Peru - Improving health care for the poor

Report No 18549-PE Peru Improving Health Care for the Poor May 28, 1999 Human Development Department Bolivia, Paraguay and Peru Country Management Unit Latin America and the Caribbean Region Document of the Worl Bank  ACRONYMS AND ABBREVIATIONS ANC Ante-Natal Care ARI Acute Respiratory Infections BCG Vaccine against Tuberculosis CISRESA Infrastructure Census (Censo de Infraestructura Sanitaria y Recursos del Sector Salud) CLAS Community-managed Publicly Financed Health Committee (Comite Local de Administraci6n de Salud) DALY Disability-Adjusted Life Year DPT Vaccine against Diphteria, Pertussis and Tetanus DGSP Office of the Director General of Health DIGESA Office of the Director General of Environmental Health (Direcci6n General de Salud Aimbiental) ENDES Demographic and Health Survey (Encuesta de Demografia y Salud) ENNIV Living Standards Measurement Survey (Encuesta de Niveles de Vida) ESAN Graduate Business School (Escuela Superior de Admmnistraci6n de Negocios) ESSALUD New name for IPSS HIS Health Information System IDB Inter-Amnerican Development Bank IEC Information, Education, and Communication IMR Infant Mortality Rate INS National Health Institute (Instituto Nacional de Salud) PPF Project Preparation Facility IPSS Peruvian Social Security Institute (Instituto Peruano de Seguridad Social) LAC Latin America and the Caribbean LHP Local Health Plans LSMS Living Standards Measurement Survey MCH Matemal and Child Health MEF Ministry of Economy and Finance (MinisterHo de Economia y Finanzas) MINPRE Ministry of the Presidency (Ministerio de la Presidencia) MINSA Ministry of Health (Ministerio de Salud) MCH Matemal-Child Insurance (Seguro Materno Infantil) NGO Nongovemmental Organizations OGP MINSA's Planning Office (Oficina General de Planif caci6n) ORT Oral Rehydration Therapy FONCODES Social Investment Fund (Fondo de Compensaci6ny Desarrollo Social) PACFO Complementary Food Program PAHlO Pan American Health Organization PFSS Strengthening Health Services Project (Proyecto de Fortalecimiento de los Servicios de Salud) PHC Primary Health Clinics PSBPT Basic Health for All Program (Programa de Salud Bdsica para Todos) PSMU Public Sector Modemization Unit (Unidad Coordinadora de Modernizaci6n del Sub-sector Pzublico de Salud) PSNB Health and Basic Nutrition Project (Proyecto de Saludy Nutrici6n Bdsica) REDES Health Networks SE Insurance for schoolchildren (Seguro Escolar) SERUM Intem Service in Rural and Urban Marginal Areas (Servicio Ruraly Urbano Marginal) SMI Matemal-Child Insurance (Seguro Materno Infantil) UNICEF United Nations Children's Fund UNFPA United Nations Fund for Population Activities USAID United States Agency for International Development WHO World Health Organization Government Fiscal Year January 1- December 31 Currency Equivalents Currency Unit= Soles Vice President: Shahid Javed Burki Country Director: Isabel Guerrero Sector Director: Xavier Coll Sector Specialist: Charles Griffin Task Manager: Daniel Cotlear  - TABLE OF CONTENTS PREFACE AND ACKNOWLEDGMENTS .........................................................;i RESUMEN EJECUTIVO EN ESPAINOL .............................................. ......................................... ii 1. EXECUTIVE SUMMARY AND INTRODUCTION ......................................................................I Reforms in Health Provision .......................................................................3 Reforms in Health Financing ......................................................................4 Management of Programs Directed to the Poor ......................................................................5 Human Resources for Health Care ......................................................................6 Proposed Refonns ......................................................................7 Outline of the Main Report ......................................................................7 2. OVERVIEW OF HEALTH SECTOR FINANCING AND DELIVERY SYSTEMS .................... 9 Overview of Health Expenditures and System Resources .....................................................................9 Health Care Access and Use ...................................................................... 11 MINSA, IPSS and the Private Sector as Providers ...................................................................... 12 Benefit Incidence of Health Expenditures ...................... ............................................... 13 3. PROGRESS IN HEALTH OUTCOMES ...................................................................... 15 4. REFORMS IN HEALTH PROVISION ..................................................................... 23 MINSA and the Distribution of Health Benefits ...................................................................... 23 MINSA: Structure and Evolution of Expenditures ...................................................................... 24 National Health Programs of MINSA ...................................................................... 25 Targeting in MINSA ..................................................................... 27 Productivity in Primary Health Clinics ..................................................................... 29 New Initiatives to Improve Management and Community Participation ............................................. 30 Equity and Efficiency in Public Hospitals ...................... ............................................... 35 5. REFORMS IN FINANCING OF SERVICES ..................................................................... 41 The Economic Barrier ..................................................................... 41 The Public Insurance Schemes ..................................................................... 42 6. STRENGTHENING PLANNING, MANAGEMENT AND INFORMATION SYSTEMS IN MINSA ..................................................................... 45 Fragmentation of Policy-making and Regulatory Functions .............................................................. 45 Weakness of the Information Systems ...................................................................... 47 7. HUMAN RESOURCES FOR HEALTH CARE ..................................................................... 51 The labor Market for Health Professionals ............................ 51 Geographical Inequality .52 Quality of Training .54 Mismatch between Skills-mix and Needs .54 8. CONCLUSIONS AND RECOMMENDATIONS FOR A REFORM AGENDA FOCUSING ON THE POOR .57 Health Priorities .57 Reforms in the Provision of Services .58 Reforming the Financing of Care .60 Reorganization of MINSA .61 Reforming Human Resource Policies .62  TABLES Table 2-1: Peru: Main Sources of Finance for the Health Sector, 1997 .................................................... 9 Table 2-2: International Comparison of Health Expenditures ........................................................... 10 Table 2-3: Health Expenditures by Subsector, 1997 ........................................................... 10 Table 2-4: Infrastructure and Staff of Health Services by Subsector ..................................................... 11 Table 2-5: Availability of Health Inputs, 1997 ............................................................ 11 Table 2-6: Peru: Use of Health Care in 1997 ............................................................ 12 Table 2-7: Importance of Providers by Region and Income Level 1997 ................................................. 12 Table 3-1: Health Indicators. Peru 1991-1996 ........................................................... 15 Table 3-2: Coverage of National Programs ........................................................... 16 Table 3-3: Indicators of Maternal and Child Health: Urban-Rural Comparison, Peru ............................ 20 Table 4-1: Improvement in Service Delivery in CLAS and Non-CLAS facilities .................................. 31 Table 4-2: Revenue from Tariffs in Lima Hospitals ........................................................... 36 Table 4-3: Bed Occupancy Rate ........................................................... 37 Table 5-1: Out-of-Pocket Health Expenditures by Quintile ........................................................... 41 Table 6- 1: Fragmentation of Nutrition Programs ............................................................ 46 DIAGRAMS Diagram 2-1: MINSA's Market Share of Outpatient Services in 1994 and 1997 ................................... 13 Diagram 2-2: Per Capita Health Expenditures by Quintiles in 1997 ....................................................... 14 Diagram 3-1: Peru: Population Growth and Urbanization ........................................ ................... 15 Diagram 3-2: Deviations from Predicted IMR in 1996 for Selected Countries ....................................... 16 Diagram 3-3: Peru: Infant and Under-5 Mortality Rates by Quintile ...................................................... 19 Diagram 3-4: Cause of Mortality by Poverty Group, Peru 1996 ............................................................ 19 Diagram 3-5: Access to Key Services by Quintiles ............................................................ 20 Diagram 4-1: Distribution of MINSA Expenditures by Population Quintile .......................................... 23 Diagram 4-2: Public Health Sector Budget 1992-1998 ............................................................ 25 Diagram 4-3: Budget Distribution for National Health Programs 1998 .................................................. 25 Diagram 4-4: Per Capita Expenditures on Health ............................................................ 27 BOXES Box 1-1: The Structure of the Health Sector ........................................................... 2 Box 3-1: Utilization of Obstetrical Services in Rural Peru ........................................................... 18 Box 3-2: Emerging and Re-emerging Diseases in Peru ........................................................... 21 Box 4- 1: Structure of the MINSA Budget ............................................................ 24 Box 4-2: Participation of Civil Society in Health Services: Experience of Other Latin American Countries ........................................................... 34 Box 4-3: The Origins of Autonomy for Hospital Managers ............................................................ 37 Box 4-4: Allocative Costs of Unregulated Hospital Autonomy ............................................................ 38 Box 5-1: Bolivia's Public Mother and Child Insurance ........................................................... 43 Bibliography ........................................................... 