A W O R L D B A N K C O U N T R Y S T U D Y 47110 Zambia Health Sector Public Expenditure Review Accounting for Resources to Improve Effective Service Coverage THE WORLD BANK WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page i A W O R L D B A N K C O U N T R Y S T U D Y Zambia Health Sector Public Expenditure Review Accounting for Resources to Improve Effective Service Coverage Oscar F. Picazo Feng Zhao THE WORLD BANK Washington, D.C. WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page ii Copyright © 2009 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First Printing: December 2008 printed on recycled paper 1 2 3 4 5 12 11 10 09 World Bank Country Studies are among the many reports originally prepared for internal use as part of the continuing analysis by the Bank of the economic and related conditions of its devel- oping member countries and to facilitate its dialogs with the governments. Some of the reports are published in this series with the least possible delay for the use of governments, and the aca- demic, business, financial, and development communities. The manuscript of this paper there- fore has not been prepared in accordance with the procedures appropriate to formally-edited texts. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/TheWorld Bank and its affiliated organizations,or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission promptly to reproduce portions of the work. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, Tel: 978-750-8400, Fax: 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, Fax: 202-522-2422, email: pubrights@worldbank.org. ISBN-13: 978-0-8213-7804-5 eISBN: 978-0-8213-7805-2 ISSN: 0253-2123 DOI: 10.1596/978-0-8213-7804-5 Cover photo: Stone scupture of two women helping each other, photo by Oscar F. Picazo. Library of Congress Cataloging-in-Publication Data Picazo, Oscar F. The Zambia health sector public expenditure review : accounting for resources to improve effective service coverage / Oscar F. Picazo and Feng Zhao. p. ; cm. Includes bibliographical references. ISBN 978-0-8213-7804-5 -- ISBN 978-0-8213-7805-2 1. Medical economics--Zambia. 2. Zambia--Statistics, Medical. I. Zhao, Feng, 1967- II. Title. [DNLM: 1. Health Expenditures--statistics & numerical data--Zambia. 2. Health Services-- economics--Zambia--Statistics. 3. Public Health Practice--economics--Zambia--Statistics. W 16 P586z 2008] RA410.55.Z33P53 2008 338.4'73621096894--dc22 2008046752 WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page iii Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Acronyms and Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Executive Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Key Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Key Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi 1. Background, Objectives, and Analytical Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Impelling Factors for the Health Sector PER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Objectives and Analytical Tools of the Health Sector PER . . . . . . . . . . . . . . . . . . . . . 2 Uses of the PER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Health System Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Health System Goals and Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Challenges at the Health Outcome Level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Challenges at Health Service Level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Challenges at the Service Delivery Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3. Sources and Uses of Funds in the Health Sector: Results of the National Health Accounts Exercises, 1999­2004 . . . . . . . . . . . . . . . . . . . . . . 17 Total and Per Capita Health Expenditures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Sources of Health Expenditures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Health Expenditures by Service Provision and Administration . . . . . . . . . . . . . . . . 24 Health Expenditure Patterns of Donors, MOH, and Households. . . . . . . . . . . . . . . 26 Financing of Service Providers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Uses of Health Expenditures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4. Budget Allocation, Release, and Spending: Results of the Public Expenditure Tracking Component of the PET/QSDS, 2005­06. . . . . . . . . . . . . . . 33 Funds Flow and Resource Envelope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 MOH Budget Process and Allocation Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 MOH Allocations versus Release of Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Allocation and Releases of District Grants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Health Facilities' Management of Vertical Funds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Health Facilities' Management of Internally-generated Funds . . . . . . . . . . . . . . . . . 50 Overall Resource Availability at the Facility Level . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 iii WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page iv iv Contents 5. The State of Health Facilities and Services: Results of the Quality of Service Delivery Component of the PET/QSDS, 2005­06 . . . . . . . . . . . . . . . . . 57 Management of Infrastructure, Utilities, and Equipment . . . . . . . . . . . . . . . . . . . . . 57 Management of Health Personnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Management of Drugs and Other Medical Supplies . . . . . . . . . . . . . . . . . . . . . . . . . 67 Clinic and Patient Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 6. Scaling Up Health Service Delivery to Reach the MDGs: Results of the Marginal Budgeting for Bottlenecks Modeling Exercise, 2006 . . . . . . . . . . 69 Health Sector Response to the Challenges: Strategy and Resources . . . . . . . . . . . . . 69 Bottlenecks in the Service Delivery Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Scenarios for Scaling Up Service Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Statistical Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 LIST OF TABLES 1. Health Facilities in Zambia, 2000s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2. Summary Statistics on the Major Diseases, 2000­05. . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3. Selected Health Service Delivery Indicators, 2000­04. . . . . . . . . . . . . . . . . . . . . . . . . 11 4. Service Coverage Indicators of Key Childhood Health Interventions in Selected African Countries, 2000s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 5. Maternal Mortality Ratio and Service Coverage Indicators of Key Maternal Health Interventions in Selected African Countries, 2000s . . . . . . . . . . . . 14 6. Government Health Expenditures (ZK Billion) Using Alternative Definitions, 1995­2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 7. Health Expenditures (ZK Billion) Under the Discretion of the Government, by Source, 2001­07. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 8. Trends in Health Expenditures by Source, 1995­2004 . . . . . . . . . . . . . . . . . . . . . . . . 22 9. Total Financing of Health Facilities, 2002­04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 10. Health Sector Resource Envelope, FY05 and FY06 . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 11. Range of Cost Estimates to Meet the Health MDGs in Zambia . . . . . . . . . . . . . . . . . 36 12. Amount of GRZ Budget Provided by DHMTs to Their Health Facilities, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 13. Health Centers' Reponses on Whether They Received an Amount More Than, Equal to, or Less Than Their Annual Allocations, 2006 . . . . . . . . . . . . . . . . . . 46 14. Uses of GRZ Budget and Basket Funds by Health Facilities, FY05. . . . . . . . . . . . . . . 48 15. Composition of Average Debt Per Facility, FY05. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 16. Donors' Vertical Funds Received by Health Facilities, FY05 . . . . . . . . . . . . . . . . . . . . 49 WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page v Contents v 17. Uses of Vertical Funds by Health Facilities, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 18. Internally Generated Funds of Health Facilities, 2005 . . . . . . . . . . . . . . . . . . . . . . . . 51 19. User Fee Practices, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 20. Percent of Health Facilities' Rate of Retained Fee Revenues, FY05. . . . . . . . . . . . . . 53 21. Uses of User Fee Revenues by Health Facilities, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . 53 22. Expected Versus Actual User Fee Revenues for OPD, FY05 . . . . . . . . . . . . . . . . . . . . 54 23. Level of Annual Resources at the Facility Level, by Major Sources, FY05 . . . . . . . . . 54 24. Health Facilities Reporting Inadequate Medical Equipment, Instruments, and Lab Test Supplies, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 25. Vacancy Rates in Health Facilities, by Cadre, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 26. Established Posts and Average Number of Actual Staff by Major Occupational Groups, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 27. Salary Management, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Annex A. MOFNP and Local Government Authorities' Health Expenditures by Recipient, 1995­2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Annex B. Donors' Health Expenditures by Recipient, 1995­2004 . . . . . . . . . . . . . . . . . . 80 Annex C. Employers' and Households' Health Expenditures by Recipients, 1995­2009. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Annex D. Total Health Expenditures by Service Provision, Administration, and Rest of the World Transactions, 1995­2004 . . . . . . . . . . . . . . . . . . . . . . . . 81 Annex E. Total Health Expenditures by Service Provision and Administration and by Type of Providers, 1999­2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Annex F. Donor Health Expenditures by Type of Providers and Administrators, 2002­04. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Annex G. MOH/CBOH Health Expenditures by Level of Care and Administration, 1999­2004. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Annex H. Household Health Expenditures by Type of Providers, 1999­2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Annex I. Health Expenditures by Line Items by Type of Facility, 1999­2004 . . . . . . . . 86 Annex J. Resources at the Disposal of a Typical DHMT, FY05 . . . . . . . . . . . . . . . . . . . . 87 Annex K. Self-Reported Versus Actual Capacity to Deliver Selected Health Services in Rural Health Centers, 2006. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Annex L. Self-Reported Versus Actual Capacity to Deliver Selected Health Services in Urban Health Centers, 2006. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 LIST OF FIGURES 1. Comparison of IMR and U5MR in Selected African Countries, Latest Year Available . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2. Trends in Neonatal, Infant, and Under-Five Mortality in Zambia, 1992, 1996, and 2001/02 Rates (Per 1,000 Live Births) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page vi vi Contents 3. Trends and Levels of Major Diseases in Zambia, 2002­04 . . . . . . . . . . . . . . . . . . . . . 10 4. Coverage of Key Health Services, 2000­04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 5. Provincial Inequities in Health Service Coverage by Region . . . . . . . . . . . . . . . . . . . 12 6. Differences in Selected Maternal and Child Health Indicators Between Rural and Urban Households in Zambia, 2000/01 . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 7. Differences in Selected Maternal and Child Health Indicators Between Lowest and Highest Asset Quintile Households in Zambia, 2000/01 . . . . . . . . . . . . 13 8. Unfilled Posts by Profession. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 9. Gaps in Numbers of Health Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 10. Trends of Child Mortality in Zambia, 1990­2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 11. Trend of Maternal Mortality in Zambia, 1996­2015. . . . . . . . . . . . . . . . . . . . . . . . . . 16 12. Total Health Expenditures in Nominal and Real Terms, 1995­2004. . . . . . . . . . . . . 18 13. Selected Health Expenditure Ratios, 1995­2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 14. Trends in Per Capita Government Health Expenditure and Per Capita Total Health Expenditure, 1995­2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 15. Trends in Health Expenditures by Source, 1995­2004 . . . . . . . . . . . . . . . . . . . . . . . . 23 16. In-Budget versus Off-Budget Donor Health Expenditures, 2000­04 . . . . . . . . . . . . 24 17. Total Health Expenditures Going to Service Provision and Administration, 2001­2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 18. MOH/CBOH Health Expenditures Going to Service Provision and Administration, 2001­04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 19. MOH/CBOH Health Expenditures by Level of Care, 1999­2004 . . . . . . . . . . . . . . . 28 20. Financing of Health Facilities by Source, 2002­04 . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 21. Health Expenditures of MOH, Mission, and For-Profit Facilities by Inputs, 1999­2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 22. Flow of Funds and Other Resources in Zambia's Government Health Sector . . . . . 34 23. Per Capita MOH Released + Supplementary Funding, by Province, FY05 . . . . . . . 38 24. Budget Release Per Capita Versus Incidence of Malaria and Under-5 Pneumonia, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 25. Cumulative Budgeted Versus Released Amounts of Total MOH Budget, by Month, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 26. Cumulative Budgeted Versus Released Amounts for Personal Emoluments, by Month, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 27. Cumulative Budgeted Versus Released Amounts for District Grants, by Month, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 28. Cumulative Budgeted Versus Released Amounts for RDCs, by Month, 2005 . . . . . . 42 29. Cumulative Budgeted Versus Released Amounts for Essential Drugs, by Month, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page vii Contents vii 30. Cumulative Budgeted Versus Released Amounts for Capital Expenditures, by Month, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 31. MOH Allocation Versus Releases to Provinces Included in the Survey, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 32. MOH Allocation Versus Releases to DHMTs Included in the Survey, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 33. Cumulative Planned Versus Actual Facility Grants by DHMTs to RHCs, by Month, FY05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 34. Cumulative Planned Versus Actual Hospital Grants, by Month, FY05 . . . . . . . . . . . 45 35. Allocated Versus Received Budgets of Selected District Hospitals FY05 . . . . . . . . . . 47 36. Allocated Versus Received Budgets of Selected Health Centers, FY05 . . . . . . . . . . . . 47 37. Health Facilities With Functional and Nonfunctional Utilities and Transport Equipment, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 38. Health Facilities with Imaging Equipment, 2006. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 39. Health Facilities with Functional and Nonfunctional Medical Equipment, 2006 . . . . 59 40. Personnel Expenditures in MOH Budget and Share of Personnel Expenditures to Total MOH Expenditures, 2000­07. . . . . . . . . . . . . . . . . . . . . . . . . . 61 41. Average Number of Hours Worked in a Week by Type of Task, 2006 . . . . . . . . . . . . 62 42. Per Capita Health Expenditures and Share of Personal Emoluments to Health Expenditures, 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 43. Summary Rates of Staff Vacancy, Absenteeism, Tardiness, Involvement in Income-Augmenting Activities, and Dissatisfaction, 2006. . . . . . . . . . . . . . . . . . . 64 44. Composite Monthly Salaries and Allowances of Clinical/Professional Health Workers, 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 45. Drug Expenditures and Drug Kits Opened Per 1,000 Patients, 1999­2005. . . . . . . . 67 46. Sequential Relationships of Health Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 47. Bottlenecks in Community-Based Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 48. Bottlenecks in Population-Based Outreach Services . . . . . . . . . . . . . . . . . . . . . . . . . . 73 49. Bottlenecks in Facility-Based Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 50. Projected Trend in Reaching Child Mortality MDGs . . . . . . . . . . . . . . . . . . . . . . . . . 77 51. Projected Trend in Reaching Maternal Mortality Reduction MDGs . . . . . . . . . . . . . 78 WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page viii WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page ix Acknowledgments he Zambia Health Sector Public Expenditure Review (PER) was written by Oscar T Picazo (senior economist, AFTH1) with additional inputs provided by Feng Zhao (senior public health specialist, AFTH1). The report consists of three analytical pieces. First, the National Health Accounts (NHA) was done separately by a team from the Ministry of Health (MOH) and the University of Zambia (UNZA) and with funding pro- vided by the Swedish International Development Agency (SIDA). Second, the Public Expenditure Tracking and Quality of Service Delivery Survey (PET/QSDS) was task-managed by Oscar Picazo.It was funded through a US$208,000 Insti- tutional Development Fund (IDF) grant of the World Bank to the Ministry of Health (MOH) of the Government of the Republic of Zambia. SIDA provided additional funding by reprogramming its funds under the Institutional Collaboration Health Policy Analysis and Health Economics Project (HEP) with the University of Zambia (UNZA) Economics Department, allowing the recruitment and fielding of additional enumerators. The PET/ QSDS report was prepared by a team from the Ministry of Health (MOH) of the Govern- ment of the Republic of Zambia (GRZ),the University of Zambia (UNZA),the World Bank, and two consultants who were involved in various stages of the work. The staff who partic- ipated in this effort were: Solomon Kagulura and Henry Kansembe of the Ministry of Health (MOH); Dale Mudenda, Caesar Cheelo, Rutendo Chitembure, and Stanley Banda of the University of Zambia (UNZA); and Catharina Hjortsberg and Felix Masiye, consultants. Third, the work on the Marginal Budgeting for Bottlenecks (MBB) model was led by Feng Zhao. This activity was funded mainly by the World Bank and UNICEF, and was car- ried out with an MBB Team from MOH, World Bank, and UNICEF. Acknowledgement is extended to other staff at GRZ and the World Bank who pro- vided technical or administrative support during the course of the PER. These are: Davies Chimfwembe and Felix Phiri of MOH; Justin Mubanga, Macroeconomic Policy Unit, and Priscilla Musole, Budget Office, of the Ministry of Finance and Development Planning (MOFDP); and Muhammad Pate, Rosemary Sunkutu, Helen Taddese, Monique Vledder, Neta Walima, and Christopher Walker (World Bank). The report was peer reviewed by Agnes Soucat, lead economist,AFTHD; David Peters, senior health specialist, HDHNE. Early comments on the approach was also provided by Laura Rose, senior economist, AFTH2. ix WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page x WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xi Acronyms and Abbreviations Capex Capital Expenditures CBO Community Based Organization CBOH Central Board of Health CSO Census and Statistics Office CHAZ Churches Health Association of Zambia DHB District Health Board DHE Donor Health Expenditure DHMT District Health Management Team DHS Demographic and Health Survey GDP Gross Domestic Product GF Global Fund GHE Government Health Expenditure GRZ Government of the Republic of Zambia HE Health Expenditure HR Human Resources IGF Internally Generated Fund IMR Infant Mortality Rate LGA Local Government Authority MBB Marginal Budgeting for Bottleneck Model MDG Millennium Development Goals MMR Maternal Mortality Ratio MOFNP Ministry of Finance and National Planning MOH Ministry of Health MTEF Medium Term Expenditure Framework NGO Nongovernmental Organization NHA National Health Accounts ORE Other Recurrent Expenditures ORT Oral Rehydration Therapy PER Public Expenditure Review PET Public Expenditure Tracking PHO Provincial Health Office QSDS Quality of Service Delivery Survey RHC Rural Health Center SWAp Sectorwide Approach TB Tuberculosis THE Total Health Expenditure U5 Under Five (Children) UHC Urban Health Center UNZA University of Zambia WB World Bank ZCCM Zambia Consolidated Copper Mines ZK Kwacha (Zambian currency) xi WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xii xii Acronyms and Abbreviations Currency Equivalents Currency Unit = Zambian kwacha US$1.00 = 4,353 as of April 26, 2007 Fiscal Year January 1 to December 31 Vice President Obiageli Katryn Ezekwesili Country Director: Michael Baxter Sector Manager: Christopher Thomas Task Team Leader: Oscar F. Picazo WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xiii Executive Summary Key Findings The longitudinal analysis of health expenditures using National Health Accounts (NHA) revealed a number of observations, some already known but provided in this report with statistical backing: I Donors: Donorfinancinghasrapidlyrisen,muchof itisincreasinglyoff-budget,and an increasing proportion of the expenditures is being devoted to administration rather than service provision. Donors' inability to fund human resource costs has led to "so much money chasing so few workers." Moreover, research activities are crowding out training expenditures. Finally, the re-emergence of vertical financing, which is often not fungible, means that the priorities of global vertical disease sponsors often supersede national priorities and needs. I Government of the Republic of Zambia (GRZ): Government contribution to overall health financing is slackening while its discretionary financing is increasing as basket funds and GRZ-managed projects have grown. The "administrative-intensity" of GRZ health expenditures has also increased, probably due to the costly separation of MOH and CBOH. Finally, there has been a dramatic increase in district allocations. I Households: Household health spending is larger than expected,used mainly in the private for-profit sector and inordinately for traditional healers. However, households' role in the financing of mission facilities and GRZ primary services is small due to low or non-existent fees at these levels. I Costanalysis: MOHfacilitieshavehigherproportionoflaborexpenditures,lowerpro- portion of drug expenditures, and lower proportion of capital expenditures relative to their for-profit counterparts. I Per capita total health spending: Zambia has higher per capita total health expendi- ture (THE) than its neighbors (Malawi, Tanzania, Mozambique) and its per capita health spending is closer to the middle-income Southern African countries. Yet, its health indicators are not any better and sometimes even worse than African IDA coun- tries. One could hypothesize that Zambia's NHA, having evolved longer, are able to capture much more comprehensively national health spending better than its neigh- bors, and this may "statistically increase" health spending. The public expenditure tracking (PET) component of the Quality of Service Delivery Survey (QSDS) showed the following: I Inter-sectoralallocation: Despiteseveraleffortstoestimatetheresourcerequirements of providing a basic package of care, little success has been achieved in dramatically increasing GRZ allocation to health. The Ministry of Finance and National Planning (MOFNP) and the Ministry of Health (MOH) do not yet have a common understand- ing of what the sector needs, nor on what proportion of the budget or of GDP the sec- tor should get. xiii WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xiv xiv Executive Summary I AllocationwithinMOH: Resourceallocationwithinthesectorisinformedbyseveral sometimes-contradictory principles. Careful resource planning to inform resource requests is not universally undertaken by District Health Mangement Teams (DHMTs). While allocation to the districts is equity-enhancing, allocation to hospitals follows principles of historical budgeting that tend to perpetuate installed capacity, even if much of that existing capacity was heavily influenced by mining rather than public- health concerns. The curative/facility-based orientation is reinforced by hospitals tend- ing to attract vertical resources far more than health centers. Only 28 of the 3,000-or-so needed health posts exist, and therefore even the district health budget is still largely oriented to higher levels of care within the district. I Geographic allocation: The above allocation rules manifest themselves clearly in the highly inequitable provincial receipt of resources, where the poorest, most remote, and least urbanized provinces receive the lowest per capita MOH releases. I Fund releases from MOH: In contrast to previous years, the cash budgeting system worked well in FY05, resulting in timely release of funds. Releases exceeded budgeted allocations in all districts and provinces included in the PET/QSDS. The key implica- tion here is the National Government's fiscal discipline; loss of such discipline trans- lates into budgetary squeeze that harms social-sector ministries. I FundreleasesfromDHMTsandpossible"localcapture"ofresources: Morethanathird of the DHMTs themselves admitted delays in releasing district grants to facilities, although this is difficult to understand given that most of them received these resources from MOH on time. In addition to the problem of delay, a fifth of the health facilities received resources less than their intended allocations. The imprest system, upon which 87 percent of health facilities depend, is also prone to delays in releasing resources to health facilities. One can surmise of the possibility of "local capture" or control of resources by influential members of the DHMTs--a phenomenon that has been observed in other countries (such as Uganda)--but establishing and quantifying the magnitude of this problem would require additional analysis of the data. I Indebtednessof healthfacilities: Aquarterof thehealthfacilitiesareindebted(mostly for utilities and transport fuel), and often rely on the DHMT for the relief of these debts. The persistence of debts is a problem. I Impact of fee abolition: How health facilities will find alternative sources of income to replace these resources is a critical issue, especially Urban Health Centers (UHCs) that are highly reliant on fees. Because these revenues are used mainly as worker incen- tives, fee abolition could reduce staff morale, unless some other incentives could take their place. The quality of service delivery component of the PET/QSDS depicted the following picture of the health delivery system: I Infrastructure, utilities and equipment: Most of the physical infrastructure is old, although staff are keeping them in good condition. Serious deficits exist in varying degrees across health facilities in utilities, communications, transport, and