Intersectional collaboration and the development of prevention infrastructures: a qualitative study

There is increasing attention to (disease) prevention within healthcare. Many countries have presented their national prevention programme in the last few years (Rijksoverheid, 2018; Bensberg et al., 2021; United Kingdom Government, 2021; Matheson et al., 2020, Naqvi, 2023). Prevention is defined as measures taken by or applied to persons who are not currently affected by a disease, intended to decrease the risk that disease will afflict them in the future. Prevention is traditionally classified into three levels: primary, secondary and tertiary (Gordon, 1983). Investing in prevention could be of great benefit to public health for several reasons. At first, people who are in good health are better able to lead fulfilling and productive lives and contribute to society (Bodenheimer and Sinsky, 2014). Secondly, implementing preventive programmes would increase the average quality of life (QoL) of the people participating. In turn, an improved QoL leads to a lower demand for care and thus lower pressure on our healthcare systems and healthcare providers’ workloads (Fitzpatrick, 2019). Thirdly, preventive interventions for high-risk groups can reduce the use of expensive medication or treatments and could therefore lead to economic benefits (Boom, 2009; Van Gils et al., 2020). These factors correspond to the goals of the quadruple aim of increasing population health, improving patient and provider experiences and lowering cost growth (Bodenheimer and Sinsky, 2014).

Despite potential benefits and the general consensus that policy strategies should focus more on prevention, there is a gap between policy and desired outcomes (Cairney and St Denny, 2020). One of the intended outcomes is the implementation of a complete range of preventive interventions for high-risk groups, because this is currently not conclusive. High-risk groups are people who are not sick yet but are at risk of getting sick because of their biological characteristics, lifestyle or life events (Gordon, 1983). For example, people with overweight, depressive symptoms, low social-economic status or people who experience loneliness or a sedentary lifestyle. Developing and implementing preventive programmes for high-risk groups is often complicated due to a lack of clarity about financing, few financial incentives, the different responsibilities of the stakeholders involved and different organisational cultures (Heijink and Struijs, 2016; Bekker and Wagemakers, 2021; Alders and Schut, 2019). For high-risk groups, selective and indicated prevention are important. Selective preventive interventions target high-risk populations and aim to prevent individuals from actually getting sick (Greenberg and Abenavoli, 2016). While indicated preventive interventions target individual people with early symptoms and aim to prevent complaints from worsening into a disorder (Gordon, 1983; Arango et al., 2018). Selective and indicated prevention partly correspond with primary and secondary prevention but divide prevention based on target group instead of disease phase (Edelman and Kudzma, 2022; Hawkins et al., 2015).

In the Netherlands, the government, municipalities, professional organisations and health insurance companies are collectively responsible for organising preventive interventions (Jeurissen and Maarse, 2021). Parties operate in a decentralized healthcare system of regulated competition (Kroneman et al., 2016). Financing collective (selective) prevention legally belongs to public responsibility (government and municipalities) and financing individual (indicated) prevention belongs to health insurance companies (Jeurissen and Maarse, 2021). However, the dividing line between selective and indicated prevention is less clear, causing ambiguity in governance regarding responsibilities and financing (Kroneman et al., 2016). Similar problems can be seen in other countries with comparable healthcare systems (Heijink and Struijs, 2016).

