IMPLEMENTATION COMPLETION AND RESULTS REPORT (IBRD-45270) ON A LOAN IN THE AMOUNT OF US$ MILLION THE REPUBLIC OF PERU FOR A

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Document of The World Bank Report No: ICR IMPLEMENTATION COMPLETION AND RESULTS REPORT (IBRD-45270) ON A LOAN IN THE AMOUNT OF US$ MILLION TO THE REPUBLIC OF PERU FOR A HEALTH REFORM PROGRAM (FIRST PHASE: MOTHER AND CHILD INSURANCE AND DECENTRALIZATION OF HEALTH SERVICES) March 6, 2007 Human Development Sector Management Unit Bolivia, Ecuador, Peru and Venezuela Country Management Unit Latin America and the Caribbean Regional Office

2 CURRENCY EQUIVALENTS (Exchange Rate Effective March 6, 2007) Currency Unit = Peruvian Nuevo Sol (PEN) PEN 1.00 = US$ US$ 1.00 = PEN 3.19 FISCAL YEAR [January 1 December 31] ABBREVIATIONS AND ACRONYMS AAA APL BONF CLAS DFID DISA DIRESA DPL EDA ENAHO EONF ESSALUD GP ICR IDB IMR INEI IRA MEF MINSA OPEC PAD PAHO PARSALUD PCU PSRL SEG Analytical and Advisory Activities Adaptable Program Loan Basic Obstetric and Neonatal Functions Local Health Administration Comitees (Comités Locales de Administración de Salud) Great Britain's Department for International Development Health Directorate (Dirección de Salud) Regional Health Directorate (Dirección Regional Salud) Development Policy Loan Acute Diarrhea Diseases (Enfermedades Diarréicas Agudas) National Home Survey (Encuesta Nacional de Hogares) Essential Obstetric and Neonatal Functions Peruvian Social Security Institute (Seguro Social de Salud) Government of Peru Implementation Completion Report Inter-American Development Bank Infant Mortality Rate National Statistics Institute (Instituto Nacional de Estadística e Informática) Acute Respiratory Infections (Infecciones Respiratorias Agudas) Ministry of Economy and Finance (Ministerio de Economía y Finanzas) Ministry of Health (Ministerio de Salud) Organization of Petroleum Exporting Countries Project Appraisal Document Pan American Health Office Health Reform Program (Programa de Apoyo a la Reforma del Sector Salud) Project Coordination Unit Programmatic Social Reform Loans School Health Insurance (Seguro Escolar Gratuito)

3 SIS SMI SNIP WHO Integrated Health Insurance(Seguro Integral de Salud) Maternal and Child Insurance (Seguro Materno Infantil) National Public Investment System (Sistema Nacional de Inversión Pública) World Health Organization Vice President: Country Director: Sector Manager: Task Team Leader: Pamela Cox Marcelo M. Giugale Keith E. Hansen Livia M. Benavides

4 PERU HEALTH REFORM PROGRAM (First Phase: Mother and Child Insurance and Decentralization of Health Service CONTENTS Data Sheet A. Basic Information B. Key Dates C. Ratings Summary D. Sector and Theme Codes E. Bank Staff F. Results Framework Analysis G. Ratings of Project Performance in ISRs H. Restructuring I. Disbursement Graph 1. Project Context, Development Objectives and Design Key Factors Affecting Implementation and Outcomes Assessment of Outcomes Assessment of Risk to Development Outcome Assessment of Bank and Borrower Performance Lessons Learned Comments on Issues Raised by Borrower/Implementing Agencies/Partners Annex 1. Project Costs and Financing Annex 3. Economic and Financial Analysis (including assumptions in the analysis) 37 Annex 4. Bank Lending and Implementation Support/Supervision Processes Annex 5. Beneficiary Survey Results Annex 6. Stakeholder Workshop Report and Results Annex 7. Summary of Borrower's ICR and/or Comments on Draft ICR Annex 8. Comments of Cofinanciers and Other Partners/Stakeholders Annex 9. List of Supporting Documents MAP

5 A. Basic Information Country: Peru Project Name: HEALTH REFORM PROGRAM (First Phase: Mother and Child Insuranceand Decentralization of Health Servic Project ID: P L/C/TF Number(s): IBRD ICR Date: 03/29/2007 ICR Type: Core ICR Lending Instrument: APL Borrower: Original Total Commitment: Environmental Category: B Implementing Agencies: Ministerio de Salud Cofinanciers and Other External Partners: Inter-American Development Bank (IADB) B. Key Dates GOVERNMENT OF PERU USD 27.0M Disbursed Amount: USD 27.0M Process Date Process Original Date Revised / Actual Date(s) Concept Review: 11/30/1998 Effectiveness: 07/02/ /02/2001 Appraisal: 09/19/1999 Restructuring(s): Approval: 12/16/1999 Mid-term Review: 09/25/2003 Closing: 12/31/ /30/2006 C. Ratings Summary C.1 Performance Rating by ICR Outcomes: Risk to Development Outcome: Bank Performance: Borrower Performance: Satisfactory Low or Negligible Satisfactory Satisfactory C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Bank Ratings Borrower Ratings Quality at Entry: Satisfactory Government: Not Applicable Quality of Supervision: Satisfactory Implementing Agency/Agencies: Not Applicable Overall Bank Performance: Satisfactory Overall Borrower Performance: Satisfactory i

6 C.3 Quality at Entry and Implementation Performance Indicators Implementation QAG Assessments Indicators Performance (if any) Potential Problem Project Yes at any time (Yes/No): Problem Project at any time (Yes/No): DO rating before Closing/Inactive status: Yes Satisfactory Quality at Entry (QEA): Quality of Supervision (QSA): None None Rating D. Sector and Theme Codes Original Actual Sector Code (as % of total Bank financing) Central government administration 3 3 Health Theme Code (Primary/Secondary) Child health Primary Primary Decentralization Secondary Secondary Other communicable diseases Primary Secondary Participation and civic engagement Primary Primary Population and reproductive health Secondary Primary E. Bank Staff Positions At ICR At Approval Vice President: Pamela Cox David de Ferranti Country Director: Marcelo Giugale Isabel M. Guerrero Sector Manager: Keith E. Hansen Xavier E. Coll Project Team Leader: Livia M. Benavides Evangeline Javier ICR Team Leader: ICR Primary Author: Livia M. Benavides Juan Prawda F. Results Framework Analysis Project Development Objectives (from Project Appraisal Document) The project is the first phase of a three-phase adjustable program loan. The goals of the overall program are to: (a) improve maternal and child health; and (b) help reduce morbidity and deaths of the poor from communicable diseases and inadequate environmental conditions. ii

7 The project development objective for phase 1 of the APL was to increase access of the poor to better quality health programs and services. This access would be improved by strengthening the demand for services and the quality of the supply. The demand would be strengthened through the implementation of a Seguro Materno Infantil (SMI, this was later renamed as Seguro Integral de Salud or SIS). The quality of supply would be enhanced by increased community participation (mainly through the CLAS system), by reducing fragmentation in the delivery of maternal and child and environmental health services and by adapting investments to local needs. Revised Project Development Objectives (as approved by original approving authority) The original PDOs and Key Indicators remained unchanged. (a) PDO Indicator(s) Indicator Baseline Value Original Target Values (from approval documents) Formally Revised Target Values Actual Value Achieved at Completion or Target Years Indicator 1 : Infant mortality rate Value 43 (nation wide) 29.2 (nation wide) 22.3 (nation wide) quantitative or 48 (project area) 34.4 (project area) 28.3 (project area) Qualitative) Date achieved 12/31/ /30/ /30/2006 Comments (incl. % Both nationwide and project area targets were surpassed. achievement) Indicator 2 : Number of pregnant women with 4 or more prenatal care visits. Value 59.9 (nation wide) 95.8 (nation wide) 75.6 (nationwide) quantitative or 32.1 (project area) 51.4 (project area) 57.2 (project area) Qualitative) Date achieved 12/31/ /30/ /30/2006 Comments (incl. % achievement) Indicator 3 : Value quantitative or Qualitative) The nationwide indicator increased 26.2 percent while the one in the project area increased by 78.2 percent. Percentage of deliveries attended by skilled health personnel 54.8 (nation wide) 27.6 (project area) 72.8 (nation wide) 36.7 (project area) 71.1 (nationwide) 50.9 (project area) Date achieved 12/31/ /30/ /31/2005 Comments The nationwide indicator increased by 29.7 percent while the one in the project (incl. % areas increased by 84.4 percent. achievement) Indicator 4 : Number of children under age 5 with ARI attended in a health facility Value quantitative or Qualitative) 8,747,589 3,700,000 7,742,457 Date achieved 12/31/ /30/ /30/2006 iii

8 Comments (incl. % achievement) Indicator 5 : Value quantitative or Qualitative) ARI targets were established originally as absolute numbers due to lack of baseline numbers during Project preparation. Absolute numbers are difficult to interpret. Number of children under age 5 with accute diarrhea treated in a health facility 4,535, ,000 4,531,800 Date achieved 12/31/ /30/ /30/2006 Comments (incl. % achievement) Acute diarrhea targets were established originally as absolute numbers due to lack of baseline numbers during Project preparation. Absolute numbers are difficult to interpret. Indicator 6 : Percentage of children between 18 and 29 months vaccinated with DPT3. Value quantitative or Qualitative) 78.30% 95% 87.41% Date achieved 12/31/ /31/ /31/2005 Comments (incl. % achievement) Indicator 7 : Value quantitative or Qualitative) DPT3 was replaced by a more complex vaccine, so that its coverage at the regional level has been reduced during the last few years. Percentage of yellow fever cases attended 80% 100% Date achieved 06/30/ /30/2006 Comments Yellow fever is a disease under weekly national surveillance. By law, the (incl. % reported cases have to be attended. achievement) Indicator 8 : At least 17 million beneficiaries receive care in the primary facilities of MINSA annually Value quantitative or 17,000,000 28,130,000 Qualitative) Date achieved 06/30/ /30/2006 Comments (incl. % achievement) The target was significantly surpassed. Number of health departments (DISAs) implementing regional and local health Indicator 9 : plans that respond to communicable diseases and environmental health problems prevalent in their respective localities Value quantitative or 80% 100% Qualitative) Date achieved 06/30/ /30/2006 Comments (incl. % achievement) All DISAs, regularly sign annual Management Agreements with the Ministry of Health, reflecting commitments with respect to regional health priorities and indicators with quantitative goals for the period iv

9 (b) Intermediate Outcome Indicator(s) Indicator Baseline Value Original Target Values (from approval documents) Formally Revised Target Values Actual Value Achieved at Completion or Target Years Indicator 1 : Percentage of attended deliveries which received oxitocin in the Project area. Value (quantitative or Qualitative) Date achieved 12/31/ /30/ /31/2006 Comments (incl. % Indicator increased by 90 percent. achievement) Indicator 2 : Percentage of deliveries attended institutionally by SIS. Value 37.6 (nation wide) 66.3 (nation wide) 65.1 (nationwide) (quantitative 18.5 (project area) 42.4 (project area) 59.9 (project area) or Qualitative) Date achieved 12/31/ /30/ /30/2006 Comments (incl. % achievement) The nationwide indicator increased 73.1 percent while the one in the project areas increased by percent (3.1 time more) Indicator 3 : Percentage of deliveries financed by SIS in the first two quintiles of population in the Project area. Value (quantitative or Qualitative) Date achieved 12/31/ /30/ /30/2006 Comments (incl. % The indicator increased by 27% achievement) Number of health facilities refurbished to comply with Essential Obstetric Indicator 4 : Function (EOF), Basic Obstetric Function (BOF) and Basic Maternal and Child services capacity Value (quantitative or Qualitative) 0 20 (EOF) 59 (B0F) 20(EOF) 54(BOF) Date achieved 12/31/ /30/ /30/2006 Comments (incl. % achievement) Indicator 5 : Value (quantitative or Qualitative) Target was met for EOF. 91 percent of BOF were refurbished. The lower target was due to additional cost from basic utilities needed for health facilities. Percentage of references with respect to the number of attentions Date achieved 12/31/ /30/ /30/2006 Comments (incl. % The indicator increased in 2004 (6%) and 2005 (7%) but fell again to 5% in The main obstacle has been the allocation of budget for fuel for v

