whether a mosque was receiving funding from a foreign and non-democratic country. In France, a governmental participant argued that mosque associations were afraid of being blamed for the radicalization of extremists. In addition, it was considered important for Imams to have a background in both Islamic theology and social work to be competent enough to deal with radicalization, but that this profile was difficult to find. At the same time, a social worker stressed that the underlying issue was not about religion.

  • Train actors together: The “Hybrid Training with Religious Community Leaders” (FHAR) at the University of Strasbourg in France was mentioned as a good practice (European Commission 2021b). Since about 2015, FHAR has been bringing together social workers and religious actors to develop new forms of trainings together for two reasons: According to observations, both social and religious actors wish to work with clients at a deeper level. Also, FHAR is based on the working hypothesis that combining socio-educational practices can help prevention efforts.
  • Focus on the local level: Since radicalization happens locally, one participant said, prevention needed to include local civil society and provide links to regional structures.

Challenging Target Groups

Finally, participants reported that additional difficulties can arise when people fear the target group or fail to understand it. According to a German governmental participant, for instance some workers in the healthcare sector were afraid of returnees from the so-called Islamic State in Syria and Iraq (IS). As such fears could result from a lack of knowledge about working with this target group, establishing closer links between counseling actors and healthcare workers could prove useful. Similarly, dealing with extremist offenders was a very sensitive area; information sharing between governmental and civil society actors needed to be effective to enable a timely reaction to any change observed. If relevant information was not shared – which was often due to the lack of a legal basis for information sharing – extremists might be able to play actors off against each other. Also, incomplete information could lead to a wrong assessment of the degree of radicalization and potential security risk. In the case of individuals with multiple problems, the label ‘radicalized’ was not expedient as it might cover other issues and prevent adequate treatment.

  • Establish multi-professional working groups: One participant said initial efforts had been made to set up multi-professional working groups to deal with IS returnees in Germany. It was too early to recommend this as a good practice, but it was clear that this kind of cooperation should be strengthened on both the structural and content level, including, for example, the health care sector, as it had great potential.
  • Build capacity: In general, people working for statutory bodies such as pedagogical institutions, youth and welfare services, and child protection services have a great need to learn more about radicalization and how to deal with vulnerable youth in the context of Islamist and other extremism. For the health sector, for example, radicalization was said to be basically “unchartered territory.” Strengthening intercultural competence and addressing Islamophobia and racism were mentioned as topics to address.
  • Reach out to target groups: Participants also stressed the need to make potential clients aware of the existing offers in tertiary prevention. At the same time, it was mentioned that people working in social work, schools and the health sector already had a heavy workload. Additional incentives were needed to encourage them to want to reach out to target groups and get involved in prevention.

Research and Evaluation

Participants agreed that tertiary prevention benefits from actors bringing multiple perspectives to the table. Yet their needs must be properly analyzed to be able to develop effective formats. The following recommendations were shared:

  • Strengthen the evidence base: One participant said more evidence was needed on mental health issues, for example regarding relevant pathologies. It would be desirable for actors in (forensic) psychiatry to generate a better data basis. Other research needs related to the development of children in Salafist family structures.
  • Establish adequate data management: Participants discussed the need to establish better ways to document cases and share information not only to make cooperation more effective but also for evaluation purposes. Multi-agency cooperation should be evaluated to further develop structures.
  • Establish a “culture of error”: One civil society participant stressed the need to speak openly about mistakes. Reports often focus too much on success in order to generate continued funding. A culture that encourages people to deal with mistakes and learn from them – for example what to do when something goes wrong – produces much better results. 
  • Reflect on your experiences: One participant said she had not taken notes during meetings because the stress level was too high. It was only later that she realized notes would have helped her to reflect on her experiences and review her own expectations.

How to Transfer Good Practices

Participants of different nationalities and institutional backgrounds stressed the general value of looking at other countries to learn from their experiences regarding the cooperation between institutions, especially with statutory bodies. Yet transferring good practices to another context can be difficult, particularly if the institutional setup or the legislation are different. One governmental participant stressed the important of travelling and observing several approaches before carefully adapting a good practice to another local context. Recent research in the Nordic countries also suggests that there need to be a lot of similarities for a transfer of good practices to be successful.

“Copy and Paste” best practices? Requirements for successful multi-agency collaboration against violent extremism

by Tore Bjørgo, Director of Center for Research on Extremism (C-REX), Professor at the University of Oslo and the Norwegian Police University College, leader of the HEX-NA project.

Multi-agency collaboration in the field of preventing violent extremism involves at least two dimensions: 1) coordination between agencies on action plans and interventions at national and local levels, and at strategic and operational levels, and 2) information sharing between agencies and practitioners on problematic issues and challenges, and on individuals at risk.

