Regulated markets and rationalised myths: an institutional perspective on value-based purchasing in the Netherlands

Abstract

In the Dutch health care system of regulated competition, health insurers are assigned the crucial role of prudent purchasers and expected to critically contract providers based on the quality and prices of their services. Thus far, however, these organisations have struggled to fulfil this role. This study sheds new light on the purchasing behaviour of Dutch health insurers. We examine how insurers perceive the context in which the value-based purchasing of hospital care should take shape, and we draw on insights from institutional theory to frame our analysis. Our findings are based on a series of semi-structured interviews (n = 18) with employees and representatives of several insurer companies whose combined market shares add up to over 90 per cent of all premium payers. Our analysis highlights an environment in which market mechanisms are tangled up with historically rooted budgeting practices, where insurers are pressured to sustain rather than critique hospitals, and where self-regulating medical professionals are firmly supported by society’s deep-seated belief in the quality of their services. Like many other organisations, Dutch health insurers tend to conform to their institutional environment. While this conformity may aid them in organisational stability and survival, it also restricts their ability to purchase prudently.

1. Introduction

Starting in the late 1980s, the health care system of the Netherlands has gradually moved from government control of the supply side towards a system of regulated competition (Schut, 1995; Boonen and Schut, 2011). A pivotal step in these market-oriented reforms was the 2006 Health Insurance Act, which reinforced competition among both health care providers and insurers, while government regulation would safeguard the accessibility of services for all Dutch citizens. The centrepiece of this legislation – primarily aimed at generating improvements in quality and efficiency via market mechanisms – was the new purchasing role for private health insurers (Maarse et al., 2016; Varkevisser, 2019). In the new system, insurers would become ‘prudent purchasers’ who critically contract providers in terms of buying high-quality care at a reasonable price on behalf of their enrolees (Boonen and Schut, 2011; Stolper et al., 2019).

In theory, organisations that act as third-party payers, such as Dutch health insurers, have a wide range of potential strategies and contract models they could adopt for the purpose of prudent purchasing (Boonen and Schut, 2011; Song et al., 2012; Conrad, 2015). Perhaps the most obvious strategy, commonplace in many countries including the USA, is that of selective contracting, whereby patients would be channelled to contracted preferential providers, which should incentivise providers to live up to desirable standards in terms of quality or prices (Boonen and Schut, 2011). Additionally, the characteristics of the contracts themselves offer options to stimulate quality and efficiency (Song et al., 2012; Conrad, 2015). Of particular interest here are value-based purchasing models: contracts designed to incentivise providers to deliver high-value care, that is, optimising the ratio between health-related outcomes and costs (Conrad, 2015; Perrin, 2017). The most prominent types of value-based purchasing contracts include pay-for-performance schemes (in which providers are financially rewarded for doing better in terms of predefined outcome metrics) (Eijkenaar, 2013) and bundled payments (BPs) for episodes of care (which should incentivise providers to optimise their care cycles for the respective episodes) (Porter and Kaplan, 2016). Overall, value-based purchasing would entail a shift in focus: from reimbursement focusing on the volume of services to reimbursement that incentivises improvements in value – with value defined as the health outcomes achieved, divided by the costs needed to achieve those outcomes (Porter and Teisberg, 2006; Porter and Kaplan, 2016).

In recent years, value-based payment models have gained a lot of attention, with studies reporting on implementation efforts around the globe (Cattel et al., 2020). When it comes to the Netherlands, however, a growing number of studies on insurer–provider contracts reveal a purchasing market dominated by contracts that, instead of focusing on outcomes and efficiency, continue to emphasise volume, capacity, and cost control through annual budget ceilings (Maarse et al., 2016; Jeurissen and Maarse, 2021; Gajadien et al., 2023). Although proponents of market-oriented reforms in the Netherlands persist in advocating selective contracting, Dutch health insurers appear to be either unable, unwilling, or both to adopt this strategy (Boonen and Schut, 2011; Stolper et al., 2019). And the same appears to apply to any other strategy that would involve insurers redirecting patients (Ibid.). While deemed critical for the functioning of the system of regulated competition (Varkevisser, 2019; Maarse et al., 2016; Stolper et al., 2019), thus far, Dutch health insurers have not been able to fulfil their role as prudent purchasers – at least not in the manner that was envisioned (Jeurissen and Maarse, 2021).

Recent work by health economists has brought up several reasons for the limited extent to which insurers have based their purchasing practices on a critical assessment of the quality and the prices of hospital services – let alone a value-based assessment of care cycles (Boonen and Schut, 2011; Stolper et al., 2019; Varkevisser, 2019). Most prominently, insurers experience a lack of trust by enrolees in the intentions of their purchasing practices, so they refrain from the more prudent strategies, and particularly avoid selective contracting, out of fear of reputational damage (Boonen and Schut, 2011). Another hurdle insurers experience concerns legal provisions that enforce insurers to reimburse enrolees for up to 80 per cent of their bills when they utilise uncontracted care (Stolper et al., 2019). On top of this, insurers experience a lack of publicly available information on the quality of health care services (Boonen and Schut, 2011; Stolper et al., 2019). According to these scholars, publicly accessible information on the measured quality of health care services would go a long way in enabling insurers in their efforts to fulfil the role of prudent purchasers of care (Ibid.).