65  PREFACE AND ACKNOWLEDGEMENTS Recent developments in Peru's health sector have been extraordinary. Since the early 1990s, both health care and health care expenditure have multiplied, with increased emphasis being given to primary care and improved access to services in poor rural and urban areas. Concurrently there has been enormous institutional innovation in the sector. Many new programs have been implemented, and many more new ideas have been proposed. This report was requested by the Government as an instrument to help it grasp where the sector stands today and what the priorities for the future should be. Its twin objectives are: (i) to help the government's efforts to formulate a strategy to continue to improve the health outcomes of the poor; and (ii) to guide World Bank and other development agency activities to help implement such a strategy. The report was prepared by the World Bank as a contribution to its ongoing policy dialogue with the Government of Peru. It is based on the findings of a World Bank mission that visited Peru in April 1998 and on numerous reports prepared during the last two years by our colleagues at the Ministry of Health. The mission members who wrote contributions were: Daniel Cotlear (task team leader and main author of this report), Mari Sol Concha (epidemiology and service provision), Tarcisio Castafieda (targeting and community participation), and Richard Webb (labor markets and primary care). Cuanto S.A. processed the survey data used in the report. The initial report was updated, based on numerous discussions held in 1998 and 1999 with officials from the Ministry of Health and the Ministry of Finance as well as interviews with members of universities, research centers and nongovernmental organizations. Laura Altobelli (public health specialist) and Jonathan Cavanagh (editor) contributed to the final version, which was patiently desktop-edited by Patricia Bernedo. The study was initiated while the Minister of Health was Marino Costa and was concluded while the Minister was Alejandro Aguinaga. Both of them, their key advisors and many officials and consultants were extremely helpful to the preparation of the study. Special thanks go to Augusto Meloni who helped us understand the sector and its issues through many hours of discussion and to Carlos Bardales, Danilo Fernandez, Pedro Francke, Ariel Frisancho, Alvaro Gaillour, Diego Gonzalez, Jaime Johnson, Ulises Jorge, Doris Lituma, Luis Manrique, Pedro Mendoza, Percy Minaya, Margarita Petrera, Nina Sotomarino, Raul Torres, Victor Zamora y Eduardo Zarate. Comments and guidance were received at the Bank from Evangeline Javier, who contributed to the report at all stages of the process and from Charles Griffin, Livia Benavides, Alex Precker, Davidson Gwatkins, Christopher Lovelace and Emesto May. Silvia Raw and Amanda Glassman from the IDB and Luis Seminario from USAID provided useful comments. In Peru, the final draft report was discussed in April 1999 with the top authorities of the Ministry of Health, the regional health directors, officials from the Ministry of Finance, the Health Commission of the National Congress, academics, NGOs and donors. i  Resumen Ejecutivo e Introducci6n El sector de salud peruano se ha recuperado rapidamente tras el colapso ocurrido entre fines de la decada del 80 y comienzos de los anios 90 como resultado de la hiperinflaci6n y el terrorismo. Esta recuperaci6n ha sido acompafnada por importantes reformas en los servicios de atenci6n primaria del Ministerio de Salud (MINSA) y ha contribuido a una mejora significativa en la salud de la poblaci6n. Este estudio evalua las principales reformas en atenci6n primaria introducidas por el MINSA y hace recomendaciones para continuar mejorando la atenci6n de salud para los grupos de menores ingresos. El gasto piiblico y privado total en el area de salud aument6 en terminos reales en mas del 50% en los tres afios posteriores a 1994. La oferta de servicios de salud aument6 de manera pronunciada, particularmente en la atenci6n primaria de salud: el numero de clinicas de salud primaria aument6 en dos tercios y se extendi6 su horario de atencion. Hubo tambien un aumento del 55% en los puestos de empleo para profesionales de salud, mayormente en atenci6n primaria. La demanda de servicios tambien se increment6. En el ambito nacional, el uso de servicios creci6 en 59% en s6lo tres anios. En las zonas rurales, se vio un crecimiento similar al promedio nacional. Los indicadores de salud tambi6n han mejorado rapidamente en los ultimos afios. Esto ha estado asociado a la mejora en los ingresos y las condiciones de vida y a la expansi6n de servicios de salud. La mortalidad infantil y la desnutrici6n infantil, por ejemplo, disminuyeron en aproximadamente un 30% durante la segunda mitad de la decada y es probable que hayan continuado mejorando desde entonces. Pese a estas mejoras tanto en los recursos como en los resultados, los encargados de formular las politicas de salud enfrentan todavia tres desafios y preocupaciones fundamentales. En primer lugar, c6mo continuar reduciendo la brecha entre el nivel de salud de los pobres y el de los otros grupos. El Peru sigue siendo un caso atipico en America Latina, puesto que a pesar de las mejoras recientes, todavia exhibe un indice de mortalidad infantil muy alto para un pais con su nivel de ingresos. Los altos indices de mortalidad infantil se concentran en la poblaci6n pobre, mientras que los indicadores de los grupos de mayores ingresos se aproximan mas al promedio latinoamericano. La poblaci6n pobre, dividida entre las zonas rurales y urbanas, es mas vulnerable a la mala salud debido a una combinaci6n de bajo nivel educativo, condiciones ambientales insalubres y acceso limitado a los servicios de salud. Los pobres se encuentran mas expuestos que el resto de la poblaci6n a las enfermedades transmisibles, muchas de ellas exacerbadas por problemas ambientales, tales como instalaciones sanitarias deficientes y transmisi6n vectorial. Aunque el acceso a servicios primarios para nifios de mas de un mes de edad ha mejorado enormemente en los ultimos afios, los pobres continuan sufriendo por la falta de acceso a servicios para madres y para recien nacidos, quienes requieren servicios de atenci6n primaria y tambien prestados en hospitales. En segundo lugar, c6mo incrementar los recursos asignados a la atenci6n de salud de la poblacion pobre. El Peru, a pesar del incremento de afios recientes, todavia sigue asignando menos recursos a la salud que la mayoria de sus vecinos. La proporci6n del ii  PIB asignada al sector de salud, 4.1% en 1997, equivale a unos dos tercios del promedio latinoamericano. La disponibilidad de medicos, 10 por cada 10,000 habitantes, representa s6lo un 70% del promedio latinoamericano. Tambien existe gran desigualdad en el consumo de bienes y de servicios de salud. El consumo per capita de bienes y servicios de salud es aproximadamente 4.5 veces mayor entre el 20% del nivel socioecon6mico mas alto que en el 20% del nivel mas bajo. Aunque gran parte de esta diferencia se debe al mayor gasto privado en atenci6n medica por los grupos de mayores ingresos, dicha diferencia se ve s6lo parcialmente compensada por el Ministerio de Salud (MINSA), el cual te6ricamente deberia servir a los sectores pobres, pero en la practica asigna a los pobres sumas similares a las asignadas a los otros grupos. El tercer desaflo es como utilizar mas eficientemente estos recursos. El gasto total en la salud asciende a aproximadamente US$2,700 millones al anio, divididos en partes iguales entre los sectores privado y puTblico (vease la casilla 1). Aproximadamente la mitad del gasto puiblico se canaliza a traves del MINSA, el cual asigna un poco menos de la mitad de sus recursos a la atenci6n primaria de salud. El sistema sufre de importantes deficiencias, reflejadas en la coexistencia de una gran capacidad subutilizada en todos los subsectores, con necesidades insatisfechas y congesti6n de algunos servicios. Las ineficiencias mas importantes se deben a la fragmentaci6n, la ausencia de competencia y la ausencia de separaci6n entre el financiamiento y la provisi6n. Tambien existen deficiencias significativas en los centros y puestos de atenci6n primaria (CPAPS), reflejadas en baja productividad y una adecuaci6n insuficiente a las necesidades locales. Casilla 1. La estructura del sector de salud El sistema de salud peruano es una amalgama compleja de varios programas puiblicos y un sector privado amplio, cada uno de los cuales es simultaneamente proveedor y fmanciador de servicios. Cada subsector sigue un curso independiente; existe poca coordinaci6n y casi ninguna competencia entre proveedores o entre fmanciadores de servicios. Los programas puiblicos principales son: (i) el Ministerio de Salud (MINSA), fmanciado por recaudaciones tributarias y pagos directos de los usuarios, que en teoria presta servicios a los sectores pobres, y (ii) el ESSALUD (anteriormente Instituto Peruano de Seguridad Social -- IPSS), financiado por un impuesto del 9% sobre planillas de sueldos, que en teoria cubre todas las necesidades de atenci6n medica de sus contribuyentes los cuales son mayormente trabajadores del sector forrmal de la economia. El sector privado se fmancia casi en su totalidad por pagos en efectivo de los pacientes y sus familias, ya que la cobertura de los seguros privados es muy reducida y los programas publicos (con muy pocas excepciones) no contratan servicios de proveedores privados. Seguin encuestas de hogares, el MINSA desempeffa un papel importante en la prestaci6n de servicios clinicos, ya que presta dos tercios de los servicios de hospitalizaci6n y 44% de las consultas ambulatorias. El sector privado tambien es muy importante como proveedor de servicios ambulatorios, tanto en zonas urbanas como rurales, y tambien para los pobres. El IPSS, que constituye el 25% del gasto nacional en la salud, presta 18% de los servicios ambulatorios y 23% de los servicios de hospitalizaci6n, concentrandose estos servicios en las ciudades principales y exclusivamente en beneficio de la poblaci6n con recursos. Se realizo un analisis de la incidencia de beneficios a fin de entender la funcion de los diversos subsectores en la atenci6n de los pobres. El consumo de la atenci6n medica es una combinaci6n de un elemento privado con el gasto canalizado a travds del MINSA y el IPSS. La mayor parte de la desigualdad en el consumno de servicios de salud resulta de la desigualdad en el ingreso, pues los grupos de mayores ingresos pueden destinar mas recursos privados a la atenci6n medica que los pobres. El IPSS contribuye a esta situaci6n al atender unicamente a los empleados del sector formal y sus familias, quienes estan concentrados en los niveles mas altos de la distribuci6n de ingresos. El gasto de MINSA se distribuye por igual entre ricos y pobres. iii  Se ha intentado varias reformas durante los anos de recuperaci6n del sector. Algunas estan