10 achievement) Indicator 6 : Value (quantitative or Qualitative) ambulances, which is not financed by SIS. Percentage of newborn children covered by the SIS weighted within the first 24 hours of birth 73 (nationwide) 64 (project area) 95 (nationwide and project area) 83 (nationwide) 74 (project area) Date achieved 12/31/ /31/ /30/2006 Comments The nationwide indicator increased 13.7 percent while the one in the project (incl. % areas increased by 15.6 percent (1.1 times more). achievement) G. Ratings of Project Performance in ISRs No. Date ISR Archived DO IP Actual Disbursements (USD millions) 1 06/27/2000 Satisfactory Satisfactory /30/2000 Satisfactory Satisfactory /19/2001 Satisfactory Satisfactory /28/2001 Satisfactory Unsatisfactory /19/2001 Satisfactory Unsatisfactory /03/2002 Satisfactory Satisfactory /10/2002 Satisfactory Satisfactory /11/2002 Satisfactory Satisfactory /10/2003 Satisfactory Satisfactory /11/2003 Satisfactory Satisfactory /05/2003 Satisfactory Satisfactory /14/2004 Satisfactory Satisfactory /29/2004 Satisfactory Satisfactory /26/2005 Satisfactory Satisfactory /01/2005 Satisfactory Satisfactory /28/2006 Satisfactory Satisfactory H. Restructuring (if any) Not Applicable vi

11 I. Disbursement Profile vii

12 1. Project Context, Development Objectives and Design (this section is descriptive, taken from other documents, e.g., PAD/ISR, not evaluative) 1.1 Context at Appraisal (brief summary of country and sector background, rationale for Bank assistance) The Peru Health Reform Project (First Phase: Mother and Child Insurance and Decentralization of Health Services (labeled as PARSALUD I), was one of a cluster of Bank-supported activities designed to improve basic health indicators in Peru. This cluster included an influential AAA which made recommendations for health policy reforms needed to obtain better results in health. These recommendations were implemented through a series of four policy based loans (Programmatic Social Reform Loans or PSRLs) and an investment loan (PARSALUD I). While this ICR will discuss exclusively the implementation of PARSALUD I, the implementation of this loan needs to be understood in the broader context of the implementation of the cluster. At the time of appraisal, the following major sector issues were facing the Peruvian health sector. (a) Low health status of the poor population. Peru's very high infant mortality of 43 per 1,000 life births in 1996 was among the worst in Latin America and the Caribbean region (LAC). The maternal mortality rate of 265 for 100,000 births observed in the period was almost one and half times higher than the LAC average and 15 times the average for developed countries. (b) Inequitable access to health. A large gap between the health status of the poor and that of the non-poor continued to exist in Peru at the end of the nineties. At the time of project appraisal, the highest mortality rates were concentrated among the poorer 40 percent of the population, where almost 60 percent of child deaths occurred. Poor households were handicapped by a combination of low education and worse environmental conditions. They also had less access to some key and basic health services. (c) Limited financing and weak targeting of health expenditures. At the end of the last decade, Peru allocated fewer resources to health than most of its neighbors. The proportion of Gross Domestic Product (GDP) assigned to health in 1997 (4.1 percent) was about two thirds of the LAC average. Per capita expenditure in health at US$90 in 1997, was about half of the LAC average. The ratio of physicians to the population of about 10 per 10,000 in 1997 was 70 percent of the LAC average. There was also great inequality in the consumption of health goods and services resulting in approximately 4.5 times higher among the richest 20 percent of the population than among the poorest 20 percent. (d) Sector fragmentation. Peru's health system was, at the time of project appraisal, a complex amalgam of disjointed public programs and a private sector, each of which went 2

13 its own way, with little coordination, integration or competition between program providers in either financing or delivery of services. In addition, there was a striking weakness of the information of health service production in Peru. At the time of project preparation MINSA had already launched a number of initiatives to increase access of the poor to health services. These initiatives provided a platform for the activities that were to receive support by the Project: Enactment in 1997 of a law to launch formally some of the key reforms the General Health Law (Ley General de Salud) which provided the normative scope for the health sector, and included the establishment of user's rights, so they would be recognized by the public and respected by service providers. Implementation of the School Insurance (Seguro Escolar) in 1997 to provide free health services and medicines for about 6 million children in the 4 to17 age cohort attending public schools. Piloting in 1999 in the regions of Tacna and Moyobamba of the Maternal and Child Health Insurance scheme (Seguro Materno Infantil SMI), a reimbursement mechanism that paid eligible providers for a package of essential health services for mothers and children under 3 years of age. These two insurance mechanisms were seen as the seed for a more comprehensive basic health package for the poor envisioned to be achieved over a ten-year horizon. Streamlining of a community participation reform initiated in The Community-Managed Health Facilities (Comités Locales de Administración de Salud- CLAS) were conceived as private-public, non-profit entities at the local level working under agreement with the Regional Health Authorities (Direcciones de Salud Regional DISAs), deconcentrated units belonging to MINSA located in each Department of the country). These agreements allowed the CLAS to receive and administer public treasury funds to provide health services through a health center or post under the framework of a locallydeveloped health plan. The CLAS aimed at improving the quality and increasing the coverage of ambulatory services at the primary health level through greater participation by the community in the planning, management and supervision of public resources destined for health provision to the poor. Rationale for Bank assistance. The 1997 Country Assistance Strategy (CAS) for Peru had as its main objective the reduction of poverty and improvement of Peru's human capital base. As part of the AAA package, a health sector study was carried out (Peru, Improving Health Care for the Poor, initiated in 1997 and published in 1999). This study made five critical recommendations: 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. These recommendations were addressed 3

14 through a complex program that included the Programmatic Social Reform (comprised of four policy-based loans between 2001 and 2004), the Health Reform Program (PARSALUD) and a Technical Assistance Loan on accountability systems for the social sectors. The World Bank (Bank) was prepared to support the medium-term financial and technical assistance requirements for the Government s health reform through a threephase Adaptable Program Lending (APL), where each phase was expected to last an average of about three years. An APL was chosen as the financial instrument instead of a single sector investment loan because it allowed for the timing and flexibility required to implement and sustain the several incremental changes needed in a long-term and complex health reform like the one the Government of Peru envisaged. The first phase (corresponding to this Project) was expected to cover from FY00 to FY04, with an expected loan amount of US$80 million equivalent, aimed at financing reimbursements to SMI, supply side investments in 10 Regions (17 DISAs) and technical assistance for the preparatory analytical work, sensitization activities and piloting required to ensure the enabling conditions to undertake more complex tasks during the second and third phases. The second phase, running initially from FY04 to FY07, with an expected loan amount of US$50 million equivalent, aimed at: (i) tackling the full implementation of the maternalchild health insurance (private and public); (ii) improving the health manpower; and (iii) enhancing the autonomy of hospitals. The consolidation of the health reform was to be included in phase III, planned for FY07-FY10, with an estimated loan of about US$50 million equivalent. Cofinanciers. The Health Reform Program, as originally designed, included independent parallel co-financing from the IDB through a separate loan of US$87 million also lasting about three-years. The IDB financing was to benefit a similar number of poor mothers and children as the Bank-financed investment in 16 other Regions (17 DISAs) of the country. In addition, the Organization of Petroleum Exporting Countries (OPEC) had committed US$8 million and the Government of Great Britain's Department for International Development (DFID) was to provide financial support and technical assistance to carry out a community health survey (under Component B of the proposed Project). As will be described below, the IDB loan was reduced to US$ 28 million, while the OPEC and DFID financing did not occur. 1.2 Original Project Development Objectives (PDO) and Key Indicators (as approved) The Project was the first phase of a three-phase adjustable program loan. The goals of the overall program were: (a) improve maternal and child health; and (b) help reduce morbidity and deaths of the poor from communicable diseases and inadequate environmental conditions. The project development objective for phase 1 of the APL was to increase access of the poor to better quality health programs and services. This access would be improved by strengthening the demand for services and the quality of the supply. The demand 4

15 would be strengthened through the implementation of a Seguro Materno Infantil (SMI, this was later renamed as Seguro Integral de Salud or SIS). The quality of supply would be enhanced by increased community participation (mainly through the CLAS system), by reducing fragmentation in the delivery of maternal and child and environmental health services and by adapting investments to local needs. Since the objective of the Project was to increase access, six access indicators were chosen as key performance indicators. These were: births attended by skilled health personnel; pregnant women receiving the recommended prenatal care; beneficiaries receiving primary care; child immunization of DPT3; children attended for acute respiratory infection; and, children attended for acute diarrhea disease As required for each phase of APLs, in addition to the key performance indicators, the Project had several trigger indicators that needed to be satisfied to determine the readiness for launching phase 2 of the APL. These consisted of five performance milestones: SMI implemented nationally; At least a third of all primary care clinics administered by CLAS; Fragmentation of mother and child and environmental programs reduced; Report on Benefit incidence of the investments of phase 1 and recommendations for phase 2; and, Loan fully committed and 80% disbursed. In addition to these performance milestones (which reflect the result of activities supported by the project), three policy actions were expected to be taken to promote policy-readiness for the second phase these actions were selected on the assumption that phase 2 of the APL would focus on the social security (ESSALUD) system of provision. These actions included: Legal framework for provision of services by ESSALUD and MINSA hospitals; Recommendations of policy studies on separation of financing and service provision in ESSALUD; Recommendations of policy studies on health manpower issues. 1.3 Revised PDO (as approved by original approving authority) and Key Indicators, and reasons/justification The original PDOs and Key Indicators remained unchanged. 5

16 1.4 Main Beneficiaries (original and revised, briefly describe the "primary target group" identified in the PAD and as captured in the PDO, as well as any other individuals and organizations expected to benefit from the project) Original beneficiary target: There were targets at the national level for the SIS component and targets at the project area level for the investment components. For the SIS component the target was to reach 1,378,000 pregnant women and 3,599,000 children less than 4 years of age 1 ; and for the investment components the target was to reach 815,000 pregnant mothers and 1,990,000 under-4 year olds in 10 departments 2 during the 4 years of project. Revised Target: There was a project restructuring in November The target population remained unchanged for the SIS component. For the investment components, the number of departments in the project area was reduced from 10 to 3 (which included the following four DISAs: Apurimac 1, Apurimac 2, Cuzco and Puno). At the time of restructuring, there was the expectation that the number of beneficiaries would be significantly less than the original beneficiary target in the project area, 339,000 pregnant mothers and 823,000 children under 4. As explained below however, that expectation proved to be unnecessarily pessimistic. 1.5 Original Components (as approved) The following four components were included in the original design of the Project and in Schedule 2 to the Loan Agreement: Component A - Strengthening health demand. The objective of this component was to reduce the economic barriers to utilization of health services through the implementation of the SMI, benefiting low income mothers and children through the: (a) reimbursement of eligible medical services provided by eligible health facilities in the Project area; (b) establishment and operation of the SMI unit and its regional offices; and (c) provision of technical assistance for the monitoring and evaluation of the SMI program implementation activities. Component B - Strengthening decentralization, policy development and institutional modernization. The objective of this component was to improve the quality of MINSA health programs through:(a) decentralization of the Borrower's health system, through the provision of the required technical assistance and training for the establishment and 1 Annex 2a of the Appraisal Document. 2 Puno, Apurimac, Cajamarca, Cuzco, Lambayeque, Piura, Tumbes, La Libertad, Madre de Dios, Lima Norte y Lima Sur. 6