In order to overcome the pitfalls of organising prevention for high-risk groups, involved stakeholders need to work together in a structured, sustainable way. In recent years, several policy interventions have been introduced in the Netherlands to stimulate intersectoral collaboration and the transition from occasional to structural collaboration in the healthcare system (Boer, 2021; Ministry of Health, Welfare and Sport, 2016). This intersectoral collaboration forms the basis for the intended prevention infrastructure. According to the WHO, intersectoral collaboration for health entails cooperation between several sectors, for example, in developing and implementing public policies intended to improve health, equity and well-being. Intersectoral collaboration makes it possible to take action that is more effective or efficient than taking action by one of the sectors alone (WHO, 2023). From an international perspective, there is growing interest in the use of intersectoral collaboration (e.g. alliances, coalitions, partnerships and joined-up governments) as a shared health governance structure to address complex, health-related issues in local communities (Hearld et al., 2019; Perri 6, 2004). Collaboration across organisations, disciplines and sectors is considered a necessary element of dealing with today’s healthcare system challenges and is increasingly addressed in healthcare policy (Dahlgreen, 1995; Bteich et al., 2017). However, collaborating is complex and requires behavioural changes, as parties have to look beyond their own sectoral borders (Karam et al., 2018; Alderwick et al., 2021; Van der Vliet et al., 2022). Having a supportive architecture (or structure) is found to be important in creating integration and collaboration (Carey and Crammond, 2015). Despite the growing attention for intersectoral collaboration in recent literature (Mondal et al., 2021), the everyday practices of intersectoral collaboration for public health policy remain largely unexamined (Chircop et al., 2015). Few longitudinal studies have been conducted to understand how strategies, attitudes and perceptions about collaborating in a network change over time (Bryson et al., 2015; Hearld et al., 2019). This results in a lack of understanding of the long-term effects of policy interventions in public health (Van Vooren et al., 2023).

The lack of a complete range of preventive interventions for high-risk groups has several causes, as explained above. Part of the solution may lie in collaboration between different responsible parties. Policy has attempted to stimulate this in recent years. The aim of this study was two-fold: First, to describe the transition towards a sustainable prevention infrastructure. And second, to gain insight into barriers and facilitators associated with intersectoral collaboration regarding organising prevention for high-risk groups.

This study was conducted in the Netherlands. Since 2016, regional collaborative networks have been encouraged and facilitated by the national government through a “Prevention Coalition grant” (Preventie Coalitie subsidie) for setting up a collaborative structure. National policy aimed to encourage intersectoral collaboration targeted at prevention for high-risk groups (Boer, 2021; Ministry of Health, Welfare and Sport, 2016). In total, the Dutch Ministry of Health, Welfare and Sports approved 20 prevention coalitions over the programme’s duration (2016–2022) (Kemper et al., 2022). Involvement of financing parties (municipality and health insurer) was a requirement for receiving the grant.

This longitudinal qualitative study was conducted over a period of five years (2016–2021), following 11 collaborative networks that focused on organising prevention for high-risk groups and received the prevention coalition grant. In this study, a longitudinal qualitative research (LQR) design is applied because this method allows following change over time. LQR can be used for follow-up studies and evaluation studies (Holland, 2007; Calman et al., 2013). In order to learn more about the complicated transition towards a sustainable prevention infrastructure and intersectoral collaboration, changes over time need to be taken into account, which is why LQR is appropriate. Reviewing the interview data and modifying interview guides between data collection points are characteristics of LQR. The research team designed the interviews to build on rather than duplicate the previously collected interview data (Tuthill et al., 2020). Interview topics concerned: network characteristics, designing collaboration, experiences with collaboration, developments in their regional prevention infrastructure and preventive interventions.

Several theoretical frameworks have been developed and used to gain a better understanding of the processes (behaviours) that play a role in collaboration between multiple sectors, such as the Framework for Collaborative Action (Fawcett et al., 2010), the CAHN framework (Steenkamer, 2020), the Bergen model (Corbin and Mittelmark, 2008) and the HALL framework (Koelen et al., 2012). One framework used previously developed behaviour models and simplified them into one coherent model, the COM-B model. The COM-B model proposes three factors that influence and interact with behaviour: capability (C), opportunity (O), motivation (M) and behaviour (B). Capability refers to an individual’s psychological and physical ability to participate in an activity. Opportunity refers to external factors that make a behaviour possible. Lastly, motivation refers to the conscious and unconscious cognitive processes that direct and inspire behaviour (Michie et al., 2011). Compared to other theoretical frameworks, the COM-B focuses on understanding the people involved in collaboration rather than focusing only on the system in which they operate, making it a suitable model to study intersectoral collaboration (Michie et al., 2011). The COM-B model recognizes that behavioural change does not occur in a vacuum, yet factors interact and mutually affect each other. The components of COM-B can be construed at any level (e.g. individual and group) (Hendriks et al., 2015). It provides a clear structure and simple classification for reflecting on barriers and facilitators. Furthermore, it was successfully applied in similar studies about intersectoral collaboration (Hendriks et al., 2015; Van der Vliet et al., 2022; Wakida et al., 2018; Gomez-Rossi et al., 2022).

