Report No. 35691-BR Brazil Enhancing Performance in Brazil's Health Sector Lessons From Innovations in the State of São Paulo and the City of Curitiba October 27, 2006 Brazil Country Management Unit Poverty Reduction and Economic Management Unit Latin America and the Caribbean Region Document of the World Bank Estatutiirios Statutory government employees covered by the RJU Fundaqlo Serviqos Especiais de Sahde Piiblica FSESP Foundationfor Special Services inPublic Health Gerente do Centro de Vigildncia Sanitiiria HealthDistrict Manager Gestlo Plena FullManagement Gratificaqlo Especial para MCdicos GEM Special Bonus for Doctors Instituto de Pesquisae de Planejamiento de IPPUC Researchand UrbanPlanning Institute of Curitiba -Dobled Curitiba Instituto de Pesquisae Planejamento Urbano de IPPUC Curitiba Research andUrban Planning Institute Curitiba Instituto Municipal de Administra@o Piiblica Ih4AP Municipal Institute o f Public Administration Laboratbrio Municipal Municipal Laboratory Lei de Diretrizes Orqamentirias LDO Budget Planning Law Medicina con Base em Evidencias EBM Evidence Based Medicine MinistCrio da Administraqlo Federal e Reforma MARE MinistryofFederalAdministrationand State do Estado Reform Movimento de Reforma Sanitiiria Healthcare reform movement Nova GerCnciaPiiblica NPM New Public Management Organizaqiio para Cooperaqlo EconBmica e OECD Organization for Economic Co-operation and Desenvolvimento Development Organiza@o Social os Social Organization Organizaqdes Sociais de Saiide oss Social Organizations inHealth Pesquisa de Rastreamento de Gestiio Nblica PETS Public Expenditure Tracking Survey Plano Incentivo a Qualidade PIQ Quality Incentive Plan Plano Operativo Annual POA Annual Operating Plan PrCmio de Incentivo Especial PI Special Incentive Award Produto Interno Bruto GDP Gross Domestic Product Programa de Agentes Comunithrios de Sahde PACS Community HealthAgents Program Programa de Incentivo ao Desenvolvimento da IDQ Incentive Program for Quality Development Qualidade dos Serviqos Programa M l e Curitibana Curitiban Mother Program ProgramaNacional de Apio B ModemizaqBo da PNAGE National Program for Support of Modernization Gestiio e Planejamento of Management andPlanning ProgramaSaiide da Familia PSF Family Healthprogram Prontuirio EletrBnico de Sahde Electronic medical records system Regime Juridic0 Unico RJU Single Juridical Regime Reino Unido UK United Kingdom RelaGlo entre Desempenho e Salirio PRP Performance-related pay Secretaria Estadual de Sahde SES State Secretariat o f Health Secretaria Municipal de Saiide SMS Municipal Health Secretariat Servidor estatutiirio Public servant Sistema de Informaqlo Gerencial SIG Management Information System Sistema Integrado de ServiGosde Salide SISS Integrated System o f Health Services Sistema Nacional de Saiide N H S National Health System Sistema Unico de Saiide sus UnifiedHealth System Sistemas integrados de Serviqos de Saiide - doble SISS Integrated Systems o f Health Services Sociedade Paranaensede Saiide da Familia FAMIPAR Family Health Society o f Paarna Termo de Compromisso de Gestiio Management Contract Unidade de Sahde HealthUnit Unidade de Tratamento Intensivo UTI Intensive Care Unit ACKNOWLEDGEMENTS This report is the result of collaboration between the World Bank's Latin American and Caribbean PREMPublic Sector Group and the LAC Human Development Network. It forms part of a series of studies the Bank has carried out to examine the quality of public expenditures in Brazil. Recent Bank studies that touch on related subjects include one on efficiency of education spendingby Brazilian municipalities (World Bank 2002) and a study on health decentralization in the State of Bahia (World Bank 2003). Other related work that is currently under way includes a study on hospital performance (World Bank 2006a, forthcoming), an overview of federal public expenditures (World Bank 2006b, forthcoming), and a public expenditure tracking survey in the health sector (World Bank 2006c, forthcoming). The task managersof the study were Yasuhiko Matsuda (LCSPS) and Jeffrey Rinne (LCSPS). The core team members included Geoffrey Shepherd (consultant), Juliana Wenceslau (LCSPS), and Maria Virginia Hormazabal (LCSPS). Francisco Gaetani, GerardL aForgia, andFernando Rojas (LCSPS) provided valuable peer reviews. The team also wishes to thank April Harding (LCSHH), Chris Pare1(LCSPS), andNick Manning for their helpful comments. Ethan Weisman (LCSPR), as Lead Economist for Brazil, provideddetailedcomments and general guidance to the team infinalizing the report. Chapter 2 is basedon the backgroundpaper "Alternative Service Delivery inSlo Paulo's Health Sector: Comparing Personnel Management & Performance inTraditional, "Social Organization" and Private Hospitals," by Jeffrey Rinne. Fieldresearch for that chapter was carried out by Vit6ria Kedy Cornetta (consultant). Chapter 3 was written by Geoffrey Shepherd, with contributions from Juliana Wenceslau and Samuel Jorge MoysCs (consultant). MoysCs also prepared a backgroundsurvey for this chapter, Avaliapio dos Incentivos aos RecursosHumanos nu AtenGio Primdria em Salide de Curitiba, 2006 (Evaluation of Human-Resource Incentives in Basic Healthcare inCuritiba). The task would not have been possible without the initial stimulation by Jerry L aForgia (LCSHH), who helpedthe team in identifying the case studies and offered critical advice and generousassistancethroughout the preparation of the report. Table of Contents EXECUTIVESUMMARY ............................................................................................................ i Improving Public Sector Performance for Higher Quality Public Spending ......................... i Primary Healthcare in Curitiba............................................................................................. Social OrganizationHospitals inthe State of SI0 Paulo....................................................... . . ...ii Overcoming Agency Problems: Conclusions andPolicy Implications................................. 111 v 1. ImprovingPublicSector Performancefor HigherQuality PublicSpending ................1 Improving Service Delivery inBrazil ................................................................................... 2 ManagingPerformance ................................................................................................... Healthcare Policy inBrazil . . .............................................................................................. 4 5 Assessing Managerial Innovations inBrazil's Health Sector ............................................... Case Selection ....................................................................................................................... 9 8 2. Social OrganizationHospitalsinthe State of S I 0 Paul0 ................................................ 11 Introduction ......................................................................................................................... 11 Performance Differencesbetween OSS and Direct Administration Hospitals ...................14 Contractingfor Healthcare in SiioPaulo's OSS Model ...................................................... 11 Human Resource Management in Siio Paulo's Hospitals ................................................... 16 Comparing Human ResourceManagement inOSS and Traditional Public Hospitals .......18 Multiple Jobs................................................................................................................. Legal Basis of Employment .......................................................................................... 16 19 Establishment Control and Staff Composition 19 Personnel Selection Salary Determination..................................................................................................... ....................................................................................................... .............................................................. 21 Performance Pay ........................................................................................................... 21 23 . Performance Evaluation and Supervision ..................................................................... Career Development...................................................................................................... 23 25 The Power of Informal Incentives(through formal and informal rules) ............................. Disciplinary Actions...................................................................................................... 26 26 Conclusions & Recommendations ...................................................................................... 30 3 . ManagingPrimaryCareinthe City ofCuritiba ............................................................ 32 Introduction ......................................................................................................................... 32 Primary Healthcare in Curitiba ............................................................................................ 32 The Approach to Public Healthcare ............................................................................... .............................................................................. 32 Primary Healthcare Outcomes....................................................................................... The MainFeatures of Management 37 The Evolutionof the Primary Healthcare System......................................................... 37 40 The Management of Primary Healthcare ManagingInformation................................................................................................... 45 Client Orientation.......................................................................................................... ............................................................................ 43 44 ManagingHumanResources......................................................................................... Specialization and Coordination ................................................................................... 47 48 ManagingPerformance ................................................................................................. 52 The Elements of GoodPerformance ................................................................................... 56 The Environment for Performance I: .................................................................... The Environment for Performance 11:Health Professionals' Networks ....................... SUS 58 The Environment for Performance 111:City Management............................................ 58 59 Challenges for Performance................................................................................................ The Environment for Performance: a Confluence of Factors........................................ 63 64 4. OvercomingAgency Problems: ConclusionsandPolicy Implications ......................... 66 Social Organizations as Corporatization ............................................................................. 66 Making the Managers Manage through Performance Contracts and Letting Managers CorporatizedHospitals in Slo Paulo............................................................................. Manage through Flexibility........................................................................................... 66 Performance-Based Pay inthe Public Sector...................................................................... 67 70 Pay for Performance inCuritiba: Not a Magic Bullet ................................................... InternationalExperience to Date................................................................................... 70 Conclusions andPolicy Implications .................................................................................. 74 Managerial Flexibility or OrganizationalDevelopment?.............................................. Clarity and Consistency of Purposes: Institutional and Informational Foundations .....76 77 .......................................................................................................................... 78 Policy Implications........................................................................................................ Lessons 79 80 APPENDIX A. Hospitals inSloPaulo ResearchSample ........................................................... ' 83 APPENDIX B ResearchQuestionnaires Applied inSlo PauloHospitals.................................. . 84 APPENDIX C. FocusGroup Interviews inSlo Paulo Hospitals................................................. 91 APPENDIX D. Survey Summary: Avaliaglo dos Incentivosaos RecursosHumanosnaAtenglo PrimAriaem Saiide de Curitiba ...................................................................................................... 93 APPENDIXE Managing Curitiba's Public Health System: Instruments andImpacts...............96 . December. 2004............................................................................................................................. APPENDIXF. Selected Management Contract Targets by HealthDistrict inCuritiba. January . 99 BIBLIOGRAPHY ...................................................................................................................... 101 TABLES Table 2.I Patient Indicators of Allocative Efficiency inOSS & Direct Administration : Table 2.2: Occupancy Rates by Hospital Type (among hospitals instudy sample) ..................... Hospitals. 2003........................................................................................................... 15 Table 2.3: Mortality Rates inDirect Administration & OSS Hospitals. 2003.............................. 15 Table 2.4: Form of Employment by Medical Speciality inSlo Paulo's Public Hospitals. 18 Table 2.5: Forms of Employment (among hospitals in study sample) ......................................... ..........15 18 Table 2.6: Staff Composition inDirect Administration & OSS Hospitals. 2003 Table 2.7: Number of Employees per Hospital Bed (among hospitals in study sample) .............20 ......................... 21 Table 2.8: Average Elapsed Time betweenInitiating a HiringProcess & Employee Arriving to Work (among hospitals instudy sample) ............................... ........................ 21 Table 2.10: Ratesof Absenteeism ................................................................................................ Table 2.9: Comparing Employment Benefits (among hospitals in study sample) 22 25 Table 2.1 1: Employee Turnover Rates (among hospitals instudy sample).................................. Table 3.2: Municipality of Curitiba: Employmentby Type of Unit and Type of Employee .......33 Table 3.1: Number of Public and PrivateHealthUnitsinCuritiba. 2005 .................................... Table 2.12: Managerial discretion over human resource decisions. by hospital type...................25 26 34 Table 3.3: Municipality of Curitiba: Average Number of Staff Employed in Health Unitsby Type of Unit..................................................................................... 36 Table 3.4: Comparison of Curitiba's Health Indicators with other Municipalities with similar epidemiologic and demographic profile.2004 ............................................................................... 39 Table 3.5: Chronology of EventsRelatedto Health Sector Management ............................................................................................... 40 Table 3.6: Curitiba: PopulationGrowth. 1940-2005 .................................................................... inCuritiba. 1941-2004 Table 3.7: Performance under the IDQ.2002-2005 ..................................................................... 43 51 Table 4.2: Replies to employee attitude surveys in selectedpublic service organizations...........71 Table 4.1: FourTypes of Public Organization .............................................................................. 73 Table 4.3: Local"Push" and "Pull" Factors inthe Choice of Healthcare Performance Management Strategies.............................................................................................. 80 Table 4.4: Summary of Policy Instruments and their Incentive Effects inthe Two Case Studies82 BOXES Box 1.1: Social Organizations ........................................................................................................ Box 2.1: Why OSS hospitals appeared in STio Paulo.................................................................... 4 14 24 Box 3.1: The Role of Protocols .................................................................................................... Box 2.2: The special incentive award........................................................................................... .......................................................................................... 46 Box 3.3: Management Contracts -One HealthDistrict's View................................................... Box 3.2: The IDQ Evaluation System 50 Box 3.4: Networks and Performance Incentives........................................................................... 55 60 Box 3.5: New Zealand Tries to Understand Curitiba ................................................................... 61 ................................................................... Box 4.2: Have Team Incentives Worked inCuritiba? Some Speculation .................................... Box 4.1: Performance Contracts inthe Public Sector 68 77 FIGURES Figure 4.1: Outline of the "Expectancy" Framework ................................................................... 74 EXECUTIVESUMMARY IMPROVINGPUBLICSECTOR PERFORMANCEFORHIGHER QUALITY PUBLICSPENDING The inadequacy of public services in Brazil today encumbers the country's economic growth and social development. The volume of government expenditure is not the principal bottleneck: at over 40 percent of GDP, Brazil's total public spending is already much higher than that of comparable middle-income countries. The vital question, therefore, is how to get greater value for public money. As a federal country, improving service delivery in Brazil calls for strengthening the incentives and institutional capacities of sub-national governments. However, it is precisely at the sub-national level where Brazil faces its greatest challenges for institutional development. Therefore, this report examines how particular managerial innovations have been applied in two sub-national jurisdictions in Brazil to deliver improved public services from available resources. From a broad range of public services, we have chosen to focus on healthcare inthis report. The health sector inBrazil absorbs approximately four percent of GDP, and its spending level is likely to rise further given the country's demographic and epidemiological profiles. Thus, improving the efficiency of public health spending can have effects beyond the quality of health services, as uncontrolled spending growth could affect the overall health of the country's public finance and macroeconomic stability in the long run. Current efforts to improve health service delivery in Brazil should be examined against the background of at least two parallel developments in the 1990s. One the one hand, at the federal level Brazil launched a public management reform agenda in the mid-1990s that subsequently influenced approaches to public sector reform throughout the country. Inspired by that reform discourse, a number of sub-national jurisdictions in Brazil have embarked upon managerial reforms to improve public service delivery. On the other hand, the health sector itself has been going through important policy and institutional transformations since the early 1990s, following the model of decentralized health service delivery. One of the approaches to service delivery improvement proposed as part of the federal public sector reform agenda in the 1990s involves the use of "Social Organizations" to deliver public services via non-government organizations certified to manage public funds for the purposes specified in management contracts. The 1990s reform efforts also included measures to make personnel management in direct public administration more performance-oriented, through legal reforms to increase managerial flexibility (e.g., managers' ability to fire staff for poor performance) and/or management instruments to enhance public servants' incentives to perform (e.g., performance evaluation and pay). Since the late 1980s Brazil's health sector has gone through considerable transformation both in terms of structure (e.g., decentralization, the adoption of the Unified Health System (SUS)), and from a reasonably coherent sector-wide policy framework (SUS) - certainly superior to other policy content (e.g., greater emphasis on primary care). The health sector in Brazil now benefits decentralized sectors such as public security or water supply and sanitation. Nonetheless, the effectiveness of the Brazilian federation in providing public healthcare is still limited, as a majority of the sub-national jurisdictions, particularly at the municipal level, suffer from weak institutionalcapacities their effectiveness andefficiency as healthcare providers. 1 Within thisbroader context, the State of SaoPaulo and the City of Curitiba offer two examples of innovative healthcare management reforms: "corporatization" of public hospitals in the State of Slo Paulo, and performance management in the Municipality of Curitiba. On the surface, these innovations appear to follow the now familiar New Public Management (NPM)prescriptions of "let the managers manage" (managerial flexibility) and "reward performance" (performance- based pay). But the case studies uncover interesting nuances that enrich our understanding of practical, context-specific approaches to performance management within Brazil's existing institutional constraints. By probing behindthe conceptual rhetoric of the reforms, the cases offer insights into "what really matters" in improving public management, particularly with regard to how policy makers can overcome so-called "agency problems." The case narratives show how boththe State of Sao Paulo and the City of Curitiba have deployed a range of organizational and managerial innovations to diminish policy makers' informational disadvantage vis-li-vis front-line service providers (doctors, nurses) and achieve greater alignment between the government's overall policy objectives in the health sector, and the objectives and incentives of the service delivery units and their staff. There is more to management improvement than simply adopting a new organizational model or new managerial slogans. Interestingly, these case studies demonstrate that that within Brazil's present legal and organizational context it is possible to make demonstrable improvements in public service delivery. SOCIAL ORGANIZATION HOSPITALS STATE OFSA0 PAUL0 INTHE In the late 1990s the government of Slo Paulo created Social Organizations in Health (OrganizaCoes Sociais em Satide - OSS) by statute to enable a formal partnership between the state and non-profit, private-sector organizations. Under this OSS model, the government provides budgetary transfers to cover the costs of runningthe hospital, but responsibility for day- to-day administration is delegated to pre-certified non-profit organizations. The State Secretariat of Health (SES) negotiates and signs a performance contract with each of these hospital managers, granting them greater flexibility than their counterparts in traditional state hospitals to run the hospital in the manner they consider best-suited to meet their performance targets. In 2004, 17 public hospitals in Slo Paulo operated as OSS. A systematic comparison of 12OSS hospitals and 10 direct administration hospitals in the State of Slo Paulo (World Bank 2006a) found OSS hospitals were more efficient and provided better quality services than direct administration hospitals. For example, the OSS hospitals inthe World Bank study offered 35 percent more patient admissions for each hospital bed and registered lower overall mortality than direct administration hospitals. What accounts for OSS hospitals' superior performance? Part of the answer is the accountability relationship between the SES and the OSS hospital. The management contract specifies the volume of different services to be performed each month (e.g., inpatient and outpatient services, medical consultations) in exchange for a specified budget (a prospective payment block contract). Ninety percent of the annual budget agreed between the SES and hospital administrator is delivered inmonthly installments. These disbursements will be reduced by 10 percent if the quantity of services delivered falls to 75-84.9 percent of the agreed targets, and by 30 percent if output falls below 75 percent. The remaining ten percent of the budget is delivered quarterly, contingent upon the hospital submitting properly coded data on their patients and the treatments or services provided. ii OSS contracts are fine-tuned through regular dialogue between hospital directors and the OSS supervisory staff of the Secretariat of Health. Adjustments are made from one annual contract to the next, but can also be made by consensus within the operational period of a given contract. Since the hospital's budget depends on meeting the pre-specified performance targets (and submitting well-organized management and performance information to the SES), the hospital managers have a clear incentive to meet the targets. Furthermore, persistent failure to perform could result innon-renewal of the OSS contract. Once accountability and performance expectations are established, the hospital manager still needs the ability to manage the hospital's resources, including personnel, to achieve the agreed performance goals. Based upon our analysis of 20 hospitals in S2io Paulo, including focus-group discussions with healthcare personnel inthese facilities, there is little or no evidence to assert that the superior performance of OSS hospitals results from higher salaries, performance pay, superior career development opportunities, or even formal supervision mechanisms. Our research did find, however, that OSS hospital managersenjoyed greater freedom in choosing a particular mix of staffhkills and in selecting personnel through less rigid recruitment processes. Rather than being requiredto follow a rigid process of competitive entry exam (concurso), whereby hospitals are only permitted to recruit those who have passed the exam in the order of their test results, OSS hospital managers are allowed to recruit staff through a more flexible process, determined by referrals, recommendations, and/or face-to-face interviews. Moreover, contrary to the norm in direct public administration, OSS managers can swiftly fire employees who fail to perform at expected levels. These findings suggest that under certain circumstances (where accountability relations are clearly defined and credibly enforced), performance improvements can result from granting front- line managersgreater flexibility inmanaging human resources. Lookingto the future, one reform option for Brazil's traditional direct administration would be to modify the rules governing concursos to allow hiringmanagers to interview a short list of pre-qualified candidates instead of forcing them to accept the top qualifiers irrespective of the individuals' likely fit with the organization. A manager's ability to discipline poor performance is also an important managerial instrument. While a "rapid path" (via rcipida) procedure was intended to enable public sector organizations to fire poor-performing employees, applicable rules result incontinued delays. Reviewing this rule to allow a speedier process of dismissal could lead to more effective human resource management, although obviously the risk of abuse must be assessedcarefully. PRIMARY HEALTHCARE INCURITIBA Like OSS hospitals in Si0 Paulo, the performance of healthcare management in Curitiba is generally considered superior to that of an average Brazilian municipality. While rigorous comparative analysis to substantiate such a claim is lacking, the overall public health outcomes in Curitiba offer primafacie evidence of the effectiveness of the city's public health system. Over almost three decades of consistent efforts to build its primary healthcare system, Curitiba has constantly innovated in healthcare and managerial practice, in a context of rapidly growing demand for healthcare (thanks to immigration and the devolution of healthcare to local government). Today Curitiba's health management includes several good practices, such as client orientation (e.g., regular use of telephone surveys to measure service quality and detect problems and streamlining of front-office processes to make a patient visit more pleasant and ... 111 convenient), as well as effective use of management information (e.g., development of a sophisticated integrated information system to manage knowledge about patients and their treatments, standardize clinical processes through integrated protocols, coordinate the different parts of the health system, and monitor performance inthe different parts of the health system). Human resources are central to effective performance management inthe health sector. Virtually all of the roughly 5,000 staff of the Municipal Health Secretariat (SMS) are tenured public servants with automatic (rather than merit-based) advancement up the seniority scale. Although they enjoy stability as well as other benefits (e.g., relatively generous public sector retirement benefits). Curitiba has tried a number of bonus schemes to motivate these public servants to become performance-oriented. The first of these schemes (PIQ, 1995) was designed to improve performance by making Health Units compete. However, this scheme failed very quickly: teams were able to cheat on poorly measured indicators and rivalry between teams began to undermine the broader unity of staff inthe SMS. A new incentive scheme to improve service quality (IDQ) was introduced in 2000. A quarterly evaluation determines whether individual employees receive a bonus. The result of the evaluation is the weighted outcome of the supervisor's evaluation of the employee, self- evaluation, SMS evaluation of the Unit (based on performance under management contracts), and community evaluation of the Unit. Inpractice, virtually everybody who does not have a record of excessive lateness or absence gets the bonus. Thus, on its face, the scheme operates only to punish extreme cases of poor performance. Nonetheless, staff believe that the IDQ has had an important impact on performance, although this impact seems to have fallen over time. It may be that, as often happens with workplace innovations, there is a temporary change in behavior until workers readjust to the norm. Inthis case, however, it also seems that the evaluation process itself (i.e., not the rewards and punishments) has contributed to identifying and solving problems in the Health Units. The culture of SMS professional staff is dominated by what might, in shorthand, be called a Health System (SUS) but often not taken seriously by states and municipalities - are used as real strategic-planning mindset. The Annual Operating Plans (POAs) - mandated by the Unified management tools in Curitiba. The POAs are accompanied by simple Management Contracts between the SMS and the Districts and between the Districts and the Health Units. Although these contracts are not formally enforced, they set targets for roughly 60 health outputs and outcomes. The targets are set on the basis of discussions between the parties. The computerized information system and the standardized definitions of procedures coming from the integrated protocols have been vital in minimizing the amount of "gaming" that can go on in the measurementof performance. The way in which Curitiba has combined a variety of managerial instruments (e.g., performance review and bonus, management information system, and standardization of basic care according to protocols) seems to be a key ingredient of its relative success. But the strategic-planning mindset that leads to the effective use of these management instruments likely has deeper roots. The history of Curitiba's administrative development is suggestive. Curitiba is famous the world over for innovative solutions in urban services, especially transport, town planning, and environmentally- and citizen-friendly policies. Most commentary ascribes its successes to the strategic-planning approach born of the systematic urban planning that started in the 1960s in Curitiba. This, in turn, was likely facilitated by the remarkable political continuity that the city has enjoyed over the past two decades. iv It would not be easy to replicate Curitiba's experience. Curitiba's system is the product of particular historical circumstances; it is complex; and the relative smallness of the SMS helps it to manage this complexity. With these cautions in mind, it would be useful for would-be replicators to understand how Curitiba has applied strategic planning: more as a modus operandi (or work habit) than a formal process. That modus operandi has pervaded the organization from the top managers to the operational level. It has involved a mindset that thinks about the future, looks for and solves problems, understands system complexity (and understands that changes in one place may create problems and opportunities elsewhere). It is empirical, experimental, and risk-taking. The formal tools of performance management seem an adjunct to, not a driver of Curitiba's strategic planning. OVERCOMING AGENCYPROBLEMS: CONCLUSIONSANDPOLICY IMPLICATIONS The two cases of public management innovations examined in this report demonstrate two divergent approaches to performance improvement in Brazil's public sector. SBo Paulo has introduced organizational innovation through contracting out hospital management to qualified NGOs in the form of "Social Organizations." This model involves devolution of managerial responsibilities from the State Secretariat of Health (SES) to each of the OSS hospitals, circumventing the well-known constraints that define human resource management in the state's direct administration. Curitiba, in contrast, managed to strengthen performance of its primary Curitiba's ability to make effective use of the estatutdrio public servants - who are often vilified care system within the existing human resource regime. We suggest (but do not prove) that as complacent in their permanent job tenure - to provide client-oriented healthcare i s largely context-dependent (i.e., the product of Curitiba's positive history of public sector development both in the health sector and more broadly). In both cases, a central challenge is to motivate staff and align their incentives with the government's broader policy objectives. Our case studies do not provide a straight-forward blueprint for other sub-national governments inBrazil, but we hope they provide a clearer picture of key elements of the reforms that other governments should consider before embarking on reform processes of their own. There are some broadly common strategic tools (which can be applied at different levels of sophistication); and there are choices to be made about the incentive systems that govern staff and management behavior. The case studies explore two types of management tools: i)those that aim to align the expectations of principal and agent and reduce information asymmetry, and ii)those that provide direct incentives to tie an employee's or manager's behavior to performance outputs. The two types are not intrinsically incompatible, but knowing how to combine these instruments is more an art than a science. Instruments for aligning expectations and reducing information asymmetry between principaland agent: There are a number of things that reform-minded governments can do to better align expectations and incentives between principals and agents. Indeed, our two studies suggest a common list of "good things" to do: 0 Invest in strategic planning by clarifying expectations and establishing performance feedback mechanisms. These processes work better when operators - the front-line troops - are directly involved. 0 Invest in better strategic management of information by standardizing processes and definitions, ensuring data quality, tapping information from the community, and providing IT systems to manage this information. V Choosing among instruments for direct performance incentives to agents: The instruments that create clearer, direct incentives for staff and managersare generally difficult to apply because they entail complicated policy tradeoffs (including a higher political profile). Our two cases suggest that extrinsic personnel incentives are more compatible with the use of external labor markets, while intrinsic incentives may be more compatible with internal labor markets (e.g., the RJU, where entry and exit are limited). 0 The OSSs in SI0 Paulo rely primarily on extrinsic incentives: the freedom and incentives to managefor managers, and the discipline of external labor markets for staff. 0 The SMS in Curitiba places greater emphasis on intrinsic motivations for staff (investing in esprit de corps, using staff appraisals to identlfy and solve problems) and maintains a more hierarchicalmanagement format. 0 Neither system has so far relied heavily on bonus schemes. (OSS hospitals have not yet tried them, and Curitiba has yet to get far with them.) This study describes the instruments that have proved important to improve service delivery in S ~ Paulo and Curitiba. It attempts to explain why these instruments are important, and how they O work in combination with others. It should be remembered, however, that our two cases are drawn from two of Brazil's most sophisticated governments. States or municipalities that lack the particular history and endowments of SI0 Paulo and Curitiba should invest in understanding what strategic planning is, and what alternative incentive systems may be applicable to their setting. Then, they should proceed, where they can, to improve their public sector performance in an experimental but consistent manner. vi 1. IMPROVINGPUBLICSECTORPERFORMANCEFOR HIGHERQUALITY PUBLICSPENDING There is a growing consensus in Brazil that improving the quality of public spending is key to accelerating the country's economic and social development. Brazil has made important advances over the last decade in reducing income inequality and improving certain social indicators. Yet the chronic deficiency of public services in Brazil is widely recognized as an ongoing impediment to more rapid social and economic gains. The volume of government expenditure is not the principal bottleneck to delivering more and better services: at over 40 percent of GDP, Brazil's total public spending is already much higher than comparable middle- income countries. The vital question is how to get greater value for public money. The quality of public spending has various dimensions, including overall allocations of available public resources across competing priorities and how particular expenditures should be funded.' Those questions are dealt with in a separate World Bank study (World Bank 2006b2). In this report, we focus on another important dimension of the quality of public expenditures: How is the provision of public services managed to get the most from available resources? As argued in recent World Bank publications on the topic (World Bank 2004, Fiszbein 2005), improving delivery of public services requires a well-functioning accountability framework that links citizendclients, policy makerdpoliticians, and service providers (e.g., doctors, teachers). This means ensuring that layers of relationships work efficiently between those who demand quality services and those who provide them. Citizens, as the ultimate beneficiaries, need to be able to influence the politicians/policy makers who have been elected to represent their interests. Politicians and policy makers, in turn, must be able to control service providers, either the government bureaucracy or other non-government agents entrusted to provide public services on behalf of the government. As we will elaborate below, this means overcoming inherent difficulties in managing layered accountability relationships (principal-agent relations in theoretical terms) that link citizens, politiciandpolicy makers, and service providers. In this report, we focus on how policy makers can alleviate agency problems - ensuring that service providers (agents) behave in ways consistent with the desires of the policy makers (principals) - through the adoption of certain managerial innovations. From a broad range of public services, we have chosen to focus on healthcare in this report. The health sector in Brazil absorbs approximately four percent of GDP, and its spending level is likely to rise further given the country's demographic and epidemiological profiles. Improving the efficiency of public health spending thus can have effects beyond the sector, improving the overall health of the country's public finance and long-run macroeconomic stability. This study analyzes two performance-oriented reforms in Brazilian healthcare: the introduction of Social Organization inHealth (OSS) hospitals in the State of S5o Paulo and the management of primary healthcare in Curitiba. The remainder of this chapter sets the scene for the two case studies by sketching the main features of Brazil's national policy framework for health, outlining healthcare management challenges, and describing the recent reforms -many inspired by modem models of performance management -that the federal government has promoted to tackle those problems. 'Designs of funding mechanisms can have effects on how the allocated funds are spent, and thus have an impact on efficiency of the spending. For example, program implementersare less likely to seek efficiency improvementsif a given level of funding is guaranteed(e.g., from earmarkedrevenues). Public ExpenditureReview (forthcoming). 