Decentralising result responsibility: Using responsibility centres in University Hospitals

(1)

Decentralising result responsibility:

Using responsibility centres in University Hospitals

By

René Poel

University of Groningen Faculty of Economics and Business

MSc. Business Administration Organizational & Management Control

René Poel s1796003 Verlengde Visserstraat 9a 9718 JA Groningen 06-11518170 r.poel@student.rug.nl Supervisor: B. Crom January 2014

Abstract

Purpose – The study aims to contribute to literature by exploring the way in which hospitals adapt internal control systems to changed external funding, especially focussing on the use of responsibility centres. Furthermore, it provides a management control design based on responsibility centres for a University Hospital in the Netherlands.

Methodology – A field study approach was conducted at 6 hospitals within the Netherlands, implications were drawn for the hospital of the case.

Findings – The paper identifies the most appropriate types of responsibility centres, accompanied by the preferred level of decentralisation. Furthermore, it identifies typical disadvantages to the use of mono-disciplinary responsibility centres in hospitals from which the main factor was found to be the reduced stimulation of multi-disciplinary collaboration.

Managerial implications – The study provides a comprehensive overview of the use of responsibility centres within the context of university hospitals, based on which a design of a management control model is suggested.

(2)

2

Introduction

A significant part of the economic activities in the Netherlands is concentrated in the health care sector, making this sector of great economic importance for the national economy. Whereas our understanding of management control in this sector is limited (Groot, 1999), additional knowledge of problems and issues within this sector, may contribute to welfare of society. Literature seems to highly concentrate on private sector firms (Frey, Homberg & Osterloh, 2013). There seems to be disinterest for public sector organizations by management scholars (Mahoney, McGahan, & Pitelis, 2009), whereas insights derived from the analysis of public sector organizations could also be of relevance for management in general (Benz & Frey, 2007).

During the 1990s, in what has become known as the ‘new public sector’, many services in advanced economies, have come under pressure to become more efﬁcient and effective, so as to reduce their demands on taxpayers, while maintaining the volume and quality of services supplied to the public (Brignal & Modell, 2000; Schippers, 2012). Groot (1999) distinguished between two basic doctrines of New Public Management (NPM): emphasising accountability for results and introducing more private sector management styles and techniques. The first doctrine leads towards a more ‘accountingization’ of public administration i.e.; the attachment of monetary values to objects (Pollit, 1993 & Power and Laughlin, 1992). The second doctrine introduces a shift towards a greater disaggregation of public organizations into units which are managed using practices derived from the private sector (Hood, 1995). Countries with predominant state health care have been searching for appropriate and effective ways to organize the health care sector (Schut, 2009). The implementation of New Public Management in the Dutch health care sector in 2005 was the attempt of the Dutch government to accomplish the goals of NPM within this sector.

Several scholars have investigated ways in which NPM can be introduced. The funding system of Diagnosis related Groups (DRG’s) or as used in used in the Netherlands, Diagnose Behandel Combinaties (DBC’s) seemed to be the most appropriate way to provoke competition in the sector (Kok et al., 2010). The DBC’s were introduced into the Dutch health care sector in 2005, replacing the old funding system in which care organizations dominantly depended on budgets from the government. The new system comprises a package of the total care per patient. There has been much debate about the appropriateness of this system (Kreis, 2006), however changes in the system seem to get into the right direction, making this the new external funding system for the Dutch health care sector (NZA, 2012).

(5)

5 have to act more like entrepreneurs, like private organizations. His research focusses on the consequences of the changed external funding on internal direction and control within hospitals. Rising costs and the subsequent increasing cuts by the government cause hospitals to focus more and more on management control systems (Crom, 2005:106).

Problem identification

The introduction of the new external funding system (currently named DOT’s), creates pressure on financial resources. Increasing social responsibilities create alters uncertainty (Asselman, 2007). The absence of a clear relation between costs, production and revenues makes it difficult for managers on different levels to agree upon the appropriate levels of budgets or appropriate results (Asselman, 2007). The new way of external funding implicates that prices are generated by markets, which increases the importance to closely monitor costs and adapt internally if needed (Schaepkens & Zuurbier, 2002). More formal, financially quantifiable information is needed to assist this managerial decision making (Hertong et al. 2005). As stated in the annual report of the case hospital in 2012: external funding becomes more variable, revenues are becoming more dependent on performance.

(6)

6

Research question development

The lack of tuning of management control to external funding methods creates control issues between controllers, medical specialists and managers in these hospitals. Currently, historical budgets are used in which no links exists between external funding and internal budgeting. This causes discussions between specialism about their financial contribution to the hospital, based on which some specialisms demand more benefits. Another implication is that changes in funding for a certain treatment can’t be translated internally in budgets or other control mechanisms.

Many hospitals implemented responsibly centres as an answer to the more performance driven way of external funding (Standaart, 2010). However literature doesn’t provide a comprehensive description of the way in which this could be done in the context of a university hospital. Current literature on responsibility centres highly focusses on the general business context, however functions and tasks within hospitals are different to such a degree that research in this area is necessary. The lack of research implicates ambiguity regarding the organisational structuring and grouping, the kinds of relevant results, the factors implicating a certain type of responsibility centre and so on. Hence, this research focusses on answering the following question:

How should a university hospital use responsibility centres, to better tune the organisation with external funding?

The following sub-questions aim at answering this research question:

 At which organisational level should responsibility centres preferable deployed?

 Which types of responsibility centres are appropriate in the hospitals context?

 Which factors should determine the most appropriate type of responsibility centre?

 To what extent can non-financial performance indicators be implemented in this system?

 Which disadvantages are typically associated with the use of responsibility centres in the context of hospitals?

 Who should be responsible for a responsibility centre?

Theoretical and practical relevance

(7)

7 university hospitals design their management control systems using responsibility centres. The focus will be on identifying the results and responsibilities within the system. In addition to this, the focus will be on the disadvantages associated with the use of responsibility centres which are typical in the hospital context. It could for instance be that the disadvantages are detrimental to such a degree that the use of responsibility centres is not preferred within hospitals. This study contributes to theory in several ways. First, it provides a comprehensive overview of relevant literature in the field of responsibility centre use in hospitals. Second, by researching the use of responsibility centres in the particular context of a university hospital. Third, by researching the way in which nonfinancial performance indicators can be incorporated in the design of a management control system using responsibility centres.