17 strengthening of the CLAS; (b) provision of technical assistance and training to the DISAs and municipalities to improve their ability to manage the delivery of preventive, promotional and curative health services within their respective jurisdictions; (c) provision of technical assistance and training for the redefinition of the role of the MINSA in the areas of: (i) formulation of health norms and supervision of their implementation by public and private health providers; (ii) ensuring equitable access to health services by low income population; (iii) strengthening mother, child and environmental health programs; and (iv) promoting and maintaining public health; and, (d) provision of technical assistance to strengthen MINSA's monitoring and evaluation capacity for: (i) conducting two types of household surveys - a health and demographic survey to produce estimates on health outcomes (including infant and maternal mortality) and an annual community health survey 3, to serve as a basis for identifying local health needs; and (ii) establishing a yearly module to the quarterly household surveys (Encuesta Nacional de Hogares - ENAHO), in order to measure the coverage of key interventions. Component C - Improving the quality of health programs and services - Investment support for regional and local health plans. This component had as its main objective improving the quality in the supply of health services through the financing of regional and local health investment sub-projects in response to health priorities at the regional and local levels. Under this component, the Project would finance health-related civil works in project areas, acquisition of medical equipment and supplies and provision of technical assistance. The Project design assumed that the Pan American Health Organization (PAHO) would provide technical assistance to selected DISAs for the formulation of their regional plan, and if needed, in the supervision of the execution of these plans. Component D - Project coordination. This component would finance the establishment and operation of the Project Coordination Unit (PCU) in the Ministry of Health. 1.6 Revised Components When the closing date was first extended, in November 2003, MINSA, with Bank support implemented the following changes: (a) The logical framework was revised ensuring that it reflected a causality model, whose design was based on international evidence on maternal and neonatal mortality. The logical framework now called for monitoring five indicators that mainly explain changes in maternal and peri-natal mortality. (b) The demand component (Component A) was strengthened to support the improved targeting of SIS affiliations and attentions to the two poorest quintiles and to include an intercultural strategy to reduce linguistic and cultural barriers among pregnant women belonging to ethnic groups. 3 This annual community survey was originally supposed to be financed by DFID, which never materialized. 7

18 (c) Components B and C of the original design were treated as a holistic set of supply interventions and limited in geographical scope to the eight Regions where the highest incidence of maternal and neonatal mortality were found. Furthermore, the interventions were targeted within the Regions and standardized under WHO standards on obstetric services, leading to the proposed improvement of the quality of maternal and peri-natal services in 79 health facilities to be equipped with basic (BONF) or essential (EONF) obstetric and neo-natal equipment and staffed with qualified and trained personnel 4. The original thematic scope to be included in PARSALUD comprising child morbidity and mortality, communicable diseases (like AIDS, tuberculosis, etc.) and inadequate environmental conditions (causing malaria, yellow fever, dengue, cholera, rabies, inappropriate handling of food, etc.) were eliminated from the scope of the Project targeting instead only maternal and peri-natal mortality. (d) Procurement was centralized in the PCU. Accordingly, a new action-plan for the periods , confirming the Project's original developmental objectives, was agreed with the Bank and loan closing date was extended from December 31, 2003 to December 15, Other significant changes (in design, scope and scale, implementation arrangements and schedule, and funding allocations) The original design of the Project included a ground-breaking characteristic it included a component that would disburse against results (i.e. it would reimburse the government a certain amount for each child and mother covered by SIS, in a way that resembled the payment of an insurance premium). The reimbursements component, which is a common feature in projects recently approved in other LCR countries, was cancelled from the loan at the request of the Government. Thus, the World Bank loan was reduced from US$ 80 million to US$ 27 million and the IDB loan was reduced from US$ 87 million to US$ 28 million. Project activities were carried out as planned, but financing for reimbursements came from the National Budget. The cancellation meant that the Bank continued to be involved, through technical assistance, in the design, planning and implementation of the SIS, but to a much lesser extent than would otherwise have been the case. The investment component was originally planned as a fund, which would finance investments on demand from communities. The communities would justify their requests by developing a local health plan related to the area of influence of a health clinic. This component was amended to develop a system of investment planning (i) based on technical criteria developed by WHO to ensure the safe provision of basic obstetric and neo-natal functions (BONF) and (ii) to exercise planning at a macro-level relevant not 4 Essential Elements of Obstetric Care at First Referral Level, WHO (1991) 8

19 only for individual clinics but for networks serving multiple clinics and including higher complexity reference services needed to provide safe motherhood services. Accordingly, the following intermediate indicators were added, in order to monitor the proposed implementation plan: (i) percentage of attended deliveries which received oxitocin in the Project area; (ii) percentage of deliveries financed by SIS in the first two quintiles of population in the Project area; (iii) number of health facilities refurbished to comply with Essential Obstetric Function (EOF), Basic Obstetric Function (BOF) and Basic Maternal and Child services capacity; and (iv) percentage of references with respect to the number of attentions. 2. Key Factors Affecting Implementation and Outcomes 2.1 Project Preparation, Design and Quality at Entry (including whether lessons of earlier operations were taken into account, risks and their mitigations identified, and adequacy of participatory processes, as applicable) There were three aspects that were clearly linked to the Quality at Entry of the Project, as follows. Client orientation and timing of the lending process. The Bank preparation team demonstrated sensitivity and commitment to results that were in the GP's best interest and consistent with the government health policies and conceptual strategies existing at the time. In particular, the Project responded to the analytical work carried out by MINSA and the Bank on the reforms needed in the health sector. Lessons learned. The Project design was built upon experiences of relevant Bankfinanced operations over the nineties, including the then ongoing Peru Basic Health and Nutrition project, as well as the recommendations from the Bank's health sector analytical work, Peru: Improving Health Care of the Poor. Accordingly, the following five lessons were included in the design: (i) ensuring that the Project's developmental objectives were linked to a clear policy framework; (ii) confirming strong GP and stakeholder commitment to the health reform at the national, regional, municipal and local levels; (iii) instituting a flexible approach to launch sector-wide incremental reform changes addressing both demand and supply side interventions; (iv) providing results-based financing; and (v) promoting the decentralization process under a context of ensuring accountability and equity. In addition, the design benefited from the lessons derived from the SMI piloting. However, the project design: (a) was complex and included too many health related issues to tackle; and (b) did not take into sufficient consideration the potential difficulties that MINSA would have in obtaining information systems and human technical support for effective management of the intended reform process. Technical and institutional aspects. The lending process was supported by various analytical work, in particular the Peru: Improving Health Care of the Poor and other 9

20 MINSA commissioned papers. The lending phase profited from a benefit incidence of health expenditures, whose main findings appear in one of the Annexes to the PAD. In addition, a financial fiscal impact of the health reform was also estimated concluding that at the macroeconomic level it would represent a net addition of 0.1 percent of GDP and compared to a no-project scenario. A social assessment was carried out to detect relevant cultural patterns of users and health providers and to understand the gaps impeding the improvement and use of health services in general, and a Participation and Social Communication Strategy, as well as an Indigenous People Plan were derived from this assessment. Still, the intercultural strategy was designed only during the implementation phase and launched from mid onwards. This strategy turned out to be a key element in reviving the financial and physical implementation progress. The design recognized the need to address the multidimensional and cross-sectoral approach necessary to improve and maintain the population s health. Thus, a DPL program was designed (PSRL I-IV) which addressed the key policy issues in all three social sectors, including those policies being supported by PARSALUD I. In addition, preparation of this Project occurred under close coordination with other relevant operations in rural education, rural water supply, rural roads and indigenous population. On the down side, there were a number of loan effectiveness conditions in the original PAD that required longer than expected to be met. First, despite the fact that there was a draft Operational Manual at the time of appraisal, this still needed to be updated and approved by the Borrower. Second, according to Section I (Readiness for Implementation) of the PAD there was not an implementation plan and schedule available at the time of appraisal. Third, no agreements had yet been signed between MINSA and each of the DISAs in the project areas, neither with INEI for the carrying of the surveys included under component B. Fourth, DISAs had not yet submitted their first-year annual regional operating plan. 2.2 Implementation (including any project changes/restructuring, mid-term review, Project at Risk status, and actions taken, as applicable) There were five factors subject to government control that hindered a more effective and timely project implementation. The first factor concerns the political uncertainty caused by the resignation of President Fujimori in September 2000, the subsequent appointment of an interim government by the Peruvian Congress and the election of President Toledo that followed in July These unforeseen political events caused significant delays in the: (i) signing of the loan agreement; (ii) compliance with the conditions of loan effectiveness; and (iii) launching of the Project and its subsequent implementation. Not all the health authorities that led MINSA during the Project, amounting to six Ministers and eight Vice-ministers, bought into the Project ethos. Accordingly, marching orders issued by one authority were at times followed by counter-marching directives of the next, creating along the way confusion and paralysis among the implementing units waiting to 10

21 see if the health policies underpinning the Project were to be ratified or modified. Also, the various Bank supervision teams witnessed the dismantling by one administration of the technical unit that was selected and trained in a previous one. This deconstruction of in-built capacity for merely political reasons was extremely detrimental to the early stages of project implementation, given that it took time to replace the skilled staff and train them. The second factor refers to the initial difficulties derived from the fusion of the SMI, previously managed by the PARSALUD administration, and the SEG, into the SIS, an autonomous institution. Initially, there was lack of coordination between MINSA, SIS and PARSALUD resulting in different approaches towards the envisaged health reform. The Bank, through the PSRL DPL loan, as well as through the PARSALUD, carried out a dialogue with the various parties, to promote a better understanding of the different roles that each institution had as part of the proposed reform. The third factor refers to the tensions existing between the central health authorities of MINSA and the regional health authorities, as part of a slow decentralization process where the delegation of functions were not clear. The Performance-based Management Agreements, explained elsewhere in the ICR, clarifying roles, responsibilities and accountability mechanisms, were to be the effective instruments addressing these operational tensions that were delaying project implementation. The fourth factor refers to the budgetary restrictions imposed by the Ministry of Economy and Finance (MEF) during the Project s early years, where counterpart resources were not released in a timely manner. As a consequence, project activities requiring counterpart funding were delayed. These budgetary restrictions were in turn a natural response by MEF to the Project s lack of physical and financial implementation progress in its early years. The fifth factor refers to the long delays in having the Sistema Nacional de Inversión Pública (SNIP) 5 approve the civil works and some large acquisition of Project-related goods. In addition, there were two major factors under the implementing agency's control which slowed down the execution pace. The first one concerned to the difficulties in establishing an effective PCU team. There were frequent changes of project management derived from the political changes at the Executive Branch. There were frequent changes of Ministers and Vice-ministers of Health, resulting in gaps and implementation delays until new decision making teams were put in place in the PCU. The design also contemplated a small PCU with no more than 10 staff whose only purpose was to coordinate and facilitate the implementation of the Project through the mainstream units within MINSA, DISAs and CLASs in order to minimize the parallel "bubble" 5 The concept of SNIP emerges in 2002 within the Ministry of Economy and enacted as a project appraisal instrument for the control of the pertinence and quality of publicly-financed investments in all the sectors of the GP. 11