1 Chapter 2 analyzes the performance of S2o Paulo's hospitals. I t describes how contracting works under the OSS model and, using data from existing studies and from a survey of 20 hospitals carried out for this report, it compares the performance of OSS hospitals and traditional direct administration hospitals. The chapter then analyzes the different employment and remuneration regimes in different hospital types and seeks to explain how these regimes contribute to differences inperformance among hospital types. Chapter 3 places performance management in Curitiba's primary healthcare system in a broader context. First, it characterizes the city's healthcare philosophy (and its historical origins). Next, it looks at management practices in greater detail, distinguishing instruments for client- orientation, knowledge and information management, specialization policies, human-resource management, and performance management. Finally, it seeks to understand how performance management inhealth was influencedby other innovative management policies in the city. The concluding chapter seeks to reinterpret the two case studies in the light of agency theory, and to draw possible lessons for reforms elsewhere inBrazilian state and municipal service delivery. IMPROVINGSERVICEDELIVERY BRAZIL IN As a federal country, improving service delivery in the Brazilian context largely means strengthening the incentives and institutional capacities of sub-national governments to enhance their perf~rmance.~ It is precisely at the sub-national level, however, that Brazil faces the greatest institutional development challenges. Diagnostic work carried out for the IDB-fundedProgramu Nucionul de Apoio h ModemizagZo da GestZo e do Planejamento (PNAGE) revealed a number of common public sector management problems acrossBrazilian states. These included: Inconsistency between structures and responsibilities, weak regulatory structures, and duplication of institutional roles; Deficient planning and policy coordination, and absence of performance management (i.e., lack of performance indicators, formalistic and centralizedplanning and budgeting); Weak human resource management, including lack of human resource planning; inadequate number and qualifications of staff; absence of career structures, professional development policies and training. In trying to strengthen public sector capacity, one of the principal dilemmas for reformers has been how to balance flexibility (for efficiency) and controls (for probity). The current federal legal framework that governs public administration tends to emphasize protection of public interests through restraining politicians' and bureaucrats' discretion in areas such as human resource management, budgeting and government procurement. These laws have not spared Brazil from patronage or corruption scandals, but they have contributed to the historical development of Brazilian bureaucracy that is relatively more professional than most Latin American and middle-income countrie~.~ In Brazil's federal design, municipalities are mainly responsible for child and primary education, local transport, basic and secondary health, garbage collection and urban planning. States are in charge o f secondary education, higher-complexity health services, and public security. Brazil's public sector began its long path toward professionalization as early as in the 1930s. This professionalization, however, is largely limited to specific areas o f public administration (e.g., economic management and certain legal professions), and is rarer at the sub-national levels. 2 Perhaps more troubling is that Brazil's common legal and organizational designs for public administration have manifestly failed to encourage efficiency or a results orientation in public administration. To ameliorate these problems, a major reform initiative was launched inthe mid- 1990swith the aim of applying some insightsof the New Public Management (NPM), inspiredby public management reforms a decade earlier in countries such as Australia, New Zealand and the United Kingdom. NPM seeks to shift the bureaucracy's attention from near-exclusive emphasis on procedural compliance toward achievement of results. It relies on market incentives, whenever applicable, and managerial practices devised by the private sector. The Brazilian version of NPM, as articulated in the government 1995 white paper Plano Diretor da Reforma do Aparelho do Estado, combines a traditional Weberian administrative model5with tenured public servants serving in a strengthened state "core" of "exclusive" state functions (e.g., tax collection), and new administrative forms for "non-exclusive" state activities (e.g., healthcare). While the reform applied to the federal government only, the proposal has had wide influence within Brazil, and has since served as a model for a number of experiments with reform at the sub-national levels. Inthis model, service delivery in sectors where the goods are contestable inthe market (Le., they could be provided by either the public or private sector) should be contracted out to a non-state entity, such as a Social Organization (see Box 1.1). Brazil's federal and state governments have adopted this model to different degrees in the area of scientific research(Associagcio Brasileira de Tecnologia de Luz Sincrotron; Ministry of Science and Technology), educational broadcasting (Associapio de Comunicapio Educativa Roquette Pinto; State of MaranhHo), health (Instituto de Promopio da Salide e Desenvolvimento Social da Micro Regicio de Irece; State of Bahia) and culture (Orquestra Sinfonica do Estado de SZo Paulo and Pinacoteca do Estado de Scio Paulo; State of Sgo Paulo). A traditional Weberianmodelof public administration refersto one where tenuredpublic servants, protectedfrom political influence, performtheir duties inpublic interests, mostly through compliance with due processes and legality of public actionsrather than with emphasis onresultsof such actions. 3 Box 1.1: Social Organizations Social Organizations are private-law entities (foundations or associations) that receive public funds to deliver services through management contracts. These organizations are not obligatedto follow public sector administrative rules and can be funded with commercial proceeds or donations, in addition to public funds. Their assets belong to the state and are non-transferable. Human resources might be civil servants (estututdrios) from other state organizations or be hired under the private sector labor regime (celetistus). Although Social Organizations are not subjected to public sector rules for procurement, contracts or financial administration, they are subject to external control by the Court of Accounts. Social Organizations can be created in one of three ways: i)from scratch, ii)conversion of an existing state organization, or iii)qualificatiodcertification of a civil society organization that already delivers services, possibly through agreements with government (convenios). To qualify as a Social Organization, the organization has to meet the following criteria: i)be non-profit, ii) set up an administrative council as a decision-making body (with members from the state, civil society, and its own members), and iii)use excessfunds exclusively for its activities. Another salient feature of Brazil's managerial reform in the 1990s are the measures proposed for making personnel management more performance-oriented. The administrative reform approved in 1998 through the lgthConstitutional Amendment included provisions to create a new legal category for public sector employment (emprego pu'blico) whereby public servants would be employed under conditions similar to private sector workers. The reform also envisioned a mechanism to allow public servants to be fired for poor performance, following systematic performance evaluations. These and other measures were intended to increase managerial flexibility and to enhance public servants' incentives for performance. While reform implementation at the federal level has remained incomplete due to political opposition and changing priorities by subsequent governments, these ideas inspired a number of sub-national governments keen to shake up their bureaucracies to improve performance. The two cases of managerial innovation covered in this report follow these reform developments from the mid-1990s through 2005. It is important to bear in mind, however, that managerial reforms in the health sector also have been conditioned by parallel developments in healthcare policy over this period. Healthcare Policy inBrazil The 1988 Constitution mandated a decentralized, universal and free health service financed from social welfare funds. This led to a series of health policy reforms under the Unified Health System (SUS) that have transformed the organization, financing and provision of health services in Brazil. The federal government's role began to shift from service provider to financer, promoter, and regulator. In the late 1980s Brazil converted its federal public health financing system to a single national health fund. In the mid-l990s, it moved away from supply-driven financing of health expenditures (based on fee-per-service) towards needs-driven funding, institutinga per capita payment for primary care services distributed directly to municipalities. This simple reform caused a vast improvement in the equity of the system. Poor municipalities suddenly had funds for primary health services on a scale they had not seen before. More recently, this capitation system was enhanced by various incentive-based programs, notably the 4 Family Health Program (PSF), through which the federal government has transferred additional funds to municipalitiesthat agreeto implement aparticular program. The most salient shift in healthcare policy was a declared emphasis on primary and preventive care rather than secondary and tertiary health care (specialized clinics and hospitals) and curative medicine. The PSFhas been a centerpiece of that primary healthcare model. Under the PSF, first introduced in 1994, municipal healthcare teams proactively provide an integrated package of health services to families in a defined community. These standardized teams work with a uniform set of equipment and procedures. Each team is responsible for providing care to a defined set of 600-1,000 families. By the end of 2005 this model was one of the largest federally- funded health programs, reaching 44 percent of Brazil's population! Only five percent of primary healthcare services are providedby the non-government sector inBrazil. In comparison to primary care, federal policy towards hospitals remains considerably more "passive." Still, federal spending on curative services far outweighs spending on preventative and public health, even though the share is declining. Managing Performance Good performance, in health as in other government sectors, depends on a host of institutional and structural factors, including: 0 human resource managementpolicies that provide an adequate supply of qualified personnel, and incentives for them to perform; a clearpolicy framework; 0 appropriate organizationaldesign, including clear definition of roles and responsibilities; 0 sufficiency and predictability offunding; 0 adequate information on costs and services delivered; 0 straightforward, agile administrativeprocesses (e.g., procurement); and 0 absence ofpolitical inteference inday-to-day administration. In this report, our objective is to describe what mattered most for performance management in specific reform settings. Recent World Bank reports indicate that managing performance in health, particularly at the municipal level, faces a number of obstacles. Notable among these studies were a public expenditure tracking survey (PETS) in a sample of state and municipal See the Ministry of Health website. This study takes the public-ownership framework of primary healthcareunder SUS as a given for states and municipalities, and examines how one state ( S o Paulo) and one municipality (Curitiba) have tackled the issue of performance within that framework. An alternative model of primary healthcare, considered best-practice by many researchers, is one in which private providers qualify to become eligible to provide publicly-funded services, and then compete to register patients. The providers are usually paid a capitation fee, though some services (which might otherwise be under-provided) are provided on a fee-for-service basis. The virtue of this system is that it creates competition and gives patients leveragethrough the market. While this model is found mostly inadvanced countries, it has also been tried in some developing countries, such as Bangladesh, Cambodia, and Guatemala. These arrangementshave not been tried in primary healthcare in Brazil, but they have some elements in common with the Social Organization model in Sfio Paulo, which is the subject of Chapter 2. This alternativemodel of private provisionwould entail adifferentset of costs, benefits,and tradeoffs from Brazil's current model. It may be in the interest of sub-national governments in Brazil to consider this alternative model as an option. 5 health secretariats and facilities (World Bank 2006c), as well as a study of costs and efficiency in hospitals (World Bank 2006a). Human resource management. The health sector relies heavily on its human resources for service delivery. The findings of the Bank's PETS in this regard are alarming. The report identified a range of problems, including: Poor work incentives reflectedinabsenteeism, incomplete work hours, excessive leave, and low salariesfor qualified personnel. Limited autonomy to manage personnel at the facility level, combined with an inability of the center (Le., state or municipal health secretariats) to control staff allocation. Employees working inone institution but accountable (through their employment tie) to another. This hampers manager's authority over employees, provoking conflicts and dissatisfaction. Absence of performance evaluations (among the hospitals surveyed only 27 percent have formal performance evaluation mechanisms). An excessive number of low-qualified personnel and an insufficientnumber of highly- qualified ones, especially inmanagement positions. Policy framework. The PETS found that the capacities of the health secretariats to plan, coordinate, and monitor health service delivery within their jurisdictions vary greatly, but are particularly low at the municipal level. Fragmented planning and budgeting processes comply with formal requirements but are divorced from implementation and do not serve as a basis for prioritization or goal-setting. The World Bank`s costs and efficiency study identified distortions in the system of resource allocation and the payment mechanism under the SUS: most of these transfers are not conditional on performance; they over-fund curative services at the expense of prevention, and over-fund complex procedures at the expense of simpler hospital treatments. Organizational design. The PETS identified a shortage of managerial autonomy and responsibility at the facility level. Many health units lack financial information, and human resource management is overly-centralized. Frequently, centralized systems contributed to low levels of budget execution. Funding. There i s a high degree of rigidity inbudget execution, accordingto the PETS. This fact may limit the scope for misuse of funds; however, it also constrains managerial autonomy needed to improve operational efficiency. Information. The PETS found that managerial information is scarce, of poor quality and little used when it is available (either for analyzing costs or controlling them). Information on expenditure was conflicting. Administrative processes. The health sector in Brazil is subjected to the same set of administrative rules as other sectors in procurement, personnel contracting, and other procedures. 6 These rules are generally designed to limit opportunities for abuse of public resources and positions. The PETS found that an emphasis on prior controls limited managerial flexibility. Political integerence The health sector is no exception when it comes to potential political interference and corruption. The extent to which corruption has penetrated the sector seems to differ by jurisdiction, probably reflecting the overall quality of governance in each locality. An additional factor that sets the health sector apart is the important role of the corporatist body of healthcare professionals - the so-called sanitaristus - who have influenced the design of SUS policy. The sanitaristus could be both sources of positive externalities (e.g., shared ideological commitment to certain approaches to public health, knowledge sharing and exchange, professional networking to facilitate inter-jurisdictional cooperation) as well as negative ones (e.g., mobilized resistance to changes deemed antagonistic to their corporatist interests). The picture of sub-national healthadministration that emerges from these diagnostic results is one wherein the capacities of individual sub-national governments are limited in setting clear policy goals based on robust information concerning the problems and performance of health units (hospitals, postos de salide). These units, in turn, have limitedincentivesto improve performance, given the rigidity imposed by various administrative rules and the budget framework. Moreover, many jurisdictions are poorly equipped to innovate managerially due to a weak complement of human resources. It is no surprise that this set of management conditions leads to poor performance in the health sector. According to the World Bank's costs and efficiency study, there is a large variation in efficiency among jurisdictions. Problems are most acute in those hospitals directly managed by governments (Le., direct administration). Too many hospitals, especially in smaller municipalities, are inefficient because they are below optimal size. The availability of these partially-functioning hospitals at the local level in turn encourages the population to seek care at the hospital level. Many hospitals are overstaffed, underutilized (in terms of beds and surgical facilities), and offer too many of some services and too few of others (World Bank 2005). Inboth hospitals and primary care, procedural rigidities lead to a shortage of materials, broken equipment, and highloss rates for medicines. Inspite of the manifest deficiencies inhealth management, there are some significant advantages for healthcare delivery that are not enjoyed by other public service sectors in Brazil. First, the policy and organizationalframework i s comparatively coherent. SUS may not be flawless, but the fact that the entire sector is structured within a policy framework is a considerable advantage over other sectors (e.g., water and sanitation) whose poor performance is often attributed (at least partially) to the absence of such a policy framework. Second, the informational infrastructure is reasonably solid. A performance orientation relies heavily on the use of information; and notwithstanding the severe problems present in many jurisdictions in collecting and managing data, the health sector is in a more favorable position than most others (e.g., public security) due to the reportingrequirements legally mandated inthe health sector.' Third, decentralization is likely to have made health spending decisions more responsive to local citizen demands. Municipalities now decide how to spend the money and resources that have been distributed more equitably between levels of government, although it is also possible that municipalities are "too responsive" to local demands, which do not take into account systemic 'Thisappearsto be driven, at least partly, by the decentralization o f the health system over the 199Os, and the concomitant information requirements established by the federal government for sub-national governments. 7 rationality in resource allocations (e.g., building hospitals in each municipality when a regional network and a referral system may be more efficient). Unfortunately, constitutional earmarking of health expenditures may encourage inefficiency as it tends to limit the flexibility of front-line managers to allocate available resources, and makes the resources non-contestable (Le., the spenders do not have to compete for allocations by showing value-for-money) (World Bank 2006b). However, earmarkingdoes provide greater predictability of public funding.8 There is at least one factor that makes service delivery more difficult in healthcare than in many other public services in Brazil: the problem of information asymmetry. Due to the highly- specialized nature of healthcare services, the customers (patients) often are not in a position to determine whether the care they are receiving is of good quality. Because doctors exercise discretion in prescribing a remedy, and because their supervisors are usually not able to observe doctors' work directly, managers are limited in their ability to monitor doctors' behavior and performance. This creates opportunities for doctors and nurses to pursue objectives that may not be fully consistent with the corporate objectives of the hospital or the government. This characteristic of information asymmetry severely complicates the task of performance management. As this cursory sketch attempts to show, taking into account all the relevant ingredients of good public sector performance is a highly complex exercise. On average, it appears that Brazil is not yet in a position to take full advantage of the potential benefits of healthcare decentralization because of the precarious institutional capacities that characterize a number of sub-national entities, as well as the legal and institutional design of much of SUS that does not encourage performance. This means that for the majority of sub-national jurisdictions, the priority agenda should be to develop the basic institutions of healthcare delivery, including adequate human resources, basic statistics and management information, and minimum administrative capacities (e.g., planning, logistics management). In jurisdictions that have made progress in these areas (such as the State of Sgo Paulo and the City of Curitiba), managerial innovations that are more demanding of incentives and information offer promising paths for improving the performance of the public healthcare apparatus. ASSESSINGMANAGERIAL INNOVATIONSINBRAZIL'S HEALTHSECTOR Policy discussions on the quality of public spending and the performance of public administration would benefit from an explicit consideration of the benefits and costs of the existing legal and organizational framework that governs Brazil's public sector, including the legal framework for governing public sector employment and personnel management (Regime Juridic0 Unico, RJU). However, reform advocates should not hold all changes hostage to alterations in the basic architecture of Brazil's public sector. Any attempt to modify that architecture would be politically uncertain, and any changes are unlikely to bring about a rapid transformation, not only because institutional reforms take time but also because the legal interpretation would likely protect those who are under the current regime from any change in their employment conditions (Le., acquired rights). Thus, while pursuing a systemic reform agenda to make the legal environment of the public administration more amenable to performance-oriented changes, and For example, the infrastructure sectors in general suffer from both weak policy frameworks and unpredictable funding arrangements. The education sector benefits from a similar degree of funding predictability, but the sheer number of teachers involved in the system, and their relatively limited professional qualifications, seem to complicate performance management in the sector. Certain economic areas (e.g., revenue administration) and social security administration may be other sectors where policy and funding arrangementsare relatively favorable. 8 while ensuring due process procedures are in place to minimize abuse and patronage, governments should pursue reform opportunities that are present notwithstanding the existing institutional constraints. Far-reaching administrative reforms proposed since the mid 1990s in Brazil have inspired a number of innovations, but their effects and effectiveness have rarely been evaluated systematically. In this report we present empirical assessments of two recent managerial reforms carried out within Brazil's current institutional setting: "corporatization" of public hospitals inthe State of Siio Paulo and the introduction of performance-based pay inthe City of Curitiba. On the surface, these innovations seem to follow the now familiar New Public Management (NPM) prescriptions of "let the managers manage" (managerial flexibility) and "reward performance" (performance-based pay). But there is more to management improvement than simply adopting a new organizational model or a new managerial instrument. The two cases uncover interesting nuances that enrich our understanding of practical, context-specific approaches to performance management within Brazil's existing institutional constraints. They offer insights into "what really matters" in improving management (in this case, of the public health sector), focusing particular attention on how policy makers can overcome so-called agency problems. The case narratives illustrate how both the State of Siio Paulo and the City of Curitiba deployed a range of organizational and managerial innovations to diminish the policy makers' informational disadvantage vis-8-vis front-line service providers (doctors, nurses) and achieve greater alignment between the government's overall policy objectives and the objectives and the incentives of the service delivery unitsand healthcare staff. CASESELECTION Siio Paulo and Curitiba are not representative of Brazil as a whole. Both are relatively more developed economically and socially than most of the rest of the country. Their human resource bases are larger and more diverse, from which governments can draw relatively better-trained professionals, a critical variable in skill-intensive sectors such as health. A supply of skilled labor is not enough, however, to deliver quality healthcare services. Without good governance, large and diverse human resource bases are insufficient for building a capable public administration. The governments of Siio Paulo and Curitiba are recognized, as well, for their relatively high institutional capacities and comparatively well-run public administrations (particularly Curitiba). Inother words, SiioPaulo and Curitiba serve as examples of what Brazilian public administration is capable of today under certain circumstances. As Siio Paulo and Curitiba are relatively well-run administrations, the main aim of our study is to identify those factors that contribute to their relatively strong performance. In the case of Silo Paulo, a parallel World Bank study on hospital performance has quantified the superior performance of that state's Social Organization hospitals (World Bank 2006a). After carefully controlling for intervening variables (e.g., size, types of services provided) among hospitals of different organizational types, this study found the Social Organization hospitals to be more efficient and provide better quality care than their direct administrationcounterparts. We are on more tenuous ground in asserting superior performance of Curitiba's health system. Curitiba certainly exhibits reasonable health outcome indicators, but available information does not allow us to attribute these positive outcomes to the performance of the city administration. Nonetheless, Curitiba has long been known for its innovative public management, especially in urban planning, and thus we believe our prima facie assertion - that Curitiba has a relatively effective public management - is reasonable. The idea of attempting to document the 9 performance superiority of Curitiba with rigorous empirical analysis was considered at the outset of the research, but discarded given the scope and the final objective of this study as well as costs involved in such a systematic evaluation. The majority of sub-national governments are not as well-endowed as these two jurisdictions in terms of basic enabling conditions for building a well-performing public sector. As such, lessons to be drawn from our study may not be directly applicable to "typical" sub-national jurisdictions in Brazil. Nonetheless, some of the insights gained from the study appear to offer useful referencesfor other sub-national governments. Therefore, the report offers its policy implications and recommendations with due caution. 10 2. SOCIAL ORGANIZATIONHOSPITALSINTHE STATE OF S A 0 PAUL0 INTRODUCTION Beginning in the late 1990s the government of Siio Paulo adopted a new management model to administer a group of state public hospitals. Social Organizations in Health (OrganizaGdes Sociais em Saiide - OSS) were created by statute to enable a formal partnership between the state and non-profit private sector organizations. Under this OSS model, the government provides budgetary transfers to cover the costs of running the hospital, but responsibility for day-to-day administration is delegated to certified non-profit organizations. The State Secretariat of Health (SES) negotiates and signs a performance contract with each of these hospital managers, committing budgetary resources from the public treasury in exchange for specific performance outputs. The managers, in turn, are granted far greater flexibility than their counterparts in traditional state hospitals to run the hospital inthe manner they consider best-suited to meet their performance targets. Silo Paulo is not alone in attempting organizational reform of this kind. OSS hospitals are illustrative of a national and international trend toward corporatization (or "autonomization") of government services that expand the so-called "non-state public sector." There are interesting theoretical arguments in the literature (discussed in Chapter 4) for why corporatization may improve financial and facilities management, and thereby improve hospital performance. Still, human resources are the mainstay of healthcare provision, and clearly the largest expense. Therefore, it is perhaps surprising that the literature on contract-style reforms in developing countries devotes little attention to labor relations within corporatized bodies. One reason for this shortcoming, as noted by Harding and Preker (2003:53), is that most governments reforming their healthcare systems have been "unwilling or unable to transfer control over labor, recruitment, salaries, staff mix, and the like and have instead left employees in the civil service, employed directly by the health ministry." Thus, even as managers have been granted greater financial autonomy and been made accountable for results, the existing public sector employment rules typically have been left in place. In Si0 Paulo, however, human resource management rules do indeed differ between OSS and traditional state hospitals. Through the OSS reform Silo Paulo has attempted to transform the incentives facing healthcare personnel at the point of service delivery, and thereby improve the efficiency and quality of care. Those personnel management differences, and how they affect - if at all - the relationship between hospital managers and their staff, are essential to understand corporatization theory in real-world hospital settings. Thus, the S2o Paulo experience is a compelling case for analysis. The central question of this study is: How do the incentives facing managers and staff working in S b Paulo's OSS hospitals actually differ from traditional public hospitals, and what is the impact of those differences? CONTRACTINGFORHEALTHCARE INSA0 PAULO'S OSS MODEL In2004, seventeen public hospitals inSI0 Paulo were administered on a contract basis as Social Organizations in Health, rather than as traditional units within the hierarchical structure of the Secretariat.' This administrative reform initially grew out of the federal government's 1998 ' Of the 17 hospitals, 14 operated under Law 846/98 as Social Organizations, while three followed the same rules through a contract betweenthe Secretariat of Health and university Medical Faculties. 11 constitutional reform, which established a legal framework for autonomous "Social Organizations." Constitutional Amendment No. 19 enabled private sector, non-profit organizations to utilize public resources (material and financial) to provide public services that are "not-exclusive to the state."" These Social Organizations would enjoy management, budget, and financial autonomy while remaining accountable to government under a performance contract. Those contracts were to specify the period of the contract, the resources provided by government, the expected outputs, the criteria for evaluation of performance, and the rights and obligations of the managers. InSBo Paulothe state government subsequently sanctioned its own Social Organization law (No. 846/98) for the state's health sector, based upon the federal law 9.637/98." The state law specified that, while governed by private sector law, only a non-profit organization could qualify as an OSS,and an OSS-administered hospital could only provide services under the SUS. These hospitals are not private. The patrimony remains publicly-owned. The first step in implementing the OSS model was for non-profit organizations to apply to the state for certificationas Social Organizations in Health. Ifan organization satisfies the criteria for certification, then it is legally authorized by the Governor to enter into a contract with the state Secretariat of Health (SES) to manage one or more public hospitals.'* (An important qualification criterion is that the non-profit must show a minimum of five years of experience administering health programs or services. SI0 Paulo certainly enjoys an advantage over other states inhaving anumber of reputable non-governmental organizations in the health field.) The next step is for the SES to negotiate a hospital management contract with the OSS. The management contracts between the SES and OSS specify the volume of different services to be performed each month (e.g., inpatient and outpatient services, medical consultations) inexchange for a specified budget (a prospective payment block contract). As the hospital is required to meet monthly performance goals, 90 percent of the annual budget agreedbetween the SES and hospital administrator is delivered in monthly installments. Performance within +15 percent of stipulated targets is permissible without affecting disbursements. However, if the quantity of services delivered falls to 75-84.9 percent of the agreed targets, then the financial disbursement is reduced by 10percent. If output falls below 75 percent, then payment is reduced by 30 percent. lo The underlying rationale for Constitutional Amendment No. 19 (June 5, 1998) was set out in the 1995 "White Paper on the Reform of the State Apparatus" produced by the Ministry of Federal Administration and State Reform (MARE) to specify the objectives and guidelines for redefining Brazil's public administration. The White Paper asserted that "rigid hierarchical standards ...controllingprocesses instead of results [produced an administrative system] shown to be stultified and inefficient and, therefore, incapable of coping with the magnitude and complexity of the challenges established by the process of economic globalization" (Brazil 1995:9-10). The proposedsolution is for the "State [to abandon] its role as executor and direct renderer of services, while preserving its task of regulator and provider or fosterer of such services" (Brazil 1995:17). Social Organizations are defined by Art. 1 of the corresponding Law 9.637 (1998) as "pessoasjuridicus de direitoprivado, semfins lucrativos, cujas atividades SZOdirigidas ao ensino,d pesquisa cientiJica, ao desenvolvimentotecnoldgico,d prote@o e preservapio do meio ambiente, h"cultura e h saride." There are a few differences between the federal rules and Ssio Paulo's Social Organization in Health (OSS) model. One difference is that under federal rules regular government employees transferring to a Social Organization (OS)would receive their former pay and could receive an additional payment from the resources of the OS. InSsio Paulo, however, a second paymentis not permitted. ComplementaryLaw no. 846 (June 4, 1998) specifiesthe conditions for an organization to qualify as an OSS. An OSS may manage more thanonepublic hospital, each with a separateperformancecontract. 12 The remaining ten percent of the budget is delivered quarterly, contingent upon the hospital submittingproperly codeddata on their patients and the treatments or services provided. This 10- percent provision reflects some "learning by doing." The OSS model, as initially drawn up in 1998, allocated 100 percent of funds through 12 monthly installments. However, since 2001, 10 percent of the agreed budget is delivered quarterly, contingent upon reporting output and quality indicators to the SES and to the Evaluation Commission. Notably, the government of Slo Paulo has demonstrated its willingness to enforce this provision. For example, during its first semester of operation, the OSS Hospital Mario Covas did not produce a proper accounting, and as a result hadten percent of its budget withheld until adequate data were delivered to the Secretariat. The OSS contracts are fine-tuned through regular dialogue between hospital directors and the OSS supervising staff of the Secretariat of Health. Adjustments are made from one annual contract to the next, but can also be made by consensus within the operational period of a given contract. The management contracts are supervised by the SES, and subsequently audited by the state's SUS Council of Health and Court of Accounts (Tribunal de Conrus). Thus, setting up an OSS reduces the day-to-day managerial authority of government over service delivery, but adds contractual accountability mechanisms based upon performance outputs. OSS governance arrangements are a hybrid of market and hierarchical control. In large part, the OSS model is designed to allow hospital managers to exercise control over the most important factor of production, namely, labor. While there may be multiple factors to explain why the government of Slo Paulo adopted this management model (see Box 2.1), at the forefront is the desire for improved staff performance. 13 Box 2.1: Why OSShospitalsappearedinSi50 Paulo The decision by the government of SBo Paulo to experiment with OSS hospitals was not motivated solely by a conviction that applying recent theories of effective human resource management would improve service delivery: there was a compelling practical motivation, as well. During the administration of Governor Mario Covas (1995-2001), the state completed construction of a number of new hospitals that were initiated during the 1980s under Governor Andr6 Franco Montoro. That presented the government with a dilemma. The federal Camata Law (Complementary Law no. 82/1995) stipulated that a state government's wage bill could not exceed 70 percent of current revenues. As Si0 Paulo already was struggling to come into compliance with the Camata limits, designating these new hospitals as OSSs provided a budget transfer to the hospital - but their salaries would not count toward the statutory limits, as convenient solution: the new OSS employees would be paid from the public purse - via the they are not considered state employees. The numbers clearly are significant. SBo Paulo's 17 contract hospitals employ over 20,000 employees, more than 20 percent of total employment in the state health sector.' The federal Fiscal Responsibility Law (Complementary Law no. 101, May 4, 2000) largely Responsibility Law (LRF)personnel expenditures - including active and retired employees and adopted the public employee expenditure rules of the Camata Law. Under the Fiscal all forms of monetary compensation, as well as employer pension contributions - cannot exceed 60 percent of Current Liquid Revenue (defined as current income minus constitutional and legal transfers and employer contributions toward the pension system). However, personnel providing services "contracted out" to private sector organizations are considered by the LRF to be "Other PersonnelExpenditures" and do not count toward the 60-percent cap. Any state that reaches 95 percent of the cap is barred from increasing salaries or employment. If the cap is exceeded, voluntary transfers and credit from the federal government is suspended. This created an incentivefor employment via cooperatives andSocial Organizations. Note: Overall, the state of Siio Paul0 employs roughly 650,000 people indirect administration and another 150,000 infoundations, autarchiesand other autonomousbodies (which includescontracthospitals). PERFORMANCEDIFFERENCES BETWEENOSS AND DIRECTADMINISTRATION HOSPITALS While the SBo Paulo OSS model is still in its relative infancy, early data on the efficiency and quality indicate that the OSS model compares favorably to the traditional hierarchical administrative model. A study by the World Bank (2006a, forthcoming) on hospital performance in Brazil examined performance data from 2003 for 12 OSS hospitals and a sample of ten direct administration hospitals in SBo Paulo of comparable size and complexity. First, the study dispelled the notion that OSS hospitals have benefited from a higher level of financial resources than their traditional public sector counterparts. The 2003 data showed no statistically significant difference in the amount of resourcesat the disposal of OSS and traditional public hospitals. The study then examined output, efficiency, and quality data for each cohort, and found that OSS performance was either the same or superior in all categories. The key finding is that OSS hospitals have produced more numerous and/or superior services with the same overall quantity of resources. 