(8)

8

Literature

This section will provide a deeper understanding of the relevant concepts within this research. First, the general concept of management control will be introduced. Second, literature on the types and criteria for responsibility centres will be presented from both a financial and an organisational perspective.

Management control systems

Management control defines the broader concept of techniques used to assist and help coordinate the process of making and monitoring the decisions within the organisation such as results control, action control, personnel control en cultural control (Horngren and Foster, 1991). It comprises all devices that can be used by senior management to control behavior of employees and managers (Merchant and Van der Stede, 2007; Malmi and Brown, 2008). MCS are conventionally perceived as passive tools providing information to assist managers, however, approaches following a more sociological orientation see it as more active, furnishing individuals with power to achieve their own ends (Chenhall, 2003). The definition of MCS has evolved from one focusing on the provision of more formal, financially quantifiable information to assist managerial decision making to one that embraces a much broader scope of information. This includes external information related to markets, customers, competitors, non-financial information related to production processes, predictive information and a broad array of decision support mechanisms, and informal personal and social controls.

Management Accounting Systems (MAS) are a key part of MCS. MAS are used to source organisations with information and to delegate accountabilities and responsibilities (Karstberg & Siverbo, 2013). Personal and clan controls are also part of the management control system definition (Chenhall, 2003). However, Malmi and Brown (2008) argue the importance to define key elements of the MCS definition. Before the different uses of MCS are discussed, first the definition requires further elaboration.

(9)

9 This study mainly focusses on what Malmi & Brown (2008) mention as, cybernetic controls like; financial measures, non-financial measures (Ittner & Larcker, 1998) and hybrids (Greenwood, 1981; Kondrasuk, 1981; Ittner and Larcker, 1998; Kaplan and Norton, 1992, 1996a,b, 2001a,b; Malina and Selto, 2001). The system of responsibility centres implicates that managers are held accountable for specific (mostly financial) results.

Types of responsibility centres

Research on responsibility centres, especially regarding the way in which responsibility centres are used in Dutch hospitals is relatively scarce (Standaart, 2010). Responsibility centres are used as a common feature developed to address a need for managerial evolution (Vonasek, 2011). Wissema, (1989) mentions it to be a management style and organisational design aiming at decentralization of entrepreneurship, used as a common feature in many for many years now (Melumad et al., 1992). The use of responsibility centres features both academic authority and fiscal responsibility departments/disciplines to instil a more entrepreneurial management culture (Linn, 2007).

Scholars referred to this management control system in general, using several corresponding concepts like: responsibility centre management, value centre management, decentralised budgeting, activity-based budgeting, value responsibility budgeting, and cost centre budgeting (Priest, Becker, Hossler, & St. John, 2002).

A base for introducing responsibility centres is the assumption that the subordinate managers is able to control the relevant results (Kruijf and Mol, 2007; Anthony & Young, 1999), also acknowledged as the controllability principle (Bhimani et al., 2012). Scholars have identified different types of responsibility centres, in which financial responsibility centres are dominant (e.g. Higgins, 1952; Prince, 1975; Kaplan and Atkinson, 1989; Melumad et al., 1992; Maddox, 1999; Bhimani et al., 2012). Literature on responsibility centres highly focusses on financial perspectives, which is why Drury (2004) and Mol (2006) argue that responsibility centres in service organisations should be complemented with other controls besides the financial instruments in order to judge quality and other aspects relevant to service organisations.

(10)

10  Cost centres – managers are held accountable for costs.

 Revenue centres - managers are held accountable for revenues.  Profit centres - Managers are held accountable for costs and revenues.

 Investment centres Managers are held accountable for costs, revenues and investments. Maddox (1999) also discusses a time dimension. He argues that profit centres are ‘units known to produce positive net revenue consistently’. Cost centres therefore might be able to generate revenue occasionally. If it is not able to produce revenues in excess of its expenditures on a regularly basis, it should be implemented as a cost centre. The degree at which these responsibility centres are deployed is not described. Responsibility centres can be found in either of extremes of centralised and decentralised organisations. Responsibility centres should preferably be deployed at the level at which results can be influences. This often means that responsibility centres are deployed at lower levels within organisations. (Bhimani et al., 2012). In costs centres, the centre’s manager has discretion over ‘internal’ decisions. These decision rights might consist of production assignments, make-or-buy decisions, input sourcing, etcetera. Most often, the general objective for the manager of a cost centre is to minimize costs, and therefore the performance evaluation likely focuses on costs achieved.

Responsibility centres in Dutch hospitals

(11)

11 responsibility centres through which decision rights and accountability are decentralised (Horngren, Datar, & Foster, 2006). The focus in these hospitals primarily shifts towards financial measurements. Management contracts and service level agreements (SLA’s) are used to incorporate both financial and nonfinancial responsibilities into goals, budgets and indicators (Duringer, 2010). These contracts are mostly based on the perspectives of the balanced scorecard.

As mentioned earlier, literature regarding the way in which responsibility centres are currently used or should be used is scarce. Non-academic books are available through, which indicate criteria and organisational designs in Dutch hospitals. The remainder of this section will be partly based on research from Durlinger (2010) and Zuurbier & Hartmann (2011).

Responsibility centres are generally structured based on distinct specialisms within hospitals (Standaart, 2010) or a clustering of activities within a specialism (Crom, 2005; KPMG, 2011) which are held accountable for tuning costs to revenues. Responsibility centres make it possible to bring professional and financial responsibilities closer to each other by providing clear relationships and responsibilities (Standaart, 2010). Chu et al., (2002) argue that the use of responsibility centres in hospitals positively influences the efficiency of departments, resulting in improved overall hospital performance.

The majority of literature focussed on the use of responsibility centres within regular business organisations. Hospitals however have multiple and very different functions. These facilitating and primary care providing functions are mainly organised in departments. Another key characteristic is found in the collaboration between specialisms (Kastberg & Siverbo, 2004). This calls for an organisational design structure which is suitable for the specific functions in hospitals. The two main perspectives can be distinguished when grouping tasks and activities: market and functional grouping. Market based grouping suggests grouping around the basis of output, client and place, whereas functional grouping comprises the bases of knowledge, skill, work process and function (Mintzberg, 1983:193). Durlinger (2011:66) describes the mono-disciplinary and the multi-mono-disciplinary responsibility centres which essentially are based on either functional and market based grouping.

