Mexico - Health Reform Under the 1995 Social Security Law : Issues and Actions

Document of The World Bank Confidential Report No. 15374-ME (Revised) Mexico: Health Reform under the 1995 Social Security Law: Issues and Actions October 15, 1996 Human and Social Development Group Country Department H F C Latin America and the Caribbean Regional Office This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. ACRONYMS AND ABBREVIATIONS CPI Cargo Performance Index DRG Diagnostic Related Groups FUNSALUD Mexican Health Foundation (Fundacion Mexicana para la Salud) GDP Gross Domestic Product GOM Government of Mexico GNP Gross National Product IDB Inter-American Development Bank IMSS Mexican Social Security Institute (Instituto Mexicano del Seguro Social) MMI Medical and Maternity Insurance MS Minimum Salaries OECD Organization for Economic Cooperation and Development PAHO Panamerican Health Organization PCP Primary Care Physician SSA Secretary of Health (Secretaria de Salud) SHCP Secretary of Finance and Public Credit (Secretaria de Hacienda y Cridito Ptblico) SMGDF Minimum Salary in the Federal District WHO World Health Organization Mexico: Health Reform under the 1995 Social Security Law: Issues and Actions CONTENTS EXECUTIVE SUM ARY ......................................................................................................................... i 1. BA CK G RO UND ................................................................................................................................ 1 2. MAJOR ELEMENTS OF THE PROPOSED MEXICAN REFORMS IN HEALTH CARE D ELIV ERY ........................................................................................................................................2 3. CONVERT PAYROLL TAX TO EMPLOYER/GOVERNMENT CONTRIBUTIONS ................3 4. INTRODUCE PUBLICLY SUBSIDIZED HEALTH INSURANCE FOR THE SELF EMPLOYED AND INFORMAL SECTOR WORKERS.................................................................8 5. OPT-OUT PROVISION (REVERSION DE CUOTAS) ................................................................11 6. DECENTRAIZATION .................................................................................................................. 12 7. EFFICIENCY IMPROVEMENTS IN SERVICE DELIVERY......................................................13 8. DISCUSSION AND SUM M ARY ..................................................................................................... 1 BIBLIOGRAPH .................................................................................................................................20 This report was prepared by Maureen Lewis (Principal Human Resources Economist). The comments of C. Baeza, F. Lysy, Z. Qureshi, J. Newman, G. Schieber, J. Londono, J. Frenk, A. Fidler, and C. Hamann are gratefully acknowledged. The appendix was prepared by John Newman (Senior Human Resources Economist. The Department Director is Mr. Olivier Lafourcard and Julian Schweitzer is the Human and Social Development Group Manager. Tables: 1. Comparison of Current and Revised Financing Scheme for Medical and M aternity Insurance ............................................................................................... 4 Figures: 1A-B. Current Employer, Government and Employee Contributions for IMSS for Medical and Maternity Insurance, by Number of Minimum Salaries.................5 2A-B. Employer, Government and Employee Contribution to IMSS for Medical and Maternity Insurance under the Payroll Reform Scheme, by Number of M inimum Salaries for YearOne ............................................................................ 6 3.. Trends in IMSS Revenue and Expenditures per Insured Worker for Medical and M aternity Insurance, 198811-198 .....................................................................7 4. Preliminary Projections of Total Collection and Government Contribution for HMI as % of GDP ................................................................................................ 17 Annexes: 1. Tables on Co-payments and Payment Systems in OECD Countries..................21 2. Health Sector Reform Studies in Mexico (1993-1995) and Bank Comments.....23 Executive Summary Health care reform is sweeping the developed and developing countries, and it is clear that the process is difficult and involved. However, that experience is invaluable for Mexico, as it provides a laboratory of trends and lessons on which to draw. One clear outcome is the fact that there is no single solution to how health care should be delivered and financed, only options that differ in their impact and effectiveness. This note provides a preliminary evaluation of the Mexican reforms in IMSS within the context of the ongoing OECD and Latin American reforms. The major elements of the reforms are as follows: * convert payroll tax to indexed employer/GOM contribution up to 3 minimum salaries (MS) with additional employer/employee contribution over 3 MS- * establish new form of publicly subsidized health insurance; * expand applicability of the reversion de cuotas (opt out provision) for employers who provide health care to employees, * decentralize (desconcentraci6n) IMSS, and merge and decentralize SSA and IMSS Solidaridad; and * efficiency improvements through out-sourcing and choice in family physician. While each of these provide new approaches to improve equity, enhance patient satisfaction, improve government responsiveness, reduce employer financial burdens, and raise efficiency in service delivery, there are some serious implications for the Mexican government in achieving these goals. Fundamentally, the adjustments initiated for each of the identified problems were addressed individually, which limits the ability of the government to integrate the incentives facing employers, employees and government in the financing and delivery of health care. However, given that the nature of implementation is not yet defined, these constraints can be addressed. First and foremost, the fiscal implications for Hacienda of the reversi6n de cuotas, new health insurance and new payroll tax arrangements are very serious, although the exact burden is not yet clear. The reversi6n de cuotas runs contrary to the trends in the other OECD countries (outside of the US) as it fragments risks and leaves the government with the highest health risks and most costly population. Preliminary analysis suggests that GOM projections of the fiscal effects of the new health insurance plan and the payroll contribution responsibility shift may be overly optimistic given experiences elsewhere and the expected Government contribution under the new law. Crude estimates based on characteristics of the reforms suggest that the additional cost of the payroll tax scheme to the government is closer to 0.6% of GDP in year one, and will rise in subsequent years. Indeed, it is the medium and long term costs that are of concern, not that of the immediate reform. To improve the success of the reforms, the following should be addressed: ii * simulate alternative scenarios of tax levels, compliance and revenue to assess fiscal impact * revise cost estimates under different scenarios, based on sound actuarial analysis, better estimates of costs of lost revenues and patients, and realistic projections of government contributions; * determine how the system will be structured and managed regarding risk pooling, revenue collection, resource allocation, resource management, etc. * set annual budges and expenditures for IMSS and ensure accountability of IMSS; * estimate service delivery and administrative costs, and establish cost containment throughout the system; * pilot all new programs, both of a financial and service delivery nature, and adjust as necessary before expanding nationally. * strengthen demand analysis to predict patient behavior and cost effectiveness of service delivery points. Decentralization (desconcentraci6n) is in a nascent stage, but it should be kept in mind that it is not a panacea nor strictly a reform. Indeed, experience in other Latin American countries suggests that decentralization is difficult and only partially successful even with careful planning, adequate resources, and strong institutions. Clarity in rules and responsibilities, defined fiscal roles, reliable financial transfers, involvement and training of lower levels of government, and accountability at all levels of government are essential for success. Inadequate attention to any of these have led to partial decentralization, worsening services, and political difficulties in other countries in the region. The service delivery improvements proposed have considerable potential, but such initiatives require careful definition to ensure the appropriate incentives for providers, patients and suppliers. These approaches have been both successes or failures depending on the chosen incentive structures. In addition to the reform-specific suggestions, additional attention is required on general topics to strengthen the health system and the IMSS reforms: * develop regulations and enforcement for private payers (including insurance companies, including indemnity, pre-paid systems and in-company services), and public and private providers (e.g., hospitals, physicians, medical schools) * initiate cost containment through multiple avenues, including expenditure ceilings, accountability of providers, co-payments by patients, * rely on demonstration pilots for key initiatives to avoid visible failures on a broad scale. Ill Health care reform is challenging and risky, but it is apparent in many contexts that without wholesale changes, improvements in equity, and in the effectiveness of financing and delivery, cannot be attained. The World Bank is prepared to assist the Government through technical assistance, existing loans, sector work or new lending. The time is opportune to pursue reform issues, and the Bank is available to participate as needed. 1. This note evaluates the health reforms of IMSS within the context of the ongoing OECD and Latin America reforms. It reviews the legislation, briefly analyzes the reforms, discusses likely impacts of these changes on the Government's fiscal burden and on the health care system, and suggests areas for greater consideration and analysis. In this process, the possible pitfalls of pursuing particular policies are emphasized to offer the government full information for decision making. 1. BACKGROUND 2. The problems confronting IMSS and the Mexican government in the health sector are serious and require equally, dramatic reform. The new initiatives, particularly the 1995 Social Security Law, address the many criticisms and weaknesses of the IMSS health services, and their implementation have the potential for making a more efficiency, effective and fair health system. However, the complexity of the sector and the multiple fronts for change make it difficult to shape an integrated program with all the needed incentives. 3. As Mexico launches a reform, the experiences from other countries can be instructive. Health reform has swept through the other OECD countries over the past decade and a half, with major restructuring of health systems and alternative financing arrangements introduced throughout. In the last five years, much of the Latin America region has embraced the need to reform health systems. Although very different in income and institutional capacity, the lessons from the OECD offer valuable experience for Latin America in general and Mexico in particular. 4. It has become apparent through multiple efforts in the OECD experience that there is no single answer to the issue of how to deliver and finance health care; every solution engenders problems. Despite this, there is a convergence in how health systems are structured and how they operate that attempts to minimize or compensate for difficulties while emphasizing characteristics that have been shown to be effective. This does not suggest that all national health systems are identical, but that individual elements are becoming more similar across countries (OECD, 1992, 1994). These will be elaborated below within the context of the proposed Mexican reforms. 5. The major goals of health reform are varied, but can be summarized as shown in Box 1. While each of these objectives is desirable, it is essential that countries rank them, for it is not possible to achieve all objectives due to their inherent incompatibility. For example, increased equity raises costs; efforts to control costs is most effectively achieved through capping expenditures, but this often reduces quality and choice. Hence the need for priorities within the context of what is affordable and feasible. 6. Country examples show the importance of tradeoffs in health care objectives. The US has designated consumer choice and provider autonomy as paramount. The result is high costs, high quality, considerable choice and powerful providers in the US, but uneven access, insufficient protection for low income households, and limited cost containment. In contrast, the UK has effectively contained expenditures, provides basic services for all, and has achieved equal access; however, patient satisfaction, quality of services, and choice are restricted. This example provides a sense of the two extremes in health care: one driven by consumer choice and multiple provider options, and the other by strong government controls and oversight. It demonstrates how 2 prioritizing ultimately defines the health system. An issue that deserves to be highlighted, is that the driving forces behind universal care commitments, in all but the US, has been the desire to spread risk evenly and provide universal access. These are the basic tenants of European health reform. Box 1 Major goals of health reform in the OECD countries * adequate services for all * equity of access * protection of income in low income households * macro-efficiency (cost and expenditure containment) * micro-efficiency (patient satisfaction, cost conscious consumers, maxinize health outcome) * consumer choice * provider autonomy * quality 7. A key issue underlying many of the reform agendas around the world are the national concepts about the role of the state in both financing and delivering health care services. These decisions are partly historical and cultural, and reflect something about the populations' perceptions of government. For example, in the US, government control of health care is unacceptable, in most of the other OECD countries it is considered essential (Blendon et. al., 1995). The decision of the role of the state is key to how to address the seven goals in Box 1, and the subsequent decisions that are therefore implied. However, there is no consensus as to the best public-private mix. 8. The objective of any country seeking to reform health care must be to adapt the existing system to the priorities and demands of the society, taking into account the above elements as well as the public-private mix. Available resources and institutional capacity also play a defining role. Thus, the experience of OECD countries are offered within the context of Mexican circumstances. The discussion below is structured to emphasize these linkages. 