14 The OSS hospitals examined by the World Bank study offered 35 percent more patient admissions for each hospital bed. Inrelation to surgical beds, patient admissions were 61 percent higher. Table 2.1 presents the comparative data from 2003 on hospital bed utilization. Similarly, in the present study the reportedoccupancy ratesfor the hospitals inour research sample suggest that OSS and privatehospitals outperform traditional public hospitals (Table 2.2). Table 2.1: Patient Indicatorsof Allocative Efficiency inOSS & Direct Administration Hospitals,2003 Avg. among OSS Avg. among direct Indicators of allocative efficiency hospitals administration hospitals (N=12) (N=10) Period between patients when hospital bed is vacant (days) 1.2 3.9 Occupancy Rate (%) 80.5 63.2 Avg. Patient Stay in Surgical Clinic (days) 4.2 5.4 Avg. Patient Stay Overall (days) 4.8 5.9 Occupancy Rate Direct Public with oss Private administration foundation (N=2) (N=7) (N=4) (N=7) 1 1 71- 80% 60-70% 1 4 1 I 81-90% 2 2 5 2 Quality indicators also show that the OSS hospitals perform as well as or better than their direct administration counterparts. For instance, overall mortality was lower in 2003 in OSS hospitals than in direct administration hospitals. Meanwhile, mortality rates were practically the same in the medical, surgical, and pediatrics sections of OSS andtraditional public hospitals (Table 2.3). MortalityRates Avg. among direct Avg. among OSS administration hospitals hospitals (N=10) (%) (N=12) (%) Overall 5.3 3.8 Surgicalclinic 3.6 2.61 Medical clinic 11.96 11.64 Pediatric clinic 2.63 2.80 The two hospital groups (OSS andtraditional) inthe World Bank (2006a) study serve populations performance indicators it should be noted that many OSS hospitals are "closed door" - meaning with similar health problems, as indicated by the patient data. However, in considering these 15 their demand is controlled though a referral process - rather than "open door" hospitals, which attend to spontaneous emergency and urgent care 24 hours a day. HUMANRESOURCEMANAGEMENTPAULO'S HOSPITALS INSA0 As a prelude to comparing human resource management in traditional, OSS, and private hospitals, it is first necessary to summarize the disparate forms of employment in Brazil's healthcare sector. LegalBasisof Employment In 1990 the Single Juridical Regime for public employees (Law no. 8.112) was approved, in compliance with a provision of the 1988 Constitution. Employees hired for a legally-established career post (cargo) anywhere in the public sector must be selected through a competitive merit process via written examination (concurso pu'blico). These statutory employees (estututurios) have legal protections against dismissal after two years of service. Each traditional direct administration hospital in Si0 Paulo has its own establishment list (padrclo de lotug&) of authorized posts. (The most recent lists date from 1994.) If there i s a vacancy, the hospital manager may hire a new employee for that post from the top of the list of concusados(ie., those who have passedthe merit exam) without further authorization from the Secretariat of Health. If there is no vacant post and yet an employee is greatly needed, the hospital may hire a concursado on a temporary basis (pro l~borio).'~ In 1974the government of Si0 Paulo approved specific legislation for employment inthe state's hospitals. Law 500174 was promoted as a means to enable hospitals to be more agile in meeting their specific and immediate personnel needs. Employees hired under Law 500 are selected via an assessment of their resume, not by the more time-consuming concurso, which often is not conducted for narrow medical ~pecia1ties.l~These "Law 500" employees receive identical salaries and benefits as estututurios, except they do not receive a reward for their academic title (licenp prernio) or length of service (sexenio)." Inprinciple, these personnel are hired only for a two-year term, with no promise of job stability. In ractice, however, their job stability and protections also became the equivalent of estututurios.18 Many healthcare professionals, whether in the private or public sector, are hired under private sector labor legislation (Consolidu@ dus Leis do Trubulho - CLT). In the public sector, selection under the CLT is via a "selective process" to occupy afincclo utividade. This selective process amounts to checking an employee's qualifications to ensure that they are suitable for the position, but does not require a competitive procedure for selection. There is no statutory l3 The pudrtio de lotag& is issued by decree by the governor. As medical technologies and procedures change there are regular conflicts as hospitals requestan expandedpadriio - typically without offering the elimination of any existing positions. If a facility has a vacancy it is entitled to hire someone to fill that post without receiving any additional permission from the governor. NB: The posts of former federal employees who work in hospital facilities that were transferred to the state do appear in the pudrZo de loqrlo. However, they are not curaos of the state. This means that when these employees retire, the hospital can not fillthe post. l4 Law 500178 specifies only a "preference" for those who have passeda concurso. Is The licenp pr2mio is an award of 90 days of leave for every 5 years of employment. The sexenio is a monetary award received after 20 years of service that is equivalent to one-sixth of the value of the four quinquenios that an employee of 20 years will have earned. l6 See Complementary Law no. 180/78. 16 impediment to firing a CLT employee "without cause." (The employee i s simply entitled to a severance payment.) It is rare, however, for CLT employees in the public sector to be fired. In addition, Brazil's labor courts often confer the same kind of protection to CLT employees (celetistus) as they do to estututurios,although the legal basis for doing so is questionable. In some cases state hospitals have met their personnel needs via cooperatives, wherein the employment relationship is not between the state (or hospital) and an individual, but rather between the state (or hospital) and a union of professional employees. The union (cooperutivu) assumes the contractual obligation to ensure that a certain number of professional hours are provided to the hospital, as specified under a contract with the facility. These personnel must completely interchangeable - contractually and in practice. The pay per hour is higher when satisfy certain qualifications, but the actual people who show up to provide the service are working through a cooperative than it i s for regular CLT or estututurio employees. However, cooperative workers receive no benefits,` and are paid only for hours worked. Thus, their pay can vary considerably from month to month. In 1999 there were approximately 23,800 doctors working as estututurios in the state of Si30 Paulo. That number represented approximately 46 percent of all doctors with a formal employment link (public and private sector combined). In addition, there were another 27,800 (54 percent) working as celetistus (Cremesp 2002:13). As shown in Table 2.4, the relative importance of different employment types varies significantly among doctors, depending upon their medical specialty. Overall, the public sector accounted for slightly more than 60 percent of all medicaljobs in S5o Paulo in2000 (Cremesp 2002:7). While we do not have detailed comparative data for all OSS hospitals, there were no estututurio employees at any of the seven OSS hospitals in our sample for this study.17 In fact, OSS employees are almost exclusively celetistus (CLT), which is the form of employment most commonly used in private sector hospitals (Table 2.5). Cooperutivupersonnel can also be found at some OSS hospitals; however, the importance of cooperatives as a form of hospital employment has diminished inrecent years, inpart as a result of a 2003 change infederal tax law that added to the costs of independent hiring, thereby eliminating the tax advantagesthat hadbeen available by hiring labor through cooperatives.18 The experience of the OSS Hospital Mario Covas is illustrative. When the hospital opened in November 2001 the director decided to hire almost all medical staff via cooperatives. As the director explained in our interview, the factors that led to this decision were short-lived. The hospital was just getting underway, and there were many unknowns. Satisfying the hospitals labor needs through cooperatives offered greater flexibility and reduced risks since cooperatives represent only a short-term labor commitment on the part of the hospital. Still, as hospital managers at Mario Covas developed firmer expectations employees - thereby establishing a direct employment relationship between the hospital and about the future, most cooperative employees were gradually converted to or replaced by CLT employee." " While it is legally permitted for a public employee to transfer to an OS, there is no incentive for employeesor OS directorsto seek sucha transfer. '*The trend away from cooperativaemployment is visible nationally,not only inSlo Paulo. Forexample, the Cearh Seccretariat of Health found that cooperatives actually turned out to be more expensive for the government, and presented greater management difficulties, than hiring someone as a staff employee. Cooperativeemployment has beentargetedfor gradual eliminationthere. There are still cooperative employees inthe specialties of anesthesiology, heart surgery, and radiology. The cooperative contract with radiologistsalso makes themresponsible for film and other inputs, as well as equipmentmaintenance. 17 Table 2.4: Formof Employmentby MedicalSpecialty inSI0 Paulo's PublicHospitals, that offer this Salaried Fee for service service (Estatutario, (no employmenttie Cooperativase I I *.- N=AR... Lei500, CLT) to hospital) empresas Others Medical specialties No. % No. % No. % No. % No. % Cardiologia I 27 56.3 24 88.9 1 3.1 2 7.4 0 0.0 Cirurgia 34 70.8 29 85.3 2 5.9 3 8.8 0 0.0 Clinica Medica 45 93.8 37 82.2 2 4.4 3 6.7 3 6.7 Gastroenterologia 19 39.6 16 84.2 1 5.3 0 0.0 2 10.5 Ginecologia 37 77.1 31 83.8 2 5.4 4 10.8 0 0.0 Hematologia 17 35.4 16 94.1 0 0.0 1 5.9 0 0.0 Nefrologia 13 27.1 13 100.0 0 0.c 0 0.0 0 0.0 Neurologia 22 45.8 20 90.9 0 0.c 1 4.5 1 4.5 Obsetericia 29 60.4 24 82.8 I 3.4 4 13.8 0 0.0 Oncologia 12 25.C 12 100.0 0 0.c 0 0.0 0 0.0 ~Ortopedia 34 70.8 29 85.3 1 2.9 3 8.8 1 2.9 Otorrinolaringologia 24 50.C 21 87.5 1 4.2 1 4.2 1 4.2 Pediatria 41 85.4 34 82.9 3 7.3 3 7.3 1 2.4 Psiquiatria 23 47.9 20 87.0 2 8.1 1 4.3 0 0.0 Urologia 21 43.8 19 90.5 1 4.8 1 4.8 0 0.0 Neurocirurgia 17 35.4 15 88.2 1 5.9 1 5.9 0 0.0 Anestesiologia 33 68.8 26 78.8 3 9.I 4 12.1 0 0.0 Medicode CTI 21 43.8 20 95.2 0 0.0 1 4.8 0 0.0 Plantonista I 42 87.5 36 85.71 2 4.8 4 9.5 0 0.0 Source: Cremesp2002 (data from a sample o' 48 public hospitals) Direct Public oss Private administration with foundation Statutory (Estatutfirio) 7 2 Emergency (Law 500174) 7 CLT 2 7 4 Cooperative 1 For Services Provided 2 4 (PrestaCZode Servigos) MultipleJobs Healthcare professionals in Brazil typically do not have just one job at one facility. Most often they hold two or more jobs simultaneously. A 2001 survey by Cremesp/Nescon found that fewer than 20 percent of doctors have a single employment tie (Cremesp 2002:9). Nurses, too, commonly hold more than onejob at one facility. WhiIe there are no systematic data on multiple job-holding by nurses, the estimates provided by the nurses we interviewed were that at least 60 percent have two employment ties. 18 InDecember 1997, Complementary Law no. 840 reduced the work week for a doctor's post inthe state public sector from 40 hours to 20 hours - which i s the same as federal doctors." Earlier that year S3o Paulo's doctors mobilized to demand a pay raise, but the state was in a dire fiscal situation. The reduction in hours (without reducing pay) was a means to respond to labor pressure without a direct additional cost to the treasury. Doctors were then able to combine two jobs (and two salaries) at different public hospitals for a total of 40 hours per week. For doctors and dental surgeons who already have an employment link with the state, it is possible to work extra shifts in addition to their regular hours, either at the hospital where they normally work, or at another facility. These 12-hour shifts are known as plunto'es.'l Most doctors work more than 40-hours per week through a combination of posts at public and private hospitals, plunto'es, private practice, and hours assisting on surgical teams on a fee-for-service basis. The 40-hour limit is for public posts only. The standard employment contract for nurses is 30 hours per week. A maximum of 40 hours are allowed at any single facility, but some nurses combine 36-hour positions at two different hospitals.'* Among nurse assistants, holding two jobs is believed to be even more common than it is for nurses. Nurse assistants can have 30, 36, or 40 hour work schedules. Many combine 30 hour schedules at two facilities. COMPARINGHUMAN RESOURCE MANAGEMENTANDTRADITIONALPUBLIC INOSS HOSPITALS A common perception of senior staff inthe Secretariat of Health-which was supportedby many of the healthcare professionals we interviewed - is that efficiency and productivity are generally superior in the OSS hospitals than in their traditional hospital counterparts. That perception is substantiated by a separate World Bank analysis of 2003 hospital performance data (2006a, forthcoming). Likewise, the data on occupancy rates and employees per hospital bed for the hospitals in our present research sample support the same conclusion: OSS performance is generally superior (Table 2.2 and Table 2.7). Most of the doctors we interviewed at OSS hospitals also maintained a job at a direct administration facility. Thus, it seems evident that traditional public hospitals are able to attract many doctors of the same caliber as those working for OSSs. Are there important differences in human resource management for healthcare professionals in OSS and traditional state hospitals that might explain the performance differences? EstablishmentControlandStaff Composition While not directly related to performance in terms of efficiency and quality of care provided, financial discipline often goes with careful use of resources, and thus cost-effectiveness and efficiency of services delivered. This indirect incentive effect of financial control depends on 2o X-ray technicians and laboratory workers also have a 20 hour work week (LC 848). 21 There is a legal limit of 12plunto'es per month. The idea of creating plunto'es for nurses, as well, has beenunderdiscussionbut is notpermitted at the presenttime. 22 This is the case for "Flavia," a neonatalintensivecarenurse at Hospital Mario Covas who participated in our focus group discussion. She works 36 hours per week for the OS and another 36 hours per week at a private hospital. A survey by the Conselho du Enfermugem found that average monthly (combined) income for nurses inSZo Paulo was R$3,500 to R$4,000. 19 who bears the responsibility of controlling cost drivers (of which human resources are the most important). Brazil's FiscalResponsibility Law sets an upper limit for government personnel expenditures (60 percent of the net current revenues for state and municipal governments). Consistent with that law, each OSS contract includes a provision mandating that the hospital wage bill may not exceed 70 percent of the hospital's overall budget. At the time of the first OSS contracts, personnel expenditures in traditional public hospitals ranged from roughly 60 percent to 68 percent, so the 70 percent cap is not meant to be onerous.23 So long as OSS directors do not exceed that 70 percent limit they are free to decide how many staff, with what skills, are appropriate to fulfill the hospital's mission. In traditional public hospitals the number of employees, by type, is given by the hospital's establishment list. The director has no responsibility or control over this allotment of personnel, as the authorization to hire additional staff at a traditional public hospital (and foundation hospital) i s made by the state Secretariat of Health. In contrast, at each OSS, like the private hospitals in our sample, the hospital's General Directorate will make the decision whether or not to hire additional personnel. Likewise, directors of OSS, private, and foundation hospitals all indicated that they have authority to determine where to make budget cuts in the event of a shortfall. As Table 2.6 indicates, OSS hospital directors in the exercise of their establishment authority have hired professional staff in markedly different proportions to what is found in the establishment lists @adrZo de lorqZo) of traditional direct administration hospitals. The data indicate that OSS hospitals rely to a greater extent on fully-qualified nurses, and less on doctors. That is precisely the kind of staffing mix that many healthcare analysts advocate to provide quality care at lower Inaddition, the number of employees per hospital bed at traditional direct administration hospitals was considerably higher, on average, than at OSS and private hospitals inour study sample (Table 2.7). Personnel hours contracted Avg. among direct Avg. among OSS Difference (40 hour equivalents) administration hospitals (OSS/direct hospitals (N=10) (N=12) administration) Doctors 203.15 143.80 (7 1%) Nurses 40.50 54.09 33% Nurse Assistants 256.81 234.12 (92%) 23A World Bank (2006, forthcoming) analysisfound that averagepersonnelcostsrepresented68% of total expendituresamong the 12OSS hospitalsexaminedinthe study. 24See, for example, Edwards,WyattandMcKee(2004). Iam grateful to AprilHardingfor suggestingthis source. 20 Employees per Traditional Public Public with oss Private Bed Foundation 2 1 4 -4.5 2 - 3.9 4.6 -5 5 2 2 5.1 -7 3 2 1 7+ 2 PersonnelSelection Personnel in traditional public hospitals are selected via written examination (concursopu'blico). Incontrast, the directors of OSS and privatehospitals included inour sample stated that personnel selection at their hospitals is based upon an analysis of curriculums, interviews, and (sometimes) practical exams. These directors unanimously affirmed that they have autonomy to hire a professional whom they consider qualified, without further interference. Incontrast, the directors at traditional public hospitals (and foundation hospitals) confirmed that they lack that authority. As the procedures for selection and appointment differ between OSS and traditional public hospitals, so does the time required from the moment a decision is made to hire someone for a position to the day that new employee arrives for work. Whereas the elapsed time for all the OSS, private, and foundation hospitals in our study was less than a month, at the traditional public hospitals that elapsed time rangedfrom one month to more than six months (Table 2.8). ElapsedTime Direct Public oss Private administration with Foundation 1month 3 2 7 4 2 months 1 6 months 1 >6 months 2 Salary Determination In traditional direct administration hospitals salaries are paid and accounted for centrally, and salary scales are also determined centrally for all categories. The director's budgetary authority, as confirmedby our study visits, is limitedto medicines and other recurrent costs. Meanwhile, the vast majority of OSS employees are hired under the private sector CLT law. In fact, while there are a few OSS managers who hire some labor through cooperatives, at most OSS hospitals all employees are celetistas. (The decision is up to the OSS director.) These hospital 21 directors can sethegotiate the pay levels for their CLT and cooperutivu staff. The realities of the labor market, not statutes, are the bindingconstraint. Focus group interviews with doctors and nurses at the hospitals in our sample revealed that, in certain instances, pay for a given post was somewhat higher at an OSS. Generally, however, the healthcare professionals we interviewed reported that their pay on a per hour basis was closely comparable at direct administration and OSS hospitals.25 Unfortunately, comparing hourly wages between traditional public and OSS hospitals is complicated by an oft-mentioned practice at direct administration hospitals whereby individual doctors and nurses reach an informal agreement with hospital administrators to receive their full pay while working less than the full number of prescribed hours. The purported rationale for this "understanding" between health professionals and their managers is the modest pay scale that applies in the traditional state system. Clearly, the practice is not uncommon, but is impossibleto quantify.26 A list of the non-salary benefits offered to employees at direct administration, OSS, and private hospitals was gathered through the study questionnaire. Benefits are somewhat better at OSS than direct administration facilities, and better still at private hospitals (Table 2.9). Private hospitals are also the arena where doctors can earn much higher pay than at OSS or direct administration hospitals. However, focus group participants consistently explained this was only true for doctors with well-established reputations working at prestigious private hospitals. Table2.9: ComparingEmployment Benefits(amonghospitalsinstudy sample) Source: Resultsof researchquestionnaireapplied ina sampleof 20 Sgo Paulo hospitals.The questionnaire appears inAppendix B. 25 "Marcos," a surgeon who joined one of our focus groups, works 24 hours per week as a celetistu employee with an OSS, another 24 hours (in2 plunt6es) as a statutory employee with a municipal hospital in SI0 Paulo, and an additional 20 hours per week in a private clinic. Interestingly, it is the municipal hospital that pays far better than the other two. NB:There is a 40 hour per week limit for a surgeon in public hospitals. Although this should apply to Marcos's situation, there is no authority that keeps track of hours worked inpublic hospitalsoperatedby different levels of government. 22 Performance Pay It is perhaps interesting to note that traditional public hospitals -andnotthe OSS hospitals -have a version of performance pay based irpon a quarterly evaluation. Unfortunately, the poor design of the Special Incentive Reward (PI) undermines its potential use as a performance incentive mechanism (Box 2.2). In all but the most extreme cases the evaluation i s apro forma exercise. Indeed, a doctor at one of the public hospitals mentioned that some of his superiors who have filled out his evaluation did not even know him, let alone his work. Likewise, nurses that work with him are evaluated by the head nurse (which seems appropriate), but without consulting him for his opinions (which seems regrettable). None of the OSSs visited for this study had a pay for performance plan, though the directors have the authority to establish such a plan if they choose. effective pay for performance system - whether in Brazil or anywhere else. In fact, pay is not It is important to acknowledge that for most jobs it is notoriously difficult to implement an closely related to performance in many organizations that claim to have merit increase systems (Baker et. al. 1988). A strong pay-for-performance scheme "motivates people to do exactly what they are told to do" (Baker et. al. 1988). However, it is often difficult to specify precisely what someone should do, which can lead to perverse outcomes. Moreover, merit-pay systems may encourage employees to expend unproductive effort in"gaming" the system to measure and evaluate output. The cost of dealing with the problems engendered by merit pay systems may simply outweigh the benefits they offer. Career Development A common theme of our interviews with surgeons is that the large public hospitals treat patients with more varied and uncommon medical pathologies than those at OSS and private sector hospitals. This feature can make the public hospitals a more intellectually and professionally stimulating place to work for many doctors. "Ana" i s one of the doctors interviewed for this study. She works 24 hours per week as a general surgeon at an OSS, another 30 hours per week, on average, at a traditional public hospital, and 12 more hours per week at a private hospital. On occasion, Ana also joins a surgery team for operations performed at other hospitals. She earns considerably less per hour for her work at the traditional public hospital, largely because one third of her time there is donated for free. Why does she work there if her hours could be fully remunerated by working elsewhere? Ana told us that the public hospital is where she did her residency, so she feels a certain emotional commitment to the hospital. But more importantly, her work there allows her to continue to develop her skills inher specialty of plastic surgery. Incontrast, her work at the OSS and private hospital is in general surgery, and does not allow her to utilize and develop those specific skills. There are instances, certainly, when OSS directors can offer employees attractive professional development opportunities. "Marta" i s an infectious disease specialist, hired as Chief of Section for infectious disease control at the OSS where she works. She also works inthe same specialty at Hospital das Clinicas (a large public teaching hospital), but with far less responsibility. Consequently, Ana earns more than three times the amount per hour for her work as Chief of Section at the OSS (R$43.90/hour) than at Hospital das Clinicas (R$13.40/hour). Still, Ana explained she does not wish to give up her employment tie with Hospital das Clinicas because 23 patients with rare diseases are treated there, and not at the OSS hospital. That exposure boosts her professional qualifications and experience. Box 2.2: The special incentive award In 1998, a special incentive award (Premia de Zncentivo Especial - PI) was created for doctors, sanitary workers (medicosaniturista) and dental surgeons employed at traditional public hospitals inS3o Paulo. (See Resolu@o Conjunta SS/SAMno. 3, May 17, 1998.) Nurses, auxiliary nurses, and other professionals directly involved in healthcare delivery are now also eligible for this PI. However, there are problems both inthe design and implementationof this award. For doctors, calculating the value of the reward is complex. Each month the human resources staff in each public hospital must tally the number of consultations, treatments, and/or surgeries performed by each doctor. Depending upon the specialty and service performed, the award is valued between R$2.00 and $2.67 per service, with a cap on the total number of eligible consultationdtreatments that ranges from 158 to 352, depending upon the employee's work day (12 hours vs. 20 hours) and specialty (e.g., psychiatry, doctor, sanitary worker). (See Resolu@o SS-111, Secretaria de Estado da Sabde, June 19, 1998.) Multiplying the number of consultationdtreatments by their unit value yields the maximum amount per month that the doctor can potentially receive for "incentive pay" that month.' The actual amount of the PIaward can be less. Eligible employees (including doctors, nurses, etc.) are entitled to 50 percent of their total potential incentivepay simply due to the fact that they are employees of the Secretariat of Health. Another 30 percent depends upon the institutional evaluation, while the remaining 20 percent is contingent upon receiving a satisfactory individual evaluation.' There is a maximum of 20 possible points on the individual evaluation form. With 11points or higher, an employee is entitled to 100 percent of the one-fifth of the PIthat is determined by the individual evaluation. With a score of 10 points the employee i s awarded 50 percent; and anything less, 0 percent. On the evaluation form, an employee receives 5 points simply by showing up to work on time. Thus, it is quite easy to accumulate at least 11points. Indeed, in the records of Hospital Brigadeiro, for the month of July 2004,96% of those evaluated receiveda score of 11or higher, while the remaining four percent received a score of ten points. Not one of the 982 people evaluated received a score below ten. The Director of Hospital Brigadeiro acknowledged that PI is primarily a reward for reasonable attendance, not efficiency or effort. Thus, the PI largely fails to serve its purported role: namely, encouraging greater effort on the part of health care employees. Inpractice, it is largely an inputto base pay and not a "reward" for performance. Note: The PI for anesthesiologists and surgeons is calculated a bit differently. They receive a certain number of points depending upon the length of time of a particular surgery. They are then awarded $R20.00 for each point, up to a maximum of 50 points. (See the Memorundo Circular STRH, April 28, 2000.) Itis perhaps a designflaw that doctors are punishedfinancially for attending a medical conference, becausethat is not compensated under the PIformula. Note: Ifon approved leave or vacation, doctors, sanitary workers, and dental surgeons are entitled to receivethe same PIaward earned duringtheir last full month of work. It ispossible that OSS directorscan offer more internal promotion opportunities basedupon merit than traditional state hospitals. Further study is needed, however, to determine the extent of this 24 difference. On the whole, professional development opportunities for doctors and nurses often appear to be greater in traditional state hospitals than at OSS facilities, largely due to the greater variety and complexity of illnesses handled by the traditional public hospitals. Performance Evaluation and Supervision OSS and private hospitals often appear to have better information than their traditional public sector counterparts regarding hospital efficiency and personnel. For example, only three of the seven traditional public hospitals in our sample were able to provide data on their rate of absenteeism; and their rates tend to be inferior when compared to the private and OSS hospitals (Table 2.10). Similarly, none of these seven hospitals could provide data on employee turnover (Table 2.11). Table 2.10: Ratesof Absenteeism Absenteeism Direct Public with oss Private (Percent) administration Foundation (N=7) (N=4) (N=7) (N=2) 0- 1 2 4 1.1-2 3 2 2.1 - 3 2 2 3.1 -4 4.1 + 1 Data unavailable 4 Table 2.11: EmployeeTurnover Rates(amonghospitalsinstudy sample) Source: Results of research questionnaire applied ina sample of 20 Slo Paulo hospitals. The questionnaire appears inAppendix B. There is little doubt that supervision can be very poor at traditional public hospitals. However, it is possible to identify traditional public hospitals where supervision appears to be excellent. The Hospital Vila Penteado maintains monthly data for all medical personnel with the type of activity performed (e.g., consultation, surgical procedure, UTI, neo-natal, & ultrasound), and the quantities for each service. These data are used for calculating the Special Incentive Reward (discussed below), but they are also used by the senior managers at the hospital to track the "productivity" of all the medical specialists, and to identify areas or individuals where some interventionor discussion is warranted. 25 Certain public hospitals, as well as OSSs, also have begun to gather systematic information from their patients on the quality of their service. The Conte Comigo ("Count on Me") program began insevendirect administrationhospitals inSeptember 2003. The statedaims of the program are to offer support to patients and visitors to the hospital, to help them get where they need to be, and to transmit criticisms and suggestions to medical personnel and managers. The Conte Comigo teams prepare a monthly report for the directors of each division within the hospital, based upon the number of surveys they gather each month (300 to 400 in the case of Hospital Brigadeiro). These teams also collect specific criticisms concerning a particular doctor or staff member. Those are forwarded to the attention of the division directors for a reply. Then, the Conte Comigo staff communicates that responseback to the patient. The seven OSSs we visited all maintained a wealth of information on the number of services performed, by hospital units and by individuals. Yet, it was not clear from our focus group interviews that doctors believeformal supervision procedures are more prevalent or rigorous at an OSS than at other public facilities. Performance contracts between the SES and the OSS specify specific outputs. However, at none of the OSS hospitals in our sample did the director then allocate a portion of those outputs to individual staff members as a performance measure. DisciplinaryActions In all of the hospitals in our sample, the typical formal means of punishment for poor performance were a written notice, suspension, and finally removal. Inthe public hospitals it was also possible (at least in principle) to deny an employee their monthly performance award, as discussed in Box 2.2. Informal incentive mechanisms, including those related to discipline, are discussed below. Table 2.12 summarizes formal human resource management authorities by hospital type. Table 2.12: Managerial discretion over humanresource decisions, by hospitaltype Note: Although managers do not have the authority to set performance incentives, there is a performance award (discussed in Box 2.2) for employees of traditional public hospitals in Sgo Paulo. Meanwhile, although OSS managers acknowledge that they have the authority to enact a monetary performance award system for their employees, none of the managers inour sample of sevenOSSs made any such awards. THEPOWEROFINFORMALINCENTIVES(THROUGHFORMALANDINFORMALRULES) Comparing key features of OSS and traditional public hospitals as employers,and then examining the employment choices of a sample of healthcare professionals reveals the complexity of 26 personal utility functions. Labor economists often adopt a narrow view of human preferences, with a parsimonious utility-maximizing assumption whereby workers make individual calculations to trade off money vs. leisure. Clearly, the employment choices made by the healthcare professionals in our Sgo Paulo focus groups reflect utility functions that are far more complex than the parsimonious model suggests. Our hospital interviews confirmed the common perception that multiple-job-holding by healthcare professionals is, in large part, a strategy to minimize risk. (The "frame" within which these professionals make their employment decisions is marked by the memory of unpredictable changes in pay and employment in both the public and private sectors.) Nevertheless, it is clear from these interviews that risk-aversion is inadequate to explain their employment choices, particularly when professionals have notjust two, but three and four employment ties. Generally we were told that posts in traditional public hospitals were better for (i)security; (ii) professional development; and (iii)flexibility with hours/schedule. A job at an OSS, meanwhile, was considered superior for (i)its "organization" (professionalism) as a place to work, and (ii)the opportunity to be affiliated with a prestigious institution. (The same was true of the foundation hospitals.) As for private hospitals, these could offer a sense of prestige, and the potential for better pay. Thus, a mix of employment ties can satisfy a range of diverse objectives. Reducing risk is merely one of them. Based upon our analysis of 20 hospitals in Slo Paulo, including focus-group discussions with healthcare personnel in these facilities, there is little or no evidence to claim the superior performance of OSS hospitals results from higher salaries, performance pay, superior career development opportunities, or even formal supervision mechanisms. If the "managerial autonomy" enjoyed by the managers of OSS hospitals i s important to explain differences in performance, then where do we see it? The key answer appears to be staff composition, the processesof staff selection, and the possibility to dismiss employees.*' When the OSS Hospital Carapicuiba (Santorinhos) prepared to open in October 1998, the director decided to hire all his staff as celetistus. Not a single concurso was held. The director and his senior managers looked to the School of Medicine for potential employees, asking for recommendations and advertising through informal networks. They asked promising candidates to come in for interviews, and hired whoever seemed to be the best fit for the organization. Department heads made their selections, which then had to be approved by the hospital director. All doctors at Hospital Carapicuiba were hiredinthis semi-informal manner. For nurses, the director chose a more formal process for interviewing and evaluating potential candidates; but again, none of the nurses were selected from a list of concusados. Five employees were hired through formal, public advertisement of the positions, but managers and the director were disappointed with the quality of the resulting hires. That experience strengthened their belief that word-of-mouth recommendations were the best way to identify the best staff for the organization. The OSS Hospital Mario Covas selects staff in the same way. In 27 An alternative explanation, which relies neither on formal nor informal human resource management rules, is a variant of the so-called "Hawthorne effect," whereby worker behavior may differ (and in this case improve) not because of the content of the reform or "experiment" employed, but simply because the workers know they are being studied (or are subject to increased attention, as in the case of a high-profile reform). In this case, however, one would expect the positive "Hawthorne effect" to be temporary, as the initial excitement or attention created by first establishing a Social Organization gradually wears away. Social Organizationshave now beenoperating inSilo Paulofor more than five years, insome cases. 27 these settings managers are empowered to search for personnel that are both highly qualified and a good fit for the organization. State hospital managers cannot select staff they believe are best suited for their facility. If they have a vacancy they must hire the person highest on the list of con curs ado^ for that position. There is no opportunity to interview even a small number of pre-qualified candidates to determine who would be the best fit for the organization. There i s no flexibility or discretion. The overall number of staff i s another aspect where the director of a traditional public hospital has no authority. At Vila Penteado the number of doctors (341) is the same today as when the hospital opened in 1991. By all accounts Vila Penteado is one of the best performing direct administration hospitals in Siio Paulo. It is unique in that this hospital has had only one director since it opened; but the director still must overcome specific challenges or obstacles to put together a dedicated staff. The establishment lists have not been changed, so turnover is the only way that hospital managers can change the staffing profile. For example, if the hospital wants to add a bum unit, the director can not simply advertise for the correct medical specialists and hire those that are needed. He can only build this new capacity gradually as those with other specialties retire or for other reasonsleave the hospital. Turnover rates were high at Hospital Vila Penteado during its first years of operation. Firing a statutory employee is exceedingly difficult. Still, the director (and other senior managers) could make it unpleasant for employees who did not perform well by regularly providing feedback and criticism, making certain that shirking would not be tolerated. Eventually personnel who were not committed to the goals and culture of the hospital sought transfers to other facilities (which helped the performance of Vila Penteado, but simply transferred the problem elsewhere). Turnover rates are now lower than in the first years as the hospital has developed a certain reputation, which tends to screen out some unsuitable candidates. However, the problem has not gone away. The director is generally pleased with new staff named to fill vacancies in pediatrics and intensive medicine; but he is regularly disappointed with almost a third of the staff in the clinical area. There is no trial period, and no simple way to discipline or fire a poor performer. One option for disciplining a bad doctor in a traditional public hospital i s to make a formal complaint to the Regional Medical Council (CRM), a professionalbody that doctors mustjoin in order to practice medicine in the public or private sector. However, this option has two important limitations. First, the CRM investigates and judges potential malpractice or ethical infractions, not complaints that a doctor is merely inefficient or unproductive. Second, the process can be extremely slow: 4-5 years from the time of the initial complaint until final resolution. The CRM will first carry out a preliminary evaluation to decide whether a disciplinary case is warranted. Then, if the CRM determines that it is justified, a disciplinary procedure is launched. Various parties can be called to testify. Penalties range from (i)confidential warning, (ii)confidential censure, (iii)public censure, (iv) suspension (usually 30 days), and (v) dismissal (disaccreditation). Many managerswill conclude that denouncing an unprofessional doctor to the CRM is simply not worth the trouble. As the president of CRM explained, "when you denounce someone you cannot be sure of the outcome; but you likely will continue to work inthe same hospital with that person until the disciplinary proceeding is finally concluded." Even if the person is eventually dismissed, the manager cannot be certain that the one named to fill that vacancy will be any better. 28 In Brazil, as in other countries, managers often are reluctant to fire, penalize or give poor performance evaluations. After all, the manager is not the residual claimant for any budget -surplus oftime, personalconflict, etc.