The mono-disciplinary responsibility centre

Mono disciplinary responsibility centres are based on one particular specialism. Within this type of centre, the policlinic, diagnostics and clinics are grouped around and are functioning for this specific specialism (Mintzberg, 1983). Bed capacity of other responsibility centres will only be used when necessary (Durlinger, 2011).

(12)

12 this specialism, however additional examination or the use of common capacity forms an exception to this (Durlinger, 2011).

Care products are primary organized around a particular patient category, implicating that result responsibility of this patient category is delegated to this specific mono-disciplinary responsibility centre. Exceptions for this way of organizing result responsibility will be limited in practice. Alignment between specialists can be organised using care paths or care programs. Most hospitals choose to create responsibility centres in a mono-disciplinary way in order to better align internal control to changed external funding (Christis, 2011). Another factor which could promote the decision to deploy mono-disciplinary responsibility centres could be the fact that goal congruence highly influences the design of a management control system (Flamholtz et al., 1985; Zimmerman, 2001 and Malmi & Brown, 2008). These authors argue the necessity to guard against the possibilities that people are directed towards doing things with unambiguous interests. Interests within disciplines however, are likely to be congruent.

The multi-disciplinary responsibility centre

Many scholars argue the need for a focus on process oriented health care (e. i.: Christis, 2011; Kastberg and Siverbo, 2013), in which multiple disciplines are represented. Their main argument is based on the fact that patients mostly need more than one specialism. This process or patient oriented approach requires inter-disciplinary alignment, making it necessary to collaborate with related disciplines. The focus in process oriented healthcare is on common responsibilities for patients and patient flows (McNulty and Ferlie, 2004). However, only a small part (approximately 15% to 30%) of patient care requires multi-disciplinary alignment (Durlinger, 2011). Different and inconsistent arguments can be found in literature regarding the distribution between mono-disciplinary and multi-disciplinary care.

The use of a multi-disciplinary responsibility centre makes it difficult to delegate result responsibility at a particular specialism within one multi-disciplinary responsibility centre, because all specialisms are dependent on each other to effectively provide care for patients. The size of affiliating patient streams determines whether a multi-disciplinary responsibility centre should be used. The more care pathways can be grouped, the more interesting it becomes to use multi-disciplinary centres.

Multi-user responsibility centres

(13)

13 traceable to their activities, whereas their activities are mostly part of the DBC’s of port specialisms. A major part of funding is done through these port specialisms. The support function of multi-user centres implies a dependency of mono-disciplinary and multi-disciplinary responsibility centres (Durlinger, 2011).

Multi-user responsibility centres can be compared with what Strikwerda, (2004) calls Shared Services Centres. These types of centres within organisations aim at: reduction of costs associated with decentralisation, accompanied with stimulating quality, professionalism, cost flexibility of supporting services. Shares Services Centres have measurable outputs with costs per unit of service provided. The next section will elaborate on the relevant criteria used for responsibility centres within hospitals.

Criteria for responsibility centres

As mentioned earlier, contingencies determine whether which management control systems are appropriate in particular manifestations (Merchant, 1984). For responsibility centres, there are no clear criteria of which scale requirements are appropriate and which production processes lend themselves for the employment of responsibility centres (Zuurbier & Hartman, 2010). The term responsibility centre can be regarded as an all-purpose word in practice, implicating that it is hard for managers within the sector to determine whether definitions between managers are used in a congruent manner. Zuurbier & Hartmann (2010), summarize success criteria for responsibility centres in hospitals based on general literature in this field. They argue that profit centres are most appropriate in the hospital context and mention four criteria for these responsibility centres. The more these criteria are met, the higher the chance of continuity and efficiency of responsibility centres. The following criteria are mentioned: (1) amenability of integral costs, (2) measurability of output, (3) uniformity in control and (4) acceptation of internal competition. The chance of continuity and efficiency are likely to increase the more a responsibility centres meets these requirement.

Amenability of integral costs means that the management of a responsibility centres is ‘in control of’ integral costs. The principle of controllability states that managers should only be accountable for items within their control (Ferrara, 1963; McNally, 1980; Roodhooft, 1995; Selto, 1988). Implications for responsibility centres in hospitals are a need for traceability of costs to processes and the availability and transparency of cost information, also from indirect costs (Zuurbier & Hartmann, 2010).

(14)

14 direct monitoring and supervision of employees (Anderson & Oliver, 1987), however the output should be clearly defined in the hospital. This could be difficult, because outputs in hospitals are often ambiguous (Burgess & Ratto, 2003) and unclear (Rainey, Backoff, & Levine, 1976; Cutler & Waine, 2005). Also, many tasks are highly interdependent, making attribution of costs difficult (Frey, Homberg & Osterloh, 2013).

Uniformity in control means that the goals of the responsibility centre are clear and that goals congruence exist between goals (Ferrera, 2013). This makes it more clear where everyone can be held accountable for.

Lastly, the introduction of responsibility centres creates competition (to some extent) between responsibility centres (Choudhury, 1986). This internal competition should be accepted internally to be able to improve management control in the hospital (Zuurbier & Hartmann, 2010).

Results responsibility

As mentioned earlier, one focus of this study is on the identification of disadvantages associated with the use of responsibility centres in the hospital context. Decentralising responsibilities draws specific implications for organisations, both positively and negatively (Prud’homme, 1995). Agency theory can be helps us in explaining the potential implications associated with decentralized result responsibility for both agents and principles (Eisenhardt, 1989). It is helpful to further elaborate on and clarify the implications for decentralization, especially in the hospital context where result responsibility is delegated to both managers and medical specialists.

