Organisational design for health integrated delivery systems: Theory and practice

Organisational design for health integrated delivery systems: Theory and practice

Health Policy 81 (2007) 258–279 Organisational design for health integrated delivery systems: Theory and practice Federico Lega ∗ Associate Professor...

1MB Sizes 1 Downloads 137 Views

Health Policy 81 (2007) 258–279

Organisational design for health integrated delivery systems: Theory and practice Federico Lega ∗ Associate Professor, Bocconi University and SDA, School of Management, C/o IPAS, 20135 Milano, Italy

Abstract Integrated delivery systems (IDS) are worldwide emerging as the dominant organizational form in the healthcare sectors. This article, drawing from international comparisons, focuses on organizational design of IDS. The analysis derives from an extensive literature review, which shows over the last years a significant lack of works on design issues, and from a number of experiences in community care settings, which provide useful insights on changes taking place in governance and delivery of health services at the local level. The frameworks discussed depict the major options of reorganization that can be observed in local integrated health systems of industrialized countries. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Organizational design; Health local systems; Integrated delivery systems

There has often been a tendency to blame individuals for poor personal performance when the real problem has been an unclear on unrealistic role [23].

1. Purpose of the work The focus of this article is on organizational design of integrated delivery systems (IDS). The analysis derives from an extensive literature review, which shows a significant lack of works on this issue [1–3], and from a number of experiences in re-organization of healthcare systems, which provide useful insights on



Tel.: +39 02 5836 2504; fax: +39 02 7003 4454. E-mail address: [email protected].

changes taking place in IDS.1 Scope of the article is to contribute to the filling of the gap highlighted by the literature review. To understand the context in which this work fits we discuss first a conceptual framework through which IDSs design issues can be addressed. In the following sections we then focus on exploring and 1 The analysis draws its conceptualizations from action researches and field investigations conducted over the last 3 years in several health systems. Among the main were involved several Local Health Authorities in Italy, some strategic health authorities and community trusts in UK, community services organization managed by local authorities in Scandinavia, Swiss, Spanish and French managed care organizations (MCOs). Along with field studies of Mount Sinai Health System (NY), Health and Hospital Corporation (HHC), UAB Health System, Kaiser Permanente and Johns Hopkins Health System (USA), extensive literature review of American HMOs and MCOs complemented the analysis.

0168-8510/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.healthpol.2006.06.006

F. Lega / Health Policy 81 (2007) 258–279

assessing the features and feasibility of different – from traditional to innovative – organizational designs for IDSs. The final section introduces as a possible area for further research the theme of the relationship between designs and people, that is managerial skills required to their implementation and functioning. However, four specifications are required to assess the context and the limits of this work before moving to the analysis.

2. First specification: the rise and the nature of integrated delivery systems Most of modern medicine can be practiced outside hospitals. It requires a high level of integration and coordination among different settings, specialties and professionals. Theoretically, only acute and sub-acute care should be delivered in hospitals or other specialized facilities for inpatient care (such as skilled nursing facilities). Such facilities should treat patients requiring some type of continuous monitoring. Yet, most pathologies are now treatable in an outpatient setting, and the steady increase in number of chronic diseases is imposing an epocal change in the way care is delivered. Costs, quality of life, and professional reasons (coordination and integration) are all requiring for different solutions from the classic hospital-centered ones. A process of “substitution” is under way. By substitution we mean a process in which there is a continual regrouping of resources across care settings to exploit the best available solutions [4]. A dominant trend includes substituting home or primary/community care for secondary hospital care (for instance, hospital at home schemes, enhancement of long term care and renal dialysis).2 Community-based services are becoming central in all health sectors of developed contries. There is a number of possible services that could be included in a list of preventive, primary and secondary 2 In WHO [4] words: “inpatient care typically consumes between 45 and 75% of resources dedicated to healthcare. There is an increasing perception that there are more cost-effective alternatives to the care currently provided in hospitals and that consequently there is scope for a further reduction in hospital services. In almost all western countries, the total number of hospitals beds fell significantly between the 1980 and the 1990s, accompanied in most cases by a shortening of length of stay. These reductions probably result from a combination of cost-containment policies, changes in technologies or treatment, and an increased reliance on primary and social care”.

259

care (which refers to specialized services deliverd by specialists) delivered outside hospital settings. Some are listed below [5]: • Primary and medical care—including basic investigations (images and laboratory) and general medicine and paediatrics (rehydration therapy, chemotherapy, outpatient treatment). • Uncomplicated obstetric care. • Specialized outpatient clinics (diabetes, ophthalmology, psychiatry, dermatologuy, ear, nose and throat, physical rehabilitation, uncomplicated cardiology, etc.). • Day surgery (gynaecology, paediatrics, orthopaedics, oral surgery, dental surgery, plastic surgery, urology, ophthalmic surgery, etc.). • First level of emergency care. • Antenatal, postnatal, and family planning clinics. • Nursery school. • Sex instruction for children, health promotion and health education for activities, marriage guidance, child guidance. • Immunization services, etc. In many countries worldwide, these services, especially in urban areas, are managed in part by general practitioners, in part by health authorities or in part by health organizations that deliver services through health centers and other facilities, including hospitals. There is a wide variety of governance initiatives adopted to coordinate the management of the complete array of health community services. Yet, over the last years a dominant organizational initiative has emerged: we refer to the integrated delivery systems (IDS). Within the general concept of IDS we refer to a number of different situations. They are “umbrella” organizations that manage the whole spectrum of services and levels of care. IDS might integrate with different combinations, community services with hospitals, home care schemes, rehabilitation facilities, nursing homes, mental health centres, etc. More specifically, IDS regroup organizations providing care at different levels: prevention and environmental health services, primary care (GPs), secondary care (outpatient services), tertiary care (general or community hospital), quaternary care (academic medical center and specialty hospital), rehabilitation (nursing homes, rehabilitation centers), and long term care (long-stay inpatient centers, home care units). IDS are vertically integrated

260

F. Lega / Health Policy 81 (2007) 258–279

organizations (by ownership, by contracts or by other networking strategies) that are aimed at building the most complete and cost-effective continuum of care for their catchment area. They manage a wide range of health activities targeted to the population enrolled in contracted plans and residents in the geographical areas where services are offered. In short, an IDS “provides or aims to provide a coordinated continuum of services to a defined population and are willing to be held clinically and fiscally accountable for the outcomes and the health status of the populations served” [6]. Examples of IDS are the Health Maintenance Organizations (HMOs) in USA, the Local Health Authorities in Italy and – to a certain extent – the Primary Care Trusts in UK. In many cases, IDS are organized in health centers or districts. Districts and health centers can have a dual nature: • from a structural perspective, they are single or a network of facilities where services are delivered to patients3 ; • from a managerial perspective, they have an organizational responsibility over the health needs of a population or territory. In this sense, they are in charge of enhancing an epidemiological picture of the health and health needs of their local population, and adopting planning and investement policies to match those needs with appropriate services. In public healthcare systems, they are accountable both 3 WHO report on primary health care in urban areas [5] reports: “For several decades, particularly since the Second World War, health systems throughout the world have tended to emphasize the development of health centres with multidisciplinary teams as the local basis for implementing integrated programmes of first-level prevention and treatment. This movement has been common to rural and urban areas, with centres that have varied both in their range of functions and in their degree of success. . . Health centres may take different forms depending on the particular conditions and circumstances in each country and city. . . Not only do they provide all health services to the population within their own catchement area, but they support the primary health care functions of smaller local health facilities in the larger area they serve. . . They also monitor community health needs, both medical and environmental”. Unfortunately, as WHO notes, at least until the beginning of the 1990s, “. . .there have been widespread complaints that many health centres were not living up to the expectations, particularly with regard to the standard of the services delivered. The hospital sector continuated to dominate and, through lack of credibility, health centres are often bypassed, even by patients with minor ailments. As a result, hospital services are frequently overloaded and functioning inefficiently and inappropriately”.

to the IDS management and to the population and authorities they serve. All in all, although they might differ in governance structures and contents, IDS represent the organizational form – for managing and delivering community services – emerging in most of the developed countries. For this reason, in this article we discuss problems of organizational design of integrated delivery systems (IDS), with particular attention on the meso-level of design, which seems to be the more complex on the field and the more neglected in available literature.