-associatedwithpenalizingorfiringanemployee inthe (see Chapter 4), so there is not a strong incentive for a manageddirector to suffer the costs interms interest of greater efficiency. Clearly, however, the costs for disciplining an employee at an OSS are much lower than in traditional direct administration hospitals. Two doctors were fired from the OSS Hospital Mario Covas in 2003, and another was fired in 2004. A fourth doctor was "dismissed" from Hospital Mario Covas at the end of his probation period. (Under the CLT legislation, managers also have the benefit of a three-month probationary period to see if a new employee is a good fit for the organization.) Each of those actions could be taken relatively quickly. Meanwhile, the director of Hospital Vila Penteado began a disciplinary process in the year 2001 to fire a doctor for malpractice. Eventually, as the process was nearing conclusion in 2004, the doctor resignedto evade an official sanction. The performance advantage, on average, of OSS hospitals appears to result largely from the ability of OSS managers to use information networks and informal incentives in personnel selection. Can this informality be abused? Certainly. On the other hand, it enables senior managers to assemble a group of employees with a common commitment to an organizational mission and culture. That collective spirit is difficult to quantify but terribly important. It helps to align the goals of principal and agent, thereby reducing monitoring costs and generating levels of performance that can only be induced, not forced. It is impossible for any manager to observe his subordinates constantly. The best hospitals are those where employees want to do good work and where they believe that their colleagues have the same commitment. Where this is true, then employees are likely to give their best ("cooperate" to achieve a common goal) in the belief that their colleagues will do the same. If, on the other hand, an employee believes that his colleagues will evade their work whenever possible ("defect"), the probability that the employee also will shirk increasessubstantially (Miller 1992).28 A professional, committed organizational culture created and sustained by skilled managers is a setting that supports the reciprocal cooperation that underlies highly effective and efficient organizations. Ifthe organizational culture is one in which cooperation is expected, then there are informal but powerful means to punish a "defector." The Hospital Vila Penteado offered an interesting illustration. Nurses often arrange among themselves to swap shifts (e.g., because one of the nurses wants a three-day weekend for a trip out of town). However, nurse supervisors must approve those requests. Generally, they do. However, as a form of punishment (and to induce cooperation) supervisors do not approve schedule changes for nurses who have showed up to work late or in other ways been "defectors." The supervisors are clear with subordinates about their reasons for approving or denying such requests. The message is "I'llcooperate with you if you cooperate with me (and the rest of the hospital staff) in delivering high-quality, efficient health services at this hospital." The Vila Penteado example demonstrates that an environment of collective cooperation toward a common goal can be induced and nurtured at direct administration hospitals, not only at OSS and private hospitals. It is important to recognize, however, that a well-meaning OSS manager has important tools at his disposal that his counterpart at a direct administration hospital lacks. And that can make all the difference. ** "The principal knows, ingeneral terms, what he wants the agent to do, but the range of possible actions that the agent can take, and the rangeof possible outcomes, is enormous." (Baker et. al. 1988:598). 29 CONCLUSIONS RECOMMENDATIONS & The descriptive data gathered from our field research at 20 SHo Paulo hospitals is consistent with the data analysis of the World Bank study on hospital performance (2006, forthcoming): OSS hospitals as a group have outperformed their direct administration counterparts. The public perception of the OSS reform is also positive, so much so that the government is exploring the possibility of extending the model to additional state hospitals. Nevertheless, it is easy to overstate the magnitude of the difference between these hospital types, and to oversimplify the explanation for the superior OSS performance. Many OSS hospitals do not offer emergency services; and many control their demand as "closed door" facilities. This is not to suggest that the superior performance, on average, of the OSS model is simply an illusion. Rather, this is merely a plea for modesty in "selling" the OSS model, noting that there are practical limitations to expanding this organizational innovation throughout the broader healthcare system. First, we must acknowledge that those large, less efficient state hospitals produce positive externalities for the health system as a whole in SHo Paulo. As our focus-group interviews demonstrated, OSS hospitals are sparedthe least common and most expensive medical cases that are handled by the large public teaching hospitals. Many doctors and nurses learn their craft at those hospitals, tackling the broadest range of medical cases. They can then take their talents to positions inOSS and private sector hospitals. The private sector, we were told, does not hire nurses without experience, and the public sector provides that experience. Second, there are political constraints to expanding the OSS model that did not have to be confronted head-on when the first OSSs were established. Recall that SHo Paulo's OSS experiment was launched in brand new hospitals. Therefore, they did not provoke strong labor opposition. Converting a direct administration hospital into an OSS is a different matter, as evident in the experience of Hospital das Clinicas Luzia de Pinho Melo. When a new wing was added to this hospital, increasing its size from 50 to 300 beds, the state initially intended to administer the "new" section of the hospital as an OSS while the "old" section would continue under traditional direct administration. It soon became clear, however, that two administrative models operating in the same facility was untenable. Thus, the SES determined to convert administration of the entire hospital and its 660 current state employees to OSS management. Several key reassurances were offered to current employees to avoid judicial challenges and mitigate labor union unrest before the OSS management contract took effect in October 2004: i) no employee would lose their current employment tie as a state employee; ii)salaries (including the PI) would not be reduced; and iii)those who did not wish to remain at the hospital administered as an OSS could receive a transfer to another state facility. One year later, only 268 (41 percent) of those 660 employees remained under the new OSS management.29 The experience of Hospital das Clinicas Luzia de Pinho Melo appears to confirm the supposition that many health sector employees prefer the less demanding work rules often found in conventional public administration hospitals, including the o portunity to negotiate their own work schedules, thereby facilitating multiple job-holding! Moreover, it suggests that widespread conversions would runup against a daunting political and fiscal obstacle: the need to 29 Data kindly provided by Dr.Nacime SalomSio Mansur, Superintendentof UNlFESP-affiliatedHosptials. 30 According to the data provided by Dr.Nacime SalomSio Mansur, the demand that employees fulfill their formal hourly schedules -- evenas they received higher pay -- was the reason given by 102doctors and 67 nurses/nurse assistants who chose to leave Hospital das Clinicas Luzia de Pinho Melo once it was converted to an OSS. Altogether, 121 doctors and 186 nurses/nurse assistants left the hospital for other public sectorposts. 30 find posts for all the disgruntled employees seeking a transfer from their newly-converted OSS. For obvious reasons, these problems were sidestepped when OSS hospitals were created from scratch. How, then, might the positive elements of the OSS model be extended? In late 2005 the Ambulatorio Maria Zelia, with 470 employees, was transferred from direct administration to OSS management, indicating that the performance contract model has not yet reached its limit in Stio Paulo's public health system. However, if we are correct in assuming that the OSS model cannot completely replace traditional public hospitals, how might human resource management in direct administration hospitals be improved, drawing on lessons from the OSS experience? The analysis presented here suggests three possibilities: i)improving procedures to discipline/fire, ii) enhancing managers' authority to hire, and iii)a more radical reform to create a new employment regime for healthcare professionals. In2002 the Procuradoria Geral do Estado de SZo Paulo promotedthe strategy of the ViaRapida ("Rapid Path") as a means to hasten administrative procedures to dismiss non-performing or shoddy public employees. However, the Complementary Law that established the Via Rdpida retains centralized disciplinary procedures, with a guaranteed right to defense that requires depositions of the relevant persons involved (e.g., managers, co-workers, subordinates). Thus, the practical effect of this policy has been minimal, at best. A true "via rdpida" will require a re- balancing of the legitimate rights and protections of the employee with the legitimate need of the state to discipline or dismiss those employees who, to the detriment of their fellow citizens, do not adequately fulfill their work-related responsibilities. In that way, non-performing ("defecting") employees could be fired in a timely manner, rather than simply being transferred. The second option centers on the authority to hire: granting public hospital managers greater discretion over staff composition and selection. For example, by allowing hospital directors to interview the top 3-5 people who sit atop the list of concursados, the merit principle would be protected while simultaneously enabling the director to shape the culture of the organization to generate greater collective commitment to the goals of the hospital. As we have seen from the field research for this study, generating cooperation does not require a pay-for-performance system. However, it does require managerial skill. Some managers can generate commitment and cooperation even under existing employment rules (e.g., the director of Vila Penteado). But heroic, highly-talented managersare scarce. Adapting discretionary features of staff composition, personnel selection and dismissal in the OSS model to direct administration hospitals should, in the hands of well-intentioned managers, generate improved hospital performance. While either of the first two options would be fraught with political and legal challenges, a third option is to craft for the state's health sector employees an entirely new employment regime that has more in common with the present-day CLT regime. To ameliorate the political opposition and legal challenges that such a proposal would engender, adherence to the new regime for current employees presumably would have to be voluntary. However, all new employees would be hired under the new statute; and gradually the management of human resources in the health sector would be transformed. We recognize that the constitutional mandate of a Single Juridical Regime for public employees in Brazil is a powerful obstacle to such a change. However, the case for reform is compelling. 31 3. MANAGINGPRIMARY CAREINTHE CITY OF CURITIBA INTRODUCTION Over almost four decades of steady development, Curitiba has emerged as an exemplary practitioner of the Brazilian model of public primary healthcare. This chapter seeks to understand how management practices, especially in the area of human resources, have contributed to this achievement. What are the various measures that Curitiba has taken to make its healthcare workers more performance-oriented, more attuned to their outputs (the services they provide) and outcomes (the impact of these services on the health of their clients)? Our particular concern i s to understand how employee incentive schemes and management contracts might have made staff more performance-oriented. Within Curitiba's primary healthcare system, this chapter concentrates on health clinics (Le., the health units that provide non-emergency preventive and curative services) and primarily in the context of Brazil's national policy of primary healthcare reform (Chapter 1). This national policy played the leading role in defining the general trajectory of primary healthcare reform at the municipal level, and we shall see in this chapter what Curitiba did to make this national policy work at the local level. While the starting point to this chapter - the relevance of employee incentive schemes and management contracts - appears narrow, the issues addressed will be broader. The kind of products an organization produces, the way it produces them, the way it is compensated, the quality of available human resources, and the environment in which it works (the way its clients relate to it and the way its owners direct its activities and facilitate its access to resources) all have an influence on the options available to managers. Thus, performance management tools must be examined in the context of the overall functioning of an organization. This chapter therefore looks not only at how healthcare is managed, but also at what the primary healthcare system produces, how medical/clinical and management issues relate to each other, and, briefly, how healthcare fits into the broader management and politics of the city. To support the analysis, we carried out a confidential survey of perceptions concerning management issues among a sample of professional staff of Health Units. The survey is available as a separate document and its results are summarized in Appendix D. The results of the survey are generally in accordance with interviews we conducted in Health Units, Health Districts, and the Municipal Health Secretariat (SMS). Inthe end, because of the complexity of the SMS as an organization, we have primarily relied on interpretation and judgment, rather than numbers, to understand how the system works and why it came to be what it is. PRIMARY HEALTHCARE IN CURITIBA The ApproachtoPublicHealthcare Under the SUS, provision of services is decentralized a far as possible to states and municipalities and jointly financed by all three levels of government. Curitiba has full responsibility for all levels of health service (Gesth Plena). In2005 its services were 60 percent federally financed, with the balance coming entirely from the municipality. A dominant part of the city's primary healthcare (general medicine) and a substantial amount of secondary healthcare (specialized medicine) are in public hands. The municipality runs 110 of 32 the city's 281 ambulatory (primary and secondary) health facilities (Table 3.1). This includes 105 Health Units (Unidades de Salide) in which 80 percent of the city's population is registered. These Units provide most of the city's primary healthcare and a large amount at the secondary level. (Table3.2 provides a breakdown of employment in the Health Units by type of unit and type of worker.) Most tertiary-level care (hospitals) is in private or non-profit hands. The city pays these hospitals on a fee-per-service basis. Table 3.1: Number of Public and Private HealthUnits in Curitiba, 2005 Public Non-profit Private Total Nivelambulatorial: Unidades corn Programa Salide da Familia 42 Unidades de Salide com especialidades 12 Unidades de Salide 24 Horas 5 Hospital geral e maternidade 1 Laboratbrio de Anilises Clinicas 1 Outras Unidades Bgsicas 44 Total Unidades de Saiide 105 17 135 257 outros 24 24 Total 129 17 135 281 Assisthcia hospitalar 3 12 15 30 33 Currentprimary healthcare inCuritiba reflects a philosophy built up over more than 25 years and owing much of its character to the Brazilian healthcare reform movement and the SUS, but also to more local developments. Based on descriptions of the system given by practitioners and from observations of how the system works, the following appear to be central elements of the primary healthcare philosophy (modelo de atengtso bdsica): 0 Primary healthcare is based on a humanized and socially-oriented practiceof medicine. To caricature the traditional model, the doctor treats the body as a machine and cures through medical technology. A newer model envisages a broader team of health professionals which also prevents and cures through a greater attention to the particular needs and situation of an individual patient and a greater understanding of hisher en~ironment.~'This can change the typical role of the doctor from a leader of a team to a specialist within a team.32 The new model is partly reflected in the principles of family medicine, where generalist doctors build long-term relationships with patients, put a patient's health in its social context, practice preventive as well as curative medicine, and are oriented to the well-being of a whole community.33 0 Primary healthcare is organized to focus on local problems. Specific areas covered by different Health Units (and micro-areas within them) face specific health problems that can be attacked systematically. For instance, an area populated by industrial workers and one populated by poor rural immigrants face different problems regarding public health or levels of violence. This local focus is operationalized through a decentralized primary-care network of nine Health Districts, each with around 10Health Unitsfor primary care. Primary healthcare emphasizes preventive practice. Curitiba does this mostly through health for population (Ca`rie Zero - Amigo Especial).Y4 specific programs, for instance programs for safe regnancies (Mtse Curitibana) or dental 0 More recently, Curitiba' s approach has been to emphasize evidence-based medicine, medical practice based on scientific evidence, distilled where possible into standard operational processes (written up, as Curitiba has done, into specific protocols for treating specific problern~).~~ Consistently since 1979, the vision in Curitiba has been to replace a traditional public health hospitals) - with a more preventive and family-medicine-orientedapproach. The application of system-combining vertical programs for specific conditions and curative medicine (emphasizing this philosophy has evolved over time (see next section). The systemhad to be decentralized as it grew and became more complex. And the family-medicine approach evolved, in the early 1990s, into the Family Health Program (PSF) model, a standardized approach to primary health 31 This is strongly linked to Brazil's healthcarereform movement (Movimento da Reforma Sanittiria) and the theory that illness is socially determined (the "social theory of medicine"). For a short history of this movement, see httu://bvsarouca.cict.fiocruz.br/sanitarista05.html. For an application to Curitiba of an a proachakin to the social theory of medicine, see astudy of children's dental health by MoysCs (2000). ,`see Porto (2001). 33 See, for instance, Canada's Four Principles of Family Medicine (College of FamilyPhysiciansof Canada at hthx//www.cfpc.cd). The Canadian government has, through the University of Toronto, worked closely with Curitiba on training infamily medicine. 34 Some of these programs are described inDucciet al. (2001). 35 "EBM is a conflation of three distinctive essences: an epochal scientific hypothesis; an ever evolving body of evidence; andan idealised professionalprocess-a way of practisingmedicine." Reilly (2004) 35 calledBasic Units- were less standardized. emphasizing teams and community outreach. Curitiba's Health Units founded prior to this -now 0 PSF teams cover a population of 3,450 on average. They have a standard structure mandated by the federal government (which provides financial support to municipal PSFs). Each PSF Unit has a Health Unit manager (ASL) and three to four teams, each operating in different micro-areas. Each team has a staff of about nine public employees (one generalist family doctor, one nurse, three nurse's aides, a dentist, a dental auxiliary, a dental hygiene technician), and typically four Community Health Agents (ACS), employed under private labor law (CLT).36 (See Table 3.) PSF Doctors (like PSF dentists) have a 40-hour week and have often made a career choice to enter family medicine. Community Health Agents -health workers with one-to-two years' training - make family and community visits and help ensure that all families are identified inthe catchment area. 0 Some Basic Units have been converted to PSF Units, but Basic Units still account for more than half of all of Curitiba's Health Units.37 They are comparably sized with PSF Units (Table 3.3), but are not constituted in teams. They serve larger populations, and do not have the same resourcesto go out into the community. They have a looser structure: typically they have the equivalent of three full-time doctors (general practitioner, pediatricians, and gynecologists working a 20-hour week). Some ten Basic Units have a specialization (for instance inHIV/AIDS) and serve clients beyondtheir area. Table 3.3: Municipalityof Curitiba: Average Numberof Staff EmployedinHealthUnitsby Type ofUnit Unidade Unidade Us All PSF UMS-24 BBsica Espec. Units B. PublicServants MCdico 4 0 5 10 5 Enfermeiro 3 15 2 2 3 Auxiliar de Enfermagem 11 66 12 11 14 Odontologo 3 6 4 4 4 TCcnicoemHigieneDental 2 1 2 0 2 Auxiliar de Consultdrio DentArio 4 6 5 2 4 Agente Administrativo 1 3 1 3 2 Auxiliar Administrativo Operacional 2 9 1 2 2 Other 0 2 1 3 1 Total 31 108 33 38 36 C. Agentes ComunikiriosdeSaGde 14 11 36 Inmany Brazilianmunicipalities, PSFstaff is employed under private labor law. 37 It isthe supply of family doctors that has preventeda faster expansionofPSFUnits. 36 The MainFeaturesof Management The Municipal Health Secretariat (SMS) has around 5,000 employees and accounts for one quarter of all municipal employment and one fifth of the municipal budget. Curitiba's primary- healthcare management policy is discussed at length below, and can be summarized as follows: The SMS pursues a strong client-orientation, by organizing its services to reach clients, by empowering the client, and by making the client's life easier. The municipality putsa strong emphasis on usingsocial and medical knowledge and it has an advanced computerized information system which supports the use of knowledge and makes a complex management system work efficiently. The SMS has concentrated its resources (and its most important innovations) in the area of primary and preventive medicine. It has further sought to reap the benefits of specialization by emphasizing problem-specific programs and by standardizing medical and management processes. It also has had to invest heavily in systems to coordinate the various vertical and horizontal, private and public agencies providing health services, and to make the primary- healthcare network the gatekeeper (porta de entrada) for these services. Human resource management remains largely within the traditional Brazilian paradigm of the Public Servant (servidor estututa'rio). There is, additionally, substantial emphasis on training and professional development. Performance management is emphasized. Strategic planning has long been a norm of management behavior in Curitiba. More recently, the SMS has developed a management contracting system. PrimaryHealthcareOutcomes The premise of this chapter is that Curitiba has a well-run health administration that produces results. The available evidence does not allow us rigorously to relate results to administrative practices. However, it is important to make some mention of the kind of evidence that leads to Curitiba's reputation as a strong performer in primary healthcare. Fully 80 percent of Curitiba's population is registered in the public health system. (30 percent of the population is privately insured, but makes partial use of the public system - for vaccinations, for instance.) The users of Health Units appear to be happy generally with the services they receive, according to regular telephone interviews with a sample of the pop~lation.~~ A recent study of the impact of public policy in Curitiba by the Municipal Institute of Public Administration (IMAP,2005) pointedto some typical achievements of the health system: Health education has eliminated locally-originated dengue fever, improved the dental health of children upto 12-year-olds, and contributed to reducingunwantedpregnancies. Vaccination programs have reduced measlesto near zero and dramatically reduced meningitis (Haemophilus Influenzae b) ininfants after 1996. 38Carried out every quarter since May 2004 by a separate government agency, this survey (Avaliago'esdu Cornunidude) elicits about 2,800 responses for Health Units (PSF and Basic Units) and over 300 for emergency Units.Inthe November 2004 survey, 87 percentof respondents found the physical installations of Health Units satisfactory; 61 percent considered the services good or very good, and 26 percent middling; and 76 percenthad their problem resolved satisfactorily. 37 0 A reduction inthe number of newborns-at-risk since 1999 is believedto be the result of anew system that Curitiba pioneered to register and track all pregnancies and make home visits. 0 The M6e Curitibana Program has reduced mother-to-child transmission of AIDS (from nine percent in2000 to zero in2003) and reduced infant and mother deaths. Sant'Ana et. al. (2002) studied the impact of the first PSF Unit, Sa0 Jose, opened in 1993. They found the following achievements: 0 The Unit contributed to a reduction in infant deaths (there were no infant deaths in this PSF area after 1995), due to better pre-natal care and a decline in the share of adolescent and unplanned pregnancies. Child nutrition also improved. 0 Health education was able to reduce the amount of hepatitis A infection (which resultedfrom children walking or playing in sewers). 0 There was a dramatic improvement in the delivery of preventive services, reflected in a rise inblood pressure tests, PAP smears, infant examinations, and Type-11diabetes tests. Partly because a nursery was attached to the Unit, there was a decline from 1994 to 1997 in under- weight and under-height children. 0 The Unit helped to encourage the local slum-dwelling population to clean up garbage, and healtheducation ledto a local demand for sewers (which were constructed in 1998). Good, accurate information played an important part inthese achievements. Consistent with these healthcare outcomes, Curitiba has gained a reputation in Brazil for the quality of its primary healthcare system. For instance, it pioneered a procedure, now adopted elsewhere in Brazil, for registration and home visits for all births (Declarqo de Nascido Vivo, DNV). Itwas Brazil's first city to computerize the electronic medical records system (Prontudrio Eletrdnico). It is Paranh's only municipality to carry out drinking-water inspections. Curitiba provides HIV/AIDS indicators information for other local governments on PMC webpage MonitorAIDS. Curitiba's clinical protocols have become references for other Brazilian cities. It has won a number of prizes and awards (including for its PSF and the Programa M6e Curitibana). Given that health services under the SUS are free, a consequence of Curitiba's success has been to draw inclients from outside Curitiba. This is a large problem for the city: 30 percent of registrations are estimated to come from people who live outside Curitiba. Curitiba performs well for selected health indicators when compared to other municipalities with a similar epidemiological and demographic profile. In 2004, Curitiba ranked either first or near the top among comparable cities in a selection of robust indicators presented in Table 3.4. For instance, Curitiba has one of the lowest infant mortality rates in Brazil with 11.2 deaths per thousand births. This is the result of intensive pre-natal care since 80 percent of births are from mothers who attended to seven or more pre-natal consultancies. It also performs at the top for non-fetal deaths by non-defined causes, with 1.1 percent. Similarly, Curitiba reduced the inpatient rate for acute respiratory infection in children under five years from 18.1 percent in 2002 to 14.4 percent in 2004. The city has a highaverage of 1.6 annual medical visits per person inbasic care. From 2002 to 2004, the interruption rate of tuberculosis treatment fell from 14.22 percent to 11.92 percent, a significant improvement. 38 The Evolutionof the Primary HealthcareSystem Curitiba's primary-healthcare workers show a pride inthe city's achievements that i s reflected in a strong sense of history: they describe today's system as the result of a quarter-century of consistent development and problem-~olving.~~Table 3.4 provides a chronology. This consistent line of development began in 1979 with the creation of a Health Department and the initiation of a healthcare model based on four principles: democratization of healthcare; extension of coverage and prioritization by level of care; integrated medicine; and community participation (Ducci et al, 2001:16). During the 1980s the city opened a number of largely standardized basic Health Units -many of these remain as Basic Units today. A body of health workers developed in parallel. Many of these workers have remained, and although they are now nearing the end of their careers, their corporate views on health and management still dominate the SMS. In the 1980s health managers put a heavy emphasis on training and on the use of foreign models. The computerization process began in 1998. The 1988 Constitution and the federal organic health laws of 1990 further embedded the principles of democratization and universality of healthcare in Brazil. These principles were in large part the product, like Curitiba's own reforms, of the Brazilian healthcare reform movement (Movirnento du Reformu Sunitdriu). This was a remarkable revolution, inthe context of Brazil's return to democracy, in which a professional group formulated a radically alternative policy, and then gained key political and bureaucratic posts to implement the new policy. In the 1990s the shape of primary-healthcare policy in Curitiba was substantially influenced by the federal government. Decentralization under the SUS gave municipalities the financial opportunity to develop their own health services, and the SUS also set the rules on both structures and processes that all municipalities hadto follow. The PSF model was particularly important. 1941-43 0 Curitiba's first urbanplan, preparedby Frencharchitect Alfred Agache; the first public health unit is founded in 1941. 1960s 0 The Agacheplan is improvedby agroupof architects andurban specialists coordinatedby the CuritibaResearchandUrbanPlanningInstitute (PPUC), founded in 1965.Through the Fundq60 Servips Especiais de SalidePliblica (FSESP), with United States support, dental clinics are establishedinsome public schools. 1970s 0 Largewaves of immigration into Curitibacreate slums (favelus) and increase the demand for public services inhealthandeducation. There is an increaseinthe number ofHealth Units. 1978 0 At the World HealthConference (Alma-Ata) some 130signatory countriesemphasizethe importanceof primary healthcare. 1979 0 Curitiba's healthcaremodelis formulatedinthe spirit of Alma-Ata andbasedonfour principles (democratizationof healthcare; extensionof coverageandprioritizationby level of care; integrated medicine; and communityparticipation).Curitiba creates a Health Departmentwithin the Departmentof SocialDevelopment(DDS) and implementsa networkof 10primary healthunitsand 13 separatedental units. 39For historicalaccounts, see: Ducciet al. (2001) onprimaryhealthcare ingeneral; SecretariaMunicipal da Salide de Curitiba(2002) on the PSF; Dercy Silveira Filho et al. (2002) on dentistry; and Venetikideset a1 (2001) on mentalhealth. 40 1986 0 Curitiba creates the Municipal Health Secretariat (SMS). 0 At the Vttt Conferencia Nacional de Salide, Brazil's health reformmovement discusses the concepts o f universalhealthcare and decentralizationthat are later incorporated inthe -- Constitution (1988) and UnifiedHealth System (SUS) (1990). 1988 0 Constitui$io Federal do Brasil (chapter on safety and health) 1990 0 Lei Orgdnica da Salide (8080,8 142), a first direct impulse for reform. 1991 0 Firstprototype family healthcenter opens (Centro de Salide Pompe`ia),basedonforeign experience, with a pediatrist, gynecologist, and generalist, and later a dentist. 0 FirstEmergency HealthUnit(24hours) opens. 0 ConselhoMunicipal de Salide (and first local councils) created inCuritiba, leading to the I Conferencia Municipal da Salide. 1992 0 Curitiba acquires responsibility, from the State, for public health and primary healthcare. It decentralizes basic healthcare administration with the appointment o f Autoridades Sanitdrias Locais (to headHealth Units)and Gerentes de Distritos Sanitdrios (to head seven Health Districts). 0 HealthUnitsintegrate medical and dental services. 0 Managerial problems (human resources weaknesses, informationdeficiencies, lack of financial resources), as well as political opposition, begin to affect the Centro de Salide Pompe`ia. 1993 0 Curitiba acquires responsibility, from the State, for hospitals. 0 SMS establishes a system to rationalize the purchase and supply of medicines (later to become Fannacia Curitibana). 0 The family-health approach is pursued again through a new initiative, also with foreign assistance, but with considerably more community participation instudy and design: Unidad de Salide Sun Jose`is opened. 0 Seven local conferences precede the II Confergncia Municipal da Salide. As a result, the family-health approach of US San JosC begins to be extended to other Health Units. 0 The city starts a telephone complaints and information line for public services (Central de Atendimento ao Usudrio) 1994 0 The IDGQ is established, a scheme o f variable salary bonuses to encourage staff to locate in out-of-the-way Health Units. 0 A Central de MarcagGode ConsultasEspecializadas is established to rationalize the process o f referring patients from Health Units to other healthcare institutions. 0 A federal PSFis proposed, and SMS develops a plan to increase PSFHealth Unitsin Curitiba. 1995 0 Curitiba acquires responsibility, from the State, for secondary healthcare (Sistema Ambulatorial). ttt Confer2ncia Municipal daSalide. A new bonus scheme, PIQ, is introduced to give a 30 percent bonus to staff insomeUnits, after discussion and approval o f the ConselhoMunicipal. Three new PSFUnitsare opened. PSF staff innew Unitsreceive a PSF bonus. Training for PSF staff with a consultancy from Toronto initiates a "process o f permanent education." D Management training for Unit leaders: Cursode Gestciode UnidadesBrisicas de Salide 1996 D Federal government enables GestcioSemiplenafor Curitiba health system. D Curitiba adheres to federal PSF. Large expansion o f PSF, with 18 new Units (new staff subject to concurso;revised level o f bonus; substantial training). 1997 D PIQterminated in 1997. D tV Confer2ncia Municipal da Salide D Intensive PSF training continues. 1998 D Federal government enables GestcioPlena for Curitiba health system. D Universities play a stronger role infamily-health training. A family-health professional 41 association(FAMIPAR)is formed. 0 Two morePSFUnits are established. 0 SMS beginswork on client-friendly redesignof processes(AcolhimentoSoliddrio). 1999 0 Following the Federal lead, Curitiba begins to employ, under private labor law, Community HealthAgents (Program deAgentes Comunitdriosde Saride,PACS). 0 IFAMIPARConference. 0 V Confer2ncia Municipal da Salide 0 Implementation of electronic medical recordssystem (Prontuhrio Eletranico) and computerization of the health municipal network 0 Implementation of supportcenters 0 Training for Acolhimento Soliddrio is started, process reengineeringis started, and clinical protocols are prepared. 2000 0 IIFAMIPARConference. 0 Extensionof PSF: eight new PSFUnitsare opened; one Health District (BairroNovo)is completely converted to PSFUnitsand is reconfigured to cover morecomplex processes. 0 Developmentof education and training onEBM(supportedby University of Toronto). 0 IDQiscreatedas anincentive schemefor performancebonusesonthebasisof evaluated performanceof the individual and the team. 2001 0 The Centro de Educagrioem Salide assumes responsibility for training. 0 The ProntuhrioEletranico de Sadde, a data warehouse/management system(incorporating medical records, systems for different healthprograms, and integrating HealthUnits with other services and functions), is fully running. VI Confer2ncia Municipal da Salide 0 Evaluation of PSF staff andpayment of PSFbonuses are rationalized. 0 Five new PSFUnitsare opened. 2002 Sistem Zntegrado de Servigoosde Saride (SISS) formalizes the establishmentof a systemfor horizontal and vertical coordination amonghealth services. 2003 0 The system of managementcontracting (Termosde CornpromissoPOA) is initiated. 0 VII ConfersnciaMunicipal da Saiide 2004 0 Regular telephonesurvey of HealthUnitclients begins. Curitiba's first prototype family-health center opened in 1991. This was also the year in which organs of social control (Municipal and Local Health Councils) were established. The process of transferring to Curitiba responsibility for health, under the SUS, began in 1992 and was completed in 1998 (Gestzo Plena). In 1993 the family-health-unit model crystallized with the creation of the Siio JosC Unit. The federal government proposed a PSF in 1994. At this point the primary-health model had more or less crystallized inits present form. Then, from 1995 to 2001, there was a strong expansion inCuritiba's primary-health system. The rate of expansion has since fallen. In the early 1990s, as the health system expanded under the health model of SUS and the PSF, Curitiba began to recognize the growing problems - actual and prospective - of management: the This recognition - the fruit of a tradition, if informal, of strategic planning - ushered in a new, system's growth was challenging the way that people, money, and information were managed. more "managerialist" era. From around 1993-94, a process of more-or-less permanent management innovations began: the system was decentralized in 1992; measures to improve the quality of services to clients were initiated from 1993; pharmaceutical purchasing was rationalized from 1993; measures to articulate (integrate) the various parts of the health system began in 1994; the first of several staff-bonus systems was introduced in 1994; there was a large management-training initiative in 1995; medical training was put on a more systematized basis from 1995; new staff (Community Health Agents) were employed under private labor law from 1999; the information system was completely integrated by 2001; management contracting was 42 introduced in 2003. Curiously, the stream of major management initiatives appears to have slowed down after 2003. This history is characterized by a high rate of innovation, first in healthcare (within a dominant guiding philosophy), then in management. Innovation occurred in an environment of rapidly growing demand for the city's health services. In the 1980s population growth was almost five percent a year (Table 3 3 , as a result of continuing (but slowing) rural immigration. The population growth rate has since fallen (2.1 percent in the 1990s and 2000s). But in the 1990s Curitiba also had to assume the challenge, under the SUS, of taking over full responsibility for the health of its citizens. (Health expenditures grew dramatically from 8 percent of Curitiba's budget in 2000 to 20 percent in 2005.) Inaddition, the demographic and developmental transitions have changed the profile of health risks as the population has aged and the level of violent deaths has risen. Table3.6: Curitiba:PopulationGrowth,1940-2005 Number Periodic Series A Series B Growthrate 1940 140,656 1950 180,575 2.5% 1960 361,309 7.2% 1970 609,026 5.4% 1980 1,024,975 5.3% 1990 1,608,15 1 4.6% 1991 1,3 15,035 2000 1,587,315 2.1% 2005 1,757,900 2.1% Source: Series A: Rabinovitch and Hoehn (1995), Table 1 Series B: IBGE THEMANAGEMENT OFPRIMARY HEALTHCARE The central aim of this chapter is to understand how Curitiba's primary healthcare system is managed, particularly human resources. To help do this, we carried out a survey of incentives for human resources in Curitiba's system (Appendix D). This survey paid particular attention to the way that performance contracting (Tenno de CompromissoPOA) and the current bonus scheme (IDQ),worked. We have grouped the various management instruments under five headings: client-orientation; information management; specialization; managing human resources; and managing performance. These instruments can aim to affect either health outcomes or management outcomes or both, and they affect each other.40 The table in Appendix E synthesizes these measures and their effects. Our focus is on managing human resources and on planning and managing performance, rather than on the other three elements. Here is an example of how different systems can interact. Information systems improve epidemiological studies, which in turn contribute to better (evidence-based) medicine, which in turn allows integrated protocols to be established. These protocols help to improve management: first, they permit a greater integration of differentmedical services; second, they standardize clinical practice to the point where, with the aid of the information system, managerscan more accurately monitor performance. 43 Client Orientation Curitiba took a series of steps inthe 1990sto give its citizens greater access to health services, to give them a measure of "ownership" and power, and to improve the quality of the services they receive. 