2. MAJOR ELEMENTS OF THE PROPOSED MEXICAN REFORMS IN HEALTH CARE DELIVERY 9. Much of the IMSS health reform is linked to the 1995 Social Security Law, but not all. Drawing on the material provided to the World Bank, the following represent the major components of the proposed reforms': * Convert payroll tax to an indexed, combined employer and GOM contribution of 27.8% of the minimum salary in the Federal District (SMGDF) per worker per month up to three The reforms are drawn from Ley del Seguro Social; "Hacia el Fortalecimiento y Modernizaci6n de la Seguridad Social" de IMSS; "Estimaci6n de Costos Fiscales Derivados de la Reforma a la Ley del Seguro Social" del Gobierno Mexicano, 30 de enero de 1996; "Overview of the Reform of the National Health System," Working Document, Secretaria de Salud (SSA), 1995. 3 minimum salaries (MS) (equivalent to N$84 in early 1986); above three MS a payroll tax kicks in (75% employer, 25% employee). * Establish new form of affordable, publicly subsidized health insurance (P$135 per month, with GOM subsidy per month equivalent to 13.9% of one minimum salary; both are indexed to the CPI). * Expansion of the reversi6n de quotas (opt-out provision) for employers who provide health care to employees. * Decentralize IMSS into 7 regional semi-independent units. * Merge and decentralize to the states SSA and IMSS-Solidaridad, and introduce block grants at some point. * Efficiency improvements through out-sourcing ancillary services; introducing free choice of family physician in private practice. 3. CONVERT PAYROLL TAX TO EMPLOYER/GOVERNMENT CONTRIBUTIONS 10. The financing of the Medical and Maternity Insurance (MMI) has been revamped under the new Social Security Law of 1995. Table 1 summarizes the current and revised financing schemes. IMSS estimates that the contributions paid by employers and employees would decline for all but those earning one minimum wage. Overall combined employer/employee contributions would decline by 33%. If current IMSS revenues remain constant, then the federal government will need to increase its contribution by the same percentage to maintain existing levels of revenue. The contributions of employers, employees and government under the past and current financing scheme are depicted in Figures 1 and 2, respectively. The current projections are based on the provisions in the Social Security Law. 11. The law states that government is expected to increase its contribution to IMSS from general revenues, but does not define the parameters of that transfer. In Article 106(iii) the law states that the federal government will finance 13.9% of one minimum salary for the population enrolled in IMSS, indexed to the consumer price index, and contribute to the sick leave fund (prestaciones economico). How much the Secretary of Finance and Public Credit (SHCP) will be expected to contribute is not clear due to (a) ambiguities in the law; (b) uncertainty as to company compliance with the law; and, (c) unknown costs of the existing IMSS system under a modified beneficiary population and a new subsidized insurance program. However, the budget is carefully shielded from inflation through systematic indexation of all transfers. 12. Article 108 states that the federal government will cover the annual budget submitted by IMSS. The parameters for the budget are not specified, nor does it define how budget decisions will be reached. It is unclear what recourse is available to IMSS in recession periods when public resources are tight. Similarly it is not obvious if SHCP has any role in negotiating or approving a budget for IMSS Medical and Maternity 4 Table 1: Comparison of Current and Revised Financing Scheme for Medical and Maternity Insurance. :Current Financing -Revised Financing* Compulsory Payroll Tax up to 12.5% Fixed CPI Indexed Contribution for 27.7% 25 minimum salaries workers earning 3 minimum salaries or less * employer 70% * employer 50% * employee 25% * government 50% * government 5% For workers earning over 3 minimum salaries, workers would pay 8% of the difference between workers' wage and 3 minimum salaries: * employer 75% * Employee 25% Compulsory 1% of Salary for Sick Leave (Prestaciones economico): * employer 70% * employee 25% * government 5% Insurance. This becomes important if meeting presented budgets jeopardizes achievement of the federal government's fiscal targets. These issues emerge most clearly if the IMSS budget rises sharply 13. On the second point, although evasion was one of the reasons for the reforms, current and future revenues cannot be predicted with great certainty because of noncompliance. While encouraging firms to take advantage of the opt out provision in the existing law, and reducing payroll taxes is likely to promote compliance, evasion is likely to remain to some degree. 14. Finally, the last point is developed further in discussions below regarding risk- pooling and the impact of the new voluntary insurance program. Those assessments suggest that the reforms are likely to raise the overall costs of the system, but it is difficult to know by how much. As Figure 3 indicates, IMSS has been in limited but chronic deficit over the past 15 years. Part of the deficit was financed through mingling of pension and health revenues. Under the new regime, this is no longer feasible. If resources are inadequate, coverage and quality are likely to suffer. Hence it is key to both more precisely measure the fiscal impact, and to address IMSS delivery issues, since it currently has few incentives to control expenditures, and its benefits are open ended. 5 Figure lA: Current Employer, Government and Employee Contributions for IMSS for Medical and Materity Insurance, by Number of Minimum Salaries $ 1406 1200 4 Employer 1000 - Employee Government 800- 600 400 - a O 200 a 01 1 MinmuSalares 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Figure 1B: Current Employer, Government and Employee Contributions to IMSS for Medical and Maternity Insurance, by Number of Mi0num Salaries. 90% 80% 70% O Government 60 % Employer 40% 20% 10% : 30% ål)-),si 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Minimum Salaries 6 Figure 2A: Employer, Government and Employee Contribution to IMSS for Medical and Maternity Insurance under the Payroll Reform Scheme, by Number of Minimum Salaries for Year One 80% 70% 60% -5- Enployer 40% a Employee 30% Government 20% - *~a~aaa 10% . 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14151617 18 19 20 21 22 23 24 25 26 27 Minimum Salaries Figure 2B: Distribution of Employer, Government and Employee Contributions to IMSS for Medical and Maternity Insurance under the Payroll Scheme, by Number of Minimum Salaries Year One. 100% 90% . 80% 70% 60% Govemment 50% 5 *Employee . Employer 40% 30% 20% . 10% or Minimum Salaries 1 3 5 7 9 11 13 15 17 19 21 23 25 27 7 15. Topics that need to be addressed by the Mexican Government to maximize the successful execution of this change include: * estimates of expected lost revenues and patients, through assessment of the number and types of firms that want to opt-out and the actuarial profile of the employees; * careful simulations of alternative scenarios of compliance and resulting revenue to evaluate the range of fiscal impact; * simulation of general revenue under different employment and economic growth assumptions; * assessment of the impact of the new scheme during the first year to monitor effects of the change. 16. Initiatives to reduce the costs of the program include: * establish mechanism that set annual budgets and expenditures and hold IMSS accountable for delivery and for remaining within budget; * establish ceilings or limitations on benefits and consider co-payments and other costs on expenditure levels; * establish incentives for cost containment throughout the system to control the volume of services provided. Figure 3: Trends in IMSS Revenue and Expenditures per Insured Worker for Medical and Maternity Insurance, 1981 - 1985 (Nominal Pesos) N$ 2500 2000 1500 Revenue per Worker -W- Expenditure per Worker 1000 500 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 Year 8 4. INTRODUCE PUBLICLY SUBSIDIZED HEALTH INSURANCE FOR THE SELF EMPLOYED AND INFORMAL SECTOR WORKERS 17. Every OECD country has some form of publicly subsidized health insurance, but they vary in the comprehensiveness of the program and the eligibility of the population. In most of Europe these take the form of compulsory payroll deductions (UK, France, Japan, Spain), employer contributions, and government subsidies (from general revenue) that are administered either by the government (Sweden, UK) or government regulated private entities (Holland, Germany). Other countries finance entirely through general taxation (Canada, Denmark). These all provide universal coverage. In the US and Switzerland insurance is largely private, although in the US, subsidized health insurance is available to low income families (Medicaid), or the elderly and disabled (Medicare), with the former financed completely through subsidies and the latter through compulsory payroll taxes, general revenues and co-payments (OECD, 1994). Moreover, in many countries, like the US and Brazil, tax deductions for private insurance offer implicit subsidies for health insurance coverage. 18. Governments provide health insurance for all, to ensure access, spread risk fairly, protect income, and regulate the sector through power over the level of expenditure and internal prices. Government control over the sector either through direct management or regulation (Germany, Canada, Holland) spreads risk across the entire population, which makes risk more predictable and manageable, thereby allowing better financial management. In Europe, funds are raised from a broad base and the commitment is to the entire society, to ensure equity, adequate services for all and protection of low income households. In contrast, the US Medicaid system is a decentralized, open ended entitlement that matches federal funds with state resources to finance health care for the poor (leaving out the near poor). State matching finances generally limit the extent of the program, although the current debate in the US Congress proposes to cut the program and shift to block grants to states for multiple programs to meet federal budget reduction targets. 19. The financial risk of a subsidized voluntary health insurance system for the near-poor and middle classes in Mexico lies with the fragmentation of risk, the likely behavior of the beneficiary population, and the resulting costs of the system, the same reservations that led to the ultimate structure of most national OECD insurance financing schemes. These are exacerbated by inherent technical and allocative inefficiencies in the current service delivery structure. 20. The response of the population to an affordable, subsidized insurance operated by IMSS is hard to predict with precision because the level and nature of beneficiary demand is unknown. Based on experiences in the OECD, the elasticity of demand for health services -- the change in the demand for services due to a change in price -- is greater than one, indicating that demand is likely to be high. Given that the costs under the current system are already more per worker than the proposed price of the new family insurance, the insurance program would be expected to be an attractive option for families, but to entail subsidies from the federal government to make up the difference between IMSS costs and the price offered low income families. The higher the demand, the larger the budget, the higher the transfers required from SHCP. Whether 13.9% per family is sufficient to cover costs given the modest premium is not clear. 9 21. The subsidized premium of P$135 is one fourth what a family in the informal sector would currently have to pay to enroll in IMSS, and represents 8.5% of current average household expenditures on health-related goods and services (FUNSALUD, 1994). Even with the SHCP subsidy, the payment to imss is below current expenditures per worker. The voluntary insurance program is targeted at those who have not had the benefit of service access in the past. The new enrollees are likely to be a higher risk population because they are sicker, have received little or no care in the past, and therefore will be more costly to treat than the traditional IMSS beneficiary. A second risk for the government is the random entry and exit of the population from the insurance program. There is an incentive for individuals to enroll when there is anticipation or existence of illness, and to withdraw when health service needs are minimal. Moreover, for those who are enrolled in the voluntary insurance program, the incentive is to overconsume (moral hazard) unless there is some method to discourage unnecessary consumption (Manning et. al., 1988). These pose serious financial issues for the Government. 22. As it is currently structured, there are inadequate safeguards for controlling either the costs or the volume of services provided, other than long waiting times in IMSS facilities, and lack of key inputs, both of which reduce quality. Without these, it is difficult to maintain desired expenditure levels. Incentives are required on the consumer and provider side to contain expenditures, including: (1) incentives for controlling volume among providers, and (2) price constraints for patients that discourage over-and unnecessary consumption of services. Unless these kinds of controls are in place, the costs of this program are likely to be high and to grow exponentially given the sicker target population, the lack of (dis) incentives within the system to promote rationale use of resources, and the absence of incentives for containing costs and volume. 23. The OECD countries have addressed these inherent problems through various measures. Use of global budgets in hospitals (UK, France, Sweden) has placed a cap on total hospital expenditure thereby restricting price or volume increase. In Latin America, most countries have ceilings for some providers, usually hospitals, but these are flexible and are often augmented when funds run out, thereby eliminating the incentive to stay within budget. Diagnostic Related Groups (DRGs) are increasingly being adopted in Europe to allocate resources (France, Spain, Sweden, Holland, Czech Republic), and to monitor hospital operations. DRG-type arrangements for budget allocation have been adapted in Chile, and Costa Rica is designing a DRG system for its hospitals. DRGs are an unbiased means of determining how to use resources and to monitor what the health sector is purchasing, which provides the information base for improving efficiency and productivity. 