rindiendo fruto, otras fracasaron. Hubo varios intentos fallidos de impulsar refornas extensas en el sector, con el prop6sito de superar las grandes ineficiencias creadas por la fragmentaci6n y la ausencia de competencia.1 La innovacion que alcanz6 el mayor exito consisti6 en la creaci6n de extensos programas focalizados. Estos apoyan la atenci6n primaria mediante el financiamiento de trabajadores de salud asignados a zonas seleccionadas con el uso de un mapa de pobreza, y mediante la asignaci6n de fondos para la nutrici6n y la lucha contra las enfermedades que afligen a los pobres. En algunas zonas, los programas focalizados han servido para experimentar con formas de organizaci6n que incentivan lograr la participaci6n de la comunidad.2 Este informe se ha elaborado con el objetivo de servir como insumo para que el gobiemo continuie desarrollando su agenda para mejorar la atenci6n de salud de los pobres. Se centra en la atenci6n a los pobres y, particularmente, en el subprograma de atenci6n primaria de la salud del MINSA, mas que en reformas de todo el sector destinadas a una mayor eficacia. Se opt6 por esta concentraci6n porque reducir la brecha entre el estado de salud de los pobres y de los otros niveles socioecon6micos es una prioridad declarada tanto del gobierno como del Banco. Al mismo tiempo, el entomo politico y econ6mico esta cambiando de manera tal que podria poner en riesgo los adelantos logrados en la prestaci6n de servicios a los pobres. La amenaza del terrorismo, que fue un incentivo importante para la creaci6n de programas focalizados, ha disminuido. En el ambito econ6mico, los altos indices de crecimiento y la relativa abundancia fiscal que facilitaron la introducci6n de los programas focalizados se han visto afectados por las repercusiones de la crisis asiatica. Un tema central que resulta de este informe y aparece reflejado en sus conclusiones y recomendaciones es que seria de gran beneficio, en este nuevo entomo y ante los desafios delineados anteriormente, mantener y profundizar las reformas destinadas a mejorar los servicios de salud para los pobres. Para que esto suceda, se tendran que resolver problemas en la prestaci6n, financiamiento, administracion y dotaci6n de personal de los servicios de salud. Estos problemas se sintetizan en este resumen ejecutivo y se describen en mayor detalle en los capitulos siguientes. Reformas en la prestacion de servicios de salud Encuestas de hogares realizadas en 1997 demuestran que el MINSA gasta una cantidad similar per capita en familias de bajos y de altos ingresos. Es decir que el gasto del MINSA es neutral en relaci6n al ingreso, en lugar de ser progresivo concentrandose en los pobres. Tambien sefialan que los pobres reciben una mayor proporci6n del gasto en CPAPS que en hospitales. Esta distribuci6n del gasto se da a pesar que el gobierno ha ILos intentos fallidos consistieron en leyes ambiciosas que intentaron separar el financiamiento de la provisi6n de servicios y reducir el exceso de empleo en el MINSA. Tambien fracas6 un intento de abrir la opci6n que los trabajadores asalariados pudiesen optar por abandonar totalmente el IPSS transfiriendo su cotizacion obligatoria a asegurados privados. 2 Ha habido tambi6n algunos avances importantes para aumentar la competencia frente al IPSS, tales como: (i) la posibilidad de optar por no incluir en el IPSS una parte de la cobertura de salud (asignando un cuarto del impuesto sobre la planilla de pagos originalmente destinado al IPSS a EPSs privadas); (ii) la anulaci6n del monopolio del IPSS sobre el seguro de riesgo para los trabajadores; y (iii) el desarrollo de sistemas de informaci6n que permiten el pago a cada hospital del IPSS a partir de los servicios producidos por el hospital, contabilizando en forma transparente cualquier subsidio que adicionalmente se asigne a ese hospital para cubrir su exceso de gasto. iv  introducido en afnos recientes importantes innovaciones que han logrado incrementar la prestaci6n de servicios del MINSA para los pobres mediante: (i) la asignaci6n de mayores recursos econ6micos a la atenci6n primaria de salud; (ii) la asignaci6n de fondos segun los niveles regionales de pobreza; y (iii) el aumento de la participaci6n de la comunidad en la administraci6n de servicios. Estas innovaciones se presentan a continuaci6n, conjuntamente con una indicaci6n de sus debilidades. Hasta la fecha no se ha hecho ninguin intento de orientar a los hospitales puiblicos, que funcionan actualmente de manera semicomercial, hacia una mayor prestaci6n de servicios a los pobres. El incremento del gasto en la atenci6n primaria se logr6 mediante la creaci6n de programas nuevos que utilizaron recursos frescos del tesoro, sin reasignar fondos existentes y sin introducir reformas en los prograrnas y servicios tradicionales. Simplificando una estructura compleja, los proveedores de servicios del MINSA se pueden dividir en tres categorias, cada una de las cuales recibe financiamiento gubemnamental a traves de canales diferentes: Hospitales Nacionales, financiados directamente por el fisco; Hospitales Regionales financiados por el fisco a traves de los gobiemos regionales; y CPAPS, financiados parcialmente a trav6s de los gobiemos regionales y cada vez m$s a traves de los programas focalizados. El presupuesto total del gobiemo para estos proveedores se duplic6 en terminos reales durante 1994-97. La mayor parte del aumento se asign6 a los programas focalizados, que recibieron un presupuesto de unos US$150 millones en 1998, o alrededor del 30% del total. El uso de fondos nuevos para fmanciar los programas focalizados posterg6 la necesidad de enfrentar a los grupos organizados en los servicios tradicionales. Esto, que fue una ventaja al permitir la rapida introducci6n de programas focalizados es hoy una fuente de vulnerabilidad. Bajo la situaci6n actual, los programas nuevos probablemente seran los mas afectados por las reducciones requeridas por una politica fiscal mas austera. Un analisis de la distribuci6n del presupuesto entre los diferentes departamentos del pais sefiala que los programas focalizados asignan una alta proporci6n de sus recursos a los departamentos mas pobres. En teminos per capita, las transferencias a los departamentos mas pobres son cinco veces mayores que las que recibe Lima (el departamento mas rico). Por el contrario, en terminos per capita, los presupuestos regionales se muestran altamente sezgados a favor de los departamentos mas ricos. Si se suman los presupuestos regionales a los recursos de los programas focalizados, las transferencias del MINSA son proporcionales a la distribuci6n de la poblaci6n. Todavia existe amplia cabida para mejorar la asignaci6n geografica del 30% del presupuesto del MINSA destinado a los programas focalizados, mediante la mejora de los mapas de pobreza y mediante un mayor control para asegurar que los recursos se distribuyan eficazmente a las zonas mas pobres. Las oportunidades mas importantes de dirigir mas recursos hacia los pobres no estan al interior de los programas focalizados, sino en los servicios tradicionales y en particular en los hospitales. Es importante hacer esfuerzos por introducir medidas que permitan a los pobres l5eneficiarse mas de los hospitales. Con el fin de mejorar la eficacia en la atenci6n primaria de la salud, se estan desarrollando nueva formas de organizaci6n con la participaci6n de la comunidad. La participaci6n de la comunidad tambi6n podrian servir de catalizador para el desarrollo de grupos organizados v  que podrian ayudar a defender los programas nuevos contra reducciones presupuestarias. Estas reformas son necesarias, puesto que la productividad es baja y las actividades se instauran frecuentemente sin utilizar una evaluaci6n diagn6stica local y a menudo se ven restringidas por reglamentos innecesariamente burocraticos. La baja productividad de los CPAPS es particularmente preocupante, ya que el promedio nacional es de 1 a 2 consultas clinicas al dia por trabajador de salud. Es probable que los datos actuales subestimen los niveles de productividad al excluir o no contar la totalidad del trabajo preventivo y externo, pero incluso los calculos conjeturales corregidos arrojan resultados muy bajos. Los CLAS son centros y puestos de salud administrados por la comunidad que implementan Programas de Salud Locales con fmanciamiento del Gobiemo. Los indicadores disponibles sugieren que los CLAS, que administran actualmente 10% de las clinicas del MINSA, han dado buenos resultados en el aumento de la eficacia de la prestaci6n de la atenci6n primaria de salud. Estos buenos resultados se han logrado pese a la creciente oposici6n de las burocracias de los gobiemos regionales, que resienten la perdida de control directo sobre los CPAPS y no proporcionan el apoyo tecnico requerido a los CLAS. El otorgamiento de mas permisos para la expansi6n de los CLAS ha estado congelado desde 1997 (pero podria reabrirse en un futuro cercano). Reformas en el financiamiento de los servicios de salud Los programas focalizados han logrado expandir la cobertura geografica de los servicios de salud, al hacerlos funcionar y mejorar su calidad en zonas remotas. No obstante, muchos de los pobres todavia no tienen acceso a la atenci6n de la salud debido a sus costos directos e indirectos. La mayoria de los proveedores del MINSA tratan de aliviar los costos para los pobres mediante exoneraciones parciales o totales del pago por concepto de servicios. Este sistema de exoneraciones tiene tres puntos debiles. En primer lugar, no existe un fondo para subsidiar medicamentos y suministros a nivel del proveedor, pero estos constituyen mas del 80% del costo directo de la atenci6n m6dica para los pobres. La mayor parte de los medicamentos y suministros medicos, financiados por el centro medico a traves de los cargos a los usuarios, se venden al usuario a su costo total m6as un margen de utilidad. En segundo lugar, la generosidad local financia las exoneraciones para los pobres. Puesto que no existe un medio para subsidiar especificamente a los pobres, cada centro fmancia los costos de esta atenci6n con sus propios recursos y toma voluntariamente la decisi6n de asignar tales recursos a la prestaci6n de servicios a los pobres. Las necesidades son mayores en algunas zonas que en otras y no existen mecanismos que compensen esta disparidad. Por uiltimo, no existe ningin criterio ni metodologia estandar para identificar a los pobres. Cada centro desarrolla su propio sistema y la mayor parte del tiempo lo aplica de manera irregular. El gobiemo ha empezado a afrontar el problema del costo para los pobres mediante la introducci6n de esquemas disefiados con el fin de proporcionar acceso a servicios cruciales para grupos seleccionados. El Seguro Escolar se cre6 en 1997 para cubrir servicios de salud y medicamentos para todos los alumnos de colegios