Although we recognise the validity of the reasons provided, this paper elicits alternative explanations. Specifically, we draw on insights from institutional theory (Meyer and Rowan, 1977; Hall and Taylor, 1996; Thelen, 1999; Scott, 2004) to offer a more historically rooted and socio-political account (rather than a purely economic one). Institutional theory constitutes a body of scholarly work that has demonstrated extensive explanatory potential regarding the way organisations behave (Torres, 1988; Greenwood et al., 2008; Deephouse et al., 2017). Hence, this study highlights an organisational context characterised by multiple and conflicting socio-political institutions, with historical roots that shape and constrain the behaviour of Dutch health insurers. The main aim of this article is to contribute to understanding the purchasing strategy and behaviour by which Dutch health insurers are trying to improve the value of health care provision. More specifically, we examine how insurers perceive the context in which the value-based purchasing of hospital care should take shape, and we draw on insights from institutional theory to frame our analysis.

This study provides relevant insights with regard to health care systems that are based on regulated competition, especially when health insurers operate as third-party payers who are expected to purchase prudently. Additionally, our analysis of the purchasing behaviour of Dutch health insurers offers insight into the ways in which the applicability of prudent purchasing strategies and value-based contract models may be constrained by the particular history and socio-political institutions of a health care system. With that in mind, the next section starts with a description of our analytical lens and ends with a subsection on the Dutch system, followed by our methods, findings, and the discussion.

2. An institutional perspective

Central to most applications of institutional theory is a focus on how social institutions influence the behaviour of organisations and the individuals within them (Hall and Taylor, 1996; Scott, 2004; Greenwood et al., 2008; Steinmo, 2008; Suddaby, 2010; Suddaby et al., 2013). While existing definitions can vary rather widely, ‘institutions’ can be usefully conceptualised as enduring modes of social exchange that reflect prevalent norms and beliefs and which have acquired a rule-like status in that they pressure for conformity (Meyer et al., 1994; Scott, 1994; March and Olsen, 2004; Greenwood et al., 2008). Accordingly, institutions function as enduring prescriptions of appropriate conduct that shape (organisational) relationships, routines, and (perceived) responsibilities (Meyer et al., 1994; March and Olsen, 2004; Greenwood et al., 2008; Mahoney and Thelen, 2010). Their rule-like status can be both formal and informal, but as individuals and organisations repeatedly conform to their pressures, institutions tend to become taken for granted as legitimate (Meyer and Rowan, 1977; Lawrence and Suddaby, 2006).

2.1 Legitimacy in organisational institutionalism

Institutional theory has become a highly influential approach in studying organisations (Greenwood et al., 2008; Suddaby et al., 2013). Key to its widely recognised merit is the explanation it offers regarding how and why organisations commonly act in ways that appear to run against their economic self-interest – i.e., why they adopt practices that hamper efficiency or other forms of technical performance (Ibid.). As outlined by two seminal papers dating back to the late 1970s and early 1980s, modern organisational environments are usually rife with institutions – that is, many organisations face multiple taken for granted and sometimes conflicting rule-like prescriptions that reflect various norms and beliefs (i.e. the ‘rationalised myths’) of the wider society (Meyer and Rowan, 1977; DiMaggio and Powell, 1983). So, the institutional explanation offered is that organisations tend to conform to their environments: they adopt practices for the sake of gaining external legitimacy (e.g. societal approval, stakeholder acceptance), thereby avoiding commotion and securing stability and survival instead (Ibid.).

To some extent, this institutional claim can be seen as an extension of a well-established fact in social psychology: individuals have a strong tendency to adopt practices and endorse opinions when they believe that these are supported by the people around them (Asch, 1956; Pinker, 2011). Similarly, organisations conform to their institutional environment to align with external expectations – i.e., to be recognised as legitimate (Deephouse and Carter, 2005). Studies have shown, for instance, a recurring trend in which hospitals are implementing changes (e.g. offering new treatments) not because of considerations around efficiency or patient demand but because of institutional mimicry and beliefs about the things ‘legitimate’ hospitals are supposed to be doing (Fennell, 1980; Van Wijngaarden et al., 2023).

It is widely recognised that organisations that are involved in or directly related to the provision of health care face an environment characterised by multiple and conflicting institutions that pressure organisations to conform their behaviour (Scott and Meyer, 1994; Currie and Guah, 2007; Van de Bovenkamp et al., 2014). In this paper, we borrow insights from organisational institutionalism, particularly the notion of legitimacy, to analyse the purchasing behaviour of Dutch health insurers.

2.2 Layering in historical institutionalism

Next to analysing organisational behaviour, a significant body of work within institutional theory has been devoted to studying the temporal and context-specific processes by which institutions emerge and evolve (Thelen, 1999). This body of work, commonly labelled ‘historical institutionalism’, unravels the (political) processes by which some institutions and not others have emerged and evolved as prevalent and legitimate in particular times and places (Thelen, 1999; Steinmo, 2008). Accordingly, it can shed light on the story behind the conformed behaviour of organisations, who, of course, do not apathetically conform to whatever pressure they may experience. They are likely to conform, however, to (what they perceive as) powerful and enduring institutions that reflect prevalent beliefs and expectations.