24. User charges are increasingly being used to reduce or raise the cost to consumers of certain services and to regulate demand. However, in virtually all OECD and LAC countries such policies include sliding scales or waivers for low income households.2 Outpatient services and drugs are the most likely to carry fees, although inpatient care, where costs are highest, are the least likely to require co-payments. The trend, however, is to greater cost sharing in general. This is also a trend in Latin America, where only a handful of countries outside of Mexico (Brazil, Costa Rica and Ecuador) do not require some form of co-payment 2 Table A-I in Annex 1 sununarizes co-payment policies in the OECD.. 10 25. The uncertainty of the costs of MMI under the new law is particularly worrisome as Mexico's fiscal circumstances are already fragile. If costs rise sharply under the new law it could place constraints on government expenditures. How this will affect IMSS is not clear. Although more flexible and more affordable for employers and employees, relying on general revenues generally engenders risk. 26. The trend in financing in most of Europe is toward earmarked taxes, since general revenues entail an annual competition for funds and budgets are uncertain. A notable exception is the US federal program for the poor (Medicaid) that is co-financed with states. State resource levels fluctuate, and strict controls on eligibility and coverage from the federal government make this an often underfunded program of lower than average quality. It may be more difficult to ensure resource levels when financing only affects a segment of the population rather than the majority or the entire population as it does in most of the OECD countries. 27. In Brazil, the merging of social security and the Ministry of Health ultimately led to the exclusive use of payroll revenues to finance pensions that left health dependent on general revenues, which fluctuate annually as does government commitment to health sector financing. Relying on annual budget agreements has proven to be risky. In the Mexican case there is a partial move in this direction as general revenues will be relied upon to supplement the payroll tax that is being reduced, and to presumably replace the pension revenue previously applied to health investments. 28. Topics that need to be addressed by the Mexican Government to maximize the successful implementation of this reform include: * undertake demonstration projects in one or two states to determine how actual costs deviate from estimated cost, what the administrative costs are likely to be, and evaluate how the program functions under existing incentives; * determine how the system will be structured and managed regarding risk pooling, revenue collection, resource allocation, resource management, etc., and vet these within the government; * re-estimate costs to Hacienda under various scenarios; * undertake a thorough demand analysis to assist in predicting patient demand for new insurance and current health seeking practices (can be done by piggybacking annual household surveys and focus group analyses); * undertake a demonstration project with alternative delivery modes (pre-paid HM4O- type arrangements have been mentioned, but not consistently, however, these only work under certain conditions and require a clear regulatory and enforcement framework). 29. Initiatives to reduce the costs of the program include: * define eligibility so as to require vesting (e.g., minimal pay-in period prior to receiving benefits) or some other form of (semi-) permanent enrollment; * define benefit ceilings and restrictions to emphasize prevention and catastrophic care and limit benefits; 11 * move all possible services to outpatient settings (e.g., cataracts, minor surgery); * determine actuarial soundness of projections in order to control benefit eligibility; * introduce a sliding scale of co-payments to encourage and discourage, respectively, inexpensive and costly (especially inpatient care) services. 5. OPT-OUT PROVISION (REVERSION DE CUOTAS) 30. The benefits of this proposal are that the government would not be responsible for the financial management and provision of health services for a segment of the population, and the evasion of IMSS payroll taxes could be resolved by placing responsibility for financing and provision with employers, effectively broadening the existing source of health care coverage. Moreover, quality and choice will likely be enhanced for the opted out sector possibly at a lower cost to them. 31. The experience in the OECD countries, Brazil and the Southern Cone countries, however, is that those who want to opt-out tend to be the population with the lowest risk of poor health, that is, those with high concentrations of the young and healthy. If that group leaves, the costs of the system will rise due to the concentration of the high risks, chronically ill and low income populations in IMSS (adverse selection), and these groups represent the most expensive members of society. Effectively, health care costs are shifted towards government instead of being spread across the healthy and the sick, as is the practice in most of the OECD where risks are pooled across the entire population. Even in the US, Medicare is mandatory for all workers. 32. Some countries either allow companies to buy their way out of the social insurance system through reduced fees, or more commonly, simply make companies provide double coverage (e.g., the UK, or Brazil where 25% of the population has double coverage) (World Bank, 1994). Maintaining the higher income, healthier populations in the social insurance system ensures a basic level of quality in IMSS, a benefit lost if the most influential and articulate populations with the highest expectations are allowed to completely leave the system. 33. A somewhat related experience of relevance to Mexico is that of the US Veterans Administration Hospitals. Technically they cover all veteran health care needs. A study commissioned by the Veterans Administration concerned about the surge of demand with the aging of World War II veterans indicated, however, that only about 4% of all eligible veterans relied on VA facilities, preferring to use public or private insurance or their own resources. This pattern suggests that public facilities are often not the preferred option if other forms of financing and delivery are available. This experience may be instructive in what IMSS enrolled firms may choose to do when given an option. It also suggests the need to consider the more radical option of separating the financing and delivery of health care services, and devolving direct IMSS provision to regulated private providers. 3 Part of the Medicare program does have an opt out clause for the portion of the program that covers outpatient services (Part B) and for which the beneficiaries pay. However, given the costs and benefits, this clause is rarely selected, and the bulk of Medicare (Part A) is compulsory. 12 34. Topics that need to be addressed by the Mexican Government to maximize the effectiveness or success of the proposal: * revise cost estimates under different scenarios; * allow IMSS to collect a partial contribution from companies that opt-out commensurate with their population size, and better health risks; * require companies that opt-out to finance a defined set of minimal services for their employees, including catastrophic care; fine companies that do not comply; * develop regulations for private insurance companies, including indemnity, pre-paid systems and in-company services; * develop effective enforcement mechanisms for these regulations; * develop regulations for providers (hospitals, physicians and nurses) in conjunction with private associations to set accreditation standards for IMSS and other public and private providers (licensing of physicians and nurses; accreditation for medical schools and hospitals) and jointly enforce standards (i.e., provide public warnings, remove accreditation, deny licenses). 35. Information needed for improved policy formulation include: * number of firms (and employees) that would opt out of the system; * estimated revenue and enrollee losses; * review the experience of firms that have historically been allowed to opt-out (e.g., banks, and firms in Monterey). 6. DECENTRALIZATION 36. Decentralization or "deconcentration" would take three forms, the decentralization of IMSS into seven semi-autonomous units, the merging of IMSS-Solidaridad and SSA with block grants to states, and the provision of a basic package of services for the indigent by SSA. All three would entail expanding capacity at lower levels and realigning functions and responsibilities commensurately. It should be kept in mind that decentralization in and of itself does not solve inherent problems. Indeed it can exacerbate them. However, an entity as large and complex as IMSS, whether public or private, is probably more effectively run on a more decentralized basis. 37. Decentralization occurs in all countries at some level, although large countries (e.g., US, Canada, Argentina, Brazil, South Africa) logically tend to rely more on decentralized systems for size and heterogeneity reasons. Recent efforts at decentralization have occurred in Argentina, Bolivia, Brazil, Chile and Colombia in the region, and is being established in Russia. Lessons are limited. Colombia decentralized effectively to the larger departments, but smaller ones lagged. Moreover insufficient budget transfers in some areas severely constrained actual decentralization. In Brazil, finances were decentralized, although transfers were often received late in the fiscal year, but roles and responsibilities have remained ambiguous, leading to overlapping functions, infighting among different levels of government and no accountability. The longest political battles have been over the rationalization of all provider salaries at the three levels of government 13 (federal, state and municipal), to conform with the highest salaries. In short, it has led to chaos that is only slowly dissipating (World Bank, 1994). 38. The potential benefits from decentralization in Mexico are probably high, but it is key that decentralization be planned and implemented in a systematic fashion, that roles and responsibilities are clear and allocated rationally (i.e., that federal roles are not passed to states that cannot effectively carry them out), that promised funds are available in a timely fashion, and that the decentralized units be trained in situ to handle the changes and responsibilities. The inter- governmental fiscal relations, responsibilities at all levels and accountability are fundamental to successful implementation of a decentralization policy. Finally, monitoring of state/unit progress and performance is key to effective implementation, and has been lacking in the many efforts throughout Latin America. Poorly implemented decentralization carries heavy political costs and can exacerbate rather than solve problems. 39. Topics that need to be addressed by the Mexican Government to maximize the possibility of success are indicated below. These apply to all three decentralization plans: * pilot decentralization in two or three states to evaluate roles and responsibilities, * establish norms and procedures for states in resource allocation and delivery; * phase-in decentralization slowly to ensure capacity at the state /unit level; * develop management and administrative capacity at lower levels of government establish norms and responsibilities to regulate relations between levels of government; * determine financial and fiscal implications based on demonstration projects; * identify legal issues and address these to ensure smooth implementation; * establish and enforce accountability at the different government levels; * ensure state, union and other stakeholder involvement in planning and implementation. 7. EFFICIENCY IMPROVEMENTS IN SERVICE DELIVERY 40. The proposal encompasses out-sourcing of ancillary (non-medical) services and free choice of family physicians, among other things. Such improvements have the potential for improving technical efficiency in hospitals and other provider sites, improving physician incentives for providing basic care, and raising patient satisfaction. However, it is equally important to have firm rules within the sector with minimal discretion to ensure fairness, and to put in place the necessary incentives to make the new plans work. 41. Improvements in technical efficiency have been part of the reform process in the OECD, although some of the proposals are not new. Selection of own family physician has been a mainstay of the British Health System since its inception and one of its most successful elements. However, the incentives physicians face are very important, as the structure of the program can take many forms. In the British Health System patients select their primary care physician (PCP) from competing private providers. The PCP coordinates all care and makes all referrals to specialists and hospitals (e.g., act as 'gatekeepers" who control access to specialized forms of care). Patients do not self-refer. Capitation rules for compensating private physicians that tie earnings to the number of patients for whom they are responsible, reward performance and skill. 14 The pre-paid arrangement, the implicit competition across physicians, the function as 'gatekeeper," and the satisfaction of patients all represent key elements in the success of the private physician model. In the US, where continuity of care and 'gatekeeping" are typically not required, the experience has been mixed. 42. In Brazil, where use of private physicians for outpatient care has been standard since 1991, the system has not worked well because it is merely a payment system that underpays physicians (well below market) for their services, does not coordinate patient care, and is not supervised or monitored at any level of government. These deficiencies can be adjusted, but are currently the victim of structural problems, and provide examples of what to avoid. The oversight function for outpatient care, which has been devolved to the municipal level in Brazil, has not worked as envisioned, partly because of the confusion of government roles (see section on Decentralization above), and partly because of the aforementioned inadequate reimbursement that is set at the federal level. How to pay the physicians and who is to be accountable for services rendered are key (World Bank, 1994). 43. Outsourcing was among the first reforms of the UK, begun in the early 1980s and currently improving the quality and lowering the costs of ancillary services. In Latin America outsourcing is growing, but lengthy experience is limited. What has been shown in Jamaica, and other countries as well, is that the specific contract, supervision of contractors, and alternative means of ensuring contractor compliance without terminating the contract are essential to effective purchase of private auxiliary services. Another issue has been the upgrading of laboratories, kitchens, laundries or other services to be privatized. In some cases there have been no bidders given working conditions, or inefficient or inoperable equipment. Pricing of outsourced services can be determined through an open bidding process using private sector prices for similar services as parameters for determining reasonable prices. 