Davies, Stephen M. MA, MSc; Tawfik-Shukor, Ali M. Biotech, MSc; de Jonge, Bob LLM
The Netherlands has a rich tradition and history in academic medicine, which can be traced back at least to the 17th century and the Golden Age of the Dutch Republic. Rembrandt van Rijn's painting, “The Anatomy Lesson of Dr Nicolaes Tulp,” offers us a vivid illustration of this history. Herman Boerhaave (1668–1738) of the University of Leiden is widely regarded as the founder of clinical teaching and of the modern academic hospital. A more recent expression of this tradition has been the development of university medical centers (UMCs), institutions that are closely analogous to U.S. academic health centers (AHCs) in their commitment to the tripartite mission of patient care, health research, and health-related education. The eight Dutch UMCs are viewed pragmatically as “constructions for cooperation between universities and academic hospitals” that facilitate their shared tripartite mission. They carry out vital functions in the national health system and are recognized as one of the foundations of the Dutch knowledge-based economy.1 As a consequence, they are held to a social contract that obliges them to serve the country's economy and national interests. The government negotiates these social tasks with the Dutch Federation of UMCs (Nederlandse Federatie van Universitair Medische Centra), with a focus on maintaining the Netherlands, a nation of 16 million people, as a key source of global knowledge in biomedical and life sciences.2
In this article, we provide a descriptive account of Dutch UMCs and their recent history that, we hope, will be of intrinsic interest. But we also seek to move beyond description by applying a conceptual framework originally developed in the United States to analyze structure, governance, and organizational dynamics. We take this approach to test the transferability of a U.S. analytical framework to another national setting. In many countries of Europe, the AHC sector is less well developed than is the case in the United States, and both organizational leaders and policy makers look across the Atlantic for evidence relevant to their goals.3 If analytical frameworks derived from the U.S. experience can be applied to other national settings, then their use may contribute to transnational learning. Understanding the limitations of such frameworks when applied in a European context may also help guard against any insufficiently critical use of U.S. models.
Through comparative analysis, we also seek to demonstrate the relevance of theories of organizational design to academic health enterprises. The questions, “Why are organizations structured as they are?” and “Which organizational designs offer the best solutions?” are central to the discipline of organizational behavior. Early paradigms of scientific management sought to identify universal principles of organizational design that applied in all circumstances. This thinking later gave way to a more flexible view in which the appropriate structure for an organization is seen as depending on the particular circumstances, or contingencies, that it faces.4 Within this “contingency approach,” some have seen organizational structure as largely determined by the external environment, whereas others, working within a paradigm of “strategic choice,” have emphasized the scope for organizations to shape, or “enact,” their own environment.5 From the late 1970s onward, some organizational behavior theorists reacted against the view that the structuring of organizations was shaped primarily by the requirement for technical efficiency or task enactment and argued instead that uncertainty and a quest for legitimacy drive organizations to adopt structures and programs that reflect “institutional rules.” In health care, these rules might cover matters such as the regulation of professions or structures for the organization of clinical services and are seen as socially constructed rather than derived from objective evidence of technical efficiency.6 Exponents in this school, which has become known as “institutional theory,” argue that it is the similarity of organizational forms and practices within sectors, rather than their variation, that calls for explanation.7
U.S. AHCs, with their characteristic of organizational diversity within certain parameters, look like fertile territory for the generation and testing of organizational design theory, yet most U.S. analysis of AHC structure avoids any explicit theoretical basis. We find instead a general consensus that there is no one preferred structural answer and a reluctance to claim that what works in one place might be transferable to another.8 In other words, a contingency approach prevails, with an implicit mix of environmental determinism and strategic choice perspectives typically discernible in the many case studies of U.S. AHCs (see, for examples, the September 2008 issue of Academic Medicine). A cross-national comparative analysis might yield fresh insights into the value and limitations of different theoretical perspectives in explaining why academic health enterprises are structured as they are and so improve the external validity of AHC studies.
Of the eight UMCs in the Netherlands, six comprise partnerships between public universities and teaching hospitals. The remaining two partnerships involve private denominational universities, one Catholic and the other Reformist, whose governance arrangements confer a greater measure of autonomy from the state. Under the Statute on Higher Education and Scientific Research or WHW (Wet op het hoger onderwijs en wetenschappelijk onderzoek), formal accountability for the UMCs is still maintained by the university and academic hospital, which remain separate legal entities, whereas the joint board of the UMC is delegated sufficient powers to fulfill the tripartite mission. The scope of this delegation, the purpose of which is functional integration, is specific to each UMC. There is thus only limited variation in the ownership and legal basis of Dutch UMCs, yet the WHW allows for different approaches to governance, ranging from tight integration under common leadership to a looser functional integration based on contractual relationships.