22 institutional arrangement. While the PCU was small at the initial stage, it grew to over 50 staff at the national level including one Project coordinator in each one of the eight participating DISAs. This growth became necessary to ensure successful implementation of projects financed by both the World Bank and the IDB, which had not yet harmonized procurement or financial management processes. The second factor refers to the lack of an effective and relevant implementation plan and schedule in the time period, hindering the translation of the Project s conceptual philosophy into concrete and targeted actions. The causality model was developed in 2003, at which point the results oriented plan was designed, guiding project implementation from 2004 onwards. 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization The PAD included baseline information at the aggregate national level, which was available during the appraisal stage. The aim was to provide the required DISAs baseline data through the 2000 DHS survey and to measure progress in health conditions, including infant and maternal mortality, through a second survey in In addition, other sources of information, both from national surveys and from MINSA statistics systems were expected to provide the necessary information for the Project monitoring system. While the launching of the monitoring was slow, towards the end of 2003, the Project developed a monitoring system, based on MINSA information systems that allowed timely reporting of key Project indicators in the eight DISAs and nationally for SIS indicators. This system was regularly used by Project staff to ensure that investments were targeted to critical areas within these jurisdictions. In addition, the system contributed significantly to monitor SIS targeting to the poorer quintiles. The project also implemented a training program on monitoring and evaluation in the eight DISAs to develop local capacity in monitoring maternal and child health indicators. 2.4 Safeguard and Fiduciary Compliance (focusing on issues and their resolution, as applicable) Financial management, procurement and environmental assessments were done during the preparation phase. The procurement assessment rated the project as average risk with respect to procurement aspects and recommendations were included in the draft Operational Manual. The financial assessment indicated that there was not an adequate project financial management system in place to handle a project management reportsbased disbursement scheme and an action plan was agreed upon with the Borrower to address these financial monitoring gaps. 12

23 The project complied with OP/BP (financial management), OP/BP (procurement), OP/BP (Disbursement), OP 4.01 (environmental assessment) and OP (indigenous peoples). This Project was subjected to annual random ex-post procurement and in-depth financial management reviews which found no major issues on procurement related matters and compliance with Bank financial management requirements. Recommendations emerging from these reviews were appropriately addressed by the Borrower. During the life of the Project, financial management reporting by the Borrower changed from the Project Management Reports (linking disbursements with the achievement of physical targets) to the current quarterly Financial Management Reports (de-linking in the reporting the achievement of physical targets from the actual disbursements). The Project was classified as Category B with respect to environmental risks. After the inflection point of November 2003, the various supervision teams regularly monitored the compliance with the environmental regulations and appropriate procedures for handling and managing medical waste at the participating health facilities, finding no relevant issues. The Project addressed key issues linked to Indigenous peoples through a number of activities linked to cultural adaptation of services, including providing for vertical birth, and training of health professionals in local culture and language. 2.5 Post-completion Operation/Next Phase (including transition arrangement to post-completion operation of investments financed by present operation, Operation & Maintenance arrangements, sustaining reforms and institutional capacity, and next phase/follow-up operation, if applicable) Given that this Project corresponds to the first of three envisaged phases of an APL, the whole issue of sustainability can only be definitely addressed in the ICR corresponding to the completion of the final phase, since many of the activities included in the first phase are to be continued under the second one. At this stage, however, there are four dimensions of the Project that lead to a ranking of the Project sustainability as likely. First, the new Government of President Alan Garcia has ratified that reducing maternal and peri-natal mortality is an unfinished agenda in the health policy of the vulnerable population, and as such, it has been granted priority in the political health platform of the new administration. In addition, the incoming administration continues to adhere to the achievement of the Millennium Development Goals, one of which is reduction of maternal and child mortality rates. Finally, the new Government has ratified its support to the Acuerdo Nacional (National Agreement) which includes goals and indicators on maternal and child mortality. Second, the SIS scheme is to be continued and strengthened. The CY05 budget for its five programs amounted to US$77 million equivalent (US$56.6 million equivalent for Programs A and C related to the Project) and the forecasted CY06 executed budget will 13

24 reach more or less a similar amount. An equivalent amount for the 2007 budget is expected. Third, the basic and essential obstetric protocols including the competency-based standards required by skilled professionals at the health centers to attend maternal and peri-natal services provided by the Project have been adopted and regulated at a national level by MINSA. Likewise, the causality chain model developed by PARSALUD is now a modus-operandi of MINSA. Fourth, the BONF and EONF infrastructure built and/or rehabilitated and equipped by the Project in 74 health facility centers is being utilized by an ever growing number of pregnant women belonging to the lowest two income-quintiles resulting from the demand financing through SIS, the promotional activities and the intercultural strategy implemented by PARSALUD. Despite this achievement, preventive and corrective maintenance of these health facilities is a pending agenda, and therefore, the mediumterm sustainability of these health facilities is likely to be compromised, unless this issue is addressed. The MINSA is currently negotiating with the Regions, as part of the decentralization process, for the inclusion of maintenance items in their budgets. With respect to the transition arrangement to regular operation, the SIS, the maternal and peri-natal health protocols and the intercultural strategies utilized to increase the effective demand of poor vulnerable pregnant women are already institutionalized in the mainstream institutions. The Bank is currently preparing the second phase, PARSALUD II, which is also expected to support the maternal and child mortality reduction objective, through both the demand and supply strategies of the first phase. In addition, this Project aims to support the GP's ongoing decentralization process. As already mentioned, appraisal is expected in May 2007 and Board approval in September Assessment of Outcomes 3.1 Relevance of Objectives, Design and Implementation (to current country and global priorities, and Bank assistance strategy) Relevance of objectives. The objective - increase access of the poor to better quality health programs and services - continues to be highly relevant in present-day Peru. The new December 2006 Country Partnership Strategy notes that in recent years Peru has leapt forward in terms of improving the key health indicators of the population, including a large acceleration in the rate of reduction of the infant mortality rate. The CPS proposes to implement a cluster of activities (including AAA, investment loans, DPLs and grants) over the 2006 to 2011 period to sustain the gains of recent years. The CPS explains that the main successes of recent years in health have been in access and that in the next stage the emphasis must be to continue improving results by improving quality and sustaining the gains in coverage. 14

25 Relevance of design. The design of phase 1 of the APL is also relevant for present day Peru as evidenced by the Government's desire to use a similar design in the preparation of phase 2 of the APL (this new project is in the pre-appraisal stage, having gone past a crucial first threshold of review by the highly demanding National System of Public Investment or SNIP). The main features of design included: (i) the implementation of the APL as part of a cluster of Bank instruments, including careful analytical work done under a AAA, intensive policy dialogue with high level economic and sectoral authorities supported by a series of DPLs and capacity building financed by the APL; (ii) the simultaneous use of components to strengthen the demand for services (through the implementation of the SIS) and the supply of services financed by an investment component; and (iii) the use of flexible investment planning instruments for regional sub-projects that led to investments adapted to different regional needs. Relevance of implementation. On the demand side, the reimbursements provided by SIS are an effective means of reducing the economic barrier faced by potential health service users. Given that reimbursement processing requires fluid transaction processes, different from the traditional flow of funds and fiduciary functions used by the national budget, the implementation was initially handled by the PARSALUD team. This mechanism was used in order to allow access to the more flexible procedures associated with Bank projects instead of the more rigid procedures associated with budget financing. After the initial phase, new norms were passed allowing SIS to use the reimbursement procedures developed under the project, but under a sustainable institutional structure. On the supply side, the provision of services through networks composed of lower complexity facilities (health posts and BONF facilities) and higher complexity facilities (EONF) has shown to be an efficient and effective strategy to respond to dispersed populations. The challenge that remains for the second phase is to ensure that the reference and counter-reference system works adequately and is subject to SIS reimbursement. As part of the decentralization process, the Regions now have the responsibility of planning and designing public investments in all sectors, including health. The Project team worked closely with the eight Regional Health Directorates to develop investment plans targeted to maternal and child health results. These plans include the detailed requirements set by the National Public Investment System (Sistema Nacional de Inversión Pública, SNIP). 3.2 Achievement of Project Development Objectives (including brief discussion of causal linkages between outputs and outcomes, with details on outputs in Annex 2) PDOs have been successfully achieved: (i) Most of the targets for the Project Development objectives of phase 1 of the APL (the key performance indicators described in the text of the PAD and in Annex 1) have been met or surpassed. (ii) The Program Development Objectives (i.e. the objectives for the three phases of the APL) have targets for 2010 and not for each phase of the program because at the time of appraisal it was decided that they could not be measured with precision or sufficiently influenced over shorter periods of time; however, existing data suggest that the targets for the program 15

26 are likely to be reached long before the original target date, and (iii) all performance milestones identified as triggers for the successful completion of phase 1 have been met. These achievements are described below. Two Program Development Objectives (or higher level outcome indicators) were utilized: the infant mortality rate (IMR) and the peri-natal mortality rate. The most recent data on IMR (IMR of 22.3 per thousand in ) shows a strong acceleration in the reduction of infant mortality compared to the trend that existed before Assuming that this trend has continued, the target for 2010 (IMR of 20) was already reached in 2005, five years ahead of schedule. The speed in the rate of reduction of IMR during is higher in Peru than in all other countries in LCR for which data is available (See Figure 1, and Figure 2). Unfortunately, no data is available for peri-natal mortality at this point, however many analysts believe that the existing gains in IMR are suggestive of gains in peri-natal mortality. 7 Figure 1 Infant Mortality Rate increase/decrease (%) between 2000 and 2004 in LCR 10% 7% 5% 0% -5% -10% -6% -8% -9% -6% -15% -14% -11% -10% -15% -20% -25% -30% -27% Argentina Bolivia Chile Colombia Ecuador Mexico Paraguay Peru Venezuela Uruguay Source: The World Bank's Health, Nutrition and Population data platform. 6 This IMR was measured by the National Statistics Institute (INEI) through the Continuous National Survey (Encuesta Nacional Continua). The reference period for the measurement of IMR gives an estimate for Infant mortality is a measure of death of children under 1 year of age; it can be decomposed into perinatal mortality (deaths under one month of age) and mortality of children one month to twelve months of age. At relatively high levels of IMR (of 60 and over) most reduction in mortality is achieved by prevention of deaths in the later months of the period (deaths due to diarrhea or vaccine preventable disease), but as IMR reaches lower levels, it's continued reduction increasingly requires improvement in peri-natal mortality. 16

27 Figure2 Infant Mortality Rate (per 1000) LCR mean* Peru * LCR's countries wich started with IMR of 30 or more in 1990 (Bolivia, Colombia, Ecuador, Mexico and Paraguay). Source: The World Bank s Health, Nutrition and Population data platform. The large achievement in IMR is the result of many factors and is not exclusively a result of the expansion in coverage. However, the expansion of coverage was prominent during the Project period and it was strengthened by the cluster of activities supported by the Bank in health. PARSALUD worked together with the PSRLs and with the AAA to support the Government's policies that strengthened demand for health services through the national implementation of SIS, and the implementation of SIS was clearly a determinant in the large expansion of coverage during the period. Coverage also benefited from the improved quality of supply in the poorest regions of the country that was partly obtained by reforms and investments that were also supported by the cluster of Bank health activities. The achievements in the key performance indicators of the project are summarized in Table 1. Targets have been achieved or exceeded for births attended by skilled health personnel, beneficiaries receiving primary care, children attended for acute respiratory infection and for acute diarrhea. Yellow fever is now considered a high risk disease and by law all cases identified and reported must be attended 8. In immunizations, the numerical target of coverage was not reached because the immunization program became more ambitious in its scope. Instead of continuing expanding coverage of DPT3 a vaccine that is simple to apply and has low logistical requirements a more complex combination of vaccine was adopted (the Pentavalent, which includes DPT3, Hepatitis B and Haemophilus Influenzae type b vaccines 9 ). This new combination of vaccines is more powerful in its impact on mortality and morbidity (and is likely to have contributed 8 Legal framework: "Directiva N OGE OEVEE". OGE (MINSA) Protocols of Epidemiological Surveillance - Part I. 9 Legal framework: "Directiva N 033-MINSA/DGSP-V