publicos entre los 3 y los 17 afios de edad (alrededor de 6 millones). El gobierno ha decidido crear un Seguro Materno Infantil (SMI), que cubriria servicios basicos para madres y nifios menores de 3 afios de edad, nutriendose de una creciente experiencia internacional que por ejemplo en Bolivia vi  aument6 la cobertura institucional de partos en un tercio en s6lo 18 meses. Ambos esquemas eliminan los pagos por los usuarios en el lugar de uso del servicio y cubren los medicamentos y otros insumos recetados. El Seguro Escolar es gratuito para los beneficiarios. El gobiemo esta estudiando la posibilidad de ofrecer el SMI, requiriendo el pago de una pequenia prima de seguro subsidiada (posiblemente gratuita en las regiones mas pobres). Se piensa introducir el SMI como un esquema aparte; una vez consolidado, se establecerfa un solo esquema de seguros publicos que incorporaria el Seguro Escolar y posiblemente permitiria la opci6n de utilizar proveedores privados -en un primer momento ambos seguros estan limitados a proveedores publicos. Pese a los resultados positivos del Seguro Escolar, que cubrio 4 millones de atenciones durante 1998 y puede haber mejorado considerablemente la cobertura de escolares, y pese a los resultados prometedores de un pequefio experimento piloto con el SMI, el gobiemo se muestra vacilante en crear y expandir el SMI. Las autoridades econ6micas estan preocupadas por el costo. Algunos proveedores de salud vacilan como consecuencia de los problemas logisticos que se han presentado en la aplicaci6n practica del Seguro Escolar. Urge afrontar estos problemas para evitar que se creen cuellos de botella en el servicio y se desacredite un esquema prometedor. En el caso del Seguro Escolar, los reembolsos son lentos y los centros de salud se han visto forzados a cubrir muchos gastos con sus propios ingresos, ya que los reembolsos s6lo cubren el costo de los medicamentos, mientras que los pagos directos por los pacientes cubren otros costos, incluyendo beneficios para el personal (tales como canastas de viveres). Administraci6n de programas dirigidos hacia los pobres Existen ineficiencias en la administraci6n de los programas claves del MINSA, lo que restringe su capacidad de establecer prioridades y prestar servicios de salud a los pobres. Muchos de estos problemas surgen de la fragmentaci6n de los programas clave y de sistemas de informaci6n deficientes. El problema de fragmentaci6n y duplicaci6n es particularmente agudo en el control de programas de salud matemo infantil, nutrici6n y salud ambiental. Ninguno de estos programas cuenta con una estructura directiva clara. En teoria, la politica y planificaci6n de las actividades matemo infantiles es una de las muchas responsabilidades de la Direcci6n General de Salud de las Personas (DGSP). Dentro de la DGSP, esta responsabilidad esta dividida en 8 programas nacionales. Dos proyectos grandes y muchos pequefios fmanciados externamente tambien financian estas actividades. Cada uno de estos proyectos y programas lleva a cabo su propia planificaci6n, desarrolla sus propios protocolos y planifica y financia sus propios programas de capacitaci6n. En el caso de la salud ambiental, existen superposiciones y a menudo duplicaciones del control de calidad de alimentos y el control de la transmisi6n vectorial. En el area de la nutrici6n, se presentan superposiciones en las funciones desempefiadas por la DGSP, el INS y muchas otras instituciones puiblicas y privadas del sector fuera del MINSA. Un sistema presupuestario fragmentado sostiene esta duplicaci6n. La informaci6n sobre los gastos y sobre la producci6n de servicios de salud en el Peru es muy debil y los intentos de corregirla sufren de falta de continuidad y de consistencia. Pocos paises en vias de desarrollo cuentan con un buen nivel de calidad en la infornaci6n vii  sobre el sector de salud, pero la mayoria tiene al menos informaci6n sobre el sector puiblico. En el Peru no existe informaci6n consolidada sobre flujos financieros del sector puiblico de salud, pues M1NSA s6lo conoce los gastos del departamento de Lima y ninguna agencia consolida los datos de otras regiones. Tampoco se conocen los datos sobre ingresos propios de los hospitales.3 Las estadisticas de producci6n de los servicios han dejado de publicarse y de analizarse desde hace varios anjos. Los calculos de consultas ambulatorias del MINSA para 1995 (las mas recientes disponibles al momento de preparar este informe) fluctuan entre 15 millones (estadisticas oficiales) y 27 millones (calculos a partir de encuestas de hogares). No existe una fuente de informaci6n oficial sobre consultas o intemamientos en hospitales. Cada programa produce sus propios datos y se efectuan pocos intentos serios de consolidarlos de alguna manera que permita conocer y monitorear las actividades en conjunto (por ejemplo las actividades de control de la transmisi6n vectorial, las actividades de nutrici6n, hasta la producci6n en laboratorios, se miden por los diferentes programas de formas que no se pueden sumar ni comparar). Ocurren problemas similares con la medici6n de suministros, incluso partidas costosas como personal, capacitaci6n o el abastecimiento de equipo, puesto que cada programa o fuente de financiamiento lleva sus propios registros y no existe una oficina de recursos humanos ni de infraestructura en el MINSA que consolide dicha informaci6n. Recursos humanos para la atenci6n de salud Muchas de las ineficacias y disparidades en el sistema de salud tienen su origen en problemas de recursos humanos. De particular importancia para la prestaci6n de servicios a los pobres son factores tales como distribuci6n geografica, mezcla de habilidades y calidad de los recursos humanos. La distribuci6n geografica ha mejorado notablemente en las ultimas dos decadas, ya que el crecimiento en la poblaci6n y los ingresos de las ciudades pequefias ha propiciado que mas medicos se establezcan en ellas. Mas recientemente, los programas focalizados han colocado a mas de 10,000 trabajadores de salud en zonas menos favorecidas. Estos programas ofrecen grandes incentivos econ6micos, pero el personal no recibe beneficios ni seguridad laboral. Pese a los incentivos econ6micos y a la sobre oferta temporal de los trabajadores de salud, la rotaci6n de personal en las zonas remotas es sumamente elevado porque los sueldos mas altos no compensan por el efecto combinado de la atracci6n de carreras profesionales vinculadas a la especializaci6n medica y el ejercicio privado de la profesi6n que ofrecen las ciudades, y la carga de vivir en un entorno cultural radicalmente diferente. Esto se ve exacerbado por la ausencia de expectativas de una carrera de largo plazo en las zonas remotas, pues los contratos son breves. En vista de este dilema, los especialistas estan buscando soluciones alternativas al problema de la atenci6n de la salud de las comunidades rurales. Una opci6n que se viene estudiando implicaria un cambio en la combinaci6n de habilidades de los trabajadores de salud locales, complementado por vinculos mas fuertes con el resto de la red de salud. Los trabajadores locales asumirian mas responsabilidades sobre el aspecto de salud puiblica, con un enfasis menor en los aspectos clinicos. Se necesitaria fortalecer las estructuras administrativas de los trabajadores locales, asi como los sistemas de 3Hay en curso un intento decrear un sistema de cuentas nacionales de salud buscando superar los problemas de medici6n de intentos anteriores. viii  comunicaciones y de referencia y, en algunas zonas, se introduciria el uso de medicos m6viles y "tecnologia de telesalud". El problema con la combinaci6n de habilidades es que la formaci6n medica no concuerda con el desplazamiento de las prioridades nacionales hacia la salud primaria y preventiva ni con el correspondiente desplazamiento hacia modelos de prestaci6n de salud en el ambito comunitario y rural. El desfase educativo se debe en parte a la atenci6n relativa prestada a diferentes patologias medicas y en parte a la ausencia de conocimientos y habilidades de orden no medico relacionados con la salud puiblica y el trabajo comunitario. Asimismo, la formacion de profesionales medicos no se ha mantenido al dia con la rapidamente creciente importancia de los sistemas, el trabajo en equipo y la informaci6n en la prestaci6n de servicios de salud y, por lo tanto, con la necesidad de desarrollar habilidades administrativas y directivas. El sistema educativo esta empezando a responder a estos desfases, pero se necesitara de mayor esfuerzo y se necesita tambien dar senales claras a los j6venes egresados y los estudiantes sobre las nuevas prioridades. El consenso entre los funcionarios del MINSA y los profesionales medicos mas eminentes es que la calidad de la educaci6n medica se ha deteriorado. Mayormente, esto es un producto de la reciente creaci6n a gran escala y poco reglamentada de facultades de medicina nuevas y el aumento en el nuimero de institutos de ense-nanza para tecnicos de salud. Las universidades deben obtener licencias, pero en los uiltimos afios los permisos de operaci6n se han otorgado profusamente y, una vez en operaci6n, las universidades no necesitan aprobaciones gubemamentales ni profesionales para otorgar titulos profesionales en medicina. Las licencias de centros educativos no universitarios no requieren aprobaci6n ni certificacion profesional. Reformas propuestas Las principales conclusiones de este informe son que las refornas introducidas por el MINSA para la prestaci6n de servicios de atenci6n primaria han rendido fruto y deben mantenerse, expandirse y profundizarse. Con ese fin, el informne principal presenta una serie de recomendaciones detalladas, que pueden agruparse bajo cinco encabezamientos claramente definidos de iniciativas de reforma: i) centrar mas la atenci6n en las necesidades de salud de los pobres al asignar fondos puiblicos; ii) reforzar la focalizaci6n entre los proveedores de salud del MINSA (mejor focalizaci6n de los programas y del gasto regional, mayor acceso por los pobres a servicios de hospitalizaci6n, mejor y mayor participaci6n de la comunidad); iii) crear mecanismos de seguro medico nuevos para financiar la atenci6n de la salud de los pobres; iv) mejorar los sistemas de informaci6n y gesti6n necesarios para ejecutar eficientemente y sin duplicaci6n los programas de MINSA para los pobres; y v) afinar los requisitos de habilidades y los incentivos de los recursos humanos a fin de mejorar la ejecuci6n de los programas orientados hacia los pobres. Descripci6n del informe principal Centrandose en los desafios detallados anteriormente, este informe tiene la siguiente estructura. El Capitulo 2 resefia brevemente el sector de salud, con el fin de colocar al 1x  MINSA y a las reformas recientes en una perspectiva mas general (flujos fmancieros, el crecimiento de desembolsos para la