44. Both proposals hold promise to improve efficiency and reduce costs. But both require careful design and implementation as well as the appropriate payment and oversight mechanisms that provide the needed incentives for quality services. Moreover, there are additional initiatives to improve efficiency that deserve consideration, most notably separating financing and delivery of services and relying on contracts between payer(s) and providers, as this permits a structure that can more easily ensure accountability; however, this implies a complete restructuring of IMSS health service delivery. Under such a scenario, IMSS could contract all services with IMSS and/or non-IMSS providers, which is the way services are increasingly financed (by government) and delivered (by public or regulated private providers) in the OECD countries (e.g., Holland, Germany, Spanish provinces, Canada and US Medicare and Medicaid). In Latin America, Brazil, Chile, Costa Rica and Nicaragua are splitting finances and delivery and implementing contracts with providers. Less radical adjustments for raising efficiency are listed below: 45. Considerations in outsourcing: * establish workable and effective contracts; * define responsibilities of contractors and the government, and consequences of noncompliance with contract; * supervise contractors carefully and thoroughly; 15 * ensure that the oversight entity has authority to manage contractor and penalize them for poor performance without terminating the contract outright. 46. Consideration in private physician selection: * prequalifying physicians; * setting mutually agreed compensation levels, functions and responsibilities; * establishing effective oversight and supervision of the program; * designing a payment system with appropriate incentives for paying the physicians; encouraging lower costs and higher quality (e.g., physician partnerships); and controlling overuse (co-payments); and * poll patients on a regular basis to determine satisfaction with services. 47. Other options for improving technical and allocative efficiency within the existing IMSS delivery model that have been shown to control costs and improve quality include: * moving simple inpatient procedures to outpatient settings to lower costs, reduce recovery time and eliminate some hospital stays; * controlling and regulating medical technology acquisition; * all payer rate setting, as in Canada, where prices are determined for the entire health care system; * improving hospital management incentives to improve technical efficiency; * closing underutilized hospitals; * contracting out the distribution of drugs and medical supplies. 8. DISCUSSION AND SUMMARY 48. The Mexican reforms represent extensive initiatives to address some of the vocal criticism of the health care system, most importantly the high payroll tax, but also dissatisfaction with IMSS delivery, highly centralized nature of both IMSS and SSA, inadequate access to coverage by the self employed and informal sector workers and high evasion of premium payments. These issues clearly deserve attention, and many of the reforms propose innovative solutions. Without these, the system is likely to suffer. At the same time, the reforms have broader financial, institutional and administrative implications that should be considered to mitigate side effects and strengthen the probabilities of attaining the desired goals. The moment is opportune to initiate fundamental change and the proposed shifts offer bold solutions. However, given the risks in any such shift, mitigating potential problems should be a priority. 49. The reforms attempt to improve equity by extending IMSS to those currently not eligible, and subsidizing their participation, thereby improving access. Quality, choice and patient satisfaction of current IMSS enrollees is contained in the opt-out provision and the proposals to improve efficiency and choice within existing programs, and quality and patient satisfaction are behind the decentralization plans as proximity to patients is expected to improve the quality and appropriateness of service delivery. 16 50. With the reform proposals the Government has followed the US approach, with fragmented coverage and divided risk. The approach separates the population into different risk- pools and charges them accordingly. The lowest income groups either must rely on SSA or on the purchase of subsidized public insurance from IMSS. The plan assumes: (1) heavy subsidies from general revenues to subsidize public insurance and to compensate IMSS for the reduced revenue from lowered payroll taxes, and to subsidize premiums of low wage earners whose contributions are eliminated; and, (2) concentration of wage earner health risks in government financed programs. Given the likely profile of remaining enrollees and the fixed costs of health care, the per worker or per patient cost will rise, raising the revenue gap and a potentially heightened role for the federal government in financing health care through IMSS. 51. The most serious implications of these reforms are the expected financial burden on the federal government. A second concern is institutional capacity for effective reform and decentralization. As discussed above, the preliminary cost measures are likely to have greatly underestimated the true financial requirements due to inadequate information on costs, on behavioral responses to new policies and on administrative arrangements, as well as overly optimistic assumptions about government financial responsibilities overtime, consumer demand and service delivery costs. OECD experience is very clear on consumer response to subsidies under open ended arrangements. The incentives for patients are to use more care, particularly of costly services. And these incentives are being put in place while contributions are being reduced and the government is to subsidize an open-ended insurance program for lower income households. 52. The current costs of health care are quite high. IMSS collects 14% of earnings (12.5% for MMI and 1.5% for health care costs of retirees), receives roughly 5% from pension revenue (N$1,242 in 1994) as well as annual transfers from the federal government (equivalent to 10% of total health expenditures and .3% of GNP in 1992), (FUNSALUD, 1994). Under the new Social Security Law some of these will be lost, and it is unclear if and how these will be recouped, and the implications if alternative sources of revenue are not found. 53. The new payroll tax is heavily subsidized for the 44% of workers who earn 1-3 minimum salaries. Calculating crude estimates of the fiscal burden of the new payroll financing scheme, based on data from the Mexican Government (see footnote 1), suggests the costliness of these reforms for the federal government. Figure 4 shows the projected pattern of total revenues under the payroll tax, and the government's contribution, as a percent of GDP4. The results of this 4 The model assumes 10% increases in IMSS enrollment, an average annual salary of US$3,336 that remains constant over the estimated time period, and, in keeping with the legal provision of phasing out contributions by employers and employees for those earning 1-3 minimum salaries, a ten percentage point increase in Hacienda's contribution each year. It, however, does not adjust for the rising contributions of those earning over 3 M.S., which may exaggerate the government's financial obligations under the reform. Figure 4 Preliminary Projections of Total Collection and Government Contribution for HMI as % of GDP 0.10 % of GDP 0.09 - 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 1996 1998 2000 2002 2004 2006 2008 1997 1999 2001 2003 2005 2007 i TOTAL: HMI Revenue + Government Contributions 18 simple model show substantial government fiscal commitment, which increases rapidly in the first decade of the new programs. The model predicts that the cost in the first year of the reform is about .5% of GDP, over and above any other transfers (see previous paragraph.), and continues to increase overtime as the government covers a larger proportion of contributions for those earning one to three minimum salaries. Even if this is an overestimate, it is representative of the direction of government expenditures under the new legislation. 54. This basic review of incomplete data, and other observations throughout the paper, suggest that the short term costs of health reform are modest in comparison to out year costs. It is in the medium and long term that government financial commitments rise sharply due to the behavioral adjustments of firms and beneficiaries, the shifts in risks, and likely increases in enrollment of the subsidized insurance. Therefore it is imperative for the success of the reforms that the areas of significant fiscal commitment be understood and steps taken to mitigate them. 55. The foregoing estimates, along with the initial estimates of the Mexican Government, deserve careful review and new projections based on sound actuarial calculations, and realistic assumptions about the behavior of providers, patients, government and employers. Indeed the lack of contingency and knowledge of consumer behavior in the mid 1960s, when the US Medicare system was designed, is projected to break the Medicare fund in the early years of the next century. Mexico can avoid that through knowledge of the experiences of other countries, more thorough analysis of this issue, assessment of alternative scenarios, careful design of the system, and built-in mechanisms to contain the volume and cost of services. 56. Institutional capacity is uneven both within the federal government and across states. To effectively decentralize, these weaknesses need to be addressed, in the: design of the decentralization; rules of financial flows and service provision; monitoring arrangements; and accountability at different levels of government. The risk is more layers of government without any benefit, and perhaps detriment, to patients. This engenders political, economic and health risks for the government and the country. Hence strategic plans, systematic implementation, training and monitoring are key. 57. Missing from the reform are some major issues: (1) implications for the SSA, and the quality, scope, and performance of its services; (2) cost containment initiatives; (3) strengthened regulation; and, (4) establishment of a payment system that can ensure the success (effectiveness and affordability) of the private physician plan. 58. Considerable efforts were made over the past few years to assess and recommend areas for SSA reform, but with little translation into reform. The SSA also requires reform and initiatives to improve client satisfaction, efficiency and quality, and to develop national standards and regulations. It is unlikely that decentralization will address these needs. However, the 19 current proposals has equity implications under federal spending patterns (SSA receives 40% and IMSS 10% of federal subsidies in health). SSA's role and relationship to IMSS and private providers, and the restructuring and reform envisioned in the past deserve to be resuscitated. 59. The issue of cost containment through incentives and regulation has been mentioned throughout this note, as a means to cope with the likely increases in the cost and volume of care demanded. It deserves priority to ensure that the health programs and insurance arrangements do not spiral out of control. 60. Although this is not discussed in depth here, the role of regulation becomes key under the proposed system since demand for private insurance and service delivery are likely to grow among the opting out firms. It deserves more attention in general, but particularly as some of the proposals are put into effect. Similarly, given the growth in medical technology, the increasing sophistication of consumers, and growing insurance coverage, the need for cost containment increases. Otherwise the costs to the government will spiral upward rapidly, with fiscal and service quality implications. The evidence on this in the OECD countries in general, and the US and Canada in particular, is clear. 61. The payment system is the structure that offers specific incentives to providers and patients. How the former are paid -- through reimbursement, capitation, salary, fee-for-service -- and whether there are co-payments will determine a good deal about how providers and consumers behave, which in turn will affect costs and volume of care. This issue requires more careful assessments, and the options are extensive, as the OECD experience suggests (see Table A-2 in Annex 1). 62. It is important to point out that the reforms represent one way to reach the stated goals, and represents an important step towards much needed change in the sector. As mentioned earlier, there are many options, each with their own set of side effects. Therefore, the reform objectives can be attained through different venues, many of which would fit within Mexican objectives and current structure. The need to translate new laws and policies into government initiatives and actions leaves considerable room for experimentation and adjustment, and or minimizing both side effects and failure with reform in the process. Despite its risks, reform is both necessary and desirable given the problems facing the sector. 63. The World Bank is open to: (1) follow up on this note, as it only lays out major issues and priority actions, without delving in any depth into the options for implementation, the design of first steps, or the phasing of change; and (2) to assist the process of reform through various avenues, including existing or new loans, sector work, and technical assistance. This can take the form of financing pilot projects, financing and/or participating in analysis, preparing terms of reference, or other tasks that would assist the government address the issues raised in this note. The time is opportune to pursue these issues, and the Bank is willing to participate to the extent that this would be useful and appropriate. 20 BIBLIOGRAPHY Blendon, Robert, Benson, John, Donelan, Karen, Leitman, Robert, Taylor, Humphrey, Koeck, Christian, and Gitterman, Daniel. 1995. Who Has The Best Health Care system? A Second Look." Health Affairs, Winter: 220-230 FUNSALUD. 1994. "Las Cuentas Nacionales de Salud y el Financiamiento de los Servicios." Documentos para el Analysis y la Convergencia No. 7. Manning, W., Newhouse, J., Duan, J., et. al. 1987. "Health Insurance and the Demand for Medical Care: Evidence from a Randomized Experiment." American Economic Review, 77:251-277. Organization for Economic Cooperation and Development (OECD). 1992. The Reform of Health Care: A Comparative Analysis of Seven OECD Countries. Health Policy Studies No. 2. OECD, Paris. Organization for Economic Cooperation and Development (OECD). 