medical equipment. High rates of nonfunctional non-medical equipment were observed, espe- cially for transport and communications, but medical equipment appear to have lower rates of downtime. Large demand exists for medical instruments, especially in rural health clinics. WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xv Executive Summary xv I Human resources: High staff vacancies (33.5 percent) especially of clinical workers (41.4 percent) are crippling facility operations. Facilities are increasingly relying on expatriate staff (especially hospitals) and volunteer staff (especially health centers).Lack of staff is exacerbated by high rates of absenteeism (21 percent self-reported) and tar- diness (43 percent self-reported), reducing the amount full-time equivalent (FTE) workers. As a result, patient queues and waiting times are long (average: 65 minutes). More seriously, the length of time being spent on patient care is being compromised. Workers are also resorting to various coping mechanisms (both inside and outside the facilities where they work) to augment their incomes. Heavy official workloads and income-augmenting activities take their toll on workers' time, and possibly well-being, contributing to high rates (44 percent of staff) of dissatisfaction among staff. I Human-resource and salary management: Salaries are highly compressed, reducing their effectiveness in providing incentives. The plethora of cash and non-cash benefits that government has provided to ease the incentive problem has become unwieldy; it also makes it ever more difficult to forecast the fiscal implications. The incentive effect of salaries is further dampened by a number of salary management problems that afflict a not-insignificant number of staff: delay in the receipt of salaries (22 percent of staff); non-receipt of the full amount of salaries (about 15 percent of staff); unauthorized salary deductions (15.5 percent of staff); and staff payment of "expediter's fee"to obtain salaries (10 percent of staff). I Drugs and other consumables: The upward trend in drug financing that made drugs relatively more available in health facilities in recent years was reversed in 2006, as the drug budget was slashed by 15 percent. More than a third of the district surveyed expe- rienced delays in the receipt of drug kits. There is statistical evidence of drug diversion. Essential and life saving drugs continue to be widely unavailable. About 55 percent of hospitals and 46 percent of rural health centers reported expired drugs. Inappropriate drugs were less of a problem, although they were still found in 14 percent of facilities. I Facilitymanagement: Afunctionalandregularsystemoffacilitysupervisionexists,but the accuracy of patient registers remains an issue. Only two-thirds of health facilities have accurate registers. I Patient perceptions: Despite the supply deficits in health facilities,they receive mod- erate patient rating of overall quality: 85 percent.A closer scrutiny of this rating reveals that although health facilities and staff receive 89­95 percent approval rating on "hos- pitality aspects" of care (friendliness, provision of information), their patient ratings on"technical"aspects of care (explaining what the procedure or drug was for) was only 40­42 percent. The Marginal Budgeting for Bottlenecks (MBB) modeling exercise revealed the following: I ZambiaisnotlikelytomeetitsMDGs,ifcurrenttrendspersist.Childhealthindicators are stagnating, and maternal mortality ratio is tending to increase. I The bottleneck analysis showed that although service availability is reasonable, a series of bottlenecks exist in the service delivery chain that severely limit service accessibility, initial utilization of the services, timeliness and continuity of services, and finally effective (high quality) coverage. This pattern is true in all three types of service packages considered (community-based, population-based outreach, and facility-based services). WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xvi xvi Executive Summary I TheMBBmodeldevelopedandcostedoutfivediscrete"steps"oroptionsforscalingup service coverage, and arrest further deterioration in Zambia's MDG indicators, espe- cially on maternal and child health. The "steps" indicate the optimal sequential nature of the proposed options, focusing first on the "big-wins" in terms of reducing under- five child and maternal mortality. I Note also that the MBB resource forecasts do away with budgeting using installed capacity (which is what MOH currently and traditionally does), and instead focuses on what needs to be done in terms of services to be funded. I The five steps,if implemented successfully,would cost US$4.36 per capita per annum under Phase I, and US$9.07 per capita per annum under Phase II. For these invest- ments, (a) child mortality is expected to decline by 24 percent for Phase I and 46 per- cent for Phase II; (b) maternal mortality is expected to fall by 5 percent for Phase I and 16 percent for Phase II. Key Recommendations Given the nature of this review, it would not be possible to provide specific recommendations for each of the host of issues that were identified. Rather, this section only highlights the major thrusts of the recommendations. Obtain National Commitment to Increase Fungible (Flexible) Financing First of all, there must be recognition at the highest levels of Government that the health sector is in a state of crisis--the increasing flows of external assistance notwithstanding--and that the crisis must be dealt with in a coordinated manner. This requires closer dialogue between MOFNP and MOH that should result in stronger agreement on the level of spending that the sector needs over the medium term. MOFNP must recognize, first of all, that most of the donor financing are non-fungible, and therefore it cannot abdicate its duty of increasing the alloca- tion to health. (The percentage share of Government health expenditure to total health expen- diture has fallen from 7.7 percent in 1997 to 4.7 percent in 2004).Allowing non-fungible donor financing to increase while fungible government financing shrinks only results in further dis- tortions in factor use (lack of funding for health workers, increasing administrative intensity, increasing off-budget financing--all classic symptoms of the imbalance between donor and domestic financing that were documented in this public expenditure review (PER). The will- ingness of a few donors to convert their assistance from projects and basket funds to budget support, therefore, must be embraced by MOH as well (which has been hesitant so far), because this means that there will be more fungible financing available to the sector. Yet, having more fungible financing also raises the challenge to the MOH to produce credible, financeable pro- grams that MOFNP and the budget-supporting donors can fund. Formulate an Overall Wage Strategy Because of its urgency, this should be top priority. The Strategic Plan for Human Resources for Health (2006­11) has been approved, but the recurrent cost implications of it remain unfunded. The human resource crisis cannot be solved by incremental approaches, which is what the government and donors have done so far (through top-ups, implicit incentives, and a WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xvii Executive Summary xvii widening array of cash and in-kind benefits),and which have not yielded lasting solutions,given that the vacancies persist and staff turnover is increasing. Such a strategy must be informed by the ongoing HRH productivity study, as well as medium-term fiscal sustainability (forecasts of GDP growth and fiscal revenues). Improve Payroll Management Although salary management issues only involved 10­15 percent of surveyed health workers, these disgruntled civil servants could easily reduce overall morale.In any case,any form of"salary pinching"and delay cannot be allowed. Cash payments to workers should be eliminated as soon as feasible, and replaced with automatic bank deposits. If it does not yet exist, an administrative order against payment and receipt of expediter's fees should be formulated and enforced. Enforce Rules on Absenteeism and Tardiness, and Provide Better Incentives for These Two Problems to be Reduced These twin problems have been ignored as the government has focused almost solely on filling up posts, rather than managing those who are already in post. A renewed campaign for being on time and fulfilling eight hours of official work is needed. The supervision teams, which visit facilities regularly and seem to do their work well, should be empowered to call to task staff who are out or late. As absenteeism and tardiness are intimately related to dual practice and other income-augmenting staff activities, a clear policy guideline to co-opt such coping behavior should be developed. Instead of off-site training, on-the-job skills acquisition should be pur- sued. Soft incentives (recognition, small-cash awards) should be provided for staff who regis- ter perfect or near-perfect attendance. Enhance Accountability and Financial Management Capacity at the District Level The PER showed that while there were no major hitches in the flow of funds from the central to district levels, the flows from DHMTs to health facilities were quite problematic. These could be due either to deliberate withholding of resources by influential DHMT members (in a form of "local capture"of resources), or bureaucratic delays due to weak fiscal management systems. (Drug diversion could also be a manifestation of undue influence by local elites, if the problem does not arise from sheer misrouting or other logistical errors,which are hard to believe)."Local capture" is an accountability issue which can be solved creatively by increasing client power, using such proven methods as publishing in national and local media the resources transferred to each district; posting the resources received at the doors of facilities, DHMT offices, or other visible areas; empowering local groups to be more vigilant; using community radios to report on the status of the fiscal transfers to districts and provinces; and using community report cards that provide active voice to local citizens. The financial-capacity issues need to be explored fur- ther to understand exactly the organizational constraints being faced. Fortify the District Health Financing Grants Although it is the equity-enhancing instrument of overall health financing, the district grant was emasculated in 2006 with large reduction in nonwage expenditures (down from 2005 by WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xviii xviii Executive Summary 13 percent) and drug expenditures (down by 34 percent). The reductions occurred even as Zambia abolished user fees for primary services that clinics and other lower-level facilities highly rely on. This reversal need to be redressed, and future allocations should take account of the following: I Unfunded and underfunded public health programs: The PET/QSDS revealed that many districts continue not to have adequate allocations for child health, maternal health, malaria, water and sanitation, and HIV/AIDS programs. A few districts do not undertake proper resource planning at all. Despite these, health facility in-charges tend to be overly optimistic about their actual capacity to deliver services. I Foregone fee revenues and expected increase in demand: Health facilities need to be paid the amount of fees they used to generate (average of ZK4.4 million for Rural Health Centers [RHCs] and ZK47.9 million for UHCs) as these were spent for a vari- ety of purposes to improve health services. More importantly, health centers and dis- trict hospitals need to brace for the expected increase in patients occasioned by the fee abolition, requiring additional resources. I Drugs: Thereductionof districtdrugallocationsin2006needtobereversed.Ideally, allocations should be based on per capita standards. I LargeRHCdemandformedicalinstrumentsandlabsupplies: ThePET/QSDSshowed a high incidence of medical consumables and instruments, for example, microscopes, audioscopes, ob-gyne instruments, gowns andprotective clothing, malaria smear, and urine test strips. Provide a Special Focus of Intervention to Address the Appalling Situation of Maternal and Child Health The overall recommendation of this PER is to support more fungible financing through increased MOFNP allocation to the sector, and increased budget support program by donors. This would deal with system-wide constraints, such as the addressing the human resource cri- sis, correcting the imbalances in geographic allocation, and improving financing to districts. However, the child and maternal mortality indicators are such that they require focused inter- ventions. The MBB simulation model specified a five-step approach ("options") that GRZ could choose from, based on the additional resources that it could muster. Elevate Hospital Investments and Financing to the Policy Agenda The relegation of hospital issues over the past years has had the unintended consequence of hos- pital financing happening by default, that is, dictated by patterns established in past invest- ments, including those of former mine hospitals. The urgency of setting testing and treatment centers for a variety of diseases (HIV/AIDS, tuberculosis, malaria, referral cases for child health, and to a lesser extent, maternal health) has resulted in hospital investments being dictated by specific disease initiatives with little overall coordination and long-term strategic vision. Mean- while, private clinics and hospitals have mushroomed with little regulation, and garnering most household health expenditures (86.4 percent of total household health expenditures went to for-profit facilities between 2002 and 2004, as shown in the NHA analysis). In the absence of any sizeable risk-pooling arrangement and fee regulation, it is unclear whether households are WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xix Executive Summary xix getting "value-for-money" for these expenditures, because they pay them at the point of vul- nerability. This patchy approach needs to be corrected. Admittedly, this is a broad topic, but these three tasks could be the spring-board for wider policy discussions: I Hospitalstock-takingexercise: Thisshouldcovernumber,distribution,andownership; patient flows and referral patterns; current physical deficits (as gleaned from the PET/QSDS); current financing; governance, management and staffing; resource requirements; sustainable number and distribution; and roles of private, mission, and government hospitals. I A new equity- and efficiency-enhancing allocation formula for government hospitals: The current allocation formula preserves the status quo, does not engender efficiency and performance, and is highly geographically inequitable (favoring the urban back- bone of the country). Suitable resource-allocation models from elsewhere in the devel- oping world should be gleaned for possible adoption in Zambia. Most importantly, GRZ should have the courage to declare some hospitals redundant, if the costs of main- taining them far exceed the service that they provide. I Risk pooling/social health insurance:A Cabinet paper has been submitted for this pur- pose. Although social health insurance (SHI) for government workers will not directly impact the welfare of the majority, SHI does have large positive collateral effects on the health system, and these "externalities" could quickly improve health services that non- SHI members could also enjoy, as has been shown in Tanzania's National Health Insur- ance Fund.The"active purchasing"inherent in fee negotiation and hospital accreditation enhances institutional performance. The insurance payments that facilities receive directly translate to better-incentivized staff, more reliable drug supply financing, and better maintained facilities that everyone benefits from, whether SHI members or not. WB Zambia_FM.qxd:WB ECarCS_FM.qxd 11/27/08 2:39 PM Page xx WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 1 CHAPTER 1 Background, Objectives, and Analytical Tools O ver the past few years, three nagging problems have bedeviled Zambia's health sector: the country is falling off-track from reaching the Millennium Develop- ment Goals (MDGs), it is facing severe financing constraints on the government front, and the health and HIV/AIDS sector is increasingly being fragmented by the re- emergence of global disease initiatives. This health sector pubic expenditure review (PER) seeks to assist the Government of the Republic of Zambia (GRZ) and its development partners take stock of the resources in the health sector and how these resources can be better used to produce better health services. Impelling Factors for the Health Sector PER The Need to Quantitatively Assess the Performance of the Health Sector and to Better Explain the Reasons Behind the Poor Health Indicators Previous stock-taking exercises and assessments of the health sector in Zambia have been largely qualitative (the Mahler Review in 1995, and the succeeding annual reviews of the SWAp). Though comprehensive and issues-oriented, these were largely descriptive and anecdotal. In contrast to the strong household-based data on utilization of services obtained through serial Demographic and Health Surveys (DHS) and Living Conditions Monitoring Surveys (LCMS), "supply-side" facility-based performance data have been, for the most part, lacking. The PER will undertake a Public Expenditure Tracking (PET) study to ease this information gap. 1 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 2 2 A World Bank Country Study The Need to Construct Alternative Scenarios for Improved Allocation and Use of Health Resources A common view in Zambia is the need to increase per capita health spending. Although there is little disagreement about this need, Zambia's per capita health expenditures are still relatively much higher compared to its neighbors. For instance, in 2003 the country's per capita health spending was US$21, about 1.6 times as high as Malawi's US$13, Tanzania's US$ 12, and Mozambique's US$12. Moreover, because of the pooled funding mechanism that has been going on for quite sometime, Zambia's health financing flows is not as highly fragmented nor verticalized as in other countries that were not in basket-funding mode (Malawi, Tanzania up until the early 2000s). Despite these, however, health indicators have continued to stall, signifying possible resource allocation and/or service delivery problems. The delivery system still remains weak and has not significantly reduced the disease bur- den. Moreover, it is unclear whether these resources have been targeted towards the indi- gent and vulnerable groups. To help GRZ and its development partners understand these issues, the PER will run the Marginal Budgeting for Bottlenecks model to explore alterna- tive scenarios for resource allocation. Objectives and Analytical Tools of the Health Sector PER The specific objectives of the PER and the corresponding analytical tools to be employed are shown in the table below: Objectives Analytical Tools To map sources and uses of funds in National Health Accounts (NHA) exercise--the the health sector results are presented in Chapter 3 To assess the allocation, release, actual Public Expenditure Tracking and Quality of Service receipt and use of public expenditure Delivery Survey (PET/QSDS)--the results are down to the facility level presented in Chapters 4 and 5, respectively To analyze alternative scenarios for Marginal Budgeting for Bottlenecks (MBB) model-- improving allocation of resources in the results are presented in Chapters 6 the health sector and options for service delivery Objective 1: Mapping the sources and uses of funds in the health sector GRZ and its donor partners pioneered "basket funding" under a program of sector-wide approach (SWAp) in the 1990s. Under this approach, participating donors pooled their resources, and these are programmed side by side with government budgetary allocations, user fee revenues, and in-kind resources for the provision of essential health services at the district level. The advent of global health initiatives such as the Global Fund for AIDS, Tuber- culosis and Malaria; Roll Back Malaria (RBM); Stop TB; the Global Alliance for Vaccine Ini- tiative (GAVI); and the U.S. Government's Presidential Expanded Program for AIDS Response (PEPFAR) has made the financing of health services ever more fragmented, neces- sitating the need to map all sources and uses of funds. Conventional wisdom is that basket funds now account for only around 25 percent of total resource flows in the health sector. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 3 Zambia Health Sector Public Expenditure Review 3 The National Health Accounts (NHA) is a well-established tool for mapping the sources and uses of funds. Zambia has a relatively longer history of producing NHA esti- mates than other countries in the region, starting in the mid-1990s (Mwikisa and others 2000), and up to the mid-2000s (Phiri and Tien 2004). The methodology and analytic cat- egories of NHA are standard and involve: a) Identification, quantification, and analysis of the sources of funds, including those from the public sector, private companies, households, and donors/cooperating partners. b) Identification, quantification, and analysis of flows of funds through financing agents, including those of the public sector, private companies, households, and donors/cooperating partners. c) Identification,quantification,andanalysisoftheflowoffundstoend-users/providers of care, including public facilities, private facilities, and others/central administration and management. d) Identification, quantification, and analysis of uses of funds by function, including inpatient care, outpatient care, and others (and where feasible, by sub-functions). Objective 2: Tracking of public expenditures Zambian discussions on health financing have tended to focus on the requirements of the basic health care package to be delivered, rather than on actual use of available resources. Through the Health Sector Committee Secretariat, formalized annual work-planning and quarterly budgeting are being undertaken for each facility and provincial health office receiving funds from the "basket" (CBOH, UNZA, and IHE 2004). However, these exer- cises are not complemented with equally important analysis of health facilities' receipt and use of basket funds, other donors' vertical financing flowing to districts, and fees that dis- trict facilities generate. To address this problem, the Public Expenditure Tracking and Quality of Service Delivery Survey (PET/QSDS) is employed in this PER. I ThePETcomponentfocusedonbudgetallocation,release,andspending,fromthe MOH down to the health facility level. The analysis tried to capture all resource flows including the GRZ budget and basket funds, funds from vertical projects, and internally generated funds. I Thequalityofservicedeliverycomponentfocusedoninfrastructure,utilities,and equipment; health personnel; drugs and other medical consumables; and clinic and patient management. Objective 3: Analysis of alternative scenarios for allocating health resources and options for service delivery Money alone is not enough and proper allocation of health resources is essential to achieve good results. Comparing alternative ways of allocating health resources would help the country to choose the most effective strategy of resource allocation and service delivery. Specifically, two issues need to be addressed in order to maximize the impact of the limited health resources. One is the formulation of essential service packages, which is important WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 4 4 A World Bank Country Study to prioritize the high-impact interventions, and about which Zambia has focused a lot of attention over the past few years. Another is the examination of the ways of services being delivered in order to identify cost-effective delivery mechanisms. I Alternativeservicepacks:Whatmakeshealthsectorcomplexisitsnatureofmultiple outcomes, outputs, and inputs. The existence of multiple interventions targeting the same disease or outcome requires careful planning, prioritizing, and targeting strat- egy. Looking into different intervention packages and identifying the most cost- effective interventions can help the country to make progress toward MDGs. I Alternative delivery mechanisms: Health services can be delivered not only within health facilities, but also through some innovative approaches such as outreach and community-based initiatives. International experience indicates that the latter is actually very effective in scaling up some key health services (e.g. immunization, bed net use). The World Development Report 2004 Making Services Work for the Poor (World Bank 2004) suggests that the poor, who normally lack access to health facil- ities, benefit more from outreach and community-based delivery. Investigating the options of service delivery has implications on both scaling up service coverage and reducing cost. Unfortunately, the increasing fragmentation of the financing health services and the recent re-verticalization of service delivery emerging from global disease initiatives have somehow resulted in the relegation of service delivery impact and effectiveness as important considerations in the allocation of resources. The Marginal Budgeting for Bottlenecks (MBB) model is an analytical tool of health service delivery, developed jointly by teams from the World Bank, UNICEF, and WHO. It was orig- inally designed to answer the following three questions: What are the major bottlenecks in service delivery? What are the resources required to remove the bottlenecks and to reach a new coverage frontier of health services? What is the impact of the marginal increase of ser- vice coverage on health outcomes? MBB provides a more holistic (rather than fragmented), cohesive (rather than piece-meal), and "horizontal" (rather than "vertical") approach to budgeting and allocation. MBB has been applied in various countries for different purposes. For instance, it was used to better allocate and use the additional resources expected from debt relief in HIPC; to improve the planning process in Mozambique; to mobilize more resources for the health sector in Madagascar; to formulate an MDG-reaching strategy in Ethiopia. MBB can be particularly useful to Zambia for its system-wide focus, bottleneck- identification approach, results-based costing, and impact-simulating methodology. Uses of the PER The results of the PER are expected to be the used for a variety of purposes, including the preparation of the health sector strategic plan, and succeeding rounds of the Global Fund request for proposals. Policy dialogue between the Bank and GRZ, both at the macro and sector levels, can also be enriched by the PER. The PER also provides critical inputs into the Medium-Term Expenditure Framework (MTEF) process, and in the assessment of the Poverty Reduction Strategy Paper (PRSP). Likewise, the PER can provide inputs to fine-tune theprocessofthepooledbasketfundingmechanismunderthesector-wideapproach(SWAp). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 5 CHAPTER 2 Health System Challenges Health System Goals and Structures While the goal of the GRZ health system has been constant, the structure has been subject to dramatic changes in over a decade. The overarching goal of the Zambian health sector is "equity of access to assured quality, cost-effective and affordable health services as close to the family as possible", and to provide cost effective quality health services as close to the family as possible in order to ensure equity of access in health service delivery and con- tribute to the human and socio-economic development of the nation." To achieve this goal, the government started decentralizing health services under a health sector reform program in the early 1990s. The intention was to transfer key management responsibilities and resources from the central MOH to the district level. In this respect, two parallel but complementary structures were introduced, namely: I The popular structures for public involvement and participation in the decision- making process, including the Central Board of Health (CBOH), Hospital Man- agement Boards, District Health Boards (DHBs), and the Neighborhood Health Committees and Health Center Committees. I The technical and management structures designed to ensure that services are implemented in a sound manner, including the management teams at MOH and CBOH, the Hospital Management Teams, and District Health Management Teams (DHMTs). The Provincial Medical Offices were reconstituted into Provin- cial Health Offices (PHOs). After years of operating under a "split purchaser/provider model" (MOH being the pur- chaser and CBOH the provider), the government decided to reunite the two functions 5 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 6 6 A World Bank Country Study Table 1. Health Facilities in Zambia, 2000s Type/Level GRZ Mission Private Total Hospitals 53 27 17 97 Health centers 1,052 61 97 1,210 Health posts 19 0 1 20 Total 1,124 88 115 1,327 Source: CBOH, Health Institutions in Zambia: A Listing of Health Facilities According to Levels and Locations, 2002. under one agency. In 2006, the Government abolished CBOH and put its functions, assets, and staff back under the MOH. This comprehensive restructuring process will effectively merge the two bodies. Thus, management and control of all public health facilities and ser- vices will again directly fall under MOH, through the PHOs. Diverse providers of health services include: public health facilities under MOH, facil- ities under the Ministry of Defense including clinics and one hospital in Lusaka, clinics under the Ministry of Home Affairs, mining hospitals and clinics, mission hospitals and clinics which are coordinated by the Churches Health Association of Zambia (CHAZ), non-government organizations (NGOs), private for profit hospitals, clinics, pharmacies/ drug shops, labs and investigation centers, and traditional healers. The private for-profit sector is growing, but mostly concentrated in urban Zambia. Table 1 summarizes the total number of health facilities. I HealthPosts:Intendedtocaterforpopulationsof500households(3,500people)in the rural areas and 1,000 households (7,000 people) in the urban areas, or to be established within five-km. radius for sparsely populated areas. The target is to have 3,000 health posts but only 20 exist. I Health Centers: These facilities include urban health centers (UHC) which are intended to serve a catchment population of 30,000 to 50,000 people, and rural health centers (RHC) which service a catchment area of 29-km. radius or popula- tion of 10,000. The target is 1,385 compared to a current total of 1,210 health cen- ters (973 RHCs and 237 UHCs). I First-levelHospitals:Thesearefoundinmostofthe72districtsandareintendedto serve a population of between 80,000 and 200,000 with medical, surgical, obstetric and diagnostic services, including all clinical services to support referrals from health centers. There are 74 first-level referral hospitals. I Second-level Hospitals: These are general hospitals at provincial level and are intended to cater to a catchment area of 200,000 to 800,000 people, with services in internal medicine, general surgery, pediatrics, obstetrics and gynecology, dental, psychiatry and intensive care services. These hospitals are also intended to act as referral facilities for the first-level institutions, including the provision of technical back-up and training functions. There are 19 second-level hospitals. Two provinces, namely Southern and Copperbelt, have five and three second-level hos- pitals, respectively. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 7 Zambia Health Sector Public Expenditure Review 7 I Third-level Hospitals: These are central hospitals for catchment populations of 800,000 and above, and have sub-specializations in internal medicine, surgery, pediatrics, obstetrics and gynecology, intensive care, psychiatry, training and research. They act as referral facilities for second-level hospitals. There are five such facilities in the country, of which three are in the Copperbelt Province. Challenges at the Health Outcome Level When studying health issues, one has to take into consideration factors both at the health outcome level, health output (health services) and health inputs (service delivery systems and key elements). Zambia faces challenges across all the three levels. Zambia has not gone through the demographic and epidemiological transition yet. Its growing population is estimated at 11.3 million (growing at 3 percent a year), with 47 percent of the total popu- lation less than 15 years old. Zambia is also one of the most highly urbanized countries in sub-Saharan Africa, with much of the population living along the traditional mining towns along the "copper belt." The epidemiological profile in Zambia features high child mor- tality, high maternal mortality, and a heavy burden of diseases mainly caused by commu- nicable diseases. Despite recent improvements in service delivery, overall health status has stagnated and the disease burden has continued to increase. The disease burden is over- run by the high prevalence of HIV/AIDS, and compounded by high poverty levels and the poor macroeconomic situation in most of the early 2000s. Table 2 summarizes the statis- tics on the recent trends for the major diseases. First, Zambia suffers high child mortality. Zambia has among the highest child mor- tality rate in the world. The infant mortality rate (IMR) and under-five mortality rate (U5MR) were as high as 95 and 168 per 1,000 live births in 2001, as indicated by the latest survey data, the Zambia Demographic and Health Survey (DHS), 2000­2001. The latest indirect estimate as shown by the World Development Indicators (World Bank) data indi- cated that the IMR and U5MR were 102 and 182 per 1,000 live births in 2004. The level of child mortality in Zambia compares unfavorably with most of other countries, higher than world as well as Sub-Saharan African average (Figure 1). When compared with surround- ing countries, both the IMR and U5MR in Zambia are only significantly lower than those in Angola and Democratic Republic of Congo, two countries severely affected by civil wars recently. Over time, Zambia has not been able to achieve a consistent decrease of child mortality. The child mortality rates declined from 1955 to 1980, followed by a notable and progressive increase from 1980 to 1996. Then from 1996 to 2002, the Zambia DHS recorded a significant decrease in the IMR and U5MR. But the most recent data indicate the child mortality rates are showing signs of increasing (Figure 2). Second, Zambia's maternal mortality ratio (MMR) is one of the highest in the world. The MMR was 729 per 100,000 live births in 2002 (ZDHS), which is unacceptably high. The direct causes of those preventable maternal deaths include postpartum hemorrhage, sepsis, obstructed labor, post-abortion complications, and eclampsia. The indirect causes are malaria, anemia, HIV/AIDS as well as delays in accessing health facilities. What makes the situation more worrisome is that the MMR has increased since 1996, in which the MMR was 646 per 100,000 live births. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 8 8 A World Bank Country Study Table 2. Summary Statistics on the Major Diseases, 2000­05 Disease Indicator 2000 2002 2004 2005 Malaria Incidence/1,000 316 388 383 373 Cases 3,591,621 4,101,169 4,328,485 Deaths 8,952 9,021 8,289 Respiratory infection, Incidence/1,000 119 148 153 161 non-pneumonia Cases 1,340,283 1,565,430 1,726,597 Deaths 1,269 1,057 1,436 Respiratory infection, Incidence/1,000 35 45 44 42 pneumonia Cases 402,643 475,389 494,040 Deaths 4,254 4,484 4,186 Diarrhea, non-blood Incidence/1,000 65 80 75 75 Cases 739,055 846,336 843,423 Deaths 2,795 2,996 2,725 Eye infections Incidence/1,000 47 43 40 40 Cases 471,743 451,346 448,280 Deaths 72 8 5 Trauma Incidence/1,000 34 42 46 46 Cases 390,869 447,278 525,039 Deaths 646 787 833 Skin infections Incidence/1,000 28 37 42 42 Cases 309,758 393,384 472,746 Deaths 135 126 125 Ear, nose, throat infections Incidence/1,000 21 25 23 24 Cases 238,403 260,058 259,877 Deaths 49 31 34 Source: MOH. Third, Zambia is affected by a heavy disease burden, which is overwhelmingly con- tributed by a small number of communicable diseases, malaria, HIV/AIDS, and TB. More worryingly, the heavy disease burden, which has adversely affected the Zambia poverty reduction and economic development efforts, has continued to increase. Based on the Health Management and Information System (HMIS) data, the incidences of the top five diseases (malaria, respiratory infection-non pneumonia, diarrhea, respi- ratory infection-pneumonia, eye infections) either increased or stagnated from 2002 to 2005 (Figure 3). I Malaria:Incidencerateshavetripledinthepastthreedecades,makingmalariathe leading cause of morbidity and mortality in the country. The entire population is at risk, especially children under five and women. The disease accounts for 37 per- cent of all outpatient hospital visits. In 2004, out of its 11.3 million population, about 4.3 million malaria cases occurred. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 9 Zambia Health Sector Public Expenditure Review 9 Figure 1. Comparison of IMR and U5MR in Selected African Countries, Latest Year Available IMR 180 160 140 births 120 live 100 80 60 1,000 40 per 20 0 Angola Africa Congo World BotswanaLesotho Malawi Namibia TanzaniaUganda ZimbabweZambia n Africa Madagascar Mozambique South Sub-Sahara U5MR 300 births 250 200 live 150 100 1,000 50 0 per Angola AfricaCongo Africa World BotswanaLesotho Malawi Namibia TanzaniaUganda Zambia Zimbabwe MadagascarMozambique South Sub-Saharan Source: World Development Indicators, World Bank, 2006. Figure 2. Trends in Neonatal, Infant, and Under-Five Mortality in Zambia, 1992, 1996, and 2001/02 Rates (Per 1,000 Live Births) 200 191 197 168 150 107 109 100 95 50 43 35 37 0 1992 1996 2001/02 Neonatal MR Infant MR U5 Mr WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 10 10 A World Bank Country Study Figure 3. Trends and Levels of Major Diseases in Zambia, 2002­04 450 400 350 300 250 200 150 Incidence/1,000 100 50 0 2000 2002 2004 Malaria Respiratory Infection: Non-Pneumonia Diarrhoea Respiratory Infection: Pneumonia Eye Infections Source: Ministry of Health, HMIS, 2005. I HIV/AIDS:Basedontheseroprevalencesurveyofthe2001/02DHS,nationalHIV prevalence rate among adults 15­49 years is 16 percent, among the highest in the world, but lower than the previously-believed figure arising from antenatal clinic data. Among women of childbearing age the prevalence even reached 20 percent, although it appears that prevalence rate has fallen among women 15­19 years of age. Indeed, HIV/AIDS control is one of the few indicators moving in the right direction in Zambia, and the country can rightfully take pride in being able to mod- estly stabilize the rate of infection. Still, the country's life expectancy has declined from 46 to 37 years mainly due to the HIV/AIDS epidemic. Moreover, the number of people living with HIV/AIDS, including orphans, have reached massive propor- tions, with as many as 12 to 19 percent of children (depending on the year and method of estimation) being considered orphans. I Tuberculosis:Thenumberoftuberculosiscasesnotifiedhassoaredfrom16,863in 1990 to 54,200 in 2002. Zambia's 2002 tuberculosis incidence rate of 668 cases per 100,000 population is the fifth highest in sub-Saharan Africa, according to the WHO TB Report, 2003. In 2002, tuberculosis prevalence (all forms) was 736 per 100,000 population while smear-positive prevalence was 353 per 100,000 popula- tion. About 49.4 percent of tuberculosis mortality is attributable to HIV/AIDS (Corbett and others 2003). Challenges at Health Service Level Zambia has made great efforts in defining and delivering a package of basic health services and some improvement has been achieved in some basic health service indicators. Based on the HMIS data, the coverage of several key services such as immunization, antenatal WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 11 Zambia Health Sector Public Expenditure Review 11 Table 3. Selected Health Service Delivery Indicators, 2000­04 Indicator 2000 2001 2002 2003 2004 Health center outpatient 0.42 0.77 0.73 0.86 0.76 per capita attendance First antenatal coverage (%) 81 88 89 95 97 Average antenatal visits (times) 3.6 3.6 3.4 3.3 3.1 Supervised deliveries (%) 39 44 49 55 61 Fully immunized children 76 86 76 74 80 under 1 year (%) New family planning acceptors 85 101 111 123 127 rate per 1000 Health center staff load 17 14 16 17 17 (patients/staff) Drug kits opened 0.73 0.75 0.69 0.73 0.93 per 1,000 patients Source: Central Board of Health. care, supervised deliveries has maintained at a high level or obtained a progressive increase since 2000 (Table 3 and Figure 4). However, the basic health service package as a whole has not been fully implemented particularly for the first and second level referral care. The improvement in key health services has not been even across provinces. Inequal- ities in key service coverage are rather salient based on the official data. As a general pat- tern, Lusaka does better in almost all health services (Figure 5). As a contrast, Northwestern and North regions lag behind in immunization, institutional delivery and postnatal care. Figure 4. Coverage of Key Health Services, 2000­04 120 100 80 % 60 40 20 0 2000 2001 2002 2003 2004 First Antenatal Coverage (%) Supervised Deliveries (%) FullyImmunisedChildrenUnder1Year(%) Source: Ministry of Health, HMIS, 2005. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 12 12 A World Bank Country Study Figure 5. Provincial Inequities in Health Service Coverage by Region 120 100 80 % 60 40 20 0 Central Eastern Luapula Coperbelt LusakaNorthernWesternSouthern WesternAverage North Full Immunization Institutional Delivery Postnatal Care Source: Ministry of Health, HMIS, 2005. Inequities also exist in residential location and socioeconomic status. Figure 6 and Figure 7 show that rural and poorer households have suffered disproportionately compared to urban and richer households. The rural/urban and poor/rich gap is particularly stark in the case of services reliant on skilled professionals such as birth attendance, although the gap is not so sharp in public-health services such as immunization. The poor child health indicators is partly explained by low coverage of child health interventions as well as continuing high rates of child malnutrition, a key risk factor in childhood illness. Zambia has the highest rate of underweight children (29.4 percent) and the third-highest rate of stunted children (43.8 percent) in East and Southern Africa, after Mozambique and Malawi. Zambia's child health interventions could be further improved to be at par with its neighbors (Table 4). For instance, Tanzania exceeds Zambia's rate of measles immunization, children receiving antimalarials, and children receiving Vitamin Figure 6. Differences in Selected Maternal and Child Health Indicators Between Rural and Urban Households in Zambia, 2000/01 100 79 84 86 80 51 60 37 40 28 20 0 % under-5 stunted % births w/skilled personnel % children immunized against measles Rural Urban WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 13 Zambia Health Sector Public Expenditure Review 13 Figure 7. Differences in Selected Maternal and Child Health Indicators Between Lowest and Highest Asset Quintile Households in Zambia, 2000/01 91 88 100 81 80 54 60 32 40 18 20 0 % under-5 stunted % births w/skilled personnel % children immunized against measles Lowest quintile Highest quintile A. Mozambique also exceeds Zambia's rate of children receiving oral rehydration therapy (ORT) for diarrhea. The rates of coverage of maternal health services are low. The proportion of births attended by a medically trained provider remains woefully low (43 percent), but espe- cially so among rural mothers (28 percent), while births by C-section (an indicator of the availability of appropriate obstetric care) is only 2 percent (Table 5). For both these maternal-health service indicators, Zambia is on the lowest rung of countries in East and Southern Africa. Recent malaria control initiatives are improving coverage rates. Household owner- ship of insecticide-treated nets has risen dramatically: according to the 2001/02 DHS, only Table 4. Service Coverage Indicators of Key Childhood Health Interventions in Selected African Countries, 2000s (Latest Year Available) Measles U5 U5 U5 6­59 Immuniz'n Children w/ Children w/ Children w/ Months Rate ARI Taken Diarrhea Fever Receiving Children Countries 2004 to Facility Receiving ORT Antimalarial Given Vit. A Botswana 90 n.a. n.a. n.a. n.a. Kenya 73 49.1 34.2 26.5 91.4 Lesotho 70 54.4 32.1 n.a. n.a. Malawi 80 26.7 35.4 31.6 85.8 Mozambique 77 55.4 46.7 n.a. n.a. South Africa 81 73.9 n.a. n.a. n.a. Swaziland 70 n.a. n.a. 25.5 n.a. Tanzania 94 45.8 36.3 53.4 94.2 Zambia 84 69.1 40.9 51.9 79.8 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 14 14 A World Bank Country Study Table 5. Maternal Mortality Ratio and Service Coverage Indicators of Key Maternal Health Interventions in Selected African Countries, 2000s (Latest Year Available) Maternal Antenatal Births Attended Mortality Ratio Coverage by a Skilled (per 100,000 Rate, Health Births by Countries Live Births) 4 Visits % Personnel C-section Botswana 100 97 94 n.a. Kenya 1,000 52 42 4.0 Lesotho 550 88 55 n.a. Malawi 1,800 55 61 3.0 Mozambique 1,000 41 48 3.0 South Africa 230 72 84 16.0 Swaziland 370 n.a. 70 n.a. Tanzania 1,500 69 46 3.0 Zambia 750 71 43 2.0 13.6 percent of households reported having at least one net, but the 2006 malaria indicator survey reports that now 50.1 percent of households have at least one mosquito net and 44.4 percent have at least one insecticide treated net (MOH 2006b). In the 2001/02 DHS, only 7.9 percent of pregnant women slept under an insecticide-treated net the night prior to two weeks preceding the survey; by 2006, this proportion has risen to 23 percent. Residual spraying is now ongoing in 15 districts, with those in Kabwe reporting the highest Figure 8. Unfilled Posts by Profession Unfilled Post % 100 80 60 % 40 20 0 ­20 Staff Total Total Doctors Nurses Wives Officers Mid Technician Pharmacists Support Clinical Lab. Professional Source: Ministry of Health, HMIS, 2005. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 15 Zambia Health Sector Public Expenditure Review 15 percentage of households sprayed (77.1 percent). Intermittent presumptive treatment of malaria for mothers only covered 35.8 percent of pregnant women in 2001/02, but by 2006, 77 percent of pregnant women were already taking anti-malarial drugs during pregnancy. Similar improvement in treatment of malaria in children has been recorded in the 2006 malaria indicator survey, where 58 percent of children with fever took an antimalarial drug. Challenges at the Service Delivery Level Delivering health services requires a functional health system, which in turns needs key ele- ments and essential inputs to make it work. At present, several challenges are affecting the effectiveness of Zambia's health system. In particular, human resources and health and health facilities are two major issues. The human resources situation in Zambia is at the state of crisis. The lack of appro- priate human resources, which is being further worsened by the high attrition rate, has severely undermined the system's capacity to deliver health services. Nationally, about 69 percent of professional posts are unfilled (Figure 8). Doctors, nurses, midwives, lab tech- nicians are in acute shortage. Significant gaps exist in the numbers and distribution of facilities required to cover the population. In total, Zambia has 1,327 health facilities including 97 hospitals, 1,210 health centers, and 20 health posts. The majority of the health facilities belong to GRZ. Based on the National Health Strategic Plan, health service delivery relies on a network of health facilities in order to cover the entire population (see Chapter 2). The existing stock of health facilities, especially at the lowest level, however, is a far cry from this ide- alized network. Primary health care facilities are particularly in shortage. For example, Figure 9. Gaps in Numbers of Health Facilities 100% 175 90% 80% 70% 60% 50% 2980 1,210 40% 30% 20% 10% 0% 20 Health Centers Health Posts Gap Existing Source: Ministry of Health, HMIS, 2005. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 16 16 A World Bank Country Study Figure 10. Trends of Child Mortality in Zambia, 1990­2015 IMR 120 Current trend births 100 80 live 60 40 MDG required 1,000 20 per 0 1990 1996 2002 2015 U5MR 250 births 200 Current trend live 150 100 1,000 50 MDG required per 0 1990 1996 2002 2015 Source: ZDHS. the target is to have 3,000 health posts but currently only 20 have been commissioned; for health centers, the target is 1,385 but currently there is only a total of 1,210 health centers (Figure 9). Based on current trends, Zambia is off-track to reach the MDGs, unless dramatic changes are instituted immediately. For the child mortality target, even using the most promising trend between 1996 and 2002, Zambia is still falling behind the required pace to reach the IMR and U5MR targets (Figure 10). It is even worse for the maternal mortal- ity target--the latest data suggests the MMR is rather increasing (Figure 11). Figure 11. Trend of Maternal Mortality in Zambia, 1996­2015 MMR 800 births 700 600 Current trend live 500 400 300 200 MDG required 100,000 100 per 0 1996 2002 2015 Source: ZDHS. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 17 CHAPTER 3 Sources and Uses of Funds in the Health Sector: Results of the National Health Accounts Exercises, 1999­2004 ambia has the longest annual series of National Health Accounts (NHA) in sub- Z Saharan Africa. This permits the analysis of health expenditure trends over the past ten years. This chapter synthesizes the findings of several NHA reports completed through the years under the auspices of the "Institutional Collaboration--Health Econom- ics Policy Analysis and Health Economics Project" of the Economics Department of the University of Zambia, with support from the Swedish International Development Agency. This chapter is organized as follows. The first section provides the overview by report- ing the trends in total health expenditures (THE) and expenditure ratios, especially ones pertaining to government. The next three sections analyze the patterns of health expendi- tures from the point of view of the sources and financing agents. The second section dis- cusses in detail the sources of health expenditures, the third section describes the health expenditures in terms of service providers and administration, and fourth section analyzes the application of health expenditures by donors, the MOH, and households. Then the chapter analyzes the patterns of expenditures from the point of view of the recipients. The fifth section examines how service providers are financed, and the sixth section analyzes the uses of health expenditures by these recipients in terms of inputs. Finally, the last sec- tion mulls the key implications of the NHA findings. Total and Per Capita Health Expenditures In nominal terms, total health expenditures increased from ZK172.7 million in 1995 to ZK1.86 billion in 2004, representing an eleven-fold increase in current prices (Figure 12). However, in real terms, THE increased by only 44.1 percent from 1995 to 2002. Total 17 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 18 18 A World Bank Country Study Figure 12. Total Health Expenditures in Nominal and Real Terms, 1995­2004 (ZK Billion, 1995 = 1.00) 2000 1,860.60 1800 1600 1,389.50 1400 1200 1,082.40 1000 721.2 800 562.8 600 328 413.8 423.7 400 243 172.7 200 172.7 197.7 211.9 223.7 188.8 192.8 198.7 248.8 0 0 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Nominal Real health expenditures as share of GDP inched up from about 5.5 percent in the late 1990s/early 2000s to 6.8 percent in 2003 and 7.2 percent in 2004 (Figure 13). At these THE/GDP ratios, Zambia occupies the intermediate range among East and Southern African countries, above the 4­5 percent rates of Tanzania and Mozambique, but Figure 13. Selected Health Expenditure Ratios, 1995­2007 (percent) 8 7.7 7.2 7.2 7 6.9 7 6.9 6.7 6.8 6.5 6.7 6.4 6.5 6.2 6 5.7 5.7 5.6 5.5 5.2 5 5 4.7 4 2.9 3 2.5 2.3 2.1 2.2 2.2 2.2 2 2 2 2 1.9 2 1.7 1.5 1.6 1.2 1 3.2 2.9 3 3.6 2.8 2.8 2.9 3.4 3.2 3 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Gov't HE/GDP - Defn. A Gov't HE/GDP - Defn. C Gov't HE/GDP - Defn. D THE/GDP Gov't HE/Total GE WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 19 Zambia Health Sector Public Expenditure Review 19 below the 8 percent rate of middle-income countries such as Swaziland, Namibia, and South Africa. During the same period of increasing THE, GRZ financial contributions to health have fallen in relative terms. As a percentage of GDP, it dropped from a consistent rate of 2.0 percent throughout the 1990s to as low as 1.2 percent in 2004. Similarly, government HE as a percentage of total government expenditures fell from an already-low share of 6­7 per- cent in the 1990s to a dangerously low rate of 4.7 percent in 2004. This pattern of low and declining government health spending underlies the major problems facing the Zambian health sector. While GRZ's share to total health spending has been going down, its discretion over resources has been increasing due largely to donor support. To analyze this trend, it is nec- essary to consider alternative definitions of "government health spending." There are no standard definitions of what constitutes "government health spending" in Zambia, as in many other African countries. This has led to confusion, especially in cross-country com- parisons. To clarify matters, the following are considered: I NHA spending by source: Under this definition, expenditures are categorized by source; thus, government health spending can be defined as: (a) MOH/CBOH expenditures (Definition A). (b) MOH/CBOH expenditures + health expenditures of all GRZ agencies, i.e., other ministries, Defense Forces, parastatals, and local government author- ities (Definition B). This may or may not include social health insurance (SHI), depending on the extent of discretion the central government or MOH has over SHI resources. However, this is not an issue in Zambia as there is no SHI. I NHAspendingby"financingagent":Underthisdefinition,expendituresarenotcat- egorized by the source or origin but by the financing agent (whomever has discre- tion over the expenditures). Thus, government spending can be defined broadly or narrowly as follows: (c) Definition B + all other health expenditures for which the government is the financing agent. Thus it includes the "basket funds" of donors under the sector- wide approach (SWAp), as well as all health projects implemented by the government, e.g., ADB, GAVI, Global Fund, JICA, OPEC Fund, and the World Bank (Definition C). (d) MOH/CBOH + Donor Basket Funds (Definition D). Using the above definitions, government health expenditures in Zambia are presented in Table 6. If one adopts a more liberal definition of government health spending to focus on financing agency and discretion, GRZ has indeed more resources at its disposal. Table 7 shows the resources available to the government, including MOH/CBOH expenditures, health expenditures of other GRZ agencies, basket funds of donors participating under the sector-wide program (SWAp), and resources from GRZ-implemented health projects. In 2004, GRZ had discretion over a total of ZK770.0 billion in resources, about 2.4 times the size of the MOH/CBOH budget, and due largely to massive increase in basket funds and donor projects managed by GRZ. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 20 20 A World Bank Country Study Table 6. Government Health Expenditures (ZK Billion) Using Alternative Definitions, 1995­2007 Definition C (MOH/CBOH + Other GRZ + Basket Funds + Definition D Definition B GRZ- (MOH/CBOH + Definition A (MOH/CBOH + Implemented Basket Years (MOH/CBOH) Other GRZ) Donor Projects) Funds) 1995 58.3 65.3 96.0 -- 1996 72.7 80.6 113.7 -- 1997 95.6 108.8 154.2 -- 1998 112.2 120.9 214.0 -- 1999 140.9 149.0 205.7 -- 2000 147.3 155.2 282.7 -- 2001 288.2 304.1 386.4 318.4 2002 342.4 350.2 545.6 408.9 2003 327.1 337.3 654.6 597.0 2004 322.4 332.8 770.0 579.4 2005 415.5 (e) -- -- 629.3 (e) 2006 estimated 557.0 (e) -- -- 682.3 (e) 2007 budget 820.2 (e) -- -- 977.5 (e) 1Note that given the plethora of government health spending definitions, it is not clear what the Abuja target of "15 percent of government spending devoted to health" means. As the numerator of the indi- cator is unclear, so is the denominator, i.e., whether it refers to total government spending, or only to government spending that is discretionary. Since 1999, nominal per capita health expenditure has consistently risen; it reached US$34.20 (in average exchange rate) by 2004 (Figure 14). This rise is mainly due to the lib- eral infusion of donor HE. This level of spending also places Zambia in the middle range of countries in East and Southern Africa, and is far higher than the US$12­13 per capita Table 7. Health Expenditures (ZK Billion) Under the Discretion of the Government, by Source, 2001­07 Sources 2001 2002 2003 2004 2005 2006 2007 MOH/CBOH 288.2 342.4 327.1 322.4 415.5 557.0 820.2 Basket funds 30.2 66.5 269.9 257.0 213.8 125.3 157.3 Subtotal 318.4 408.9 597.0 579.4 629.3 682.3 977.5 GRZ-implemented donor 52.1 128.9 47.4 180.2 -- -- -- health projects Other GRZ 15.9 7.8 10.2 10.4 -- -- -- Grand Total 386.4 545.6 654.6 770.0 -- -- -- WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 21 Zambia Health Sector Public Expenditure Review 21 Figure 14. Trends in Per Capita Government Health Expenditure and Per Capita Total Health Expenditure, 1995­2007 (US$ Ave. Exchange Rate) 40 35 34.2 30 25.5 26.5 25 23.3 21.9 21.3 22 20 19 19 17.5 17.6 17 15 14 13 13 12 12 12 11 11 11 10 10 10 8.1 8 8 8 6.8 87.8 87.5 6.4 6.1 6.2 5.9 5 4.9 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Definition A Definition C Definition D per cap. Total HE HE of Tanzania, Mozambique, and Malawi. As a point of interest, Zambia's per capita HE now slightly exceeds the level that the Commission of Macroeconomic and Health has deemed sufficient to finance a basic package of health care (US$33). While per capita THE has risen since the late 1990s, per capita government HE (Definition A) has fallen to US$5.90 in 2004, after peaking at US$8.10 in 1995. In contrast, the most liberal definition of GHE (Definition C, which includes MOH, other GRZ, basket funds, and GRZ-implemented donor health projects) shows an uptrend. Indeed, the combined forecast for MOH bud- get and basket funds (Definition D) shows more optimistic rising figures in nominal per capita HE. Sources of Health Expenditures Changing Pattern of HE Sourcing The advent of global disease initiatives in the late 1990s and early 2000s as well as the stag- nation of government health spending have shaped the pattern of the sourcing of health expenditures in Zambia. In 2004, total health expenditures reached ZK1.8 billion (Table 8), 11 times its nominal size in 1995. During the decade, donor health expenditures grew nearly 41 times to its current level of ZK790 million. Private health expenditures multiplied eight- fold to reach ZK735 million. Government health expenditures grew only five-fold--the slowest nominal expansion among the sources--to reach ZK409 million by 2004. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 22 22 A World Bank Country Study Table 8. Trends in Health Expenditures by Source, 1995­2004 (ZK Billion) Private Other All Years GRZ Donors Employers Households Private Private Total 1995 65.3 19.0 26.6 58.8 2.9 88.3 172.7 1996 80.6 43.4 33.4 81.6 3.9 118.9 243.0 1997 108.8 72.9 40.1 101.7 4.5 146.3 328.0 1998 120.9 95.1 57.8 130.9 9.0 197.7 413.8 1999 149.0 37.9 50.3 178.4 8.2 236.8 423.7 2000 155.2 101.1 59.2 223.8 23.6 306.5 562.8 2001 304.1 104.3 54.7 248.2 10.0 312.9 721.2 2002 350.2 336.2 62.5 308.9 24.7 396.1 1,082.4 2003 337.3 528.4 89.8 399.6 34.4 523.7 1,389.5 2004 332.8 790.1 128.9 528.9 79.9 737.7 1,860.6 Rapid Dominance of Donor Spending Reflecting the differential growth rates, donors now account for a massive proportion of health expenditures, with serious consequences on the way these resources are being allo- cated and spent. Over the past three years (2002­04), about 38 percent of THE have come from donors, 24 percent from government, and another 38 percent from the private sector. This has not always been the case. In 1995, donors only accounted for 11 percent of THE, and government accounted for 38 percent, while the private sector contributed 51 percent (Figure 15). Indeed, today Zambia's health sector has become the most donor- reliant, with the share of external resources to THE exceeding that of any country in the region, with the possible exception of Mozambique and Rwanda, which both came out of a war. For instance, in Uganda the share of donor resources has remained stable at around 29 percent; in Tanzania, it exceeded 30 percent in the early 2000s but has since stabilized; and in Mozambique it exceeded 40 percent in the early 2000s, but has also stabilized. In Zambia, the staggering and still-increasing amount of donor resources sitting on what is essentially diminishing government resources raises important public finance questions that beg to be properly articulated and analyzed. Declining Share of GRZ Health Expenditures GRZ health expenditures are coursed through the Ministry of Finance and National Planning (MOFNP), local government authorities (LGAs), and other public funds (see Annex Table A). The Treasury's funds are mainly channeled through the Ministry of Health (MOH), which takes a substantial 96.9 percent of all MOFNP expenditures for health. The rest (3.1 percent) go to the Ministries of Defense, Education, and Home Affairs, other public institutions and, up until the late 1990s, to nongovernmental orga- nizations (NGOs). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 23 Zambia Health Sector Public Expenditure Review 23 Figure 15. Trends in Health Expenditures by Source, 1995­2004 (ZK Billion) 2000 1800 1600 1400 1200 1000 800 790.1 600 528.5 336.2 400 104.3 200 72.9 95.1 37.9 101.1 19 43.4 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 GRZ Donors Employers Households Other private Health Expenditures by LGAs for the Most Part Unknown LGAs operated clinics and health centers in their jurisdictions until 2000, when these were taken over by MOH. The bigger LGAs, such as the Lusaka City Council, continue to imple- ment public health programs. For most LGAs, however, the health expenditures are not separated from other expenditures, making it difficult to isolate health spending. A survey of all 72 LGAs need to be carried out to determine their health expenditures, although it is bound to be minor. Increasing Off-budget Expenditures The proportion of donor health expenditures going to MOH and other public institutions has declined from 84 percent in 2000 to 59 percent in 2004 (see Figure 16, derived from Annex Table B). In contrast, the proportion of off-budget donor expenditures began to reach 40 percent or more since 2002. This trend is expected to continue, fueled largely by the continuing large inflow resources from global and bilateral disease initiatives which are typically implemented as parallel financing arrangements by donors themselves (as financ- ing agents), or through grants with NGOs, community-based organizations (CBOs), or hired professional contractors. Note that the tables and figures refer only for health, nutri- tion, and population expenditures, and excludes HIV/AIDS expenditures that are not directly related to health service provision. If HIV/AIDS expenditures are included, the proportion of off-budget spending increases even more dramatically, as these are coursed largely through NGOs and CBOs. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 24 24 A World Bank Country Study Figure 16. In-Budget versus Off-Budget Donor Health Expenditures, 2000­04 (percent) 83.9 90 79 80 70 58.2 60 58.8 60 50 41.8 40 41.2 40 30 16.1 21 20 10 0 2000 2001 2002 2003 2004 In-budget donor HE Off-budget donor HE Employers' Health Expenditures Until 2000, the Zambia Consolidated Copper Mines (ZCCM) was a major source of employer-provided health benefits. With its privatization, its health facilities have been taken over by MOH or privatized. Other employers have since loomed as important source of health financing so that by 2004, employers already commanded expenses amounting to ZK128.9 million (Annex Table C) representing 17 percent of all private health expendi- tures, and 7 percent of total health expenditures in the country. Household Health Expenditures These are largely out-of-pocket expenditures incurred by individuals seeking care at GRZ or private facilities (paid as user fees), or purchasing drugs and other medicaments from drug stores, or paying traditional healers. They also include contributions to health insur- ance schemes. Household health expenditures have been growing rapidly since the late 1990s, impelled in part by a widening choice of private providers and over-the-counter drug purchases. Recent data show household health spending accounts for 71 percent of all private HE, and 28 percent of THE. Health Expenditures by Service Provision and Administration Health expenditures increasingly are being devoted to administration. Disaggregating health accounting data into service provision and administration reveals the increasing "bureaucratization" of health services. Expenditures on administration have quickly jumped from 14.8 percent share in 2001 to 30.8 percent in 2004 (Figure 17, based on Annex Table D), that is, nearly a third of all health expenditures in Zambia are being used for activities not directly related to health service provision. Instead of achieving economies of WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 25 Zambia Health Sector Public Expenditure Review 25 Figure 17. Total Health Expenditures Going to Service Provision and Administration, 2001­2004 (percent) 90 80 70 60 50 85.2 40 75.8 74.8 68 30 20 30.8 10 21.9 23.3 14.8 0 2001 2002 2003 2004 Service provision Administration Note: Rest-of-the-world health expenditures are excluded. scale on administration, the fragmentation of financing and service delivery appears to have increased administrative costs. The increasing administrative-intensity during this period can be explained by two major factors: (a) the separation of CBOH and MOH turned out to be duplicative and costly; and (b) donors' sponsorships of a plethora of NGOs and CBOs and local and international contractors entailed administrative and trans- actions costs not seen previously. A review of health expenditures on service provision shows the following (see also Annex Table E): I Theroleofgovernmentfinancinginserviceprovisionhasbeenhighlyvariable,at least in terms of shares to total expenditures going to health services. The GRZ share to service provision varied from 40 percent in 1999 to 75 percent in 2002, until it stabilized back to around 41­43 percent in later years. Also, GRZ has massively increased the resources going to district health services. The share of districts to total GRZ spending on health service provision rose from 33 percent in 1999 to 65 percent in 2004. If one excludes Defense Force and ZCCM hospitals, districts' share to GRZ health service provision expenditures rises from 45 percent in 1999 to 66 percent in 2004. I HE going to private for-profit hospitals has risen dramatically from 18 percent in the late 1990s to 30 percent by 2004. Similarly, household purchases from drug stores and chemists are becoming dominant. Their share to total HE on service pro- vision spiked to 23­25 percent in the 1999­2000, but have since stabilized to around 13 percent. I Mission (non-profit) hospitals have far smaller role in generating funding than conventional wisdom suggests. Their share to total HE is at most 1 percent, derived from their fee operations. On the other hand, traditional healers' share in service provision is big; it swung from 8­15 percent during the period. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 26 26 A World Bank Country Study A review of health expenditures on administration reveals that GRZ health administra- tion has grown big, accounting for 46 percent of total administration costs of the sector by 2004. Even more so, administrative costs of donor-supported vertical projects--as reflected in the items "other providers' health administration" and "other related services and institutions"--incurred 50 percent of total administrative costs in the sector. The NHA methodology records these institutions under administration, as they basically receive funding from donors which they then use to support existing health facilities. However, some of them are beginning to implement their own health services, so in fact, not all of their expenditures should be treated as administrative. Health Expenditure Patterns of Donors, MOH, and Households How Do Donors Spend Their Resources? Annex Table F provides data on the recipients of donor resources in the health sector. Although only three years of NHA provide consistent data on detailed expenditures, they do show clear trends. I MOH remains the biggest recipient of donor funds, with some small amounts going to other public institutions. Together they captured 57.4 percent of all donor health expenditures in the period 2002­04. I Other traditional service providers such as mission hospitals, private-for-profit hospitals, ambulatory health care providers, and drug stores and chemists together captured a small proportion (2.3 percent) of donor health expenditures. Indeed, as individual categories, each received minuscule shares of less than 1 percent. I Nontraditional providers--labeled under NHA methodology as "other public health providers," "other providers' health administration," and "other related ser- vices and institutions,"--together received a large share (33.3 percent) of donor health expenditures. In common parlance, these are the nongovernment organiza- tions (NGOs) and community-based organizations (CBOs) involved in a wide range of advocacy, social mobilization, information and education, and service provision. For a significant part of their operations, these NGOs/CBOs also make use of GRZ, mission, and sometimes for-profit health facilities. I Research activities have rapidly grown, and now account for 3 percent of donor health expenditures. However, training is given a very short shrift, with only 0.3 percent of donor expenditures devoted to it, despite Zambia's current human resource crisis in the health sector. How Does MOH/CBOH Spend its Resources? Since 2001, the percentage of MOH/CBOH health expenditures going to service provision has gone down while the percentage going to administration has gone up sharply (Figure 18, derived from Annex Table G). By 2004, administrative expenditures accounted for 37.6 per- cent of total MOH/CBOH expenditures. Among administrative costs, general administra- tion at MOH/CBOH consistently expanded since 2001, rising in share from 63.4 percent to WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 27 Zambia Health Sector Public Expenditure Review 27 Figure 18. MOH/CBOH Health Expenditures Going to Service Provision and Administration, 2001­04 (percent) 90 80 70 60 50 40 79.6 69.8 68.3 30 62.4 20 30.2 31.7 37.6 10 20.4 0 2001 2002 2003 2004 Administration Service provision 93.6 percent of total administration in 2004. As mentioned previously, the creation of to separate entities (MOH and CBOH) during this period largely accounts for this adminis- trative intensity. Administrative spending on "other providers of health administration" and "other related services" revolving around NGOs/CBOs rose markedly in the early 2000s, garner- ing as much as 30.5 percent of MOH/CBOH administrative expenditures in 2001. This type of spending has since declined at MOH as many of transactions and activities probably migrated to donors' auspices. Expenditures on training and research together remained stable at a nominal value of ZK5-6 billion a year. However, in terms of share to total administrative expenditures, research and training have gradually eroded from a combined 8.2 percent share in 1999 to only 1.6 percent in 2004. In addition, the composition has changed drastically as training expenditures declined acutely from ZK3.9 billion to an insignificant ZK433 million in just six years, while research expenditures more than doubled during the same period. In fact, research now exceeds training expenditures almost by a factor of 10. MOH spending by level of care (Figure 19) reflects a dramatic re-allocation away from tertiary hospitals (which declined in share of service provision from 40.3 percent to 19.2 per- cent in six years) towards district and ambulatory health services (which rose in importance from 49.7 percent of service provision to 67.3 percent). The share of secondary hospitals to total service provision remained fairly stable at 7­10 percent a year. How Do Households Spend Their Resources? Despite the overwhelming extent of government health providers, they receive a surpris- ingly small share of household resources for health (Annex Table H). From 1999­2004, only 5.4 percent of household health expenditures went to GRZ health facilities, mostly to tertiary hospitals (about 2­3 percent yearly). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 28 28 A World Bank Country Study Figure 19. MOH/CBOH Health Expenditures by Level of Care, 1999­2004 (percent) 2004 19.2 8.3 69.5 3 2003 20.3 10.4 67.3 2 2002 20.3 9.8 67.6 2.5 2001 17.6 10.3 70.1 2 2000 32.3 9.4 56 2.3 1999 40.3 7 49.7 3 0% 20% 40% 60% 80% 100% Tertiary hospitals Secondary hospitals District + ambulatory Mission hosp + othrs Households use most of their resources on the private health sector, notably on pri- vate for-profit hospitals (which received 44.2 percent of household health expenditures in 2004) and drug stores and chemists (29.1 percent). The large household spending on drugs reflects MOH facilities' inability to dispense most of the prescribed drugs, forcing house- holds to find them elsewhere. Other providers such as mission hospitals and other public health providers capture an insignificant share of household health expenditures (at most 2 percent). Over the period 1999­2004, private providers together received a whopping 71.4 percent of household health spending. Traditionalhealersalsofigureprominentlyamonghouseholdsseeingcare.Inthesixyears since 1999, they received 21.1 percent of all household spending on health. Indeed, household spending on traditional healers is 4 times the size of their spending on GRZ facilities. Financing of Service Providers What Sources of Financing Do Health Facilities Rely On? Table 9 shows the types of health facilities and their sources of financing over a three-year period, 2002­04. The matrix shows the continuing simplicity of Zambia's health service payment system, with many zeroes in the cells, reflecting lack of contractual and financial obligations between potential payors and health facilities. For instance, employers have only established contractual arrangements with for-profit hospitals, and none with any MOH health facilities. While all MOH facilities are all heavily reliant on the MOH budget, as expected, they become less so the higher the level of care. Instead, they become gradually more reliant on other forms of payment (user fees from households, resources from other private sources), as shown in Figure 20. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 29 Zambia Health Sector Public Expenditure Review 29 Table 9. Total Financing of Health Facilities, 2002­04 (ZK Billion) Facility MOH/ Other House- Emplo- Other Type CBOHa GRZ Holds yers Private Donors Total MOH tertiary Hospitals 252.4 0 33.5 0 38.3 0 324.1 MOH secondary Hospitals 120.5 0 8.0 0 10.5 0 139.5 MOH district Hospitals 808.7 0 17.8 0 48.4 0 874.9 For-profit hospitals 2.3 0 490.4 281.2 23.5 11.0 808.3 Mission hospitals 19.1 0 0.6 0 4.5 8.8 32.6 Ambulatory Care 57.7 0 0.5 0 1.7 14.6 74.5 Other Providers 37.9 0 13.3 0 7.8 73.6 132.6 Other institutions 5.9 34.1 3.4 0 1.0 293.9 338.2 aThis includes in-budget resources of donors coursed through the Ministry, e.g., the basket funds. All other off-budget (parallel financing) activities of donors are reflected under "Donors" in the table. User fees from households are a minor source of revenues for primary and secondary MOHfacilities(2­6percent),andevenlesssoformissionfacilities(about1percent).However, they are quite important for tertiary MOH facilities (10 percent of their revenues come from fees). In 2006, GRZ abolished user fees for primary health services and medical care services in rural areas. In contrast to MOH and mission health facilities, for-profit hospitals rely almost exclu- sive on user fees (60.7 percent of their resources) and employer payments (34.8 percent). Under a Memorandum of Understanding between the Churches Health Association of Zambia (CHAZ) and GRZ, government is supposed to provide only 75 percent of the full funding requirements of mission facilities; they're supposed to seek additional fund- ing elsewhere. It is difficult to verify whether mission facilities receive the amount they Figure 20. Financing of Health Facilities by Source, 2002­04 (percent) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% MOH 3* hosp 77.9 10.3 0 11.8 MOH 2* hosp 86.7 5.7 7.6 MOH district hosp 92.4 2 5.5 For-profit hosp 0.3 60.7 34.8 2.9 1.4 Mission hosp 58 1.9 13.5 26.6 Ambulatory care 77.5 0.7 2.3 19.6 MOH/CBOH Households Employers Other private Donors WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 30 30 A World Bank Country Study request from government, nor whether this is sufficient. It is clear that the MOH subven- tion only accounts for 58.0 percent of their total resources, and that they also rely on pri- vate financing (13.5 percent), donor parallel financing (26.6 percent), and user fees (a paltry 2 percent, similar to the rate generated by district hospitals). The "other providers" and "other institutions" includes nontraditional entities such as NGOs, CBOs, contractors, and the like which have come into the scene in full force espe- cially with the advent of global and bilateral disease initiatives. They rely heavily on paral- lel financing from donors. Uses of Health Expenditures Cost structures derived from the health expenditures of various types and levels of health facilities reveal interesting differences between MOH, mission, and for-profit health facil- ities. The MOH expenditure ratios below should not be treated as ideal or normative pat- terns but rather as actual cost structures based on the existing state of affairs, which is bedeviled with problems. For instance, the high level of staff vacancies, the shortage of drugs and other supplies, deficits in transport and medical equipment, and the relatively poor state of health infrastructure requires the exercise of caution in using the above ratios for planning purposes. However, the cost structures for mission and for-profit facilities do provide interesting comparators by which to benchmark the MOH cost ratios. Among MOH health facilities, secondary hospitals appear to be the most labor-intensive, incurring an average of 64 percent of expenditures on personal emoluments (Figure 21, derived from Annex Table I). Contrary to expectation, tertiary hospitals are less labor intensive than secondary hospitals, using up only 61 percent of expenditures on employee Figure 21. Health Expenditures of MOH, Mission, and For-Profit Facilities by Inputs, 1999­2004 (percent) For-profit 44 9 28 2 12 5 Mission 48 0 21 4 22 5 MOH 3* 61 0 21 2 14 2 MOH 2* 65 0 19 3 11 2 MOH 1* 50 0 21 7 18 4 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% PE Administration Drugs, etc. Transport Other rec. Capital Note: MOH and mission facilities' ratios are averaged from 1999­2004 data while those of for-profit facilities are averaged from 2002­04 data. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 31 Zambia Health Sector Public Expenditure Review 31 costs. Primary facilities are the least labor intensive, with only 50 percent of their expendi- tures going to labor costs. The large proportion of costs going to labor in MOH facilities is somewhat surprising, given the many vacancies in facilities as a reflection of the human resource crisis. However, such large share of labor could only reflect the increase in the wage bill which began to increase sharply in 2000, and again in 2003 (which included generous housing allowances). Compared to their MOH counterparts, non-profit mission facilities have a lower pro- portion of expenses incurred on labor (48 percent). It is not clear if this is because they have less staff, they have lower salaries and benefits, or the staff are simply more efficient. For-profit facilities have the lowest proportion of personnel costs among the different types of facilities (44 percent), if administrative costs are excluded from labor costs. But even if administrative costs of about 9 percent are included to PE, labor costs would only be around 52 percent of total costs, compared to 61 percent for secondary and 64 percent of tertiary MOH hospitals. This large difference in the labor cost structure between for-profit and MOH health facilities should encourage government to investigate why, despite the large proportion of MOH resources devoted to labor, the human resource crisis persists. Consistent across levels of care in MOH and mission facilities, drugs and other con- sumable supplies account for around a fifth (19­21 percent) of expenditures. However, for-profit facilities spend a greater proportion on drugs (28 percent), mainly because they spend a smaller proportion on labor, as discussed above. The proportion of expenditures incurred on transport varies inversely with the level of care. MOH primary facilities use up 7 percent of their expenditures on transport, com- pared to 3 percent for secondary and 2 percent for tertiary facilities. Mission facilities, which are mostly of the secondary level, incur 4 percent of their expenditures on transport, while for-profit facilities use up 2 percent on transport. The relative "transport intensity" of primary care may have to do with the magnitude of outreach they are expected to per- form, and the geographic distances they need to cover. Other recurrent expenses (ORE) represent a range of inputs, activities, and transac- tions that are difficult to summarize. Among MOH facilities, primary clinics tend to use the largest proportion of expenditures on ORE (a hefty 18 percent), compared to 11 per- cent for secondary and 14 percent for tertiary hospitals, and 12 percent for for-profit hos- pitals. Mission facilities are a clear outlier, incurring as much as 22 percent on ORE. Capital expenditures are given a short shrift across all types of MOH facilities; the pro- portion spent on repairs, construction, and the like is only 2­4 percent. This low level of spending is reflected in the poor state of repair of many these facilities. In contrast, both mission and for-profit facilities spend a greater proportion (5 percent) on capital expenses. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 32 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 33 CHAPTER 4 Budget Allocation, Release, and Spending: Results of the Public Expenditure Tracking Component of the PET/QSDS, 2005­06 his chapter presents the findings of the pubic expenditure tracking in the health T sector that was undertaken using FY05 financial data. The endeavor covered all major resource flows to the health facility including the GRZ budget, basket funds, vertical project funds, and internally generated funds. The analysis focused on amounts allocated, predictability of releases, differences in allocation and released or received amounts, and uses of received resources. The study proceeded optimistically at the cen- tral and facility levels, but it soon became obvious an information "black box" exists around the District Health Management Team (DHMT), which is the weak link in the fis- cal information chain. Thus, while it is easy to track resources from MOH to districts, and finally how the resources are used in health facilities, what happens to these resources while they are received and allocated by the DHMT to the different health centers and dis- trict hospitals is quite opaque, and the "information fog" is hard to clear up because of lack of information, lack of standardized terminology, and related issues. This chapter is organized as follows. The first section describes the flow of funds in the health sector and the resource envelope for FY05 and FY06. The second section dis- cusses the GRZ budget process and allocation principles. The third section analyzes MOH allocations versus actual releases by various types of expenditure items. The fourth section focuses on the role of district health management boards (DHBs) and DHMTs that are mandated to allocate the district grants (non-wage/non-drug resources) to lower-level facilities under them (health centers and district hospitals). The fifth section focuses on donors' vertical funds that reach the facilities. The sixth section focuses on the facilities' own internally generated funds. The seventh section builds a comprehen- sive picture of resource availability at the level of typical health facilities (RHC, UHC, and hospital). 33 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 34 34 A World Bank Country Study Funds Flow and Resource Envelope The flow of funds in Zambia's health sector is a complicated and fragmented system, where salaries,drugs,andotherrecurrentexpendituresaredisbursedseparatelybydifferentagencies. MOFNP provides salaries directly to the Provincial Health Offices, which then remit these to their health centers, first-level district hospitals, and second- and third-level hos- pitals (Figure 22). Figure 22. Flow of Funds and Other Resources in Zambia's Government Health Sector Ministry of Finance & National Planning Donors/ Cooperating Partners Basket F Nonwage MOH- S u Budget managed a n projects l d V a s e r r i Medical t e Stores, Ltd. i s Ministry of Health Funds c a Non- D wage Non- r D l u r Budget wage g u (District Budget P s g Grants) r Provinces s o Salaries Vertical j 2nd & 3rd Level proj. e Hospitals c S t a s l District Health Management Teams a r i Nonwage Drugs e Expend- s itures Health Centers and District Hospitals User Other fees IGFs Income-generating projects, Households Communities, Businesses WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 35 Zambia Health Sector Public Expenditure Review 35 MOFNP provides the budget for other recurrent expenditures directly to MOH. Cooperating partners (as donors are called in Zambia) also provide the budget for other recurrent expenditures through their basket funds to MOH. These basket funds are allo- cated in tandem with GRZ funds, and are managed and monitored closely within the MOH framework. For drugs, MOH allocates specific funding for the Medical Stores Ltd. (MSL), the paras- tatal in charge of drug procurement and distribution. MSL distributes drugs to DHMTs largely through a supply-driven ("drug-kit" system), and to second- and third-level hospi- tals through a combination of "drug-kit" and demand-driven requisition system. MOHthenprovidesrunningcosts(usingGRZandbasketfunds)toDHMTsandsecond- and third-level hospitals. The DHMTs, in turn, provide running costs and drugs (which they obtain from MSL) to the health centers and district hospitals under them. Additional resources come from the following: I Separate projects implemented by MOH, and following GRZ procurement and fiduciary systems or donor-determined systems. If implemented using GRZ sys- tems, these flows follow the usual channels as described above. I Vertical projects implemented by donors or their financing agents or contractors which, for the most part, lie outside the MOH procurement and fiduciary systems, but which may be implemented by MOH facilities. For the most part, these pro- vide in-kind support, but as will be shown in the survey, they are increasingly pro- viding cash support directly to facilities in ways that are still not well-understood or documented. I The internally-generated funds (IGFs) of health facilities, including user fee rev- enues, training revenues, community donations, revenues from income-generating projects, and the like. These are not well understood or documented, but they are quite significant, as will be shown in the survey results. The resource envelope in Zambia's health sector was about ZK899.8 billion for FY05 and about ZK1,080.8 billion for FY06. The details are shown in Table 10. Note that these amounts only include monies where the GRZ/MOH is the financing agent; it excludes ver- tical projects. It also excludes IGFs, although formally, the MOH health facility is the financing agent for them. The FY05 and FY06 resource envelopes show the following patterns.2 There has been an 11 percent increase in flexible health funding from ZK612 billion in FY05 to ZK681 bil- lion in FY06. The gainers in FY06 were MOH nonwage expenditures (up by 31 percent), third-level and second-level hospitals (up 8 percent for wage component and 55­59 per- cent for nonwage component), and grants and other payments (up by 41 percent). The losers were expenditures on district nonwage expenditures (down by 13 percent) and dis- trict drug expenditures (down by 34 percent). Indeed, drug allocations declined all levels of care, but particularly in districts. Changes by types of inputs shows a massive increase in capital expenditures (39 percent), a modest increase for wage and non-wage expendi- tures (8 and 9 percent respectively), and a 15 percent decline in drug expenditures. 2. This paragraph draws on the analysis done by Yates (2006). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 36 36 A World Bank Country Study Table 10. Health Sector Resource Envelope, FY05 and FY06 (ZK Billion) GRZ + Basket Funds Non- Subtotal Donor Headings Wage wage Drugs Capital GRZ + Basket Projects Total FY05 (US$ 1 = ZK4,500) MOH HQ 4.4 51.6 29.7 0.1 85.8 288.1 373.9 3rd-level hospitals 57.5 13.2 4.2 -- 74.9 -- 74.9 2nd-level hospitals 41.2 13.5 4.2 -- 58.9 -- 58.9 Districts 164.4 142.7 12.7 24.1 343.9 -- 343.9 Training institutions 3.8 15.8 -- 2.0 21.6 -- 21.6 Grants & other payments -- 26.6 -- -- 26.6 -- 26.6 Total 271.3 263.4 50.8 26.2 611.7 288.1 899.8 FY06 (US$ 1 = ZK3,500) MOH HQ 4.8 67.4 29.1 0.3 101.6 399.8 501.4 3rd-level hospitals 62.2 20.4 2.8 17.5 102.9 -- 102.9 2nd-level hospitals 44.7 21.5 2.8 -- 69.0 -- 69.0 Districts 178.1 123.8 8.4 36.9 347.2 -- 347.2 Training institutions 4.1 16.0 -- 2.8 22.9 -- 22.9 Grants & other payments -- 37.4 -- -- 37.4 -- 37.4 Total 293.9 286.5 43.1 57.5 681.0 399.8 1,080.8 The resource envelope is still less than the financing required to meet the health MDGs. Several costing studies were recalculated in the IMF "Country Case Study Zambia" (2006) to present all projections in terms of the same baseline GDP and GDP growth used in the FNDP. The results, shown in Table 11, indicate that the annual costs of providing basic care and reaching the health MDGs is in the range of US$30­32 per capita, or around Table 11. Range of Cost Estimates to Meet the Health MDGs in Zambia Total Cost Average Per Capita 2005­2015 Annual Cost Annual % of GDP in US$ in US$ Cost (FDNP Basis Million Million in US$ Projections) CBOH (2004) + Kombe & 4,444.9 404.0 31.1 2.9 Smith (2003) costing Mphuka (2005) costing 4.403.6 400.3 30.8 2.9 CBOH (2004) + NAC (2006) 5,369.1 488.0 37.5 3.5 costing FDNP Core Costs 4,578.8 416.2 32.0 3.0 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 37 Zambia Health Sector Public Expenditure Review 37 3 percent of GDP.3 If the NAC costing on HIV/AIDS interventions are used, the figures rise to US$37.50 per capita health expenditure, and 3.5 percent of GDP. Recent experience shows that GRZ allocates only about 2 percent of GDP to health, and even projections in the Medium-Term Expenditure Framework do not see the per capita health expenditure and health/GDP ratios rising dramatically anytime. These indicate: the need for closer dia- logue between the MOFNP and MOH on what the former allocates to the sector and what the latter sees as the level of resources needed in the sector; and the need for stronger pri- ority setting if the ideal levels of spending and health/GDP ratios cannot be achieved. Important caveats must be kept in mind in interpreting the above resource require- ments. The costing models do not take full account of the likely costs of addressing the human resource crisis, including needed actions on adding posts, filling posts, increasing wages and/or consolidating benefits, and training. Also, the costing models deal mainly with recurrent cost requirements, but as will be shown in the PET/QSDS survey, there are severe deficits in infrastructure, utilities, equipment, and systems as well. MOH Budget Process and Allocation Formula The planning and budgeting process was formalized by then-CBOH in 2005, and calls for bottom-up approach in priority setting. According to this formal process, the budget preparation and schedule within the MOH are as follows. Around July, MOFNP issues the overall budget guidelines to MOH, which then transmits them to Provincial Health Offices (PHOs) and District Health Boards (DHBs) and District Health Management Teams (DHMTs). In late July, the DHMTs schedule a briefing meeting with their respective com- munity representatives. The communities discuss their health priorities up until early August, when they collect their proposals and present these draft requests in a meeting at their Health Centers. In early September, all the Health Center requests are aggregated and presented in a meeting with the DHMT. Feedback is obtained, and incorporated into the draft, which gets finalized by the DHMT into an action plan by early October. In late October/ early November, DHMTs present their Action Plans to the DHBs. In early November, DHBs submit their Action Plans to the PHO, which reviews it and provides feedback. Any revi- sions are incorporated in late November. In early December, the Final Plan is submitted back to the PHO. In December, the PHO submits the Final Plan to the MOH. The budget allocation formulae for health combine equity-enhancing and status- quo-maintaining elements whose overall effects may prevent achievement of greater geographic equity. I Districts:DHBs/DHMTsreceiveadirectgrantfromMOHfornon-wageandnon- drug recurrent spending. The MOH abandoned incremental budgeting for this component of its budget in the mid-1990s when it adopted per-capita allocations to improve equity. In January 2004, further refinement was introduced in the formula by the inclusion of a material-deprivation index (Kabaso and Tembo, 2004). Thus, 3. The FY06 resource envelope of ZK1,080.8 billion translates to per capita health expenditures of about US$26.85, using an exchange rate of ZK3,500 per US$1. However, if the "pre-appreciation" 2005 exchange rate of ZK4,500 is used, this figure goes down to US$20.90. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 38 38 A World Bank Country Study the formula for district grants now consist of four variables: population weighted for price of fuel (to reflect the large distances that primary health care workers need to cover), population weighted for epidemics, population weighted against popu- lation density (it is cheaper to move around in urban areas), and presence of a bank. Note that the district grant formula only affected 23 percent of the total GRZ + Basket Budget in FY05 and only 18 percent in FY06. So only a rather small percentage of the budget is subject to this equity-enhancing feature. I Second- and third-level hospitals: Allocations to this level of care continue to be based on the health facilities' number of beds (historical budgeting). If the hospi- tals were constructed not based on some equity considerations to start with--and some of them weren't, especially the mine hospitals which are now under MOH-- then the continued use of historical budgeting ignores present-day equity consid- erations. Indeed, there are quite a few instances where mine and government hospitals co-exist in the same neighborhood (for example, Kabwe). There have been discussions about hospital allocation especially in light of the MOH's continuing burden of maintaining the mine hospitals, but in practice, for- mal criteria for second- and third-level allocations that reflect equity and efficiency aspects have not been developed. The planned introduction of a social health insurance program for civil servants should make this an urgent concern. Note also that this "status-quo-maintaining" part of the budget (wage and nonwage allocation for second- and third-level hospitals) ate up 21 percent of the GRZ + Basket Budget in FY05 and 24 percent in FY06. The persistence of historical budgeting in hospital allocation means that Zambia has not been able to reduce provincial inequities. Figure 23 shows the per capita values of MOH Figure 23. Per Capita MOH Released + Supplementary Funding, by Province, FY05 (ZK) Lusaka 42,540 42,858 Copperbelt 39,532 40,279 Southern 28,748 29,219 Zambia 26,024 26,369 Central 21,128 21,227 Northwestern 20,838 20,840 Western 20,075 20,303 Eastern 16,998 17,369 Northern 15,265 15,551 Luapula 15,271 15,462 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 Total Minus training institutions WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 39 Zambia Health Sector Public Expenditure Review 39 releases and supplemental funding to provinces in FY05 (total releases as well as total minus releases to training institutions in the province). Three provinces (Lusaka, Copperbelt, and Southern) have per capita releases twice or more than the poorest provinces of Luapula, Northern,andEasternprovinces.Indeed,LusakaandCopperbeltprovinceshavealmostthrice the values of Luapula and Northern provinces. Note also that the best-resourced provinces in health also happen to be the richest provinces in terms of household income, and the most urbanized, so the MOH allocation merely perpetuates and exacerbates the underlying socioe- conomicinequityinsteadofredressingit.Thethreeleastresourcedprovinceshaveacombined 2005 population of 3.9 million, representing 35 percent of Zambians. The geographic budget allocation does not do service to the burden of common dis- eases and indeed runs counter to disease incidence. Figure 24 shows budget release per capita for each province matched against the incidence rates for Zambia's two common diseases, malaria and pneumonia. The trendlines, especially for malaria (Zambia's num- ber one leading cause of morbidity and cause of visit to health facilities), graphically depict the inconsistency between resource allocation and need. The same inconsistency can also be shown using charts of service coverage indicators, such as full immunization, institu- tional delivery, and postnatal care. Resource planning is not universally undertaken by DHMTs to inform their resource requests. While MOH has long emphasized district planning, a cursory examination of the 21 districts included in the survey showed that at least three DHMTs (Kaoma and Shangomo in Western Province and Lusaka Urban) did not undertake resource planning of the type that would inform accurate resource needs. These districts merely plugged in the same amounts for different health sub-programs into their budgets. Nor is this practice uncommon. Not all DHMTs in the survey provide the full complement of necessary public health services. Those with "missing" health services (no budget was allocated to them, nor were there any amounts received) are as follows. Figure 24. Budget Release Per Capita Versus Incidence of Malaria and Under-5 Pneumonia, 2005 (ZK) 45,000 500 35,000 400 25,000 300 15,000 200 5,000 100 ­5,000 n al n n n n 0 C'belt ther Lusaka Zambia Centr ester Easter Souther N'wester W Nor Luapula Budget release per capita (ZK) Malaria incidence per 1,000 pop'n U5 pneumonia incidence per 1,000 Linear (Malaria incidence per 1,000 pop'n) Linear (U5 pneumonia incidence per 1,000) WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 40 40 A World Bank Country Study Figure 25. Cumulative Budgeted Versus Released Amounts of Total MOH Budget, by Month, FY05 (ZK Million) 500000 400000 300000 200000 100000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Budgeted Released I Firstreferralservices--Mufulira; I Childhealth--Mufulira; I HIV/AIDS--LusakaUrban; I Malaria--Mufulira; I Maternalhealth--ChingolaandMufulira; I Waterandsanitation--Chingola; I Mental health--Lusaka Urban, Chingola, Luangwa, Kasama, Chinsali, Chilubi, Mpika, Nakonde, Livingstone, Kalomo, Namwala, Siavonga, Sinazongwe, Mongu, and Sesheke; and I Oralhealth--Ndola,Chingola,Mufulira,Luangwa,Chilubi,andSesheke. MOH Allocations versus Release of Resources The "cash budgeting" system in place since 1993 to control government spending and inflation has had a checkered record, but seems to have worked well in 2005, the focus year of the PET/QSDS survey. The system was intended to keep actual spending in line with rev- enues, and this seemed to have worked well in the early years. However, problems occurred in the first half of 2000s as the government was prone to take on additional expenditure commitments during the year, resulting in large budget overruns and payments arrears. ThismacroproblemmanifestedinlineministriessuchasMOHthroughexpendituresqueezes and unpredictability of monthly cash releases, especially on non-wage and capital spend- ing. An IMF assessment (2006) indicates that the cash budget system was in effective oper- ation since 2005, but that adverse shocks occurred again in 2006 due, among others, to the shortfalls in kwacha-denominated donor support because of the large exchange rate appre- ciation (ZK4,500 to ZK3,500) and election-related spending not included in the budget, both of which adversely affected the MOH. Overall, MOH budget releases were made in a predictable and timely fashion in FY05.4 This is shown in Figure 25. However, there were significant differences in the pattern of 4. The budget release data for this section drew from the work of Mumbwali (2006). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 41 Zambia Health Sector Public Expenditure Review 41 Figure 26. Cumulative Budgeted Versus Released Amounts for Personal Emoluments, by Month, FY05 (ZK Million) 300000 250000 200000 150000 100000 50000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Budgeted Released releases across types of budget items (see Figures 26 to 30). Personal emoluments were released very predictably. RDCs were released quite predictably until May, and then they dramatically slackened. Releases for grants began to exceed budgeted amounts starting May, and it continued on this trajectory for the rest of the year. Budget releases for essen- tial drugs were also relatively reliable until May, when they also slowed down. Budget releases for capital expenditures were the most erratic; these budgets were not released until well into the year. Capital budget releases picked up in May but then slackened through the rest of the year. Subsequent analysis for the early part of 2006 also showed similar trends (DfID, 2006). Provinces and DHMTs received their allocations in full. MOH budget releases exceeded budget allocations in FY05 for all provinces and districts and provinces included in the survey, as shown in Figure 31 and Figure 32. Figure 27. Cumulative Budgeted Versus Released Amounts for District Grants, by Month, FY05 (ZK Million) 80000 70000 60000 50000 40000 30000 20000 10000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Budgeted Released WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 42 42 A World Bank Country Study Figure 28. Cumulative Budgeted Versus Released Amounts for RDCs, by Month, 2005 (ZK Million) 60000 50000 40000 30000 20000 10000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Budgeted Released Figure 29. Cumulative Budgeted Versus Released Amounts for Essential Drugs, by Month, FY05 (ZK Million) 15000 10000 5000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Budgeted Released Figure 30. Cumulative Budgeted Versus Released Amounts for Capital Expenditures, by Month, FY05 (ZK Million) 18000 16000 14000 12000 10000 8000 6000 4000 2000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Budgeted Released WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 43 Zambia Health Sector Public Expenditure Review 43 Figure 31. MOH Allocation Versus Releases to Provinces Included in the Survey, FY05 (ZK Million) 35,044.50 Southern 38,871.90 20,689.40 Northern 22,761.60 15,178.50 Western 16,983.30 55,431.80 Copperbelt 66,272.50 61,353.80 Lusaka 70,865.30 0.00 10,000.00 20,000.00 30,000.00 40,000.00 50,000.00 60,000.00 70,000.00 80,000.00 Release Allocation Figure 32. MOH Allocation Versus Releases to DHMTs Included in the Survey, FY05 (ZK Million) Kafue 140.7 137.1 Lusaka Urban 570.2 559.1 Luangwa 35.4 34.6 Mufulira 94.4 87.8 Ndola 330.1 323.7 Chingola 103.2 100.7 Mpongwe 346.8 321.9 Mongu 271.4 265.9 Kaoma 306.9 298.7 Shangombo 147.4 144.4 Sesheke 141.2 138.3 Kasama 283.3 279.5 Mpika 283.2 270.7 Nakonde 135.2 132.4 Chinsali 250.6 245.5 Chilubi 135.6 132.8 Sinazongwe 148.2 128.2 Siavonga 92.6 90.2 Kalomo 294.8 286.3 Namwala 150.3 147.2 Livingstone 46.4 45.2 0 100 200 300 400 500 600 MOH Allocation MOH Releases WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 44 44 A World Bank Country Study Allocation and Releases of District Grants Among the different segments of the PET/QSDS exercise, this section provided the least reliable data. Survey response rates were uniformly low across the different types of health facilities. Facility in-charges often did not have solid information, and many did not have a good grasp of the financial figures' orders of magnitude. Even with pretest- ing, it is clear that the terminologies used were interpreted variously (planned, allocated, released, and so forth) Moreover, there is no clear financial reporting especially at the lower levels. When tallied, survey responses revealed wide variances. Confirming and tri- angulating the survey results with other existing sources of information showed major discrepancies, and in many cases, there were no corroborating data to begin with. In any case, we report what is "salvageable" with the caveat that caution must be exercised in interpreting the information. The timing of MOH releases on district grants is predictable. At least for FY05 and early FY06, district grants were provided on time. Indeed, by about the middle of the year in 2005 (May), the released amounts already exceeded the budgeted amounts, as shown in an earlier chart. More than a third of DHMTs reported delays in the release of district grants to health facilities. Funds are released on a monthly basis from DHMTs to the health facilities. How- ever, sometimes delays occur. Out of the 20 DHMTs queried, seven (or 35 percent) reported delays in releasing funds to the health centers and district hospitals under them, 50 percent did not, and 15 percent provided no categorical answer. However, if the monthly receipts of health centers were cumulated and charted as in Figure 33, the DHMT releases seemed to be predictable (assuming, of course, that health facilities got their dis- trict grants in full, which seems to be supported by less-than-robust data). To be sure, the response rate to this question is low. Also, the "actual" figures seem suspiciously in line with the "planned" figures that it is likely the difference between planned and actual receipts were misunderstood, or that some of the enumerators simply divided the annual figures by 12, instead of obtaining the actual monthly figures from facilities. A similar exercise was performed for hospitals and shown in Figure 34, and the same pattern was obtained. Figure 33. Cumulative Planned Versus Actual Facility Grants by DHMTs to RHCs, by Month, FY05 (ZK Million) 1000 800 600 400 200 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Planned Actual WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 45 Zambia Health Sector Public Expenditure Review 45 Figure 34. Cumulative Planned Versus Actual Hospital Grants, by Month, FY05 (ZK Million) 14000 12000 10000 8000 6000 4000 2000 0 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Planned Actual How much do DHMTs provide the health centers under their responsibility? Table 12 shows the amount of GRZ budget provided by DHMTs to their health centers, as gathered from the survey. The sizes of health centers, as inferred from the resources they provided, vary greatly. Leaving out the clear outlier districts, a typical health center would receive about ZK35 million to ZK90 million (about US$8,000­20,000). Four out of five health centers received the planned amount of resources that were allocated to them by their respective DHMTs. This issue was explored using two ques- tions, the first requiring a simple yes/no answer from the in-charge, and the second asking specific financial figures. Recall that in general, 2005 was a year when the health sector and most of the health programs and facilities received releases more than their original budgetary allocations, as discussed earlier. This is borne out in this survey (Table 13), where 56 percent of health centers received equal to, and another 22 per- cent received more than, their annual budget allocation (a total of 78 percent for both categories). Thus, only 22 percent of health centers received amounts less than their allocations. Many health centers and district hospitals could not provide good data that would categorically show that they received the amounts released to them by DHMTs. About 84 percent of health centers reported receiving amounts of releases that were less than a quarter of what their respective DHMTs having paid them. Only 9 percent of the health centers reported having received amounts of releases within 5 percentage points (+ or ­) of what their respective DHMTs indicated as having released to them. Finally, about 7 percent of health centers reported receiving amounts more than what the DHMTs indi- cated releasing, but these may be outliers (caused by data reporting or entry errors) since the observed variations are too large (greater than 200 percent). Figure 35 recovers data that are "cleaner" than most for district hospitals, while Figure 36 does the same for health centers. It would seem from this small sample that 50 percent of the health cen- ters did receive their full allocations while another 50 percent received less than their full allocations. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 46 46 A World Bank Country Study Table 12. Amount of GRZ Budget Provided by DHMTs to Their Health Facilities, FY05 (ZK Million) No. of Total Amount Health of GRZ Budget Average Centers Provided to Amount per under Health Centers Health Centers Province DHMT DHMT (ZK Million) (ZK Million) Lusaka Kafue 6 211.9 35.3 Luangwa 4 37.6 9.4 Copperbelt Ndola 10 1,016.1 101.6 Chingola 6 463.3 77.2 Southern Livingstone 4 665.5 166.4 Namwala 4 392.9 98.2 Siavonga 5 31.9 6.4 Sinazongwe 4 8.8 2.2 Northern Kasama 7 211.6 30.2 Chinsali 6 430.4 71.7 Nakonde 4 857.2 214.3 Chilubi 4 366.2 91.6 Mpika 3 217.6 72.5 Western Sesheke 6 277.7 46.3 Shangombo 5 446.9 89.4 Mongu 11 510.8 46.4 Kaoma 5 210.4 42.1 Total 94 6,356.8 1,201.2 Average 6/district 373.9/district 70.7/HC HowdidhealthfacilitiesactuallyspendtheGRZresourcesandbasketfundstheyreceived? Table 14 shows the expenditure patterns of surveyed health facilities, by type of expenditures. The absence of banks in many districts necessitates an imprest system where delays sometimes occur. Only 16.2 of the health facilities surveyed had bank accounts: 11.4 percent Table 13. Health Centers' Reponses on Whether They Received an Amount More Than, Equal to, or Less Than Their Annual Allocations, 2006 RHCs UHCs All HCs Allocated vs. Received Amount (n = 76) (n = 43) (n = 119) Allocated > Received 22% 21% 22% Allocated = Received 57% 56% 56% Allocated < Received 21% 23% 22% WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 47 Zambia Health Sector Public Expenditure Review 47 Figure 35. Allocated Versus Received Budgets of Selected District Hospitals FY05 (ZK Million) Lewanika DH 432.4 Yeta DH 36.7 Chilonga DH 381.4 Zimba DH 210 0 100 200 300 400 500 600 700 800 900 Received Allocated among RHCs, 8.3 percent among UHCs, and 47.6 percent among hospitals. These low rates of bank-account ownership, even among hospitals which are located in urban areas, explains the need for an imprest system. Some 86 percent of RHCs, 89 percent of UHCs, and 88 percent of hospitals operate such a system (or 87 percent of all surveyed facilites). Of these facilities having an imprest, only 32 percent of RHCs, 30 percent of UHCs, and 40 percent of hospitals said they received their monthly imprest in a timely manner. Figure 36. Allocated Versus Received Budgets of Selected Health Centers, FY05 (ZK Million) Kahare 4.2 4.2 Nkole Mfumu 5 3 Siatwinda 10.1 10.1 Siavonga HA 6 3.5 Kalilo 4 8.4 Kabundi East 7.6 7.6 Nkwazi 12 5 Sathya Sai 0.7 0.7 Masala New 12 10 Ndeke 8.4 8.4 Mushili 1 1 Kaloko 8.4 7 Chipokota 1 1 Mwembeshi 21.7 12.8 0 5 10 15 20 25 Received Allocated WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 48 48 A World Bank Country Study Table 14. Uses of GRZ Budget and Basket Funds by Health Facilities, FY05 (percent) Expenditures RHC UHC Hospital All Personal emoluments 28.4 -- 38.0 29.7 Drugs & medical supplies 7.3 10.3 8.4 8.8 Non-medical supplies 4.1 -- 5.7 4.4 Workshops, conferences, etc. 7.5 24.1 16.0 17.2 Repairs & maintenance 8.8 21.7 0.3 5.1 Food & catering 6.6 18.0 2.3 5.7 Utilities 2.7 -- 4.1 3.2 Fuel & transport 11.6 -- 7.6 6.2 Payments for TA, consultancies 1.8 -- -- 0.1 General charges 2.7 16.7 8.9 10.1 Payments of debt -- -- -- -- Capital purchases 0.6 7.1 2.5 3.3 Total 100.0 100.0 100.2 100.0 This means that about two-thirds of health facilities operating on an imprest face delays in financing their operations. A quarter of the health facilities are indebted. Thirty-two health facilities (or 25 per- cent of those surveyed) reported having debts of various kinds in FY05 (Table 15). Most of the debts are for transport fuel and electricity and other utilities. More than half of the hos- pitals and more than a third of the UHCs had debts. Many of these health facilities, includ- ing 10 percent of district hospitals and 3­4 percent of health centers borrow from the DHMT to pay off part or the full amount of their debts. Table 15. Composition of Average Debt Per Facility, FY05 (ZK Thousands) Items RHC UHC Hospital All Composition of debts Electricity 52.2 2,205.1 11,813.1 4,301.8 Other utilities -- 14.3 2,264.7 643.2 Transport fuel 440.7 22,535.9 10,033.1 12,805.2 Drugs & medical supplies -- 1,799.4 1,924.8 1,313.4 Food supplies 287.2 26.8 4,202.1 1,274.4 Staff-related debts 55.6 3,767.5 143,597.6 42,050.7 Others 281.1 6,481.8 80,450.4 25,541.5 Total 1,116.8 36,830.7 254,285.8 87,930.3 Health facility borrowings from DHMT to pay off debts Median amount 281.1 2,205.1 10,033.1 4,301.8 Maximum amount 440.7 22,535.9 143,597.6 42,050.7 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 49 Zambia Health Sector Public Expenditure Review 49 Health Facilities' Management of Vertical Funds Despite the high visibility of vertical disease funding in the global and national scenes, their presence is not felt strongly by many peripheral GRZ health facilities. Only a tiny minor- ity of health centers reported having received cash or in-kind receipts from these donors/projects. Only 4 percent of health centers have received resources from the Global Fund, and only 1 percent have received support from GAVI, UNICEF, WHO, DANIDA, and AfDB. Strangely, although many of these projects have a public-health orientation, they are far more palpable in hospitals. For instance, 20 percent of hospitals report having obtained support from the Global Fund; 10 percent, from the World Bank; and 5 percent, from the UNICEF. Collectively, only 35 health facilities in the survey (19 RHCs, 11 UHCs, and 5 hospitals) have actually received support from these vertical projects; 110 health facilities have not. Vertical resources as a proportion of overall resources of the few health facilities that do receive them is significant. For the very few health facilities (35 in the survey) that do receive support from vertical funding, the proportion of such support to total resources available for the health facility is about 15 percent, a not insignificant share. As Table 16 shows, of the 19 RHCs that received vertical support, each received an average of ZK5.8 million. The average amount received per facility rises to ZK50.5 for UHCs, and to a whop- ping ZK266.2 million for hospitals. Vertical resources as a proportion of total health facilities' resources are modest. If the same amount of available vertical financing were shared across all health facilities, the aver- age amount going to each type of health facility would be quite low, as shown in the above table. Thus, the dilemma of vertical funding is that by its nature, it needs to focus on a few favored health facilities. This limitation is dictated by two factors: the lumpiness of health investments requires focusing on a few sites, rather than such investments being dissipated in many sites; and the need to show impact requires restricting the interventions and the population coverage. However, in the process, vertical funding leaves a trail of system-wide inequity, with many health facilities being unable to benefit from the large funding. In this study, as much as 80 percent of health facilities (136 out of 171) did not benefit from any vertical support. In this light, basket funding and budget support programs have poten- tially far more equitable impact. Table 16. Donors' Vertical Funds Received by Health Facilities, FY05 (ZK Million) Items RHC UHC Hospital All Total amount of vertical financing received 110.215 555.425 1,330.863 1,996.503 No. of facilities that received 19 11 5 35 Average amount for each recipient facility 5.800 50.493 266.173 57.042 All health facilities in the survey sample 90 40 18 171 Average amount if total vertical financing 1.224 11.000 73.937 11.675 was spread to all facilities in each category of sample WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 50 50 A World Bank Country Study Table 17. Uses of Vertical Funds by Health Facilities, FY05 (percent) Expenditures RHC UHC Hospital All Personal emoluments 0.6 7.3 2.4 3.1 Drugs & medical supplies 6.4 10.2 8.3 8.5 Non-medical supplies 5.8 18.7 2.2 4.7 Workshops, conferences, etc. 10.2 Negl. 7.4 6.4 Repairs & maintenance 2.3 16.4 8.7 9.7 Food & catering 3.5 Negl. 5.5 4.7 Utilities 28.4 2.5 8.4 7.9 Fuel & transport 1.6 Negl. Negl. Negl. Payments for TA, consultancies Negl. Negl. Negl. Negl. General charges 24.5 Negl. 37.2 31.6 Payments of debt 7.6 21.3 0.3 3.5 Capital purchases 2.4 Negl. 4.0 3.3 Others 6.6 23.6 15.6 16.6 Total 100.0 100.0 100.0 100.0 Surprisingly, health facilities have used vertical resources in a broad variety of expen- diture categories, although much is still hidden under "general charges." The results shown in Table 17 somehow casts doubt on the conventional wisdom that vertical resources are inflexible and could only be used for specific and narrow line items. To be sure, each donor imposes its own specific rules on the use of project funds, but it is clear that health facili- ties have used these resources in more creative ways than originally thought, even for the payment of debts, for instance. Nevertheless, for the most part, the fact remains that ver- tical financing is often not fungible across a wide range of disease interventions and pro- grammatic uses, and since activities "follow the money," vertical financing tends to distort national priorities in preference for what global sponsors dictate. The above discussions dealt only with cash support provided by vertical projects. It did not include in-kind support (drugs, commodities, technical assistance, training, informa- tion and management systems, research, monitoring and evaluation, and so forth) that donors support under their vertical projects. It is well-recognized, though not quantified, that these in-kind assistance far outweigh the financial support provided by vertical pro- jects. In this survey, only nine facilities reported having received in-kind support from donors, mostly in the form of medical and non-medical supplies and workshops. Health Facilities' Management of Internally-generated Funds Internally-generated funds (IGFs) are often ignored in public expenditure reviews, but they are an important source of revenue for the health facilities. IGFs include revenues from user fees, prepayment schemes and health insurance reimbursements, referral and medical fees, WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 51 Zambia Health Sector Public Expenditure Review 51 Table 18. Internally Generated Funds of Health Facilities, 2005 (ZK Million) RHC UHC Hospital All IGFs Total Ave. Total Ave. Total Ave. Total Ave. User fees (low-cost & 239.7 4.4 1,574.1 47.9 2,676.6 104.1 4,490.5 51.9 high-cost) Prepayments and health 2.0 Negl. 144.5 4.7 153.9 5.2 300.4 0.5 insurance Referral & medical fees Negl. Negl. 341.2 11.0 459.6 18.4 800.8 7.4 Rev. from hospital-affiliated Negl. Negl. 3.4 0.1 3.4 0.1 6.8 Negl. training inst. Income generating projects 19.6 0.4 Negl. Negl. 37.3 1.5 56.8 1.1 Community donations Negl. Negl. 27.4 0.9 579.0 35.3 606.4 Negl. Others 1.0 Negl. 3.0 0.1 287.2 20.4 291.2 20.2 Total 262.3 4.8 2,093.7 64.6 4,197.0 185.0 6,553.0 81.1 revenues from hospital-affiliated training institutions, income-generating projects (IGPs) of health facilities, and various forms of community donations, including those coming from churches, mosques, and other faith-based organizations, businesses, community- based organizations, and philanthropies. Table 18 shows different types of IGFs generated by each type of health facility. User fee revenues are the largest source of IGFs of health facilities. User fees account for 91 percent of RHC, 75 percent of UHC, and 64 percent of hospital IGFs. The declining importance of fees in hospitals has to do with their broader base of other IGFs. Thus, pre- payments and health insurance, and referral and medical fees, become more significant as one goes up the level of care. Moreover, community donations and a broad category of "others" rise dramatically in hospitals. Aside from fee-based revenues which are governed by official policy, it is unclear how these other significant amounts of IGFs are managed at the facility level, given that these hospitals do not have autonomy. Zambia abandoned user fees for lower-level facilities (health centers) and for services deemed of a public health nature in 2006. Even before fees were abolished for health cen- ters and for public health services in hospitals, the fee program was already stumbling along. As the survey showed, only 4 percent of RHCs and barely a fourth of the hospitals charged fees. Fee programs were strongest in UHCs where 58 percent of them had formal fee programs generating considerable revenues. In most cases (60­80 percent), fees were set by the district health boards or DHMTs, with informal approval of neighborhood health committees where they exist. Only a fifth of the facilities admitted that "fee rates have always been like this." Most facilities have copies of the official fee guidelines, although this is not universal. Most facilities also display the applicable fee schedules. However, only very few health facility staff explained to the patients what the fees were for. Before fees were abolished, a small proportion of rural patients and a large proportion of urban patients paid fees. Table 19 summarizes several features of the fee program in health. Fees for outpatient department (OPD) services is about ZK2,500 for RHCs, WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 52 52 A World Bank Country Study Table 19. User Fee Practices, 2005 Items RHC UHC Hospital All % of facilities where copy of official fee 71 73 65 71 guidelines exist % of facilities that displayed applicable user fees 69 71 90 72 % of facilities that charge fees to some categories 4 58 25 23 of patients % of patients who reported staff explained what 2 12 8 7 fees are for % of OPD patients who paid 16 49 8 -- (per Facility Questionnaire) % of OPD patients who paid 6 40 25 -- (per Patient Questionnaire) Ave. amount of OPD fees paid ZK2,718 ZK3,675 ZK7,886 ZK3,960 (per Facility Questionnaire) Ave. amount of OPD fees paid ZK2,484 ZK3,791 ZK8,041 ZK4,607 (per Patient Questionnaire) % of patients who paid registration card 8 27 9 15 Ave. amount paid for registration card ZK403 ZK2,084 ZK2,863 ZK1,807 % of patients who paid separately for supplies 1 3 1 2 not available in facility Ave. amount paid for unavailable supplies ZK3,346 ZK3,420 ZK3,125 ZK3,328 ZK3,500 for UHCs, and about ZK4,000 to 4,500 for hospitals. There is a very close corre- spondence between what the patients and the health facilities reported on the fees charged. Even before it was abolished, only a small proportion of patients (less than 10 percent) paid OPD fees in RHCs; this proportion shoots up in UHCs where as much as 40­50 percent of the patients pay. Strangely, the proportion of paying patients drops in hospitals, where only about 8­25 percent of patients pay, depending on whether the information is obtained from the patients themselves or the hospital. In addition to the fees themselves, some patients are asked to pay the registration card (presumably for first-time patients). In even rarer cases (1­3 percent), patients pay separately for supplies not available in the health facility. Drugs and other supplies are in shortage in many health facilities, and patients often buy these outside the clinics. The low percentage of this phenomenon in the survey can only be explained by the fact that the patients are still in the health facility when they were interviewed, and have not actually bought the missing drugs and supplies, if ever. Revenue retention was not universal. Full fee retention at the point of service and col- lection is still not universal in Zambia, even in hospitals. Only 39 percent of facilities sur- veyed retained 100 percent of their fee revenues (Table 20). Fee revenues are mostly used for staff enhancements. Table 21 shows that almost 47 per- cent of user fees and other internally generated funds were spent on personal emoluments and capacity-building activities (workshops, and such) that have a salary-augmenting effect on health workers. RHCs are less prone to resort to these two activities (36 percent), compared to hospitals, which use up more than half (54 percent) of their revenues on WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 53 Zambia Health Sector Public Expenditure Review 53 Table 20. Percent of Health Facilities' Rate of Retained Fee Revenues, FY05 (percent) Rate of Fee Retention RHC UHC Hospital All 100 percent 41 29 50 39 75 percent 6 13 -- 8 50 percent 1 9 5 4 25 percent 4 13 -- 6 0 percent 29 20 5 23 No response or D.K. 19 16 40 20 "salary-augmenting" items. UHCs do not directly use their fees for PE, but still use up nearly a quarter of these revenues on workshops. None of the fee revenues was ever spent on the payment of facility debts, although a modest portion was used for utilities (3 per- cent) and fuel and transport (6 percent). About a tenth of these revenues were used to pur- chase drugs and medical supplies, and another 4 percent for non-medical supplies. The measurement of leakage in user fee program is highly reliant on "self-reported" assumptions that are tenuous and prone to variability. Table 22 puts together various assumptions derived from the survey in order to compare actual versus expected fee rev- enues. One presumes that if actual revenues are less than expected revenues, then there must be leakage. In this case, however, expected revenues are well within the actual rev- enues reported by health facilities, by a significant factor. It is clear that this arithmetic method, though useful, is not strongly reliable. A careful study and audit would be neces- sary to conclude whether or not the fee program leaks resources. Table 21. Uses of User Fee Revenues by Health Facilities, FY05 (percent) Expenditures RHC UHC Hospital All Personal emoluments 28 -- 38 30 Drugs & medical supplies 7 10 8 9 Non-medical supplies 4 -- 6 4 Workshops, conferences, etc. 8 24 16 17 Repairs & maintenance 9 22 0 5 Food & catering 7 18 2 6 Utilities 3 -- 4 3 Fuel & transport 12 -- 8 6 Payments for TA, consultancies 2 -- -- Negl. General charges 3 17 9 10 Payments of debt -- -- -- -- Capital purchases 1 7 3 3 Total share 100 100 100 100 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 54 54 A World Bank Country Study Table 22. Expected Versus Actual User Fee Revenues for OPD, FY05 (ZK Million) Items RHC UHC Hospital Ave. no. of OPD patients per month 368 663 317 Proportion of patients paying user fees .16 .49 .08 No. of patients paying user fees per month 59 325 25 No. of patients paying user fees per year 707 3,898 304 User fees (ZK) 2,718 3,675 7,886 Expected user fee revenues per year, ZK Million (a) 1.922 14.327 2.400 Proportion of patients paying registration card .08 .27 .09 No. of patients paying card per month 29 179 29 No. of patients paying card per year 353 2,148 342 Registration card (ZK) 403 2,084 2,863 Expected revenues from regist. card per year (b) 0.142 4.477 0.980 Expected total fee revenues per year, ZK Million (a + b) 2.064 18.804 3.380 Actual low-cost user fee revenues, average, FY05 4.332 29.612 76.316 Overall Resource Availability at the Facility Level Zambia's fragmented health financing system requires an effort to put together a compre- hensive picture of each typical facility's sources of funds. Table 23 summarizes the findings of the expenditure tracking in terms of the various sources of financing and how they end up at the service delivery level. Still, the table is full of caveats. For GRZ, PE is not always included in the DHMT grant, but lies outside it for a number of facilities and districts. Drugs often come in-kind as drug kits, and weren't monetized and included in the table. The imprest system is the purview of the DHMT/DHO, and it is hard to track how much of the resources go specifically to the facility requesting such imprest. Vertical projects often provide in-kind resources, which cannot be monetized. Thus, although the table below is comprehensive, it still not comprehensive enough. Table 23. Level of Annual Resources at the Facility Level, by Major Sources, FY05 (ZK Million) One Typical One Typical One Typical One Typical RHC UHC Hospital Facility Sources Amt % Amt % Amt % Amt % GRZ allocation 27.145 72.4 19.427 19.9 575.110 52.5 113.366 44.7 Basket funds 4.358 11.6 2.625 2.7 261.189 23.8 47.464 18.7 Internally-generated funds 4.783 12.8 64.649 66.2 184.968 16.9 81.101 32.0 Vertical projects 1.224 3.3 11.000 11.3 73.937 6.8 11.675 4.6 Total 37.510 100.0 97.701 100.0 1,095.204 100.0 253.606 100.0 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 55 Zambia Health Sector Public Expenditure Review 55 Table 23 shows that UHCs rely heavily on internally-generated funds, specifically user fees. This may be a response to the surprisingly low financial support they obtain from GRZ and basket funds allocation. Alternatively, it could also be hypothesized that such alloca- tions are small because of the anticipated large expected fee revenues every year. Also, ver- tical projects are a minor factor in a typical RHC. These projects tend to provide greater support, in relative terms, to UHCs and hospitals, in contrast to their often-bandied public-health and rural orientation. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 56 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 57 CHAPTER 5 The State of Health Facilities and Services: Results of the Quality of Service Delivery Component of the PET/QSDS, 2005­06 his chapter presents the results of the quality of service delivery component of the T PET/QSDS. A full-blown report providing the detailed results of the PET/QSDS was written separately,5 and only the summary is provided in this chapter. The first section deals with infrastructure, utilities and equipment; the second with health person- nel; the third with drugs and other medical consumables; and the final section with clinic and patient management. Management of Infrastructure, Utilities, and Equipment The Zambian National Health Accounts show that over the past years, MOH has devoted low level of spending to capital. While for-profit health facilities spend as much as 5 per- cent of their resources on capital expenditures (capex), MOH facilities only spend 2­4 per- cent. Primary MOH facilities have witnessed increasing allocation to capex, where its share consistently rose from 2 percent in 1999 to 5 percent in 2004. However, for both secondary and tertiary facilities, this share has been irregular and has tended to decline, from 5 per- cent in the early 2000s to 1­2 percent in the mid-2000s. For FY05, the data on MOH releases show that capex was squeezed midway through the year. In contrast to these trends, MOH dramatically increased capital spending in the 2006 budget, perhaps in an effort to arrest further erosion of its capital estate. The PET/QSDS provided a graphic picture of the gravity of problems with the health sector's capital assets. Despite their age, most of the health facility buildings appear to be 5. "Zambia Public Expenditure Tracking and Quality of Service Delivery Survey in the Health Sector: Findings and Implications." MOH, UNZA, WB, and SIDA. 57 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 58 58 A World Bank Country Study in good shape, and most are maintained in hygienic condition by the staff. However, sig- nificant deficits exist in many areas. I Less than half (43 percent) of RHCs and only 82 percent of hospitals have electric connection. Only UHCs have almost universal access to electric power (95 percent). I Serioustransporthandicapsexist,especiallyinRHCs,whereonly5percentofthem have a car, ambulance, or animal-drawn vehicle. Roughly 30­40 percent of hospi- tals do have these, and 62 percent have a 4-wheel drive. I Forcommunications,mostfacilitiesrelyonland-basedphoneand/or1-or2-way radios that often conk out. Electronic connectivity is extremely low (only 7 percent of health facilities have e-mail); but even in mostly-urban hospitals, electronic con- nection is low (27 percent). I Only 62 percent of hospitals have incinerators, and only 32 percent of UHCs and 24 percent of RHCs have them. I Only half of the facilities have a mothers' waiting area. The high rates of nonfunctional utilities and transport equipment erode actual access to them. Capital assets with high rates of non-functionality include solar panels, generator sets, land-based phones, 1- or 2-way radios, and bicycles and motorcycles (Figure 37). Hos- pitals have high rates of nonfunctional transport: one out of three hospitals have non- working ambulance; one out of three hospitals have nonworking four-wheel drives; one out of seven hospitals have nonworking trucks; and one out of five hospitals have non- working motorcycles. Clearly, asset maintenance needs to be given greater budget alloca- tion and management attention. Many hospitals continue not to have the complete complement of medical equipment. Only 71 percent of hospitals have an x-ray and only 62 percent have a sonogram machine Figure 37. Health Facilities With Functional and Nonfunctional Utilities and Transport Equipment, 2006 (percent) Electric connection 62 1 Solar panel 38 16 Generator set 18 6 Land phone 19 18 1- or 2-way radio 44 20 Ambulance 6 3 4-wheel drive 9 3 Motorcycle 23 9 Bi/tricycle 54 8 0 10 20 30 40 50 60 70 Working Not working WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 59 Zambia Health Sector Public Expenditure Review 59 Figure 38. Health Facilities with Imaging Equipment, 