salud y el acceso a los servicios, y un analisis de los subsectores beneficiados a partir de encuestas de hogar). El Capitulo 3 describe el progreso en los indicadores del estado de salud, asi como un analisis de las diferencias entre los niveles y las necesidades de salud de los pobres y del resto de la poblaci6n. El Capitulo 4 describe al MINSA en mayor detalle y analiza las reformas clave en la prestaci6n de salud, tales como: el mayor enfasis en la atenci6n primaria, el uso de la orientaci6n a grupos especificos y los incentivos nuevos para mejorar la atenci6n de grupos pobres, la gesti6n y la participaci6n comunitaria. Tambien plantea dos problemas fundamentales en los hospitales pliblicos. El Capitulo 5 analiza los obstaculos de costo y presenta dos esquemas actualmente en desarrollo para superar estos obstAculos: el Seguro Escolar y el Seguro Matemo Infantil. El Capitulo 6 examina las deficiencias en la administraci6n de los programas orientados hacia los pobres, con enfasis en la fragmentaci6n y la ausencia de informaci6n. El Capitulo 7 analiza los aspectos relacionados con los trabajadores de salud. Primero analiza las tendencias principales en los mercados laborales para estos trabajadores y luego profundiza en temas de disparidad geografica, calidad de la formaci6n y la discordancia entre la mezcla de habilidades y las necesidades. En el Capitulo 8 se presentan las conclusiones y recomendaciones principales. x  Peru: Improving Health Care for the Poor 1 1. EXECUTIVE SUMMARY AND INTRODUCTION The Peruvian health sector has recovered rapidly after collapsing in the late 1980s and early l990s as a result of hyperinflation and terrorism. Total public and private spending on health rose by over 50% in real terms in the three years after 1994. The supply of health services increased sharply, especially in primary care: the number of primary care clinics increased by two-thirds and their hours of operation were extended. The employment of health professionals rose by 55%, mostly for primary care positions. The demand for services also increased. Nationwide, the use of services grew by over 55% in only three years. In rural areas growth was almost 90%. Health outcomes have also improved rapidly in recent years. Infant mortality and child malnutrition, for instance, fell by almost a third during the first half of the decade and are likely to have continued to improve since. Despite these improvements in both inputs and outcomes, policy-makers face three key challenges and concerns. First, how to continue to reduce the large gap between the health status of the poor and that of the non-poor. Peru is still an outlier in Latin America, exhibiting a very high IMR for a country with its level of income. The high rates of infant mortality are concentrated among the poor. Indicators for the higher income groups are close to the average for Latin America. The poor, who are split between rural and urban areas, are made especially vulnerable to poor health by a combination of low levels of education, poor environmental conditions, and scant access to health services. The poor suffer to a greater extent than the rest of the population from communicable diseases, many of which are exacerbated by environmental problems such as poor sanitation or vector infestation. While their access to primary services for children over one month old has greatly improved in recent years, the poor continue to suffer from lack of access to services for mothers and young infants, who require primary and hospital-based services. Second, how to increase the resources assigned to provide care for the poor. Peru continues to allocate fewer resources to health than most of its neighbors. The proportion of GDP assigned to health, -4.1% in 1997- is about two thirds of the Latin American average. The ratio of physicians to the population, -10 per 10,000- is only 70% of the Latin American average. There is also great inequality in the consumption of health goods and services. Per capita consumption of health goods and services is approximately 4.5 times higher among the richest 20% of the population than among the poorest 20%. Whereas much of that difference is to be explained by greater amounts of private expenditure on health care by the rich, it is only partially offset by the Ministry of Health (MINSA), which in theory should be serving the poor, but in practice directs similar amounts to both the poor and the non-poor. Third, how to increase the efficiency in the use of these resources. Total health expenditures are around US$2.7 billion per year divided in equal parts between the  Executive Summary and Introduction 2 private and public sectors (see Box 1-1). Approximately half of the public expenditures are channeled by MINSA, which assigns about a fifth of that to primary health care. There are very large inefficiencies in the system, reflected in the coexistence of extensive underutilized capacity in all sub-sectors with unmet needs and pent-up demand for some services. There are, above all, widely recognized inefficiencies due to fragmentation, lack of competition, and lack of separation between financing and provision. There are also significant inefficiencies in Primary Health Clinics (PHC), reflected in low productivity and insufficient adaptation to local needs. Box 1-1 The Structure of the Health Sector Peru's health system is a complex amalgam of several public programs and a private sector, each of which tends to go its own way, with little coordination or competition between program providers, in either fnancing or delivery of services. The main public programs are: the Ministry of Health (MINSA), financed by tax revenues and co-payments by users and providing services theoretically directed to the poor; and the Social Security Institute (formerly IPSS, currently ESSALUD), financed by a 9% payroll tax and theoretically covering all the health care needs of its formal-sector contributors. The private sector is financed almost entirely by out-of-pocket expenditures by households, as private insurance coverage is very small and the public programs (with very few exceptions) do not purchase services from private providers. Household surveys show that MINSA plays a major role in the provision of clinical services, providing two-thirds of inpatient services and 44% of outpatient consultations. The private sector is also very important as a provider of outpatient services, in rural as well as urban areas and for the poor as well. IPSS, which accounts for 25% of national expenditure in health, provides 18% of outpatient services and 23% of inpatient services, all concentrated in the main cities and serving exclusively the non-poor. A benefit incidence analysis was performed to understand the role of the different subsectors in serving the poor. Consumption of health care combines a private element with benefits channeled by MINSA and IPSS. Most of the inequality in health care consumption results from the assignment of greater amounts of private expenditure to health care by the rich than by the poor. This is reinforced by IPSS, which serves only formal sector employees and their families, who are concentrated in the higher echelons of the income distribution. MINSA expenditures reduce the overall inequality, but are not large enough or sufficiently well targeted among the poor to significantly equalize expenditures. The importance of MINSA is largest for the bottom 20%, who obtain 70% of their health benefits from this source compared to 20% for the richest quintile. Several reforms have been tried during the years of recuperation of the sector. Some were successful, others failed. There were several failed attempts to pass legislation to introduce widespread reform in the sector to overcome the considerable inefficiencies created by the fragmentation, lack of competition and lack of separation between financing and provision that characterize the sector.4 The most successful innovation consisted of building large 4 The failed attempts included draft legislation to separate financing and provision and reduce excess employment in MINSA. There- was also a failed attempt to allow formal sector workers to transfer their compulsory payroll contribution from IPSS to private insurers. While the attempts of "big bang" reform failed, important progress was achieved in introducing reforms to increase competition to IPSS, including: (i) the possibility of opting-out of IPSS for part of health coverage (assigning a fourth of the payroll tax originally earmarked for IPSS to private "health promoters"; (ii) the elimination of IPSS's monopoly over occupational risk insurance; (iii) the creation of insurance policies in IPSS for independent workers; and (iv) the development of information systems that allow for payment of each IPSS hospital based on services produced, and make explicit any additional subsidy paid to the hospital.  Peru. Improving Health Care for the Poor 3 targeted programs to support primary care by allocating health workers based on regional priorities set by a poverty map, and by assigning funds for nutrition and to combat the diseases of the poor. In some areas the targeted programs have successfully piloted community participation. This report is produced as an input to be used by the Government to continue to develop its agenda to improve health care for the poor. Its focus is on poverty, and particularly on ways to improve MINSA's primary health subprogram, rather than on sector-wide reforms to increase efficiency. This focus was chosen because reducing the gap in health status between the poor and the non-poor is a stated priority for both the Government and the Bank. The emphasis on incremental change, instead of widespread reform, was chosen because recent history suggests that the conditions are not ready for widespread reform but that much can be achieved with incremental changes, and these could become part of a broader reform in the medium term. At the same time, both the political and economic contexts are changing in a way that could jeopardize the progress achieved in providing services to the poor. The threat of terrorism, which was a major incentive for the creation of some targeted programs, has receded. The high rates of growth and relative fiscal abundance that facilitated the introduction of the targeted programs have been affected by the reverberations of the Asian crisis. A leitmotif that emerges from this report and is reflected in its conclusions and recommendations is that much can be gained, in this new context and in response to the challenges outlined above, by sustaining and deepening the reforms directed toward improving health care for the poor. For that to happen, key outstanding issues in providing, financing, managing, and manning health services have to be resolved. These