While such institutions indeed tend to have an enduring character, they do have an origin, and they do change over time (Thelen, 1999; Mahoney and Thelen, 2010). For both their origins and their changes, historical institutionalists point towards the significance of path dependency: institutions are seen as ‘the legacy of concrete historical processes’, and once in place, they change in response to circumstances, ‘but in ways that are constrained by past trajectories’ (Thelen, 1999: 381 and 387). Although they occasionally develop through radical shocks, the constellation of institutions is generally considered a matter of gradual evolution, dependent on local challenges and past trajectories (Mahoney and Thelen, 2010).

Of particular interest here, is what is referred to as institutional layering: the type of gradual change in which new arrangements do not supplant existing institutions, but are grafted onto or placed alongside them, thereby altering the significance of existing arrangements and the degree to which the older institutions affect behaviour (Mahoney and Thelen, 2010; Van de Bovenkamp et al., 2014). This type of institutional evolution tends to occur in environments where defenders of the status quo (e.g. medical professionals) are powerful enough to preserve existing institutions (e.g. professional self-regulation) but cannot prevent that slight amendments and incremental modifications will be made (e.g. through market-oriented reform policies) (Mahoney and Thelen, 2010; Van de Bovenkamp et al., 2014).

Organisations that operate in the Dutch health care system do so in a layered institutional environment: they face a hybridity of rule-like arrangements with deep historical roots (Helderman et al., 2005; Schut, 1995; Van de Bovenkamp et al., 2014; Van de Ven, 2015). Let us highlight the most significant parts of these roots, with special attention to the role of health insurers.

2.3 The Dutch legacy

In the early 1900s, the Dutch system predominantly operated through private initiatives, with private health care providers and private though generally not-for-profit insurers (including sickness funds), and with little to no government intervention (Van de Ven, 2015). From the perspective of insurers, the system was largely a free market, but one that was already ingrained with elements of professional self-regulation that reflected the dominant position of medical professionals (Torres, 1988; Maarse and Jeurissen, 2020). For instance, it was due to pressures from the National Medical Association that sickness funds would only be available to people with lower incomes (Maarse and Jeurissen, 2020). Moreover, when sickness funds sought to eliminate the obligation to contract all medical professionals in their regional working area – that is, engage in selective purchasing – medical professionals were able to prevent this from happening (Van de Ven, 2015). And although there were various government attempts to establish a national insurance scheme in the form of mandatory sickness funds, all such state efforts proved fruitless before the Second World War (Helderman et al., 2005). At least in part, this was due to ‘fierce opposition from the medical profession’ (Jeurissen and Maarse, 2021: 6).

The situation changed drastically in 1941, when the German occupying forces imposed a compulsory sickness fund scheme that covered people under a certain income threshold (then about 40 per cent of the population) (Maarse and Jeurissen, 2020; Van de Ven, 2015). Sickness funds, which used to be private and financially independent organisations, were now transformed into executive government agencies (Van de Ven, 2015; Jeurissen and Maarse, 2021). These impositions left their imprints: in the decades after the war, the Dutch state was able to take on a much larger role than during the pre-war era, and the 1964 Sickness Fund Act (Ziekenfondswet) was essentially a copy of the scheme imposed by the Germans in 1941 (Jeurissen and Maarse, 2021). Regionally operating sickness funds, obligated to contract all providers in their region, now cover about two-thirds of the population (Van de Ven, 2015). People who fall above the income threshold could acquire private health insurance.

Starting in the 1960s, the health care system increasingly incorporated corporatist arrangements – characteristic of Dutch politics and policy-making – in which decision-making is based on seeking consensus through collective bargaining with and between powerful interest groups (e.g. associations of provider organisations, medical specialists, and insurers) (Schut, 1995). From the 1970s onwards, faced with the need to control costs, the state stepped in and enforced top-down rationing by regulating prices, capacities, and budget caps (Helderman et al., 2005; Van de Ven, 2015). By the mid-1980s, however, doubts about the effectiveness of large-scale state interventions led to the gradual introduction of more market-oriented policies (Helderman et al., 2005; Schut, 1995). In 2006, these gradual reforms culminated in the Health Insurance Act, by which the government shifted responsibility for cost-containment to providers and private insurance companies. This Act ended the separation between sickness funds and private insurance schemes: all health insurers now belong to the same regulated market, where they are mandated to offer the same basic insurance package, and where they are encouraged to compete nationally and purchase prudently.

3. Methods

To better understand the purchasing practices by which Dutch health insurers strive to improve the value of hospital care, we conducted a series of semi-structured interviews (n = 18) between February and July 2022 with a variety of insurer employees and representatives. More specifically, we asked insurers about their perspective on the utility of value-based contract models for hospital care and how they perceive their overall ability to fulfil the role of prudent purchaser within the context of the Dutch health care system. An interview guide and topic list were constructed for the purpose of this study (for the English translation, see Supplementary file 1).