1994. The Reform of Health Care Systems: A Review of Seventeen OECD Countries. Health Policy Studies No. 5. OECD, Paris. World Bank. 1994. The Organization, Delivery and Financing of Health Care in Brazil: Agenda for the 90s Report No. 12655-BR. Brazil, Peru and Venezuela Department, World Bank, Washington, D.C. 21 Annex 1 Table A.1: Co-payments in Publicly Financed Health Care Services of OECD Countries, 1993 (US$) iGeneral SeilsDrg,inpatien..:. X-ry n .......7Pattoe _______________ Caetl Australia for 25% of bills for 71% of bills, maximum US$11 nil Included in average of US$5 an average of per prescription specialists' US$8 bills Austria 20% of the population pay 10% or 20% US$2.50 US$6 same as doctors Canada nil nil discretion of nil nil provinces Denmark nil except for under 3% of the 0/25/50% nil nil population Finland US$0.17 US$17 60% in excess of US$22 nil US$8 Greece nil nil 10%/25% - Iceland' US$17 plus 40% of the rest US$9 of the cost 0,12.5%,25% Italy nil 0, US$3 plus 50% - 30% Japan Employees, 10% of all costs; dependents, 20%; self-employed, 30%. Luxembourg 5% 5% 20% flat rate New Zealand Outpatients US$2-US$8 with nil Outpatients stop loss US$3-US$17 Norway 25% if on blue nil X-ray US$11 ticket, maximum US$43 per prescription Portugal US$91-US$213 0/30/60% US$30 Sweden" - US$6-US$19 - first drug US$15 US$8 - then US$1 each Switzerlandb 10% 10% US$7 US$7 10% Turkey nil nil 20% US$10 a day nil United States' 20% in excess of the US$100 US$676 deductible same as deductible 1st 60 days doctors Notes: a) Maximum for the year in the charging scheme. b) Yearly deductible of SF. From 1986 higher deductibles can be chosen. c) Lower deductibles if in HMOs. Source: (OECD, 1994) 22 Annex 1 Table A.2: Type of Payment System for Primary Care and Cost Sharing Practice for OECD Countries .ii.::jAnnua TypedoflPaymn toPrm .. CstSarn :i: -: doctor visit ::;;CareProvider .. p...cap.t.,..... __ _ _ _ _ _ _ 98*9 ..... ...... Australia 8.9 Fee-for-service 25% extra billing Austria 5.8 Fee-or-service 20% of the population pay 10% or 20% Belgium 7.6 Fee-for-service Self employed pay full-cost Canada 6.8 Fee-for-service None Denmark 5.6 28% capitation, 63% fee-for- None service, 9% other Finland 3.5 Salary $0.17 France 7.2 Fee-for-service 25% including, extra-billing Germany 11.5 Fee-for-service none Greece 5.4 Salary None Iceland 4.2 Salary $9 Ireland 6.6 Fee-for-service if higher income; None if lower income capitation if lower income Italy 11.0 Capitation None Japan 12.9 Fee-for-service 10%, 20%, 30% Luxembourg - Fee-for-service 5% Netherlands 5.5 Fee-for-service if higher income; None if lower income capitation if lower income New Zealand - Fee-for-service Extra-billing Norway - 35% salaries, 65% fee-for- 30% costs for selected items service Portugal 2.8 Salary None Spain 6.2 Salary/capitation None Sweden 2.8 Salary $6-$9 Switzerland - Fee-for-service 10% of cost Turkey - Salary None United Kingdom 5.7 Capitation None United States 5.3 Fee-for-service Extra-billing Note: a) Reviewed in OECD (1992) Source: OECD HEALTH DATA (1993). Annex 2 HEALTH SECTOR REFORM STUDIES IN MEXICO (1993-1995) AND BANK COMMENTS I. Introduction 1. Health reform has been discussed actively in Mexico since 1993. In anticipation that a new Government would embark on a program of substantial health reform, four health sector studies were completed immediately prior to the elections of 1994. These studies recommended a comprehensive set of reforms and, in two of the cases, presented plans for different phases of reforms. 2. This paper presents, in a summary format, what each study identified as the reasons why a reform is needed and the main recommendations for the sector. We also summarize the plans for different phases of a reform offered in two of the studies. After commenting briefly on the arguments for reform, the paper highlights some points that should be considered in any plan designed to implement the recommendations. The paper also argues that it will be necessary to identify in more detail the instruments to be used to bring about the proposed changes. In the section on possible phases of a reform, some examples of the actions that might be needed are offered. 3. This paper was produced as part of Bank's technical assistance to Mexican authorities in the field of health sector reform. This assistance program started early in 1993 and included technical assistance inputs provided by IDB, PAHO/WHO, and the Bank coordinated the program through mid 1995. This assistance program represented a departure from the Bank's normal procedures in conducting sector work. Instead of the Bank being responsible for producing the reports, the report was very much the product of the government, reflecting mainly the position of the SSA. The Bank played an advisory role and participated in several seminars leading up to the preparation of the final report. H. The Studies 4. The four health sector studies completed in 1994 were under the direction of four different groups. The final reports were: 1. Economia y Salud: Propuestas para el Avance del Sistema de Salud en Mexico, prepared by the Fundaci6n Mexicana para la Salud (FUNSALUD) under the direction of Julio Frenk, currently Executive Vice President for Economics and Health in the Fundaci6n Mexicana para la Salud. 2. Saludpara Todos los Mexicanos, prepared by the Grupo Asesor en Salud consisting of representatives from SSA,SHCP, CEDESS,FUNSALUD and the Grupo de Prospectiva, under the direction of Jaime Sepulveda, currently Director of the National Institute of Public Health. 3. Retos y Propuestas, prepared by the Fundacion Mexicana Cambio XXI Luis Donaldo Colosio under the direction of Juan Ram6n de la Fuente, currently the Minister of Health. 24 4. Propuestas de Reforma, prepared by IMSS. 5. In addition, at the request of President Zedillo, IMSS prepared a diagnostic study of its entire operations which was completed in March 1995. III. Motivation for Reform 6. The motives for reform of the different studies are summarized in Table 1. Judging from the table, there would appear to be considerable consensus among the first three studies as to the nature of the current problems facing the sector. Foremost among the concerns are: equity, gaps in coverage, low efficiency, patient dissatisfaction and the precarious financial situation of IMSS. There also appears to be a consensus that, if no action is taken, the problems are likely to get worse rather than better due to higher demand for more expensive health services on the part of an aging and increasingly urbanized population. With the exception of the fourth study (the IMSS study) which focuses on the particular problems of the social security institution, the studies do not reveal the priorities they attach to fixing each problem. Since funds will be limited, there will necessarily be trade-offs. 7. While there appears to be a consensus on the nature of the problems, there still remains considerable uncertainty about the magnitude of the problems of coverage and of efficiency of delivery of services within the SSA and IMSS. Before moving on to discuss the recommendations of the different studies we briefly comment on the treatment of efficiency and financial unsustainability of the IMSS health services. We argue that further studies of efficiency should be a high priority and that there is reason to believe that the financial situation of the IMSS is even worse than the dire picture painted by the existing studies. Equity and gaps in coverage 8. The studies present compelling evidence that there exist substantial differences in the health status of the population and in the distribution of health resources across the country. For example, while the burden of disease (measured in Disability Adjusted Life Years) in the three richest states is only 8 percent higher than in established market economies, in the poorest states it is 65 percent higher. the differences in health status are due not only to differences in socioeconomic characteristics of the states and a higher private expenditure on health in the richer states, but also to public expenditures that favor the richer states. For example, in the poor states in the Bank's PASSPA II project, per capita SSA expenditures are 25 percent lower than the national average, although the burden of disease in those states is significantly higher than the national average. This pattern is repeated across the country. Per capita public health expenditures are positively associated with a state's per capita GDP and negatively associated with a state's burden of disease. Thus, instead of compensating for underlying socioeconomic differences, public expenditures exacerbate the situation. Funding allocation rules based on historical allocations with yearly adjustments perpetuates these inequalities. 9. Although there can be disagreements as to what constitutes coverage, by almost any measure there are significant numbers of poor Mexicans, mainly from rural areas, who 25 consume low levels of health services. The low consumption is related in part to: high travel costs due to geographic isolation; poor quality relative to the direct and indirect costs of the service; cultural barriers and a poor referral system. Equity and gaps in coverage 10. The studies present compelling evidence that there exists substantial differences in the health status of the population and in the distribution of health resources across the country. For example, while the burden of disease (measured in Disability Adjusted Life Years) in the three richest states is only 8 percent higher than in established market economies, in the poorest states it is 65 percent higher. The differences in health status are due not only to differences in socioeconomic characteristics of the states and a higher private expenditure on health in the richer states, but also to public expenditures that favor the richer states. For example, in the poor states in the Bank's PASSPA II project, per capita SSA expenditures are 25 percent lower than the national average, although the burden of disease in those states is significantly higher than the national average. This pattern is repeated across the country. Per capita public health expenditures are positively associated with a state's per capita GDP and negatively associated with a state's burden of disease. Thus, instead of compensating for underlying socioeconomic differences, public expenditures exacerbate the situation. Low efficiency 11. The studies argue that the Mexican health system considered as a whole is inefficient and that service delivery within the IMSS and SSA is also inefficient. The evidence that the system as a whole is inefficient is gleaned from comparing Mexico's health outcomes given the amount of money spent to those in other countries. This aggregate evidence is fairly convincing that Mexico could do better. While there is a perception that service deliyery within each one of the major health institutions is inefficient (and this is probably true), there appears to be little hard evidence on the extent of inefficiency. Indeed, data are lacking both on the outcomes side and, especially, on the cost side. 12. What is known about efficiency comes mainly from studies of service delivery within the IMSS system. The major sources of inefficiency within the IMSS are identified in the IMSS diagnosis as: * Health services being delivered at the inappropriate level, i.e. services that should be done at the primary level being done at the secondary level and those being done at the tertiary level that should be done at the secondary level. * High costs of purchasing and distributing medicines * Expensive medical equipment being used at low intensity. * Activities of low cost-effectiveness being performed Health services being delivered at an inappropriate level 13. The IMSS diagnosis states that there are 1,495 primary level centers that absorb 26 percent of the total budget and should (but not necessarily do) provide 85 percent of the consultations. There are 227 secondary level hospitals that absorb 54 percent of the budget 26 and 42 specialty hospitals that receive 20 percent of the budget. The study estimates that N$ 1,100 million is spent on patients who are sent to the second level that should be treated at the first. At the exchange rate prevailing at the time of the study, this was over US$ 300 million, an amount roughly equivalent to the deficit in the IMSS health care system. 14. The above figures imply that if patients could be redirected to the appropriate facility for care, that the health system would enjoy substantial savings. While it is undoubtedly true that there would be savings, unless the study made an adequate distinction between fixed and variable costs (something that is not always done in many hospital cost studies), it is likely that the cost savings are overestimated. In many hospital cost studies, these savings are calculated by first dividing total consultations by total costs to arrive at a per consultation cost at primary and secondary level facilities. Because the secondary level facility has higher fixed costs, the per consultation cost is almost always higher than in the primary facility. For each consultation that could be redirected from the secondary to the primary level, it is assumed that the system could save the difference between the per consultation cost at the primary and secondary levels. This reasoning is incorrect. If the fixed costs related to the primary and secondary facilities must be paid regardless of where the patient receives attention, the savings from shifting attention back to the primary level is simple the difference in the variable cost of having the service provided at the primary instead of secondary level. For example, if an individual received stitches for a cut at the primary level instead of the secondary level, this would free up the time of the physician at the secondary level to treat another patient. The consequence may be a reduction in waiting times at the secondary level, but would not necessarily result in a substantial savings in cash outlay for the system as a whole. Only if the variable costs of delivering the service at the two levels were dramatically different or if were possible to reduce the overall fixed costs in the system (by avoiding having to build as many secondary level facilities in the first place or closing existing secondary facilities) would there be significant savings for the system as a whole. If it were possible to reduce the overall number of employees (a quasi-fixed factor), there would also be savings in the system. High costs of purchasing and distributing medicines 15. The Diagnostico prepared by the IMSS identified the following as the principal weaknesses in the system of purchasing and distribution of medicine, diagnostic inputs and curative supplies: * Multiple errors in information and control, caused in part by a computer system that is almost 15 years old * High cost of inventories that outweigh the savings from purchasing in large volumes * Delays in distributing medicine and supplies * Inadequate systems of feedback to ensure that supplies more closely match demands 16. The IMSS diagnosis of the problem is excellent and worth stressing because it undoubtedly applies to the SSA as well. The excessive attention paid to economies of scale in centralized purchasing is common to procurement specialists but is bad economics. The IMSS diagnosis points out the high costs associated with delivering the products, maintaining inventories, losses due to expiration of dated medicines, and inappropriate mix 27 of medicines. Good economic reasoning would require that all the costs involved in getting medicines to the point of service be taken into account. Expensive medical equipment being used at low intensity 17. Evidence that expensive medical equipment is used with low intensity suggests that past purchasing decisions were inefficient. However, sunk costs are sunk costs and using the existing medical equipment more intensively in the future does not, by itself, indicate a gain in efficiency. It may be more efficient to avoid the variable costs of using the equipment when it is not really justified on medical grounds. 