3. Second specification: governance, organization and operations: three design issues Healthcare organizations are faced with the reality of design to an increased extent [2]. In the last two decades, several environmental forces have required healthcare organizations to shift from their traditional closed self-referential attitude for openness towards expectations related to their stakeholders [4]. Choices on priorities, services delivered, organization schemes, are all increasingly becoming objects of negotiation with the stakeholders. This process has involved all levels of organizational design: governance, structure design, and service delivery. At the macro-level, many actions have regarded modifications in the governance model, driven by changes in ownership or by pressures exerted by stakeholders (e.g. government, patients, purchasers, other providers, employees, etc.). This is the case of a government looking to play a stewardship role in healthcare organizations, or a community looking to be represented on the governance bodies. In the search for a governance framework to balance the organization’s goals with the expectations of all stakeholders (that is balancing the degree of autonomy desired with the degree of accountability required), what we have experienced is a continuous re-adaptation of governance structures. Some examples are private organization boards, including representatives of citizens and of government, public organization boards, including representatives of private non-profit organizations and of citizens. Further, powers of the board might be increased or decreased with respect to those of the CEOs or of the clinical directors. Basically, governance models of healthcare organizations are often

F. Lega / Health Policy 81 (2007) 258–279

redesigned according to dominant political ideology (role of private and public in health services) and to expectations raised by communities, by organized lobbying groups and by influential employed professionals, such as clinical directors, chiefs of specialties, etc. [7]. Along with the redesign of the governance structure, many restructuring processes have taken place at the micro-level [8]: re-engineering of health services, delivering procedures, adoption and adaptation of emerging working philosophies as patient focused care, continuous improvement, and total quality, etc. In a context of emerging consumerism, coupled with financial pressures to reduce costs and improve efficiency, and with social and professional pressures to increase quality (effectiveness and evidence-based medicine), attention to health services delivery design has became a key issue [9]. Between the macro- and the micro-levels, a third field of healthcare organization redesign is represented by the meso-level. It involves the selection of choices on the organizational design, which means choices on labor division: how work tasks are grouped in units for coordination and accountability needs and how units are grouped in hierarchical layers for the same reasons. A formal structure with its supporting organizational mechanisms (planning and control, human resources management, information systems) is the result of the meso-level design [7]. A commonly accepted definition of organization design states that it “refers to the way in which the building blocks of the organization – authority, responsibility, accountability, information and rewards – are arranged” [10]. What is surprising is the lack of interest that scholars and practitioners of the healthcare sector have demonstrated on this issue [1,2]. Without a doubt, healthcare organizations’ performances are highly influenced by their governance model, which is primarily related to their ability to minimize the gaps between unreasonable and reasonable expectations and between the expectations and the goals pursued. Their performances also heavily depend on their delivery processes, mainly for what concerns operational efficiency and the patient’s (consumers) satisfaction. Overall performances, though, are “bounded” by the organizational design, which shapes the links between front desk operators and the strategic apex. To use a metaphor, the movements of a finger are determined by the impulse given by the brain, but they

261

will be appropriate only if the nerves, the muscles and the bones convey properly to the impulse.4 While much has been written on micro- and macro-level design for healthcare organizations, few studies have investigated the meso-level. This is an issue we need to investigate at large, as healthcare organizations are growing bigger and more complex – as in the case of integrated delivery systems (IDS) – in an more and more unstable and demanding environment (managed care, regulations, consumers expectations, etc.). An environment that needs to be taken into account for the implementation of new and more effective organizational designs [8,11,12]. For years, they have adopted a design labelled professional bureaucracy. It is characterized by the search for standardization of procedures and products through the so-called pigeonholing process: the organization seeks to match a predetermined contingency to a standardized program. So, they organize around the skills and knowledge of their professionals who are in charge of categorizing or “diagnosing” the client’s (patient) need and apply, or execute, the matching program or procedure [13]. Organization around skills is what determines a functional design [14]. The complexity of modern healthcare organizations, which reaches the highest point with IDS, and the rising expectations of the external environment have made, in many circumstances, obsolete and inappropriate the traditional structures adopted by these organizations – either disciplines-based or facilities-based – are pushing for a new approach to their design. As they become more complex in a more unstable environment, they face the problem of coping simultaneously with needs to differentiate, as with the identification of functional disciplines-based units, and integrated work, as with the re-grouping in different ways of knowledge and skills [15]. They have, then, the difficult task of rethinking their design and allowing at the same time for higher integration. Some examples are, the reorganization of units around markets, areas, products or clients, without loosing their edge on specific knowledge and the 4 As Dixon [23] writes “this is in no way meant to imply that the skills, knowledge and capacity of the unit managers are not just as important as organizational characteristics. But we now have a good deal of evidence and experience about the conditions under which effective work can be carried out, in units or elsewhere”. That is, organizational structure enhances or constrains managers’ performances.

262

F. Lega / Health Policy 81 (2007) 258–279

skills required by a continuously developing medical science. The reasons as to why an IDS should look for one or another structure design are discussed in next paragraphs.

4. Third specification: contextual, rational and non-rational forces driving the design process Since the object of the analysis embodies the reasons for which an IDS might be designed, the framework also offers the opportunity to outline the forces driving the design process, at least the ‘rational’ ones. This fact requires some further explanations. On one side, is necessary to recognize that the primary factors driving the design process in different countries are quite not the same. The peculiarities of the context (dominant ideology, legal and institutional framework, etc.) play a fundamental role in determining the drivers for designing. Looking at individual countries, we can identify different priorities: for instance, some choices might be oriented by cost reasons, while others seem to be more driven by clinical practices. Therefore, different aspects of the general framework through which we analyze design issues for IDS in this paper might assume different relevance for the reader according to the priorities dominant in his or her country. A contextualized reading might emphasize one set of issues instead of another. On the other side, as in any decision-making process involving services of public interests, and therefore several stakeholders (community, politicians, professionals, etc.), a redesign of an IDS could be the outcome of a technical (rational) process aimed at evaluating projected benefits and drawbacks, or the result of a political debate in which, for instance, authorities in power sustain a policy (ideology) of ‘one best way’ forcing the IDS to redesign independently by any other reasoning and evidence. In countries with public-based health systems, politicians and bureaucrats are not immune to management fashions, and the desire to differentiate themselves from the policies pursued by previous authorities in power might lead to the extensive adoption of course of actions which in some cases have a good rational basis, while in others they have much less of one, as in the privatization-era, or in the (still on-going) period pro-closure of small rural hospitals [25]. So politicians, bureaucrats and citizens play a key role in

influencing the consolidation decision-making process and its technical or political basis. Furthermore, another non-rational force for consolidation can be also recognized in that some hospital executives seek redesigns to increase their power, prestige and benefits linked to their future status as managers of larger organizations [26]. In this work, non-rational drivers are not addressed, and the focus is exclusively positioned on rational–technical drivers for IDS design. This is because it is undoubtedly true that even the most politically influenced decision, such as the reconfiguration of responsibilities and services managed by an IDS, should be based on a technical analysis in order to be rationally approved and pursued. The aim of this work is to contribute to the consolidation of the technical basis to take those sound decisions.