0 Reaching the client The construction of a decentralized system of Health Units linked by public transport began in the 1980s and was consummated in the 1991 decentralization into Health Districts. The Family Health Program (PSF), later reinforced by the introduction of Community Health Agents, was designed to create a more permanent link between health provider, the patient and its family. Beyond the better health outcomes that might be expected from this approach, decentralization also has management implications by shifting responsibility for results down the chain of command, to Health Districts, Units, and outreach workers. Empowering the community. Since 1991, tripartite Municipal and Local Health Councils (and later District Councils) have met every two years to discuss healthcare services. The Municipal Council has set up several commissions to make recommendations on specific topics (for instance on women, old-age, or the budget). It votes on the four-year Municipal Health Plan and on the health chapter of the annual Budget Planning Law (Lei de Diretrizes Orpmentdrias) and reports on budget exe~ution.~'Local populations seem to consider that Councils are an effective channel for complaints andfor proposing improvements. 0 Empowering individual clients. Curitiba's healthcare clients have also been individually empowered by a telephone complaint and information system for city services (Central de Atendimento ao Usudrio (CAU), established in 1993) and, more recently, by twice-yearly telephone client surveys on Health Units (Avalia~o'esdu Comunid~de).~~ Survey results at the level of the individual Unit provide community evaluations that form part of the current system of personnel evaluation (which in turn has consequences for the IDQbonus). Our survey indicates that Health Unitmanagers are well aware of these evaluations and make use of them to improve client services, Empowering the community and individual clients allows the municipality to benefit from local knowledge and preferences with respect to health issues and services. It also allows the municipality, the monopoly supplier of a free service, to gauge whether it is providing the right service. 0 Making the client's life easier. In 1998, SMS proposed a set of actions, under the title of Acolhimento Soliddrio (roughly translated as "friendly welcome"), to reorganize the work process (and re-work the culture of the staff) inorder to make access to services an easier and to dispose of cases more effe~tively.4~The intention was to shift to a more quality-oriented approach across the broad range of activities of the Health Units. Protocols were developed to prioritize cases as clients entered the Unit in a way that minimized queuing. Unit work 41Because there is prior consultation, the Council has never failed to approve a budget. The consultative process is now said to operate smoothly. But to some extent, the Municipal and Local Councils have become vehicles for party politics. Our survey found that, at the more hands-on level of management, especially performance contracting, the role of the Local Health Councils was more modest, though it did provide information about the community and identify problems. 42See footnote 38 above. 43 The aims, instruments and results of Acolhimento Soliddrio are described on the SMS website (httD:Nwww.curitiba.pr.Pov.br/saude/). 44 processes - in particular the nursing function - were then reorganized and staff intensively trained to apply the new procedures and to adopt a more client-oriented culture. The new practices were not introduced without resistance, but the SMS reports that changes in behavior and processes have reduced queues, improved the rate at which problems are resolved, andreduced complaintsto the CAUe4 By emphasizing service quality and a participative or consultative approach, these instruments of client-orientation also have the political function of "selling" their service directly to voters. ManagingInformation Curitiba's approach to primary healthcare has required two principal forms of knowledge: social knowledge about clients and the environment they live in (c.f. the social theory of medicine), and medical knowledge of epidemiology (which is often relatedto social knowledge) and medical and clinical "best practice" elsewhere in the world. The municipality has developed an increasingly sophisticated information system to manage social knowledge and some medical knowledge, as well as to manage its healthcare system. Social knowledge. The expansion of primary healthcare services has been planned on the basis of area-by-area socio-economic analysis. Central to the way Health Units function is the emphasis on better knowledge of individuals and families. Community health agents play an important role inthis. Medical knowledge. Given world advances in understanding of health problems and treatment technologies and the permanent experiment in effectively organizing public health services, the local collection and application of medical and clinical knowledge is a constant endeavor for any health authority. This effort occurs through systematized epidemiological activities, programs linked to universities and training institutes, and so on, such as the intensive training relationship with the University of Toronto. More recently, a particular thrust in Curitiba has been the development of integrated protocols. These are manuals for clinical procedures in various areas (mental health, dental health, physiotherapy, hypertension, adolescence, and so on) which also serve the purpose of helping integrate primary healthcare (the entry point for clients) with specialized medicine and hospital services. These protocols are part of the evidence-based-medicine (EBM) approach that Curitiba has adopted. In addition to reflecting best health practice, protocols may have economic effects: by standardizing clinical practice, they permit (in theory at least) greater economies of scale and make it easier for managersto measure performance (see Box 3.1). The information system. Curitiba has a very advanced information system (under construction since 1988), built on a number of computerizeddata bases that communicate with each other. At the heart of the system is the Prontudrio Eletrdnico, a completely computerized patient medical- records system, with medical, dental, and nursing modules, established in 1999 (see Cazura Xavier and Shimazaki, n.d.). Each person registered in the system has a magnetic card (Cartlio Qualidude Salide) which allows them access to their records and to health services. The Prontudrio is an instrument in clinical management because it is compatible with the Integrated Protocols. Complaints about Health Units have to be reported to the Local Health Council. See Ducci et al. (2001:109-115>for adescription of the processby which complaints are followed up. 45 The Prontua`rio is in turn integrated with other IT systems (some of which provide an interface with private providers), including the Central de Regulqo (which allocates hospital beds for specialized-medicine and hospital resources financed by the federal government), Central de Procedimentos (which sets appointments for consultations, examinations, and therapies), Laboratorio Municipal, Pharmacy, Centro de Epidemologia (where disease information is obligatorily notified, then transmitted to federal Ministry of Health systems), Sistema de Znfomzqrlb Gerencial (to monitor performance), and various municipal systems outside of SMS (citizen identification, human resources, etc.). Box 3.1: The Role of Protocols The SMS sees the advantages of protocols as standardizing medical information and providing consensus among teams. The adoption of protocols was negotiated between technical staff, directors and scientific societies. Inthe beginning, the doctors were resistant but after the entry of scientific societies in the discussion, they changed their position. Indeed, they came to see protocols, with their support from scientific societies, as a protection against claims of medical malpractice. Protocols also improve the procurement process since only a pre-determined set of medicines is purchased for the entire network. This permits economies of scale. The SMS guarantees the availability of drugs in the Health Units' pharmacies and guide doctors to prescribe them, given scientific evidence. The SMS also believes that protocols save public funds that could be lost in lawsuits against the state. Another advantage is that protocols are part of the SMS management information system. A doctor who attends a pregnant woman, for instance, knows all the exams the patient has already undergone (Le. pre-natal exams, sonograms) and what others she still needs. The protocol's indicators are monitored and will later feed into the reports about the accomplishment of POA's targets. Sometimes doctors do not fill inthe information properly or do not follow the agreed procedures. To check on this, designated doctors (who are allowed access to confidential information) monitor the use of protocols in a sample of Health Units.The intention is for each Health District to have one doctor specialized inmonitoring protocols. According to those working in the sector, the information system has had clear health and managerial outcomes (Cazura Xavier and Shimazaki, n.d.): It has systematized nursing practice and led to changes in the way dentistry is done (and requiredthe development of dental health protocol training). B y facilitating the monitoring of individuals, families, and populations at risk, the system has enabled coordination of health actions; epidemiological information i s more timely; laboratory results are returnedquicker. The system has integrated different services: clients find that their appointments are set more efficiently (and that they have greater choices); there is better management of hospital bed allocations. 46 Inaddition, Curitiba's system of performance management, emphasizing performance contracting and teamwork is greatly facilitated by accurate measurement of more standardized processes.45 But, as the managers of this system recognize, implementation has posed, and continues to pose, management challenges. Specializationand Coordination A complex system (doing many different things to produce a service or range of services) must create an internal specialization of tasks so that each thing can be done well. But the system must also design a way of making all these separate parts work together. Vertical Specialization. The municipality is the dominant producer of primary and preventive health services, many of which are considered public goods, more appropriate for government p r o d ~ c t i o n . ~The municipality also finances and regulates (rather than produces) much ~ secondary and most tertiary health services, which are largely private goods. The shift to preventive and primary medicine is also the central feature of Brazil's (and Curitiba's) strategy to control the costs of public health. Horizontal specialization has been promoted in two ways: concentration on specific problems and diseasesby means of Programs and programmedactivities (Aten@ progrurnudu) for groups of patients (diabetics, for instance); and the standardization of processes, both administrative (occurring in conjunction with computerization) and clinical (as a result of the Protocols). Specialization should improve efficiency through greater economies of scale.47 Standardization also facilitates managers' ability to monitor performance. Coordination of services: the process of coordinating the many aspects of a complex system of healthcare provision can be said to have begun in 1994 with the establishment of the Central de Murcu@o de ConsultusEspeciulizudus, and to have culminated in 2002 in an integrated system of services (SisternuZntegrudu de Servips de Sulide), underpinned by communicating computer systems and by integrated Protocols. The primary-healthcare network i s the gatekeeper for the entire public health system. B y controlling the access of patients to other services and tightly coordinating the different levels of services, the SMS seeks to provide appropriate and timely health services to citizens and to rationalizethe use of public resources. 45 Our survey respondentsoverwhelmingly (93 percent) acknowledged the importance of the information system in identifying problems. They saw its greater impact in epidemiological knowledge, faster service, team productivity, and reaching contractual targets. 46 Public goods are generally available to all consumers at the same time, and one person's consumption does not reducethe supply available for anotherconsumer. 47The Aravind EyeHospital inIndia is a well-known exampleof economies of scale through specialization - in this case in cataract surgery (See Rangan 1994). With less than one percent of the country's ophthalmic manpower, Aravind performs about five percent of all cataract surgeries in India." (httD:Nwww.aravind.ordabout/index.htm). 47 Managing HumanResources The SMS has a workforce ofjust over 6,000 (Table 3.2): 0 There are almost 5,000 public servants - three times the number of the early 1990s. These are permanent career employees. 40 percent of them are professionals (Superior); 44 percent are at the Intermediate level; and 16percent at the Basic 0 Just over 200 of the career professionals are managers who occupy, for the duration of the appointment, a politically-appointedpost. 0 Almost 1,200 people are employed under private labor law (CLT) as Community Health Agents. Careers of public servants inhealthcare.The structure and rules for public servants in Curitiba are constitutionally mandated and resemble those for other Brazilianjurisdictions. Entry into the career is based on academic qualifications and a written public examination (concu~~o pliblico) and the top candidates assume an Office (cargo) defined by a narrow occupational group: doctors, dentists, and nurses are offices at the Superior level, and auxiliary nurses at the Intermediate level. After a three-year probationary period, they receive lifetime tenure.49 New entrants start at the bottom of a career and regularly and automatically advance within and between grades. There is no possibility of lateralmovement outside the given Office.'' Healthjobs are professions in themselves: people typically become doctors, dentists, and nurses for life. In the public sector, the most obvious career-development path typically available is to gain, temporarily or permanently, a management post (a political appointment). Withinthe SMS, there is one career move that has aspects of career development: the move from a Basic Unit to a PSF Unit (via an internal public examination) results in more money and a different (more team- and community-oriented) job. There is a heavy demand for doctors inBasic Unitsto pass to PSF Units. For most health professionals, satisfaction comes from the job itself, not advancement prospects. In the SMS, part of this satisfaction also comes from the frequent opportunities for training and professional development. (We come back to job satisfaction in a discussion below on performance culture.) Most health professionals have the choice to work in the public or private sectors. For doctors and nurses in particular, the jobs in each sector may be quite different because of the public sector's emphasis on primary healthcare and preventive and family medicine. Inpractice, nurses, auxiliary nurses, and dentists are in adequate supply, and the public sector pays them at rates that are competitive with the private sector. Some of these professionals typically have made a career choice basedon their preference for what the public sector does in healthcare. As a result, there has been a reasonable stability of public sector healthcare employees, which has probably helped foster teamwork and increaseexperience inHealth Units. Doctors are different. First, the profession exercises an effective control on entry into medical schools. This has raised the price of doctors and lowered their numbers on the Brazilian labor 48Three-quartersof those at the Intermediate level are auxiliary nurses. They havejust beenupgraded (with a corresponding salary rise) from the Basic level. To consolidate this change, auxiliary nurses must now undergo aprocessof training and certification. 49Following a 1998 constitutional amendment, public servants can in theory lose their jobs for poor performance; inpractice, this is hardly known. 50Some auxiliary nurses hope to move up to the higher Career of nurse, for instance by going to night school, then taking the public examination. 48 market. Second - and notwithstanding the Brazilian healthcare reform movement - the preference of most doctors (reflected in the availability of university training courses) is to specialize rather than practice general, or family medicine. However, the federal PSF has since 1996 developed a new labor market for generalist doctors: doctors who take up posts in PSF Units have a greater tendency to stay in the public sector than doctors in Basic Units. And in recent years, federal funds have been used to force the universities to begin to adapt medical and nursingcurricula. The scarcity of doctors has implications for the turnover of doctors and the way they are employed. 30 percent of doctors appointed to the SMS leave in the first year, usually for the private sector. Curitiba's doctors (like doctors elsewhere in Brazil and in many poorer countries) hold multiple jobs - an average of two to three jobs (and up to five), each formally requiring 20 hours of work a week. Often, this multiple-job strategy seeks to combine the benefits of higher salaries in the private sector (in clinics and hospitals) with the higher prestige, greater learning opportunities, and non-salary benefits (particularly pensions) of the public sector. The doctor's work week in a Basic Unit is 20 hours; ina PSF Unit it is 40 hours, a reflection inpart of a bonus scheme that is discussed be10w.~' Nonetheless, doctors seem to be the weakest link in Health Unitteams becauseof turnover andmultiple-job holding. Bonus schemes. The SMS has had several systems for bonuses on top of basic salaries. There have been three systems whose purpose has simply been to raise the level of remuneration in an area where the labor market was supplying too few qualified people. From the beginning of the PSF, all public servants in PSF Units have been eligible for a bonus funded by the federal government. Subject to receiving a satisfactory annual evaluation: doctors and dentists receive 80 percent of their basic salary, nurses 60 percent, and auxiliary nurses 50 percent. Repeated absence and lateness disqualify staff from their bonus. A less-than-satisfactory evaluation two years in a row would lead to staff being movedelsewhere in the SMS. This happensto almost no staff. Thus, the PSF bonus acts as a general incentive to attract staff to the PSF Units, more than to stimulate their performance while they are there (except inextreme cases).52 The first scheme independently introduced by SMS was the IDG, a bonus introduced in 1994 to encourage staff to work in dangerous and less accessible areas of the municipality. Health Units were rated in three categories according to the level of difficulty faced by personnel. This systemprovedto be controversialbecauseit causeddisputes with HealthUnitsthat were rated in the least difficult category. The IDG was soon abandoned (and now the factors that gave rise to it have more or less gone away). In2002, SMS introduceda Special Bonusfor Doctors (Gratificap7oEspecialpara Mkdicos, GEM)equal to 20 to 40 percent of the basic salary. Since 2002 the turnover of doctors has fallen, and it is thought that this incentive is largely responsible. 51Health workers failing to complete their prescribedhours does not appear to be as commoninCuritiba as inotherjurisdictions. 52The PSF Units are staffed through an internal competitive process (concurso plus interview) open to qualified SMS staff. This system was introducedto provide for a legitimate appointment procedure that helps to counteract any disincentive effect that Basic Unit staff feel as a result of getting lower pay for similar work. Our survey suggestedthat, surprisingly perhaps, there was little sense of a differenceineffort andperformancebetweenPSF-Unitand Basic-Unitstaff. 49 The SMS has also experimented with two schemes to encourage better performance. The first of these was launched in 1995, when the SMS felt the need to create a performance incentive to support the drive for greater quality in health services. The Quality Incentive Plan (PIQ) established a bonus for all SMS Units, with the exception of PSF units. Based on performance according to a small set of indicators (health outcomes and health and management outputs), the staff in around 10 percent of the units were rewarded with a 30 percent bonus. This scheme turned out to be even more problematic than IDG. The focus of the teams became one of simple compliance with a checklist. Moreover, senior managers in the SMS perceived that the team- based nature of the incentive provoked rivalry among teams and undermined the unity of the system as a whole. The PIQ was terminated in 1997. The current scheme, the Incentive Program for Quality Development (IDQ), was created in 2000, at a point where there was a rapid expansion of PSF Units. Each quarter, SMS workers are evaluated to see whether they qualify for a 20 percent bonus of their monthly base salary (in the case of a few Unitsthe bonus goes as highas 40 percent). All SMS employees are eligible except some political appointees (cargos em comisslio). The IDQ evaluation is based on four elements: the supervisor's evaluation of the employee, self-evaluation, SMS evaluation of the Unit, and community evaluation of the Unit (see Box 3.2 for more detail^).'^ Box 3.2: The IDQEvaluation System The IDQ is based infour elements of evaluation. Each element receives a score of 0-100which is monthly bonus, but there are specific "excluding factors" - repeated absence, penalties received, weighted into a final score of 0-100. An employee with a score above 80 points will receive the abnormal leave periods, or late arrival - which disqualify the employee from receiving the bonus. The four elements (and the weighting) are as follows: Individual Evaluation of the employee (weighting 45%). The immediate superior makes a preliminary evaluation halfway through the quarter, then again at the end of the quarter. The criteria are: knowledge, professional attitude, interpersonal skills, work quality and compliance with administrative rules. This evaluation provides a feedback to employees about their performance and indicates specific actions to enhance quality and productivity. A minimumof 80 points is necessaryto be eligible for the bonus. Self Evaluation (weighting 5%). At the end of the quarter, each employee assesses hidher own performance, behavior and productivity in relation to the expected outcomes, using the same criteria. Health Unit Evaluation (weighting 35%). This evaluation, at the end of every quarter, assess whether the performance targets set by the Health Unit were achieved. The targets analyzed are performance indicators from the management contract (Temzode compromisso). Community Evaluation (weighting 15%). The SMS uses quarterly sample telephone surveys (Avaliap?es da Comunidade) to ask about the quality of treatment and services. The Units with scores above 70 percent ("excellent'' and "good") are eligible to participate in the IDQ program. Over 90 percent of employees win the bonus each quarter (Table 3.7). Most of the others are disqualified because of "excluding factors" (leave, absence, lateness, etc.). Prolonged absence 53In a separate scheme, Community Health Agents (ACS) can also earn a 20 percent bonus for having reached the work targets set for them. 50 (ufustumento) is the overwhelming cause. The number of eligible employees excluded by low performance is very small: around 20-30 people in the first two quarters of 2004, none in the subsequent three quarters, rising to more than 100 in the last three quarters of 2005. SMS reported that low performers are known and usually are the same people. Their managers try to assign these employees to other jobs, to which they might be better suited. The SMS is also reported to be raising standards of performance over time (this is consistent with the rising number of poor-performance cases in 2005).54 Table 3.7: Performance under the IDQ,2002-2005 ~~~ ~ Bonusnot received,by categoryof disqualification: Poor Bonus perfor Late- Year Trim Participants received mance Leave Absence Penalty ness Total 2002 I n 1,584 93.3% 6.7% I11 1,592 92.7% 7.3% Iv 2003 I 3,002 94.4% 5.6% I1 3,080 95.5% 4.5% 111 3,076 90.7% 9.3% Iv 3,400 91.7% 8.3% 2004 I 3,700 93.5% 0.5% 4.1% 1.2% 0.1% 0.7% 6.6% 11 4,582 90.8% 0.6% 6.5% 0.7% 0.3% 1.2% 9.2% 111 4,842 90.5% 0.0% 6.8% 0.8% 0.3% 1.7% 9.6% Iv 4,835 91.3% 0.0% 6.4% 0.8% 0.3% 1.5% 8.9% 2005 I 4,865 90.8% 0.0% 6.3% 0.8% 0.0% 2.1% 9.2% I1 4,738 93.6% I 111 4,9 18 87.8% 2.6% 7.2% 0.5% 0.2% 1.7% % 10.4 Iv 4,864 89.6% I 2.8% 5.3% 0.4% 0.1% 1.8% % Source: IMAP(2002-2003) and SMS (2004-2005) Those disqualified from the bonus for poor performance account for a very low share of total participants (less than 1percent in 2004, but risingabove 2 percent in2005). Where so few cases are caught, it i s reasonable to suppose that the scheme is only catching very extreme cases of poor performance and that it is not actingto stimulate better performance ingeneral (inthe way a sales commission does, for example). However, the results of our survey seem, at first sight at least, to contradict this. Ninety percent of respondents thought that IDQ had an influence on behavior when it was introduced. Consistent with the patternof qualification and disqualification inTable 3.6, this influence was considered greatest in the area of workplace behavior (discipline, punctuality, and so on). But it was also perceived to have improved motivation and, to a lesser extent, work organization and productivity. The survey also suggests that respondents feel that the impact has fallen over time: only 61 percent believed that IDQ at present continues to provide an important influence on behavior. 54 Itis often observed that evaluators in Brazil (as elsewhere in Latin America) are reluctant to give someone a low performance evaluation. This may be because the prevailing culture value equal treatment more than rewarding merit; or it may reflect a fear that loyalty and friendship are more likely to be rewarded than merit. 51 Sometimes, management innovations tend to have a short-lived effect: they enhance performance initially when they are still a novelty and provide a spotlight on workers, but as they lose their novelty, workers adjust their effort back to the norm: the IDQ may be one such innovation. But according to our survey, the IDQ may also have had a more valuable side effect that, hopefully, will last longer. It seems that the IDQ process may act as a motivator through the evaluation process itself. This process is perceived to be fair and coherent and to contribute most by promoting ernployee-supervisor dialogue (rather than by direct reward/punishment, except punishment on account of "excluding factors"). For instance, it is perceived to help solve team conflicts. Collectively passing the IDQ test is a major preoccupation of teams (72 percent say this is "always" so and 25 percent "frequently" so). When a problem is perceived with a staff member, the team acts to solve it. Thus, it may be that the enthusiasm (if waning) that staff show for the bonus as a management tool relates to evaluation as a tool for identifying problems, rather than the bonus as a performance-enhancing reward system. Managers. All managers inthe SMS (around 215 posts) are political appointees (Le., they serve at the pleasure of elected official^).^^ By tradition, these managers, including the Secretary, are also public servants from the SMS. (They tend to come more from the ranks of nurses and dentists than doctors.) When they no longer have the political appointment, they are free to resume their normal career (office) within the Secretariat. Managers are appointed at four levels: at the top (the Municipal Secretary of Health); then Directors; then District Health Supervisors and Coordinators; and finally, Health Unit Managers (ASLs). People occupying these positions get small increments to their salaries as public servants. The structure of management is very flexible: there are only four levels in the hierarchy and managers can rise up from the ranks of professionals, and then return to being a professional. The tradition of appointing all managers from within the SMS also means that managers, though politically appointed, share the same culture as their non-manager colleagues. This reflects the limitedinfluence of party politics onhealthcare inCuritiba. ManagingPerformance Strategic planning and systems thinking have been a central habit of mind among health managers since a Health Department was created in 1979. Upto the early 1990sthe main aim of planning was to define and implement the primary healthcare model and to develop appropriate human capital. From the beginning of the 1990s to around 2004, a central theme was to make performance central to management. From the mid-1990s to the early 2000s, another central target was the physical expansion of primary healthcare capacity. Strategic planning has been practiced both informally and formally. Among the foremost examples of formal strategic planning was GERUS, a management course for leaders of Health Units undertaken in 1995 with the aim of setting a track for reorganizing basic health~are.~~After a year-and-a-half of reflection, GERUS generated a number of proposals that substantially shaped the managerial innovations of the 1990s, including the development of the information system and of performance targeting. Many aspects of Curitiba's healthcare-management system are informed by systems thinking: for instance, information systems have consciously changed management 55There are also over 300 nurses in the Health Units (Table 3.2) who, though they are not formally managers, frequently play a leading role inteams (doctors generally do not). 56GERUS was a national program, supported by the Pan-American Health Organization, and applied locally by Curitiba (see Bertussi, et al. 1996). 52 methods (and vice versa); the introduction of integrated protocols has similarly contributed to changes inmanagement methods. Strategic planning has also become a more formal, routinized activity which has paved the way for formal performance management. Health planning in Curitiba is mandated by federal legislation (Law 8080). Curitiba has a four-year plan which is a product of the MunicipalHealth Council meeting in its Conference (the Curitiba Municipal Health Plan 2002-2005). The plan sets out desired outputs and outcomes: how health services are to be delivered, how services are to be managed, and how major medical problems are to be tackled. The plan concentrates far more on primary health than tertiary. It is detailed in indicating a number of objectives for health-system outputs and outcomes, but these objectives tend to be fairly general and often unmeasurable (for instance: "Population satisfaction with health system" or "Consolidate the System of Food and Nutritional Control). The plan does not quantify these objectives (many of which are difficult to measure), nor does it describe the steps that need taking or estimate the money needed to attain them. It is the fruit of strategic thinking. For instance, a principal concern is to chart the reorganization of primary healthcare according to the principles of the IntegratedSystem of Health Services (SISS). But because it does not really chart how objectives are to be attained, it is something less than a strategic plan.57 The Municipal Health Plan (four-years) is translated into more specific actions in the Annual Operating Plan (POA), which uses the same format to set targets for the coming year. In turn, a number of quantitative indicators have been developed since 2000 to help, where quantification i s possible, to make the strategic POA targets more operational. Some 60 indicators are enshrined inannual Management Contracts (Temzo de Compromissode Gestrio)between the SMS and the Health Districts, and between the latter and the Health Units. The Management Contracts appear to provide an effective means of reinforcing the performance orientation of the staff in primary healthcare. This is quite a rarity inBrazil and in many other countries. Management contracting. The Management Contracts are simple: they summarize the general obligations of principal and agent towards each other and set specific performance indicator targets. For the contracts between SMS and the nine Health Districts there were 55 indicators in 2005 and for the contracts between Health Districts and Health Unitsthere were 63. No rewards or punishments are directly associated with performance under the contract^.^^ However, 35 percent of the points awarded individuals under the IDQevaluation system depend on the Unit's performance under the Management Contract. We discuss this further below. Targets are agreed in discussions between the SMS and the Health Districts and between the latter and the Health Units. Some of the indicators are agreed upon with the Ministry of Health, regardingtargets of coverage, visits and procedures. Target-setting takes into account differences between areas (for instance, the more middle-class the neighborhood, the more inhabitants are likely to use private health services) and differences of installed capacity. According to our interviews and survey results, targets are fixed between Health District and Health Unit with a mix of negotiation and imposition. But the Health District is open to re-negotiation when there 57Strategic planning is basedon a vision of how the world is likely to change inthe medium- to long-term, andaims to shape an organization's future by defining its goals andhow it will achieve them. It is a process of continuous learning. 58 Thus, they have the characteristics of a relational contract, a quasi-contractual agreement getting its strengthnotfrom formal sanctions, but from the shared needs of principal and agent to do businesstogether over the longer term. Unlike some relational contracts, the SMS Management Contracts are, through the agreed targets, highly specific about some aspects of expectedperformance. 53 are difficulties in fulfillment. This would suggest that there is a reasonable balance between the "top-down" direction and "bottom-up" participation that many consider necessary for good management. The effectiveness of the contracting system depends on the quality of the indicators, so that participants have less opportunity to "game" the system.59 The construction of an advanced information system from 1999 (the Prontua`rio Eletrdnico) led to faster and more accurate information, while the Protocols developed under the IntegratedSystem of Health Services (SISS, 2000) led to more standardized definitions. Every month the management information system transmits performance data from the Health Unitsto their Health Districts. Every quarter the data are sent from the Districts to the SMS, which carries out a comparative analysis and reports to the City Council and the Municipal Health Council on the progress of the POA. The number of indicators has grown over time and the SMS continues to seek improvements inthe indicators (in the case of tuberculosis, for instance).60 The impact of management contracting. The most important value of management contracts appears to lie in the way they provide signals to individuals and teams about performance expectations, identify problems they need to solve, and promote teamwork. (See the comments from one Health District in Box 3.3.) The survey found a (surprisingly) high level of enthusiasm for management contracts: an overwhelming number of respondents found the exercise useful (and not excessively costly to them). 0 Expectations: the process of negotiating contractual targets helps fix staff expectations: 85 percent of survey respondents found the contracts substantially or entirely useful in establishing priorities. 0 Problem-solving: the process of monitoring and discussing performance under the contracts feeds back into the organization of work: it is an instrument of strategic planning at the local level. The Districts regularly consult with the Unitson their performance. The SMS, in turn, meets with the Districts. These meetings examine indicators, identify poorly-performing Health Units, and identify problems. Following these meetings, the SMS and the District provide help to the Health Units through training and exchange of ideas. Eighty-one percent of survey respondents found the contracts substantially or entirely useful in identifying problems, and 72 percent found the contracts useful in solving problems. The survey also found that management contracting helped create links with a broad variety of other municipal-government services; this is, presumably, a particular instance of the problem- solving function. 59Under Britain's National Health System, one of the indicators to assess hospital performance is the amount of time patients admitted to hospital wait before they get to a ward. One imaginative hospital got aroundthis indicator by declaring that the corridor where the patients were waiting was a ward. 6oAppendix Fprovides some illustrationof the indicators with target and performancedata for eight of the indicators relevant to contracts between the SMS and the Health Districts. Those indicators with area populations intheir denominator are subject to the distorting factor that many non-Curitibans use the city's public health facilities. There is no weighting of the indicators, or provision of key indicators, but it appears that some indicators, such as those relating to safe pregnancies and infant deaths, are more closely watched than others. For all the targets, including the eight inAppendix 3, there is a wide range of over- and under-fulfillment. This may be because it is so difficult to set accurate indicators, but it also clearly suggeststhat monitoringperformance under targets is not apro-forma activity. 54 Teamwork: management contracting is part of a broader set of management instruments involving staff evaluation and incentives (IDQ) and Community Evaluation. Taken together these instruments support a teamwork approach, both in everyday operations and in problem- solving. Teams are well aware of Community Evaluations and make use of them. This chapter has argued that the IDQ probably does not function as a performance incentive in the classic manner; but the survey emphasizes that it is a vehicle for staff evaluation and for the perception - 92 percent of respondents, shared by staff of PSF and Basic Units alike - that identification and solution of problems that relate to team performance. There was a strong teams worked well. According to the survey, ensuring that staff gain the IDQ bonus is a major preoccupation of teams: 50 percent of an individual's evaluation depends on team performance (35 percent on performance under the management contract and 15 percent on the Community Evaluation). Teams, when they perceive a problem, will, most often, talk to an under-performingcolleague or else adjust work processes. e Competition among Health Units: Health Unit managers are reasonably aware of how other Units are performing. But the survey suggestedthat staff generally hesitate to openly judge the performance of other Health Units (and, if they judge them, hesitate to find them better or worse). It can be inferred from this that there i s no strong sense of rivalry between Health Units and that trying to foster inter-unit competition as a motivation for better performance may be misplaced within the particular culture of Curitiba's public administration.61 Box 3.3: Management Contracts One Health District's View - A meeting with the supervising staff in one health District provides a view of how management contracts contribute to performance. Their main impact i s to provide accurate information, establish sharedobjectives, and underpin trust and problem-solving. The system could not work without the quality of information provided by today's information system. One of the things that defeated the PIQ bonus scheme of 1995 was that teams could cheat becausethe quality of information was poor. Infact, when the new information system was established in2000, levels of measuredperformance inthis District fell. The level of trust established between District and Unit is high. District supervisors have all worked in Units and may go back to working in them. When a problem arises in a Unit, trust paves the way to cooperation, not defensiveness. Various approaches to solving work-process problems have been developed: visits to units (especially working with nurses), group discussions among nurses, and so on. When one unittold the District that it could not achieve a particular target related to the number of cases of hypertension registered, the District suggested novel ways to do detective work to find the clients or, if necessary, to explain the shortfall. A performance culture. Performance management is associated with a thriving performance culture among health workers in Curitiba's public sector. Staff behave as team players and 61A similar argument is often made with respect to Brazilian public servants' cultural aversion to being evaluated and compared with others (e.g., Vaitsman 2001). However, as a generalization, such a cultural argumentshouldbe received with skepticism, as blanket statements are often made inthe nameof culture. 