The Netherlands is a decentralized unitary state in which health policy is decided at the national level with some delegation of health system management to local government (provinces and municipalities). The health system is characterized by a mix of regulated competition and market-oriented, incentive-based health care. Health care has historically been dominated by the voluntary sector, but it also includes public and private providers. Reforms in 2006 replaced a mix of public and private insurance with a single system in which all citizens are required to enroll with one of a number of competing health insurers. Extensive government regulation of this insurance market guarantees open enrollment and a ban on differential pricing by age, sex, or health status (an element of funding is provided from a national risk equalization fund). This approach to insurance, which seeks to reconcile the discipline of markets with a commitment to social solidarity, has aroused considerable interest as a possible model for reform in the United States.9,10 A similar balance is sought in the hospital sector, where the Dutch government has shown concerns with safeguarding the UMCs' special public functions (research, education, and specialized care) while maintaining competition with other care providers for those services that can be contested in a regulated market environment. Maintaining this balance has required a range of funding and regulatory mechanisms.
The goal of this study was to analyze the different approaches to governance found in Dutch UMCs and to consider the implications of this for theories of organizational design in relation to academic clinical enterprises. The study followed a comparative case study design, involving documentary desk research and semistructured interviews with UMC leaders. Interviews were initially structured using the typology developed by Weiner and colleagues11 in 2001 for the U.S. context, testing the extent to which an analytical framework developed for one national setting could be applied to another. Weiner et al employed three dimensions to construct a typology of medical school–clinical enterprise relationships. The first, clinical enterprise organization, is the extent to which the clinical affiliates in an AHC represent an organized delivery system, as opposed to a range of autonomous providers to which the medical school is linked by nonexclusive contracts and agreements. The second, academic–clinical enterprise integration, is the extent to which the medical school organizes the clinical practice activities of its faculty or otherwise influences the activities of its clinical partner(s). The third looks at the authority position of the chief academic officer (usually head of the medical school) over the clinical enterprise, including his or her role in the governance of the clinical enterprise and influence or control over resource allocation. These three dimensions are then combined to develop a typology with eight archetypes for medical school–clinical enterprise relationships.
We further developed this framework on an inductive basis as the data we gathered in the Netherlands yielded new insights. Clinical enterprise organization proved to be a dimension of limited variation because of the tight linkages between the university hospital and the medical school in the Dutch conception of the UMC. Also, the regulation of health care provider markets by the Dutch Healthcare Authority (Nederlandse Zorgautoriteit) means that university hospitals have specific roles (general and specialized acute care) in a national system that is, in effect, an organized delivery system for the entire population of the Netherlands.12 In contrast, it became apparent that organizational arrangements and dynamics below board level constituted an area of considerable focus and experimentation in the Netherlands.
In view of these considerations, the analytical framework we finally adopted was multilevel. We first considered structure and governance at the highest level, which required analysis of the composition, functions, accountabilities, and relationships of UMC boards and of the structure and ownership of UMC entities. The role, position, and power of medical faculty dean (chief academic officer) were also explored. At the level below this, we explored how functional integration for delivery of the tripartite mission was achieved at the subboard level. Finally, we considered some of the detailed organizational mechanisms used to bind together different missions within the UMC. The discussion of findings in the following sections is organized within this multilevel analysis.
After examining the literature and consulting with UMC leaders in the Netherlands, three centers were selected as case studies: Radboud University Nijmegen Medical Center, UMC Groningen (UMCG), and the Academic Medical Center of the University of Amsterdam (AMC). These cases were chosen because they exhibit differences in governance and because they differ across historical, geographical, and cultural contexts. UMCG is located in the provincial capital of the rural north of the Netherlands; AMC is in the nation's capital, itself part of the dense conurbation of the Randstad, home to some 7.5 million people; and Nijmegen is a relatively small provincial center in the east of the country. Radboud is one of the two UMCs linked to private denominational universities, in this case Catholic. AMC has a particular history as having developed from municipal and charitable foundations in the capital city, whereas UMCG's origins lie in a public university and voluntary teaching hospital.