18. Using equipment at low intensity may adversely affect quality. Experience in Canada has demonstrated that there is a health benefit from performing complex operations at a centralized location. High volumes are associated with low death rates and low costs. Evidently, a minimum volume of operations appears to be needed for proficiency. 19. The more important decisions that would affect efficiency in the use of medical equipment would be: a) the decision to sell off underused medical equipment; and b) the decision on when to acquire new equipment. New equipment should not be acquired as long as there are cheaper ways of providing services by using existing equipment more intensively or reallocating equipment across different facilities. It would also be useful to study how well the predictions for use originally made to justify the equipment are borne out in practice. If predictions are systematically overly optimistic, it may be worthwhile rethinking the process by which equipment needs are identified. Studies of inefficiency should be a high priority 20. Identifying the sources of inefficiency within the IMSS and SSA should be a high priority. If SSA and IMSS facilities are to operate as autonomous units competing against multiple providers they will have to know where they are spending their money and how to reduce costs while maintaining quality. Moving to a competitive system will not automatically provide information on where the costs are being incurred and what to do about it. It will simply provide more pressure to become more efficient as the IMSS and SSA autonomous units will lose affiliates and hence resources if they do not. It will still be up to management to identify the sources of inefficiency and to take action. For that they will need information and it is best to start building the information systems today. 21. It will be important to decide how best to approach studies on efficiency. First, the studies should try to exploit variations in outcomes across the system. For given resources, some units will have good outcomes, others poor outcomes. A statistical analysis could provide a first cut on the differences. It would then be useful to complement this statistical analysis with a qualitative analysis, visiting the extremely good and bad performers to try to understand why the different outcomes were obtained. Trying to reduce the variation in outcomes given resources, turning poor performers into average performers and average performers into performers closer to the top of the scale is one promising way to tackle the efficiency problem. Second, some decision should be taken as to the most appropriate combination of health facilities over which one detects and treats efficiency problems. 28 Should it be all primary clinics statewide or nationwide? Should it be a single hospital? If one of the sources of inefficiency is that services are being delivered at an inappropriate level, this would argue that efficiency problems be considered within a vertically integrated set of primary, secondary and tertiary units. What the most efficient autonomous unit is for dealing with efficiency problems has obvious implications for the decentralization process. If the most efficient autonomous unit lies not at the state level, but at a more disaggregated level, then decentralizing decisions only to the state level will not yield substantial gains. Unsustainable financial situation of IMSS. 22. The IMSS Diagnostico presents a bleak picture of the financial situation of the IMSS health system. It estimates that the quota required to cover costs given the current level of services and without considering a potential increase in users is 13.5%. The current quota is 12.5%. Historically, the quota for health services was set too low - at a level to cover the primary beneficiary without taking into account the costs of treating other family members. In the past, this deficit had been made up by subsidizing the health services with funds that had been collected for pensions and day care and by transfers from Hacienda. Given the crisis in pensions, cross-subsidization is no longer an option. Nor is raising the quota, as it would raise labor costs and reduce competitiveness. At the same time, as the population is aging the demand for more costly health care is rising. 23. As serious as the situation appears to be, there are some reasons to believe that the situation is, in fact, worse than diagnosed by IMSS. First, the projections of the aging of the population are presented for Mexico as a whole. As the affiliates and families of IMSS are wealthier and have lower fertility, they are undoubtedly older on average than the general population. Because those who are not yet in IMSS are on average younger than those who are, adding members might be expected to generate a positive cash flow, at least in the short term. However, adding these types of workers is likely to prove difficult and one cannot assume that, under the current system, expansion rates for IMSS coverage will continue as they have in the past. In urban areas among salaried workers, IMSS coverage is already around 75 percent, but this varies dramatically with the size of the firm. Coverage in firms with 100 or more workers is 92 percent, while coverage in firms with 1-5 workers is 28 percent. Coverage among self-employed and piece rate workers is still low, but including these groups will probably require other means of collecting payment. The record of the IMSS in special programs targeted to non-salaried workers is not particularly good. This program had a deficit of N$ 950 million. Either cuotas are set too low to cover costs, perhaps to increase coverage or there is more evasion of cuotas which is difficult to control. Thus, it may prove administratively more expensive to expand membership and may require additional subsidies IV. Recommendations of the Studies 24. The main recommendations of recent health sector studies are summarized in Table 2. The main recommendations common to the different studies are to: 29 * Define a package of essential health services * Decentralize * Separate financing from provision of health services 25. These recommendations are similar to those in sector studies conducted in other countries, especially those conducted after the influential WDR on Health. It is important to recognize that while there seems to be a convergence in the type of recommendations emanating from sector studies throughout the world, there has yet to be a country that has gone from the beginning to the end of the reform process. Different countries are arrayed along different points on a time line, with some variations in the approaches taken. Therefore, for Mexico it will be important to take note of some practical issues that have been confronted or will have to be confronted. 26. Below we raise some points to consider if Mexico were to try to implement the recommendations common to the different studies. Define a package of essential services 27. The recommendation to define a set of preventive and essential clinical services that would be guaranteed for all follows a suggestion made in the 1993 WDR on Health. If the definition of a package of essential services is introduced into the existing health care system simply as a means of defining priorities, it is unlikely to have much of an effect. For years, health sector studies have recommended that more emphasis be given to primary over tertiary care, not just in Mexico but in other countries. With the essential package, how one arrives at the priorities may be more sophisticated, but defining priorities, by itself, will not bring about changes to the system. 28. However, if the package of essential services is to play a central role in contractual arrangement between a principal who would like to see the essential package consumed and the agent or agents charged with ensuring that the package is consumed, then the introduction of an essential package of services could change the way the health system operates. It would be the introduction of contractual arrangements together with the notion of defined benefits that would be most responsible for the change, rather than the particular choice of services that make up the package. 29. If it is envisioned that the essential package would be delivered through contractual arrangements, then the treatment of the essential package in the four different studies has been incomplete. It is not sufficient merely to list the interventions that one would like to see take place. It is also important to identify: a) what is the defined population that is to receive the services; b) how is a family to be identified as being from the defined population (through a system of affiliation with a provider?) c) how will the increase in coverage be phased over time; d) what are the performance standards for delivering the services; e) what is the mechanism to be used in defining whether the agents have performed satisfactorily in delivering the package; f) how will adequate performance be monitored; g) what role will consumers play in determining whether performance has been adequate; h) what rewards will be provided for adequate performance and what sanctions for inadequate performance; i) 30 what dispute resolution mechanisms are available if there is a disagreement over whether the package has been adequately delivered; j) what is the price that the principals are willing to pay for adequate performance in delivering the essential package; and k) who bears the risk if delivering the essential package turns out to be more costly than expected. 30. There is a second sense in which the treatment in the different studies of the package of essential services has been incomplete. There is no discussion of what is to be done when, even at a zero direct price for the services, there is insufficient demand for a service considered to be highly cost-effective and included in the package of essential clinical services. This may be the case, for example, with prenatal care. One option to increase consumption of prenatal care is to generate an implicit negative price for the service. This does not have to involve a cash payment to those who receive prenatal care. It can be achieved by linking a non health-related benefit to going for prenatal care. For example, nutritional supplements could be distributed through the health clinics or, if an adequate administrative mechanism could be developed, the supplements could be delivered through other channels only with proof that a woman has sought prenatal care. Alternatively, an implicit negative price could be generated by bundling different health services. If there is a service, say attended births, for which there is substantial demand at a positive price, then a reduction in the price of the attended birth could be given to those who attend a specified number of prenatal care sessions. 31. The government will also have to decide how they will limit public expenditures on health interventions that are not part of the basic package. This is essential if public funds are to be freed up so that the basic package can be guaranteed. This is particularly important as it is quite likely that the government willingness to finance interventions in the basic package would be expected to induce a substitution of public for private funds. Presumably because they are so cost effective, some private individuals would have been willing to privately finance some of the interventions within the essential package even without the government guarantee. 32. Public expenditures should be limited by controlling inputs used in the interventions outside of the basic package and not by attempting to define what a doctor operating in a public facility may or may not treat. The key inputs to control are: a) specialized facilities; b) human resources, especially the training of specialists; c) nongeneric drugs; and d) specialized equipment. As discussed in the previous section on efficiency, containing costs may involve managing the introduction of new and expensive technologies. In Canada, new medical technologies are first introduced to the larger hospitals and filter down to other hospitals when demand exists. 33. If the number of specialized facilities are limited so as to contain costs, this will give rise to rationing of services. One way of reducing the waiting list while still avoiding constructing new facilities would be to issue vouchers that would enable clients to use private facilities. Such a system has been in use in Germany for many years. The vouchers would be issued for a given service at a fixed fee. Some proportion of the waiting patients would be expected to accept the voucher and go to the private sector. If the patient in the private sector wanted to pay additional fees for additional services (a more comfortable room, etc.), they would be free to pay the difference. 31 34. Finally, if guaranteeing the consumption of a package of essential services is to be the goal for health policy, then a new objective replacing the current notion of coverage based on physical distance is needed. The number who are not covered is derived as a residual, obtained after subtracting the number of people covered by IMSS, an estimate of how many people receive services from the private sector and an estimate of the number of people who live more than 60 minutes away from a public SSA health facility. By this definition, there are roughly 13 million Mexicans without access to health care. 35. While this definition of coverage based on distance to a public clinic is useful if the goal is to ensure that all Mexicans live within 60 minutes of a public health facility, it is not particularly helpful in assessing how close Mexico is to a goal advocated in all of the studies - that of guaranteeing a package of essential services to all. Without a visible benchmark to remind the public and policy makers how far they have to go to meet the goal, it may prove difficult to marshall political support for the needed budgetary allocations, especially given the budgetary inertia and the financial pressures that are likely to prevail in Mexico over the next few years. 36. One possible way of defining the goal is by whether those who need health services actually utilize them, whether from public or private providers. An operational way of defining this concept is that, for a given intervention included in the essential package, the probability of receiving the intervention given that a person is ill would be unrelated to a person's income, place of residence or any other systematic factor. The information required to calculate these probabilities could, in principle, be obtained from household surveys. Because it is in rural areas where Mexico is furthest from the goal, to save costs it might be useful to restrict the survey to rural areas. 