55 problem solvers becausethey have intrinsic motivation in their belief in the values and objectives of the SMS. Extrinsic motivation - "Ido it because I'm materially rewarded" - i s always important, but it does not appear to dominate in Curitiba. According to the survey, respondents reported that they were motivated by extrinsic factors (job stability and nearness to home), as well as intrinsic factors (job content, learning opportunities, training opportunities). They reported that salary levels and flexible hours were less important. This sense of professional motivation did not differ much between PSF and Basic Units. The general enthusiasm of professional SMS staff for their work comes through in interviews. It is also expressed in the astounding amount of effort staff voluntarily put into preparing professional pa ers for the biennial meetings on public health that SMS organizes for primary-healthcare staff! Port0 (2001, page 5) provides some quotes from staff working on the Programa Mc7e Curitibana that capture the culture: "I love my work. It's engrossing (nurse's aide)"; "If someone says that they come to the PSF for the love of it, it isn't true. Butafter they see what is going on here, they get to love it (health agent)." It is by no meansclear that SMS management practices alone created this thriving performance culture. However, it is clear that a performance culture is inthe air of Curitiba. We discuss this inthe next section. THEELEMENTSGOODPERFORMANCE OF We have postulated in this chapter that Curitiba has a primary healthcare system that has performed well, and we have sought to explain the ways in which the system is well-managed. In this final section, we synthesize the main characteristics of this performance, discuss some of the factors that may help explain how this good performance came about, and try to draw out a few implications from this case. We suggest five salient characteristics of good performance in Curitiba. First, Curitiba adopted innovative health policies that have worked. These policies, many of which came from the SUS, in turn dictated much of how the health system had to be organized: decentralized, client- oriented, using standardized procedures. These features (by accident?) turn out to be compatible with performance management. Second, the managers of the SMS understand how to manage complexity. They have been able to design and manage complex, interacting systems and to use information technology to help coordinate them. (The table of Appendix E tries to capture some of the ways in which different systems, whether intended as instruments of healthcare or management, interact.) A sophisticated information system appears to be an important condition for coordinating different systems. It i s of particular interest here because reliable information underpins the successful management-contracting system. Third, and addressing the concerns that originally prompted this chapter, good performance is driven by a strong performance culture. This culture is characterized by a shared belief in Curitiba's primary-healthcare mission, an aptitude for teamwork, and an orientation to uncover and solve problems. This chapter has touched on some of the elements of this culture, but cannot explain it completely. Secretaria Municipal de Salide de Curitiba (2004) lists 240 resumCs o f professional presentations for the 2004 meeting. 56 0 Formal planning and management contracting did not create the culture, but they support it because they are an added tool to fix expectations, solve problems, and foster teamwork. Indeed, Curitiba's system of management contracting provides a good example of relational contracting at work. 0 Bonus systems in Curitiba have not worked, in a direct way at least, to promote performance. One of them, PIQ, sought to promote competition between teams, but this competition proved destructive and the scheme was soon abandoned. The current scheme, IDQ, deals with egregious performance problems (lateness, etc.), but we do not think it has directly affected the performance of most workers (even though our survey respondents thought it did). However, it has had important indirect effects: staff evaluations under the IDQ have reinforced the performance effects of management contracting; it is a preoccupation of teams to solve individuals' performance problems so that team objectives (performance under the management contract and a satisfactory evaluation by the community) are met. 0 Thus planning, management contracts, and the IDQ (through its staff-evaluation function) turn out to be interesting instruments that support a performance culture, but they do not explain it. The explanation, we surmise, has to lie inenvironmental and historical elements which we discussbelow. Fourth, some of the reforms we have mentionedhelpedto mitigate the principal-agentproblem that characterizes the health sector: inasmuch as professional health workers have specialized knowledge, it is difficult for their managers (principals) to control what they do (as agents), just as it is difficult for their patients (as principals) to control what professional health workers do (as agents). Curitiba modified the principal-agent relationship through reforms that standardized the product (medical knowledge, clinical procedures), improved the information system to capture this standardized information, and used this standardized information to clarify the performance expectations that managers had of health workers. These reforms somewhat centralized discretion over medical decisions, while other reforms partly decentralized implementation decisions. Municipal and Local Health Councils, the telephone complaints and information system, and community evaluations (telephone surveys) have also helped to reduce the asymmetry of information between health workers and patients. Finally, one might speculate that a process of uninterrupted growth and change has been important in helping to drive this culture. Today's system is the result of a consistent line of development since 1979; this has required constant innovation in the medicalklinical and managerial fields, as well as capacity expansion (in response to the city's growth as a result of immigration and of the municipalization of health), Permanent change has meant a successionof challenges and required a permanent attention to problem-solving. Permanent change has also meant a continuing increase inthe size of the workforce. This has provided greater opportunities for career development and may also have helped minimize the negative impact of tenure rules that protect under-performing staff. The Curitiba case raises two related questions. First, where did its performance culture come from? Second, what allowed this culture to thrive under a civil service regime that is not ostensibly oriented to performance? At this stage we can only offer a mix of analysis and hypotheses (in the hope that these hypotheses might fuel subsequent work on the policy reform process inhealth). We suggest three separate sources of influence on performance: 0 The Brazilian government, through the SUS, provided a workable policy (a set of operational rules) and a guaranteed level of health financing. 57 0 The sanituristus constituted a professional network that, we hypothesize, provided expertise (a set of working beliefs and practices consistent with SUS) and contributed to a strong work ethos. 0 We also hypothesize that the Curitiba city government provided political continuity and the tool of strategicmanagementfor the health sector. We discuss each of the three influences inturn. The Environment for Performance I:SUS In the 199Os, the SUS, implemented by the Ministry of Health, earmarked and decentralized healthbudgets on the one hand, and provided a policy model for primary healthcare reform on the other. Some of the features of basic healthcare management that can be observed across many municipalities reflect health and management philosophies which initially stemmed from the ideas of the Brazilian healthcare reform movement and were articulated and refined by the federal government under the SUS. Common to this health philosophy are: humanized and socially- oriented medicine, a focus on local problems, and preventive medicine. Common to the management philosophy are: the move to highly-structuredteams usingstandard procedures (PSF and Community Health Agents); decentralization within cities; Health Units as gatekeepers to more complex curative services; clearing houses for appointments and hospital beds; formal health planning; institutions of social control (including municipal health conferences); and telephone hotlines. The influence of the SUS on Curitiba's primary healthcare system is clear. The Environment for Performance 11:Health Professionals' Networks Curitiba's healthcare sector has been affected by two overlapping sets of networks (people or groups that associate to pursue a collective objective). The first is composed of the various medical professions; the second is the sanituristus, a group of primary-healthcare workers in the public sector which grew out of the healthcare reform movement. Doctors and other healthcare workers are among the most powerful professions in Brazil, as in other countries. Their specialized knowledge and their organization have given them a presence in Brazilian government that is probably stronger and more cohesive than any other professional group. This is a major reason why health spending is constitutionally guaranteed. Like many professional groups, the objective of healthcare professional groups is to serve their members' interests, in particular by maintaining a self-regulating monopoly. Given the asymmetry of information between health professionals on the one hand and their political masters and the public on the other, controlling entrants to the profession and maintaining a set of behavioral standards are meant to assure the public that the professional is acting in public interest. The process of standardization that has been occurring through the PSF (and even more markedly in Curitiba) threatens to weaken this monopoly by increasing the information available to politicians and the public. It is therefore perhaps no coincidence that doctors, the most highly organized of the healthcare professionals, have been less enthusiastic than other healthcare professionals about primary-healthcare reform. The sanituristas are not a formal group, but bring together, more or less, the primary-healthcare workers of the public sector (and cut across the healthcare professions). Apart from a common interest in securing more resources for the sector, the objectives and impact of this network have not been the same as those of the professions, particularly doctors. Initially the network had the characteristics of a policy community (a network involved in a particular area of public policy). 58 Born in the 1960s and 1970s, before Brazil's return to democracy, the healthcare reform movement adopted a very political view which linked health to social conditions and to democracy, and took on the responsibility for political action to pursue its beliefs.63 It was this movement - inthe absence of an articulated popular demand for reform - that dictated the health clauses of the 1988 Constitution. Later, this led to the SUS and a well-defined model for primary- healthcare reform. The movement's members sought positions in government from which they could protect and implement the new policies. With this policy objective largely achieved, the network of suniturisrus that grew out of the healthcare reform movement also became a community of practice (a network of people who collaborate to share ideas about, and find solutions to, common problems). The sanirarisras circulate within a large, mobile labor market for workers in public health at all levels of government. There are substantial interchanges of people and information through this job market, through formal organizations - such as the Association of Family Health and the National Graduates in Collective Health - through associated Congresses and other meetings, and through Council of Municipal Health Secretaries (CONASEMS) and the Brazilian Association of academic faculties that teach inthe area of primary healthcare. The network of sunituristus brings to the SMS in Curitiba a group of workers with a common experience, a common understanding of the problems and of the tools to solve them, and the bonds of belonging to the same network. It i s particularly important that most of today's senior managers were all products of this network and lived through Curitiba's healthcare reforms. It is reasonable to infer that this movement has been most helpful in imparting specialized knowledge to Curitiba's healthcare workers, providing a strong working ethos that motivates performance and the sharing of knowledge, and thereby facilitating their work in teams. (See Box 3.4 for a more theoretical argument about how networks mightaffect performance incentives.) The Environment for Performance 111: City Management Curitiba has an outstanding city management that is known the world over. Quite a lot has been written about Curitiba, most of it describing and praising the innovative solutions to problems of urban services, much of it ascribing these achievements to urban planning (see, for instance, Gnatek 2003). But there is less analysis of management methods and on the particular politics that allowed Curitiba to flower. Curitiba is best known for its innovative public transport system - based on planned bus routes and dedicated bus lanes linked to urban zoning (see Santoro and Leitmann 2004). But it is also famous for its environmental policies: it recycles 13 percent of its trash; it i s the city with the largest proportional amount of green spaces in Brazil; it has developed industrial zones with low- environmental-impact industries. Curitiba i s citizen-friendly in other ways, as well: it introduced the first pedestrian streets in Brazil; it has a network of one-stop citizens' shops (Ruu du Cidadaniu). It is reported that 99 percent of Curitibanos say they would not want to live anywhere else (McKibben 1995). 63See FundaggoOswalsoCruz (n.d.). 59 Box 3.4: Networks and Performance Incentives Miller (1992) argues that economic incentives are insufficient for managing complex organizations because agents can always find ways to cheat their principals (i.e., information is asymmetric). Thus "political" mechanismsto improve credible commitment, hence cooperation, become important.Usinggame theory, he points out that it can be inthe mutual interestof agents to cooperate if they believe that they have a long-term future in the organization (which makes cooperation and its fruits a "repeated game"), if they engage in activities that create a shared confidence that others will cooperate, and if they have some reciprocal ability to punish non- cooperators. Horizontal cooperation in smaller teams, he argues, can be reinforced by informal social rituals of reward and punishment, while vertical cooperation between subordinate and superior can be reinforced by creatingenoughtrust to allow each side to reveal more information. Inculcating a sense of common identity and purposeis central to this. Miller applied his analysis of mechanisms of cooperation to the private firm. H i s analysis also applies to public hierarchies; but networks outside the hierarchy can also help foster cooperation. Indeed, this can often be one of their mainpurposes. (On networks, see Thompson 2003). An economic network is a group of people (or organizations) who associate in order to pursue common objectives that have the character of public goods for that group (goods available for all of that group to consume regardless of who paysand who does not pay). Suchgroup public goods include: the public reputation of membersfor acting ethically andknowledgably; the mutual trust of members so that they may cooperate with each other (in work teams or through inter-firm contracting); andthe advocacy of policies of interestto members. The public-goods nature of their objectives means that networks have to ove ome a free-rider problem: members can benefit without contributing to the provision of the public good. Networks therefore need to find instruments to encourage members to conform to expected behavior (notably, to act consistently with the desired reputation or act to share knowledge). To the extent they can do this, networks clarify the expectedstandards of behavior of their members; restrict membership to those more likely to conform to these standards, and to a number consistent with profitable collective action; educate entrants to create an ethos consistent with these standards; punish non-conformance with the standards (ultimately by excluding members from the network); and support standards of behavior by force of law (to make it costly to be excludedfrom the network). Conforming behavior in networks thus results from a mix of motivations that are extrinsic (economic incentives such as the fear of exclusion) and intrinsic (ethos). (Note that according to Miller's analysis, the intrinsic motivation is consistent with long-term self-interest.) If the network functions effectively, this mix of motivations producespredictable standards of behavior that, by benefiting the group, benefit its members. Curitiba's period of intensive urban innovation began in the 1960s, with the creation of a city planningagency, the ResearchandUrban Planning Instituteof Curitiba (IPPUC).64 The strategic- planning approach of the urban planners has become a dominant paradigm of city management. 64OnIPPUC see Campbell(2006). 60 Formally, IPPUC produces a master plan for the city which it monitors and updates constantly. But Curitiba's strategic planning can also be characterized as a fairly "soft" approach to management, inasmuch as it emphasizes a culture, or mindset - about imagining a future and seeing how to organize and coordinate resources to get there - rather than the "hard" rules of systems for human resource management or financial management. This strategic-planning approach was characterized by the report of a study tour from New Zealand that visited Curitiba in2003 (New Zealand, Parliamentary Commissioner for the Environment2002) and illustratedin interviews with Curitiba's two longest-serving mayors since the 1960s (Gnatek 2003). The New Zealand study tour (which was intent on learning lessons for sustainable urban development) concluded that the forces behind city management in Curitiba were quite different from those in New Zealand. Box 3.5, quoting from their report, provides a thoughtful outsider characterization of the "Curitiban Way." Box 3.5: New Zealand Tries to Understand Curitiba In 2002 a study tour of 20 New Zealanders visited Curitiba to learn lessons for the sustainable found -"a healthy ecology, a vibrant economy, and a society that nurtures people." New Zealand development of New Zealand's towns and cities. They were highly impressed by what they is acknowledged as one of the leading exponents of performance management in the world. It is, therefore, interesting that discoveringthe way that Curitibanos went about runningand improving their city should come as something of a culture shock for New Zealanders. "Talking with people who were on the tour i s critical because what this report cannot do is adequately articulate the initial difficulty the group had inrelating to the `Curitiban Way`. This i s the focus Curitibans have, for example, on relationships, quality of life, systems thinking, master plans, long-term visions, smart public/private partnerships and a bias for action. It was difficult to relate to the way Curitibans have approached their development. As a group of Kiwis we have had nearly 20 years exposure to a diet heavily laced with process orientated legislation and policies (an outputs as opposed to outcomes focus), intense application of commercial competition models, separation of policies from delivery functions, effects based environmental legislation, and the rights of the individual having supremacy over the collective (community). Initially we were too focused on the processes, the accountabilities, a mechanistic frame of reference. Our challenge was to realize that in many ways they have had a very different set of values drivingtheir evolution." "They value relationships highly. They create long-term visions of what might be; delivering results to all intheir society by iterations. They champion and grow leaders and are very focused on people and enhancing the quality of their lives and maintaining consistent governance. Above all they focus on getting things done in ways that constantly turn every challenge faced into an opportunity somewhere in society. This very systems thinking approach is woefully absent in many New Zealand social and environmental programmes." Source: New Zealand, ParliamentaryCommissioner for the Environment(2002) Notwithstanding Curitiba's "soft" approach, the New Zealand study tour report observed some "hard" elements inCuritiba's approach to management: 0 The city is aware of the interconnectednessof different city needs and services, hence the need to integrate different sectoral policies. IPPUC takes a leading role in this. 61 0 Policy innovations adopt flexible, experimental approaches, using evaluation and consultation with the community to adjust policies. 0 Policy tools are not necessarily original, but they are applied pragmatically and combined ininnovative andeffective ways.6s 0 Good information underpins other policies. The city puts significant resources into urban researchand making its results available. 0 The city invests in listening to its citizens, through regular public hearings, hotlines, and an array of web-based resources. Citizens are keen users of these channels to report problems. It is also clear that Curitiba has invested significantly in human resources. The city places a strong emphasis on meritocratic entry of new public servants, educational qualifications, leadership, and training.66 The politics of city management. Curitiba's achievements have been secured in a singular political environment. There has been a remarkable political continuity: technocrats have played a powerful role as political entrepreneurs, and they have been able to provide consistent, coherent leadership. Urban planners (Le., technocrats) first defined policy trajectories, then came to dominate the city's political life and ensure political continuity. Since 1971, Curitiba's two longest serving mayors (Jaime Lerner and Cassio Taniguchi) have spent 20 years between them in that office. Both came into politics from IPPUC. With the experience of several decades of improving city services, citizens understand that they have a stake in the system and that they have a voice, and they have voted for continuity. Evans (1997) suggests that Curitiba's benign form of constructive, democratic politics has been underpinned by some particularities. He mentions two necessary factors. First, he agues (quoting Ames and Keck 1997) that Curitiba did not have the dominant traditional oligarchy that typified much of Brazil. The power of traditional elites was further diluted by new immigrants arriving in the later nineteenth century to become independent farmers. The lack of dominant elites encouraged the growth of collective institutions and the provision of public goods. Second, political entrepreneurship, which legitimized planning, was also a necessary part of the explanation. City management and healthcare management. The management style of SMS clearly bears much in common with other parts of the city's administration. The same strategic-planning approach is central. Policies are applied flexibly and progressively. The interconnectedness of the different systems that combine to produce effective health services is well understood. Good information is emphasized. Client feedback is important. Human resources are nurtured. And like city leadership, the leadership of the SMS has been technocratic and it has benefited from continuity. It is plausible to argue that public healthcare has benefited greatly from Curitiba's management traditions. 65 In similar vein, Evans (1997) argues that bus transportation is a clever public-private partnership "carefully shapedby the visible handof public planningand regulation." 66 Curitiba's Secretariat of Human Resources (SRH) administered a merit-evaluation system for all municipal staff from 1991 to 1997. Evaluation results influenced the speed of promotions, but did not affect levels of remuneration. The systemcame to an end becauseof a decision by Brazil's Supreme Court that evaluation of performance was unconstitutional. SRH is now introducing a new evaluation system which will determine promotions and provide for bonuses. The system is based on evaluation of competencieswhich have beenmappedby each sector. 62 It can also be argued that Curitiba has good political reasons to emphasize the provision of social services. Evans (1997) has pointed out that Curitiba's politicians have long had to manage the challenge of conflicts resulting from immigration into the city. Curitiba has seen a very highrate of immigration - at times the highest inBrazil - some of which is, of course, the consequenceof its very success as a city. Hence, the city has had to find ways of winning over the newly-arrived poor in the urban periphery. In this light, the effectiveness and expansion of health services can be seen as a political imperative in the struggle to acquire the political support of new immigrants. The Environmentfor Performance:a ConfluenceofFactors Our discussion of the different sources of influence on performance has inevitably been speculative and should ideally be confirmed through further study, particularly of the policy- reform process. Bearing in mindthe tentative nature of our observations, we can observe a fairly complex pattern of influence inwhich there are importantcomplementarities: While the SUS in many ways provided the lead in policy design, this lead was supported by the network of sunituristus who brought a set of working beliefs and practices consistent with SUS. The leadership of the SMS, inturn, emphasized the importanceof learning from foreign models. Similarly, the SUS provided a guaranteed level of external health financing, while the city government showed a strong commitment to providing additional funds. The mitigation of the principal-agent problem was partly effected through the standardization process inherent in the design of SUS primary-healthcare policies, while SMS policy innovations (standardization of clinical procedures and ICT) also contributed. Contributions to the performance culture came from several directions. We hypothesized that Curitiba city management traditions were important in establishing the approach to strategic management (and the ability to deal with complex systems) characteristic of the SMS. The leadership of the SMS refined several instruments to support the performance culture: a relational-contracting system, team incentives (IDQ), and telephone surveys of users. The network of sunituristus, we suggested, made an important contribution (unusually important, perhaps, for a network) to a strong working ethos. The city government may also have contributed inthis respect. We also hypothesize that the Curitiba city government provided crucial political continuity for the health sector. Thus, several distinct sources of policy change worked together to create a performance-oriented system. Each source may have been necessary, but not sufficient. In the case of two sources of influence, the SUS and the network of sunituristus, all Brazilian jurisdictions could, to some extent, benefit. This suggests that the role hypothesized for the Curitiba city government cannot easily be ignored. The relationship between the working ethos in the SMS and the regime for public servants is worthy of some final speculation. Brazil's regime for public servants produces a meritocratic civil service that should be the envy of most other Latin-American countries. Curitiba reinforces this regime by paying its health workers reasonably, and they continue to benefit during their careersfrom a good training system. So the public service statute does not appear, at first sight, to have proven a problematic basis on which to buildaperformance culture. But many in Brazil and 63 elsewhere criticize, with some justification, the traditional policy model of the tenured Public Servant as one that does nothing for perf~rmance.~~ I s this the case in Curitiba? Clearly, the working ethos is strong in the SMS, and it is tempting to think that this strength is sufficient to counteract the incentives to under-perform that mightreside in the public servant statute. But it is equally plausible to argue (consistent with the incentives model presented in Box 3.4) that a strong performance ethos and employment stability provide complementary conditions for creating an incentive to cooperate: the statute creates the belief among workers that they have a long-term future in the organization, while the ethos creates a sharedconfidence that others will cooperate inthe future. CHALLENGESFORPERFORMANCE How easy would it be to replicate the Curitiba experience? Replicators need to exercise caution for several reasons. First,the story we have told is one of complexity and nuance: a set of interacting systems developed over 25 years and underpinned by a strong and enduring performance culture which is as much driven by informal values as it by formal design. Second, a department with 5,000 staff i s by many standards modest in size, hence more easily managed: coordination mechanisms are easier and smallness makes it more likely that the most important senior staff - notably managers and nurses - will know each other personally. Third, Curitiba's experience occurred within a particular set of historical conditions, including the city's broader reform experience, the Brazilian healthcare revolution, and the rapid expansion of demand for primary healthcare inCuritiba. With that caution in mind, would-be replicators need to understand how healthcare reform, systems thinking, and performance culture have interacted in Curitiba's case and how today's system is the result of a long process of investment in ideas. Inother words, it would be useful to understand how Curitiba has applied strategic planning. It has applied it more as a methodology (or work habit), than as a formal process - a methodology that has pervaded the organization from top managers to the operational level. It has involved a mindset that thinks about the future, looks for and solves problems, understands systems complexity (and understands that changes in one place create problems and opportunities elsewhere). It is empirical, experimental, and takes risks. The formal tools of performance management are an adjunct to, not a driver of, strategic planning. What are the problemsthat Curitiba's healthcaremanagersneedto be thinkingabout? "If it isn't broken, don't fix it." Butbeyondthis, several points may be worth making. 0 It is worth considering whether the largest threat to Curitiba's system might come from a fall inthe rate of expansion and innovation inthe system. This neednot necessarily happen since Curitiba's population continues to grow rapidly and Curitiba will continue to face the same drama of escalating costs and uncontrollable demand as other public-health organizations. (Moreover, there is probably a significant reform agenda insecondary and, especially, tertiary healthcare). But a casual impression is that the rate of expansion and innovation may have fallen in the last couple of years. Lower rates of expansion of the workforce may make it more difficult to absorb poor performers who have tenure. Lower expansion may also reduce job satisfaction: new professional opportunities and new problems to solve would be fewer. 67See, for instance, Bresser Pereira (1999, chapter 16). 64 0 performance comes from outside the sphere of health management - from the attitudes that We suggested in this final section that part of the strong working ethos that motivates staff workers derive from the network of sunituristus,a network inturn rooted inthe health reform movement. To the extent this so, this ethos may gradually fade over time as the health reform movement goes further back into history and the current generation of leaders retires. Maintaininghistorical levels of intrinsic motivation may therefore provide a challenge. 0 Having had the experience of dysfunctional rivalry created by the team-based bonuses of PIQ, it seems that the SMS was sensibly cautious, when it introducedthe IDQ, to ensure that the new system of bonuses for individuals only punished staff in extreme cases. It should continue to be cautious about introducing bonus schemes that reward individuals or teams. But its experience may enable it to design more effective schemes in the future. The experience may suggest that, with good measurement of outputs, a more gradated and selective bonus (distinguishing the best, the worst, and the middling employees, and providing for different levels of reward), and a less generous average bonus, team bonuses mightwork for Curitiba. 0 Finally, if Curitiba's performance has been supported by the expertise and ethos that the network of sunituristus has brought with it, it should also be pointed out that professional networks can have a downside: they can become a source of myopia when the domination of one view of what a policy should be discourages other points of view to be presented. The general reluctance to consider private-sector solutions in primary care, contrary to international trends, may be an example of such myopia. There is also a downside to guaranteed funding of health spending under the SUS: managers are relieved from being as conscious of costs as they would otherwise be. Several factors may make for a financially less rosy future - rising costs of medical technology, the epidemiological transition, the pressure of "health immigrants" on the city's primary-healthcare capacity. Brazil's and Curitiba's healthplanners may have to consider radical alternatives for rationing demand. 65 4. OVERCOMINGAGENCY PROBLEMS: CONCLUSIONSAND POLICYIMPLICATIONS Our study began as an assessment of the effects of specific organizational and managerial innovations - with a focus on human resource management - and concludes with a nuanced interpretation of some of the elements conducive to enhancing public sector performance. A central message is that there is no magic bullet, whether in the form of an organizational change or a cutting-edge managerial instrument. Enabling environments matter, and so does the manner in which change measuresare introduced. The two cases we have examined offer interestingand complementary insights about how to design and implement managerial innovations. A particular challenge in strengthening performance management, independent of a specific organizational setting, i s policy makers' and managers' ability to overcome the agency problem that characterizes the relationships between those who demand performance and those who deliver, and to elicit voluntary cooperation by the employees to pursue the organization's corporate objectives. In this final chapter, we situate these cases within a broader context of economic and management theories, as well as recent international developments in public management reforms; and we present general policy implications. The problem of information asymmetry - patients who cannot determine the quality of the services they receive and managers whose ability to monitor the provision of services by their staff is limited - is particularly acute in the health sector. The relative success of performance management in Si30 Paulo and Curitiba owes much to the governments' effectiveness inreducing this information asymmetry, a main driver of the principal-agent problem. Our Siio Paulo case study has shown that the relative effectiveness of the OSS hospitals is due partly to the more flexible human resource management authority these hospital managers can exercise. Meanwhile, primary healthcare in Curitiba operates within the generally rigid public sector model. However, the case study suggests that under certain (perhaps exceptional) circumstances, the much-maligned direct administration model can support accountability for performance. In the next section, we will reinterpret the two cases from the point of view of agency theory, and highlighthow SiioPaulo and Curitiba addressedthe agency problem. SOCIAL ORGANIZATIONSAS CORPORATIZATION Making the Managers Manage through Performance Contracts and Letting ManagersManagethroughFlexibility Creating Social Organizations to run public hospitals in Brazil is a form of "corporatization" of public services. Corporatization attempts to utilize the "high powered" incentives of markets - whenever feasible - incontrast to the weak and often untoward hierarchical incentives typical of government bureaucracies. Hospital corporatization in SI0 Paulo i s consistent with this prescription: government retains ownership of the hospital and public responsibility for healthcare, but introduces organizational reforms intended to improve performanceby mimicking private sector incentives. Procedural (ex ante) controls are reduced, while simultaneously focusing managerial attention and accountability on outputs.68 The objective is to allow the managers of public hospitals to exercise the same authority as their private sector counterparts, but without losing government control over the quantity and quality of services provided, or the population served. 68 See L.R. Jones (2004) for a recent debate on the meaning and merits o f "new public management." The article also has a number of references to recent studies o f NPM-style reforms (and counter-reforms). 66 A key characteristic of corporatization is that it separates the payer (government) from the provider (in this case, hospital management). Principal-agent theory helps to explain the roles of these two actors The payer (Le., the principal) specifies the goals or outputs desired. Operational decisions are then left largely to the discretion of the provider (i.e., the agent). Procedural monitoring by the principal is kept to a minimum, as a management contract supplants hierarchical control as the means through which public resources are allocated and performance accountability measures are specified. SI0 Paulo's Secretariat of Health has delegated decision- making authority to the managers of particular public hospitals. Thus, there i s a key principal- agent relationship between the Secretariat of Health (principal) and the hospital manager (agent). In addition, there is a principal-agent relationship between hospital managers (principals) and their staff (agents). The rationale for payer-provider splits is to encourage greater mission clarity and to better align the individual incentives of managers and service delivery personnel with public policy goals. Like most attractive reform ideas, implementation presents numerous challenges. Management contracts are inherently incomplete, as it is impossible to anticipate every contingency and specify all actions ina contract in a way that would be enforceable. It is difficult for the principal (the Secretariat of Health) to anticipate actions by the agent (hospital managers) to keep costs down or improve quality.69 In order to achieve her objectives, the principal must rely upon the agent to exert adequate effort and apply his discretion appropriately. However, the principal is limited in her ability to monitor the agent's actions. Information that is known to the agent can be masked or withheld from the principal (e.g., information related to the effort expended by the agent). Thus, ameliorating these information asymmetries, or reducing their importance, is a North as the "cost of measuring and enforcing agreements" ( 1 990:14) - can be prohibitive when central challenge of contracting between principal and agent. Transaction costs - defined by puttingcontract-style reforms into practice. CorporatizedHospitalsinSHo Paulo If introducing a public sector performance contract in commercial sectors is difficult (see Box 4.1), the chalIenges in the health sector are even more acute. In a fully commercial sector, the market transmits information on service/product quality to customers, who in turn will exercise their choice of "exit" to weed out poor performers. In the health sector, the "customers" (patients) are not always in a position to judge the service quality ex ante or even ex post. This market failure increasesthe need for a well-designed contract. As government has a fundamental interest in the quality as well as the quantity of health services delivered, the health sector presents an archetypal "multitask agency problem" (Holmstrom and Milgrom 1991): hospital managers are expected to meet or exceed suitable service standards while simultaneously controlling costs and responding to demands that are difficult to predict, let alone control. This exacerbates the problem of incomplete contracts; and effective performance will depend even more on discretion of the front-line managers and service providers over whom the principal (the government) has only limited control. 69 The prominent role o f firms as economic actors - despite their "low powered" internal incentives - can be understood as a practical response to mitigate the problem o f information asymmetry and the difficulty o f coordinating group efforts (Miller 1992). 