To what extent can a successful practice of multi-agency collaboration be copied and implemented in a different setting? Or, to put it differently, which conditions need to be similar to make the transfer of a good practice successful? These are among the issues we are exploring in an ongoing research project titled “Nordic Multi-Agency Approaches to Handling Extremism: Policies, Perceptions and Practices (HEX-NA).” The overall project objectives are to map comparative differences in Nordic institutions and their approaches to countering violent extremism, and explore prerequisites for successful interagency collaboration, with three sub-projects: Policies (a cross-country comparison of legal frameworks and institutional setups); Perceptions (cross-country and cross-professional group comparison of stakeholder and public perceptions, exploring trust and willingness to share sensitive information), as well as Practices (cross-country and cross-city comparison of practices and implementation). So far, only the sub-project on Policies is finished and has published its report (Sivenbring & Malmros 2019).

Some main findings include that

  • Policies in the Nordic countries are more different than expected.
  • Interagency collaboration is far more developed, extensive, and institutionalized in Denmark and Norway than in Finland – and even less so in Sweden.
  • Police are less involved in interagency collaboration in Sweden than in the other Nordic countries.
  • What can explain these differences? The legal frameworks that facilitate interagency collaboration and information sharing differ considerably between the Nordic countries. Denmark has legislation that permits (and regulates) sharing of sensitive personal information for crime preventive purposes between agencies. Confidentiality regulations are more restrictive in the other Nordic countries. Police laws in Denmark and Norway have a much stronger emphasis on prevention and on collaboration with other agencies, inducing the police to be more proactive and preventative than is the case in Sweden and Finland, where the style of policing is more reactive.

There are also differences in organizational models for interagency collaboration. The SSP model for collaboration between school, social services, and the police, originally developed in Denmark from the mid-1970s, has been implemented in rather different ways and degrees in the Nordic countries. Denmark has implemented the SSP model in all municipalities, and it is also at the core of the so-called Infohouse model that was later established to handle violent extremism. Norway’s variety of the model (SLT) is implemented in about half of the municipalities. Finland has recently adopted a similar model (Anchor). The police are a lead actor in interagency collaboration in Denmark and play major roles in Norway and Finland. However, the police are less involved in such collaboration in Sweden because it is assumed that trust in teachers and social workers will be undermined if they share information or collaborate with the police. Consequently, the SSP approach is least developed in Sweden. Different professional cultures are also important. This graph displays different logics on interagency collaboration in the Nordic countries (Sivenbring and Malmros, 2019):

Good practices of interagency collaboration thus require some basic conditions to be fulfilled:

  • Legislation facilitating partnership and information sharing
  • A high level of trust in public institutions among the population and relevant minorities
  • A high level of trust between practitioners in relevant agencies

These findings as well as similar experiences shared by other participants led to a discussion about the preconditions for a successful transfer of good practices. The question was debated whether it is necessary to identify similar conditions or ‘match-making’ criteria. Some participants said that it might be helpful to consider which countries and areas actually have similar conditions.

Multi agency cooperation in practice

Addressing these challenges and implementing some of the good practices in the respective local context has the potential to contribute to a better cooperation with statutory bodies and lead to a more effective and sustainable tertiary prevention work. The following three examples show how cooperation with different governmental structures can be put into practice.

VVSG: a trusted umbrella organization in Belgium

The EU-funded project EMMA has the objective to develop common standards for multi-agency workings in Germany, the Netherlands, and Belgium and to establish national peer-to-peer networks. In Belgium, it is led by VVSG, an umbrella organization that includes the police (in Belgium, the mayor always holds the position of local chief of police as well) and acts as a bridge and mediator between political, police, and civil society actors. As local authorities in all regions und municipalities in Belgium have been obliged by federal law since 2018 to establish multi-agency workings, central coordination units for (tertiary) prevention of extremism exist on the local level (the so-called LIVC-R units). These units deal with cases of prevention at an early stage, and the police will only be involved when dealing with potentially violent extremists. The project includes a process evaluation carried out by the University of Geneva with the objective of developing a self-evaluation tool that includes a self-learning dimension and can adapt to different practices.

What the Belgian experience shows, according to a participant, is that multi-agency workings need a strong mandate to work effectively. They should be placed at the center of prevention work and have a clear legal basis regarding their role and objectives. For example, the action plans of each city in Belgium were publicly accessible. Another advantage of the Belgian system was that the multi-agency workings are communal entities that are legally regulated and thus sustainable. Yet some challenges remained: The role of the coordinator had not been clearly defined and included too many responsibilities.

NeDiS: a network in the penitentiary system in Germany

The administrative unit NeDiS (Network for the Deradicalisation in the penitentiary system) was established in 2016 and is led by the Ministry of Justice in the German state of Hesse. So far, it is the only comparable entity in Germany (DPT n.d.). The NeDiS concept includes four pillars: identification, prevention, deradicalization, and coordination. Different competences are pooled, including Islamic science, psychology, law, and administration. The objective is to prevent extremist proselytization in the penitentiary institutions in the German state of Hesse and support disengagement and deradicalization. This includes helping clients take up or continue an education, debt or drug counselling, and teaching democratic values and media competence. The actual exit work has been outsourced to the civil society organization Violence Prevention Network (VPN), since it is better equipped to deal with radicalization and its staff finds it easier to gain the clients’ trust than do employees at penitentiary institutions. VPN also works closely with security agencies and has access to the Information and Competence Centre against Extremism, located in the Hesse Ministry of the Interior. VPN staff have an employment contract which includes provisions for data protection and requires them to be checked for reliability and security, which creates a basis of trust. VPN and the security agencies share information about how cases develop in reports and case conferences. Clients who are likely to need monitoring after release are discussed regularly, including at a meeting scheduled three months before they leave prison.