Since 2016, each year, the number of regional collaborative networks with a “Prevention Coalition grant” has increased. Using purposive sampling, new coalitions were approached annually to participate in an interview. Based on distribution across the country, size of the region and variation in involved health insurers. Over the years, data saturation was reached with the following 11 networks in total: Interviews were held with representatives of municipalities, health insurers, municipality health services and programme managers involved with the network. Networks were followed over time and therefore, approached annually for an interview (re-interviews). Two to six interviews per network were conducted during a five-year study period. In addition to the networks distributed over the years, seven people involved with providing(?) a specific preventive interventions, four national stakeholders (e.g. the umbrella organisations of health insurers and municipalities), six healthcare insurers and 13 municipalities were interviewed about their experiences with intersectoral collaboration in general. Interviews were repeated every year around September–December (Figure 1). In total, 86 semi-structured interviews were held online by telephone or face-to-face. Interviews were audio recorded, and the length of the interviews varied between 30 and 90 min.

Interviews were transcribed verbatim. Except for the first two data collections in 2017 and 2018, summaries were used for those interviews. During the study, in total, nine researchers were responsible for conducting and analysing the interviews. The analysis consisted of two steps. Transcripts from every interview were coded by two researchers using MAXQDA (versions 18, 20 and 22). Any disagreements between the two researchers were discussed to ensure consensus.

Step one included thematically analysing the data after each year of data collection (cross-sectional). A code tree was drawn up based on the interview guide. Deriving from what the researchers found in the transcripts, the code tree was supplemented with new codes. Thematic analysis enabled us to find themes and patterns within and across the data set of interviews per year and was mainly inductive. Step two included a longitudinal analysis, which focused on differences and similarities between the yearly evaluations. Data from all the consecutive years were included and analysed together through a deductive approach (Figure 1). The COM-B model was used in this step to identify and categorize facilitators and barriers associated with intersectoral collaboration. Longitudinal results were produced using an iterative process to summarize, analyse and consolidate themes. The transition towards a sustainable prevention infrastructure was analysed by listing objective events and pulling them apart from the experiences.

This study was approved by the Ethics Review Board (ERB) of Tilburg University according to the Medical Research Act (reference number EC-2017.97). The interview respondents were informed of the study’s purpose, mode of participation and confidentiality. Respondents gave written or oral informed consent prior to the interview.

The first part of the results section describes the transition towards a sustainable prevention infrastructure in the collaborative networks and the changes over time based on the yearly interviews. The second part describes barriers and facilitators based on stakeholders' experiences in setting up and maintaining intersectoral collaboration.

At the start of the study, few collaborative networks in the field of prevention for high-risk groups existed. The interviewees recognized its importance, but real collaboration was still in its infancy.

Networks started with identifying local health issues and choosing themes and target groups accordingly. Networks also inventoried existing interventions in the region and worked towards a plan of action or cooperation agenda. In the regions where a Prevention Coalition network emerged, the municipality and health insurer often already had a good collaborative relationship, according to the interviewees, which often prompted them to apply for the grant and further develop their plans. In the initial phase of these new networks, the intensity of collaboration between municipalities and health insurers varied greatly.

On a national level, according to interviewees, collaboration was further stimulated by implementing preventive interventions, a regional working structure, national conventions and an ambassador. In 2019, the implementation of four different preventive/lifestyle interventions was stimulated by the national government, namely: combined lifestyle intervention (CLI), social prescribing (SP), falls prevention and diabetes prevention. Municipalities and health insurers were obligated to cooperate regarding the implementation. Many collaborative networks indicated that they embraced these interventions. The CLI was the only one included in the Health Insurance Act.