67 Box 4.1: Performance Contracts inthe Public Sector What are some of the essential ingredients of successful contracting within the public sector, where the market forces are weak and where independent mechanisms of arbitration and dispute resolution (e.g., courts) are unavailable? According to a World Bank study (1995) on "contracting in" for state-owned enterprise management in developing countries, effective performance contracts depend on the ability of the government (as the principal in the principal- agent relationship vis-&vis public enterprises) to (i)address the problem information asymmetry to monitor and measure the public enterprises' efforts and performance; (ii)design mechanisms for rewards and penalties; and (iii) make a credible commitment to honor the contract. The study found that performance contracts were common in public sectors all over the world; but effective ones - in this case those that led to improvements inpublic enterprises' operational performance measuredas their total factor productivity - were quite rare. Ina significant number of the cases examined in the study, performance actually worsened after the introduction of performance contracts. Often managers of the public enterprises would manipulate their information advantages and agree on "soft" performance targets. Very few of the performance incentive mechanisms were linked to robust measures of performance (partly due to the information problem just mentioned). Many of the public enterprise managers, as political appointees, were not subjected to rigorous assessments of their performance, nor did they have the flexibility or authority to make necessary management changes. Most importantly, performance contracts between the government and public enterprises usually did not specify a neutral arbitration mechanism, and thus left the interpretationof the contracts at the discretion of the government. I Source: World Bank (1995). In such a situation, why should a hospital manager, in the exercise of his delegated discretion through corporatization, be expected to behave in ways consistent with the preferences of the principal? From an economist's point-of-view an efficient solution is to confer upon the agent (in this case the hospital manager) the rights of a "residual claimant." In short, the manager is granted the rights to the surplus (or part of it), so that he has a strong incentive to monitor the activities of his staff to eliminate shirking and ensure efficient service^.^' A number of corporatization reforms have included provisions that give managers (and occasionally staff) a material interest in residual resources. Clearly, a prospective payment contract provides a powerful incentive to reduce costs if a significant share of any savings can be retained by managersas the residual ~laimant.~' This is the case with Silo Paulo's OSS hospitals inthat they are permitted to retain a surplus budget and carry it over into the following financial year. However, this does not create strong individual incentives for hospital directors or staff since they are not allowed to retain this surplus for themselves (e.g., performance bonus), as sometimes occurs in the private sector. In both traditional and OSS hospitals mana ers are paid a fixed salary without bonuses for producing more or better health services for less.7$ 70 Another way to create the same incentive is to awardperformancepay for managers based uponmeeting ''annualtargets,which may include cost reductions. Cost reimbursementschemes offer incentives to maximize the quantity of services delivered, provided the reimbursement rates are adequate. However, cost reimbursement generally does not provide an ''incentiveto reducecosts. The manager of an OS may receive a salary that is greater than that of a typical hospital. However, there are no financial performance incentives. Managerscannot appropriatepart of any surplus. 68 Another way to make hospitals (if not their managers as individuals) residual claimants is to let them earn revenue under market conditions rather than relying solely on a budget allocation (Harding and Preker 2003). However, this is not true of the OSS model in SI0 Paulo. OSS hospitals are part of the national SUS healthcare system and are not allowed to charge patients for medical services. In addition, the hospitals are not directly reimbursed by the SUS for those invoices: the SUS funds go directly to the state treasury. Moreover, OSS hospitals are expressly prohibited from establishing agreements with private health plans. The OSS are entitled to raise and retain revenues from their parking facility or a cafe on hospital grounds; but these revenues are insignificant as a share of the overall hospital budget. Thus, OSS hospitals are wholly dependentupon agreed transfers from the state trea~ury.7~ The arrangement in SZo Paulo does include "penalties," however. As reported in Chapter 2, OSS hospitals must meet their performance targets (within a permitted range) in order to receive their contracted budget transfers. They also must submit to the SES a pre-specified set of information regarding hospital operations, which i s a clever mechanism to ameliorate, if not to eliminate, the SES' information disadvantage vis-&-vis the OSS hospitals. SZo Paulo has adopted another measure to counter the problem of incomplete information. The OSS contracts in Siio Paulo are fine-tuned through regular dialogue between hospital directors and the OSS supervisory staff of the Secretariat of Health. Adjustments are made from one annual contract to the next, but can also be made by consensus within the operational period of a given contract. Provided that the purchaser receives sufficient information about quality, and the provider cares about renewing the contract, "relational contracts" such as these provide a framework for continual improvements (Milgrom and Roberts 1992). Selecting established NGOs - as SI0 Paulo has done - to lead corporatized hospitals can help to support these relational contracts, as these NGOs generally care about their reputation. Regular dialogue also increases the likelihood for natural alignment of goals between government and the service provider, which reduces the risk that service quality will suffer in the face of the monitoring challenges described above. Naturally, if an OSS fails to complete its contract, it risks losing the renewal of the concession. And for its part, the government must be careful not to lower service quality unwittingly by ratcheting downward its budget transfers to OSS hospitals when renegotiating annual performance contracts. Thus, both parties are interested in sustaining a long-term contractual relationship. What gives the SI0 Paulo government an important advantage over many other sates in Brazil is the presenceof a number of NGOs that are capable of, and may be interested in, taking over management of any given hospital (i.e., the government has a viable "exit" option). This option is rarely used, but already one NGO lost its OSS contract over a dispute concerning the hospital's performance andthe adequacy of government f~nding.'~ 73A "residual claimant" may not be needed where the number of patients to be treated and the cost of treatment are known in advance (e.g., for routine, non-emergency procedures). Then, fixed budget contracts may work well, as it may be relatively easy for the principal to monitor basic quantities such as these. The risk, though, in contracting for routine procedures and more sophisticated prospective payment contracts, is that quality may suffer or resources may go underutilized. For instance, a hospital may close wards and leave people untreated if they already have met their numerical goals under a performance contract. (This is not merely a theoretical concern. At the OSS hospital Mario Covas our interviews revealed that mammogram equipment had gone underutilized, not for lack of demand but because the hospital had already met its numerical target for mammograms under the contract.) The challenge, then, is how to expand the decision-making power of hospital managers without threatening the mission of Brazil's Unified Health System (SUS). 74This was the caseof the Hospital Geral de Itapevirunby Associa@o Sanatorinhos in2005. 69 A key premise of the coporatization model as a solution to improveorganizational performance is that managerial flexibility is granted to the manager of the corporatized organization. To meet the organization's specified performance targets, the chief executive officer needs the ability to adjust the mix of resources(inputs) in the way he deems the most appropriate. Thus ex ante input control is relaxed in exchange for ex post accountability for results. The SI0 Paulo case study reveals that OSS managers do indeed enjoy higher levels of discretion in different dimensions of hospital management, particularly staffing. They are free to recruit professionals of their choice without following the rigid public exams (concursos). In extreme cases, they are also able to dismiss staff with performance problems, which is nearly impossible in direct administration hospitals. To the extent that hospitals depend for their performance on their staff`s cooperative behavior with each other, it is tremendously important that directors can recruit professionals that they deem are good fit for the organization. Our study could not investigate in detail how medical professionals work together as a team inside each of the OSS hospitals, or how that relates to the hospitals' organizational performance. But, to the extent cooperative collective behavior is conducive to achieving organizational goals, the directors' ability to mold their own teams must be an essential ingredientof effective hospitals. PERFORMANCE-BASED INTHEPUBLICSECTOR PAY Following current management trends, a common tool for incentive alignment is to link individual employees' monetary rewards (e.g., salaries, bonuses) to measures of performance. Performance-based pay was not utilized in OSS hospitals in Siio Paulo. However, the Municipality of Curitiba provides an interesting case inthis regard. International Experience to Date Since the beginning of the so-called New Public Management (NPM) reforms in the 1980s' a number of countries have introduced variants of performance-related pay (PRP) in their public sectors. PRP schemes typically aim to achieve one or more of the following objectives: Foster individual motivation by appealing to employees' pecuniary incentives; Improve the attractiveness of the pay package, especially at managerial level, vis-a-vis the private sector; 0 Contain the overall wage bill by de-linking salary increases from automatic career progression with seniority; 0 Compensate for the loss of job security in those countries where tenured civil service systems are abolished; Counter the public perception that public servants are overpaid and unaccountable (OECD 2005). Introducing performance-related pay may help to overcome the problem of goal incompatibility in principal-agent relations. An organization needs its employees to be motivated to pursue the organization's corporate objectives, but individual employees are driven by a variety of motives, including many that may not be compatible with the organization's goals. A simple example of how a PRP scheme can align the incentives of the individual with the interests of the organization 70 is a sales comrni~sion.~~However, for many professions, technical challenges in designing an effective pay for performance scheme are daunting. It is often difficult or impossible to measure employees' effort and performance accurately. This problem is particularly well-known in the public sector, where employees' outputs are not always tangible or easily measurable. (See Table 4.1 andWilson 1989.) Table4.1: FourTypes ofPublicOrganization Typeof Characteristics Examples Management organization (USgovernment) practices IRisks ($permitted) Production Managerscan Tax collection, postal Reward good IThe measurabledrives observe service, social performance. out the unmeasurable. processes security (retirement Non-production ("outputs") and claims) agencies (e.g., police) outcomes. try to poseas Procedural Managerscan Health & safety Highlevels of observe administration, armed surveillance of damage morale outputs, not forces inpeacetime procedures outcomes. Craft Managerscan Armed forces in Goal-oriented Freedomof operatives observe wartime, inspection, management(ensure can lead to a minority outcomes, not criminal detection, that good outcomes acting opportunistically outputs. antitrust occur); promote investigation, Army professional ethos Corps of Engineers. Coping Managerscan Teaching, policing, Effective Anad hoc alarm observeneither some diplomatic managementis very systemcan produce a outputs nor activities. difficult, andhas to sense of unfairness, outcomes. rely on ad hoc alarms hence labor relations that somethinghas can be difficult. gone wrong and ensuingpunishment. owe: Geoffre ihepherd(2003), a iptedfromWilson (1989 Agreeing on a manageable number of performance indicators as the basis for determining staffs performance and thus their pay i s a complicated task, especially in the public sector where the "bottom line" takes multiple dimensions (e.g., efficiency, client orientation, service quality, compliance with legality and due processes, etc.). Inevitably, a small subset of "what really matters" is captured as performance indicators either because these are considered truly higher objectives, or simply because it is easier to measure staff performance along these dimensions. This can lead to the oft-mentioned problem of skewed employee efforts (i.e., prioritizing those aspectsof the work that are captured by the pay-linkedperformance indicators), including blatant manipulation of indicators and/or work methods (e.g., arresting more suspects than appropriate when police officers' pay is linked to the number of arrests as a performance indicator). In addition, PRP schemes, especially when workers' efforts and outputs are not easily and objectively measurable, tend to generate discontent and even conflict among employees because of the sense of unfairness they may generate. A recent OECD review of performance-related pay 75 See Milgrom and Roberts (1992) for a more extensive discussion of the theoretical foundation of incentive pays. 71 in its member countries reported that "extensive staff surveys, conducted notably in the United Kingdomand the United States, showed that despite broad support for the principle of linkingpay to performance, only a small percentage of employees thought their existing performance pay schemes provided them with an incentive to work beyond job requirements and in many cases they found it divisive" (OECD 2005:5). Based on a survey of public servants' perceptions of performance pay in selected UK public agencies, Marsden (2003) reports that while a majority of public employees (in the general civil service and N H S trust hospitals, but not in schools) consider performance pay to be a good principle, they also see performance pay as sources of jealousies, and a weak generator of incentives to perform (Table 4.2). Finally, motivating staff to raise their effort is often insufficient for improving their performance. There are a host of other variables that intervene in the process of public service production other than staff effort. Performance improvements may be prevented by staffs lack of capacity to carry out thejob, or misspecificationof the organization's goals (see Figure 4.1). 72 Table 4.2: Replies to employee attitude surveysinselectedpublic service organizations NHStrust Schools ----- ---_ hospitals ---- _ _ I ~ - ~ _ - _l_l Question: % ineach cell ~ I -~_"-ll_ [nland Inland ~ Employ Individual Group replying `agree' or `agree Revenue Revenue ment PRPtrust PRFJ strondv' 1991 1996 Service trust Pay and work orientations PPa good principle 57 58 72 62 52 29 42 Motivation: perceived incentive PP gives me an incentive to 21 18 12 32 22 8 10 work beyondjob requirements PPgives me an incentiveto 27 20 20 36 19 9 11 show more initiative in my job PP means good work i s 41 19 24 47 34 38 40 rewarded at last Motivation: perceived divisiveness PPcausesjealousies 62 86 78 61 51 58 70 PPmakes staff less willing to 28 63 52 22 19 51 54 assist colleagues PPhas made me less willing 10 30 26 19 14 7 4 to cooperate with management Relations with management: non-manager replies: Management use PPto reward 35 57 41 41 27 Na Na their favourites There is a quota on good 74 78 74 57 36 48 45 assessments" Linemanager views: PPhas reduced staff 20 45 39 30 27 Na Na willingness to cooperate with management PPhas increasedthe quantity 22 42 28 52 34 Na Na of work done N(total replies) 2,420 1,180 290 680 900 1,050 860 Response rate (%) 61 30 33 28 21 51 21 Note: based on five-point Likert scales: `strongly disagree', isagree', `no view'. agree' and `agree strongly'. NAHT: National Association of Head Teachers (mainly primary schools); SHA: Secondary Heads Association (mainly secondary schools). Source: Reproduced inMarsden (2003) 73 Figure 4.1: Outline of the "Expectancy" Framework Obstacles: Obstacles: Inadequateskills Poor performance measurement Weak goal setting 1 Mgt. lack necessary resources 0 Poorcoordination Mgt. badfaith Effort Performance Reward Obstacles: Obstacles: No scope to increaseeffort Performancerewardsnot valued Very tight management Other motivationsmore important Already work at max. Conflicts with other motivators Mgt. motivesdistrusted Payfor PerformanceinCuritiba: Not a MagicBullet The perception of fairness is critical for successful application of performance pay. Perhaps because of the long history (and often continued practice) of patronage in Brazil's public sector, Brazil's public servants appear to even less convinced than their OECD counterparts that a PRP scheme will generate higher staff motivation and better organizational performance, even when they accept the basic principle of performance pay (Vaitsman 2001). Various forms of pay-for- performance have been implemented in the Brazil's public administration, In most cases, however, these are PRP schemes in name only. Based on pro forma performance evaluations, they often are nothing more than a disguised form of salary top-up conferred to virtually everyone within the organization. This was precisely what we found in Siio Paulo's direct administration hospitals, as well as in some of Curitiba's previousbonus schemes. Brazil's public sector suffers from another structural characteristic that complicates the effective use of incentive mechanisms: the common practice of multiple job-holding by healthcare professionals. By holding multiple jobs, an individual spreads his work-related risks across differentjobs. This risk diversification is likely to dilute effects of any single mechanism applied in a given job. For example, many job seekers in the public sector are motivated by stability, which in Brazil is largely guaranteed if one enters the public service as an estututa`rio. Highly- trained medical professionals also seek professional development and reputation among their peers. More often than not, large public hospitals where a broad range of cases are treated - including some of the most complex - become convenient laboratories for upwardly-inspired doctors to gain experience and establish their credentials. The Brazilian public sector offers 74 generous pension benefits as Perhaps as a result of the medical professionals' successful pursuitof their corporate interests, public hospitals usually offer doctors flexible work schedules. Many doctors take advantage of this situation and holdjobs elsewhere, partly to compensate for the relatively low salaries from the public sector. Designing an organizational incentive framework for workers who are affiliated with more than one organization and motivated by different attributes in each i s a complex challenge for human resource management. What drives a particular doctor in terms of her professional development goals may not necessarily fit the public policy goals of the public hospital in which she works (e.g., she may be interested in pursuing an obscure specialization that would benefit only few patients or have limited public health impact). When someone is working for a public hospital because of the stability that the estatutdrio regime offers and is holding anotherjob inthe private sector for monetary reasons, there may be little the manager of the public hospital can do to motivate this person with performance-related pay. The same doctors, when they work in the private sector, may respond to more purely pecuniary incentives, and yet not see this employment as the primary avenue for their career development. These difficulties raise doubts about the applicability of performance-related pay in Brazil's institutional and labor-market contexts. However, we found in Curitiba an example of a seemingly well-run performance incentive scheme. Our initial interviews with city officials prior to launching the study indicatedthat this scheme hada positive effect on healthcare professionals' motivations and performance. Despite the well-known difficulties of making performance pay work inthe public sector, the idea that pay-for-performance was working well inCuritiba seemed plausible given Curitiba's reputation as a well-structured, high-performance public administration. Following a closer look at the data, it appears unlikely that this performance- based pay scheme should have a strong impact on employee motivation, as virtually everyone in the system receives the performance bonus. Yet, paradoxically, an overwhelming majority of the health sector staff inCuritiba opines that the latest incentive pay mechanism, IDQ,has a positive effect on their performance. To make sense of this apparent paradox, we have offered a nuanced interpretation of how multiple instruments have played a role in fomenting and sustaining a performance culture in Curitiba, even within the (generally inhospitable) institutional environment of the Brazilian public sector. Our findings point to the central importance of the performance review process as a participatory forum in which managers and staff jointly assess units' performance, identify possible problems and seek solutions. Similar to the regular reviews of performance contracts in Sgo Paulo's OSS hospitals, these processesprovide an opportunity for health sector authorities to identify potential performance problems, further align objectives between the authorities and the front-line providers inthe regional units, and buildshared ownership of both the problems and the solutions. Other managerial instruments play important complementary roles to this performance management process centered on the regular performance reviews. Citizen feedback is regularly sought and used directly in the performance-evaluation process. A well-crafted management information system provides vital infrastructure without which robust monitoring and measurement of performance indicators (the core of the whole process) could not be sustained. Standardization of medical treatments (through the protocols) and their recording in the 76The pension reform in 2003 has made the public sector retirement benefits less generous, but these new benefits apply only to those who have entered the public sector after the enactmentof the reform, and thus do not cover the majority of the active-duty public servants. 75 management information system significantly reduce the costs of monitoring the behavior (Le., performance) of front-line service providers. This observation that pay-for-performance is not by itself a magic bullet but can have positive effects on performance as part of an array of performance-oriented managerial instruments i s consistent with the lessons drawn from the experiences of OECD countries that have attempted PRP schemes over the last two decades. While the limitations of pay-for-performance are widely acknowledged, at least when applied in the public sector, the OECD review also recognizes that introducing PRP schemes can be "a window of opportunity for wider management and organizational change" (OECD 2005:6). PRP gives managers an added incentive to manage effectively and stimulates them to fully endorse a goal-setting approach. Goal-setting and appraisal provide the motivation for the kind of one-to-one contact between employees and their line managers in which the manner of working can be discussed and explained. Goal-setting works well when accompanied by more interaction between manager and staff member so that any reduction of formal controls is substituted by informal control. Another level at which the performance appraisal element of PRP has emerged as critical lies in the scope it offers to link broader organizational objectives to those of individual employees (OECD 2005:6). Curitiba's experience also offers interesting insights concerning how to promote a team-based performance culture. OECD experience appears to indicate that pay-for-performance, at least in the public sector, is more successful when it is team-, rather than individual-based. Team-based incentives may produce positive results to the extent that these generate better intra-team cooperation. But this can come at the expense of inter-teamcooperation, as Curitiba's experience with the team-based PIQ showed. That intra-organizational cooperation i s essential seems obvious; but how to elicit this is less so. Simple reliance on extrinsic incentives, including team- based bonuses, does not seem to be the answer. (See Box 4.2.) CONCLUSIONSAND POLICYIMPLICATIONS The two cases of public management innovations examined in this report demonstrate two divergent approaches to improved service delivery in Brazil's public sector. In both Si30 Paulo and Curitiba a central challenge facing policy makers is how to motivate staff and align their incentives with the government's broader policy objectives. In other words, how can the hierarchical organization of public administration work better through non-traditional management of human resources? Si30 Paulo has introducedorganizational innovation through contracting out hospital management to qualified NGOs in the form of "Social Organizations." This model entails devolution of managerial responsibilities from the State Secretariat of Health (SES) to each OSS manager. It is a model that circumvents the well-known constraints of the RJU system. Curitiba, in contrast, has managed to strengthen performance of its primary-healthcare system within the existing human resource regime. , 76 Box 4.2: Have Team Incentives Worked in Curitiba? Some Speculation Primary healthcare in Curitiba, particularly in PSF Units, depends on teams - small groups of people with complementary skills and committed to a common purpose, set of performance goals, and approach for which they hold themselves mutually accountable. To what extent have team incentives been tried and worked? Two very different schemes have been attempted in Curitiba, and they contain valuable lessonsfor what gets people to work together. The first incentive (PIQ, 1995-97) was an extrinsic incentive explicitly focused on team performance. It enabled the staff of the best-performing 10percent of all Basic Unitsto win a 30 percent bonus. PIQ was judged to have failed because the scramble to secure a high score generated a "checklist mentality" among Units. The relatively large size of the award created a indicators were poorly measured- a sure recipe for resentment among the losers. The PIQ thus strong incentive for teams to win, but the PIQ also contained an opportunity to "cheat" because undermined the unity of the SMS. The IDQ (since 2000) is, on the face of it, an extrinsic incentive to individuals. But there is a strong reasonfor teams to ensurethat individual members perform well because 50 percent of the evaluation of the individual is based upon the teams' performance (35 percent on performance under the management contract, 15 percent on the community's evaluation of performance). Since very few people fail to qualify for the bonus (1 to 2 percent) we do not believe that the IDQ provides a strong extrinsic incentive to promote individual effort (beyond avoiding lateness and absences). By the same token, the IDQ should provide little incentive for team members to ensure that individual members perform well. Yet our survey results indicatethat passing the IDQ test is a major preoccupation of teams; and for that reason they devote increased attention to solving problems that arise with individual team members. Thus, in contrast to the PIQ, staff attest that the IDQ functions as an effective team-based incentive scheme. There is something of a puzzle here. It may be that the incentive is largely intrinsic: the satisfaction of passing the test of the management contract and community evaluation. Unlike the PIQ, however, the apparent team incentive related to the IDQ does not involve rivalry with other Units because it does not function as a zero-sum game. Rewards are more modest. And the scheme is also based on more trustworthy information. Each of these features contributes to a greater perception of fairness on the part of employees. The IDQ experience also suggests that what most clearly drives team performance at the moment in Curitiba are the intrinsic incentives and attitudes of trust and cooperation that follow from the culture of the health sector and the culture of problem-solving (strategic planning) in Curitiba. Clarity andConsistencyofPurposes:InstitutionalandInformationalFoundations In spite of the divergent approaches adopted in these two cases, there are several characteristics they share in common. One is the relative clarity of purpose imposed on both Ssio Paulo's OSS hospitals and Curitiba's regional primary care units through performance contracts with the central authorities (i.e., state and municipal secretariats of health). Clear specification of performance goals is a necessary foundation for any performance management. Direct administration hospitals operate without explicit performance goals. Interestingly, both in S3o Paul0 and inCuritiba, the performance of healthcare organizations is reviewed on a regular basis. These regular sessions serve to fine-tune the content of the management contracts as well as 77 regularly monitor unit performance. In that way the units' performance goals are better-aligned with both the health secretariats' strategic objectives and the units' abilities to deliver on those targets. As an added benefit, regular reviews of goals and actual performance attenuate the information asymmetry between the principal and the agent, a fundamental challenge of any performance management system. Explicit specification of performance targets and meaningful monitoring are only possible with robust information systems. Both Silo Paulo and Curitiba have benefited from investments in this area. Inthe case of Curitiba, development of standardized treatment protocols has contributed to reducing the cost of monitoring regional units' performance. Standardization of service delivery procedures, where possible and desirable, can ease the costs of monitoring the performance of front-line service providers, and is therefore an appealing option to consider. A clear policy implication of these two case studies is that governments should invest in establishing explicit performance goals, whatever those may be, in order to form shared expectations between the policy maker (e.g., education secretariat) and service providers (e.g., schools). Of course, what really counts is substance rather than form. Performance goals must be clear, relevant, and readily monitorable. And they must represent a shared commitment between the principal and the agent. The regular reviews and revisions of performance goals practiced by both Sgo Paulo's OSS hospitals and Curitiba's primary healthcare system is a good practice of "relational contracts" that other jurisdictions would do well to emulate. ManagerialFlexibility or Organizational Development? Once performance targets are clearly set and an arrangement for regular monitoring and measurement of performance is inplace, managers still must "manage" inputs in order to produce targeted outputs (and outcomes). Here, the two cases present somewhat nuanced pictures of what matters most in managerial flexibility. The SI0 Paulo case, in particular, points to hospital directors' ability to recruit staff of their own choosing as an important element of fomenting intra- organizational cooperation. It is our hypothesis, having discarded a number of alternative explanations, that in directly recruiting staff, hospital directors select those individuals that they deem best-suited to the culture of the organization and to mutual cooperation toward achieving the service goals of the organization. The Curitiba story places less emphasis on formal managerial flexibility, since the model has developed within the framework of tenured estututa`rios. The expansion of the sector (and some staff attrition) may have allowed the Health Secretariat a degree of flexibility through increased recruitment. However, the essence of the Curitiba "model" is not one of flexible labor relations where public servants are subjected to the forces of labor market discipline. Here, what seems to matter most is the staffs esprit de corps, a consequence, it appears, of their belonging to the network of self-identified professionals, reinforced by clearly defined policy goals, a performance-oriented cultural legacy, and processes - such as the IDQ- that together reinforce knowledge-sharing and teamwork to further a common policy vision. In short, mechanistic applications of extrinsic incentives have not been the main drivers of performance improvements in Sgo Paulo or Curitiba. Rather, both cases highlight the role of organizations as organic environments within which managers and staff must find common goals (through carefully devised organizational structures and management processes) and seek ways to cooperate with each other through an admixture of diverse managerial and information tools. Our conclusions are similar to Miller (1992:237-238) in his assessment of how inherent "managerial 78 dilemmas" between individual employees' self-interests and the corporate interests of the principal may (sometimes) be reconciled: It is to be expected that hierarchies only rarely and briefly achieve anything that may be regarded as a full resolution of the problems of information asymmetry, team production externalities, and market power. Rather, hierarchies are political settings in which people continually struggle to achieve the potential made possible by specialization and cooperation. They do so as purposive, rational actors who are aware that it i s their own conflicting self-interest that i s the primary obstacle to the achievement of their shared goals. The tools for dealing with this dilemma are the classic political tools: the enforcement of social norms, political leadership, and the credible constraint of hierarchicalauthority. Lessons What do these findingshold for other sub-national jurisdictions inBrazil? Of the success stories, which elements are transferable to other contexts? As we have already noted, there is no single magic bullet to improve the quality of public services. The details of an appropriate reform measure and its application depend considerably on local contexts. Thus, while OSS hospitals have worked well in Siio Paulo, the model may be unsuitable for many states and municipalities, particularly those with few health-sector NGOs that have a demonstrated commitment and experience in delivering healthcare services. The capacity of state/municipal secretariats to design and enforce performance contracts is crucial, as well. But that capacity can be learned and fostered. Findingreputable NGOs to partner with the government to improve services may be the bigger challenge. Performance-related pay is another popular concept in Brazil that purports to leverage service delivery improvements. However, international evidence and our Curitiba case study strongly suggest that PRP alone is a troublesome tool to align incentives between principal and agent. (In fact, it can often be counterproductive.) The Curitiba case does suggest that if PRP is used in concert with other tools it can facilitate the process ofjointly identifying problems, agreeing clear goals, and enhancing performance, even if the monetary incentive effect is difficult to sustain over time. Given that a performance-oriented culture does not emerge spontaneously, reformers should be encouraged to consider adopting (and adapting) the tools that were applied to good effect in Slo Paulo (Social Organizations with clear specifications of performance expectations and rewards and sanctions for performance) or Curitiba (systematic performance reviews based on agreed, monitorable, team-based indicators, with modest performance-related pay playing a supplementary role). The appeal and suitability of specific tools will be determined by the "push" and "pull" factors present in a given setting. "Push" factors are those that propel management reform in a given direction. Meanwhile, "pull" factors are pre-existing characteristics or competencies without which a given solution is likely to founder. As an illustration, Table 4.3 presents the primary "push" and "pull" factors identified in our Slo Paulo and Curitiba case studies. Policy makers should be aware that no single instrument in isolation is likely to bear fruit. Moreover, Siio Paulo and Curitiba already possessed two of Brazil's most sophisticated governments before they launched the managerial reforms described here. Where enabling conditions are less favorable than in,580 Paulo or Curitiba, then modest steps anda degree of caution should infuse any reform 79 strategy. A steady, strategic, experimental approach i s best to turn around an organization's performance orientation. "Push"Factors "Pull"Factors [stimulate change ina certain [support the "good fit" of a given direction] solution] Sao Paulo 0 Dissatisfaction with service Vibrant NGOpresencein [contractingmodel] quality health sector; many potential 0 Human resources wage bill actors with experienceand constraint (LRFlimits) ability to compete to administer "public" hospitals as Social Organizations Curitiba 0 Sunirarisras pursuing a 0 Cultural tradition of strategic [directadministration particular preventive care planning and problem-solving managementwith strategic model planningand performance 0 Unintended consequences reviews] of earlier performance- related-pay scheme Policy Implications Beyond the general observations above, our case studies do not provide a straightforward blueprint for other sub-national governments in Brazil to follow. Still, there are some broadly common strategic tools (which can be applied at different levels of sophistication); and there are choices to be made about the incentive systems that govern staff and management behavior. The case studies explore two types of management tools: i)those that aim to align the expectations of principal and agent and mitigate information asymmetry, and ii)those that provide direct incentives to tie an employee's or manager's behavior to performance outputs. The two types are not intrinsically incompatible; but knowing how to combine these instruments i s more an art than a science. Instruments for aligning expectations and reducing information asymmetry between principal and agent: There are a number of techniques that reform-minded governments can employ to better align expectations and incentives between principals and agents. Indeed, our two studies suggest a common list of "good things" to do: 0 Invest in strategic planning by clarifying expectations and establishing regular feedback on performance. These processes work better when operators - the front-line troops - are involved inthe planning-and-evaluation cycle. 0 Invest in better strategic management of information by standardizing processes and definitions; ensuring the quality of data; tapping information from the community; and providing IT systems to manage this information. Choosing among instruments for direct performance incentives to agents: The instruments that create clearer, direct performance incentives for staff and managers are generally difficult to apply because they entail complicated policy tradeoffs (including a higher political profile). Our two cases suggestthat extrinsic personnel incentives are more compatible with the use of external 80 labor markets, while intrinsic incentives may be more compatible with internal labor markets (e.g., the RJU, where entry and exit are limited). 0 The OSSs in SZo Paulo rely primarily on extrinsic incentives: the freedom and incentives to manage for managers, and the discipline of external labor markets for staff. 0 The SMS in Curitiba places greater emphasis on intrinsic motivations for staff (investing in esprit de cops, using staff appraisals to identify and solve problems) and maintains a more hierarchicalmanagement format. 