5. Fourth specification: people matters It is finally worthwhile to note that while we can generalize “archetypes” of IDS designs, consequences and outcomes depend not only on the choices on structure or operating mechanisms, but also on a number of factors that make it difficult to draw general conclusions. Reorganizations following a new design of an IDS can be challenged by several factors. Drawing from Lefkovitz’s [41] work we can distinguish barriers to change in systemic and attitudinal factors. Systemic barriers include logistical structure, turfism, incompatible financial incentives, lack of support from organizational mechanisms (information systems, budgeting process, etc.). Attitudinal barriers refer to conceptual orthodoxy and “either-or thinking”. Here we will point out and briefly discuss one of them. It refers to conceptual orthodoxy and is the refusal of the management approach and the resurgence of the clinical need approach, as a consequence and reaction of health staff to the increasing managerial responsibilities delegated or imposed to them. In many countries and organizations, this shift has indeed produced spread sentiment of opposition and the refusal between the medical staff and other health professionals, both inside the organisations and in the external political arena. Managerial responsibilities and roles, such those of District managers, Department managers, etc., need

F. Lega / Health Policy 81 (2007) 258–279

adequate competencies, which typically are not present in the educational background of the health staff. More, those responsibilities have often to cohabitate with more traditional professional tasks: in this situation, the two roles are potentially conflicting and that can therefore induce individual psychological stress. In other cases, managerial objectives and responsibilities are set simply too high, i.e. at levels not realistic nor feasible. This maybe a good solution for the IDS managers, at least in the short term, but it loads the line managers and the organization with inappropriate and wrong pressures, asking them decisions that should be taken by the top management, generating a sense of frustration and rejection. Because of this situation, professionals refuse to deal with those tasks, which they judge undue and undeserved, they feel themselves inadequate to play that role, and they progressively adopt a strategy of disengagement from the organizational objectives. This trend can produce, as final result, a sort of entrenchment of staff behind the more comfortable tradition and value of their professions: the attempt to have “bureaucratized professionals” (clinician–manager) has so led to the result of having “bureaucratic professionals”, that is professionals that operate exclusively along the prescription of their own profession. This means the resurgence of an approach based exclusively on clinical needs, probably self-referential, with minimum attention for the organizational dimensions of activities. Consequently, in this situation is easy to experiment severe conflicts between the clinical and the administrative staffs and roles (i.e. between clinical director – Chair – and the administrator of the department, between the Chair and the Chiefs of the specialities regrouped in the same department).5 To avoid such a situation and the risk to have bureaucratic professionals, IDS have to work in parallel at least on three dimensions: 5 A similar situation, for instance, is the one outlined by Som [36] in his paper on UK doctors’ response to clinical governance. Som’s research “indicates that doctors are not enthusiastic about clinical governance and it is not receiving wholehearted support from doctors because they feel that clinical governance is a managementled initiative imposed without adequate consultations. The research points out the tension between an organisation (wishing to bring clinical care within a management framework) and doctors (who are resisting managerial efforts to replace the old framework of Bureu professionalism”. This context highlights the importance of support of clinicians for successful changes.

263

• develop adequate managerial competences for the clinical staff asked to take roles that imply managerial responsibilities; • design the organizational structure with the clinicians, especially those called to play managerial roles; • manage effectively the internal communication, in order to inform and involve all key actors. In other terms, the spectrum of designs that may produce benefits or drawbacks, and the degree of their impact, is to be evaluated with regard to the actors involved. Unsurprisingly, the main actors that will influence the magnitude of positive and negative reactions generated by the IDS design are represented by the professionals. In this perspective, though this work does not focus on change management issues, it is of the paramount importance to underline that design and implementation should not be considered as separate issues: success of implementation depends heavily on the design process and in the managing of change necessary to implement desired design (time, actors involved, groups, tactics, etc.) [27–33]. Actors called to design formulate the hypothesis which will be tested, reformulated, fine-tuned during the implementation process. As change/implementation involves consensual validation of premises and hypothesis of which relies – that is sharing of perceptions between organizational actors and designer about what is the problem, how we name it and how we frame it (a “meeting of minds”)– much of the design activity involves managing the process through which ideas and proposal about new designs are discussed, shared, experimented and finally adopted [34,35]. This complex activity is based on the ability of matching unique organizational situations/problems with theoretical cognitive maps so as to name, frame and change appropriately the context. Change should not be viewed by clinicians as a compulsion, but rather as an opportunity. That is why, along with a deep understanding of techniques and archetypes of designs, people matters [37–40]. 6. Back to the analysis: organizational needs and prominent factors for designing integrated delivery systems For the scope of this work, we assume that IDS are organizations resulting from mergers or other formal

264

F. Lega / Health Policy 81 (2007) 258–279

partnership agreements stipulated by providers who accept to be governed by one single board and/or CEO. This type of IDS has a tight governance structure.6 We will not discuss the case of IDS with loose governance structures based on informal alliance agreements stipulated by independent providers, weak alliances that often do not allow for effective integration, cooperation and reorganization of the IDS7 [16,17]. In some health sectors, like in USA, IDS were autonomously developed to cope with managed care, mainly as a result of vertical integration between providers (e.g. Preferred Provider Organization – PPO – and Point of Service Organization – POS) and between them and insurers (e.g. Health Maintenance Organizations – HMOs). In other health sectors, like in UK, France, Spain and Italy, IDS were institutionally created under different national health systems (NHS). For instance, we refer to regional, provincial, district or local health authorities, strategic health authorities, community trusts, etc. They are in charge of providing care for the population resident in the geographical area under their responsibility. In the last years, the creation of internal quasi-markets through the separation of providers (acute hospitals, community hospitals and trusts and primary and secondary care practices/trusts) and payers (Health Authorities) is challenging the initial nature of Health Authorities as an IDS. Nevertheless, in most cases they still represent and act as IDS since they often retain the ownership, and they govern directly or indirectly the providers through various formal agreements/mechanisms (provision contracts, strategic planning processes, licensing, etc.). How they can guarantee an effective coordination and efficacy of services directly delivered or purchased calls for the attention to the design of top and middle management responsibilities, which means looking at the organization mesolevel design. 6 As Longest [24] writes “the emergence and growth of IDS reflects a fundamental shift in how healthcare services are organized and delivered. In essence, the healthcare industry is changing from a past in which almost all healthcare organizations were organized independently to a future in which many, perhaps most, healthcare organizations will be structurally integrated with others. If the integration is not through ownership arrangements, then it will be through agreements and affiliation that create virtual IDS”. 7 In other words, an IDS is not just the sum of several independent providers with a common brand, but it’s their reconfiguration to better serve its population.

Through their design and functioning, IDS try to meet three complementary needs [5]: • to put an end to the inappropriate and inefficient use of secondary and tertiary referral care and the resulting pressure on scarce resources; • to ensure that full use is made of primary health care services; • to provide cost-effective services to enable the maximum health gains to be obtained from restricted budgets. issues which are becoming more and more relevant with the emerging of managed care contexts are based on financing through capitation. In such contexts, it is fundamental to manage demand patterns in order to: • relocate care and cure delivery at the most costeffective level (from hospital, to day service, to outpatients services, to GPs); • stimulate referrals to the provider within the system that has the best expertise and skills to cope with a specific problem. Therefore, in our view a conceptual framework of reference for designing IDS should take into account the following organizational needs: Integration (clinical integration).8 We can refer to Conrad’s [18] words to define the concept of clinical integration: “it is the coordination of care of a given person over time and is the sine qua non of vertical integration. . . Fundamentally, patient care coordination occurs for the individual person over time. Thus, true vertically integrated care demands a system capacity to plan, deliver, monitor, and adjust the care of an individual over time. Second, the essence of a system is the ability to “aggregate” individual care coordination 8 Charns [1] distinguishes four phases in the IDS development in the USA health sector: (1) the “cottage industry” phase, preIDS formation, characterized by independent facilities (before the 1970s); (2) the horizontal integration phase, with the pooling of facilities/actors at similar stage of “production processes”, like multihospital systems or physician group practices (mid-1970s to mid1980s); (3) the vertical integration phase, with the creation of network among hospitals, physicians, nursing homes, insurances, community services, etc. As result emerge HMOs, MCOs, IDS, etc. Vertical integration occurs mainly through ownership (late 1980s to mid-1990s); (4) the clinical integration phase, where vertical integrated systems look for approaches and tools to improve their capacity to deliver services in an integrated fashion (late 1990s till now). Integration is searched both at operations and structure level.