28 to the acceleration in the improvement of outcomes), but is more demanding in its application and logistics. Regional and local health plans were institutionalized in 2002, as prescribed in the General Health Law (No of 2002). Each Region carries out a Health Situation Analysis (ASIS) which describes the sanitary conditions annually. The ASIS describe prevalence of communicable diseases in the region, as well as the climatic, social and cultural conditions. When relevant, the documents include description of environmental health problems, for example, Apurimac I ASIS 2004 includes detection and control of food and water contaminations and Cusco ASIS 2005 includes air pollution indicators. In addition, each region signs with the central Government Annual Management Agreements (Acuerdos de Gestión) in which they agree to meet certain targets with respect to specific health results which include the prevention and treatment of communicable diseases as well as other prevalent health problems. Table 1 Project Development Objective Indicators Indicators Baseline (2000) Target Result (2005) Achieve Percentage of births which had more than or equal to 4 antenatal visits by a skilled professional (Project area) 32.20% 51.36% 57.20% YES Percentage of deliveries attended by a skilled health professional (doctor or trained midwife). (Project area) 27.60% 36.70% 50.90% YES Number of cases of acute respiratory infections (IRA) in the children under age 5 attended (annual)* 8,747,589 3,700,000 7,742,457 see below Number of cases of acute diarrhea (EDA) in children under age 5 attended (annual)* 4,535, ,000 4,531,800 see below Percentage of children between 18 and 29 months vaccinated with DPT3 (Project area)** 78.30% 95% 87.41% NO Number of beneficiaries receive care in the primary facilities of MINSA annually 17,000,000 YES Increase in the number of municipalities with DPT3 coverage for children under age 1 ** - 44% - see below Percentages of yellow fever cases attended *** - 80% 100% YES Percentages of health departments (DISAs) which are implementing regional and local health plans that respond to communicable diseases and environmental health problems prevalent in their respective localities**** - 80% 100% YES * IRA and EDA targets were established originally as absolute numbers due to lack of baseline numbers during Project preparation. Absolute numbers are difficult to interpret, particularly for these two illnesses. ** DPT3 was replaced by a more complex vaccine, so that its coverage at the regional level has been reduced in the last years. Data at the district or municipality level is not available. *** Yellow fever is a disease under weekly national surveillance. By law, the reported cases have to be attended. **** All DISAs, regularly sign annual Management Agreements with the Ministry of Health, reflecting commitments with respect to regional health priorities and indicators with quantitative goals for the period. 18

29 The performance milestones designed as triggers have all been met: The SIS (which replaced the SMI) was implemented nationally in 2002, ahead of the scheduled date of A third of primary care clinics were administered by CLAS. This strengthened the quality of service in these clinics. 10 In early 2006, during the final months of the project, the Government changed the labor legislation giving CLAS employees the status of civil servants. The Government will need to ensure that quality gains are not lost over the next few years as the new legislation is implemented. Fragmentation of the formerly "vertical" programs was reduced. The MINSA area responsible for the vertical programs was restructured into age-based groups and the Integral Health Model was instituted. The budget was integrated for all vertical programs. The information systems still need to be unified through an integrated MINSA monitoring system. This system is being supported by a separate investment loan (Accountability for Decentralization in the Social Sectors TAL). The loan was fully committed and disbursed. An assessment of the results of the project was completed by the government to produce lessons and recommendations for the formulation of phase 2. The lessons have been incorporated into the design of PARSALUD 2. An analysis of the incidence of health policies (with an application to the project area as data permited) is presented as part of Annex Efficiency (Net Present Value/Economic Rate of Return, cost effectiveness, e.g., unit rate norms, least cost, and comparisons; and Financial Rate of Return) There have been important gains in effectiveness, efficiency and equity of public spending on health associated with the initiatives supported by the Project, as shown in Annex 3 and summarized in this section. Coverage of SIS (for the maternal and child programs of SIS, which were the main focus of project support) increased significantly. While the targets in the PAD were considered ambitious at the time of appraisal, they were surpassed and sustained over time, using funding by the treasury. By the end of the Project, 503,000 pregnant women and 1,254,000 children (younger than 5 years of age) were expected to benefit from the SMI, while the actual numbers were, during the last year of full Project implementation (2005), 649,000 pregnant women and 3,444,000 children. Efficiency of existing health sector resources was raised by increasing their utilization. While there were some investment in works and equipment during the Project period (including those financed by the Project), there were small increases in overall public 10 CLAS have been shown to have greater productivity per employee, longer hours of operation, higher levels of patient satisfaction and greater levels of community participation (see RECURSO, p ) 19

30 health expenditures (see Annex 3). Hence the greater demand generated by SIS was satisfied by achieving higher levels of productivity in previously underutilized facilities. SIS contributed significantly to improving the targeting of public health expenditure through a direct and an indirect effect. Directly, SIS targets the poor through geographical and means tested procedures. A benefit analysis of all the large social programs in Peru, found that in health SIS is the best targeted program (with a concentration coefficient of -0.08, compared with a concentration coefficient of 0.21 for expenditures in MINSA hospitals (see RECURSO p. 15). Indirectly SIS also contributes to improving the distribution of expenditures of other MINSA facilities: by financing the co-payment that allows the poor entry to a hospital, it directs the (usually larger) subsidy to the facility to the poor beneficiary. Annex 3 also presents significant estimates for the net present value of the Project and shows that it had a low fiscal impact: The mother and child components of SIS have a relatively small cost (5% of public health expenditure or 0.07% of GDP), and this increase has been easily absorbed in the budget. Investment expenditures in 2005 were only 0.4% of public health expenditures and their recurrent cost implications are also minimal Justification of Overall Outcome Rating (combining relevance, achievement of PDOs, and efficiency) Rating: Satisfactory The project is rated as satisfactory, considering that relevance is judged to be satisfactory, achievement of PDOs is judged to be satisfactory (but as explained, the attribution for the outcomes is likely to be stronger for the cluster of Bank activities that had those PDOs than for each of the activities individually), and efficiency is also judged satisfactory. 3.5 Overarching Themes, Other Outcomes and Impacts (if any, where not previously covered or to amplify discussion above) (a) Poverty Impacts, Gender Aspects, and Social Development The Project did not directly measure poverty or gender indicators. However, the main beneficiaries of this Project are women and children and positive impacts were determined with respect to maternal and child health. Thus, the Project is very likely contributing to overall objective of poverty reduction, with a particular impact on women. (b) Institutional Change/Strengthening (particularly with reference to impacts on longer-term capacity and institutional development) 11 Source: PIU PARSALUD I, SIS and SIAF/MEF. 20

31 SIS is without doubt the most significant institutional innovation of the Peruvian health sector in the last decade. It consists of a system of reimbursements for services produced by health providers to poor users. It includes systems of geographic targeting, identification of poor users and defined benefit packages. The system was expanded and developed with support from the project and is today an essential instrument of the health system, fully institutionalized in its normative and in its financing. (c) Other Unintended Outcomes and Impacts (positive or negative, if any) 3.6 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops (optional for Core ICR, required for ILI, details in annexes) 4. Assessment of Risk to Development Outcome Rating: Low or Negligible The Risk to Development Outcome is considered to be low, both for the demand and supply side components linked to the maternal and child health results. On the demand side, SIS is an institution that is strongly supported by the current Government, as shown by the continuous budget support provided and recognized as a valuable service by the population. On the supply side, the new health standards and reference systems have been made official within MINSA. Furthermore, the second phase of this APL and an upcoming HD DPL program are expected to continue supporting the maternal and child objectives. 5. Assessment of Bank and Borrower Performance (relating to design, implementation and outcome issues) 5.1 Bank Performance (a) Bank Performance in Ensuring Quality at Entry (i.e., performance through lending phase) Rating: Satisfactory PARSALUD was identified in mid-1998, appraised in September 1999, negotiated in November 2000 and approved by the Bank s Board on December 16, Project preparation activities were financed by a Japanese Grant (JPN PE) amounting to US$778,000 all of which, minus US$177.45, were actually spent. There was strong coordination among the Bank and IDB's preparation teams. Preparation missions were carried out jointly. Each Bank prepared its own separate project appraisal documents, whose technical contents by and large coincided. There were some 21

32 differences, especially with respect to the expansion rate of the SIS (where the IDB took a more gradual approach) and with respect to the kind of support that was needed to strengthen decentralization and modernize MINSA. The first coordinator of the PCU ensured the merging of the two appraisal documents into a joint version to allow the PCU to work on one project description instead of two. More than eight months elapsed between the preparation mission carried out in November 1998 and the pre-appraisal missions of July The lack of a Decree by the GP formally enacting the SMI (which lacked an institutional and organizational framework) and clarifying some operational issues, which were crucial for the Project design, was the main reason for this delay. The Project preparation process complied with most quality at entry criteria. The Project was developed following detailed sector work and was an integral part of the Bank s assistance to the health sector. The strategic relevance of the Project design, which addressed the demand for health services, has shown to be critically important for the improvement of health results. The economic analysis carried out during project preparation was sufficient to support this strategic approach. Further to this, Bank fiduciary assessments were carried out as per institutional requirements. On the other hand, the following difficulties were faced during the preparation process: (i) while institutional arrangements were clearly identified, the design of the Project lacked a concrete action-plan and timetable, including activities for the continued strengthening of the SMI's accounting and internal control systems; (ii) the monitoring and evaluation arrangements were not sufficiently ready for most of the large list of output and impact indicators included in Annex 1 of the PAD; (iii) detailed baseline information and mid-term benchmarks for the primary health centers at the participating DISAs were not provided at appraisal; (iv) although the Borrower had prepared a draft Operations Manual by appraisal, the final version was greatly delayed; and, (v) the preparation process coincided with the end of an elected administration, and effectiveness conditions overlapped with change in technical teams, resulting in a delay of approximately 19 months. (b) Quality of Supervision (including of fiduciary and safeguards policies) Rating: Satisfactory The Project suffered significant delays in signing (15 months after Board approval) and achieving effectiveness (19 months after Board approval). These delays were mainly caused by frequent changes in the Executive Branch of the GP and the resulting unsettling political scenario that was consequently unleashed, explained in greater detail under section The Project, launched during the November 16, 2001 supervision mission, was subjected to a total of 11 field supervision missions during its life, including one mid-term review (September 25, 2003). On average, two field supervision missions were carried out 22

33 annually during the six and one half year life span of this Project. The diverse supervision missions included the appropriate skill-mix to carry out their envisaged terms of reference. A total of four task team leaders (TTLs) were in charge of the lending and supervision of this Project with tenures ranging from a minimum of 10 months to a maximum of about four years. There was sufficient continuity between TTLs so that Project supervision was not affected. The day-to-day supervision of this operation benefited from the presence of sector-related Bank staff in the office in Lima. The project electronic files include 16 PSRs/ISRs, implying that there is not necessarily a one-to-one correspondence between field supervision missions and reports, as continuous oversight of this investment was carried out on a continuous basis by the Bank Peru office staff. Supervision reports were updated, on average, every semester, and formal Bank missions led by Washington staff were undertaken less often. The quality and pertinence of the information in the PSRs/ISRs is a mixed-story during the life of the Project. As there was significant delay in the development of the Project monitoring and evaluation system, developmental and intermediate outcome indicators were not included in the first thirteen PSRs. From PSR 14 of April 26, 2005 onwards, these indicators were reported. Accordingly, Bank management lacked information as to the trend in achieving the project developmental objectives for five of the six and half years of the life of this first phase of the APL. The Bank supervision teams assigned a satisfactory rating to the achievement of developmental objectives in the totality of the first thirteen PSRs and justified its ranking by a qualitative narrative in the reports. These ratings are not strongly supported with hard evidence. On the other hand, it is important to recognize the concerted effort by the Bank supervision teams since 2004 to have the Borrower address the collection of information, analysis and production of outcome indicators reports in a more regular and consistent fashion. On the other hand, the Bank supervision teams displayed significant objectivity, downgrading the ratings of implementation progress, monitoring and evaluation and management when implementation stalled and increasing the Project's risk rating. Once the Project was removed from the Bank's "Problem Project List" by the mid 2002, some of these ratings were improved accordingly. The various TTLs assigned to this operation displayed appropriate sensitivity and flexibility to adjust the Project to the existing contextual changes leading to the partial cancellation of the loan, the two amendments to the LA and the two extensions of closing dates. The cancellation of $53 million from the original 80 million occurred between Board approval and the signing of the loan, requiring that the team address the restructuring of targets, costs and implementation schedule. Also, discussions held with the Borrower in the latter part of 2003 to recommend the first extension to the closing date, led to the turning and inflection point of the life of the project, as detailed in section 6.6 of the ICR. Once evidence was made available to the Bank during the second semester of 2004 about the degree to which some of the agreed first phase triggers were 23