Responsibility centres from an agency theory perspective

The agency theory, like transaction cost theory, can be seen as an economic organisation theory in which principals and agents are central stakeholders (Jensen & Meckling, 1976). The agent executes commands for the principal based on a contract (Melumad et al., 1992). An important task for the principal comprises the communication of his interests. Examples of these interests in hospitals can be seen in the transfer of financial, quality- and safety risks that come together at the board of directors (Zuurbier & Hartmann, 2011). Incentive structures and control systems are used by principals to direct the agents behavior in the right direction based on incentives. A variety of types of management controls could be incorporated to direct behavior (Malmi & Brown, 2008).

(15)

15 (Cvitanik et al., 2013). The problem of adverse selection lies in the fact that the agent possesses information which is not available for the principal, making it difficult for the principal to judge whether this information is collected in his interests. The moral hazard problem can be explained by the fact that the principal is confronted with the risk that performance evaluation of the agent is based on incorrect choices. This could be a problem, whereas the principal is only able to perceive these results and no actions. The principal has to judge whether results are based on the environment or by actions of the agent. Agency theory assumes that principals are remotely involved into daily activities in the organisation of the agent. Agents will possess more information about activities and contradicting interests make that agents are unlike to lift this information asymmetry. Principals will introduce monitoring using management information or responsibility information to gain relevant information (Jensen & Meckling, 1976).

The agency relations can be seen in hospitals between hospital management and department/responsibility centre management. Internal department budgets can be seen as informal contracts between these two stakeholders. Zuurbier & Hartmann (2010) summarise the benefits and risks as follows for hospitals:

Bhimani et al., (2012) add the benefits of management development and learning. Provision of more responsibility to employees is likely to promote the development of an experienced pool of management talent.

Vonasek (2011) argues that responsibility centres apply principles aiming at enhancing flexibility and adaptability of organizations. Strauss, Curry, & Whelan (1996) and Whalen, (1991) elaborate on these principles. Responsibility centre budgeting is likely to enhance decision-making authority by managers; ‘moving decisions closer to the locus of their impact makes them accountable and decisions more pragmatic’, by decentralisation of responsibilities (Vonasek, 2011: 499). The second principle comprises the development of an organisation-wide decision information system; ‘when both administration and managers cooperatively deﬁne the information to be produced and have access to the same reports, communication between them is improved’ (Vonasek, 2011: 499). Another principle is that responsibility budgeting provides clear decision rules and understandings of rewards and sanctions reduces decision response times (Vonasek, 2011).

(16)

16 Processes like performance evaluation can be enhanced, using responsibility centres, whereas managers can be held accountable for results of their specific centre which are controlled by them. Also a sense of group commitment is created using responsibility centre budgeting; ‘each responsibility centre is aware that the success of the institution is dependent upon their individual and collective success and understands the authority, responsibility, and resources assigned to them reﬂect that’ (Vonasek, 2011: 499).

Agents in hospitals

In hospitals, knowledge about the right way to perform tasks lies in lower hierarchical layers close to the patient, which increases the chance of better decisions and higher motivation. The choice between more or less decentralization should be based on consideration of these benefits and risks, or solutions to the risks (Zuurbier & Hartmann, 2010). However, these benefits can only be exploited when the right persons are held accountable and when balance with performance measurement and compensation exists.

Decentralisation implies a high degree of freedom of action of therapists, nurses and managers, which raises the question who has which authority to make decisions and to which persons responsibility should be delegated. Current literature on responsibility centres assumes general management roles as leaders of responsibility centres (Bhimani et al., 2012), however responsibilities within hospitals are delegated to both medical specialists and general managers using dual management structures (Pool, 2006). Literature provides an extensive debate about the differing interests of medical professionals and directors and managers in care organizations (Wanrooy, 2001 and Weggeman, 2008). Porter, (2006) expresses this dilemma in his book: The real proof of success is better patient results (quality versus cost), not compliance with processes specified by outside experts or administrators. Hospitals have a long history within this tension between medical professionals and directors, and this eventually limits the control possibilities, partly influenced by the clinical responsibility of physicians (Standaart, 2007). This makes it necessary to elaborate on the agents within hospitals with regard to a decentralised structure.

(17)

17 strands. The first strand consists of those that retain traditional arrangements and have a large contractual relationship between the hospital board and the ‘partnerships’ (representing doctors). Here doctors in the partnership may undertake managerial work in relation to their firm, but remain disconnected from the hospital as a whole (Neogy and Kirkpatrick, 2009). This form is mainly seen in general hospitals.

Second, hospitals are experimenting with ‘Physician in the Lead’ models with departments based on specialisms, that are typically managed by a clinician (along the lines of a clinical directorate model). This is usually with some (often limited) responsibility for budgets. Some hospitals are experimenting with divisional style structures managed by clinicians, whilst others retain traditional partnership/contractual arrangements that keep doctors at arm’s length from hospital governance (Neogy and Kirkpatrick, 2009). Divisional structures are often seen in university hospitals (UMC’s) in which specialisms are bundled based on functional affinity. These divisions are usually directed by a medical department manager, a nursing manager and a business manager (Durlinger, 2011). Responsibility for results can be either evaluated within a division or directly with the board of directors.

Brink et al. (2010) argue that dual management by both a business manager and a medical manager is key in implementing responsibility centres. Organisational and professional orientations are required to achieve the goals within a responsibility centre (Moller & Kuntz, 2013). Durlinger (2011) distinguishes different control models that are appropriate for the different types of responsibility centres. The mono-disciplinary responsibility centre, for instance is preferably controlled in a dual model: a combination of a medical specialist and a business manager. Within this situation, a choice can be made whether the medical specialist is held accountable for the results of the responsibility centre. Responsibility in dual management structures can be either based on one two competence-profiles that apply to both the specialist-manager and the business specialist-manager. The case of one profile, leaves room for mutualisation of tasks without detraction from joint integral result responsibility. Subcategories of the balanced scorecard are usually split up in situations in which two profiles are set. A medical manager can for instance be responsible for the client perspective (i.e. patient satisfaction), while a business manager is responsible for the employee perspective (i.e. employee satisfaction) (Zuurbier & Hartmann, 2011: 118).

(18)

(19)

19

Methodology

This section describes and gives an overview of the methodology used to generate insights for this particular case in practice, which helps us to answer the research question. The methods used and the data collection procedure will be explained. The process of data collection will be further elaborated and concluding a description of the case will be provided.