F. Lega / Health Policy 81 (2007) 258–279

and clinical processes into a system level capacity to plan, deliver, monitor and adjust structures for coordinating the care of populations over time. The coordination of care for individual patients is a necessary, but not sufficient, condition to realizing system-level clinical integration”. As a matter of fact, along with integration at the operational level, an appropriate design of governance – macro-level – and organization structure – meso-level – are necessary to support and enhance clinical integration. Unfortunately, it seems that vertical integration has been implemented, with mixed results, mainly due to a focus centred on the governance level, especially with reference to the organization’s ownership [1], while actions at micro- and meso-levels have been largely ignored. Anticipation: Emphasis should be placed on preventive medicine. Early detection and cure of diseases, particularly the chronically degenerative ones, improve chances of a better quality of life for the patient, and contains the costs’ escalation due to the disease severity. Proximity: Services should be delivered as close as possible to the patient. This becomes economically viable only if the largest part of health needs are filtered and tackled by general practitioners and other first level physicians, which act as gatekeepers. They are the most diffused points of service in the system. Filtering and keeping the patient as much as possible at low-intensity care level should limit costs and improve the patient’s “comfort”. Patients are treated in a known environment, either at their home or at the physician’s ambulatory. Territoriality: Some healthcare expenses must be monitored and governed on a geographical basis. As a matter of fact, the following categories of activities and expenses represent the interlocked dimensions that must be planned and controlled with reference to the whole community or sub-parts of it that are served by the IDS: pharmaceutical expenses, hospitalization rates, volumes of out-patient diagnostic exams, volumes of out-patients specialist visits and therapies, rehabilitation, long term care and home care expenses. For example, a decrease in hospitalization rates (and expenses) generally translate into higher post-discharge pharmaceutical expenses and in more follow-up specialist visits, out-patients exams, home care accesses or long term care beds occupation, etc. [4,5]. In other terms, if patients are not treated through

265

hospitalization, they should find answers to their problems through the community services, and vice versa.9 The fact that health consumes are interlocked requires the IDS to investigate the opportunity of defining organizational responsibilities of enrolled sub-populations (markets or geographical area, such as areas, zones, districts, etc.). Eventually, market–area managers would be accountable for planning coordinated actions over the different lines of service, so that interventions to reduce inappropriate hospitalization do not generate unreasonable increases in other health expenses, or on the contrary, when rationing in pharmaceutical coverage generate substantial increases in hospitalization. Fitness: Due to a different geomorphology (urban, sub-urban, country, mountain, seaside, etc.) and composition of their population, different geographical areas require a different mix in healthcare services delivered, both in terms of specialties and logistics. Some areas require more babies and children or women services, others more elderly, home care or mental health services. In some areas, services can be grouped in one central multi-specialty delivery site (health centre, out-patient department, etc.). Other areas require different single or multi-specialty delivery sites located across the territory. Availability of transportation means and times required for reaching the sites affect such decisions. Accountability: Different local authorities (municipalities, health authorities, etc.) and a number of other stakeholders (patients’ organization, private non-profit organizations, purchasers, suppliers, etc.) operate on the territory covered by the IDS. How is the IDS accountable to all of them? Different political parties governing local authorities might have different ideologies and expectations, and require separate lines of responsibility inside the IDS to cope with them. Especially in public-based systems, each local authority (town, city, etc.) might want to “negotiate”, on behalf of its citizens, the health services offered in its area. So, the IDS organizational design should also consider the needs of meeting those lines of accountability. 9 Obviously, decrease and increase might be proportional or not: it depends on the level of inefficiency and inappropriateness in the system. High level of inappropriateness might allow for reduction in hospitalization without generating relevant new activities and costs on the rest of the system.

266

F. Lega / Health Policy 81 (2007) 258–279

Manageability: A large size of an IDS (in any dimensions, such as population served, geographical area covered, workforce employed, etc.) requires a more refined and articulated organizational design supporting delegation of power and coordination needs. This means providing for intermediate hierarchical levels, for coordination roles, and for a multi-level budgeting system. A design that hopefully will close the gap between the strategic apex and the line units, which tend naturally to duplicate and fragment self-referentially on professional basis.10 Without intermediate-middle management positions to decentralize responsibilities, in large IDS, top management struggles to supervise the increasing multitude of line specialty units [19]. The question is whether this middle level, in which lowerlevel line units are regrouped, should be structured according to a traditional facility or to a disciplinebased criteria or according to other more innovative – for the health sector – design dimensions. In short, the organizational design of an IDS should be defined according to its organizational needs determined both by external forces acting upon it and by its internal complexity factors. Regulations (e.g. the financing mechanism), competition, political (authorities) and/or social (citizens) expectations represent external forces. Internal complexity is mainly linked to the organization’s size (employees, population served), to the diversification of products (levels of care, types of services offered), to the geomorphology of territory, to the logistics and to the connected opportunities for economies of scale (concentration of services in few centres, as in urban areas) as opposed to the need for decentralization (since proximity to population requires points of access diffused, as in rural areas). It is the necessity to cope with external forces and with complexity factors that lead to an organizational design more oriented towards one organizational need or another. The result might be a design supporting more integration or differentiation of line units, centralization or decentralization of decision-making, closeness or openness to the environments, and innovation or standardization of processes. Given this conceptual framework, the next paragraph outlines and examines the main design alternative for IDS. 10 Rational reasons are to guarantee a manageable span of control and the ongoing specialization required by clinical practice, irrational reasons are often based on power dynamics.

7. One design to fit one IDS A combination of internal complexity factors and external forces might well be numerous. Having to fit a specific internal and external environment, IDS designs could be, then, almost infinite. Field analysis shows such variety [3]. This statement is consistent with a contingency approach to design, whose aim is to meet a IDS specific organizational needs. Nevertheless, to support management in designing, some sort of categorizing is possible and necessary. Therefore, although the number of dimensions that we could refer to when classifying health services are quite large, we can circumscribe the main alternative bases for units’ designs to the following: functional/specialist knowledge (disciplines), facility, geography, clients group, and line of product. As discussed in the previous paragraphs, from the beginning, many IDS have adopted functional structures, either a facility structure or a disciplines-based design (with departments such as nursing, social work, physical therapy, imaging, etc.). Functional structures were traditionally developed because it was considered easier to manage specialists if they were grouped together in the same department (informally labelled “silos”) and if their head had training and experience in that particular discipline. Functional or disciplinebased structures promote skills specialization and know-how economies, it reduces duplication of scarce resource, it ensures homogeneous delivering procedures across IDS catchment’s area, and it facilitates decision-making and communication because managers share expertise and culture with their subordinates. On the contrary, functional structures emphasize routine tasks, that encourage short time horizons, it fosters parochial perspectives by managers, it reduces communication and cooperation among different units, and it obscures accountability for overall outcomes, especially for clinical outcomes when they depend on multidisciplinary actions. Most of the time, there is a large fragmentation in care paths due to a lack of coordination: this is the case of complex diseases such as Alzheimer, diabetes, cancer, hypertension, etc. With emerging complex systems and patients, functional designs have in many cases failed to answer the needs for integration, anticipation, accountability, territoriality, fitness and manageability. IDS have then started to look for new organizational structural forms [20].