34 being attained, the Bank supervision teams were able to build a case in support of the financing of a second phase. Consequently, preparation activities of PARSALUD II were started in the second semester of 2005, aiming to appraise the proposal on or about February 2007 and take it to the Board in May There was a good working rapport between the Bank and the Borrower as well as a close and coordinated effort with the IDB, producing extremely good results for the Project. Up to mid-2004, the co-financing of these two multilateral institutions was divided geographically by DISAs. The Bank supervision teams from mid-2004 onwards were responsible for convincing the IDB and the Borrower to adopt a more holistic financial approach. Accordingly, for example, the intercultural strategy financed by the Bank was utilized in the eight DISAs, and ambulances financed by the IDB loan were provided to participating health facilities in all DISAs. IDB financing will allow the bridging to the second phase given that the Bank loan closed on June 30, As noted above, while this Project was under implementation, the Bank also supported the Government of Peru through the PSRL DPL series, with the same technical team. This strategy led to important synergies in terms of ensuring that policies agreed upon in the DPL were implemented in the Health sector with adequate financing. This was particularly relevant for the SIS financing given that, after partial loan cancellation, it was fully financed by the National Treasury. (c) Justification of Rating for Overall Bank Performance Rating: Satisfactory Based on the above lending and supervision observations and ratings, the ICR rates the overall Bank performance as satisfactory. 5.2 Borrower Performance (a) Government Performance Rating: Satisfactory At the lending stage. This Project was prepared simultaneously with the IDB-financed Project, under the agreement that there would be a common Development Objective focused on maternal and child mortality and that both loans needed to be implemented under one Project Coordination Unit. The counterpart team was composed of a group of qualified consultants selected from the then ongoing IDB-financed project. Project preparation activities, of both proposals, were coordinated by the Unidad de Coordinación para la Modernización del Subsector Público de Salud within MINSA. Some of the consultants in the preparation team were familiar with the implementation of the previous Bank-financed Basic Health and Nutrition Project, thus they had previous 24

35 experience with Bank operations, procedures, especially in procurement and financial management. The Project design benefited from the findings derived from the SMI piloting in the DISAs of Tacna and Moyobamba, especially with respect to the reimbursement mechanism and their required information system and the formation of health networks with applications of integrated health care protocols for mothers and children. Other pilots were carried out to test the feasibility of introducing in the Project design the concept of performance-based payment. No major issues surfaced during preparation. Hence the Government s performance is considered satisfactory. At the implementation stage. This and the next section of the ICR assess the project along four distinct periods in its implementation cycle. The original four-year implementation period (December 1999 to December 2003) was increased to six and half years (to June 30, 2006) and transcended four administrations. The loan closing date fell one month short of the inauguration of the newly elected government of President Alan Garcia. The first period, covering from Board approval in December 1999 until the arrival of the elected President Toledo in July 2001, was characterized by continuous political changes, including Ministers and Vice-ministers of Health (some with tenures of less than one year) and corruption scandals hindering effective compliance with the original Project implementation schedule. This period was plagued with health policy uncertainties, orders and counter-orders leading to a highly unsatisfactory rating by the Bank of the Borrower's performance, listing the investment in the Bank's Problem Project Category and considering the possibility of canceling the loan. During this period, the Project had to be restructured, while keeping the same developmental objectives, due to the partial cancellation of the Bank and the IDB loans. The second period, covering from July 2001 until November 2003, was characterized by unsatisfactory physical and financial implementation progress. The few activities undertaken by the project, mostly hiring of consultants to support the development of local health plans and some civil works, were highly unfocused and carried out in a context of lack of priorities. Total disbursements of loan proceeds by the third quarter of 2003 amounted to about US$7 million. This period was also characterized by high turnover of qualified technical teams at all levels of the Project's decision making process, mainly for political reasons (presumed affiliation to the Fujimori regime) without replacement with other qualified staff, largely for lack of institutional capacity to do so. During this period, the SMI (previously managed by PARSALUD) and the SEG (which had its own management team) were fused into SIS, an autonomous agency in MINSA. At this point, difficulties typically faced by new institutions arose. The various sector institutions (SIS, MINSA, and PARSALUD) were not clear on their roles in the health 25

36 reform process. In order to align the PCU to MINSA, the Health Minister, in December 2001, decided to modify the Project's oversight arrangements included in the PAD by vesting these functions in a Coordinating Committee. The Committee was composed of MINSA department heads, as well as the SIS Director, chaired by the Vice Minister of Health and with Secretariat support from the PCU. Although coordination improved slightly, the Project lacked sufficient political support to launch significant reform activities. As a result of a badly-managed decentralization process, the relationship between the national (MINSA) and regional (DISAs) institutions was tense. Roles and responsibilities with respect to the reform process were not clear and the lack of effective accountability mechanisms was not appropriately addressed during this period. In order to address this national-regional issue, MINSA inaugurated in 2002 a system of Performance-based Agreements (Acuerdos de Gestión) with all the DISAs, establishing observable and measurable targets for health care coverage for each region and holding them accountable for reaching these targets, while giving them greater flexibility in managing their budgets. This was seen as a positive and constructive move to foster decentralization and to decrease micromanaging procedures from the center. These agreements have been functioning ever since, and together with the improvements of the health management information systems producing relevant and timely data since the end of 2004, have become powerful tools to redirect and target public resources addressing specific areas and health issues. MINSA was not fully successful in fully integrating all vertical programs. The budget was integrated and managed by one administration unit. The maternal and child programs were also integrated at the facility level, but this was not the case for other vertical programs such as vaccination, malaria, TB etc. Monitoring systems were never integrated and each area in MINSA manages its own system. The Project s rating for Implementation Performance was upgraded to satisfactory in mid At the time,, the Borrower had satisfactorily complied with the: (i) establishment of a functioning Coordinating Committee for PARSALUD under Ministerial Resolution No dated December 10, 2001; (ii) hiring of almost all the key PCU senior staff; and (iii) approval of the Annual Operating Plan and the Investments Manual. Finally, by early 2003 it had become obvious that the SIS's maternal and child programs were not adequately covering poor, mostly rural areas, which was considered critical for achieving of the Project's developmental objectives. The targeting was eventually corrected during the third and fourth periods, through intensive promotional activities increasing access to the SIS among vulnerable population and introducing intercultural strategies to remove linguistic and cultural constraints to childbirth. The third period, running from November 2003 until the appointment of the last PCU coordinator in March 2004, was characterized by a significant improvement in implementation following a refocusing of Project interventions. Some of the outcomes initiated in this third period, but better observed in the fourth one, are: (i) improved 26

37 targeting of SIS to poor and vulnerable populations (where the PSRL dialogue also contributed); (ii) a more effective articulation between PARSALUD and SIS via satisfactory implementation of technical assistance to strengthen the SIS (financial, auditing, information, etc) and close monitoring of key outcome/output indicators in the participating DISAs; (iii) immediate budgetary increases for PARSALUD; and (iv) monitoring overall sector reforms progress including performance agreement with public providers, decentralization and community participation. The fourth period, covering from March 2004 until the end of the first phase of this APL, is characterized by the Project s satisfactory physical and financial progress as reported by all of the Bank's supervision missions since early 2004 onwards. The physical accomplishments are detailed in sections 3.2 and Annex 2 of the ICR. Rating: Weighting the highly unsatisfactory performance of the Government up to 2003 with a positive one since November 2003, the ICR team ranks the Government implementation performance as satisfactory. (b) Implementing Agency or Agencies Performance Rating: Satisfactory The PCU had five coordinators: (i) the first one, in place during the Project s appraisal and loan negotiation, lasting until October 2001; (ii) the tenure of the second one lasted 3 months, from October 2001 to January 2002; (iii) the third one lasted 7 months, from January 2002 to September 2002; (iv) the fourth one was in place for 16 months (September 2002 March 2004); and (v) the fifth one was in charge from March 2004 until December These frequent changes in command, and the short tenure of some of the PCU coordinators, compounded with all the political uncertainties and structural issues described before, resulted in much discontinuity that caused significant delays with the compliance of agreed implementation schedules. In addition, these changes also caused high turnover of key trained personnel within the PCU, with consequent impact on Project implementation. The negative impact of the above-mentioned changes in the PCU is reflected in the history of the Project s financial and physical implementation progress. The staff involved in Project preparation was not the same as those responsible for its early implementation, resulting in an initial disconnect. The first loan disbursement of one US million dollars took place on January 10, 2002, about 25 months after loan approval and six months after loan effectiveness. Two months later, a reimbursement of US$1.07 million was made by the Bank against retroactive financing for eligible expenditures incurred after the November 1999 appraisal mission. This implies that for more than 30 months after the completion of the appraisal mission (September 1999 to March 2002), very few Project activities were implemented. The period 2002 and 2003 was characterized by a sluggish disbursement pace averaging about US$4 million per year. At the time of the original loan closing date, only one third of the loan proceeds had been disbursed. Starting in early 2004, loan disbursements 27

38 began to show a more vigorous pace averaging about US$8.4 million per year, twice the amount shown in the period. The action plan and schedule prepared by the PCU contributed to speed up the physical and financial implementation of the Project. This action plan was also seen as a requirement to bring the first phase of the APL to a smooth and timely completion, and to finance preparation activities for the second phase of the APL. The ICR recognizes these efforts made by the implementing agency, which seem to have contributed to better health-related outcomes and full disbursement of the total loan. The monitoring and evaluation capacity of the PCU, absent during the first years of project implementation, improved by 2004 once the causality model was established, and the process to collect the baseline information in the participating DISA started. By 2004, the Project's monitoring and evaluation scheme was assisting in the programming and targeting of Project-financed investments in the eight participating DISAs. After two failed attempts to select a consulting firm, a qualified consultant was selected. The final report for PARSALUD was delivered in December (c) Justification of Rating for Overall Borrower Performance Rating: Satisfactory Based on all of the above-mentioned observations, borrower performance is rated as satisfactory. 6. Lessons Learned (both project-specific and of wide general application) (a) Political volatility has a detrimental effect on Project implementation and impact. Frequent changes of political appointees frequently result in changes in priorities that lead to confusion in Project teams and delay in project implementation. Given that political support to Project Development Objectives is critical, timing the preparation and implementation of a new investment with the political cycle is of important to ensure a favorable political context. (b) Use of a variety of lending and non-lending instruments can enhance the implementation of critical sectoral reforms. The use of AAA, DPLs and Investment loans allows for widespread dialogue with various stakeholders leading to greater support to reforms. In particular, AAA products are critical in promoting consensus in the policy dialogue. In addition, when budget support or restructuring is a critical factor in a reform, the combination of the DPL, where the Ministry of Finance is the main counterpart, and the investment loan, where the implementing agency is responsible for the reform, ensure that all parties are involved in the process. 28