Exchanging knowledge about successful interventions, experiences and best practices took place through national conventions. In addition, a national ambassador was appointed in 2019 by the Ministry in order to advice potential regional collaborative networks based on experiences in other networks. A regional working structure of health insurers and municipalities was formulated by its umbrella organisations in early 2020. In this structure, regions consisting of multiple municipalities were designated to cooperate with a particular health insurer, primarily the largest health insurer in the region. If the working structure matched the existing way of working, it was found to be a stimulus. If the structure did not fit well with the health insurer’s self-established structure, it was perceived as counterproductive, according to the interviewees. The obligatory regional working structure stimulated cooperation because municipalities and health insurers were, for the first time, jointly responsible for making agreements about prevention for individuals at risk.

A rising trend can be seen over the years in which municipalities and health insurers were increasingly able to find each other, according to the interviewees. They made agreements on organising prevention for high-risk groups through regional cooperation agendas and arrangements regarding preventive/lifestyle interventions, such as CLI and SP. The number of prevention coalition networks has increased considerably over the last five years as well as their list of cooperation partners. Respondents mentioned that more parties joined the network, also outside the healthcare sector (such as housing associations, (special) education and employers). When the grant period reached its end, the prevention coalition networks tried to sustain what they had achieved.

This section elaborates on promoting and impeding factors divided into the three COM-B themes: capability, opportunity and motivation. In Table 1, an overview of the facilitators and barriers is presented.

Several factors influenced one’s capability to collaborate around prevention for high-risk groups. The sheer basis of working together is getting to know and, more importantly, understand each other’s language. This basis is of great importance as a starting point in formulating a shared vision or ambition. In 2016, parties experienced major differences in culture and language between different types of organisations on a local level. This made working together complex. It was found conducive when, within networks, knowledge about successful interventions, experiences and best practices was exchanged. A major threat to the continuity of the collaboration was a change in the people involved, for instance, the change of councillors after municipal elections. Multi-year agreements exceeding political terms were found to be a strategic solution to this threat.

Another solution was appointing regional coordinators within municipalities and health insurance companies (part 1) who ensured that, despite personnel changes, parties continued to work together and learned to know each other. On this basis, mutual understanding about responsibilities and interests grew and tasks were better allocated. As a result, parties became increasingly familiar with each other and the relationship became stronger.

Intersectoral collaboration became increasingly common over time; a number of factors related to “opportunity” are highlighted below. The Prevention Coalition grant facilitated the establishment of a collaborative network in which parties from the medical and social sectors met regularly and could make agreements about collaboration and organising prevention for individuals at risk. However, getting and keeping all parties involved in the network was found to be labour-intensive, especially as the network grew over time.

General practitioners were recognised as important collaboration partners in those networks because they could refer people to preventive interventions. However, in practice, they were not always well engaged. Building, maintaining, expanding and sustaining collaboration requires time to invest in these relationships. Interviewees indicated that structural financing was needed to enable collaboration across sectors and make long-term commitments. After all, grants were only temporary and often fragmented per high-risk group. According to the interviewees, health insurers and municipalities had limited financial possibilities or incentives to invest in collaboration with regard to prevention, and municipal cutbacks put the budget for prevention even more at risk. Lack of manpower and an insufficient budget at municipalities to organise cooperation were major barriers. Most interviewees were, over the years, increasingly convinced that legal adjustments were necessary to compel and enable involved parties to implement preventive interventions.

There were several factors that influenced the motivation to set up an infrastructure around prevention for high-risk groups. Firstly, there had to be a shared sense of urgency. Over the years, interviewees increasingly indicated that prevention was essential in order to keep healthcare sustainable and affordable. The parties realized that they needed each other and should make agreements about prevention in order to collectively achieve goals and keep healthcare accessible.

With regard to making agreements, it was found helpful if trust between parties had grown. On an organisational level, the sense of urgency was affected by the priorities of network parties, defined by their tasks and obligations. Secondly, motivation was negatively affected by the long-term effects of prevention on health and healthcare costs and the uncertainty of whether these effects would actually be realized. Network parties wanted and needed more insight into health outcomes to stay motivated and be able to make informed decisions. However, overall, motivation to collaborate grew among the various parties.