F. Lega / Health Policy 81 (2007) 258–279

Some authors [3] analyzed this phenomenon and proposed a continuum of alternatives comprising of the traditional organization of an IDS by facility and the innovative form based on service lines. According to Charn and Tewksbury [15], clinical service lines may be defined as a family of organizational arrangements based on outputs rather than on its inputs. Organizing around outputs creates a service line structure consisting of people in different disciplines and professions, and who have a common purpose of producing a comprehensive set of clinical services. Other authors have referred to the same concept as product line management [21,22]. Between the facility structure and the service line structure, the continuum proposed by Parker et al. [3] includes variations based on service line integrators, task forces, an internal reorganization of facilities, service line teams, and matrix arrangements. IDS could benefit from an increasing integration in their operations when adopting those variations in their structure. In our view, the framework proposed by Parker et al. is partial, since it does not take into account other designing needs than the integration one and it does not include the geographical dimensions as a designing criterion. Furthermore, we argue later that the service line design does not enhance the expected high level of integration since it is still based, although indirectly, on professional lines.11 However, we consider it as a 11

As a matter of fact, this argument is indirectly already present in Parker et al. study. In their analysis, based on telephone interviews with senior executives at 14 healthcare systems, Parker et al. [3] found as perceived advantages of service line design the following: improved focus on planning and decision making; enhanced skills for career mobility; reduced interfacility friction; improved accountability (on clinical outcomes); reduced costs; improved focus on clinical areas. On the contrary, disadvantages most frequently cited were: new blockages and backlogs in decision making; stakeholders’ differing expectations of service line objectives; increased political maneuvering (for internal competition reasons); difficulties in giving service line managers sufficient authority; difficulties in implementing change. Of particular interest for our analysis are the first two disadvantages. In the words of one of the interviewed, as reported by the authors, “service lines are just a different set of silos, creating new blockages and backlogs”. In this sense, service lines (or product line management) tend to reproduce some of the unwanted consequences of functional organization (fragmentation, barriers to knowledge and information sharing, self-referentiality). Stakeholders’ differing expectations are reported with reference to the problems created by service line grouping operations of previously competing specialty centers in the same market. Again, barriers to integration are raised by the different centers, in the attempt to protect their

267

starting point for our exploration of alternative organizational designs for IDS. Our analysis, based on reorganization recently attempted by IDS to better respond to their specific needs, is focused on innovative designs placed along a continuum, which can be considered as an extension to the one proposed by Parker et al. One end of the continuum is in fact represented by the product line structure. We label this design “P-design”, where P means product (service) lines-based. The other end is associated with population, area or market-based designs. We label these designs “G-designs”, where G means geography based. The P-design refers to structures and responsibilities centered around two types of outputs: 1. diseases/pathologies, such as diabetes, Alzheimer, pneumonia, transplants, etc. 2. clients, such as women, elderly, mental health patients, etc. P-designs look for multi-professional integration around differentiated programs and services, while Gdesigns look for integration and accountability on a geographical/population basis. As a matter of fact, in P-design, specialization is still central since those types of designs are functional or divisional schemes built around professional knowledge (required for a specific group of client, specific line of products or broad clinical area). While in G-design we are referring to designs seeking multi-specialty integration built around discrete geographical zones (labeled as districts, areas, etc.).12 The implications and consequences of one or the other design of the organization are quite relevant. As already noted by Dixon [23] more than 20 years ago, “to the extent that the structure is based on dimensions that are basically institutionally or professional in character (facilities, disciplines and products), there is the tendency to reinforce the existing patterns of service and to prevent developments which challenge institutional and professional boundaries. To the extent that the structure is based on dimensions such as areas of competitive advantages (intellectual capital, knowledge, etc.). In our view, the problem of stakeholders’ differing expectation assumes great significance since it is not just related to internal relationships but it affects also the relationships between the IDS and its external stakeholders, such as local municipalities, health authorities, citizens’ groups, patients’ organizations, etc. 12 For a detailed discussion of characteristics of functional and divisional structures see Duncan [14].

268

F. Lega / Health Policy 81 (2007) 258–279

health care, health care groups or programmes (clients and populations, areas or markets), there is an increased likehood of identifying needs that are not being met, of producing innovative solutions and of providing services that cut across jurisdictional boundaries”. To reduce or enhance undesired or desired consequences, corrective mechanism might be introduced in the two “extreme” designs, providing for second level areas/market responsibilities in P-designs or for disciplines/products/clients responsibilities in G-designs. These “mixed” designs converge towards the middle of the continuum where a matrix design (M-design) is positioned. A matrix design combines first level responsibilities of P- and G-designs. Each of the five designs located along the continuum, which differ in terms of authority over personnel and control over resources, are in the next section, and analyzed in detail and assessed with respects to its capacity to meet IDS complexity factors and organizational needs, as outlined in the previous paragraph. The designer’s attention must be focused on avoiding gray areas of accountability: each time two or more dimensions are used as the basis for defining areas of responsibility at the same level within the same organization, they are bound to be areas of “overlapping” which could logically fit within one area or another [23]. This is, for instance, the case of community services (such as mental health, drugs and alcohol rehabilitation, hygiene and prevention, etc.) when in the same

IDS there are geographical responsibilities (districts, zones, areas), and crossing functional ones (i.e. a Mental Health Unit). Who reports to the psychiatry, or to the District manager or to the chief of Mental Health Unit? When should we opt for one design or the another? 7.1. “P-designs” and their features This cluster of design refers to the divisional structures with responsibilities built around product/service lines or customer segments. Personnel and resources are grouped on such a basis. Drugs and alcohol dependency, mental health, hygiene and preventive medicine, long term care, primary care, women’s health, babies and children’s health are all examples of divisional units (Fig. 1). A divisional structure is also labelled a selfcontained-unit structure, since all or most of the resources necessary to accomplish a specific goal are set up as self-contained units headed by a division manager. Divisional or products/clients-based structures recognize the needs for inter-units cooperation, foster a greater orientation toward overall outcomes and clients, and ensure accountability by managers and so promote delegation of authority and responsibility. On the contrary, divisional structures may use skills and resources inefficiently, impede specialists’ exposure to others with the same specialties, and promote

Fig. 1. Example of a P-design.

F. Lega / Health Policy 81 (2007) 258–279

departmental objectives as opposed to overall organizational goals. Borders between disciplines, products and clients are very blurred in the health sector. For example, mental health might refer to a discipline, to a segment of clients or to a line of products. This explains why divisional structures tend to reproduce similar patterns to those of functional structures. Furthermore, P-designs are highly centralized as the geographical dimension is considered not relevant and there are no diversified market responsibilities defined. In a complex health context, this design, which focuses on the supply side, generates a self-referential approach, supports the development of control systems exclusively based on productivity measures, does not create any incentive at monitoring and governing demand patterns and overall expenses, and generates a significant drawback. And, middle management is not held accountable for governing the whole demand of care by the population of reference. As discussed in previous paragraphs, such demands present relevant interdependencies among the sub-sectors. The only ones with a broad picture are the top executives, but in complex contexts, it might prove difficult for them to guarantee the required coordination and integration among service lines. Middle managers’ task is primarily to respond to the health demand involving their departments, no matter what the appropriateness of that demand or the distribution of the services delivery is. Equity and accessibility are not perceived as key issues. It is not critical if 100 patients treated were all living in the same area, while patients in other areas suffer a shortage of service. What counts is productivity, the total number, 100, and its performance must be considered efficient (at least acceptable). And, it is not a problem if out of the 100 treated, 25 might have been treated better in a different context or with other specialists (e.g. from the hospitals to the GP, from the GP to the home care service, etc.). Still, 100 is the goal and the manager’s usual comment is “people were in need, what was I supposed to do, not to treat them?”. Rationing services, and its consequences, are not an issue. Also, it does not matter if the population is not homogeneously distributed across the territory (e.g. in situations where there is a higher relative percentage of elderly in some areas and of young in other areas). In most cases, the accessibility of specific services will be tailored on the needs of the specialists (through a classic close-system self-referential