key issues are summarized here and described in more detail in the following chapters. REFORMS IN HEALTH PROVISION Household surveys for 1997 show that MINSA expenditures per capita benefit households of higher and lower incomes to a similar degree, instead of concentrating on the poor. They also show that the poor receive a larger proportion of the expenditure in Primary Health Clinics (PHC) than they do from hospital services. This pattern in the distribution of expenditures exists despite the introduction of successful innovations to increase the focus of MINSA service provision on the poor by: (i) increasing funding for primary health care; (ii) targeting funds according to regional poverty levels; and (iii) increasing community participation in running the services. These innovations are discussed below, with a brief indication of their remaining weaknesses. To date there have been no attempts to direct the public hospitals, which at present function in a semi-commercial way, to provide more services to the poor. The increased funding for primary care was achieved by the creation of new programs using fresh funds and involving no reorientation of existing funds. MINSA service providers can be divided into three categories, each of which receives government financing through different channels: National Hospitals, funded directly by the treasury; Regional Hospitals funded by the treasury through regional governments; and PHC, funded  Executive Summua and Introduction 4 partially through the regional governments and increasingly through the targeted programs created after 1994. The overall government budget for these providers doubled in real terms during 1994-97. Most of the increase was assigned to the newly created targeted programs, which received a budget of around US$150 million in 1998. The use of fresh cash to fund the targeted programs postponed any conflict with the powerful organized groups in the traditional services. Today the new programs are likely to take the larger cuts required by a tight fiscal stance. An analysis of the distribution of the budget to the different Departments in the country shows that the targeted programs assign a large proportion of their resources to the poorest departments. In per capita terms, transfers to the poorest departments are five times larger than for Lima (the richest departmnent). By contrast, in per capita terms the regional budgets are highly skewed in favor of the richer departments. Aggregating the regional budgets and the targeted programs, MINSA transfers are proportional to the distribution of the population. There is still ample room for improving the geographical targeting of the 20% of the MINSA budget which is assigned to the targeted programs, by improving the technical tools used to target and by strengthening controls to ensure that resources are effectively deployed to the poorer areas. To improve targeting of the overall MINSA budget would require politically more challenging changes, involving a reassignment of the budget directed to the hospitals. New forms of organization, involving community participation, are being developed in an effort to increase effectiveness in primary health care. The reforms may also catalyze the development of organized groups that could help defend the new programs from budgetary cuts. These reforms are necessary, as productivity is low, activities are often implemented without use of a local diagnostic assessment, and they are often bound by unnecessarily bureaucratic rules. The low productivity of PHC is particularly worrisome, with a national average of 1-2 clinical consultations per health worker per day. Existing data underestimates productivity levels by excluding or undercounting preventive and extramural work, but even corrected guesstimates remain in a very low range. The CLAS are committees of community members who administer public facilities to implement population-based local health plans, financed by the Government. Most existing indications suggest that the CLAS, which today operate 10% of MINSA clinics, are successful as a means of increasing the effectiveness of primary care delivery. This success has been achieved despite a growing opposition from the regional government bureaucracies, which resent the loss of direct control over the clinics and are not providing the required technical support to the CLAS. The granting of new permits for the expansion of CLAS has been frozen since 1997 (but may be reopened shortly) REFORMS IN HEALTH FINANCING The targeted programs have succeeded in expanding the geographical coverage of health services by making services available and by improving their quality in remote locations. However, many among the poor remain without access to health care because of its direct and indirect costs. Most MINSA facilities make an attempt to address the cost barriers for the poor by providing partial or total exemptions from payment for services. This system  Peru: Improving Health Care for the Poor 5 of exceptions has its shortcomings. First, there is no fund to subsidize drugs and inputs at the provider level, and these constitute over 80% of the direct cost of health care for the poor. Most drugs and medical inputs, which are financed by the establishment out of revenues from user charges, are charged to the user at full cost plus a mark-up. Second, exceptions for the poor have to be financed by local generosity. As there is no instrument to have subsidies "follow the poor", each establishment finances the lost revenues from its own resources and the decision to assign these resources to provide services for the poor is voluntary. Third, there are no standard criteria or methodology to identify the poor. Each establishment develops its own system and applies it erratically most of the time. The Government is beginning to tackle the cost barrier by introducing schemes designed to provide universal access by selected groups to key services. The Seguro Escolar was created in 1997 to cover health services and drugs for all children aged 3-17 attending public schools (around 6 million). The Government has announced the creation of a Seguro Materno Infantil (SMI), which would cover a package of basic services for mothers and for children under 3 years of age, following in the footsteps of expanding international experience in this area (a successful experience in Bolivia increased institutional coverage of births by a third in only 18 months). Both schemes eliminate co-payments by patients at the point of use of the service and cover prescription drugs. The Seguro Escolar is free of charge to beneficiaries. The Government is considering a small subsidized insurance premium (possibly free of charge in the poorest regions) for the SMI, which it plans to introduce as a separate scheme. Once consolidated, it would establish a single public insurance scheme incorporating the Seguro Escolar and possibly allowing a private provider option. Despite positive results from the Seguro Escolar, which covered 4 million consultations during its first year of implementation and may have significantly improved coverage of school children, and despite promising results from a small pilot for the SMI, the Government is hesitant about expanding the SMI. The Ministry of Finance is concerned about its cost. Some health providers are hesitant because of logistic problems encountered in the implementation of the Seguro Escolar. These problems need to be addressed urgently to avoid creating bottlenecks in service and discrediting a promising scheme. In the Seguro Escolar, reimbursements are slow, and facilities have been forced to cover many expenses from their own revenues as reimbursements cover only the cost of drugs whereas the co-payments by patients covered other costs, including benefits for the staff (such as food baskets). MANAGEMENT OF PROGRAMS DIRECTED TO THE POOR The ability of MINSA to successfully prioritize and deliver health services to the poor is reduced because of inefficiencies in the management of key programs. Many of these problems arise from fragmentation of key programs and from weak information systems. The problem of fragmentation and duplication is especially acute for interventions in programs for maternal and child health (MCH), nutrition, and environmental health. None of these programs has a clear leadership structure. In theory, MCH policy and planning is one of many responsibilities of the Direccion General de Salud de las Personas (DGSP).  Executive Summary and Introduction 6 Within DGSP, this responsibility is split into 8 national programs. Two large and many small externally financed projects also finance MCH activities. Each of these projects and programs does its own planning, develops its own protocols, and plans and finances its own training programs. In environmental health, there are overlaps and often duplications in food quality control and vector control. In nutrition, overlaps occur in functions performed by DGSP, INS and many other public and private sector institutions outside MINSA. This duplication is underpinned by a fragmented budget system. The weakness of the information on health service production in Peru is striking and past attempts to correct this have lacked continuity and consistency. Few developing countries have good quality data for the health sector, but most have information for the public sector. In Peru there is no attempt to consolidate information about financial flows or production services for the public sector. No agency collects or monitors information about expenditures incurred by all MINSA providers, or by all MINSA programs, or even by all externally funded programs.5 Production statistics are no longer collected and published regularly. Estimates for MINSA ambulatory consultations for 1995 (the most recent available) run from 15 million (official statistics) to 27 million (household survey estimates). There is no official source for inpatient consultations. Each program produces its own data and there are few serious attempts to consolidate it in a way that would allow monitoring of activities at an aggregate level (e.g. vector control activities, nutrition activities, even laboratory production are measured by different programs in ways that cannot be aggregated or compared). Similar problems exists with the measurement of inputs, even for high-cost items such as staff, training, or the provision of equipment, since each program or funding source maintains its own records and there is neither a human resources nor an infrastructure office in MINSA to effectively consolidate such information. HUMAN RESOURCES FOR HEALTH CARE Human resource issues are at the root of many inefficiencies and inequities in the health system. Especially important for the provision of services to the poor are issues of geographical distribution, skills mix, and human resource quality. Geographical distribution has improved noticeably during the last two decades as physicians have been attracted to small cities as their population and income grew. More recently, the targeted programs have placed over 10,000 health workers in less-favored locations. These programs provide large financial incentives but workers have no benefits and no security in the job. Despite the financial incentives and the existence of a temporary oversupply of health workers, turnover in remote locations is exceedingly high as the higher incomes do not overcome the combined effect of the pull of professional careers, tied to city-based medical specialization and private practice, and the burden of life in a radically different cultural environment. This is exacerbated by the lack of expectations of a long-term career. This conundrum is leading specialists to look for alternative solutions to the problem of serving rural communities. One option under discussion would involve a change in the skills-mix for local health workers complemented with stronger links to the rest of the ' There exists an ongoing effort to produce national health accounts attempting to overcome the technical weaknesses of previous efforts.  