Research design

A qualitative problem solving research was conducted using an empirical field research method which emphasises on understanding the context of the subject, allowing for richness of understanding (Cooper & Schindler, 2008). The exploration of phenomena within a contemporary context is facilitated by using a qualitative research method, making it the most appropriate method for this study (Eisenhardt, 1989; Yin 1989, 1993, 1994; Baxter and Jack 2008). Additional to this, the major focus of this study is more on the description and explanation of phenomena than measurement and prediction, and therefore a qualitative approach seems most appropriate for this research (Fitch, 1994a). Data were acquired at other hospitals for several reasons: (1) to gain deeper understanding on the way in which other hospitals design a management control model using responsibility centres and (2) to gain deeper understanding in the specific barriers and facilitators to successful use of the system.

As stated before, this study focusses on the design element of MCS within this university hospital. This study focusses on the first three phases as defined in the regulative cycle by van Strien (1997). These phases comprise problem definition, analysis and diagnosis and finally a solution-redesign will be made.

Sampling

Primary data about the current way in which internal control is applied in one of the University Medical Centres (UMC’s) in the Netherlands were collected from between October and January 2014. A variety of hospitals other hospitals were approached to acquire relevant additional data. This data were collected in 6 other hospitals in the Netherlands. Controllers and

(20)

20 general managers in both, academic hospitals (4 in total) and general hospitals (2 in total) were interviewed to collect data. Interviews were conducted by two researchers, including the author whereas the second researcher also used part of the data for own research purposes. Interview data was supplemented with (internal) documents which were acquired after the interviews and with annual reports.

The selection of hospitals was based on the characterization of the organisation, either general hospitals or university related hospitals. Two general hospitals were included based on a proposed differing perspective on responsibility centres. Respondents were selected among various dimensions (including rank, departments, hospitals) to find variation in data, as recommended by multiple authors (e.g. Strauss, 1987; Corbin and Strauss, 2008). In total, 9 in-depth interviews were conducted that lasted between 40 and 65 minutes. The selection method is acknowledged as theoretical sampling. Respondents could be mangers or controllers (or other comparable functions) involved on a departmental level, sectorial level, or centre level who were involved in the design process with regard to responsibility centres. Another factor taken into account was the geographic distribution of these hospitals in contemplation of potential biases due to geographical differences.

Data collection

The main source of data collection for this research were in-depth semi-structured interviews with relevant stakeholders within the hospital. A semi-structured interview approach was used, while using a list consisting of specific questions and at the same time leaving sufficient room for additional information (Van Aken et al., 2012). Semi-structured interviews have advantages on getting insights in a problem and in potential relations. This main source of data is, if possible, backed up by data from observations and records to provide more stronger substation of constructs and hypotheses (Eisenhardt, 1989). The interview protocol, as recommended by Bryman and Bell (2003), was developed by the researcher and can be found in appendix A. In total 7 interviews were conducted which lasted about 55 minutes.

Qualitative data analysis

(21)

21 Charmaz, (2006) recommends the conduction of a coding process consisting of the three stages of open, axial and selective coding. This method takes codes from data analysis which are subsequently compared with categories that emerge around certain topics (Hutchinson, 1993). The text “Deployment of responsibility centres induces the focus within specialisms” for instance was coded as `internal focus´. Atlas.ti software was used for the coding process, in which the different organisations were separated using numbers (Muhr, 2004). Open coding consisted of a categorisation of comparable texts. Axial coding included the investigation of categories within their conditional context. During the second stages, relations and connections between categories could be identified. This way a model could be created that details conditions that give rise to a phenomenon’s occurrence (Brown and Stevens, 2002). The third stage of coding entails the selective coding process. Corbin and Strauss (1998) argued that this stage consists of refining and integrating categories around a central explanatory concept or category.

As mentioned, secondary date were used in the appearance of documents. A variation of documents were analysed, consisting of internal reports, annual reports and memos. Documents were examined, categorised and compared, based on which suitable documents were chosen to serve as a basis for the analysis of the current situations within hospitals. Aspects like organograms, performance indicators and other relevant data were analysed.

(22)

22

Results

This chapter is divided in three parts. First, a general description of the case hospital will be provided, supplemented with the present design of the management control system. The focus will be on the current reporting structure, the organisational structure and the budgetary process. Second, the findings from the analysis of documents and the interviews will be presented. Thereafter, a new design using responsibility centres will be presented. The findings and the design solution will be described based on the topics concerned in the research questions.

Introduction of the hospital

The case being analysed for this research concerns the management control system of one of the University Medical Hospitals (UMC) in The Netherlands. This hospital consists of over 10.000 employees which work in the areas of patient care and academic research for which they closely collaborate with a university. The core tasks of this hospital comprise: patient care (academic care, top-referent care¹, top-clinical care², and regular patient care), education and academic research.

The current control situation

The divisional structure is deployed in the current organisational structure in which specialisms which are related or close to each other are grouped (Appendix B). These sectors are directed by one sector director, which is assisted by two staff directors which are responsible for HRM and management control. These directors are assisted by managers which have direct contact with a specific departmental manager. Departments are managed by a dual management consisting of a departmental manager (a medical specialist) and a business manager.

The board of directors specifies budgets for the specific departments based on historical budgets. The following perspectives are taken into account on corporate level: financial, client, internal process and learning and innovation performance indicators. The board also specifies non-financial results for the total sector. Departments can decide which indicators to include within their reports. Results of departments are evaluated within sectors, between department management and sector management. Facilitating services are grouped within sector E. Cost prices of facilitating services are translated into budgeted costs for the profit centres.

(23)

23 that the department manager reports to the board of directors, whereas the manager reports to the sector director. This implicates that the departmental manager is able to arrange things directly with the board of directors, without sectorial boards discretion. Differences in the way in which a sector director organises the structure can be seen between sectors.

As mentioned, no direct linkage exists between external funding and internal control; external funding and internal budgeting are highly separated. Budgets are based in historical data and adjustments are made based on changes in volume. Price changes are compensated in the budgets. This lack of transparency causes discussions about budgets between departments. Some specialisms are convinced that their department contributes a major part to revenues of the hospital which makes that they also expect to benefit from this in a financial manner. The absence of this link implicates that it is impossible to ‘really’ perform control based on loss-generating activities on variances in revenues based on negotiations with care insurers.