269

approach), determining in many cases over or under treatment situations in both areas. P-designs support anticipation, proximity and manageability needs, but do not meet complex IDS territoriality, fitness and accountability needs. Therefore, when can this organizational design be adopted successfully? Presumably, IDS serve a small and geographically concentrated population located in a territory without geomorphologic relevant differences (mountains, seaside, etc.) and with political homogeneity. In other words, we refer to IDS as those serving one single undifferentiated market. When geographical and political differentiation increases, IDS need to move toward a market-based diversified structure. In fact, as the territoriality, fitness and accountability needs increase, IDS should look for the introduction of geographical responsibilities. The P-design can be progressively modified with the introduction of market responsibilities, where markets represent sub-areas (populations) in which the IDS territory can be divided. For instance, in IDS, a market responsibility might correspond to a district. Fig. 2 shows the nine districts in which an Italian Local Health Authority (LHA) is organized. The LHA is one of the three of the metropolitan area of Milan, Italy. It is in charge of providing health services to the northern population (1.3 million of citizens) of Milan hinterland (Lombardy Region, Italy). In this case, the choice of the districts reflect differences in geomorphologic characteristics (southern districts are urban and heavily populated, while northern ones present hills and medium size towns, eastern ones are rural, less populated and with longer distances between villages) and in political parties in power. In each district the delivery of health services is planned and organized according to the demands and needs of the population. A first option to modify the P-design by introducing market responsibilities is represented by the introduction of area responsibilities as the second level of management (Fig. 3). In this structure, specialization remains the leading criteria for labor division, but area responsibilities are second-level criteria for personnel and resources sub-grouping. We still expect a high degree of homogeneity in services delivery thanks to the centralized departments (procedures, protocols, therapies, etc.), but paired with a better response to the fitness need through a “tailored” design of services in the different areas.

270

F. Lega / Health Policy 81 (2007) 258–279

Fig. 2. The Districts of LHA no. 3 of Milan.

This modification provides a better focus on markets, but does not adequately support the territoriality and accountability needs. It does not provide both for responsibilities in monitoring and governing overall demand patterns and interlocked expenses, and for decentralized community and political organizational interfaces. Also, it does not solve the fragmentation issue. Complex diseases that require multidisciplinary interventions are still “black holes”. In this view, it is the structure that fits medium-size IDS, where population has homogeneous trends in health service demand across the territory, and geomorphology (with its consequence on accessibility, transportation, etc.) is not an issue. However, the size of the IDS – in terms of personnel – should be enough to allow its decentralized allocation to areas without compromising (too much) economies of scale associated with its previous centralized management. If greater flexibility is required because one specialist must cover more than

one geographical area, a different structure could be implemented. This is the case of a P-design modified with market coordinators. The coordinators assume transversal roles and responsibilities that are aimed at ensuring the best possible answer to the health needs of their market’s population (Fig. 4). On one hand, they are in charge of stimulating and monitoring the performances of specialized departments in their territory, and on the other hand, they are responsible for assessing the needs met or unmet. Market coordinators can be introduced when territorial characteristics require a different attention to different areas and when IDS size, in terms of employees, does not allow for a further sub-division of specialized departments/units in area responsibilities. Market coordinators are weak roles, out of the hierarchical chain, and are required to use their expertise and leadership skills to influence top management and department managers’ policies. In this context, they play a techno-

F. Lega / Health Policy 81 (2007) 258–279

271

Fig. 3. P-design modified with market responsibilities.

structure role, having the task of monitoring needs, performances and healthcare expenses in their market, and providing information to influence decisionmaking by department managers. Therefore, they support the IDS in coping with the territoriality need, ensuring that interlocked activities and expenses (phar-

maceutical expenses, hospitalization rate, volumes of out-patients diagnostic exams, volumes of outpatients specialist visits and therapies, rehabilitation, long term care and home care expenses) are planned, managed and controlled with reference to the geographical area, where the interconnections are stronger.

Fig. 4. P-design modified with area/market coordinators.

272

F. Lega / Health Policy 81 (2007) 258–279

Although without formal organizational power, they can also play a role in ensuring the fitness of the IDS by trying to influence line managers’ policies on: • integration of fragmented services, specifically the ones relative to multidisciplinary diseases; • mix and location of services delivered, so as to make them meet the market needs as best as possible. In any case, the last word stays with centralized department managers, which are responsible for the operating budget and production objectives. Their contribution, instead, to support the IDS need for accountability towards its stakeholders is very weak. This is better met by a design encompassing the next step in reinforcing the market coordinators’ roles: we refer to the P-design modified with market managers acting in a quasi-market context (Fig. 5). The shift from a coordination role to a management one means the market manager has the power to define the providers’ production in its area: the system works through a budgeting negotiation between market managers and department managers. Mix, location, production level, and quality of services are contracted and “bought” by the market managers on the basis of market or internal transfer prices. In this view, the organization’s financial resources are first assigned to market managers, often on a per capita basis and weighted to consider differentiation factors of markets’ populations and territories such as age, prevalence of diseases, geomorphology, seasonality, etc. The market managers are then in charge

of programming and contracting with unit managers accountable to the delivery of services. This design defines a line of responsibility which supports the IDS needs of territoriality and accountability quite well. The fitness need is indirectly met through the budgeting negotiation, although conflicts might arise between unit/department managers and market managers when resources are constrained resulting in services delivery processes partially unfitted. It is common that in a period of rationing of expenses unit managers “defend” their performances, sustaining the impossibility of meeting the whole market’s health needs due to resource shortages. After all, they are the ones in charge of defining the technical and operational choices regarding service delivery. Market managers try to influence them in the budgeting process. In this context, it is important that market managers play a role in the performance assessment and definition of the production bonuses of unit managers, so as to reinforce their authority in budgeting negotiation and goals setting. 7.2. “G-designs” and their features The second pole of the continuum is represented by a market-based design. In this structure, organizational activities are grouped around the geographical areas where they are delivered. From an organizational point of view, it is still a divisional model, one that is used in industrial contexts, but quite unusual and innovative when referred to the healthcare context. In

Fig. 5. P-design modified with area/market managers acting in quasi-market context.

F. Lega / Health Policy 81 (2007) 258–279

fact, with this design, the relevance of specialization – connected to professions, products or institutions as basis for units – is clearly subordinated to integration needs, and specialists are managed by a head which might have a different background and expertise. In each market division, the manager should be able to govern the full spectrum of health services, including the hospital located in that specific district. The marketdistricts are the IDS operational line. Multidisciplinary and multi-professional teams are in charge of assessing the health needs, planning, and delivering services. Top management discusses goals and budgets with market managers, who have direct control over personnel and resources. In this framework, centralized departments play a techno-structure role being in charge of ensuring (Fig. 6): • The quality of specialized human resources working in the districts, providing training, guidelines, procedures, benchmarking data, etc. They are the professional points of reference where specialists operating in the different districts gather to share their expertise, to acquire new knowledge, to develop their skills and competencies. They have the task of

273

bringing the same quality standards to the specialists operating in the different districts. • The homogeneity and quality of services delivered in different districts, so as to guarantee to the people residing in different areas the same opportunities of access to care. In this context, they are required to use their expertise and leadership skills to influence market managers’ policies on mix, location and access to services. For instance, diabetes’ patients should be treated according to the same clinical pathway in all districts, possibly the one most appropriate and evidence-based. What could and should change is the organizational path, which will be tailored to the district specificities (geomorphology, transportations, facilities, etc.). Dixon [23] has observed that one of the initial failures of the district reorganization of UK national health system was due to the enforcement of one micro-organizational design for all, which was not applicable in many contexts. Market divisions can also be provided with administrative personnel to enhance their autonomy through the partial decentralization of operating procedures, such as low-volumes purchases, human resources

Fig. 6. Example of a “G-design”.