37. The notion that there could be nonsystematic variation across individuals in the probability of receiving the intervention reflects other factors, such as the intensity of the illness and particular preferences of individuals that cannot be observed or necessarily affected by policy decisions. What is of most concern for policy makers is whether there are systematic differences that could be affected by policy. In Mexico it is undoubtedly the case that a person's income and place of residence do matter for utilization of health services. An initial step has been made with a recent study on demand for health services based on the 1994 National Health Survey. However, more remains to be done and we don not yet have any time-series information on whether the problems are getting worse or better. Establishing baselines and tracking achievements of major investments such as the ones contemplated in the World Bank supported Basic Health II would be desirable. Decentralize 38. The discussion on decentralization of health services in Mexico has focused mainly on what powers to delegate from the central to the state level. Less attention has been given to a key question related to the industrial organization of the sector, namely what is the most 32 efficient autonomous unit to take decisions. Is it a single clinic? a single hospital? or an integrated entity consisting of a hospital and affiliated clinics capable of providing all the clinic services included in a package of essential services or other packages that might be chosen by families? Decentralizing decisions to the states will not yield the hoped-for efficiency gains if the most efficient autonomous unit to take decisions is something different. 39. Another way of framing the question is to ask where are economies of scale exhausted. The FUNSALUD and IMSS reports explicitly recognize that the public institutions are too large while the private sector in Mexico is too atomistic. The IMSS study, in particular, provides several examples where the sheer size of the institution creates difficulties and proposes splitting up the institution into more manageable units. 40. How the decentralization is managed is also likely to affect the ease with which the system could move to a system which separates financing from provision. If decentralization stops at the state level, the task of converting public facilities into smaller HMO-type organizations that could compete with the private sector in the delivery of pre-paid plans would probably be no easier than if this task were initiated under the current centralized system. If, however, the decentralization had passed on budgets and decision making authority to smaller integrated entities that consisted of public hospitals and clinics, the steps required to have these public entities then compete with private entities may not be as large. Separate financing from provision 41. Separating financing from provision necessarily requires that the activities being provided be governed by explicit contracts. The four studies have argued that separating financing from provision is the most promising way to introduce greater efficiency into the system and have stressed the positive incentive effects that are created. Less attention has been paid to how the contracts should be structured so as to minimize the transaction costs of achieving the positive incentive effects. 42. Some of the relevant issues were mentioned above in connection with the discussion on the delivery of an essential package. Additional questions are: 1) How should services be bundled into contracts? With a task such as outsourcing of laundry in a hospital, this is fairly simple. With health services it is more difficult, but not impossible. HMOs and insurance groups work with defined packages of services. The IMSS proposal for reforms also mentioned the possibility of creating clearly differentiated packages of services. 2) Who should be a party to the contracts? At issue is not just whether the public financing institution should enter into contracts with private entities, but also what should be the appropriate level of integration of services. 33 If the SSA would like to contract for the delivery of a package of essential services, should it enter into contracts with clinics and hospitals separately or enter into a contract with a single entity consisting of a hospital with affiliated clinics that could offer all services. In the latter case, it would be the responsibility of the managers of the integrated services to ensure that the right service is delivered at the right level of care. Which arrangement results in the lowest transaction costs for a package of essential services? Very little is known about the transactions costs associated with alternative contractual arrangements in Mexico, or for that matter in other countries that are attempting to separate financing from provision. For this reason, policy makers may want to consider pilot testing some of these arrangements in a state, possibly Aguascalientes that has already introduced several innovations in its health care financing. 3) What is the appropriate size of the contract? It must be small enough so that more than one entity could bid for the contract. Thus, one might observe many potential bidders for a contract to provide an essential package to 100,000 residents in a defined rural area. However, if the contract is to provide basic care for 40 million Mexicans, only the existing IMSS institution could bid for that contract. 4) Is a model based on separation of financing from provision feasible for rural areas? The studies note that ensuring adequate services in rural areas (however this is defined) is likely to remain an important task for public health institutions, regardless of whether the health system is reformed or not and regardless of the nature of the reforms that are ultimately adopted. They also recognize that a different approach may be required in rural areas because there is a limited choice of provider. Most of the institutional reforms are directed to an urban setting where multiple providers already exist and where incomes are higher. In rural areas, there is still the possibility of putting the delivery of services out to bid, that is, to grant a franchise for the delivery of services as is suggested in the FUNSALUD study. While this is a possibility, it is worth noting that contracting out services in rural areas involves some additional complications as compared with the urban setting. The main complication is that private contractors would not have an incentive to invest in specific assets (for example a rural health clinic or training of health staff) unless they would be guaranteed that they could recoup the cost of the investment. The longer the length of the contract, the more likely it is that they could recoup the investment and, hence, the more willing they would be to make investments in specific assets. However, if such a lengthy guarantee is given, that decreases the beneficial effects of being able to award the contract to someone else if performance is not adequate. If the public sector will be involved in delivery of services in the rural sector (at least in the immediate future), the relevant questions to ask are what are the best approaches to guarantee the consumption of an essential package of services and how much will this cost. Should there be mobile units, rural health technicians, additional health clinics with more 34 doctors, or should services be put out to bid to potential private providers who do not currently operate in rural areas? In answering these questions, one should keep in mind the following: a) If there is uncertainty about which approach is superior in a given setting, there could be potentially large returns to testing alternatives in a rigorous way; and b) The central level of the SSA should make available the experience of what has worked and what has not under different settings, but should allow some local discretion to adopt the approach that they think best fits their circumstances. 5) Will alternative organizations of services within a public institution be needed for the separation of financing from provision to yield the expected gains? 43. In the US there is considerable realignment of the health care industry as providers are forced by major purchasers of health care services to become more efficient. Hospitals are merging. Large hospitals are becoming affiliated with smaller clinics to provide an integrated packet of services. The relative position of independent specialists and large teaching hospitals is threatened. These dynamics are being driven by market forces even though no health reform took place. However, in a large monolithic public institution such as the IMSS or the SSA, there are no market forces that would suggest that more complex services should be delivered in a health clinic, that hospital stays should be minimized even if that leads to lower occupancy rates or that the most efficient autonomous decision making unit should be a large hospital with affiliated clinics. The leaders of those institutions would have to decide that these changes were needed and take administrative decisions that would lead to those changes. V. Possible Phases of a Reform 44. It is unlikely that the recommendations could be carried out in a single comprehensive reform. It is more likely that the reforms will be carried out in phases. Because it is uncertain how long the system might remain in each phase, it is important that the actions taken at any one step lead to a stable system or, at least, to a system where a problem does not explode. For example, if changes made during the first phase of a reform will provide an incentive for the growth of private health insurance, then the first phase should also involve the development of regulations of the private sector to avoid well-known problems with private health insurance. 45. Tables 3 and 4 present the phases of reform proposed by the FUNSALUD and SSA studies. As with the motivation and recommendations, there appears to be some consensus in the sequence of activities, as evidenced by the titles each study chose for the three phases. Both studies make the point that the three phases could also be viewed as three alternatives for the health system. That is, they argue that the actions taken during each phase would give rise to a sustainable system. 46. In reviewing the recommendations of the four studies and the two phased plans, in particular, it is worth keeping in mind the objectives of the studies. Prepared during an election year, the four studies identified problems within the health sector and provided a 35 vision of what a desirable health system for Mexico would be. Although there was considerable consensus on the problems and recommended solutions, these studies did not specify a detailed mechanism on how to move from the initial conditions to the desired outcomes. That task awaited a political decision on the direction and extent of reform. 47. If a reform is to be put in practice, the reform process needs to be instrumented in greater detail. Table 5 briefly describes actions within nine key areas that the government might want to consider taking over three different time periods: the near term (A), the short- run (B), and the medium term (C). How long these periods should be is best defined by the government, but a tentative reading might be for the near term to be the first year, the short run within three years and the medium term after three years. 48. In the legal area, it is urgent to prepare the legal proposal for constitutional adjustment, and it is important that work be started on the legal framework leading to a possible draft law (if time and conditions so allow). In the short run, the spacing of regulations should be clearly defined in order to begin implementing the set of decisions and actions in the financial and organizational areas indicated in the following paragraphs. 49. In the financial area, it is urgent to define the mobilization of resources for reform (the possible complementarity among payroll, general and prepaid taxes), such as the adoption of a generalized per capita formula to allocate resources to the different regions of the Secretariat's and IMSS' system, and the start-up of an anti-tax evasion plan. The first steps should be taken to start up a standardized system of reintegration by service groups, gradually leading to the adoption of a prospective budgeting system. The decentralized use of resources should be guided by these general regulations and be registered under a standardized accounting plan that allows proper follow-up and analysis of resources used. In subsequent steps, management agreements could be put in place with the various social security institutions and the main public providers, using a system of performance incentives for administrators and management groups. 50. In the organizational area, critical decisions deal with the regionalization of public purchasing systems in social security and the Secretariat of Health, with a view to their later integration, the definition of mechanisms for identifying beneficiaries and the involvement of the universal (or basic) plan in relation to the poorest people and the promotion of the market for comprehensive health service providers. Initially, the strategy should strengthen the capacity to regulate providers, promoting their autonomy under the previously defined general incentive system, defining the framework of their contractual relationships and trying to strengthen the quality of care they provide. Subsequently, they should be able to adopt an accurate benefit plan for health insurance, generate initial comprehensive contracting (as an institutional alternative to quota reversal) and the gradual adoption of a standardized social security card. Thereafter, the gradual integration of state public funds and social security funds could be started, by allowing the use of local funds to finance health insurance. The new institutionalized control and supervision could be started (separate from the system's authorities responsible for its control and execution), a general framework for competition and consumer protection could be adopted, and the increasing options for partial or total freedom of choice could be put in place. 