Peru: Improving Health Care for the Poor 7 health network. Local workers would be more in charge of the public health aspect of the job with less emphasis on the clinical aspects. Management structures for the local workers, communications and reference systems would need to be strengthened, and in some areas the use of mobile physicians and "telehealth technology" would be introduced. The skill-mix problem is that the content of medical training has not kept up with the shift in national priorities toward primary and preventive health, or with the corresponding shift to community and rural health delivery models. The educational gap is in part a matter of the relative attention given to different medical pathologies, and partly of non-medical knowledge and skills related to public health and to community work. At the same time the training of medical professionals has not kept up with the rapidly increasing importance of systems, teamwork and information in health delivery, and therefore, with the need for administrative and managerial skills. The educational system is beginning to respond to these gaps, but additional effort will be needed. The consensus among M4NSA officials and medical professional leaders is that standards in medical education have been falling. This is mostly associated with the recent large-scale and little regulated creation of new university faculties in medicine and the increase in the number of teaching institutes for non-physician medical workers. Universities must obtain licenses, but permissions to operate have been granted liberally in recent years and, once in operation, universities need no further governmental or professional approval to grant professional titles to physicians. The licensing of non-university teaching centers is not subject to professional approval or certification. PROPOSED REFORMS The main conclusions of this report are that the reforms introduced by MINSA in the provision of primary care services have been successful and should be sustained, expanded and deepened. To accomplish that, the main report concludes with a chapter that makes a number of detailed recommendations, which may be grouped under five distinct reform initiative headings: i) establishing a sharper focus on the health needs of the poor in the allocation of public funds; ii) reinforcing that focus on the poor among MINSA health providers (improved targeting, increased access by the poor to hospital services, enhanced community participation); iii) creating new insurance mechanisms to finance health care for the poor; iv) improving the information and management systems needed to run MINSA's programs for the poor efficiently and without duplication; and v) refining human resource skill requirements and incentives to better serve poverty-oriented programs. OUTLINE OF THE MAIN REPORT Focusing on the challenges outlined above, this report is structured as follows. Chapter 1 is a brief overview of the health sector designed to put MINSA and the recent reforms in a more general perspective (financial flows, the growth in health expenditures and in access to services, and an analysis of which subsectors benefit different income groups). Chapter 2 describes the progress in health outcomes, analyzing differences in health status and needs between the poor and the rest of the population. Chapter 3 describes MINSA in  Executive Summary and Introduction 8 greater detail and analyzes the key reforms implemented in health provision including: the increased emphasis on primary care, the use of geographical targeting, and new initiatives to improve management and community participation. It also discusses key issues in public hospitals. Chapter 4 analyzes the cost barrier and discusses two schemes currently being developed to overcome this barrier: the Seguro Escolar and the Seguro Materno Infantil. Chapter 5 examines ongoing shortcomings in the management of the programs directed to the poor, emphasizing fragmentation and lack of information. Chapter 6 examines issues related to health workers. It first analyzes the main trends in the labor markets for these workers and then discusses issues of geographical inequality, quality of training and the mismatch between skills-mix and needs. Chapter 7 brings together the main conclusions and recommendations.  Peru: Improving Health Care for the Poor 9 2. OVERVIEW OF HEALTH SECTOR FINANCING AND DELIVERY SYSTEMS Peru's health sector is a complex system is a complex amalgam of the following sectors: * the Ministry of Health (MINSA), which finances and delivers care, in theory directed to the poor; * the Peruvian Institute of Social Security (previously known as IPSS, recently renamed as ESSALUD), which finances and delivers care to formal sector workers and their dependants; * several smaller public programs, including one for each branch of the military and another for the police, each of which finance and deliver care; * a large private sector in terms of both delivery and financing of care; and * several NGOs which mostly support the delivery of services by MINSA. What follows is, first, a description of health expenditures and inputs in the sector as a whole. A second section describes the health infrastructure and staff available, their growth rates in the 1992-1996 period, and how they are distributed among the different subsectors. The third section documents the weight of the different subsectors in the provision of services. Fourth, a benefit incidence analysis is presented, describing how public and private expenditures are divided among the different income groups. OVERVIEW OF HEALTH EXPENDITURES AND SYSTEM RESOURCES During 1988-1993, financing for health fell drastically as a result of the drop in personal income and the collapse of all public expenditures brought about by the hyperinflation of the late 1980s and the stabilization policies used to control it in the early 1990s. Total spending in health has rapidly recuperated, increasing from US$1.6 billion in 1994 to over US$2.7 billion in 1997. Adjusting for inflation, this represented an increase of over 50% in real soles in only three years. Despite very rapid growth of GDP during these years, health expenditures grew faster than GDP, indicating a relatively high income elasticity (estimated at 1.29). Table 2-1 Despite this rapid recuperation, Peru: Main Sources of Finance for the Health Sector, 1997 health expenditures in Peru continue to be low by any standard. USSmillion % of total % ofGDP As a fraction of GDP, health Household spending 1,316 49 2.0 expenditures were only 4.1% in Govemment (taxes and loans) 690 26 1.1 Employers' contributions to health 1997 (Table 2-1). Estimates of the insurance 682 25 1.0 average for Latin America range Total 2.688 100 4.1 from 5.5%-7.3%.6 Per capita Source: Budget data and ENNIV 1997 6 The low estimate is by the World Bank for 1994, and the high estimate is by PAHO for 1995. Only 6 countries in LAC spend as little on health as Peru.  Overview of Health Sector Financing and Deliverv Systems 10 expenditures in health, at US$90 per capita in 1997, are also about half of the Latin American average. Table 2-2 compares health expenditures for several countries in 1994. Table 2-2 International Comparison of Health Expenditures (circa 1994) Peru Bolivia Brazil Colombia Ecuador Mexico Nicaragua Panama (1994) (1994) (1994) (1994) (1994) (1995) (1994) (1994) Health to GDP ratio (%) 3.7 7.1 4.6 7.4 5.3 4.2 8.6 67 Per capita expenditures on health (US$) 80 54 163 143 78 182 37 199 Per capita expenditures on health $PPP) 156 194 264 477 259 311 174 457 Public share (%) 60 58 40 40 39 56 61 70 Source: Health and Population Indicators. World Bank, 1999. Notes: Conversions to USS based on offlcial exchange rates. Purchasing Power Parities (PPPs) are exchange rates used to convert local currency into US dollars taking into account price differences across countries. MINSA is financed mainly by tax revenues, but it also obtains funds from external loans and from user fees. IPSS is financed by a 9% payroll tax paid by employers of formal sector workers. The smaller public programs are entirely funded by tax revenues. As in other countries of similar income level, in Peru about half of the health financing is provided by households, mainly as out-of-pocket payments. Private insurance covers less than 1% of the population and is often used as a complement to IPSS. Table 2-3 MINSA, IPSS and households have all Health Expenditures by Sub-Sector, 1997 increased their spending substantially after 1994 (Table 2-3). The largest Real Increase 1994 Per Capita Spending increase was in households' out-of- 1997 (USS) pocket payments for health care which MINSA 35 28 doubled in real terms during 1994-97.7 IPSS 19 105 MINSA increased its spending by a Private 10 54 third in real terms over this brief Total 55 90 period. IPSS had the smallest increase, but even that meant a substantial Notes: The real increase was estimated in 1994 soles increase in real terms in only 3 years. Table 2-4 provides information on the physical configuration of the service delivery system in terms of personnel and facilities. MINSA has an important share of the assets and the staff in the sector. It runs the majority of the Primary Health Clinics (PHC), and while it accounts for less than a third of hospitals, these include most of the large hospitals, giving MINSA over half of the beds. It also employs over half of the health care professionals in the country, including 39% of physicians, most nurses, midwifes and auxiliary nurses. IPSS, with a stronger emphasis on secondary and tertiary service, has a small proportion of primary clinics, but a significant share of hospital beds and 7There is agreement that growth in private consumption was very rapid in this period, there is some disagreement about the magnitude of this growth, some of the apparent growth may be due to improved measurement.  