274

F. Lega / Health Policy 81 (2007) 258–279

selection, technology and infrastructure maintenance programs, etc. Part of the administrative functions and personnel are kept centralized for scale economies’ purposes. The G-design enhances accountability with respect to clients’ needs and to other stakeholders’ expectations, it creates better opportunities in building the continuum of care and a multidisciplinary approach to complex diseases, it allows for inter-specialty flexibility in using common resources (e.g. nurses, facilities, technologies, etc.), and it ensures same standards of service quality for district populations. One very important aspect connected to the market divisions is their accountability to municipalities, local authorities and other political organizations operating in its territory. Districts are a clear reference point for administrators lobbying for their citizens. Building good relationships between the market manager and the administrators enhances opportunities and possibilities for changes in the services provision (for instance, closure and transformation of hospital in outpatient centers, reconfiguration of points of delivering, etc.). On the contrary, the G-design creates a risk of differentiation in quantity, quality and accessibility of services across markets – which should be prevented by the monitoring and control exerted by techno-structure units – and increases the rigidity in the specialists’ utilization. Professionals assigned to a district can be reappointed to other areas uniquely either by collaborative agreements between market managers or by top-management enforcements decisions. However, since all market managers have goals to achieve, it is quite difficult that they agree on depriving their market of a specialist to lend it to another district in need. Action by top executives is often necessary to manage temporary or permanent reallocation of personnel, since it normally involves a renegotiation of objectives with the “lending” market manager. Markets-based design fits well in medium-large IDS, serving a large and geographically dispersed population living in a territory with geomorphologic relevant differences (mountains, seaside, etc.) and/or with different political parties in power. A IDS is in charge of multiple differentiated markets. In the G-design, provision of services is guaranteed by work teams. The second hierarchical layer of a market/district can be then defined according to four alternatives micro-structure designs (Fig. 7):

1. replication of discipline-based units, where specialists and other employees are regrouped in functional teams; 2. multidisciplinary customer-based teams, where specialists and other employees are regrouped in teams in charge of providing services for a specific segment of customers; 3. multidisciplinary integrated health centres, where specialists, general practitioners and other employees are regrouped locations providing care to all kinds of patients. 4. programs-based organizations, where specialists, general practitioners and other employees are regrouped in teams around clinical processes, as care pathways. Teams are in charge of improving the disease management, especially of chronic patients, through a better continuum of care. Each team defines a disease manager (frontline operator) for each patient, and is responsible for activating the specialist and other professionals to be involved in the care process on the basis of the patient needs assessment done by the team’s evaluation unit. Variations among the four basic models provide other alternatives. 7.3. Matrix design The matrix design involves a structure with a dual authority system based on both disciplines–products– clients and market responsibilities (Fig. 8). Organizational activities are contemporarily regrouped under both lines of authority, which have the same relevance in the eyes of top management. The second level is of two-boss managers, that is to say those managers who assume responsibility in a market and in a specific discipline, a product line or a client segment, representing the key factor in this design. All the physicians, paramedics and the other professionals are subordinate to a two-boss manager. The bosses play a technical role, which means they are involved in the planning and budgeting process, but do not directly manage organizational resources. They participate equally in goal setting, performance evaluations, and supervision. Resources are managed by the two-boss managers, who negotiate their objectives with both bosses. Two-boss managers do not depend hierarchically on one of the two bosses. In this case, the matrix structure would recollapse towards one of

F. Lega / Health Policy 81 (2007) 258–279

275

Fig. 7. Alternative market/district or health centres micro-structure designs for community services.

the two poles of the design continuum, working as a markets-based or a disciplines–products–clients-based structure. All first level managers refer and respond to top management on two-boss manager performances, but only top management should determine the two-boss manager career. This is to avoid a biased influence from one of the bosses over the decision-making of the twoboss manager. The matrix design recognizes the organization’s need for a contemporary multi-focus, which might involve specialization, product, market or client results. To function properly, the matrix design requires very refined operating mechanisms, such as planning, budgeting and control system, human resources management system, information system, etc. The dual authority determines a complex decision-making process, with the risk of looping and conflicts arising

between the two bosses, who strive to obtain from twoboss managers the utilization of (scarce) resources for their goals. Moreover, it generates relevant costs due to dual accounting, budget, control, performance evaluation and reward systems. Some authors [15] argued that this design is so complex and it is not necessary to achieve the desired integrative benefits. In our view, given its features, the matrix design might be adopted by very large IDS, facing an extremely complex environment characterized by the presence of several differentiated sub-markets and with enough staff to be efficiently and effectively reallocated into the two-boss units. In small and medium IDS, the size of the staff would hinder the possibility of its reallocation into the numerous two-boss units determined by the matrix scheme. The G-designs are, therefore, preferable.

276

F. Lega / Health Policy 81 (2007) 258–279

Fig. 8. Matrix design.

8. From boxes to people The discussion of different design alternatives has highlighted the contingent approach required to match IDS specificities with the most effective structure of responsibilities. Table 1 synthesizes the relationships between complexity factors, the design choice and the organizational needs met. However, it is important to underline that identifying the right design is never that simple, nor a mechanistic task. This is the case for three relevant reasons. First, any organizational structure shaping responsibilities requires mechanisms to reinforce the formal authority defined by the hierarchy: these mechanisms are the organizational support systems (e.g. the planning, budgeting and controlling process, the human resources management system, the information system, etc.). Such mechanisms should be designed coherently with structure in order to reinforce them. Differently, they might create opportunities for the increase of spaces and influence of informal authority contributing to de-legitimatizing the desired organi-

zational design. IDS and its comprising organizations have histories and contexts for which this might prove difficult – at least in the short time – and to develop the organizational mechanisms required for the functioning of the desired design. Second, as Mintzberg [42] once underlined, health organization “are constantly reorganizing, which means shuffling boxes around on pieces of paper. Somehow, it is believed that by rearranging authority relationships, problem will be solved. But all this may reflect no more than the frustration of managers in trying to effect real change in the clinical operations. Architecture might offer a much more effective solution: if even a fraction of the efforts that are put into moving positions around on charts went into moving people around on floors, there might be an awful lot more collaborative activities. . .”. In this view, the focus of this work is both its strength and its limit. It is explicitly aimed at support policy-makers or managers in designing IDS, while it seems considering the people element as a variable to be accounted for during the implementation. However, it is clear that this

F. Lega / Health Policy 81 (2007) 258–279

277

Table 1 IDS characteristics and most fitted design Design

P-design P-design modified with market responsibilities P-design modified with area/market coordinators P-design modified with area/market managers acting in a quasi-market context G-design Matrix design

Complexity factors Size of the IDS in terms of employees and/or served population

Differentiation in territorial geomorphology

Political and social different expectations

Population concentration

IDS organizational needs that might be met

Small Medium–Large

Low Medium

Low Medium

High High

A; P; M I; A; P; F; M

Small

Medium

Medium

High

I; A; P; T; M

Medium–Large

High

Medium

Medium

I; A; P; F; T; M

Medium–Large Large

High High

High High

Low Low

I; A; P; F; T; R; M I; A; P; F; T; R; M

I: integration; A: anticipation; P: proximity; T: territoriality; F: fitness; R: accountability; M: manageability.

is not the case, as organizations are made of people. Therefore, and third, as already discussed in paragraph 5, people might be or might not be ready for the roles the organization assigns to them and they might be or might not be supportive of expected change [20]. This might affect the IDS design in a dual way. On one side, the IDS might be in the situation of presenting features which lead to the matching of more than one design. How to choose? By looking at peoples’ skills, and evaluating if there are managers up to the roles required in the different design. It is irrelevant, however, to deeply discuss this issue that has been extensively analyzed by organizational behaviour scholars, but it might be a topic for future research. On the other side, the one-best match on the paper might not be pursued immediately because the IDS managers are not ready or not willing to manage the roles they are required to play (especially when they are drawn from the ranks of clinicians). After all, traditionally the design of healthcare organizations has been determined primarily by the staff self-interest, whereas the distribution of medical work and the way in which doctors are organised have been influenced more by professional interests, financial inducements, rivalries and career aspirations than by analysis of the needs of the patients.13 In general, the influence of power and politics on organizational design choices is well known 13

See the leading article, A flawed system, Lancet i, 463–464, 1980.

in literature and is extremely important when looking for change [43–45]. The least in this case is to deploy a plan of organizational development to manage the transition from present design to the one desired by preparing and “escorting” the managers towards their new roles through adequate training and tutoring. This is a common critical issue in any organizational change.14 People development is the key to successful or unsuccessful redesigns. In this article, we have analyzed the factors which affect the theoretical design of IDS structures. However, a good assessment of both IDS complexity factors and middle managers’ skills is fundamental for successfully designing choices. Design should match the complexity factors, but its success

14 According to Farrar [46] and Eastman and Fulop [47] change in healthcare organization is difficult because the structure of incentives might be inadequate and cultural differences persist. In most case clinicians are drawn from the ranks of doctors who will most likely view management as an inferior and less rigorous area of endeavour and one that does not even remotely compete with the status or recognition gained from medicine (especially academic medicine). As proof of this resistance, other authors, especially clinicians, have clearly argued against the clinician–manager role. Chant’s [48] words summarised this view: “if the present practising doctor–manager ideas are pressed, I fear many of us who should be primarily concerned with our patients’ well-being will find our energies dissipated in resource allocation disputes with our colleagues, for which we have little knowledge, minimum training and little stomach”.