¹ Top-referent care is specialist patient care which is associated with special diagnosis and treatment,

making it the ‘last resort’.

(24)

24

Findings from data collection in other hospitals

In this section, an analysis will be provided on the degree to which and how responsibility centres are used in current organisational structures in hospitals.

The focus will be on the aspects of (1) the level at which responsibility centres are employed, (2) the characterization of these centres, both financial and organisational (3) the result responsibilities within these centres, (4) the use of non-financial result responsibilities and (5) the disadvantages associated with responsibility centres.

A demarcation must be made between the types of hospitals which were visited, as one of the respondents mentioned. General hospitals are structured around partnerships, bringing along some implications for management control:

Controller General Hospital 1: ‘’Partnerships are likely to have different interests from the perspective of the hospital. These interests are aligned in an ideal situation, however in practice, this is often not the case.’’ and ‘’General hospitals only perform regular patient care and more managerial terms are more applicable for these hospitals.’’

In total, two general hospitals and four university hospitals were visited. Within these hospitals, differences were seen in the degree of use and differences in the way of use of responsibility centres.

The definition of a responsibility centre

Respondents were asked for their definition of responsibility centres. The first noteworthy finding is the difference in the definition of a responsibility centre between hospitals. A respondent in an university hospital stated that he saw responsibility centres as a ’all-purpose word’. He stated:

Controller University Hospital 1: ‘’I find it difficult to define the term responsibility centre. In my definition, a responsibility centre could be everything, I’m also result responsible but does that make me a responsibility centre?’’

Whereas a controller in General Hospital 1 provided the subsequent statement:

“The centres as used in this hospital function as separate business units, almost like BV’s.”

(25)

25 Organizational level at which responsibility centres are deployed

The most common organizational level of at which hospitals in this study use responsibility centres was found to be at the level of (port)specialisms. These centres are preferably organised as distinct departments within the organisation. Capacity units like operating rooms and labs are also used as responsibility centre. The majority of respondents agree that this lowest level within the organisation is the most appropriate and natural. A respondent in a university hospital stated:

Controller University Hospital 1: ‘’In our management philosophy, specialisms are the most natural entity to deploy as responsibility centres. It is possible to create a divisional structure, using specialisms that are highly related, however conflicting interests are likely when responsibility centres are deployed at such levels”.

A respondent in a general hospital which originally used a clustering mentioned:

Controller General Hospital 1: “We used to have care clusters, a bit like the multi-disciplinary philosophy. We decided to split things up due to a lack of innovation and an conflicting interests”.

The number of responsibility centres in hospitals within this study ranges from 39 in a relatively small general hospital to 62 departments in a university hospital. Some hospitals cluster specialisms which are related for the purpose of controllability for the board of directors, whereas responsibility centres or departments directly report to the board of directors:

Controller University Hospital 1: “We aggregate related pieces to make it manageable for the board of directors…”.

Aggregation in this sense means the grouping of activities, the addition of an hierarchical level. University hospital 3 chose to create multi-disciplinary centres. Examples of dominant design criteria were: strategic focus topics, patient oriented, functional affinity, comparable sizes and the medical specialist as responsible director. The choice for multi-disciplinary centres in this case means more a higher level of centralisation. Result responsible teams are positioned under the responsibility centres and report directly to centre management. For instance, within the Cardiovascular Centre, 12 teams are formed around specific syndromes like heart failure or arrhythmias. Performance of these teams is measured using an operational scorecard.

Characterization of the responsibility centres

(26)

26 disciplinary, multi-disciplinary etc.). Hospitals within this study prefer the use of profit centres, if possible. Bhimani et al., (2012) state that managers of profit centres should be held accountable for both costs and revenues. One respondent indicated:

Controller General Hospital 1: “Profit centres are preferable deployed when both costs and revenues can be traced to a certain treatment, e.g. specialism”.

Controller University Hospital 4: “Externally focussed specialisms generate the revenues and the costs associated with treatments, which is why we deploy them as profit centres”.

Hospitals in this study use both mono- or multi-disciplinary centres combined with the use of multi-user responsibility centres. Multi-user responsibility centres are mainly seen as labs, operating rooms and shared service centres. Multi-user centres are arranged as cost of profit centres. University hospital 4 choses to make a distinction between a focus on internal and externally focussed departments. Multi-user centres are arranged as internal departments and are funded by demand from externally focussed departments.

Prerequisite for the use of as a profit centre is the knowledge about and traceability of costs and revenues per treatment. Most hospitals are moving or prefer a shift towards a system in which departments are used as profit centres. Hospitals are used to base budgets on historical information. The degree to which cost and revenue information is available varies between hospitals. One respondent of a university hospital indicated:

Controller University Hospital 1: “We do have a diversion system based on which DBC revenues can be traced, we are at the starting point of confronting real costs with real revenues. The philosophy has been that medical professionals should be bothered with such information and tasks associated”.

This university hospital was not able to structure the organisation around responsibility centres. The general hospital, however also indicated problems with the allocation of DOT revenues to specific responsibility centres:

Controller General Hospital 1: “The translation from DBC’s to the centres could easily be done, making it possible to use profit centres. The transition to DOT-funding, however made it impossible to allocate revenues, which constrained us to the use of cost centres”.

Respondents indicated registration of care products as key facilitators for success when using responsibility centres.

(27)

27 and the data provided in the interviews. Recurrence of some subjects in the analysis of both these data sources stimulated the decision to incorporate these factors in this section.

Disadvantages associated with responsibility centres

Theory indicated the potential disadvantages resulting from decentralisation of organisational units, as described by agency theory. The focus of this study is however not to demonstrate the most common problems associated with decentralisation, but more on the extraction of factors that are typical and potentially detrimental in the context of a hospital. Subsequent factors could be extracted from the interview data and an analysis of the documents:

 Reduced stimulation of multi-disciplinary collaboration

Based on theory, two main perspectives regarding the structuring of responsibility centres could be distinguished, mono-disciplinary and multi-disciplinary centres (Durlinger, 2010). Several questions within the semi-structured interview focussed on these two perspectives to extract why choices for a certain perspective are made. The majority of respondents mentioned difficulties for collaboration between specialisms within a care path or within multi-disciplinary care centres when using mono disciplinary responsibility centres.