Since our aim was to gain both a profound and generally representative picture of the perspective of Dutch health insurers, we intentionally interviewed respondents with five types of roles: buyers (n = 4); strategy and policy officers (n = 5); medical advisors (n = 4); upper management of purchasing (n = 3); and chief executive officer (CEO) (n = 2). Moreover, we interviewed employees from six insurance companies and multiple representatives (n = 15, including two CEOs) of the four largest health insurers, who, in 2022, together held 85 per cent of the insurance premium market (Nijhof et al., 2024). We also interviewed two employees (a buyer and a medical advisor) from health insurers with a relatively smaller market share and one policy officer of the national association of Dutch health insurers (Zorgverzekeraars Nederland). Respondents were contacted via e-mail by way of purposeful sampling, initially based on the authors’ professional network, later to be complemented by way of snowball sampling. Interviews were held through online videoconferencing and lasted 73 minutes on average, the recordings of which were transcribed verbatim.

Interview transcripts were analysed within two rounds of coding. In the first round, the first author marked fragments that stood out in light of our research objective; these fragments were shared with and discussed among all authors in the form of analytical reports. This, in combination with our theoretical framework, laid the groundwork for the second round of coding in which all of the transcripts were analysed thematically, using both theory-driven deductive codes (e.g. ‘legitimacy’, ‘layering’) and codes that refer to themes that emerged inductively (e.g. ‘budgeting’, ‘DBC prices’). See Supplementary file 2 for our final coding scheme.

It should be noted that although our interview guide (Supplementary file 1) formed the semi-structured basis for all interviews, as the data-gathering stage proceeded, we increasingly incorporated member checks by presenting preliminary insights based on previous interviews to new participants. We used this form of member check to strengthen our findings and prevent that one or a few of the most outspoken or persuasive participants would skew our interpretations. Furthermore, all cited participants have given written permission to cite them accordingly after being presented with the abstract and full results section of the manuscript we first submitted to this journal.

4. Findings

4.1 Lacking legitimacy

Interestingly, insurers perceive the idea of value-based purchasing – in which considerations about the outcomes and costs of services would inform purchasing behaviour – to be a core component of their role within the Dutch system of regulated competition. However, actually fulfilling the role of prudent purchaser in general, and value-based purchasers of hospital care in particular, is considered an arduous task. In line with previous studies, our analysis revealed a perceived lack of available information that constrains insurers’ ability to apply prudent purchasing practices. Moreover, Dutch health insurers appear unable to enforce the availability of the information they desire.

I think that based on the starting points of the Health Insurance Act as they were originally formulated, it is a clear assignment for insurers to make the connection between quality and costs, and payment. Definitely. In theory, yes. The biggest problem – I will just start right away – the biggest problem: that it is very difficult to unveil and obtain proper quality information. If you want to pay for good quality of care you will therefore need good information on quality; and obtaining that information is very complicated (4).

Operating underneath these perceived constraints due to a lack of availability of information, however, we also found a profound lack of organisational legitimacy: Dutch health insurers feel that for them to make judgement calls about quality of care would not be seen as appropriate conduct by the wider society.

Then you really quickly get into a very difficult situation. Because who are you to judge that as an insurer? And you also don’t even have society’s mandate to make choices like that (11).

Dutch health insurers refrain from making independent judgement calls about the quality of care; they would much rather have medical professionals setting their own standards. So, although the Health Insurance Act (2006) officially (i.e. on paper) paves the way for insurers to set their own norms for the purpose of prudent purchasing (e.g. with benchmarks for outcome measurements), its practical effects have not dismantled the dominance of professional self-regulation:

Something which we always try to pay close attention to, especially when it is about actual medical issues, is that we are not going to step into the position of the doctor. So, we are sometimes asked to set a norm ourselves, yeah that is, we would much rather have the professional field setting a norm […], you want there to be recognition from within the field, about what is good and what is not (10).

Next to the historically rooted dominance of professional self-regulation, purchasing behaviour is also constrained by (perceptions of) prevalent beliefs and expectations – the rationalised myths – of the wider society. In line with institutional theory, Dutch health insurers conform to such beliefs.

In this country, we have such a deep-seated belief that all the care that is prescribed is relevant and of good quality. So, yes, if you say that, within that [context], insurers think very carefully about what they get themselves into, also because politicians are so highly responsive to it; yes, that very much determines the playing field. And that is also the reason why we are really not going to start any large projects anymore […] around quality indicators or around transparency (18).

So, for Dutch health insurers, having information available does not automatically translate into the ability to incorporate that information in prudent purchasing practices. By contrast, these organisations find themselves restricted in fulfilling the role of prudent purchaser because they do not experience society’s mandate to make judgement calls about the measured performance of providers in terms of quality or outcomes. Moreover, doing so is perceived to entail risky exposure to public scrutiny:

That is the fear that reigns among health insurers not to make hard choices: because health care professionals or health care organizations will then go to the media, and you just know that as a health insurer you are 3-0 behind if you appear in the media. Because who are you going to believe, the doctor or the insurer? […] I still notice that this plays a big part. Sometimes I am very much inclined to just make a very logical choice, of which I think ‘come on, we are in the driving seat, we now just need to make a choice’, and then my colleagues respond very reluctantly: ‘yeah, but the consequence will be that we appear in the media’ (6).