278

F. Lega / Health Policy 81 (2007) 258–279

depends on the ability of people to understand, share and play the expected roles. In the end, the redesigning of the structure is just one element – although fundamental – of a complex and long change process aimed at progressively aligning the organization and its people to the projected design. Successful implementation starts with the engagement of people (especially clinicians) from the inception of the redesign process. References [1] Charns M. Organization design of integrated delivery systems. Hospital Health Services Administration 1997;42(3):411–30 [fall]. [2] Fischbacher M, Francis A. Managing the design of health care services. In Davies H, Tavakoli M, Malek M, Neilson A, editors. Managing quality: strategic issues in health care management. Aldershot, Ashgate; 1999. [3] Parker V, Charns M, Young G. Clinical service line in integrated delivery systems: an initial framework and exploration. Journal of Healthcare Management 2001;46(4). [4] World Health Organization (WHO), Regional Office for Europe. European health care reforms: analysis of current strategies. Copenhagen: World Health Organization; 1996. [5] World Health Organization (WHO), Technical Report Series. The role of health centres in the development of urban health systems. Geneva: World Health Organization (WHO); 1992. [6] Shortell SM, Gillies RR, Anderson DA, Morgan Erickson K, Mitchell JB. Remaking healthcare in America: building organized delivery systems. San Francisco: Jossey-Bass; 1996. [7] Shortell SM, Kaluzny DA. Health care management: organization design and behaviour. 4th ed. Albany: Delmar; 2000. [8] Leatt P, Baker GR, Helverson PK, Aird C. Downsizing, reengineering and restructuring: long-term implications for healthcare organizations. Frontiers of Health Services Management 1997;13(4):27–39. [9] Kassirer JP. The next transformation in the delivery of health care. The New England Journal of Medicine 1995;5:52–4. [10] Leatt P, Shortell SM, Kimberly JR. Organizational design. In: Shortell SM, Kalunzny AD, editors. Health care management: organization design and behavior. III ed. Delmar Publishers; 2000. [11] Herzlinger RE. Market-driven health care. Reading, MA: Addison-Wesley Publishing Company; 1997. [12] Mossialos E, Le Grand J. Health care and cost containment in the European Union. Aldershot, Ashgate; 1999. [13] Mintzberg H. Structuring in fives: designing effective organization. Englewood Cliffs: Prentice-Hall; 1983. [14] Duncan R. What is the right organization structure? Decision tree analysis provides the answer. Organizational Dynamics 1979;(Winter):59–79. [15] Charns M, Tewksbury L. Collaborative management in health care: implementing the integrative organization. San Francisco: Jossey-Bass; 1993.

[16] Herzlinger RE. Healthcare Advisory Board (1994) Network Advantage, Washington, DC. In: Market-driven Health Care. Reading: Addison-Wesley; 1997. [17] Fulop N, Protopsaltis G, Hutchings A. Process and impact of mergers of NHS trusts: multicentre case study and management cost analysis. BMJ 2002:325. [18] Conrad AD. Coordinating patient care services in regional health systems: the challenges of clinical integration. Hospital and Health Services Administration 1993;38(4):491– 508. [19] Currie G, Procter S. In: Rushmer RK, Davies HTO, Tavakoli M, Malek M, editors. The role of middle managers in realising human resources strategy: evidence from the NHS. Ashgate, Aldershot, UK: Organisation Development in Health Care; 2002. [20] Griffith JR. Championship management for healthcare organizations. Journal of Healthcare Management 2000;45:1. [21] Charns M, Smith L. Product line management and continuum of care. Health Matrix 1989;VII(1):40–9. [22] Leander WJ. Product line management. Formalizing clinical product lines as part of patient focused restructuring. Review of Patient-focused Care Association 1993:2–5. [23] Dixon M. The organisation and structure of units. In: Wickings I, editor. Effective unit management. London: King Edward’s Hospital Fund for London; 1983. [24] Longest B. Managerial competence at senior levels of integrated delivery systems. Journal of Healthcare Management 1998;43:2. [25] Maynard A. Hospital mergers: blissful ignorance or high risk taking? British Journal of Health Care Management 1997;3(10):512. [26] Snail TS, Robinson JC. Organisational diversification in the American hospitals. Annual Review of Public Health 1998;19:417–53. [27] Schein EH. Process consultation, vol. I. Reading: AddisonWesley; 1988. [28] Watzlawick P, Weaklund JH, Fish R. Change: principles and problem formulation. New York: W.W. Norton; 1974. [29] Quinn JB. Strategies for change. Homewood, IL: Richard D. Irwin; 1980. [30] Ackerman L. Development, transition or transformation: the question of change in organizations. OD Practitioner 1986: 1–8. [31] Jick TD. Managing change: text and cases. Homewood, IL: Irwin; 1993. [32] Argyris C. Overcoming organizational defenses: facilitating organizational learning. Needham, MA: Allyn & Bacon; 1990. [33] Argyris C. Knowledge for action: a guide to overcoming barriers to organizational change. San Francisco, CA: Jossey-Bass Publishers; 1993. [34] Schein EH. Organizational psychology. Englewood Cliffs, N.J: Prentice-Hall; 1965. [35] Weick K. The social psychology of organizing. 2nd ed. Reading, Mass: Addison-Wesley; 1969/1979. [36] Som CV. Nothing seems to have changed, nothing seems to be changing and perhaps nothing will change in the NHS: doctors’

F. Lega / Health Policy 81 (2007) 258–279

[37] [38] [39]

[40]

[41]

response to clinical governance. International Journal of Public Sector Management 2005;18(5):463–77. Leung WC. Managers and professionals: competing ideologies. BMJ Career Focus 2000;321(S2):7266. Mathie AZ. Doctors and change. Journal of Management in Medicine 1997;11(6):342–56. Parry J. Making sense of executive sensemaking: a phenomenological case study with methodological criticism. Journal of Health Organization and Management 2003;17(4):240–63. Wallace LM, Freeman T, Latham L, Walshe K, Spurgeon P. Organisational strategie for changing clinical practice: how trusts are meeting the challenges of clinical governance. Quality in Health Care 2001;10:76–82. Lefkovitz PM. The continuum of care in a general hospital setting. General Hospital Psychiatry 1995;17(4):260–7.

279

[42] Mintzberg H. Toward healthier hospitals. Health Care Management Review 1997;22(4):9–18. [43] Perrow C. Complex organizations. 3rd ed. N.Y: Random House; 1986. [44] Pfeffer J. Power in organization. Marshfield, MA: Pitman Publisher; 1981. [45] Pfeffer J. Managing with power: politics and influence in organizations. Boston: Harvard Business Press; 1992. [46] Farrar S. NHS reforms and resource management: whither the hospital? Health Policy 1993;26(2):93–104. [47] Eastman CJ, Fulop L. Management for clinicians or the case of ‘bringing the mountain to Mohammed’. International Journal of Production Economics 1997;52(1/2):15–30. [48] Chant AD. Practising doctors should not manage. Lancet 1984;1(8391):1398.