39 (c) Ensuring implementation readiness conditions prior to negotiations leads to a more effective and timely physical and financial execution cycle. The strategic design needs to be clear and have a detailed implementation plan in order launch the Project as soon as it is approved. (d) The use of a clear evidence-based causality model can strongly enhance the achievement of Project Development Objectives. Causality models are likely to guide and target cost-effective interventions. Along these models, the availability of monitoring data, as mentioned above, feeds the correction and targeting mechanism as needed. (e) Intercultural strategies are vital when beneficiaries belong to diverse indigenous groups. Demand-side interventions are critical to ensure appropriate use of services. In highly diverse countries, such as Peru, demand-side interventions need to address cultural barriers together with the economic and geographic barriers. Cultural barriers need to be well diagnosed and need to include language training when necessary. (f) Community ownership of project activities is likely to contribute to the sustainability of a health reform process through participatory maintenance of the goods provided and promotes decentralization under a context of ensuring accountability and equity. (g) Rehabilitation of health facilities needs to pay greater attention to overall utility capacity, in order to ensure that new facilities have sufficient availability of electricity, water and sanitation services. As the Project only financed maternal and child health services during the initial stage, rehabilitation and refurbishment designs did not include additional financing for increased capacity in basic utilities, thus resulting in overwhelmed facilities. Once this problem was detected, increased capacity of basic utilities were included in later rehabilitations. 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners (a) Borrower/implementing agencies The borrower provided comments linked to two main issues: Attribution of results to the Project intervention. As explained in section 3.2, this Project was part of a larger World Bank program that included a series of DPLs (PSRL I-IV) and a AAA. In addition, other cooperating agencies such as the European Union and Care have been implementing specific Projects linked to maternal and child health in some of the Project areas and have, surely, contributed also to results. Thus, the observed results are clearly a product of a long-standing policy on maternal and child mortality which has been supported by the World Bank, the IDB, and other agencies. Implementation of the Project monitoring and evaluation (M&E) system. The Project s M&E system was clearly innovative and influenced significantly the decision making process in the Project area. A particular attribute of the system was the fact that it used existing administrative systems to calculate indicators. Unfortunately, at the central level 29

40 this system was not as valued as in the regions. The main reason for this is the large number of existing monitoring systems at the central level, which are not built under the same platform and have different stakeholders. Under the follow-on Project, PARSALUD II, the Government is requesting support for the unification of the various administrative systems in order to develop a solid M&E system at the MINSA level. (b) Cofinanciers The IDB was the main cofinacier for this Project. They reviewed the document and had no major comments. (c) Other partners and stakeholders (e.g. NGOs/private sector/civil society) NA 30

41 Annex 1. Project Costs and Financing (a) Project Cost by Component (in USD Million equivalent) Components Appraisal Estimate (USD M) Actual/Latest Estimate (USD M) Percentage of Appraisal STRENGTHENING HEALTH DEMAND STRENGTHENING DECENTRALIZATION POLICY DEVELOPMENT AND INSTITUTIONAL MODERNIZATION IMPROVING QUALITY OF HEALTH PROGRAMS AND SERVICES PROJECT COORDINATION Total Baseline Cost Physical Contingencies 2.60 Price Contingencies 2.80 Total Project Costs Front-end fee IDB Front-end fee IBRD (b) Financing Source of Funds Total Financing Required Type of Cofinancing Appraisal Estimate (USD M) Actual/Latest Estimate (USD M) Percentage of Appraisal Borrower INTER-AMERICAN DEVELOPMENT BANK INTERNATIONAL BANK FOR RECONSTRUCTION AND DEVELOPMENT OPEC FUND

42 Annex 2. Outputs by Component Component A Strengthening Health Demand: Implementation of the SMI/SIS. This component had as its main objective to reduce the economic barrier through the implementation of the Seguro Materno Infantil (SMI), specifically through (a) financing of reimbursements, (b) support to the SMI administration, and (iii) monitoring and evaluation of the SMI. Prior to the signing of the loan, the GP requested both Banks to reduce the loan amounts in the amount of the SMI reimbursements and committed to finance the equivalent through the national treasury funds. In January 2002, the Borrower, in the context of Law 27657, integrated the SMI, which had been previously managed by PARSALUD, and the School Insurance Program under the newly enacted SIS. This legal and structural change was reflected in the second amendment to the LA approved by the Bank in May The autonomy status granted by the Law to SIS, initially caused difficulties in the relationship with MINSA, while roles were being defined. However, by mid SIS, MINSA and PARSALUD started working in a more coordinated fashion. At this time, the sector and MEF became aware of the relevance of SIS in the increase in the demand for health services, as can be observed in the indicators, and ensured the adequate allocation of treasury funds. The result was that the actual allocation was greater than originally planned in the Project design ($164 million versus $218 million). The Bank allocated about US$825,417 equivalent of loan proceeds to finance: (i) the administration of the SIS from November 2001 to October 2002; (ii) the provision of international and national qualified technical assistance to ensure SIS's institutional arrangements and the issuing of its internal operational regulations; (iii) the training of its nationally located staff and personnel in its 34 deconcentrated units in the country; (iv) the acquisition of computers and peripheral equipment for the entire SIS management information network in the country; (v) improvements in the SIS's supervision schemes; and (vi) the production of promotional brochures to increase the access of the most vulnerable eligible population. PARSALUD also contributed to the design and implementation of the SIS information system, which allowed SIS to correct the initial problems faced with the targeting of SIS. The SMI scheme increased its coverage from 5 DISAs in 2002 to the entire country when converted into the SIS scheme in The fact that SIS is now operating in all the DISAs implies compliance with one of the triggers agreed with the Bank to move to the second phase of the APL. The number of affiliates to the SIS scheme in its 5 programs increased from 5.8 million in 2002 to a little over 11 million in 2005 (89.7 percent increase). The SIS coverage of the Project's targeted population (program A - children 0-4 years old- and C pregnant women - of the SIS) increased from 2.55 million in 2002 to about 4.1 million in 2005 (60.8 percent increase). The budget allocated to SIS for its entire 5 programs increased from US$47.2 million equivalent in 2002 to US$77 million equivalent in 2005 (63.1 percent increase in nominal terms), while for the SIS's programs A and C the increased 32

43 went from US$4.4 million equivalent in 2000 to US$6.7 million equivalent in 2002 under the SMI scheme (52.3 percent increase in nominal terms), and then from US$27.6 million equivalent in 2002 to US$56.6 million equivalent in 2005 under the SIS figure (105.1 percent increase in nominal terms). Given the current macroeconomic conditions of Peru, The percentage of newborn children covered by the SIS weighted within the first 24 hours of birth in project areas during the period increased by 15.6 percent (from 64 to 74 percent), while at the national level the increase in the same period was 13.7 percent (from 73 to 83 percent). This result is still short form the target set at appraisal of 95 percent by the end of the first phase of the APL. The number of days elapsing between the submission request to SIS and the actual receipt of reimbursement by health facility, which is taken as a proxy of SIS's efficiency, decreased from about 180 days in 2002 to 91 days in This current delay is still short of the target set at appraisal of a maximum of 25 days interval, a figure that might likely result too optimistic against other international benchmarks associated to good practices. It is also important to ensure that SIS does not generate perverse effects with respect to the medical community favoring more expensive, but unnecessary protocols, to increase the reimbursement amount to their health facilities. An area for reflection for phase 2 of the APL is to assess the possibility of including in the SIS reimbursement package the transportation costs from the health facility to the patient's home which currently are excluded as eligible expenditures. The Project also financed the development and implementation of an intercultural strategy including two variants of the Quechua language. PARSALUD selected the universities of San Antonio Abad in Cusco and Huamanga in Ayacucho to provide an awareness training program for the staff of 74 health establishments in the project areas in order to remove the linguistic and cultural barriers in assisting beneficiary population belonging to these ethnic groups requiring maternal and peri-natal health services. These intercultural training programs are now being extended to other relevant medical and nursing schools in the country located in primarily populated ethnic groups. Components B and C. Strengthening decentralization, policy development and institutional modernization and improving the quality of health programs and services. As a consequence of the strategy modification agreed with the Bank in November 2003: (i) significant less consideration was provided by the Project to the decentralization and empowerment of the CLAS envisioned in the original design to produce and implement regional health plans; and (ii) a more integrated and holistic 1 The number of days elapsing from the submission of the request for reimbursement to SIS until the actual deposit of funds to the DISA bank account is 34 days. It takes almost two more additional months for the DISAs to allocate those funds to the corresponding health center. This last portion of the reimbursement circuit seems to be the bottleneck in need to be further addressed by GP to make the reimbursement process more efficient. 33

44 supply-side intervention of the original components B and C targeted to eight healthnetworks (DISAs) was carried out. The revised strategy, emerging from the maternal and peri-natal death causality model mentioned before, centered its supply-side intervention along the following three levels of interventions local, regional and national. At the local level, the Project financed the strengthening of 8 health-networks in the 8 participating DISAs comprised by 59 health facilities with BOF and 20 health facilities with EOF through civil works, equipment and furniture and training. The Project completed 188 works (84 against the Bank loan and 104 against the IDB loan). From the 84 works financed by the Bank, 17 were new constructions of complete maternal and/or peri-natal health facilities, 40 were new health facility rooms and 27 were rehabilitation of existing facilities. About 22 out of the 84 Bank-financed works were completed before the key November 2003 date and the rest afterwards. It is important to distinguish this time inflection point because, the works completed before November 2003 were carried out under a decentralized procurement process managed by the DISAs and considered of a significant lower quality and pertinence than the ones completed afterwards under a centralized procurement scheme. The increased pertinence of the works undertaken after November 2003 was underpinned by the causality model, while the increased quality of the works lies in the pre-feasibility requirements set by the SNIP from 2002 onwards before approving any civil work investment under the Project. The 79 health facilities received basic and essential obstetric equipment and furniture including, inter-alia: (a) equipment for childbirth at the health facility and at the homes of pregnant women (when, due to distance from the facility and/or cultural values, the delivery occurs at home under the care of a skilled staff); (b) equipment for maternal and peri-natal emergencies; (c) laboratory equipment; and (d) equipment for pre-natal and post-natal visits for pregnant women and their children. Six out of the 20 EOFs included in the Project received full equipment to undertake their assigned tasks according to the causality model protocols in use, while the rest of the EOFs received partial equipping. One weakness detected is the existence of a very fragile and incipient policy in the DISAs to ensure appropriate and timely corrective and preventive maintenance of these works and their corresponding equipment. This is an area where the second phase of the APL should insist vigorously. The staff assigned to the health facilities in the 8 networks received continuous training related to the medical protocols regulating the basic and essential obstetric functions to reduce maternal and peri-natal mortality rates. In addition, staff was recruited to these health facilities according to competency-based standards set by the above-mentioned protocols. At the regional level, the Project financed the provision of technical assistance to participating DISAs for the: (i) strengthening of the Management Information System for the Handling of Medical Supplies (Sistema de Gerenciamiento de Medicamentos SISMED) to ensure a three-month supply of basic and essential obstetric functions related medicines in all the health facilities belonging to the DISA; (ii) strengthening of the 34

45 DISA's Management and Logistical Information System (Sistema de Información Gerencial SIGA) to control non-medical related acquisitions, inventories, capital goods including DISA's managed real estate; (iii) training for DISA's staff; and (iv) establishment of Regional Performance-based Annual Health Management Agreements (Acuerdos de Gestión Regional) between the regional authority, the DISAs and MINSA. The yearly targets set in the performance-based agreements, providing clarity in the roles and responsibilities between DISAs and MINSA, as well as accountability mechanisms, are being monitored and updated periodically with the information produced by SIS, PARSALUD and MINSA. Finally, the Project contributed in the totality of DISAs in the country (in project and non-project areas) to integrate the mother and child health programs under their responsibility, which were functioning as vertical and uncoordinated activities. This achievement is part of another trigger to move into the second phase of the APL. At the national level, the Project financed the: (a) issuing of national regulations referred to the maternal and peri-natal protocols, the functioning of the networks and the profile standards of health facilities' staff; (b) the provision of technical assistance for the strengthening of MINSA's Management Information System for the acquisition of medical supplies; and (c) provision of technical assistance for the strengthening of the MINSA's Epidemiology Information system where a Monitoring and Evaluation Unit has been established. However, the Project felt short in supporting MINSA to implement an integral health approach (transversal services according to risk groups children, youngsters, pregnant women and adults) by diminishing the fragmentation of the different vertical services currently administered by the various Directorates within the Health Ministry and functioning in an uncoupled fashion. With the change of strategy in November 2003, the Project disengaged its control over the process to achieve one of the agreed triggers to start the second phase of the APL, namely the one refereed to the percentage of health posts and health centers administered by the CLAS (the target set at 30 to 33 percent). This trigger was met under the PSRL III benchmark. After a lengthy delay in entering an agreement with INEI, the Project was instrumental in 2002 in having ENAHO include in its annual survey relevant information concerning maternal and peri-natal health. PARSALUD was also influential in having the health demographic survey (ENDES) be carried out on an annual basis instead of every five years. ENDES was not financed by PARSALUD, rather receiving significant support from the United States Agency for International Development. For reasons beyond the control of PARSALUD, DFID ended up not financing the annual community surveys included in the original design of Component B. The total amount disbursed by the loan on these two components amounted to US$ million equivalent. 35