Controller university Hospital 4: “Deployment of responsibility centres induces the focus within specialisms, which is potentially impeding collaboration between specialisms”.

The focus within mono-disciplines makes that responsibility for results within care paths is not always clear and interests are ambiguous. A struggle between the two approaches to organise care in a mono-disciplinary versus a multi-disciplinary perspective is clearly visible in all hospitals. This struggle is commonly seen as a choice between these two perspectives:

Controller University Hospital 1: “From my personal perspective as a controller, I see it as a choice which has to be made. Budgetary and financial responsibility should not be mixed in departments and horizontal care centres, then we would have a kind of matrix structure, making it hard to maintain an overview”.

 Span of control of the board of directors

(28)

28

Controller University Hospital 1: “We introduced divisions in order to keep things controllable for the board of directors”.

Hospitals make choices between divisions, sectors or themes as a way of clustering activities. The choice for a clustering of specialisms seems to be contingent upon the amount of departments, whereas larger university hospitals consisting of 60 or more departments choose to cluster activities while the smaller hospitals didn’t see this need.

Differences could be seen in whether departments within a cluster were reporting towards a director of the division or whether these departments report to the board of directors:

Controller University Hospital 2: Themes are just used as a clustering of departments. This is different than divisions in which a divisional director is responsible for the departments. We don’t use it that way which makes it harder for the board of directors, however bottlenecks within departments will directly show up”.

In this example, evaluation of results is not done within the theme but directly with the board of directors. The theme only functions as base for multi-disciplinary counselling.

 Information asymmetry

Contact with the external environment happens separately for each responsibility centre. This draws facilitating implications for tuning external funding (the external environment) with internal control systems. Every centre has its own contacts with the external environment making it hard to behave like one hospital. A controller from General Hospital 1 mentioned the following:

“Contact with the external environment is direct for the centres, which facilitates the translation to internal control within a specific responsibility centre. However, one could question whether this is most efficient for the hospital as a whole, because some centres do have better contacts with the external environment, resulting in asymmetry of information for the board of directors”.

Use of nonfinancial performance indicators

(29)

29

Controller University Hospital 2: “Control based in financial results remains most important, but it is not the only thing”. and “Financial indicators are most important, because health insurers won’t change funding based on safety indicators”.

The relation between financial and nonfinancial indicators was mentioned in the interviews. A distinction can be made between primary and secondary results.

Controller General Hospital 1: “The primary and most important results used for control are the financial results, however we are working on a system in which quality and for instance patient satisfaction are taken into consideration”.

As expected, a difference can be seen in nonfinancial indicators used in general versus university hospitals as a consequence to differences in tasks and functions between these hospitals. Document analysis reveals that university hospitals mainly focus on Research & Education (Dutch word: Onderwijs en onderzoek; O&O). Indicators like the number of medical degrees, promotions and top publications are set for departments within university hospitals.

Both financial and nonfinancial performance indicators are mostly considered using a balance scorecard approach or by a management contract.

Controller General Hospital 1: “A management contract consists of themes in which the board of directors specifies indicators concerning quality, safety and so on”.

Agents in hospitals

(30)

30

Redesign of the management control model

As can be deduced from the case description, this hospital is searching for a way to introduce a more coupling between external funding and internal control. The introduction of responsibility centres as a way of decentralisation is likely to come along with advantages aiming at a clearer link. Major advantages concern: (1) more local information, aiming at better decisions, (2) increased motivation and involvement and (3) flexibility through the possibility to quickly respond to changes Zuurbier & Hartman (2010). Especially the first and the third advantage make that the system of responsibility centre is an appropriate mechanism to create a link between external funding and internal control. Analysis of the conducted interviews indicated the recognition of these benefits accompanied with the motivation to implement responsibility centres. As mentioned before, the choice for this system should be based on an consideration of both advantages and disadvantages of decentralisation.

Preferences from within the case hospital implicated the consideration of the following factors in designing a new system: (1) adequacy of application within the current organisational structure and (2) the degree to which the design facilitates a coupling between external funding. Implementing responsibility centres based around specialism highly matches both factors. The adequacy of application within the current structure is important, whereas major changes are perceived as impossible and costly by the board of directors.

Structuring of this section is based on the topics concerned in the research questions. The level of deployment of responsibility centres

(31)

31 Characterization of responsibility centres

As mentioned in the theory section, the characterization of responsibility centres can be either financial type (e.g. cost or profit centre) and the organisational structure type (e.g. mono-disciplinary or multi-mono-disciplinary). As mentioned, deployment of mono-mono-disciplinary responsibility centres van be integrated within the current organisational structure in which specialisms are organised around specialisms. Mono-disciplinary responsibility centres are preferably deployed as profit centres and multi-user responsibility centres are preferably deployed as cost centres. Responsibility centres grouped around specialisms are preferable arranged as profit centres, whereas both revenues (by means of care product registration) and costs can be influenced. This way, incentives like good registration are delegated to most appropriate level.

Factors influencing the type of responsibility centre

The controllability of results determines the type of centres which is appropriate. Profit centres, for instance can’t be (optimally) deployed when the allocation of revenues to a certain centres is impossible. Prerequisites for the implementation are availability of transfer prices and the possibility to allocate revenues. Informants indicated that the hospital is currently busy calculating and allocating costs and revenues to the specific departments. The availability and use of market based transfer prices for the services of sector E is also preferred to increase efficiency.

Incorporation of non-financial measures

Both financial and nonfinancial responsibilities should be clearly formulated for these responsibility centres and management contracts could be used to formalise responsibilities. Whereas responsibilities for results decentralised to agents at lower levels in the organisation, results to be achieved should be clear in order to guide behaviors at these level. These contracts should preferable be based on a hybrid balanced scorecard approach in which both financial and non-financial indicators can be incorporated. This implicates for instance that indicators concerning education and research should be translated aggregated to the level of departments. Indicators should be evaluated and unwanted variances should be used to modify underlying behaviour or activity that inﬂuenced the variance.