This study followed collaborative networks over time to gain insight into experiences (perceived barriers and facilitators) with intersectoral collaboration regarding organising prevention for high-risk groups. Furthermore, it aimed to describe the transition towards a sustainable prevention infrastructure in the local healthcare system. The results suggest that the transition towards a sustainable prevention infrastructure has started. Policy has stimulated intersectoral collaboration in the last few years. Organisations from the health and social sectors were found to be genuinely motivated and willing to work together as a result of a shared sense of urgency. Collective coordination of preventive interventions is becoming increasingly common. However, many barriers are still present and ask for new policy interventions on responsibilities and financing. For a complete overview of the facilitators and barriers, we refer to Table 1. In the following sections, findings will be discussed and compared to earlier research.

The creation of regional prevention coalitions has been encouraged by the Dutch government in the past few years. Over time, networks grew, collaborated more, and implemented intersectoral interventions that best suited their region and residents. The obligatory regional working structure stimulated cooperation because municipalities and health insurers were, for the first time, jointly responsible for making agreements about prevention for individuals at risk. Previous research also shows that assigning increasing stakeholder responsibilities is essential for a sustainable prevention system. For example, in Denmark, municipalities are responsible for ensuring that local communities provide healthy environments and activities and promote well-being and prevent disease (Schmidt et al., 2019). Likewise, in Germany, health insurers are explicitly given the (financial) responsibility to invest in prevention in settings such as neighbourhoods and schools (Hildebrandt et al., 2012). Future research should further examine in what way collaboration is likely to succeed in a prevention infrastructure. However, finding ways for organisations that are organised differently to work together is eternal and universal, according to Perri 6. Differences between and within countries are not random or necessarily based on culture, but rather on fundamental organisational limitations and on conflicts and competitions with each other (Perri 6, 2004).

When it comes to organising and financing prevention, primacy lies with politics. The extent and manner in which attention is paid to prevention, especially for high-risk groups, is strongly related to the prevailing political view of citizens' own responsibility (RVS, 2023). In recent years, the challenges regarding prevention have been considered increasingly urgent in the Netherlands. There seems to be a small political shift towards more “social” policies; is the need for prevention finally being recognised? However, despite increased focus on prevention and numerous prevention policies, spending is not yet increasing (Van Gils et al., 2020).

Previous research on intersectoral collaboration found similar barriers and facilitators related to capability, as shown in the present study. Having a shared vision of the problem to be addressed was also an important success factor in health promotion activities, according to Van Dale et al., (2020). Others found that a new (long-term) vision helped collaborative partnerships (Steenkamer et al., 2020; Van der Vliet et al., 2022; Buist et al., 2019). According to Alderwick et al., differences in decision-making processes were an important barrier concerning capability. In the present study, this barrier is comparable to differences in organisational culture. Furthermore, a lack of accountability between organisations was found to hinder collaboration (Alderwick et al., 2021). In the present study, this was described as a lack of clarity about responsibilities.

According to the interviewees, an important facilitator related to opportunity was “receiving finance for setting up a collaborative structure”. Arranging additional funding for collaboration is often recognised as beneficial in the literature (Steenkamer et al., 2020; Van der Vliet et al., 2022). Others argue that partnerships would be too busy collecting, distributing and accounting for those funds or grants (Eisenstein and Chang, 2017). Furthermore, according to Siegel et al., the overwhelming reliance on short-term partnership financing is not aligned with the task of long-term, multisector integration and therefore, sustainable change in the healthcare system (Siegel et al., 2018). This raises the question of whether networks survive after the end of the grant period. Another important factor concerns building relationships and being committed, especially if parties do not have a history of collaborating (Steenkamer et al., 2020). Cooperating partners from different sectors each come from a unique culture of thinking and communication, and they may fail to find common goals and benefits or lack of commitment (WHO, 2018; Storm et al., 2016).