Structure and board-level governance
Moving away from the traditional designs, new teaching hospitals built in the Netherlands in the postwar decades included centralized functional units that promoted integration of services, flexibility in use of facilities, and sharing of expensive equipment. However, the technical and logistical challenges associated with the creation of these institutions proved easier to solve than the challenge of developing an internal organizational structure that could successfully integrate the hospital and academic communities. The situation was so dismal in the 1980s that the Netherlands Court of Audit concluded that one of the teaching hospitals was “de facto ungovernable,” with a dysfunctional board and lack of leadership at every governance level.13 At that time, each medical school and academic hospital was separately managed and reported to different governing bodies. This led to a lack of organizational coherence, management conflicts, complicated reporting arrangements, difficulties in recruitment, and organizational inefficiencies at all levels, prompting a search for new ways to improve functional integration.14,15
From the mid-1980s to the early 1990s, pressure was mounting for organizational change and performance improvements. As an illustration of the challenges involved in responding, the process of negotiating new governance arrangements for UMCG took nearly two decades. Radboud also debated various integration scenarios at length, including either creating a medical university (i.e., splitting the medical school away from the parent university) or bringing the hospital under the management of the medical school. Both options were eventually abandoned because neither medical school nor hospital was prepared to submit to each other, and the separation of the medical school from the university was judged undesirable. In the case of the AMC, the hospital was highly supportive of closer integration as early as the mid-1980s; however, the medical school faculty was apprehensive about increased integration with a partner with a budget and staff nearly seven times its size.13,16 It was not until 1994 that UMC governance took on its current form, with the AMC becoming the first UMC. Across the Netherlands, different approaches to integration subsequently emerged, depending on local views and the relative power and size of university faculty and academic hospital. These ranged from quasi-mergers, in which a high level of authority was delegated by both medical school and teaching hospital to the UMC, to more formalized affiliation through strengthened contractual arrangements.15 The WHW allows for a diversity of UMC governance arrangements, classifying UMCs as “constructions for cooperation between universities and academic hospitals” with the goal of achieving a joint tripartite mission. The common element in all the solutions that emerged is that of a de facto UMC in which medical school and academic hospital remain separate de jure but are governed by a strong, integrated board with significant academic representation.
The most common model (see Figure 1) involves a two-tier governance structure within a not-for-profit entity, in which a UMC board is accountable to a supervisory board, members of which are nominated by the minister of science, culture, and education. The exceptions to this arrangement are Radboud and the Free University of Amsterdam UMC, where UMC boards have been established under the supervisory boards of existing university foundations. In these cases, appointments to the supervisory boards are by church authorities, conferring a greater level of autonomy from the state. Universities are strongly represented on supervisory boards (or, in the case of the denominational universities, control these boards) and also on UMC boards.
The precise nature of academic representation on the board and the powers of the dean are generally indicative of the nature of the pre-UMC relationships between the two organizations and their respective amounts of power. The AMC illustrates this, in that the dean of the medical school is also chair of the UMC board. Paradoxically, this can be attributed to the historical fact that the university faculty was in a weak position during the period leading up to UMC establishment, the medical school being in administrative turmoil and financial difficulties. As a consequence, the chairman of the board was appointed dean of the university faculty, establishing a precedent that has been upheld since. In the case of Radboud and UMCG, the dean is appointed as the vice president on the board of the UMC. In all cases, it was felt that by placing the dean on the UMC board, education and research would be on an equal footing with patient care. In reality, UMC board members share and rotate specific responsibilities for each component of the tripartite mission. The extent to which these responsibilities are integrated differ between UMCs; for example, the AMC and Leiden UMC are highly integrated, with a high level of formal university representation in UMC governance, whereas Maastricht UMC (the last UMC to be formed in the Netherlands) has a stronger division of responsibilities and a rather contractual relationship between the university and academic hospital.
In summary, the model of UMC governance that has emerged is one of delegation of powers by the partner organizations to a UMC board, whose focus is the achievement of excellence across the tripartite mission. In some UMCs, this has been accompanied by the combining of the roles of chief academic officer and chief executive officer of the hospital in a single position. However, the underlying legal entities of medical school and hospital remain with separate accountabilities: to the ministry of education and the ministry of health and insurers, respectively. Positioning this within Weiner and colleagues'11 framework, we can observe that the Dutch system has evolved over the past two decades from one in which an “alliance partner” archetype was dominant to one characterized by the “owner” archetype. However, in the Dutch case, de facto common ownership has been achieved by the delegation of powers to an intermediary board, rather than by medical schools directly owning hospitals, or vice versa. This has happened to different timescales across the Netherlands, and the degree of structural integration is less evident in some centers, which may better fit the “coalition” archetype. This reflects a legislative context that is flexible enough to allow a diversity of governance and contractual arrangements to exist. The remainder of this paper explores this diversity at subboard levels.