Rather than risking public exposure, Dutch insurers avoid making hard choices when it comes to the purchasing of hospital care: they practically never (dare to) apply their own quality standards. Instead, they steadfastly submit to the standards set by medical professionals and thereby conform to the historically rooted institution of professional self-regulation and society’s deep-seated belief in the quality of care in the Netherlands.

4.2 Misleading prices

In the Dutch system of regulated competition, hospitals directly bill health care insurers for the services they have provided to their respective members through a system of diagnosis treatment combinations (diagnose behandel combinaties, DBCs). In theory, and as intentioned by the market-oriented reforms of the 2006 Health Insurance Act, between-hospital differences in DBC prices for the same service should indicate either varying production costs, different profit margins, or both. This would make claims data a useful tool for prudent purchasing in general and value-based payment models in particular. However, as indicated by the quote below (from someone who used to negotiate contracts with insurers from the side of a hospital), this line of reasoning does not hold water when it comes to hospital-insurer contracts in in the Netherlands.

I now actually consider the whole purchasing story claptrap (handjeklap). […] For example, suppose that I as an insurer have hips and knees as a point of focus. Then I go to the hospital, and I say: I see some practice variation regarding hips and knees. And then the hospital says ‘okay, I see your point […] I can do something about the prices’. […] What I will then do – I’m now in the hospital role – I have drawn up that price list myself: everything I have agreed on with that insurer regarding hips and knees, that is fixed, but all the other care products, those are variable. […] And in the end, we have agreed on a certain number for our overall contract value. So, all those other products, those I will adjust in price to compensate. […] So, there is really no relationship to be found between production price, sales price, and final contract value. Because in the end, I recalculated and compensated all of it to come full circle (7).

In practice, purchasing negotiations between insurers and hospitals, which usually take place annually, are first and foremost about budgets: the total amount for which hospital ‘A’ can bill insurer ‘B’ next year. In the vast majority of cases, it is only after this total amount is agreed upon that DBC prices get discussed – that is, if those prices even become a matter of debate at all. Although the Dutch system of regulated competition officially contains a lot of openly negotiable DBC prices, their de facto function on the purchasing market is restricted by the institutionalised practice of budgeting.

So everyone calculates down towards a budget from which the fixed costs and the salaries all need to be paid. And the DBCs are but a means, and the rates are just a means to round up that budget. Well yes, then you always have a hassle: whether you work with ceilings, how to deal with overproduction, gross prices and net prices of DBCs, about all of that. But in the end, we never really got rid of the old system of function-oriented budgeting (5).

Before the 2006 Health Insurance Act, Dutch hospitals were funded through ‘function-oriented’ budgeting (functiegerichte budgettering), in which the government determined annual budgets for each hospital based on their ‘functioning’ in terms of accessibility parameters, overall capacity, and previously delivered volume of services (Broertjes, 1992). The Health Insurance Act was meant to displace this type of funding: hospitals would no longer have their budgets guaranteed; instead, their income would depend on their performance and negotiations with insurers. But as of yet, the Insurance Act and the gradual expansion of openly negotiable prices that followed it (from 10 per cent total hospital funding in 2006 to 70 per cent in 2012) have not replaced the annual cycle of budgeting. Rather, it has grafted a new market-oriented layer (of free pricing) onto the historically rooted practice of budgeting.

Again, with the Health Insurance Act and with the DBC system, the original purpose was that we would start to purchase care, and that we would purchase care based on quality and outcomes. We just haven’t done that for the past fifteen years. We’ve drifted in all directions, but we have nevertheless still been involved with budgeting. A lot of old buyers have simply gone along with the system, and they still continued to make agreements on budgets with those hospitals. Calculated differently than before, but essentially it still comes down to that (4).

Strikingly, the switch to openly negotiable DBC prices – a pivotal part of the market-oriented reforms that have shaped the Dutch system of regulated competition – has primarily altered the administrative method of billing as part of the enduring practice of hospitals being budgeted annually by their financiers.

4.3 Consensual constraints

Next to the restrictions imposed by a (perceived) lack of available outcome data, a lack of relevant prices, and legitimacy and reputation issues, the purchasing behaviour of Dutch health insurers is also constrained by corporatist arrangements. Most prominently, this concerns the national administrative accords, which result from negotiations among the most dominant interest associations within the Dutch health care sector. These accords outline the overarching health policy directions for the next four years. Both the ‘Main Direction Accords’ (Hoofdlijnenakkoorden or HLA) and more recent ‘Integral Care Accords’ (Integraal Zorgakkoord or IZA) were repeatedly brought up by several respondents in relation to the purchasing behaviour of insurers.