On the motivation part, we found that enthusiasm from parties involved, as expected, contributed to collaboration. Having common goals, creating a win-win situation for the partners in the collaboration and recognising shared problems lead to higher motivation. High motivation is also an important theme in other studies (Van der Vliet et al., 2022; Mondal et al., 2021; Corbin et al., 2018). Mutual trust is the foundation for collaboration and can grow over time. The importance of communication, trust and good relationships between collaborating parties is frequently cited in the literature (Mondal et al., 2021; Corbin et al., 2018). Provisionally granting trust and respect for first steps generates momentum to develop more trust and respect ahead (Kerrissey and Singer, 2023). Another factor that influenced organisations’ motivation to join the network was a growing societal movement about a broad view of health (De Bekker et al., 2023). This movement increased parties’ willingness to engage in broader partnerships and jointly focus more on prevention. This led to more parties participating outside the healthcare sector (such as housing associations, (special) education and employers).

Overall, in this study, many facilitating factors for intersectoral collaboration within prevention were found, especially in the capability and motivation component of the COM-B model. Stakeholders are genuinely motivated and willing to work together and have the necessary knowledge. However, all three COM-B themes are required in order to change behaviour and establish the actual execution of that desired behaviour. Barriers concerning financing (opportunity) are currently holding back intersectoral collaboration and should be recognized as a conditional factor. To bridge this gap, policymakers need to adopt new financing methods that create an enabling environment for multisector partnerships (Siegel et al., 2018). Recent developments in Dutch prevention policy concern a specific payment for prevention, which is part of the national Healthy and Active Living Agreement (Gezond en Actief Leven Akkoord) (GALA, 2023). The first insight about this regulation is that it meets the need for clarity about responsibilities and the need for allocated resources. Although a certain tension between the two financing parties (municipalities and health insurers) will probably remain due to their legal responsibilities (Van Duijn et al., 2022). Future research has to show whether it has an effect on the offer, availability and usage of preventive interventions. Based on this study, it is considered necessary to make the recommendations below.

Recommendations for a sustainable prevention infrastructure:

1. Establish responsibilities around prevention for high-risk groups, for instance, through legislation;

2. Create structural financing to support and sustain intersectoral networks and

3. Create economic incentives for financing bodies to implement effective disease prevention and health promotion programmes.

Through LQR, many different networks and stakeholders were interviewed over the course of five years. Because of the design of the study, the reliability of the results is high. The shift to a sustainable prevention infrastructure is a long-term transition; a longitudinal design fits the bill. LQR may contribute to increased awareness among researchers about the importance of time (Carter, 2006; Calman et al., 2013). In this study, it was observed that the Prevention Coalition grant provided an initial push to work together. Next, the networks ran into all sorts of obstacles when working together. Those obstacles emerged over time, becoming more and more specific. This allowed the researchers to make more targeted recommendations for policy (Kemper et al., 2022). A disadvantage of a long follow-up time is the high probability of staff changes. Networks in this study had to deal with frequent changes of contact persons, which caused knowledge to be lost as well as the cooperation relationship that had been built up. Usually, an LQR benefits from the development of a history between the researchers and the participants (Carter, 2006); this was only the case to a limited extent in this study.

The transition to a sustainable prevention infrastructure has been set in motion. The longitudinal nature of this study allowed us to look at changes over time. Because of the element “time”, trust grows, parties get to know each other better and the challenges regarding prevention are considered more urgent. This study shows that collaboration regarding prevention is becoming increasingly common. The relationship between key stakeholders, like municipalities and health insurers, is generally stronger than five years ago. The working structure ensured more regional and local agreements. Knowing the factors that facilitate or hinder intersectoral collaboration can help stakeholders improve or develop their collaboration infrastructure. Many facilitators were found, but eventually, a prevention infrastructure will not be sustainable if structural financing and clarity about (legal) responsibilities are not arranged.

We wish to thank all stakeholders for their input and involvement, especially the people who participated in the interviews for sharing their time and experience. Additionally, we appreciate all who have assisted in collecting the data and data transcription. Specifically, we would like to thank Natascha van Vooren for providing valuable feedback on the manuscript.