Management structures and organizational dynamics
With dozens of departments and thousands of employees in each UMC, the implementation of strategy needs to be devolved to a management structure that can coordinate the tripartite mission. As already discussed, board-level governance has evolved as variations on a theme, but below this level many UMCs have been continuously reevaluating their internal organizational structures in an attempt to coordinate and optimize patient care, education, and research activities.
Both Radboud and UMCG are shifting from management through divisions (larger clinical and research groupings of several departments; e.g., diagnostic services) toward management through departments (smaller groups typically based around medical specialties; e.g., imaging, pathology, endoscopy). In Figure 1, a typical structure is shown as including both divisions and departments. In these two UMCs, several divisions were originally set up, each headed by a director.17,18 The rationale for this was the integration of departments and decentralization of decision making from the board to improve organizational performance. However, the underlying departments remained powerful and resisted efforts by divisions to strengthen cooperation and integration, resulting in persistent silos that adversely affected care, especially in the flexible use of facilities and beds. In addition to departmental resistance, the divisions typically lacked the leadership attributes and skills to take on their roles as facilitators and coordinators. In the case of UMCG, original ideas of devolving budgets to the divisions have been scrapped because they lacked the ability to coordinate patient care, education, and research among their respective departments. In the case of Radboud, the divisions are seen as having limited influence, as patient care is managed by the heads of departments, education is managed top-down with disregard to the formal organization, and research is undertaken in groups of informal principal investigators, mainly enabled through external funding.
In these UMCs, the divisions have increasingly been seen as an additional “virtual” layer of governance, complicating accountability and hindering communication between clinical departments and boards. Divisions shirk responsibility over departmental problems, yet the UMC board is held accountable, as they have appointed divisional directors. Interdivisional cooperation is also stifled because divisions did not want to relinquish what little authority they do possess. A recent internal review at Radboud highlighted these challenges, stating that “the informal organization has become too dominant, endangering quality … as roles, powers and responsibilities are not clear.”19
Both UMCs now recognize that the departments, with their critical mass of budget and staff, “run the show” and represent their core business and competences. UMCG is hoping that the divisions gain leadership experience in coordinating and “running a tight ship” and will be able to manage power struggles between the various departments in times of budget cuts. The board hopes this will free up some of their time to discuss issues of innovation with departmental chairs. Until this happens, the board is considering reducing the number of chairs while maintaining an open line of communication at the departmental level.
At Radboud, it was decided that the departments (after merging those lacking critical mass) should report directly to the board. As of 2009, the divisions will be abolished, research will be coordinated in six institutes, and education will be managed by seven directors, while the 50 remaining heads of department are responsible for the management and output of all three parts of the mission. Thus, the board, in addition to maintaining external relationships within a network environment, will assume a very strong and direct position within the internal dynamics of the organization.
In contrast, the AMC's divisional internal governance structure has led to improved coordination between departments, mainly due to the way in which divisions are structured, empowered, and organized. Each of the nine divisions has a staff office and budget and is managed by a board consisting of a chair (professor), a business manager, and a nursing manager.20,21 These divisional boards are responsible for safety and quality of patient care, while quality of research and education is monitored by their respective institutes.
After consultation with departmental chairs, division heads are typically chosen by the board for their demonstrated leadership capabilities. From a specialty perspective, division heads are usually drawn from the more powerful departments and are supported by the business and nursing managers. This structure—having a divisional doctor and nurse—cascades down at all other hospital levels, whether be it wards, outpatient clinics, or departments.21 Placing respected professionals in the lead at divisional levels ensures sufficient emphasis on research, education, and patient care. To avoid organizational “myopia” and the creation of silos, business managers are responsible for funding and staffing multiple divisions (nurses, but not doctors, can be rotated) and are therefore incentivized to maximize global efficiency and effectiveness.
The search for an optimal configuration of clinical management structures is not, of course, unique to academic hospitals, but it is made more complex in this setting by the challenge of integrating education and research with the delivery of clinical services. The degree to which this integration can be achieved through management structures will be directly related to the positioning of a partnership within Weiner and colleagues'11 typology, as will the presence or absence of other integrating mechanisms.
Mechanisms towards optimizing the tripartite mission
Through strong university representation on the supervisory and executive boards, the UMC ensures that policy goals relating to education and research are given equal weight to patient care and that there is no divergence in strategic priorities between the university and hospital. In recent years, universities have insisted on clearer accountability, audits, and more comprehensive reporting on how lump-sum funds transferred from the faculty are used by UMC departments, whose performance is actively monitored and benchmarked.