Peru. Improving Health Care for the Poor 11 physicians. The private sector is very large. Over 200 small formal private clinics employ a third of the physicians. Many of the physicians in public service also provide private services in single practitioner offices and other inforrnal arrangements (these are not included in Table 2-4). HEALTH CAR1E ACCESS AND USE The years since 1992 have seen a sharp increase in the supply of health services, especially in primary health clinics (Table 2-4). The number of clinics increased by almost two thirds between 1992-96. FONCODES (Peru's social investment fund) alone rehabilitated or built 1, 213 primary care establishments for use by MINSA between 1992 and 1997 at a cost of about $30 million. Other agencies also contributed to this expansion of infrastructure. Growth has been rapid both in MINSA (where the number of establishments grew by 51%), in IPSS, and in the military. Additionally, MINSA and IPSS have established policies that have lead many clinics to significantly increase their hours of operation. Table 2-4 The availability of health Infrastructure and Staff of Health Services by Subsector professionals has also in 1996 increased rapidly. Recruitment Total % Increase MINSA IPSS Mllitaryt Private' Total in professions such as 1996 1992-1996 Police 1)midwifes and auxiliary nurses, Infrastructure 1% distribution 1996) PHC Facilifies 6,717 61 86 3 2 9 100 who are required by primary Hospitals 472 4 30 15 4 51 100 clinics, has been particularly Hospital Beds 42,979 n.a. 67 14 6 13 100 rapid. Despite this expansion, Staff the low level of financing Doctors 24,708 50 41 18 7 34 100 assigned to health care Nurses 16,139 45 58 23 10 9 100 Midwifes 5,105 120 77 12 3 8 100 continues to be reflected in a Dentists 2,622 89 53 13 13 21 100 low availability of physicians Technical Staff 44,742 n.a. 66 13 11 10 100 (only 70% of the Latin 11 'Prnvate" refers exclusively to the corporate Aninics. Soumces: CISRESA 1992 and CISRESA 1996. Amencan average) and of other health workers such as nurses Table 2-5 and dentists (Table 2-5). Availability of Health Inputs, 1997 (per 10,000 population) While Peru also has a lower Peru LAC Andes number of beds per inhabitant Physicians 10.3 14.9 13 than the Latin American Nurses 6.7 7.4 14 average, it has more beds than Dentists 1.1 5 3.8 neighboring countries, and Hospital Beds (per 1,000) 1.8 2.6 1.5 there is significant excess Source: PAHO, Basc Indicators, 1998. capacity (only 52% occupancy). Many Latin American countries invested heavily in health infrastructure during the 1990s, attempting to regain ground lost during the crisis of the 1980s. Some countries, such as Bolivia and Honduras have witnessed large growth of infrastructure accompanied by a proportionately much smaller expansion in effective coverage (in Honduras  Overview of Health Sector Financing and Delivery Systems 12 infrastructure grew by 40% and production increased by only 24%). Peru has been much more successful in translating physical expansion into improved coverage and effective use. Coverage for immunization and other preventive services has improved markedly (see Chapter 3). The use of ambulatory services -number of consultations-increased by 59% during 1994-97 (Table 2-6).8 Inpatient services grew by 10% during that period. Table 2-6 MINSA, IPSS AND THE PRIVATE Peru: Use of Health Care in 1997 SECTOR AS PROVIDERS Ambulatory Consultations in 1997 MINSA has a very large role in the Total Increase Per capita (million) 94-97 provision of most types of health Total 63.5 59% 2.6 services. It is in charge of providing Lima 22.2 99% 3.2 Other Urban 23.8 34% 2.7 most public health interventions Rural 17.5 58% 2.0 including organizing the whole Source: Cuanto SA, based on ENNIV 1994 and 1997 immunization program and all vector control activities and sharing with municipalities responsibility for controlling the quality of water and food. MINSA also has a large role in the provision of clinical services, providing two-thirds of inpatient services and 44% of outpatient services. IPSS is the second most important provider of inpatient services, but has a much smaller role for the provision of outpatient consultations. By contrast, the private sector provides over a third of outpatient consultations, but only a small fraction of inpatient services. Table 2-7 Importance of Providers by Region and Income Level, 1997 (%) Ambulatory Hostitalizations Military/ Military/ MINSA IPSS Police Private Total MINSA IPSS Police Private Total Lima 36 20 4 40 100 53 26 8 13 100 Other Urban 40 25 2 33 100 67 26 0 7 100 Rural 60 5 1 34 100 81 12 0 7 100 Quintile I (poorest) 68 4 1 28 100 85 0 0 16 100 Quintile 2 52 12 1 34 100 71 27 0 3 100 Quintile 3 53 15 1 32 100 65 26 0 9 100 Quintile 4 40 23 4 33 100 63 29 0 8 100 Quintile 5 (richest) 26 25 3 47 100 58 18 8 17 100 TOTAL 44 18 2 36 100 65 23 2 9 100 Source: ENNIV, 1997 The best definition of access would relate use to need. Unfortunately, LSMS surveys use only self-perceived illness as an approximation to "need." Self-perceived illness is influenced by expectations that tend to grow with income and education and has been shown to be poorly associated with burden of disease. Using self-perceived illness as an indicator would show that the largest increase in illness occurred in Lima and that Lima has the worst health status, followed by the Sierra (highlands) and, lastly, the Selva or Amazon jungle region (the opposite of what most analysts believe to be the case). By income level, the equally absurd pattern emerges that the richest quintile has the worst health status and the poorest quintile the best.  Peru: Improving Health Care for the Poor 13 There exist some regional differences. MINSA is most important in rural areas of the Sierra (highlands) and the Amazon (Table 2-7). While it is less important in urban areas, even there it is the single most important provider of health care. IPSS provides very few services to rural populations. The private sector is slightly more important in Lima than in the rest of the country, but it provides about a third of outpatient services in all regions (in remote rural areas this includes traditional practitioners). There also exist differences by income group. MINSA is most important as a provider of services for the poor (Table 2-7). However, it is also significant as a provider for the non-poor. Lower income groups get most of their consultations from PHC centers, while users in the higher income groups use hospital outpatient facilities. MINSA's market share grew between 1994-97 for all kinds of services. This increase was especially large for the top 40% of households (Diagram 2-1). IPSS beneficiaries are concentrated among the richer 40% of the population. Private providers are most important for the rich. However, they provide almost a third of outpatient consultations for the population as a whole. Diagram 2-1 BENEFIT INCIDENCE OF HEALTH MINSA's Market Share for Outpatient EXPENDITURES Services In 1994 and 1997 (% of total) Per capita consumption of health UrbanCoast goods and services is about 4.5 Rwal Coast UrbaniSier,s times higher in the top quintile than Rural Slera - in the bottom quintile (Diagram Urban Amazon 2-2). This difference, while very Rural Anazo . large, is less than the difference in total consumption (a ratio of 1:7.5). Qudinle_I_(bottom)_ The consumption of health care Quintile2 combines a private element with Quintile 3 1994 benefits channeled by MINSA and Quirtile4 * 1997 Quntileb I97 IPSS. Most of the inequality in 0 20 40 60 SO health care consumption results S.urae ENNI 1994and 1997 from the assignment of greater amounts of private expenditure to health care by the rich than by the poor (top to bottom quintile ratio of health expenditures of 1:10). This is a common pattern as health care is a good whose consumption normally rises with income or faster. 9-Benefit incidence analysis combines the cost of providing public services with information on their use in order to generate distributions of the benefit of government spending. This has become an established approach since the path-breaking work by Meerman (1979) on Malasia and Selowsky (1979) on Colombia. The benefit incidence analysis presented in this study involved a three step methodology. First, estimates were obtained of the unit subsidy of: (i) consultations in PHC, (ii) ambulatory consultations in hospitals, and (iii) inpatient consultations in hospitals. These estimates were obtained based on 1997 budgets and 1997 ENNIV utilization data. Second, the unit subsidy was distributed to individuals who were identified from ENNIV as users of the service. Third, individuals were aggregated into five quintiles (each including a fifth of the population) based on the total per capita expenditures of the household.  Overview of Health Sector Financing and Deliverv Systems 14 Diagram 2-2 Per Capita Health Expenditures by Quintiles in 1 997 7 400 - 20,% - 7% 350 0 3 50 -M IN SA tra nsfe rs as _ 3 0 0 % of health, onsum ption _ % 300 - > 5% 2 5 0 -M IN rans ers as 3 _ 200 -(right hand scale) | 3% Uz1 5 0 1 : l> _ 100 - L70% 1 0 _ - - 2% 50 _ 0 0 % Poorest Q2 Q3 Q4 Richest I 13 M IN SA MIPSS I P rivate I The inequality in the consumption of health benefits is reinforced by IPSS, which serves mostly formal sector employees and their families and these are concentrated in the higher echelons of the income distribution.10 MINSA expenditures reduce the overall inequality, but are not large enough or sufficiently well targeted among the poor to significantly equalize expenditures. The distribution of MINSA expenditures is almost proportional to the population, in the sense of similar nominal amounts being directed to the poor and to the non-poor. The poorest 20% of the population receive a slightly lower per capita subsidy than the rest of the population. However, in proportion to their income this benefit is significantly higher for this group than it is for the rest of the population: for the poorest it adds about 5% to their total consumption; for the richest quintile, it adds only 1% to their total consumption. The importance of MINSA is largest for the bottom 20% who obtain 70% of their health benefits from this source compared to 20% for the richest quintile. In this chapter the focus has been on financing and provision of health services. We now turn to a discussion of the progress made and ongoing shortcomings in health outcomes. '