What you see now is that because of that Accord – the negotiations are now taking place, this summer – that will determine the frameworks of what will happen at the purchasing table in the coming years. […] For example, there will be an agreement on a certain percentage of digitalization of care; or there will be an agreement on an overall percentage of decline or growth. That will be directly translated, one-to-one, to all parties at the table. ‘If we nationally agree on 0.3, then that also applies to me’. […] Those Accords tremendously frame the behaviour at the table, and also what people think they may be able to accomplish. And that applies to both parties. It paralyzes us (18).

So, the consensual agreements made in the national accords (HLA, IZA) directly set the parameters for the annual contract negotiations between hospitals and insurers. Of particular interest here are agreements on total annual spending, which are ‘basically indicating the financial frameworks’ (2) for the annual negotiations. In practice, both parties can invoke what the representative interest associations of either side have agreed upon. While these agreements can thereby reduce financial uncertainty, they also have a restrictive effect on purchasing behaviour, and they reinforce the already institutionalised practice of budgeting. Hence, the Dutch institution of consensual decision-making through corporatist arrangements, which reappears in the national accords (HLA, IZA), effectively regulates behaviour on the Dutch purchasing market of hospital care, and constrains insurers’ (perceived) ability to purchase prudently.

4.4 A restricted market

For several reasons, insurers experience a sense of responsibility for the financial fitness and survival of hospitals within their core regions (where they have a large share of the insurance premium market). One is that they are legally required to safeguard the accessibility of hospital services for their enrolees. Another one is related to the perceived lack of legitimacy mentioned before: Dutch health insurers feel that if a hospital (almost) goes bankrupt, or otherwise ceases to exist, the general public will perceive that as illegitimate conduct by untrustworthy insurers. Within this context, insurers sometimes behave in ways that contradict the principles of prudent and value-based purchasing.

One exemplary case, which was brought up in multiple interviews, concerned a hospital that came into financial trouble because it received considerably fewer referrals from general practitioners (GPs). In agreement with the largest insurer of the region, these GPs were granted more time per consultation, and this extra time turned out to be a highly effective way of lowering the local demand for specialty care. However, when it became publicly known that the hospital was missing vital streams of income, the insurer cancelled its agreement with the GPs in an attempt to regenerate the demand for specialty care and save the hospital from bankruptcy. Interestingly, several respondents argued that in this particular case, it would be logical to let the hospital go bankrupt or otherwise hold its own. There were other hospitals located nearby, so accessibility of care would not have been an issue here. However, there were socio-political factors at play that made the idea of hospital bankruptcy ‘politically totally unfeasible’ in the eyes of insurers (10).

Public opinion is already mentioned. Yeah, that just has enormous impact. And then I don’t just mean the public as in people who read the paper, but also simply the political domain. So, if there is a hospital that gets into trouble or where there are doubts about its continuity, then that is something that immediately gets discussed in parliament, and of which it is also directly said: ‘it cannot be true that this hospital at that location would no longer be there’. And that thereby also gives tremendous pressure towards parties such as the insurer (11).

In regions where they enjoy a large share of the premium payer market, insurers have an indirect but vested interest in keeping hospitals financially healthy. This avoids unwanted commotion and thereby serves their own stable survival. And instead of experiencing a strong bargaining position (which may typically be expected in a competitive market), Dutch health insurers consider a large regional market share to make it inevitable and absolutely necessary to reach a contractual agreement with hospitals in those regions.

With an organization where you are the largest, where you really are by far the largest, that is a kind of marriage that you have deal with. Then it is better to make something of it from which you can both get certain benefits (6).

Ironically, this has created a purchasing market in which insurers avoid the most critical and prudent purchasing strategies in exactly those regions where it could have the most impact on the behaviour of providers and the medical decisions of their enrolees. This helps explain, for instance, why Dutch insurers consider the most obvious prudent purchasing strategy available to them – selective contracting – to be ‘on the brink of death’ when it comes to hospital care (1). After all, it would only be an effective strategy in those regions where an insurer has a lot of potential patients to channel, but those are exactly the places where political pressures and societal expectations more strongly prescribe less stringent behaviour.

4.5 Value-based budgeting?

In today’s purchasing market of Dutch hospital care, any type of prudent or value-based purchasing would have to take form within the institutionalised confines of a budgeting system. This way of funding has not just become the historically rooted norm, it is the primary mechanism by which hospitals ensure that their annual costs are covered, and the main way by which insurers control their total expenditure on hospital care.

I have been involved in so many innovative contracts, if you do not agree with each other on how big the [budget] is, then everything is pointless and meaningless. […] So, everybody knows that this is how the game is played, and that is the why we have said: those contracts start with shared understanding of the infrastructure and the required budget, and after that we fill in the rest (18).

Although a large market share, as described in the previous section, indeed limits insurers’ (perceived) options in terms of prudent purchasing, it often also implies a forceful interdependence of resources between hospitals and insurers. And this interdependence can set the stage for what appears to be the most prevalent type of alternative insurer-hospital payment: multi-year contracts that include agreements on the overall quantity of services but can also involve agreements on quality. It should be noted that insurers make these types of agreements primarily in their core regions.