0 Neither system has so far relied heavily on bonus schemes. (OSS hospitals have not yet tried them, and Curitiba has yet to get far with them.) Of the two management tool types, steps to better align expectations and reduce information asymmetry are generally less sophisticated and entail fewer risks of "unintended consequences" Brazil - which lack the installed administrativecapability of SZo Paulo and Curitiba -may find it than direct performance incentive strategies. Thus, the majority of sub-national jurisdictions in necessary to focus first on basic statistics and management information, as well as minimum administrative capacities (e.g., simple planning, logistics management). The government must have a clear policy objective that can be translated into meaningful performance targets/expectations at the sectoraVoperationa1level. In the beginning, when the government's experience with designing proper performance indicators is limited and its capacity for information management is weak, it is preferable to focus on a limited set of performance indicators: far better to have a few indicators that can be reliably tracked and serve as references for the government and service providers in their regular performance review discussions than to propose myriad indicators without the ability to track them on a timely and reliable basis. As our case studies demonstrate, there are gradients of sophistication in each set of tools. Consider strategic planning, for example. Clarifying expectations can be tackled at a minimumof four different levels: i)informal or formal discussions; ii)formal strategic or indicative planning; iii)relational contracting; and iv) enforceable contracts. Similarly, feedback can take place through simple discussions (informal or formal) or through a formal evaluation process. Once again, what really matters is the political commitment behind the expressed performance expectations. The choice of instruments is secondary. Governments with limited institutional capacity are well-advised to be cautious about adopting formal contractual tools, as it will prove particularly difficult to set appropriate performance targets ex ante, which may result in frequent contract revisions (reducing the credibility of contracts an enforceable tool) or its complete disregard by the actors involved. Similarly, mechanistic applications of performance incentive tools, especially performance- related pay, should be approached with caution by governments with weak administrative capacities. The conditions under which performance-related pay can be effective are rather stringent (e.g., performance itself is reliably monitored, managerial "courage" to apply both positive and negative incentives, etc.). When such conditions are not met, PFW schemes become a formality without real impact on staff behavior, or even produce negative consequences. Onthe other hand, development of professionalism and esprit de cops is a longer-term endeavor that is not within easy reach of any particular government administration. As the federal government has done, conscious efforts to strengthen professionalism of certain "core" functions through careful recruitment strategies will pay off inthe medium to long run. 81 Table 4.4 provides a concise depiction of the performance approaches utilized in Sgo Paulo and Curitiba, and their effects. Other sub-national jurisdictions in Brazil can learn from these experiences, but must be careful to choose techniques within the reach of their capabilities. Natureof incentive Aligning incentives (reducing Rewardingperformance instruments informationasymmetries) Strategicplanning Clarifying expectations CU: goal claritication through "relational contracts." SP: goal clarification through relational and enforceablecontracts. Feedback on goals attainment CU: problem bolving culture: detecting probleins and problem- solving through contract diwussions. SP: problemdetection andsolving through contract discussions; accountability (contract fulfillment). Making complex systems work CU: inve\tment in combination of ;L s) stems, leading to better processes and information. Instrumentsfor strategic management Information system CU: good informationsystems contribute to goal clarification. SP: good information systems contribute to goal clarification. Standardization CU. evplicit standardization measures aid clarity of infomiation nnclgoal-setting. Informationfrom community CU: community information reveals Instrumentsfor inducingpersoi el toperform Staff evaluation CU: personnelevaluation ;in aid to detecting problems and problem- solving. Extrinsic incentives: diferential CU: not very strong results froni remuneration bonus systems w fat, but potential for sniull-team collaboration incentives Extrinsic incentives:freedom to SP: freedom to hire (usedmore hire andfire than freedom to fire). Intrinsic incentives CU: inkesttnetit injoh satisfaction -espritclecorps(sometension t\ith extrinsic incentive\). Appropriation of surplus SP: OSSscanretain savings. Enforceable contract S P poorly performingOSSs lose franchise. Freedom to use resources ISP: OSSs canoptimize resource use and retain savings. C U Curitiba primary care SP: SI0 PauloOSS hospitals 82 APPENDIX A. HospitalsinSiioPauloResearchSample' The study sample includes seven traditional state hospitals, seven OSS, four private hospitals, and two public hospitals supported by foundations. At each of the hospitals visited, small focus group interviews were conducted with doctors and nurses - outside the presence of their managers - to gather their observations about the hospitals where they work, and explore the reasons for their employment choices. The questionnaire and core focus group questions are presented in Appendix B. HospitalCharacteristics No. Legal Status Direct administration 07 Public (with Foundation) oss 02 07 Private 04 Type of Services Secondarykertiary 18 Tertiary 01 "Quaterndrio" 01 Size (by no. of beds) Small < 100 02 Medium 101-200 06 Large 201 -400 11 Extra-large >400 01 Year of inauguration Public (with Foundation) 1888-1980 02 Direct administration 1948-1998 07 Private oss 1965-2oO4 04 1998-2003 07 Source: Research questionnaire -"Caracten'sticas dos Hospitais." Note: Inaddition to these 20 hospitals where a uniform instrumentwas applied, valuable information was also obtained duringpreliminary visits and interviews at the following hospitals: 0 Brigadeiro (direct administration) 0 Vila Penteado (direct administration) 0 Mario Covas (OSS) 0 Carapicuiba (OSS) 0 Sumari (foundatioduniversity) 83 APPENDIX B. ResearchQuestionnairesAppliedinS5oPauloHospitals I-DADOSDAINSTITUICAO Nome: Enderego: CEP: FAX: e-mail Naturezajuridica da InstituigIo: Pdblico ( ) Privado( ) Caracten'sticada InstituigIo: AdministragIo Direta 0 Autarquia 0 FundagI o oss 0 0 Outras ( ) . Quais? Data de inicio das atividades I I Nome do Diretor Geral Formagio / Perfil 0 Possui Curso de AdministragIo Hospitalar ou Equivalente? 0 HAquanto tempo exerce o cargo de diretor? 0 Como foi feita a sua indicaGIo para diretor? 0 Existe algum instrumentoou mecanismo que avalie seu desempenhono hospital? Qual? 0 Voce considera 2004 um ano de exito para o Hospital?Por que? 0 Quais foram as principais metas adotadas?Ecomo foram selecionadas? 0 Epara 2005 quais as metas que voce pretende atingir? I1-PERFILDOHOSPITAL Porte do hospitalem relagIo ao node leitos instalados: Pequenoat6 49 leitos 0 MCdio 50 - 149 leitos 0 Grande 150-500 leitos 0 Extra-grande, acima de 500 leitos 0 Complexidade do Hospital no Sistema de Sadde Tipo do Hospital: Geral 0 Especialidades 0 Quais? Taxa de ocupagIo (%): Taxa de permanencia (%): Ndmero de Leitos Operacionais: Ndmero de Leitos por Especialidade: Ndmero de funcion6rios por leito: 84 Dados da Produqgo do Hospital (Exercicio 2004): Ndmero de consultas ambulatoriais: Ndmero de consultas emergenciais: Ndmero de Exames (SADT): Ndmero de Internaqaes: Ndmero de Cirurgias: 111-WORMACAO ACESSO ECONTROLE 0 Existe um banco de dados no hospital com informaq8es sobre pessoal? 0 Quem administraesse banco de dados? 0 Com que frequCncia C atualizado? 0 Como C feita a coleta de informaGdes? 0 Que tip0 de informaSdes consta desse banco de dados? 0 De que forma a Dire@o do Hospital se utiliza desses dados e que decisdes s50 tomadas em rela@io aos mesmos? IV-CONTROLEDEPAGAMENTO 0 Qual a fonte de recursos para o pagamento de pessoal? 0 HAdiversas fontes derecursosparao pagamento de pessoal? 0 Alguma vez houve dificuldades financeiras para pagar o pessoal?Eque medidas foram tomadas? 0 Quando hA cortes orGamentkios o diretor tem autoridade para decidir onde cortar? 0 HAalgumaesp6ciede oqamento participativo? V -POLiTICA DERECURSOSHUMANOS 0 Existe a Area de recursos humanos: S I M 0 NAO ( ) [Anexar o organograma] 0 Quais as formas de contrata@io de pessoal? Estatutdrio 0 EmergCncia (733,3131) 0 TemporArio (Lei 500) 0 CLT 0 Cooperativa 0 Terceirizaqgo ( )em que Area: Presta@o de services 0 0 Quais serviSos siio contratados pelo hospital? Limpeza 0 SeguranCa 0 AlimentaGBo 0 Lavanderia 0 Outros 0 0 A contrataggo para umou mais serviCosacima descritos C recente (nos dltimos 3 anos)? S I M ( ) NAO ( ) 85 Taxa de absentismo (%): Taxa de rotatividade (%): Ndmeros de A@es Trabalhista (Exercicio de 2004): Tipos de A@es Trabalhista: Qual a contribuiqiio dos serviqos abaixo relacionados para o bomdesempenhodo hospital, pontue de uma dez: Recursos Humanos 0 Recursos Materiais 0 Recursos Financieros 0 Existem atividades ou a@es que ajudem a fortalecer o compromisso dos funcioniuios com as metas do hospital? Quais os principais obsticulos com relaciio ao Recursos Humanose que procedimento voce adota para remove-los? VI-RECRUTAMENTO ESELECAO Quais as formas de recrutamento utilizadas pela InstituiGiio? Recrutamento interno 0 Recrutamento externo: Jornais 0 Diirio Oficial 0 Outros ( )Quais? Quais as formas de seleqiio de pessoal utilizadas pela Instituiqiio? Concurso pliblico 0 Prova pritica 0 Entrevista 0 Dinlmica 0 Teste Psicol6gico 0 Aniilise de Curriculo 0 Seleciio 0 Outros ( ) Quais? 0processoseletivo Cexecutadopela: Pr6pria instituiqao 0 Empresa contratada 0 Outras ( )Quais? 0 Qual a dura@o que existe entre o processo de seleGloat6 o profissional comepr a trabalhar? 3 dias 0 1semana 0 1 mCs 0 2 meses 0 6 meses +6 meses 0 0 Outros 0 86 0 Os resultados sobre a seleggo variam muito de acordo com a forma legal de contrataggo (Estatutlrio, Emergsncia (733,3 13I), Temporlrio (lei 500), CLT, Cooperativa, Terceirizagio)? 0 Quais s50 as vantagens e desvantagens destes regimes do ponto de vista do diretor? VII -CONTRATACAO 0 Quem toma decisdes sobre a contrataggo de funcionlrios? 0 0 queohospitalfaz quando hdnecessidadedecontratar funcionhrios?A quemserefere: Secretlrio da Salide 0 Cooperativas 0 Contratos temporlrios 0 Outros ( )Quais? Voce tem autonomia para contratar um determinado medico quejulgue competente? Qual a especialidade mais dificil de contratar? Recrutamento dessa especialidade 6 realizado de forma diferente? 0 hospitaltem autonomia parademitir funcionlrios?Qualoprocedimento? Quais os fatores que levam a demissdes de funcionhrios? HAoutras maneirasde punirfuncionkios pormaldesempenho? HAinterferhcia de sindicatos/cooperativas quando hddemissdes?Quais sgo? Qual o nlimero de profissionais treinados e capacitados no exercicio de 2004?Indicar o ndmero de profissionaistreinados e capacitados por categoria e carga horlria (Anexo 11) VIII -PLANEJAMENTO DERECURSOSHUMANOS 0 A instituiggo possui plano, cargos e sallrios? SIM( ) NAO( ) 0 Realizapesquisade mercado para manter o equilibrio salarial interno? SIM( ) N A O ( ) 0 Quais os beneficios praticados pela Instituigb (vale refeiggo, vale transporte, cesta bdsica, plano de sadde, plano odontol6gico, bolsa de estudos, outros)? XIX -AVALIACAO DEDESEMPENHO 0 Existe instrumento de avaliaggo do desempenho do profissional: S I M ( ) NAO ( ) Emcas0 positivo, especificar a periodicidade e os critdrios de avaliaggo. 0 Existe premiagao por produtividade? S I M ( ) N A O O Emcas0 positivo, quais os criterios? Equais premios? 0 0resultadodaavaliaggo dedesempenhosubsidia: Dispensa do profissional 0 Readaptaggo do profissional 0 Treinamento e capacitagio 0 Promoglo 0 Revisgo dos critCrios de seleggo 0 Outros ( ) Quais? 0 Existe plano de carreira na instituiggo? S I M ( ) NAO() 87 X -CONTROLE/ DISCIPLINAS 0 Qual o procedimento informal adotado pel0 Diretorao profissional (medico / enfermeiro) pel0nlo cumprimmto da carga horiria ou atraso no plantlo? 0 Qual a atitude / aglo do diretor quando o profissional (medico / enfermeiro), se nega a participar ou colaborar para o desenvolvimento das atividades? 0 Nos tiltimos 3 anos houve algum processodecorrente de err0 medico e qual a atuaglo do CREMESP? 0 Nos tiltimos 3 anos houve algum processodecorrente de err0 medico e qual a atuaglo do COREN? 0 Quais procedimentos slo mais eficazes para disciplinar os funcionirios? XI-CONTROLEDEPAGAMENTO 0 Como slo definidos os salirios dos profissionaisdo hospital? 0 Existe flexibilidade de pagar os profissionais com base em seu desempenho? Existe competiglo e conflito entre os profissionais? Que procedimento C tomado pela direglo do Hospital? XI1-CONTROLE EXTERN0 0 Existe controle que fiscaliza a execuglo orgamentbria / folha de pagamento / gestlo de pessoal/ qualidade de servigo? Se sim, como 15a rela@o com a direglo do hospital? 0 Os funcionirios fazem parte do orglo? 0 As decisaesdessesorglos tem impact0 direto na gestiio das atividades do hospital? 0 Quando os clientes tem queixas com relaglo aos servigos oferecidos a quem se reportam? 0 As queixas feitas slo levadasem consideraggoe que providencias slo tomadas? 0 Hiumaouvidorianohospital?Ea quemestasereporta paraa discusslo dosproblemas? 0 Os problemas em geral slo resolvidosde que forma? 88 Quantifica$io dos Recursos Humanos [referencia dezembro 20041 Fisioterapeuta MCdico Nutricionista Programador ~~Terapeuta Ocupacional Outros Nivel Universitirio (*) Sub-Total Nlvel Universitirio Sub-total Nivel MCdio 3. Nivel Bisico Manutenqfio Motorista Vigia Outros Nivel Bisico (* ) Sub-Total Nivel Bisico TOTAL, OBS: Acrescente linhas, se necesskio 89 TREINAMENTOEDESENVOLVIMENTO (EXERCfCIO DE2004) MVELUNIVERSIT~IO Nome do Curso/rreinamento: Carga horhria: Ndmero de ProfissionaisTreinados: Institui@ocertificadora MVEL~ D I O Nome do Curso/Treinamento: Carga horhria: Ndmero de profissionais: Institui$lo certificadora N~VELBASICO Nome do CursoRreinamento: Carga Horhria: Ndmero de Profissionais: OBS: Acrescentar linhas para cada curso/treinamento realizado 90 APPENDIX C. FocusGroup Interviews inSQoPauloHospitals PESQUISA QUALITATIVA DEGRUPOFOCAL, Pesquisa de grupo focal, com mCdicose enfermeiros das instituiqdes hospitalares Questdes: Existe politica de formaqiio e capacitasiio na instituigiio onde trabalha? Dentre as aqdes de forma@o e capacitagiio, voce destacariaalguma que valoriza o profissional e que priorize a melhoria da qualidade dos serviqosprestados? Quem coordena as aqdes de formaGiio e capacitaqiio na institui@o? Destaque as a@es de formaqiio e capacitaqiio voltadas a informaqiio e orientagiio para a populaqiio quanto aos serviqos ofertados? Quais as instituiqdes formadoras parceiras do process0de capacitaqiio de sua institui@o? Existe investimento para o desenvolvimento de pesquisae ensino? Quando existe treinamento / forma$io a solicitaqiio C feita por parte dos profissionais ou C uma imposi@o da Diretoria? Existe uma determinada frequCncia para realizagiio desses treinamentos? Como voce identifica o cumprimento da legisla@o de pessoalem sua instituigiio? Como siio realizadas as contrataqdesde pessoale quais os critCrios de sele@o? Como siio realizadas as demissdesna sua institui@o? Como se dd o cumprimento da carga hordria dos profissionaisque trabalham na sua instituiqiio? Quais as medidas disciplinares adotadaspor sua instituigiio?Quais as medidas para estimular a cooperagiio com as metas do hospital? A institui@o adota mecanismo de incentivo aprodutividadehom desempenho? Caso positivo, quais os critCrios de mensuraqiio? Quais os mecanismos de promoqio e evolu@o funcional usualmente adotado por sua institui@o? Existe Plano de Carreira na sua Institui@o? HApossibilidade deascensiioprofissional paravoce nestainstituiqiio? Quantos vinculos empregaticios voce tem. Quais siio? Qual a vantagem e desvantagem em trabalhar no setor Pliblico? Qual a vantagem e desvantagem em trabalhar numa OSS?Enumhospital de administraqiio direta? Qual a vantagem de trabalhar no setor Privado? Quais os fatores que motivam o profissional medico para que tenham multiplos empregos (classifique de 1a 10de acordo com suas prioridades) --- aprendizadoimanter-se atualizado na sua especialidade 0 saldrio 0 estabilidade 0 -- prestigio institucional 0 ligaq6es academicas(residgncia, mestrado, professor) 0 - outros 0 Total (10) 91 Quais os fatores que motivam seu emprego no OSS? (classifique de 1a 10de acordo com suas prioridades) - aprendizado/manter-seatualizado na sua especialidade ----- estabilidade 0 salkio 0 0 flexibilidade de horkio 0 ambiente/organizaqiio/ordemsuperior 0 prestigio institucional 0 - perspectiva de ascensiioprofissional na instituiqiio -- ligaqdesacademicas(residhcia, mestrado, professor) 0 0 outros 0 Total (10) Hospital de administraqiio direta? (classifique de 1 a 10de acordo com suas prioridades) --- salkio aprendizadolmanter-se atualizado na sua especialidade 0 0 estabilidade -- flexibilidade de horirio 0 0 ambiente/organizaqb/ordemsuperior 0 0 -- perspectiva -- prestigio institucional de ascensao profissional na institui@o 0 ligaqdes academicas(residencia, mestrado, professor) 0 outros 0 Total (10) Hospital privado? (classifique de 1a 10de acordo com suas prioridades) - aprendizado/manter-seatualizado na sua especialidade 0 - salkio -- estabilidade 0 0 flexibilidade de horkio - ambiente/organiza@o/ordemsuperior 0 --- prestigio institucional 0 0 perspectiva de ascensiioprofissional na instituiqiio 0 1igaCdes academicas(residencia, mestrado, professor) 0 - outros 0 Total (10) 92 APPENDIX D. SurveySummary: Avalia@o dosIncentivosaosRecursosHumanos naAtenqiioPrimariaem SaGde de Curitiba InJuly-August 2005, as part of its study of the managementof primary healthcare inCuritiba, the World Bank, with the help of the Secretaria Municipal de Sabde, carried out a survey of perceptions that senior staff inHealth Units had about the performance incentives they face. The survey was sent, through SMS's internal mail system, to 372 eligible staff (nivel superior) in a sample of Health Units. 254 responded (a satisfactory 68 percent response rate). The 31 units in the sample were chosen to give adequaterepresentation to: all Health Districts; the distribution by Basic Units. (Emergency Units - 24 Horas - were excluded form the survey.) The survey was type of senior staff (ASL, doctor, nurse, dentist); and the breakdown between PSF Units and confidential and replies were coded, entered, and analyzed by the World Bank. The survey questions were selected, after discussion with the SMS, on the basis of field ' interviews that sought to understand the management system and performance incentives overall. We also went into the study with a particular prior interest in the way that the remuneration-based incentive IDQ worked. 29 multiple-choice questions tackled the following themes. 0 The uses, benefits and costs of Management Contracts (Termo de Compromisso de GestZio / POA) and the way these are negotiated. 0 The extent to whichpolitical influence is exerted on management contracting and on staffing. 0 The impact of the Incentive Program for Quality Development (IDQ salary bonus) on staff behavior andperformance and on teamwork and the role inIDQ of community evaluations. 0 The extent to which the above factors, as well as belonging to a PSFUnit versus a Basic Unit, contribute to teamwork and staff motivation. Here is our summary of the survey results, with some commentary. SMS's managernent-contracting system is perceived to be working, and professionals have bought into thisper$omzance-oriented culture. 0 95 percent of respondents find management contracting useful in general -for organizing work processes, identifying and solving problems, and clarifying objectives and establishing priorities (Grbficos 6-10inthe survey report). 0 85 percent found that the benefits of management contracting exceeded costs (Grbfico 11). 0 93 percent found that the information system helped identify problems (GrAfico 12). The system helped in a variety of ways relating to health information and to management: epidemiological information, faster service, team productivity, reaching targets (Tabela 1). 0 61 percent were ignorant of how other Health Units were performing (Grifico 13). But this picture changes when different types of staff are considered: a large majority of managers (ASL) do know how other Health Units are performing, while it is practitioners (nurses, dentists and, in particular, doctors) who do not. Among the managers (Le. those with most opinions), those that profess a view tend to think they are performing similarly to other units (GrAfico 14). Commentary: It is the managers, not the practitioners, who regularly meet with their supervisors at the District level, so it makes sense that the managers are more aware of how other Health Units are performing. But there appeared to be a general hesitation to judge the performance of other Health Units (and a hesitation to find them better or 93 worse). It can be inferred from this that there is no strong sense of competition between Health Units. a Targets are fixed between District and Unit with a mix of negotiation and imposition (Grbfico 15). But the Health District is open to re-negotiation when there are difficulties in fulfillment (Grbfico 16). a Performance contracting is an instrument that creates links with a broad variety of other municipal-government services (Grbfico I 7 and Tabela 2). The role of the ConselhoLocal de Sau'de in management contracting appears to be on the modest side. a The involvement of the Local Health Council in management contracting is positive, but limited (Grbficos 18-21). The Council's greatest contribution is to provide information about the community and to identify problem (Tabela 3). The survey provides ambiguous information on external political influences. a Many staff consider that external political influences are substantially present in setting targets, appointment decisions, and staff assignments to units (Grhficos 23-25). Managers, doctors and dentists share something of a pattern of bi-modality: they either feel that political influences are strong or that they are absent or weak, but relatively few feel they are moderate. (Nurses are the group most unambiguously convinced that political influences are strong.) Commentary: This is the one part of the survey where responses were not broadly in line with interview findings. This may be becausethe survey was confidential. Yet the replies are difficult to interpret. The questions may have proven ambiguous to respondents (if only because managers are political appointees and because production targets are political targets). The bi-modal pattern of responses raises questions about the usefulness of the responsesto this question. IDQ is perceived to stimulate staff pegormance, partly because, through the staff-evaluation process, it gets teams to help solveproblems of individuals. There is an overwhelming view (90 percent) that IDQ had an influence on behavior when it was introduced(Grbfico 26). This influence was felt to be: greatest in the area of workplace behavior (discipline, punctuality, and so on) and motivation; middlingfor work organization and productivity; and least (but still with some impact) for improving salaries and team work (GrAficos 27-32). 61 percent of respondents believed that IDQ continues to provide an important influence on behavior (Grhfico 33). The individual evaluation process that is part of the IDQ is perceived to contribute most by promoting employee-supervisor dialogue; it contributes somewhat to training and solving team conflicts and can have effects on employees losing bonuses. It is not perceived to result intransfers to other units(Grhficos 34-36 andTabela 4). Employees generally think of the individual evaluation process that is part of the IDQ as a fair instrument -transparent, impartial, and coherent (Grbficos 37-39). Clearly, passing the IDQtest is a major preoccupation of teams (Grbfico 40). Teams, when they perceive a problem with a staff member, will, most often, talk to that person or else adjust work processes(Tabela 5). 94 Commentary. Since the IDQ statistics (Table 7) clearly indicate that virtually the only staff to forego the bonus are those with problems of workplace behavior (mostly punctuality), it is difficult to believe that it functions as a simple performance bonus (as a sales commission does, for instance). (Note that workplace behavior is perceived to be where IDQ had its greatest impact.) Beyond this, it is a vehicle for staff evaluation and for the identification and solution of problems that relate to team performance. Sometimes, management innovations tend to have a short-lived effect (known as a Hawthorne effect) as they lose their novelty and workers adjust their behavior: the IDQ may be one such innovation. There is, indeed, a prima facie indication of a substantial decline in the perceivedeffectiveness of IDQ (Grbfico 33), though this is not completely clear. Community evaluations (part of the IDQ evaluation process) are useful in making services more client-oriented. 0 Teams are well aware of Community Evaluations (Avaliaqdes da Comunidade) and make use of them (Grbficos 41 and 42), though seven percent are unaware and 17 percent make little or no use of them. Most use is made of them in improving client service, better resolution of cases, and reducing waiting times; less use in improving physical installations or handling peak demand (Tabela 6). Respondentsshow a strong sense of professionalism (which also suggests a substantial sense of unity across different types of Health Unit). 0 Professionals report that they are motivated more by job stability, elements contributing to their professional standing (job content, learning opportunities, training opportunities), and nearnessto home than by salary or flexible hours (Gr~icos43-49). 0 Respondents perceive the differences between PSF and Basic Units as residing in salary levels, team methods, types of service provided, length of working week, and to some extent availability of resources. Perceptions of differences in professional standing were far less marked(Tabela 7). Teamwork is perceived to work well, across different types of unit, but the exact ingredients of this are not clear. 0 There was a strong perception(92 percent), shared by staff of PSF and Basic Unitsalike, that teams worked well (Grbfico 50). It is not clear what are perceived to be the most important factors inthis (Tabela 8), but good management and clarity in task attribution appear to be at the top (a common position inBasic and PSF Units). 95 APPENDIX E. ManagingCuritiba'sPublicHealthSystem: Instrumentsand Impacts Instruments Intenddpossiblehealth Intendedpossible impacts managementimpacts Client-orientah n Reachingthe 0 Geographical 0 Easier client access 0 Decentralization client decentralization via to health services. devolves Districts and HealthUnits accountability. (1991); Unitsare linked by public transport. 0 Family medecine- ProgramaSadde da Familia (1993). 0 Programade Agentes Comunidrios de Sadde- PACS (1999). Empowering 0 Municipal and local 0 Community and 0 Community and the client HealthCouncils (1991). individuals provide individuals provide 0 Central de Atendimento local knowledge. expectations and ao Usukio (1993): information onservice complaints system. quality. 0 AvaliaGBes da Comunidade (2004): telephoneevaluation system. Makingthe 0 Acolhimento Solidirio 0 Reduction of clients' 0 Management of SMS' client's life (1998): client-friendly transactioncosts (e.g. external relations. easier processre-engineering. waiting times). Use of inform sn Social 0 Planning of services 0 Area knowledge knowledge linked to area-based allows prioritization knowledge and analysis. and concentration of 0 Decentralized Health resources, hence Units(especiallyPSF more effective Units)andthe useof services. Agentes Comunidrios de 0 Knowledgeof Sadde. individuals and families permits more effective services. ~ Medical 0 Evidence-basedmedicine 0 More effective 0 Process knowledge (EBM): clinical procedures. standardization, 0 Epidemiological through protocols, activities. can reap benefits 0 Clinical protocols from economies-of- (from 1998). scale and can 0 Training (see below). facilitate performance measurement. Information 0 Computerized information 0 Improves accuracy 0 Improves accuracy technology systemfrom 1988. andtimeliness of and timeliness of 0 Cart20 Qualidade Salide clinical information. managerial andProntukioEletr6nico Improves information. 96 Facilitates horizontal information (EBM), and vertical thus contributes to coordination. more effective clinical processes. performance-based I I I Strategic areas Instruments Intended/possiblehealth Intendedpossible impacts managementimpacts Speciulization Choices to 0 Municipal production of 0 Vertical 0 Emphasison primary produce or primary health care and specialization, with an and preventive financehegulate regulation and financing emphasis on healthcarecuts health services of much secondaryand preventative demandfor (or most tertiary health healthcare, allows allows more care. SMS to concentrate specializeduse of) 0 Shiftofemphasis from on public-goods secondaryand curative/secondary& aspectsof health (for tertiary services. tertiary to instanceinprograms preventative/primary likeMIe Curitibana medicine. andCirie Zero - Amigo Especial). Horizontal Problem-specific 0 Specialization and 0 Specialization and specialization programs: MIe concentrationof process Curitibana, specific resources can lead to standardizationcan diseases, ,etc. more effective clinical provide benefitsfrom 0 Standardizationof processes. economies-of-scale Health Unitprocesses, and can facilitate notably Protocols performance measurement. ICoordination of Coordination ("integraqiio") 0 "uma rede integrada 0 Coordination reduces of differenthealth services, de pontosde atenGI0 unnecessary with Health Unitas que prestaassistsncia (secondary and gatekeeper: continua B populaqIo, tertiary) activities. 0 Central de MarcaqIo de no tempo certo, no Consultas lugar certo, como Especializadas(1994). custo certo e a 0 SistemaIntegradade qualidade certa" Serviqos de Salide (2002). anugement 0 Careerrules: merit- 0 Training keepsstaff 0 Career rules produce basedentry; tenure; familiar with medical competenceand inflexible career- advances (including shield staff from developmentrules. EBM). politics, but do not 0 Training: permanent promote professional training; U. performance. Toronto training in 0 Training givesjob family health andEBM, satisfaction (inthe from 1995 absence of career development). 0 Non-contingent bonuses 0 Non-contingent (supply problems (PSF bonuses mitigate bonus; IDQ,1994; specific labor-market GGEM, 2002) shortages, e.g. getting 97 Contingent bonuses as staff to work in performance incentive outlying Units). (PIQ, 1995; IDQ, 2000) Contingent bonuses only directly affect extreme cases of performance. Managers Managers are political Management is appointments chosen technocratic and well from career public integrated into SMS sevants Strategic Instruments Intended/possiblehealth Intenddpossible areas outcomes management impacts Performance m8 Planning * Informalor ad hoc 0 Widespread strategic- strategic planning planning problem- (e.g., GERUS, 1995) solving culture. Municipal Health Plan 2002-2205 andAnnual Operating Plan (POA). Management Non-enforceable Management contracting Management contracting: clarifies ContractsPOA priorities; identifies (2000), with c. 60 problems; and (with targets. IDQevaluation) 0 Monitoring of POA reinforces teamwork. 98 APPENDIX F. SelectedManagement Contract Targets byHealthDistrictin Curitiba, January December, 2004 - ~~ ~~ ~ Target Population Coverage Target Actual fulfillment (number) (percent) (number) (number) (percent) (=AxB) (=D/C) ~~ Coberturade ConsultasMidicas SantaFelicidade 183,152 150% 274,728 376,088 137% BoaVista 225,696 160% 361,114 418,020 116% Boqueir2o 192,858 150% 289,287 426,639 147% Port20 286,052 130% 371,868 464,966 125% Pinheirinho 161,40 1 150% 242,102 249,426 103% Cajuru 194,113 85% 164,996 170,802 104% Matriz 204,5 16 40% 81,806 80,783 99% BairroNovo 139,587 150% 209,38 1 355,080 170% Totauaverage 1,587,375 126% 1,995,281 2,541,804 127% CoberturadeProc. Odontolbgico1Habitante SantaFelicidade 183,152 150% 274,728 188,974 69% BoaVista 225,696 160% 361,114 247,419 69% BoqueirIo 192,858 150% 289,2 87 180,629 62% Portixo 286,052 130% 371,868 317,211 85% Pinheirinho 161,401 150% 242,102 250,064 103% Cajuru 194,113 85% 164,996 160,959 98% Matriz 204,5 16 40% 81,806 76,380 93% Bairro Novo 139,587 150% 209,38 1 170,669 82% ToWaverage 1,587,375 126% 1,995,281 1,592,305 80% Cobertura VacinaldeMenores de 1ano BCG - SantaFelicidade 2,726 2,78 1 102% BoaVista 3,314 3,318 00% Boqueir2o 3,211 3,014 94% Portixo 4,477 3,653 82% Pinheirinho 2,968 3,032 02% Cajuru 3,158 3,112 99% Matriz 2,163 2,956 37% Bairro Novo 2,789 2,6 17 94% Totauaverage 24,806 24,483 99% Coberturade VisitasDomiciliares SantaFelicidade 45,788 204% 93,408 235,095 252% BoaVista 54,424 204% 111,025 292,538 263% BoqueirIo 48,214 204% 98,357 264,758 269% Port50 71,498 204% 145,856 385,603 264% Pinheirinho 40,350 204% 82,314 310,518 377% Cajuru 48,528 204% 98,997 356,505 360% Matriz 51,129 100% 51,129 61,879 121% BairroNovo 34,897 204% 71,190 225,338 317% ToWaverage 394,828 191% 752,275 2,132,234 283% 99 Target Population Coverage Target Actual fulfillment (number) (percent) (number) (number) (percent) (=AxB) (=D/C) Coberturade NovasInscrigcjesde Gestantes SantaFelicidade 2,098 100% 2,098 1,979 94% BoaVista 2,554 50% 1,277 2,556 200% BoqueirIo 2,793 100% 2,793 2,661 95% Portzio 3,117 100% 3,117 3,295 106% Pinheirinho 2,284 100% 2,284 2,830 124% Cajuru 2,432 100% 2,432 2,693 111% Matriz 951 90% 856 874 102% Bairro Novo 2,447 100% 2,447 2,647 108% TotaYaverage 18,676 93% 17,304 19,535 113% Concentrag20de examesanti-hivpara 1OWo&s gestantes SantaFelicidade 1,979 1,75 1 88% BoaVista 2,556 2,060 81% BoqueirIo 2,66 1 2,267 85% Portzio 3,295 2,880 87% Pinheirinho 2,830 2,473 87% Cajuru 2,693 2,2 18 82% Matriz 889 684 77% Bairro Novo 2,647 2,397 91% TotaYaverage 19,550 16,730 86% 4nalisar 10Wodos dbitos infantis dentrodos critkriosp/am'lise SantaFelicidade 30 23 77% BoaVista 31 31 100% BoqueirIo 23 23 100% Port50 54 54 100% Pinheirinho 40 39 98% Cajuru 42 41 98% Matriz 25 24 96% BairroNovo 30 27 90% rotal/average 275 262 95% CoberturaTotalde Inscripio de Hipertensos SantaFelicidade 16,989 62% 10,533 10,243 97% BoaVista 20,856 50% 10,428 12,594 121% Boqueirso 19,715 50% 9,858 12,016 122% Portzio 20,643 70% 14,450 14,616 101% Pinheirinho 13,798 65% 8,969 10,405 116% Zajuru 17,256 53% 9,146 10,105 110% Matriz 12,443 35% 4,355 4,554 105% BairroNovo 13,623 100% 13,623 9,829 72% rotal/average 135,323 60% 81,361 84,362 104% Source: SMS 100 BIBLIOGRAPHY Abrantes, Alexandre. 1991. "Contracting Public Health Care Services in Latin America." In Alexander S. Preker and April Harding, eds., Innovations in Health Service Delivery: The Corporatization of Public Hospitals. Washington, D.C.: The World Bank. Ames, Barry and Margaret Keck. 1997. "The Politics of Sustainable Development: Environmental Policymaking inFour Brazilian States." Journal of Interamerican Studiesand WorldAfsairs, 39(4): 1-40. Baker, George P., Michael C. Jensen, and Kevin J. Murphy. 1988. "Compensation and Incentives: Practice vs. Theory," The Journal of Finance, 43(3): 593-616. Bertussi, Debora Cristina, Rosangela Scucato, Samuel Jorge MoysCs, Solange Maria Petris, and Vera Ldcia Martins. 1996. "Implanta@o do Projeto GERUS na Secretaria Municipal de Salide de Curitiba -Parani." RevistaEspappara a Sau'de,5(5). Brazil. Presidency of the Republic. 1995. "White Paper: Reform of the State Apparatus." Brasilia: Presidency of the Republic, Chamber of State Reform, Ministry of Federal Administration and State Reform. Bresser Pereira, Luiz Carlos. 1999. Reforma del Estado para la Ciudadania: La Reforma Gerencial BrasileAa en la Persnectiva Intemacional. Buenos Aires: Universidad de Buenos Aires/CLAD. Campbell, Tim. 2006. IPPUC: The Untold Secret of Curitiba: In-House Technical Capacityfor SustainableEnvironmental Planning, UrbanAge Institute, Draft 10/23/06. Cazura Xavier, Elziane, and Maria Emi Shimazaki. n.d. A Experigncia de Curitiba com o Prontudrio Eletrbnico - A Ousadiaem Inovar. httu://www.hu.ufsc.br/IX CIBS/trabalhos/arauivos/l6.~df Chalkley, Martin, and James M. Malcomson. 1996. "Contracts for the National Health Service." The Economic Journal, 106(November): 1691-1701. Conselho Regional de Medicina do Estado de Sao Paulo (Cremesp). 2002. "Mercado de Trhbalo Medico no Estado de SBo Paulo." SBo Paulo: Cremesp, Ndcleo de Estudos de SaGde Coletiva, UniversidadeFederal de Minas Gerais. Dercy Silveira Filho, Antonio, Luciano Ducci, Mariangela Galvao Simao, and Sylvia Palermo Gevaerd, eds. 2002. Oz Dizeres da Boca em Curitiba: Boca Maldita, BoqueirGo, Bocas Sauddveis.Rio de Janeiro: Centro Brasileiro de Estudos de Sadde. Ducci, Luciano, Maria Alice Pedotti, Mariangela Galvao Simao, and Samuel Jorge Moyses, eds. 2001. Curitiba: A Sau'dede BragosAbertos. Rio de Janeiro: Centro Brasileiro de Estudos de Sadde. Edwards, Nigel, Sylvia Wyatt, and Martin McKee. 2004. "Configuring the hospital in the 2lSt century." Policy Brief No. 5. European Observatory on Health Systems and Policies. 101 Estado de SZo Paulo. FundaSZodo DesenvolvimentoAdministrativo(FUNDAP).2004. "Por que OrganizaGdesSociais naSacde?' SP.gov, l(2): 10-17. Evans, Peter. 1997. Sustainability, Degradation and Livelihood in Third World Cities: Possibilities for State-Society Synergy. Paper Prepared under the auspices of the Research Group on States and Sovereignty for UN21Project, Conference on the GlobalEnvironment, New York, November 15, 1997. Fiszbein, Ariel, ed. 2005. Citizens,Politicians, and Providers: TheLatinAmerican Experience with ServiceDelivery Reform. Washington, D.C.: World Bank. FundaQio OswalsoCruz.n.d. "Sanitarista: ReformaSanitaria", Bibliotecavirtual SergioArouca. htt~://bvsarouca.cict.fiocruz.br/sanitaristaO5 .html Gnatek, Tim. 2003. Curitiba`s UrbanExperiment.December. httu://www.ubs.org/frontlineworld/fellows/brazi11203/ Harding, April, and Alexander S. Preker. 2003. "A Conceptual Frameworkfor the Organization Reforms of Hospitals." InAlexander S. Preker andApril Harding, eds., InnovationsinHealth Service Delivery: The Corporatization of Public Hospitals. Washington, D.C.: The World Bank. Holmstrom, Bengt, and Paul Milgrom. 1991. "Multitask Principal-AgentAnalyses: Incentive Contracts, Asset Ownership, and Job Design," Journal of Law, Economics, & Organization, 7(special issue): 2452. IMAP (Instituto Municipal de Administra@o Pliblica). 2005. Avaliapio das Politicas Pu'blicas Municipais de Curitiba - 1997a 2004. httu://www.imau.curitiba.Dr.nov.br/ Jones, L.R. 2004. "`New Public ManagementHas Been Completely Discredited,Thank God,"' InternationalPublic ManagementReview, 5(2): 148-172. Kohn, Alfie. 2002. "Why Incentive Plans Cannot `Work." Harvard Business Review on Compensation,pp.29-49. HarvardBusinessReviewPublishingCorporation. Marsden, David. 2003. "Renegotiating Performance: the Role of Performance Pay in Renegotiating the Effort Bargain." Discussion Paper No. 578, Centre for Economic Performance,LondonSchoolof Economics, London. McKibben, Bill. 1995. "Curitiba." Hope, Human and Wild. Boston: Little Brown and Co. [excerptedin Yes! Magazine, Summer 19991. Milgrom, P. andJ. Roberts. 1992.Economics, OrganizationandManagement.EnglewoodCliffs, N.J.: Prentice-Hall. Miller, Gary J. 1992.ManagerialDilemmas. Cambridge: CambridgeUniversityPress. MoysCs, Samuel Jorge. 2000. "Oral Health and Healthy Cities: an Analysis of Intra-Urban Differentials in Oral Health Outcomes in Relation to `Healthy Cities' Policies in Curitiba, Brazil. Thesis submitted for the degree of Doctor of Philosophyof the University of London 102 Department of Epidemiology and Public Health, The Royal Free and University College MedicalSchool, UniversityCollegeLondon. New Zealand, Parliamentary Commissioner for the Environment. 2002. Showing the Way: Curitiba: Citizen City. Wellington: Parliamentary Commissioner for the Environment. httD://www.ece.govt.nz/reeorts/allreeorts/l 877274 06 2.edf North, Douglas. 1990. Institutions, Institutional Changeand Economic Performance. New York: Cambridge UniversityPress. OECD. 2005. "Performance-related Pay Policies for Government Employees." OECD Publishing. Porto, Dora. 2001. Programa Mcie Curitibana: Ciclo de Premiagcio 2001. In HClio Batista Barboza and Peter Spink, eds., 20 Experitncias de Gestcio Pliblica e Cidadania. Programa Gestiio Pdblicae Cidadania, FundaGiioGetdlio Vargas, Siio Paulo. Rabinovitch, Jonas, and John Hoehn. 1995. A Sustainable Urban Transportation System: The SurfaceMetro" System in Curitiba, Brazil. Washington, D.C.: EPAT/MUCIA. Rangan, V. Kasturi. 1994. "The Aravind Eye Hospital, Madurai, India: In Service for Sight." Harvard Business School Case 593-098. Reilly, BrendanM.2004. "The essenceof EBM", BMJ, 329:99 1-992.htte://bmi.bmiiournals.com Sant'Ana, Ana Maria, Walter W. Rosser, and Yves Talbot. 2002. "Five years of family health care inSiio JosC." Family Practice, 19: 410-415. Secretaria Municipal da Sadde de Curitiba. 2002. 10Anos de PSF Curitiba: A Histdria Contada Por QuemFaz. Secretaria Municipal de Sadde de Curitiba. 2004. Cademo de Resumos: 8" Encontro de Salide Coletivade 13a 16de Julho de 2004. Shepherd, Geoffrey. 2003. "Understanding Public Organizations: an Aid to Reform in DevelopingCountries." mimeo. Simon, Herbert A. 1987. "Rationality in Psychology and Economics." InRobin M. Hogarthand Melvin W. Reder, eds., RationalChoice: The Contrast betweenEconomicsand Psychology. Chicago: Universityof ChicagoPress. Sojo, Ana. 1996. "Potential andLimits of HealthManagement Reform in Chile," CepalReview, 59(August): 127-145. Thompson, Grahame. 2003. Between Hierarchies and Markets: TheLogic and Limits of Network Forms of Organization. Oxford: OxfordUniversityPress. Vaitsman, Jeni. 2001. "Gerencialismo, cultura e expectativas entre servidores pdblicosde sadde." RevistadeAdministrapio Pliblica, 35(1): 29-47. 103 Venetikides, Cristiane H., Danuza M. Maceon, Maria Alice C. Pedotti, and Samuel Jorge Moyses, eds. 2001. Sau'deMental em Curitiba, Centro Brasileirode Estudosde Saude, Rio de Janeiro. Wilson, James Q. 1989.Bureaucracy: What GovernmentAgenciesDo and Why They Do It. New York: Basic Books. WorldBank. 1995.Bureaucratsin Business. Washington, DC:The World Bank. WorldBank. 2002.Brazil -Municipal Education:Resources, Incentives,and Results. ReportNo. 24413-BR. WorldBank 2003. Brazil -Decentralization of Health Care in Brazil. ReportNO.24416. WorldBank2004. WorldDevelopmentReport 2004: Making ServicesWorkfor Poor People. Washington, DC. WorldBanWOxfordUniversityPress. WorldBank2005. Brazil -Addressing the Challengeof Non-CommunicableDiseases in Brazil. ReportNo. 32576. WorldBank. 2006a(forthcoming).In search of Excellence: StrengtheningHospital Performance in Brazil. WorldBank. 2006b (forthcoming). Brazil -TheLevel and the Quality of Public Expenditures and Implicationsfor Economic Growth: Public ExpenditureReview (tentative title) WorldBank. 2006c (forthcoming). Brazil -Managing Health Expenditures in the Brazilian Federation: Public Expenditure Tracking Survey (tentative title). 104