46 Component D. Project Coordination. The ICR highlights the PCU's role in ensuring: (i) the functioning of a monitoring and evaluation unit capable of updating the Project's five relevant indicators mentionedbefore on a quarterly basis in each one of the 8 participating DISAs; and (b) compliance of the Project's physical and financial execution with respect to the implementation scheduled agreed for the period at the time of the first extension to the closing date. Accordingly, and despite the unsatisfactory performance of the PCU prior to November 2003, this Component is considered satisfactory. The total amount disbursed by the loan on this component amounted to US$2.25 million equivalent. 36

47 Annex 3. Economic and Financial Analysis (including assumptions in the analysis) The Health Reform Project showed significant gains in effectiveness, efficiency and equity of public spending on health during the Project. The results described below show important gains in: (a) better use of public resources through programs which have propoor expenditures; (b) significant increase in the number of beneficiaries of these programs; (c) considerable improvements in institutional birth coverage and reduction in infant mortality rates; (d) low fiscal impacts in comparison to the national budget allocated to the public health sector. Impact of SIS on beneficiary population. Coverage of SIS (Integrated Insurance System "Seguro Integral de Salud"), with respect to the plans C and A (pregnant women and children less than five years of age), has increased significantly (See Table A3-1 and Figure A3-1). While the figures in the PAD were considered somewhat ambitious, the targets were surpassed and sustained over time, with full financing by treasury funds. Table A3-1. Beneficiary Population of the SMI/SIS (Thousands) Year Appraisal Estimates Children Less Than 5 Years of Age Actual/Latest Estimates Children Less Than 5 Years of Age Pregnant Women Pregnant Women , , , , ,444 Source: Project Appraisal Document and SIS central database 37

48 Figure A3-1 Beneficiary population of SMI/SIS ( ) 4,500,000 4,000,000 Beneficiary population 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000, , Year Appraisal Estimates (P.women and children less than 5) Trend Actual/Latest Estimates (P.women and children less than 5) Source : Project Appraisal Document and SIS central database. SIS had a significant impact not only in coverage but in improving the efficiency of existing health sector resources by increasing utilization of public facilities nationwide. While public facilities were financed in eight regions of the country to improve the supply and quality of health services through investment support, there was no significant expansion in human resources. Thus, the elimination of the economic barrier, through SIS, resulted in an increase in service utilization. The Project also contributed to gains in efficiency and progress towards equity through the better allocation of the SIS resources. SIS was designed to cover the poorest areas of the country, rural and urban marginal zones. The intervention of the Project, with support of the policy dialogue through the PSRL, helped to increase the targeting of SIS and the resulting increase in public resources directed to the provision of essential health services to the poor. In the case of institutional deliveries, the percentage of women in the first quintiles went from 44 percent to 56 percent (see Figure A3-2). Furthermore, in the Project Area, the targeting to the poorest quintile for Plan C improved significantly, as shown in Figure A3-3. Finally, 60% of women who benefited from SIS live in rural areas (where the incidence of maternal mortality is highest) 1. Coverage in the rural area is around 85%, significantly larger than the coverage in the urban area (see Figure A3-4). 1 Source: PARSALUD I final evaluation 38

49 Figure A3-2 Percentage of deliveries financed by SIS in the first two quintiles of population in the Project area 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Year Rest of population First tw o quintiles Source: PIU PARSALUD I Figure A3-3 Beneficiary population of Plan C (pregnant women) by quintile of poverty in the Project Area Quintile 1 Quintile Source: SIS 39

50 Figure A3-4 Percentage of coverage by the Plan C (pregnant women) nationwide 100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% SMI SIS Rural Urban Source: ENDES At the national level, the targeting of SIS has improved significantly. As shown in Figure A3-5 significant progress in reducing the population of poor women not covered by Plan C (pregnant women) from 85%, at the beginning of the SMI/SIS program, to 29% by the end of In addition, there has been a significant reduction in the filtration of SIS resources to the non poor, from 13% in 2000 to 6% in This reduction represents savings of around a US$ 25 million for the year 2005 only. Figure A3-5 Beneficiary Population of the SIS Plan C Percentage Year Non beneficiary poor population Coverage of non poor population Source: Final document of the PARSALUD I evaluation 40

51 Overall Project impact. In order to assess effectiveness and impact of the Project as a whole, it is convenient to look at the improvement of the development indicators in the Project areas as compared to the rest of the nation 2. Project areas showed a steeper increase in institutional deliveries as compared to the overall national trend. For example, while the percentage of deliveries attended by a skilled professional in 2006 improved 30 percent nationwide as compared to 2000, the change in the Project area reached 84 percent, that is 2.8 times more (see Results Framework). As shown in Figure A3-4, it is possible to attribute the increase in the indicator to the Project intervention, otherwise, if the Project areas had maintained the national growth rate, it is probable that this indicator would not have improved as much (Figure A3-6). Finally, the impact observed in the infant mortality rate is remarkable. Again, the changes in the intervention area were greater than the national average, 41 percent and 32 percent, respectively (1.3 times). In this case, if this indicator had kept the national rhythm of growth in the project area, it would have achieved smaller reductions in the indicator (Figure A3-7). Figure A3-6 Percentage of deliveries attended by a skilled health professional Percentage Project area National trend Year Source: PIU PARSALUD I 2 It was not possible to compare Regions subject to the Project interventions with similar Regions that were not covered by the Project due to difficulties in finding appropriate matches. Therefore, the analysis compares Project area versus national data. This will tend to underestimate the Project impact as national data include the Project area (which has shown significant increases) and Lima, which is well covered by health services. 41

52 Figure A3-7 Number of deaths before 12 months per 1,000 living births Number of deaths Project area National trend Year Source: PIU PARSALUD I The results also display improvements in terms of poverty incidence in the Project Area. As it shows in the Figure A3-8, there has been a sustainable increase in the number of deliveries attended by a skilled health professional in the poorest quintile since the beginning of During 2005, the achieved results in this population group were greater than the results for the richest quintile, an important change with respect to The original gaps represented inequalities mainly produced by the economic barriers to utilization of health services for the poor population groups. Figure A3-8 Number of deliveries attended by a skilled health professional in the Project Area by quintile of poverty quintil Q1 Q2 Q A smaller gap Source: SIS 42

53 There are similar achievements in the rural areas. The increase in the coverage of deliveries attended by skilled professionals has been 25% larger than the observed increase in the urban areas since 1998 (see Figure A3-9). Figure A3-9 Coverage of deliveries attended by a skilled health professional nationwide 100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% SMI SIS 20 Urban area 25 Rural area Source: PIU PARSALUD I There were also differences within the rural areas (see Figure A3-10). The trends describe a faster growth of this indicator in the Project Area (the poorest regions). While the coverage in the non project area was greater, the gap was significantly reduced towards the end of

54 Figure A3-10 Coverage of deliveries attended by a skilled health professional in Rural Area 50% 45% 40% 35% 30% 25% SMI SIS 20% Non project area 15% 10% 5% 0% Project area Source: PIU PARSALUD I It is possible estimate the economic benefits using the figures on infant mortality. Forecasting the situation with and without the Project, with the Project there are 42,883 additional births attended by a skilled professional and, because of the smaller infant mortality rates, about 4,318 infant deaths avoided from an overall of 568,502 births between 2001 and , which represents 23 percent of the total births. In sum, there were effectiveness gains in the project. As a result, the project has a net present value of almost US$ 358, Table A3-2 shows an estimate of the fiscal impact of the project in terms of its annual costs and budget of the Public Health Insurance SIS for the Plans A and C over the period The table also compares these expenditures with the MINSA's budget and total expenditures in health services provided by the regional governments. It can be 3 We have considered the period between 2000 and 2004 because the calculations used the ENDES data to determine some of their assumptions 4 The assumptions taken were: 1) In order to value a healthy life, it was used the GDP per capita in 2004, which is US$ ) The situation without the project was projected between 2001 and 2004 with a linear regression. 3) The situation with the project between 2001 and 2004 was a linear projection with the current values of ENDES 2000 and ) Since we have the trend of the expenditures between 2002 and 2004 from the PIU-PARSALUD, the umber of deaths avoided for 2001 was evenly distributed using the evolution of the expenditures. 44

55 seen that the project represented, on average, only about 0.2 percent of the public health expenditures and about 4.8 percent of the SIS Plans A and C. Table A3-2 Fiscal Impact of the Project Project Cost 1/ 4,124,999 4,345,563 7,362,362 15,125,501 SIS Plan A and C 2/ 15,500,754 23,468,892 97,215, ,092, ,113, ,688,158 Public Health Expenditure (PHE) 3/ 2,643,584,510 2,994,356,220 3,196,841,544 3,260,168,856 3,774,526,750 3,982,882,688 % Project Cost / PHE 0.13% 0.13% 0.20% 0.38% % Project Cost / SIS Plan A and C 4.20% 3.10% 3.60% 8.20% Source: 1/ PIU PARSALUD I 2/ SIS 3/ SIAF-MEF Sustainability of the supply component of the Project includes the maintenance of the infrastructure and equipment. A very low fiscal impact would be expected, given the low percentage cost of the Project with respect to the health sector budget, which shows a historically positive trend. 45

56 Annex 4. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members Names Title Unit Responsibility/Specialty Lending Supervision/ICR Patricia M. Bernedo Senior Program Assistant LCSHH Jose Pablo Gomez-Meza Sr Economist (Health) LCSHH Pablo Augusto Lavado Junior Professional Associate LCSHS-DPT Luisa Maria Yesquen Solari Program Assistant LCSHH (b) Ratings of Project Performance in ISRs No. Date ISR Archived DO IP Actual Disbursements (USD M) 1 06/27/2000 Satisfactory Satisfactory /30/2000 Satisfactory Satisfactory /19/2001 Satisfactory Satisfactory /28/2001 Satisfactory Unsatisfactory /19/2001 Satisfactory Unsatisfactory /03/2002 Satisfactory Satisfactory /10/2002 Satisfactory Satisfactory /11/2002 Satisfactory Satisfactory /10/2003 Satisfactory Satisfactory /11/2003 Satisfactory Satisfactory /05/2003 Satisfactory Satisfactory /14/2004 Satisfactory Satisfactory /29/2004 Satisfactory Satisfactory /26/2005 Satisfactory Satisfactory /01/2005 Satisfactory Satisfactory /28/2006 Satisfactory Satisfactory c) Staff Time and Cost Lending Stage of Project Cycle Staff Time and Cost (Bank Budget Only) USD Thousands No. of staff weeks (including travel and consultant costs) FY FY FY FY

57 Supervision/ICR FY FY FY FY FY Total: FY FY FY FY FY FY FY FY FY Total:

58 Annex 5. Beneficiary Survey Results (if any) NA 48

59 Annex 6. Stakeholder Workshop Report and Results (if any) 49

60 Annex 7. Summary of Borrower's ICR and/or Comments on Draft ICR 50

61 51

62 52

63 53

64 54

65 55