I think we have also come to learn that these types of discussions should actually be poured into multi-year agreements, in which you together have the starting point: we are about to face a huge care demand, the financial frameworks are limited […] let’s agree that this is the financial framework […]. Well, and if you are past that point, and say, in confidence, ‘okay, 100 will remain 100’ […] then you already have a very different conversation (2).

These multi-year contracts usually consist of prospective lump-sum payments for an agreed-upon quantity of service provision that covers three to five years (occasionally more) and sometimes include a pay-for-performance component to incentivise hospitals to reach certain benchmarks. Essential, however, is their stabilising effect on the interorganisational insurer-hospital relationship: these multi-year deals remove financial insecurity and thereby form the foundations for collaborative partnerships. At least to some extent, this can enable insurers to discuss and potentially affect operational matters within hospitals.

That just gives a lot of peace, because then you no longer have that annual circus about ‘does it need to be 99 or 101’ […]. And that really is a very different way of talking to each other. Instead of arm-wrestling about finances, we are suddenly discussing the composition of care, accessibility, yes – because you have created financial rest with each other (2).

Within a context of institutionalised budgeting and regional market share differences, Dutch health insurers employ multi-year contracts to foster a cooperative relationship with hospitals. Rather than the ‘annual circus’ of ‘arm-wrestling about finances’ (2), in their core regions insurers seek to build partnerships.

We look for collaboration with them, long-term and intensive, with hospitals but also with [other] mostly large providers. […] We say: let’s remove the dynamic of annual negotiations, because we are, so to speak, just recalculating full circle and rounding up [the budgets of] each other; we have work it out together. We are also sort of condemned to each other. […] It is better to do that in a form of co-creation and in a form of collaboration (14).

At least in their core regions, the historically rooted practice of budgeting appears to have formed a basis for a type of purchasing in the form of multi-year contracts with lump-sum payments. Multiple respondents referred to these types of contracts as part of a purchasing strategy by which Dutch health insurers try to improve the quality and value of hospital care. By contrast, and even though we explicitly asked respondents about their views on BPs, this contract model was deemed too risky and too complicated for widespread implementation. And while BPs would break with the traditional budgeting system, the multi-year contracts fit right in. Even though one may question whether they are truly value-based, the stabilising effects of such multi-year agreements may very well make these contracts the most viable form of prudent budgeting.

5. Discussion

With the commencement of the Health Insurance Act in January 2006, the Dutch health care system officially transitioned to a model of regulated competition, marking a major regulatory overhaul. In the new system, private health insurers were assigned a particularly crucial role: as prudent purchasers, they would become critical drivers of value-based competition among providers. The main aim of this article is to shed new light on Dutch health insurers’ purchasing behaviour by examining how insurers perceive their role as value-based purchasers of hospital care. Our findings reveal several institutional factors that restrict insurers’ (perceived) ability to purchase prudently. Although the Health Insurance Act indeed established a regulated market system in the Netherlands, this did not wash away the existing institutional environment characterised by professional self-regulation, regionally concentrated health insurers, corporatist arrangements, and global budgets. Instead, it grafted a new layer of market mechanisms onto the existing institutional context.

While earlier accounts of Dutch insurers’ purchasing behaviour have relied on an economic rationale (Boonen and Schut, 2011; Maarse et al., 2016; Schut and Varkevisser, 2017), our study draws on insights from institutional theory (e.g. Greenwood et al., 2008; Meyer & Rowan, 1977) to provide a more historically informed, socio-political interpretation of insurers’ purchasing practices. It should be noted, though, that our intention is not to argue against the potential validity of other (economic) reasons; our intention is to demonstrate the added (complementary) value an institutional perspective can offer.

For instance, our findings confirm the commonly put forth explanation regarding the lack of value-based purchasing practices: that insurers do not have the information they need (Boonen and Schut, 2011; Stolper et al., 2019). However, our analysis also revealed that it is not merely a lack of information that is restraining insurers; we have found a more profound lack of legitimacy at play. Dutch insurers believe that for them to make judgement calls about the quality or value of care would not be in accordance with society’s expectations. In this institutional context, having information available does not automatically translate into the ability to incorporate that information into prudent purchasing practices. This helps explain why insurers, who are well-aware of the generally positive volume-to-outcome relationship (e.g. Chioreso et al., 2018; Huo et al., 2017), virtually never set their own volume thresholds, but steadfastly conform to the norms set by the medical profession (e.g. yearly hospital minimums of 20 and 50 surgical procedures for rectal cancer and colon cancer, respectively). Moreover, recent studies have shown that Dutch health insurers actually do have data available to them that can reveal considerable between-hospital performance differences for the same treatments (Schepens et al., 2023; Schepens, 2024). As it turns out, however, multiple years of data showing significant between-hospital differences do not warrant insurers to make a judgement call on behalf of their enrolees – only the professionals themselves can legitimately evaluate their work.