The Safety House: a holistic approach in the Netherlands

As part of a larger holistic multi-agency approach to the prevention of extremism in the Netherlands, the so-called “Safety Houses” are neutral spaces located in different municipalities and funded by the Dutch government. When a case of possible radicalization is reported to the municipality or the police and the concerns are found to be valid, the case is referred to the local Safety House. Depending on the case and the required expertise, a decision is taken on who will be involved in the case. Teachers, probation or police officers, mental health personnel, or youth workers will then work together on a case and develop a plan on how to help the person concerned. Interventions can include support with his or her financial situation, family status, mental health, or ideology. The Safety House also keeps track of relevant data, such as basic personal information, and tracks the intervention. Every six weeks, representatives of the municipalities involved meet to exchange information and discuss new laws and trends, and share their experiences of what works and what does not.

Once a case is taken up by the Safety House, the person concerned receives a letter with an explanation and is told that he or she has the right to consult his or her file. When the person takes up this offer, the Safety House discusses the case with him or her and obtains more information that, according to internal rules, can be shared within the Safety House. Overall, however, only limited data can be collected as the specific data protection legislation applicable to this work has not been passed yet. Obtaining information from other actors is difficult, too. The municipalities are unsure of how much they can share with other institutions, and police are restricted to what has been agreed in a confidentiality agreement with the Safety House. However, as one participant put it: “We know that if we really need information, we will get it.”


In tertiary prevention of Islamist extremism and terrorism, effective case management does not only require specific civil society and government exit programs and security services. A range of additional agencies and actors from the social, educational, justice, and public health sector play an important role in disengagement and deradicalization work as well.

At the InFoEx workshop, practitioners, researchers, and government officials from several European countries stressed the importance of a holistic approach to tertiary prevention. Nevertheless, multi-agency cooperation – whether ad-hoc or in multi-agency settings like the Danish Infohouse – in their experience needs to overcome significant hurdles. To address these challenges, it helps to accept existing differences. Actors need to work toward a clear definition and understanding of different roles and objectives, for example by drawing up working definitions together. Similarly, good experiences have been made with developing common guidelines to regulate cooperation and information exchange. Another option is to consider establishing multi-agency structures instead of ad-hoc cooperation. Existing and well-functioning structures and cooperation should be promoted as good examples; at the same time, it is important to create formats for regular peer-to-peer exchange. Cooperation on specific cases should focus on involving local actors, but there should also be a structured exchange on trends and strategic decision between the involved actors, separating operational and strategic aspects. Capacity building for governmental actors is necessary, and it should include not only specialized training but also trust building exercises. Finally, all cooperation should include elements of monitoring and evaluation.

Strengthening the role of statutory bodies in multi-agency cooperation can significantly improve tertiary prevention efforts. They add valuable expertise regarding access to housing, education, (mental) health, and social (re)integration. Diverse push and pull factors contribute to radicalization processes; similarly, efforts to support deradicalization and disengagement are made easier if a variety of important needs can be addressed. By acknowledging that (de)radicalization is not only about security, such an approach also avoids stigma and leverages existing resources in governmental institutions. Through such efforts, tertiary prevention ultimately becomes more effective and sustainable.


InFoEx is a joint project of the German Council on Foreign Relations (DGAP) and the Research Center for Migration, Integration and Asylum of the Federal Office for Migration and Refugees (BAMF). InFoEx brings together good practices and scientific findings from tertiary prevention of Islamist extremism in Germany and abroad. The aim of the project is to compile empirical findings on radicalization and deradicalization processes with a focus on their practical applicability for prevention work. To this end, the BAMF Research Centre initiated a network of research fellows who are embedded at local advice centers and research institutions, partnering with the BAMF Advice Centre on Radicalisation. Together with counselors working in these local advice centers, these research fellows represent the core members of InFoEx.


The seventh InFoEx workshop in the fall of 2020 focused on the cooperation between civil society and governmental actors in tertiary prevention of Islamist extremism. The ongoing restrictions imposed because of COVID-19 made it necessary to organize the workshop in a virtual format. Among the around 30 participants were network partners of the BAMF Advice Centre on Radicalisation from civil society and government institutions as well as practitioners and academics from Belgium, France, Germany, the Netherlands, and Norway. To align the workshop with the needs of its stakeholders, research fellows embedded at local advice centers and at research institutions in Germany shared – in agreement with practitioners at their local advice centers – specific information needs and questions prior to the workshop. External experts were invited accordingly. At the workshop, the participants were able to share their experiences and views of the challenges and good practices of the cooperation between civil society and governmental actors.

Bibliographic data

Koller, Sofia. “Cooperation in Tertiary Prevention of Islamist Extremism.” German Council on Foreign Relations. September 2021.

DGAP Report No. 15, September 15, 2021, 20 pp.

Themen & Regionen