Barriers which should be taken into account in the design

(32)

32 avoid causing unnecessary harm to the patient (Bitter, 2013). However, respondents indicated that decentralisation could cause the disciplines to increasingly focus on their own results, possible at the expense of multi-disciplinary collaboration. Multi-disciplinary collaboration in the current structure is already seen in the appearance of deliberation and meetings between specialisms that are highly interlinked. The new design should incorporate incentives aiming at stimulation of multi-disciplinary collaboration.

Deploying responsibility centres at departmental level implicates a relatively large span of control of the board of directors. The span of control of the board of directors could be reduced using the current divisional structure. Directors from these divisions could perform the evaluation of results of departments, however it is important that results and amenability of results are delegated to the level of the departments. This structure could also function as hatch between the centres and the board of directors concerning the symmetry and availability of information.

(33)

33

Discussion

The main objective of this discussion is to discuss the most important findings and to confront the results of the analysis with existing literature.

Defining responsibility centres

The mixed findings related to the definitions made by controllers in hospitals confirm the proposition that the term responsibility centre is perceived as an all-purpose word by practitioners in the field. The defining concepts and boundaries tend to be specific and clear in business organisations, however controllers in hospitals tend to be vague and confused about the concept. This makes it difficult to develop the knowledge base, essential to refine and move the field ahead.

The level of deployment of responsibility centres

A preference can be indicated to focus on the level of specialism, when comparing hospitals choices regarding the degree of decentralisation and the inherent degree at which responsibility centres are deployed. This finding matches theory concerning responsibility centres. Theory indicates that the degree of decentralisation should: “maximise the excess of benefits over costs bearing in mind organisational context factors” (Bhimani et al., 2012). Responsibility for results is preferably delegated to low levels in the organisation where results can be amended. Grouping specialisms within a division in which results are combined is likely to provoke dissonance due to differing interests between specialisms when results are combined within a division.

The characterization of responsibility centres

As mentioned, the characterization of responsibility centres can be based on two dimensions. The organisational structure type (e.g. mono-disciplinary or multi-disciplinary) as argued by Durlinger (2011) and the ‘financial’ type (e.g. cost or profit centre) as argued by Bhimani et al., (2012). The organisational structure type is highly linked with the decentralisation choice, however the focus in this section is on the choices regarding a certain grouping type, which makes it preferable to elaborate in this characterization.

(34)

34 respondents confirm that jointly performing tasks in patient treatment asks for cooperation between specialists, which is why a process orientation approach could be beneficial. One university hospital choose to focus on multi-disciplinary responsibility centres. This hospital is convinced that multi-disciplinary collaboration is essential to improve efficiency and that organising management control around care paths facilitates this.

Capacity departments such as operating rooms, labs and other facilitating departments which provide services to general specialisms are commonly employed in the presence of what Durlinger (2010) distinguishes as multi-user responsibility centres.

Another relevant aspect in the design of a structure based on responsibility centres is the ‘financial’ type of centre, as described by Bhimani et al., (2012). The choice for the focus on a certain level of decentralisation doesn’t prescribe the preferred type of responsibility centre. The types of responsibility centres can be found at all levels of (de)centralization within organisations depending on the degree of controllability of costs, revenues and/or investments. A typical finding related to these types of responsibility centres is that the revenue and investment centre were not used in hospitals in this study. A possible explanation can be found in the specific nature of functions hospitals, which are not found in business organisations. Bhimani et al, (2012) suggest that revenue centres are commonly used in sales departments. Sales departments or other corresponding departments are not used in hospital settings, implicating the irrelevance of the use of this type. Non-use of investment centres could for instance be explained by the fact that investments in buildings are irrelevant for specialisms. Investments are not amenable/controllable for managers of departments.

Factors influencing the type of responsibility centre

Choices between the commonly used profit or cost centres, seem to depend on the controllability and the subsequent traceability of costs and revenues between specialisms. Hospitals need to have clear overviews of cost prices and internal transfers prices, as described by Asselman, (2010). When focussing on the revenue side, external care product funding needs to be allocated to the responsibility centres which are involved. The revenue side seems to be complicated and subject to changes in the system by external stakeholders. This currently complicates the provision of revenue information.

(35)

35

Incorporation of nonfinancial indicators

Responsibility centre management highly focusses on budgets and financial measurement systems, acknowledged as cybernetic controls (Malmi & Brown, 2008). It is important to determine both results and responsibilities for each responsibility centre. Freedom of decisions unrelated to responsibilities should not exist (Zuurbier & Hartman, 2012). Control based on nonfinancial indicators in hospitals is executed using a hybrid system in which both management contracts and balanced scorecards are commonly used. Respondents argue the use of a balanced scorecard approach to be appropriate within the hospital setting as a way to combine both financial and nonfinancial indicators. This finding corresponds with literature (Grigoroudis et al., 2010). The difference in nature and functions between general and university hospitals is typically translated into performance indicators. Indicators regarding academic funding, promotions, top publications and medicine diplomas are examples of relevant translations of the functions of university hospitals. These examples of nonfinancial performance indicators are preferably incorporated into management contracts, either based on the balanced scorecard approach or not.

Disadvantages associated with responsibility centres

There are numerous macro and micro barriers and disadvantages associated with decentralisation in organisations in general. Major contextual, institutional and professional factors were briefly described in the theory section. Factors specific to the hospital environment were found during the analysis of the data.

As mentioned, the most commonly mentioned disadvantage associated with the use of responsibility centres in mono-disciplinary perspective is the reduced stimulation of multi-disciplinary collaboration. Making specialists accountable for own results implicates that they are highly motivated to focus on results within the mono-discipline. This potentially influences the degree to which specialisms search for collaboration needed within a care path. Respondents didn’t mention this implication to be detrimental, as collaboration and alignment between medical specialists also happens within these structures. However, not considering the incentive for collaboration could for instance lead to a reduced focus on quality, patient satisfaction and innovation within a care path. The inclusion of incentives aiming at the stimulation of collaboration between disciplines could be beneficial for the organisation and for patients in several ways, as literature indicates that collaboration between disciplines is key in hospitals.

Decentralising result responsibility: Using responsibility centres in University Hospitals