A further mechanism used by UMCs is the establishment of research and education institutes, which coordinate these strands of mission across clinical and academic departments. Institutes coordinate and monitor multidisciplinary interdepartmental research and education programs and are the mechanism for allocating core research funding and research time, thus providing a platform for individual investigators to seek grant funding, as illustrated for Radboud by Table 1. This allocation is based on the performance of individual researchers and has been criticized as too favorable to established researchers.22
The education institutes generally have a formalized system to plan and coordinate curriculum modules. Quality of education is also maintained through performance-managed competition, whereby various departments and staff compete for teaching allocations. Quality is assessed periodically by a panel of peers and students, which offers advice for improvement and advice on the future allocation of teaching funds. Education is structured around clinical problems, with a focus on applied learning using mixed methods including lectures, tutorials, practicals, clinical demonstrations, and, increasingly, simulation and computer-assisted education.
Staff appointments provide another means of optimizing the tripartite mission. Departmental research appointments are under intense scrutiny not only from their board but also from institute directors and the dean of the faculty.14 Senior appointments are made by the departments from the university faculty, after approval by the UMC executive boards and research and education institute directors. The dean is generally growing in importance in research policy matters and the appointment of professors.
Finally, the university plays a lesser role in discussing clinical matters at the UMC's board level, and it is generally felt that research and education activity at the institute and departmental levels do not require much attention from the board. Therefore, the bulk of the board's time is spent discussing clinical matters. Nevertheless, the board pays particular attention to maintaining the caliber of research and educational staff and is heavily involved in professorial appointments.
Discussion and Conclusions
The key rationale behind the creation of UMCs in the Netherlands was to integrate the governance and administration structures of the different organizations jointly responsible for the tripartite mission. The approach that has emerged, after long deliberation, is one of delegation of powers by the partner organizations to a UMC board, whose focus is the achievement of excellence across the tripartite mission. Universities and academic hospitals remain as separate legal entities in the background with ultimate accountability to separate ministries and regulators. The composition of the UMC board is controlled by a supervisory board, which is appointed by the government (or by the relevant church in the case of the two denominational universities). Both supervisory boards and UMC board of directors include strong academic representation from the university and dean of the medical faculty, respectively. The degree of academic representation on the board, and the position of the dean within it, are generally indicative of the nature of historical relationships between the two organizations and their respective amounts of power at the time of UMC establishment. Below the board level, there is considerable variation and ongoing experimentation in search of the most effective structures and management arrangements for delivering the tripartite mission.
The typology developed by Weiner et al,11 with its three dimensions, provided a useful framework for analysis, although it required modification to reflect the different national context. Compared with the United States, variation in clinical enterprise organization is not great, reflecting the fact that the Netherlands has a national health system (albeit one using diverse insurers and providers) in which UMCs have a specific role. To capture the diversity of approaches in the Netherlands, it also proved necessary to expand the framework to include governance arrangements at different levels in the organization. The degree of transferability of the typology suggests that many of the challenges involved in delivery of the tripartite mission through university–clinical enterprise partnerships are common to different national settings, regardless of differences in national health care systems.
Finally, what conclusions can we draw from this study about theories of organizational design as applicable to academic clinical enterprises? Two interpretations seem possible. The first would be that the relatively limited range of UMC governance structures that have emerged in the Netherlands supports an interpretation of environmental determinism, in which a combination of the legislative framework and the intrinsic challenges of integrating the tripartite mission create an environment in which organizations must have specific attributes, within a quite narrow range of theoretical possibilities, to prosper. A more contentious view might be that what we have observed is a process of “mimetic isomorphism,” in which UMCs have modeled themselves on organizations that are perceived to be successful, looking both at pioneers within the Netherlands (most notably the AMC) and internationally. It is certainly the case that the academic health sector exhibits many characteristics identified by theorists as likely to lead to isomorphism, including uncertainty about the relationship between ends and means, ambiguous goals, and reliance on academic credentials in choosing personnel.23 Fully exploring the validity of these different interpretations is beyond the scope of this article, but the material presented here does suggest that a more theoretically informed study of academic clinical organizations, both within and across nations, would be a rich seam of inquiry.
The authors wish to thank Prof. Eduard C. Klasen (dean and member of the board of directors at Leiden University Medical Center) for his kind advice and facilitation of the study.
Direct expenses of this study were met from funding by the Commonwealth Fund of New York City, originally awarded to Mr. Davies as a Harkness Fellow in 2001.
The opinions expressed in this article are those of the authors alone and do not reflect the views of their employing organizations.