36 Table 5. Strategies AREA DECISION ACTIONS TIMING LEGAL STRATEGY a. Constitutional Modify articles Prepare alternatives A b. Legal Prepare draft law Prepare alternatives B d. Regulatory Prepare decrees and Identify agenda B instructions Establish working groups FINANCIAL STRATEGY a. Resource 1. Substitute or Modify payroll tax A Mobilization supplement payroll Definition of taxes replacement taxes and complementary quotas 2. Fiscal control of Implicitly eliminate B taxes subsidy on employers' quotas Anti-tax evasion Decree A b. Efficient allocation 1. Domestic market Training to provide A of resources using for services basic services market instruments Reimbursement for B diagnostic groups 2. Domestic insurance market Training adjusted by A risk and reinsurance funds c. Decentralization of 1. Resources for Formula for allocating A resources regions resources according to needs Regulations for use of B 2. Resources for local funds (and providers prospective budget) Unified accounting plan B 3. Unification of accounting procedures d. Cost controls 1. Cost disclosure and Management agreement C rationalization with IMSS, ISSTE and Secretariats 2. Performance For administrators B incentives For management C I groups _ ORGANIZATIONAL STRATEGY 37 AREA DECISION ACTIONS TIMING a. Separation of 1. Unification of Permit the use of C service provision and Funds regional public funds to financing finance insurance 2. Constitution of public purchasers A. Social Security A. Regionalization of A IMSS purchasers B. Decentralized Regulation of local A public health funds and charts of responsibility 3. Autonomy to General Hospital B medical providers Statute a. Public b. Private 4. Strengthening of Popular councils and B social control organizations for basic authorities health care 5. Develop model Sample contracts B contracts 6. Unified invoicing Sample forms B and payment mechanisms 7. General statute on Consumer protection C health care decree competition b. Creation of market 1. Development of Regulation of HMOs A for competitive, comprehensive and insurance. comprehensive organizations (HMOs) Promotion organizations 2. Development of Sample contracts B sample contracts for Gradual contracting B total risk 3. Free choice of Regulation C affiliation 4. Development of Regulation C family medical practice groups 5. Development of Sample contracts B sample contracts for partial risk 6. Free option for Gradual contracting B basic family care 7. Supervision model Definition of C institutional authorities, and implementation of C procedures c. Definition of benefit Universal plan A plans Insurance plan B 38 AREA DECISTON ACTIONS TIMNG Supplementary plans C d. Identification of Standard social Definition of card B beneficiaries security card Systematization of B reading and usage Standardized family Definition A registration system Implementation in poor B areas e. Quality assurance General Statute on Provider Accreditation B Quality Assurance Plan Professional licensing C plan Public information B systems VI. Tasks for Various Authorities 51. The paper concludes with some thoughts about the possible role of the government and the World Bank in the health reform process. The government as a whole, perhaps under the coordination of the Health Cabinet as its logical authority, is responsible for three critical tasks: the preparation of a conceptual proposal for reform (essentially a document based on the current Overview of the Reform of the National Health System), the final identification of the strategic agenda for reform, and the coordination of legal advisers on constitutional, legal and regulatory matters. 52. The Secretariat of Health would be responsible for the conceptual and operational development of decentralization in the context of the process of universalizing health insurance. This represents decisions on formulas to distribute national resources, the configuration of local funds as purchasers-coordinators of local public and private providers, the activation of social control authorities, and the progressive introduction of competition and freedom of choice, especially in larger cities. It is critical to make the organizational and incentive implications operational in order to proceed with the basic health care package and to keep it from being only a recommendation with little effectiveness. The design of fairly significant pilot programs, in the organization of universal health insurance systems in any state, or in the community organization of basic health care, would be most useful. Finally, the growing involvement of the system's administrative agency in quality issues should start to be exercised in tasks such as a general statute on quality assurance (with accreditation, standards and basic reports), a statute on hospital autonomy, and the development of sample contracts. 53. In the IMSS, it would appear useful to concentrate attention on carrying out the effective regionalization of the system and the progressive autonomy of providers, accompanying the adoption of efficient budgetary and payment mechanisms. It is therefore 39 essential to develop a per capita formula, adjusted in terms of risk, in order to transfer resources to regional agencies as well as to develop budgeting and payment systems for regional adoption. It would be useful for performance plans to be available to regions and to the Institute's providers, as a mechanism for planning the change. The IMSS could invest resources in scientific studies of effectiveness that allow other countries' experiences to be adapted in the designs of service plans and minutes of understanding. It also plays a critical role in the gradual establishment of service options, whether partial (such as family physician or outpatient group practice, or professional risks) or comprehensive (as an alternative to quota reversal). 54. Hacienda is responsible for taking the leadership role in establishing a fairer and more efficient incentive framework and in promoting competition and free choice within the system. The first urgent task would be to evaluate the impact of the various reform alternatives, using and adapting models that have been applied in other reform processes. The second task would be to work with the SSA to modify the procedures for overall allocation of resources to regions, in order to break long-standing inertia and to increasingly benefit the poor and needy. In line with the above, a unified accounting, budgeting, invoicing and payment plan should be adopted in the various public agencies that are in the process of decentralization, using IMSS' experience as much as possible. It would also play a more active role in defining the means to stratify the target populations for public subsidies in various areas, including health. Finally, it would play a more active role in promoting decision-making mechanisms that would allow a gradual development of competition and freedom of choice. The most important initial decision is the definition of a decree regulating comprehensive health service providers (or HMOs) so that they can enter the market under efficient and fair conditions. 55. The Bank can play a dual role during this time of change in Mexico. First, it can carry out the PASPA II project within the context of the reform process, for which it must substantially develop organizational and financial designs in order to carry out the universalization of the basic package. In so doing, it should prioritize technical assistance for the definition of the organizational options of domestic agencies and of basic service providers , as well as the general framework of incentives and financing that make it sustainable. For example, the project could contract with foreign HMOs and other private health care firms to assist the SSA and IMSS in creating combinations of public sector facilities that could operate as viable providers in a multiple provider system. The project could also finance studies that examine the efficiency of the system and pilots that would reveal information about the transaction costs of moving to a system that separates financing from provision. 56. Second, in view of its increasing participation as an interlocutor in the overall health reform process, the Bank could mobilize additional technical assistance resources (or try to allocate some of the resources from current loans) to support the set of health reform decisions that will be affected by overall and social security reforms. It would be particularly important to advise on the model for evaluating the reforms as a whole, the overall mechanisms for allocating resources to regions (per capita models), and adjustments to a competitive model for decentralization. The Bank, especially the center, could sponsor seminars to facilitate exchange on the how-to of reform. To be most useful, the seminars 40 should be organized for countries that share initial conditions and are roughly trying to achieve the same goals. 41 Evaluation: to change all workers' contributions to a single quota financed by workers. FAVORABLE CRITICAL EVALUATIONS ARGUMENTS ARGUMENTS 1. LABOR MARKET Decreases average Depends on average labor cost. quota level. If affiliation is mandatory, and the quota level is equivalent to the current average cost, the average labor cost does not decrease. Does not affect Effects may be the Elasticities may be employment opposite (depending on different for various (ambiguous effects) who pays the market types of workers. If and increases formal price and the elasticity elasticity of supply is salaries (if overall of supply) greater for those less supply is not elastic, qualified, the effects but substitution with would be to expand the informal sector is salary structure and high). worsen employment possibilities for those less qualified. Increases fairness for May be unfair This is an empirical two reasons: issue. - it reduces current - under-declaration of under-declaration payroll tax should not (evasion) be greater than that in income tax or IVA. - it reduces the lack of - Would these persons affiliation by persons demand IMSS services, better able to pay even if they were free? (avoidance) Reverses Informalization also The empirical informalization: depends on magnitude of the effect encourages development of markets is uncertain. participation in formal for goods. sector - Mexican wages are not low for its level of economic development. - Alternative financing (through IVA) may maintain incentives to informalization in order to avoid the IVA. Improves conditions True. for progress in productivity because it 42 FAVORABLE CRITICAL EVALUATIONS AROUMENTS ARGUMENTS decreases marginal taxes for the largest product. 2. HEALTH SERVICES MARKET Creates incentives to Only for the The magnitude of the expand affiliation. informal sector with impact on affiliation is greater ability to pay, an empirical issue. because the effect on income may be highly significant for the average population. Expansion of affiliation depends on federal subsidies for the poorest people. Generates stable yields, Stability depends on: The relative stability of because public income - the share of prepaid the various sources of fluctuates less than resources (insurance is financing is an wages. highly elastic to empirical issue. income). - source of public resources (IVA is highly elastic). Generates sustainable Costs may be excessive. Examine in more detail resources for the Reaching 70% of the the quota level that is expansion of the health population with a quota both sustainable (no system. of N222/month more than 5% or 6% of represents 10.7% of the GDP) and compatible GDP, not counting with costs of providing public health the service. expenditures and expenditures for the poor. Quota becomes a The efficiency incentive This may be achieved benchmark of the depends on the level of by other means. overall budget, which the benchmark. promotes efficiency. Breaks labor ties and True, but it takes away allows expanded a sense of ownership coverage. and responsibility. Sensibly finances May be achieved by affiliates of insurance other means (direct of special programs: fiscal subsidy). eliminates identity conflict and increases the subsidy's i _ _transparency. By increasing For such a large, stable Adverse selection may coverage, it decreases population, the effects be more related to 43 FAVORABLE CRlTICAL EVALUATIONS ARGUMENTS ARGUMENTS adverse selection of adverse selection may temporary affiliation against insurance. be smaller. movements and the minimum periods for certain costly operations. Facilitates the The reversal of quotas May be achieved by development of private is a system to promote other, fairer and more insurance systems, with competition that lacks efficient means. the reversal of quotas fairness and efficiency. and subsidies on demand. 3. PUBLIC FINANCE Decreases collection Not much, because the Effect may not be costs. payroll tax remains for significant. other insurances managed by the IMSS. Increases collection Depends on the relative The effect on various because of efficiency in efficiency of the incomes may not be simplified collection collection of other significant. mechanisms. taxes. Collection loses the solidarity of wealthier groups. Field for subsidies. This field may initially Depends on restricting be very large and could their size (avoiding become a precedent for their dynamic growth) later expansion. and the possibility of focusing them on the neediest populations (avoiding their generalization). Reduce the Some documents It is necessary in any deductibility of suggest the opposite: case to reduce the state business tax quotas. generalize the subsidy on excess deductibility of spending. voluntary insurance. Concentrate the The maintenance (or This is a political distributive role of increase) of the judgment on the public finance instead system's current advantages and of keeping it in a single progressiveness depends possibilities of fairness. tax. on the possibility of additional reforms in other taxes. Barbara Diallo M:\LEWIS\FINALMX.DOC May 8, 1996 10:41 AM 43 FAVORABLE CRITICAL EVALUATIONS AROUN1ENTS ad%erse selection of ad%erse selection may temporary affiliation against insurance. be smaller. movements and the minimum periods for certain costly operations. Facilitates the The reversal of quotas May be achieved by development of private is a system to promote other, fairer and more insurance systems, with competition that lacks efficient means. the reversal of quotas fairness and efficiency. and subsidies on demand. 3. PUBLIC FINANCE Decreases collection Not much, because the Effect may not be costs. payroll tax remains for significant. other insurances managed by the IMSS. Increases collection Depends on the relative The effect on various because of efficiency in efficiency of the incomes may not be simplified collection collection of other significant. mechanisms. taxes. Collection loses the solidarity of wealthier groups. Field for subsidies. This field may initially Depends on restricting be very large and could their size (avoiding become a precedent for their dynamic growth) later expansion. and the possibility of focusing them on the neediest populations (avoiding their generalization). Reduce the Some documents It is necessary in any deductibility of suggest the opposite: case to reduce the state business tax quotas. generalize the subsidy on excess deductibility of spending. voluntary insurance. Concentrate the The maintenance (or This is a political distributive role of increase) of the judgment on the public finance instead system's current advantages and of keeping it in a single progressiveness depends possibilities of fairness. tax. on the possibility of additional reforms in other taxes. W � � N м и w � а Н W 1 � i1] Са rz н •• W Н о � 22W' t'J W СЛ О д -и а н н .. � �, о а э Н zaa � о о � с� �аН