Another illustrative example of an institutional constraint is the enduring practice of budgeting. This historically rooted way of funding hospitals, which insurers have consistently applied and thus reinforced as legitimate, has now become taken for granted and virtually unavoidable in the eyes of insurers. Indeed, budgeting has become so institutionalised within the Dutch health care system, that it may be considered an institution in its own right – an enduring arrangement of exchange that has acquired a rule-like status in that it pressures for conformity (Meyer et al., 1994; Scott, 1994; March and Olsen, 2004; Greenwood et al., 2008). Consequently, our findings indicate that any type of value-based payment would have to take shape within the constraints of this budgeting system. And on a more general level, it appears that within the regulated market of Dutch hospital care, at least for the moment, the idea of prudent purchasing will essentially come down to the practice of prudent budgeting.

It is within this context that Dutch health insurers have come to view multi-year agreements with prospective lump-sum payments as a logical and prudent extension of their usual contracts. These agreements, which some have labelled ‘global budgets with a new look’ (Jeurissen and Maarse, 2021: 74), reflect the entrenched budgeting mentality that prioritises financial predictability and stability. Nevertheless, they may also enable insurers to promote their prudent concerns within hospitals, as several of our respondents explicitly pointed towards these contracts as a way for purchasing to become value-based. Yet, whether these multi-year deals truly represent a shift towards value-based purchasing, or rather the latest iteration of the budgeting paradigm, will ultimately depend on the specific characteristics of the contracts and the criteria one uses to assess them.

For some scholars, including some of the most prominent voices on value-based health care (e.g. Porter and Teisberg, 2006; Porter and Lee, 2013; Porter and Kaplan, 2016), these types of contracts would fail to meet the criteria for value-based payment. While Porter and colleagues continuingly advocate BPs at the level of care cycles for medical conditions (Ibid.), these multi-year contracts apply to the hospital as a whole – covering low-value as well as high-value care cycles. Even if such contracts would include a lot of pay-for-performance elements designed to incentivise outcome improvements, this would make better outcomes increasingly more expensive, which runs against the idea of improving value (Porter and Teisberg, 2006). Rather than rewarding hospitals for their overall performance, Porter and colleagues emphasise the importance of rewarding high-value providers with more patients, specifically at the level of medical conditions (Ibid.). Instead of pay-for-performance, value-based purchasing would focus on incentives to direct patients towards high-value care cycles – patients-for-performance.

In line with previous studies, however, we have found that the idea of insurers channelling hospital patients through purchasing practices is of questionable viability in the Netherlands and practically futile when selective contracting would be the mechanism to do so (Boonen and Schut, 2011; Jeurissen and Maarse, 2021). Interestingly, this has been the case since the early 1900s: Dutch health insurers have never been able to use their buyer position selectively concerning hospital care; specialist physicians have always been able to fend off third-party interference with patient streams (Van de Ven, 2015).

This study has at least two important limitations. First, our findings rely on a limited number of purposively sampled respondents, which may have created some bias – even though we have sought to counter this by interviewing respondents with various roles and through member checks. Second, while the institutional perspective we have applied here has certainly proven useful for analysing health insurers’ purchasing behaviour, our focus on institutional factors may have left other factors underexplored. Relatedly, it was beyond the scope of this study to examine the relative weight of institutional factors. To be clear, though, our objective in this study was not to devalue the validity of other factors or perspectives but to promote the recognition of institutional factors and the added value an institutional perspective may offer.

The lens of historical institutionalism, for instance, can draw out connections between the regionally concentrated (and regionally constrained) health insurers of today and the regionally restricted sickness funds of the twentieth century. While sickness funds have played a vital role in the Dutch health insurance landscape during most of the 1900s, it was not until 1992 that they were allowed to operate nationally. Before then, the funds were tightly confined to their separate working areas, where they acted as ‘regional monopolists’ (who were prohibited to contract selectively) (Jeurissen and Maarse, 2021: 20). But the early 1990s saw a set of market-oriented policy decisions that sparked a wave of mergers across the by then fragmented health insurance landscape. By the time the Insurance Act came into force (2006), what had once been a diverse mosaic of sickness funds and private insurance schemes had now largely consolidated into a handful of regionally dominant insurance companies (Vonk, 2013). When we take into account the additional mergers that have occurred since then, the market shares of the four largest insurance companies, including their regional concentrations, have remained noticeably constant to this day (cf. Nijhof et al., 2024). And just like their twentieth-century predecessors, their negotiations with hospitals have consistently been region-oriented, budget-focused, nonselective, and conforming to professional self-regulation. Although our intention is certainly not to argue against the validity of economic reasons that may explain these patterns – nonselective global budgeting may indeed offer the most rational and prudent approach in this situation – we do intend to promote the explanatory potential of a more historically informed account, which may point towards alternative or complementary factors. Specifically, an institutional perspective can help illuminate the past trajectories that have shaped what insurers consider appropriate and rational courses of action – that is, what it means to purchase prudently on this particular market.

Within the regulated market of Dutch hospital care, health insurers face more than just technical challenges (e.g. a lack of information; legal restraints). These organisations operate in an environment where market mechanisms are tangled up with historically rooted budgeting practices, where insurers are pressured to sustain rather than critique hospitals, and where the dominance of a self-regulating medical profession is supported by a mythical belief in the quality of its services.