2006 (percent) 4 RHC 3 11 UHC 11 71 Hospitals 62 16 All 14 0 20 40 60 80 X-ray Sonogram or ultrasound (Figure 38). In general, a little more a two-thirds (67­76 percent) of hospitals have anes- thetic equipment, blood bank, oxygen supply, and laboratory equipment. Thus, around a third of the hospitals continue to operate without these basic imaging and other medical equipment. Medical equipment and instruments are better maintained than non-medical equip- ment. Health facilities had surprisingly low rates of nonfunctional medical equipment. Only 2­4 percent of health facilities reported non-functioning medical equipment in their possession, such as x-ray, sonogram, refrigeration equipment, anesthetic equipment, lab- oratory equipment, blood bank, and oxygen supply (Figure 39). Large demand exists for medical instruments by RHCs. Rural health centers are clearly the least provided among the three types of health facilities included in the survey. Because Figure 39. Health Facilities with Functional and Nonfunctional Medical Equipment, 2006 (percent) X-ray 12 4 Sonogram 12 2 Refrigeration eqpt. 90 4 Sterilization eqpt. 62 8 Exam beds 84 6 Anesthetic eqpt. 12 2 Lab eqpt. 36 4 Blood bank 10 1 Oxygen supply 18 2 0 10 20 30 40 50 60 70 80 90 100 Working Not working WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:43 PM Page 60 60 A World Bank Country Study Table 24. Health Facilities Reporting Inadequate Medical Equipment, Instruments, and Lab Test Supplies, 2006 (percent) Items RHC UHC Hospital X-ray 33 75 33 Sonogram 50 50 23 Lab equipment 54 56 50 Anesthetic equipment -- -- 53 Blood bank -- -- 43 Oxygen supply -- -- 64 Height measuring device 54 45 19 Microscope 74 55 10 Audioscope 87 68 33 Surgical instruments for ob-gyne 57 66 19 Gowns and protective clothing 54 50 14 Malaria smear 76 58 14 Urine test strip 81 74 19 of the large inadequacies at this level, RHCs expressed the highest demand for such items as height measuring devices, microscopes, audioscopes, surgical instruments for ob-gyne, gowns for protective clothing, malaria smears, and urine test strips (Table 24). Management of Health Personnel MOH personnel expenditures steadily increased in nominal levels until 2005; it dipped in 2006 but is expected to rise dramatically to ZK340.9 billion in 2007. Reflecting these trends in absolute levels, PE as a share of MOH expenditures peaked at 46 percent in 2005, and fell to 39 percent the following year (Figure 40), although it is estimated to garner 50 per- cent of the MOH's budget in 2007, the highest-ever share. MOH PE/GDP is about 1 per- cent of GDP. The vacancy rates are undeniably high at 33.5 percent (even higher at 41.4 percent for clinical staff; see Table 25), and the rapid staff turnover especially at RHCs has become untenable. For this reason, the HRH Strategic Plan (2006­11) suggests a substantial increase in recruitment and in staffing establishments. The PET/QSDS, however, showed that skewed patterns persist in established posts (for example, the burden of having "on the books" so many low-skill posts in RHCs, and so many administrative posts in UHCs, as shown in Table 26). The staffing pattern as reflected in these established posts need to be carefully reviewed before any large-scale recruitment. This is all the more important because as shown in the NHA analysis, MOH facilities are far more labor-intensive than either mis- sion or for-profit facilities, even with the large shortage of MOH workers. Alternatively, MOH should set explicit criteria on the types of posts/cadres that should be filled or created as urgent, i.e., professional staff and critical administrative staff in rural areas. Failing to do so would result in bloated administrative and low-skill cadres (because they are easier to fill) WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 61 Zambia Health Sector Public Expenditure Review 61 Figure 40. Personnel Expenditures in MOH Budget and Share of Personnel Expenditures to Total MOH Expenditures, 2000­07 (ZK Billion and percent) 350 100 90 300 80 250 70 200 60 50 50 150 44 46 43 40 39 40 100 30 20 50 10 75.2 139.5 175.9 236.5 265.2 244.1 340.9 0 0 2000 2002 2003 2004 2005 2006 2007 PE in ZK Bn PE/MOH Exp (%) even as professional staff may continue to be in short supply. It would also seem reasonable that, given the increasing share of health expenditures going to administration (as shown in the NHA analysis), central MOH HQ should receive less priority in recruitment. Half of the staff surveyed complained of the workload, and very little time is being devoted to patient care. About 47 percent of health workers claimed working long hours. Table 25. Vacancy Rates in Health Facilities, by Cadre, 2006 (percent) RHC UHC Hospital No. of No. of % of No. of No. of % of No. of No. of % of Estab Vacant Posts Estab Vacant Posts Estab Vacant Posts Cadre Posts Posts Vacant Posts Posts Vacant Posts Posts Vacant Doctors 11 10 91 58 22 38 85 50 59 Clin officers 110 64 58 136 59 43 111 59 53 Medical licentiates 15 13 87 12 5 42 24 18 75 Midwives 109 55 50 282 90 32 179 63 35 Nurses 215 92 43 577 131 23 695 344 49 Env health officers 76 30 39 37 9 24 14 6 43 Pharma, etc. 18 12 67 34 7 21 37 17 46 Dentists, etc. 13 13 100 44 9 20 23 9 39 Lab, x-ray tech, etc. 15 12 80 48 13 27 76 37 49 Physio, etc. 8 8 100 15 3 20 47 34 72 Administrative staff 48 24 50 280 24 9 215 79 37 Other staff 292 55 19 485 62 13 594 152 26 Total 930 388 42 2,008 434 22 2,100 868 41 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 62 62 A World Bank Country Study Table 26. Established Posts and Average Number of Actual Staff by Major Occupational Groups, 2006 RHC UHC Hospitals All Staff No. % No. % No. % No. % Established Posts Prof'l/clinical staff 590 63.4 1,240 61.8 1,291 61.5 3,124 62.0 Administrative staff 48 5.2 280 14.0 215 10.2 543 10.8 Non-clinical, non- 292 31.4 485 24.2 594 28.3 1,371 27.2 administrative staff Total 930 100.0 2,005 100.0 2,100 100.0 5,038 100.0 Average Number of Staff Prof'l/clinical staff 5 57 25 60 52 58 16 57 Administrative staff 0 0 6 14 8 9 3 11 Other staff 4 44 11 26 30 33 9 32 Total 9 100 42 100 90 100 28 100 Long hours of work, however, could not take the place of an "extra hand." The overall effect is that the available time of available workers is being rationed to patients, and so the amount of time being spent on patient care is being squeezed, given that there are other competing tasks that the health worker has to do (Figure 41). The shorter time spent on patients has implications on quality that have not been analyzed. Moreover, the long hours worked manifests itself negatively in the following day as tardiness (due to their being "on- call" the previous night) and tiredness, in a persistent and vicious cycle. Figure 41. Average Number of Hours Worked in a Week by Type of Task, 2006 RHC 25.3 6.5 5.2 2.9 6 UHC 33 6 2.4 1.51.6 Hospitals 30.6 6.9 2.1 1.8 1.2 All 29.6 6.5 3.2 2.1 2.9 0 10 20 30 40 50 Patient care Administrative Social mobilization Training Others WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 63 Zambia Health Sector Public Expenditure Review 63 Figure 42. Per Capita Health Expenditures and Share of Personal Emoluments to Health Expenditures, 2006 (US$ and percent) 60 57 50 44 36 40 30 18 20 14 22 10 0 GRZ GRZ+Basket Funds GRZ+Basket+Vertical Proj. per. Cap. HE (US$) PE/HE (%) Note: The data and table are borrowed from Par Eriksson of SIDA who presented these data during the HR Roundtable in Zambia in 2006, held at the Swedish Embassy. Paradoxically, the MOH and mission facilities' HR shortage is worsening at a time when the health sector is being flooded with donor resources. The causes are obvious, though not often heeded. I First,verticalprojectsrarely,ifever,providedirectsalarysupport(althoughsome do provide allowances in various forms, and "top-ups" for selected staff). Belatedly, the Global Fund through Round IV has allowed the funding of health systems strengthening including human resource development. (Curiously, the Global Fund has funded NGO project staff from the very beginning, but not government staff.) Indeed, most of the other large vertical initiatives (such as PEPFAR) lie out- side the purview of government, even though they involve the MOH service deliv- ery system and rely on MOH health workers. I Second,thebasket-fundingcooperatingpartnersstillhaven'tcreatedafundtosup- port personal emoluments directly. I Third,MOHhasbeenunabletoadjusttotheemergingeraofbudgetsupportthat could have increased the funding for human resources overall, preferring instead health-sector-specific support that it can control (IMF 2007). Moreover, efforts in the 1990s to de-link health workers from the civil service so that they can be pro- vided higher salaries failed. I Thecombinedresultofthesetrendsisdepictedstarklyfortheyear2006inFigure42. As total per capita health expenditures increase with the addition of more funding into the health system, the proportion of PE to total health expenditures declines (even as the proportion of PE to MOH expenditures increases, as mentioned above). In short, it is the inability of the basket funds and vertical financing to formally6 finance PE that causes "so much money chasing so few workers." The imbalance in 6. The word "formally" is important, because as was shown in the chapter on "Budget Allocation, Release and Spending," vertical funds are being used by health facilities to incentivize health workers through one form or another. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 64 64 A World Bank Country Study Figure 43. Summary Rates of Staff Vacancy, Absenteeism, Tardiness, Involvement in Income-Augmenting Activities, and Dissatisfaction, 2006 (percent) Vacancy rate 33.5 Absenteeism rate, count 9.6 Absenteeism rate, self-reported 21 Tardiness rate 43 Rate of involvement in income-augmenting 21 activities Dissatisfaction rate 44 0 5 10 15 20 25 30 35 40 45 this factor ratio has not been properly analyzed, including its tendency to inflate wages in the "flexible-wage" sector (NGOs, contractors) vis-à-vis the "fixed- wage" sector (government, missions). Indeed, a proper labor dynamics analysis is needed. Absenteeism and tardiness severely restrict the actual number of full-time equivalent ("real") workers, and these twin problems must be tackled head-on (see Figure 43). Based on staff count, some 9.6 percent of staff were not in the health facility during the survey. Absenteeism was highest among doctors (31 percent), clinical officers and medical licen- tiates (20 percent), nurses and midwives (14 percent), and other clinical staff (14 percent).7 However, the rate of self-reported absenteeism was higher, with 21 percent of staff admit- ting having been absent at least once during the previous month of the survey; five days was the average number of days' absent. Self-reported tardiness was also high (3 percent overall) at an average of four days' absence per month. Human resource discussions in Zambia have overly focused on the need to fill vacancies, and have relegated the issue of staff absenteeism and tardiness to the background. The rates of absenteeism and tardi- ness derived from the PET/QSDS imply a total loss of 4,108 working days per month. Thus, if absenteeism and tardiness were fully eliminated, Zambia would gain virtually 187 full-time equivalent staff, enough to staff 2 hospitals, or 4 UHCs, or 21 RHCs. HR systems need to have a better handle on the absenteeism and tardiness problems, and how to deal with it. 7. The high rate of absenteeism among health workers is not unique to Zambia. Indeed, it is pretty common in the developing world. For primary health facilities, Reinikka and Svensson (2006) find the rates of absenteeism to be as high as 35 percent in Bangladesh, 27 percent in Honduras, 40 percent in the 19 states of India, 40 percent in Indonesia, 19 percent in New Guinea, 23 percent in Peru, and 37 percent in Uganda. Lindelow (2006) reports 19 percent absenteeism rate in Mozambique. Absenteeism studies and methods need to be standardized: it matters whether the survey team announces or not their visit to the health facility. In the case of the Zambia PET/QSDS, the visit of the survey team was announced, which could explain the relatively lower absenteeism rate compared to the other countries. However, the health worker survey found higher rates of self-reported absenteeism. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 65 Zambia Health Sector Public Expenditure Review 65 Figure 44. Composite Monthly Salaries and Allowances of Clinical/Professional Health Workers, 2005 (ZK Million) 0 1 2 3 4 5 6 Doctor 3.778 2.456 Pharmacist 3.072 1.049 Lab scientist 2.688 0.973 Tutor 2.430 0.970 Sr. nurse 1.683 0.892 Nurse 1.142 0.355 Midwife 1.142 0.355 Clinical officer 1.142 0.355 Lab technologist 1.142 0.355 Pharmacy technician 1.142 0.285 Lab technician 0.981 0.255 Salary Allowances An overall wage strategy is needed. The salary structure is highly compressed (see Fig- ure 44) and although the allowance system has given a reprieve in decompressing such structure, it is not the best way of dealing with the problem. While the retention scheme was a right stop-gap measure at the beginning of the human resource crisis, it involved only a tiny minority of staff. The fragmented cash allowance and in-kind benefit system need to be consolidated. The wide variety of allowances and benefits only caters to a small segment of the health workforce, and it is difficult to forecast the budget implications of such a wide range of benefits. GRZ salary management needs to be strengthened. The PET/QSDS revealed a num- ber of weaknesses in this system, including the following: I Inconsistency in the number of posts actually filled: If reckoned as a residual of vacant posts, the filled posts from the survey would be 3,348 staff. However, if reckoned in terms of staff count from the absenteeism component of the survey, the filled posts would total 3,885 staff. The difference between these two figures is 537 staff. I Nonreceiptofthefullamountofsalaries:About15percentofstaffdidnotgetalltheir salaries in the previous 12 months (Table 27), a higher percentage of them in hos- pitals. The unpaid salaries for these staff can be as high as 3­5 months.8 8. Studies on unpaid salaries of civil servants in the developing world are few. In Kogi State, Nigeria in 2001­02, the average number of months that salaries were unpaid was 5 months for all health workers. It ranged from 1 month in Ogori Magongo District to 10 months in Dekina District (Das Gupta, Gauri, and Khemani, 2003). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 66 66 A World Bank Country Study Table 27. Salary Management, 2006 Percent of Staff RHC UHC Hospital All % of staff who received all salaries 85.4 87.7 82.3 85.4 due the past 12 months % who did not receive all salaries 14.6 12.3 17.8 14.6 due the past 12 months Ave. no. of months not paid 4 3 5 4 % who received all salaries on time 28.7 16.7 19.8 21.9 % who experienced delays in 71.3 83.1 80.2 78.1 receipt of salaries Ave. no. of months delay 1 1 1 1 % who received salaries in cash 11 10 10 10 % who had salaries automatically 88 90 90 90 deposited in the bank % who received all salaries 90.9 86.0 75.0 84.5 net payable % who received less than net payable salary, 9.1 14.0 25.0 15.5 without consent or understanding Ave. amount of salary missing (ZK) 72,444 239,133 244,278 189,015 % who recovered missing portion of salary 18 0 25 21 % who paid "expediter's fee" to obtain salary 6 8 13 10 I Delayinthereceiptofsalaries:Moststaff(78.1percent)experienceddelayofabout one month in receiving their salaries. A third of those who experienced this prob- lem said it is due to "systemic delays," and half cited "other reasons." I Unauthorizeddeductionsinsalaries:About15.5percentofstaffreceivedanamount less than their net payable salary without their consent or understanding. This is highly prevalent in hospitals where 25 percent of staff cited this problem. The miss- ing portion of salaries is not an insignificant amount: it averaged ZK189,015, and the missing amount tends to be higher in hospitals compared to health centers. About 21 percent of staff who experienced this problem reported they eventually recovered the missing portion of their salary. I Payment of "expediter's fee" to obtain salaries: A tenth of the staff reported paying a facilitation fee to obtain their salary; this problem occurs mostly in hospitals. One can surmise that this problem occurs among those staff who continue to receive salaries in cash (10 percent of staff), since the salaries of most staff (90 percent) are automatically deposited into their bank accounts. Due to time limitations, staff productivity was not assessed in this study. However, it is critical that this be done (the raw data already exists from the PET/QSDS) to understand better the input-mix of service provision, and to provide better evidence on how health workers should be incentivized. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 67 Zambia Health Sector Public Expenditure Review 67 Management of Drugs and Other Medical Supplies The visible improvement in drug availability could be reversed with the surprisingly sharp decline in the drug budget for FY06. Since the early 2000s, the drug budget has been increasing, and this has resulted in a positive impact on drug availability as drug kits opened per 1,000 population rose steadily since 2002 (Figure 45). Thus, it is surprising that for FY06, the MOH + Basket Fund Budget slashed drug allocations by as much as 15 per- cent, with the districts bearing the brunt of the reduced allocations by as much as 34 per- cent. These budget reductions come at a time when health facilities are still smarting from very visible drug shortages as will be documented in this section. The PET/QSDS data revealed the following drug distribution problems: I Although MSL drugs are supposed to arrive on a predetermined schedule, more than a third of the districts reported delays in receiving drug kits, some as frequently as seven times a year. The drug kits arrive relatively on-schedule in hospitals, but a third of the UHCs and RHCs experience delays. I Evidence of drug diversion can be inferred in 25 percent of the facilities receiving drug kits in excess of what DHMTs claimed they distributed, and 39 percent of facilities receiving drug kits fewer than what DHMTs claimed they distributed. Only 36 percent of facilities reported exactly the same amount that DHMTs reported they distributed. I Essentialandlife-savingdrugsarewidelyunavailable.Coartem,amalariadrug,was out of stock on average for as long as 9.5 weeks in RHCs while ampicillin, an antibi- otic, was out of stock on average for 7.4 weeks in UHCs. Life-saving drugs such as insulin and dextrose were also unavailable in many hospitals during the survey, although their lengths of stock-out were shorter. Figure 45. Drug Expenditures and Drug Kits Opened Per 1,000 Patients, 1999­2005 (ZK Billion) 120 1.2 1.08 100 1 0.93 80 0.8 0.73 0.75 0.73 0.69 60 0.6 40 0.4 20 0.2 13.5 44.8 62.8 88 80.5 113.5 0 0 1999 2000 2001 2002 2003 2004 2005 Drug exp. Kits per 1000 pts. Note: The drug expenditure data were taken from the NHA series, as broken down into primary, sec- ondary, and tertiary facilities. It is likely that this series also included expenditures from vertical pro- jects, which are not included in the GRZ + Basket Funds budget. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 68 68 A World Bank Country Study I More than half (55 percent) of hospitals and about half (46 percent) of RHCs reported having expired drugs. Inappropriate drugs were less of a problem; still, 14 percent of all health facilities reported having them. Clinic and Patient Management Actual capacity of facilities to deliver health services falls short of their self-reported capac- ity. Health facilities tend to be over-optimistic about their capacity to provide care. For instance, health centers consistently pride themselves in being able to deliver basic services such as family planning, antenatal and postnatal care, child immunization, child health/IMCI, and adult malaria treatment and diagnosis. Some 93­100 percent of RHCs and UHCs reported their ability to dispense these. However, using availability data on drugs, consumables, equipment, and instruments from the survey, it is clear that actual capacity is several percentage points lower than claimed capacity. Using indicators on the availability of drugs, vaccines, and key consumables, it can be shown that: I Only ~87 percent of RHCs and ~95 percent of UHCs could provide family plan- ning services (rather than the self-reported 98 and 100 percent, respectively); I Only~89percentofRHCsand~84ofUHCscouldprovideantenatalandpostna- tal care (rather than the self-reported 99 and 95 percent, respectively); I Only ~77 percent of RHCs and ~84 percent of UHCs could provide child immu- nization services (rather than the self-reported 94 and 95 percent, respectively); I Only ~86 percent of RHCs and ~81 percent of UHCs could provide child health/ IMCI services (rather than the self-reported 93 and 97 percent, respectively). Regarding facility management and supervision, the PET/QSDS revealed that data accu- racy of patient registers is still an issue. While having such system is high among health facilities, only two-thirds of them have accurate data. There is a functional system and reg- ular timing of facility supervision where the supervisory team does a variety of tasks and provides feedback and follow-up. Regarding patient management, the PET/QSDS revealed that Facilities are quite close to patients (60 percent of them are within 30 minutes walking distance to a health facility), but waiting times are long (average of 65 minutes). Indeed a few patients wait hours to obtain service. Needless to say, the long queues and waiting times are a direct effect of the human resource shortage. Most patients simply walk to a health facility incurring no trans- port costs, but a quarter of them do entail sizeable transport costs. Although a very small percentage of patients (1 percent) incur hotel and food costs as they sleep overnight in town to access care, their average costs for these are high (average of ZK36,875). Overall patient rating of quality is moderate, at 85 percent. Some 89­95 percent of patients rated health workers highly on the "hospitality" aspects of care (friendliness, pro- vision of information), but only less than half of them (40­42 percent) rated health work- ers well on the "technical" aspects of service delivery (for example, explaining the purpose of the procedure, exam or drug). Despite the drug shortages found in the survey, only about 8 percent reported not receiving any medication after the clinic visit. Of the 92 percent of patients who did receive medication, only 74 percent said health workers explained what the medication is for. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 69 CHAPTER 6 Scaling Up Health Service Delivery to Reach the MDGs: Results of the Marginal Budgeting for Bottlenecks Modeling Exercise, 2006 A bout 15 years into the Zambian health sector reform process, the country has made remarkable progress particularly in improving access to some basic health services and formulating health policies and strategies as well as its implementing mecha- nisms. However, as discussed in the earlier sections of this report, these efforts have not translated to a significant improvement in health status, and Zambia is facing tremendous challenges in reaching its MDGs. Chapter 2 of this report documented the challenges fac- ing the health sector in delivering health services and organizing health inputs in order to improve health outcomes, and describes the country's strategies and efforts to scale up health service delivery. Chapter 5 further portrayed the deficits in service delivery in terms of key inputs. Based on these findings, this chapter evaluates the options of maximizing the impact of health resources on outcomes by examining the bottlenecks along the service delivery chains and effective service delivery arrangements. Finally, it provides the evidence from a simulation analysis on specific service delivery scenarios and their cost and impact. Health Sector Response to the Challenges: Strategy and Resources National Health Strategic Plan (NHSP) Realizing these challenges itself, the country's current NHSP, 2006­10, the fourth in the series of strategic plans implemented under the health sector reform, emphasizes the importance of focusing on a defined list of national health priorities, which include resolv- ing the human resource crisis; addressing national public health priorities; and ensuring thatprioritysupportsystemsandservicesreceivethenecessarysupport.Oneoftheimportant steps to achieve the objectives of the NHSP is to continue to finalize and deliver the Basic Health Service Package, which is not fully implemented mainly due to both financial and 69 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 70 70 A World Bank Country Study human resource constraints. It is estimated that the Basic Health Service Package for the primary and second levels of care costs about US$12 per capita. The cost will go up to US$18 per capita if tertiary care and ARV treatment is included. The estimated cost is rea- sonable when compared to international standards. To help reach this target, the govern- ment is also planning to increase its health budget to 15 percent of the total government budget in the near future as requested by the Abuja target. Options for Maximizing the Impact of Health Spending To improve health outcomes in Zambia, one of the priorities for Zambia as well as devel- opment partners is to improve effectiveness and efficiency of health resources. Interna- tional experience has proven that resources are necessary but not sufficient solution to better health. Ineffective spending is not uncommon in the field of internal health as health is such a complex field with multiple outcomes, outputs and inputs as well as the interac- tions among them. It is imperative for Zambia to avoid mistakes made by other countries and to learn for its own lessons in the past. To align resources to results, the country needs to look into the following questions: I Whetherresourcesarespenttowardsolvingthemainhealthproblems? I Whether resources are spent on services that not only are cost-effective but also have an impact on the main health problems? I Whetherresourcesarespentonanefficientandintegratedservicedeliverysystem to deliver the services? I Whetherresourcesarespentoninputsthatarekeytotheservicedeliverysystem? I Whetherresourcesarespentoneffectiveservicedeliveryarrangements? Choosing Efective Service Delivery Modes Among these elements, Zambia performs well in identifying health problems by using its information and M&E systems; defining health priorities under the process of the National Health Strategic Plans; and prioritizing a list of cost-effective interventions through the for- mulation of the Basic Health Service Package. However, to achieve the impact at the health outcome level, Zambia will need to further focus on building a functional service delivery system with proper service delivery arrangements and necessary inputs. One important step to build a functional service delivery system is to explore and invest on effective ser- vice delivery modes in order to align inputs towards results. Health services can be deliv- ered not only within health facilities, but also through outreach or mobile arrangements, and by communities themselves. Based on international experience, health services can be classified according to three types of service delivery mode: I Family or community-based care. This consists of interventions that include pre- ventive measures and the management of maternal and childhood illnesses. They can be delivered by households or communities themselves under guidance from health professionals. Insecticide treated bed nets (ITNs) for pregnant women and children under five, condom use, breastfeeding, and oral re-hydration therapy (ORT) are some examples of interventions that are family or community based. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 71 Zambia Health Sector Public Expenditure Review 71 I Population-oriented outreach services. These services are delivered to all the popu- lation, regardless of whether or not people are currently sick. They are usually deliv- ered through periodic outreach or scheduled clinical services. This delivery mode includes preventive care interventions, such as immunizations, antenatal care, fam- ily planning, TB treatment. I Clinical-basedindividualcare.Theseactivitiesincludealltypesofindividualcurative care interventions that need to be delivered at a health facility and by a trained health care professional. They are offered in a continuous manner so that they can respond to unpredictable situations, such as a sudden illness, or the delivery of a baby. Traditionally, facility-based care has been the focus of health service delivery and drawn most resources, but each mode has its unique advantage and can be useful for different ser- vices. For example, for services with standardized procedure and that apply for a segment of population instead of just sick individuals, the outreach mode can be particularly effec- tive, especially when access to health facilities is limited. The promotion of healthy behav- ior and domestic hygiene practices undertaken in household and community settings is a critical complement to professional health services. Each service delivery mode requires a different set of inputs. For outreach services, the availability of vehicles, essential drugs and supplies, qualified nurses, and the participation of communities are the basic requirements to improve service coverage. The success of community-based services is based on the dis- semination of health knowledge, the supply of essential materials, and the follow-up by community health workers. Meanwhile, for facility-based services, qualified health profes- sionals, infrastructure, equipment, and drug supply are essential to ensure quality of care. International experience has shown that delivering a set of essential services through community-based service and outreach arrangement are very cost-effective in reducing mortality (particularly child mortality) and controlling infectious diseases. A study of 12 countries (Benin, Burkina Faso, Ethiopia, Ghana [north], Madagascar, Mali, Mauritania, Mozambique, Niger, Nigeria, Rwanda, Senegal [west]) shows that, at a low cost, community- based and outreach services can potentially reduce child mortality by 43 percent and 16 per- cent, respectively. While the rebuilding of health facilities will take some time and requires a large investment, the development of community-based and population-based outreach service delivery should go in parallel with facility-based service delivery. And even facili- ties should not just be points of service provision, but serve as bases to support outreach teams and community-based services. Bottlenecks in the Service Delivery Chains To deliver good-quality health services to the users, a smooth service flow is required along the five dimensions of the service delivery chain, which are: availability, accessibility, uti- lization, continuity, and quality. I Availability:Theavailabilityofcriticalhealthsysteminputssuchasdrugs,vaccines, supplies, and human resources. I Accessibility:Referstothephysicalaccessofhealthservicestotheclients.Thisincludes the presence of trained human resources at the community level, the number of WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 72 72 A World Bank Country Study villages reached frequently by outreach services, and the time taken to reach a facil- ity providing basic and emergency obstetric and neonatal care services. I Utilization:Itindicatesthefirstuseofmulti-contactservices,suchasthefirstante- natal contact or BCG immunization. I Continuityofcoverage(oradequatecoverage):Itrepresentstheactualcontactsofthe target population with health services compared to optimal contacts. This applies to services that require multiple visits, such as DPT3 immunization that require three shots, or antenatal care where women should have at least three visits to a health facility during their pregnancy. I Quality of coverage (or effective coverage): This measures the quality of care by assessing the skills of health workers. The skills are assessed in terms of workers' ability to examine beneficiaries, diagnose, provide the requisite interventions, use the equipment appropriately, and advise properly. The five determinants are sequential, following the logic of service delivery flow. First, ser- vices have to be available in a given area. Second, the service locations have to be physically accessible to users. Third, to receive health benefits, potential users have to actually use the services. Fourth, potential beneficiaries should utilize the services with continuity, for example, follow the complete schedule of care such as three doses of DPT. Finally, poten- tial users should utilize the services in a correct and effective matter (see Figure 46). In a perfect world, services are delivered smoothly and effectively when key inputs are available, services are widely accessible, the quality of care is good, and the utilization rate by patients is high. In this case, the investment will achieve a high return and users will obtain maxi- mal benefits from the services. In reality, bottlenecks can occur in any of the five dimen- sions to hinder service delivery. Consequently, this not only stops services from reaching their intended target, but also makes resources invested in other dimensions ineffective. So, it is imperative to understand the nature of the bottlenecks, which is not an easy task, as different service delivery modes may face different bottlenecks. A bottleneck analysis conducted on all the three service delivery modes in Zambia indi- cates that a series of bottlenecks exist in the service delivery chain and different delivery modes face different problems. The analysis first clustered a list of key services under each service delivery mode to represent each mode; then identified indicators for the five dimen- sions (availability, accessibility, utilization, continuation, and quality) under each key service; finally used the weighted averages of all services under each Figure 46. Sequential Relationships mode to demonstrate the gaps in of Health Services each of the five dimensions. 5. Quality I Community-based services: The 4. Continuity bottlenecks exist in all five dimensions but the quality of care 3. Utilization (or effective coverage) presents 2. Accessibility the largest problem. It is esti- 1. Availability mated that community-based Population eligible for services services are provided to areas covering about 68 percent of the WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 73 Zambia Health Sector Public Expenditure Review 73 Figure 47. Bottlenecks in Community-Based Services Community-based Services 80% 70% 60% 50% 40% 30% 20% 10% 0% Availability Accessibility Initial utilization Timely and Effective continuous coverage Source: MBB Team's analysis based on available survey and official data. households, about 62 percent of households have reasonable access to the services, on average 51 percent are actually using the services, 36 percent used the services in a timely and continuous matter, and only 17 percent received effective and qual- ity services that would achieve the potential effectiveness of the services (Figure 47). I Population-based outreach services: The availability and accessibility indicators rea- sonably stand at 79 and 71 percent each. However, the levels of utilization, conti- nuity, and effective coverage are low at 42, 29, and 27 percent respectively, indicating that major bottlenecks exist in these dimensions (Figure 48). I Facility-based services: While availability is reasonable, accessibility is a concern with only slightly more than 50 percent of households having access to the defined list of services. In addition, utilization, continuity, and quality are very low, repre- senting significant bottlenecks in these dimensions of service delivery (Figure 49). Figure 48. Bottlenecks in Population-Based Outreach Services Population-based Outreach Services 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Availability Accessibility Initial utilization Timely and Effective continuous coverage Source: MBB Team's analysis based on available survey and official data. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 74 74 A World Bank Country Study Figure 49. Bottlenecks in Facility-Based Services Facility-based Services 80% 70% 60% 50% 40% 30% 20% 10% 0% Availability Accessibility Initial utilization Timely and Effective continuous coverage Source: MBB Team's own analysis based on available survey and official data. Tobuildasmoothservicedeliverychain,investmenthastobebalancedamongthefivedimen- sions. Making a strong investment in one part of the chain while ignoring the others will be a waste of resources, and will not help the delivery of services. For example, expanding health infrastructure in order to improve access to services must go together with the availability of health staff, drugs and supplies, as well as the utilization and quality of the services. Building a health facility without proper staff, medicine and equipment, as well as demand-side incen- tives,willonlyunderminethevalueoftheinvestmentandnotincreaseservicecoverageforthe needy.Onlyhigheffectivecoveragecanmaximizetheimpactonhealthoutcomes;butindoing so, availability, accessibility, utilization, and continuity need to be systematically improved. Differentbottlenecksrequiredifferentsolutions.Forexample,solvingtheproblemoflow availability and accessibility requires an expansion of health services. However, low utilization andpoorqualityneedtobeaddresseddifferently.Carefullystudyingthenatureofbottlenecks can help target resources effectively and ensure the smooth operation of service delivery. Scenarios for Scaling Up Service Coverage This report recommends the country to consider the following steps in two phases. Phase I will aim at increasing most the essential service coverage by 30 percent (with adjustments made on the coverage targets to be consistent with the National Health Strategic Plan) and Phase II increases the coverage by 50 percent. Phase I will need about three years to com- plete and Phase II may take longer. A simulation analysis, using the Marginal budgeting for Bottlenecks tool, was conducted to estimate the cost and impact for each step and phase. I Step1:Undertakesocialmobilizationandbehavioralinterventionsaswellassup- ply essential materials to households through community-based interventions; I Step 2: Organize outreach and mobile teams to provide a set of standardized ser- vices to populations without access to health facilities; I Step 3: Expand the primary health care network to provide preventive and basic curative care; WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 75 Zambia Health Sector Public Expenditure Review 75 I Step 4: Strengthen the first level referral care that can provide comprehensive and emergency health care; I Step5:Improvethesecond-levelreferralcarethatcanprovidespecializedcare. Step 1 This aims at reducing child mortality, controlling communicable diseases, and improving environment health. To that effect, it is essential to scale up community-based services (such as breast feeding, bed net use, and so forth). The implementation of this step requires a suf- ficient number of community health volunteers and health promoters to actively dissemi- nate health knowledge and promote healthy behaviors. It is estimated that one volunteer is needed for about every 1,000 population in order to carry out the required workload. Other inputs required include training, incentives, and essential materials (such as mosquito nets). While Zambia is experiencing a human resource crisis for health professionals, it has a good base of active community health workers. Therefore, the development of community health interventions can be started relatively quickly as it is not conditional on the expansion of health infrastructure. The challenge will be how to establish a sustainable incentive mecha- nism to motivate community health workers. Many countries have issues on the proper form and amount of the incentives, and in the case of Zambia this will need to be decided in function of the local context. The simulation result indicates that the additional average annual cost for this step would be modest at $0.85 per capita per year for Phase I and $1.66 for Phase II, and if properly implemented, it would potentially reduce under-five mortality by 13 percent for Phase I and 20 percent for Phase II, and maternal mortality by 0.5 percent for Phase I and 0.7 percent for Phase II (see matrix below). Cost and Impact of Step 1 Step 1: Undertake Community-Based Social Mobilization and Behavioral Reduction Reduction Cost (US$ per Interventions in U5MR in MMR capita p.a.) Phase I 12.8% 0.5% 0.85 Phase II 19.7% 0.7% 1.66 Step 2 This includes efforts to take a package of highly effective and standardized services (such as immunization and antenatal care) outside health facilities and bring it to the households and communities. This is particularly important in Zambia, as significant percentage of the population does not have a reasonable access to health services. The country has achieved a significant success in delivering immunization services in communities, and this effort should be expanded to integrate other interventions and deliver a broader package. This step requires putting together an outreach team with at least two qualified nurses or other types of health professionals to visit communities totaling about 5,000 people in a timely manner. To ensure quality and efficiency, the outreach team needs to work together with community health workers and its support teams in health centers or health posts. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 76 76 A World Bank Country Study Key inputs to make the outreach team functional include vehicles, essential drugs and sup- plies, and training. The cost and impact of this step are summarized in the matrix below. Cost and Impact of Step 2 Step 2: Scale up Population-Based Reduction Reduction Cost (US$ per Outreach Services in U5MR in MMR capita p.a.) Phase I 9.2% 0.3% 0.73 Phase II 14.0% 1.0% 1.88 Step 3 This aims at expanding the health facility network so that the majority of the population has reasonable access (less than 2 hours walking distance) to basic primary health care (including IMCI). This step requires investment effort and sufficient recurrent budget to ensure that the required human resources and other inputs are in place. At this level, health services provided in facilities include both individual curative treatment and preventive services. This step involves facilities, health professionals, equipment and recurrent costs (salary, supplies). The cost and impact of this step is shown in the matrix below. Cost and Impact of Step 3 Step 3: Expand Clinical Primary Health Reduction Reduction Cost (US$ per Care (including IMCI) in U5MR in MMR capita p.a.) Phase I 14% 2% 0.16 Phase II 16% 8% 1.10 Step 4 This aims at improving the first referral care, which is particularly crucial for complicated cases and maternal complications. Individual treatment and emergency obstetric care needs to be provided at this level, and at least one general doctor or qualified clinical technician needs to be installed. To implement this step, Zambia will need to produce/introduce more physicians and deploy more physicians to the district level. The cost and impact of this step are summarized in the matrix below. Cost and Impact of Step 4 Step 4: Strengthen the First Level Reduction Reduction Cost (US$ per Referral Care in U5MR in MMR capita p.a.) Phase I 2.5% 0.8% 0.36 Phase II 7.2% 5.2% 1.08 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 77 Zambia Health Sector Public Expenditure Review 77 Step 5 This focuses on the delivery of comprehensive emergency obstetric care and specialized care. Specialists are needed at this level. The cost for this step is higher as it requires con- structions, salary and incentives, equipment and drugs (see matrix below). Cost and Impact of Step 5 Step 5: Improve the Second Level Reduction Reduction Cost (US$ per Referral Care in U5MR in MMR capita p.a.) Phase I 0.2% 1.9% 2.25 Phase II 0.3% 2.7% 3.39 Overall Impact Altogether, the five steps, if implemented successfully, would have a significant impact on child and maternal mortality. They would reduce child mortality by 24 percent for phase I and 46 percent for Phase II, and maternal mortality by 5 percent for Phase I and 16 per- cent for Phase II. This achievement would cost Zambia, on average, an additional US$ 4.38 per capita per year for Phase I and US$9.08 for Phase II. Total Cost and Impact of all Five Steps Reduction Reduction Cost (US$ per All Five Steps in U5MR in MMR capita p.a.) Phase I 24.4% 4.9% 4.36 Phase II 45.7% 15.9% 9.07 Figure 50. Projected Trend in Reaching Child Mortality MDGs U5MR 250 Current trend 200 births 150 live 100 Phase I 1,000 per 50 Phase II 0 1990 1996 2002 2010 2013 2015 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 78 78 A World Bank Country Study Figure 51. Projected Trend in Reaching Maternal Mortality Reduction MDGs MMR Current trend 800 700 600 births 500 Phase I live 400 Phase II 300 100,000 200 per 100 0 1996 2002 2010 2013 2015 Zambia needs to make a commitment and take concrete steps to scale up the coverage of essential health services and to improve the population's health status. If successful, the improved health will no doubt boost the economy, improve productivity, and provide sta- bility to the country. The simulations are by no means a prescription for the country's future health strategy to reach the MDGs. Rather they provide examples of how this evidence-based simulation exercise could help the country's policy development. Other policy scenarios, if requested by the country, can also be incorporated in to future simula- tion exercises. These steps will help Zambia accelerate the pace of reducing child mortal- ity and make it back on track to reach the child mortality MDGs (Figure 50). They will also reverse the worsening trend of the MMR and make it possible to start to lower the level of maternal mortality towards the MDGs (Figure 51). WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 79 Statistical Appendix Annex Table A. MOFNP and Local Government Authorities' Health Expenditures by Recipient, 1995­2004 (ZK Billion) MOFNP LGAs Other Non- Local MOH/ Ministries & government MOH/ Autho Years CBOH Public Inst. Inst. Total CBOH Rities Total 1995 58.3 2.1 3.4 63.8 0 1.5 1.5 1996 72.7 2.8 2.5 78.1 0 2.5 2.5 1997 95.6 5.2 2.6 103.3 0 5.5 5.5 1998 112.2 3.7 3.9 119.8 0 1.2 1.2 1999 140.9 5.2 0 146.0 0 3.0 3.0 2000 147.3 6.2 0 153.4 0 1.8 1.8 2001 288.2 5.2 0 293.8 10.3 0 10.3 2002 342.4 7.7 0 350.2 0.8 0 0.8 2003 327.1 10.2 0 337.3 0 0 0 2004 322.4 10.4 0 332.8 0 0 0 79 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 80 80 A World Bank Country Study Annex Table B. Donors' Health Expenditures by Recipient, 1995­2004 (ZK Billion) MOH/CBOH & Other Donors as Financing Public Institutions Agents, incl. NGOs Years Amount % Amount % Total 1995 17.6 92.2 1.5 7.8 19.0 1996 9.3 21.4 34.1 78.6 43.4 1997 16.7 22.8 56.3 77.2 72.9 1998 63.1 66.6 32.0 33.4 95.1 1999 29.2 74.4 9.7 25.6 37.9 2000 84.9 83.9 16.2 16.1 101.1 2001 82.3 79.0 21.9 21.0 104.3 2002 195.6 58.2 140.6 41.8 336.2 2003 317.3a 60.0 211.2 40.0 528.5 2004 437.2b 58.8 352.8 41.2 790.1 a MOH/CBOH alone is ZK309.9 billion. bMOH/CBOH alone is ZK413.2 billion. Annex Table C. Employers' and Households' Health Expenditures by Recipients, 1995­2009 (ZK Billion) Households Employers Households Other as Financing Health Years ZCCM Employers Total Agents Insurance Total 1995 16.4 10.2 26.6 58.8 0 58.8 1996 26.0 7.5 33.4 81.6 0 81.6 1997 30.6 9.5 40.1 101.7 0 101.7 1998 30.8 27.0 57.8 130.9 0 130.9 1999 27.8 22.5 50.3 178.4 0 178.4 2000 31.9 27.2 59.2 223.8 0 223.8 2001 0 54.7 54.7 248.2 0 248.2 2002 0 62.5 62.5 308.9 3.6 308.9 2003 0 89.8 89.8 395.3 4.2 399.6 2004 0 128.9 128.9 524.3 4.6 528.9 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 81 Zambia Health Sector Public Expenditure Review 81 Annex Table D. Total Health Expenditures by Service Provision, Administration, and Rest of the World Transactions, 1995­2004 (ZK Million) Service Provision Administration Rest of the World Years Amount % Amount % Amount % Total HE 1995 135,120 78.4 37,317 21.6 0 0 172,437 1996 202,411 83.3 40,556 16.7 0 0 242,967 1997 278,700 85.2 48,302 14.8 0 0 327,002 1998 314,550 76.0 99,209 24.0 0 0 413,759 1999 343,528 80.9 81,217 19.1 0 0 424,745 2000 462,791 82.2 100,043 17.8 0 0 562,838 2001 614,673 85.2 106,545 14.8 0 0 721,218 2002 820,947 75.8 236,783 21.9 24,707 2.3 1,082,437 2003 1,038,802 74.8 324,538 23.3 26,168 1.9 1,389,508 2004 1,265,153 68.0 573,317 30.8 21,088 1.1 1,860,558 Notes: "Administration" includes general administration, research, and training. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 82 82 A World Bank Country Study Annex Table E. Total Health Expenditures by Service Provision and Administration and by Type of Providers, 1999­2004 (ZK Billion) Items 1999 2000 2001 2002 2003 2004 Service Provision MOH tertiary hospitals 48.1 59.7 61.5 91.6 105.8 126.8 MOH secondary hospitals 8.1 18.6 31.6 39.3 51.5 48.6 MOH district facilities 45.7 87.4 215.9 258.2 279.2 337.6 Defense Force hospitals 5.1 6.0 5.1 5.5 5.9 10.0 ZCCM hospitals 31.0 36.4 0.0 0.0 0.0 0.0 Private for-profit hospitals 63.2 87.7 140.7 171.0 255.3 382.0 Mission hospitals 2.8 4.5 5.8 9.5 12.6 10.8 Nursing and residential care Negl. Negl/ Negl. 0.3 0.5 1.3 Ambulatory health care 8.3 6.5 5.4 17.2 39.0 18.3 Traditional healers 36.3 45.6 52.4 72.2 87.7 103.4 Drug stores & chemists 86.3 108.5 96.2 107.3 131.2 162.1 Other pub. Health providers 8.9 1.8 -- 48.9 70.1 64.2 Subtotal, Service provision 343.5 462.7 614.7 820.9 1,038.8 1,265.2 Administration GRZ health administration 59.4 74.0 49.8 147.9 174.6 261.8 Other providers' health adm. 7.7 7.0 30.1 21.8 36.6 74.2 Other related svcs & inst. 8.3 12.1 21.1 45.2 82.9 210.3 Research 1.9 2.4 4.5 17.2 22.4 20.7 Training 4.0 4.5 0.2 4.6 8.1 7.3 Subtotal, Administration 81.2 100.0 106.5 236.8 324.5 573.4 Rest of the world 0 0 0 24.7 26.2 21.0 Grand total 424.7 562.8 721.2 1,082.4 1,389.5 1,860.6 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 83 Zambia Health Sector Public Expenditure Review 83 Annex Table F. Donor Health Expenditures by Type of Providers and Administrators, 2002­04 (ZK Billion) Items 2002 2003 2004 Total 02­04 % Share MOH and other public institutions 195.5 317.3 437.2 950.0 57.4 Donors acting as financing agents (i.e., parallel financing) I Defense Force hospitals 0 0 3.4 3.4 0.2 I Private for-profit hospitals 1.4 3.0 6.6 11.0 0.7 I Mission hospitals 2.5 3.5 2.8 8.8 0.5 I Ambulatory health care 4.3 7.3 3.0 14.6 0.9 I Drug stores & chemists 0 0 2.7 2.7 0.2 I Other public health providers 48.9 70.0 64.2 183.1 11.1 I Otherproviders'healthadministration 8.4 12.8 52.4 73.6 4.4 I Research 13.1 19.5 16.6 49.2 3.0 I Training 1.1 2.7 1.1 4.8 0.3 I Other rel. svcs. & institutions 39.8 70.4 183.7 293.9 17.8 I Rest of the world 21.1 21.9 16.4 59.5 3.6 Sub-total, parallel financing 140.6 211.2 352.8 704.7 42.6 Grand total 336.2 528.5 790.1 1,654.8 100.0 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 84 84 A World Bank Country Study Annex Table G. MOH/CBOH Health Expenditures by Level of Care and Administration, 1999­2004 (ZK Billion) Items 1999 2000 2001 2002 2003 2004 Sources of Funds MOFNP 140.9 147.3 288.2 342.4 327.1 322.3 Donors 29.2 84.9 82.3 195.5 309.9 413.2 Total 170.1 232.1 370.6 538.0 637.0 735.5 Application of Funds Service provision, by level of care MOH tertiary hospitals 41.0 51.6 53.3 76.1 88.2 88.1 MOH secondary hospitals 7.1 15.1 31.4 36.9 45.3 38.3 MOH district hospitals 42.9 83.7 207.8 243.4 261.5 303.9 ZCCM hospitals 0.2 0 0 0 0 0 Private for-profit hospitals 0.4 0.3 0.5 0.5 0.7 1.1 Mission hospitals 2.0 3.2 5.3 6.5 6.5 6.1 Nursing and residential care 0 0 0 0.2 0.5 0.8 Ambulatory health care 7.8 5.9 4.7 11.8 31.5 14.5 Drug stores & chemists 0 0 0 0 0.9 5.6 Other public health providers 0.3 0 0 0 0 0 Subtotal (a) 101.8 159.7 303.1 375.4 435.1 458.3 Administration MOH/CBOH health administration 59.4 73.6 49.3 146.1 173.0 259.1 Statutory boards 0 0 0 1.3 1.1 1.9 Other providers of health admin. 3.9 2.6 20.5 8.8 19.4 9.8 Other related svcs & inst. 0 0.4 3.2 1.6 2.9 1.4 Subtotal (b) 63.3 76.6 69.8 156.2 193.4 270.8 Research 1.8 2.2 4.4 4.1 2.8 4.2 Training 3.9 4.3 0.5 0.8 2.8 0.4 Subtotal (c) 5.7 6.5 4.8 4.8 5.6 4.6 Grand total (a + b + c) 170.8 242.9 380.9 537.9 637.0 735.0 Errors and omissions (0.7) (10.7) (10.3) Negl. Negl. (0.5) WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 85 Zambia Health Sector Public Expenditure Review 85 Annex Table H. Household Health Expenditures by Type of Providers, 1999­2004 (ZK Million) Items 1999 2000 2001 2002 2003 2004 MOH tertiary hospitals 4,310 6,644 8,158 9,546 10,223 13,709 MOH secondary hospitals 552 1,830 1,954 2,229 2,158 3,600 MOH district facilities 2,052 2,751 2,923 4,357 5,511 7,886 ZCCM hospitals 2,976 4,510 0 0 0 0 Private for-profit hospitals 38,095 50,192 80,868 102,103 154,493 233,766 Mission hospitals 172 179 86 89 280 265 Ambulatory health care 325 131 138 150 176 181 Traditional healers 36,250 45,616 52,400 72,230 87,687 103,444 Drug stores & chemists 86,289 108,506 96,178 107,362 130,338 153,760 Other pub. Health providers 5,492 0 0 0 0 0 Service provision 176,513 220,359 242,705 298,066 390,866 516,610 Administration 1,888 3,418 5,446 7,280 4,472 7,677 Health insurance 0 0 0 3,580 4,227 4,616 Grand total 178,401 223,777 248,151 308,926 399,565 528,903 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 86 86 A World Bank Country Study Annex Table I. Health Expenditures by Line Items by Type of Facility, 1999­2004 (ZK Million) Personal Drugs & Other Recur. Capital Years Emol. Supplies Transport Expend. Expend. Total MOH Primary Facilities 1999 28,032 5,881 4,982 5,819 1,009 45,722 2000 34,659 29,155 7,301 13,509 2,757 87,381 2001 107,050 45,869 13,013 37,946 12,050 215,927 2002 134,734 55,602 14,240 52,110 10,600 267,285 2003 146,607 44,696 14,611 57,220 16,018 279,152 2004 155,244 77,969 20,526 68,171 15,667 337,577 Ave. share 50% 21% 7% 18% 4% 100% MOH Secondary Facilities 1999 5,086 1,320 485 1,200 28 8,118 2000 8,951 4,983 883 2,927 856 18,600 2001 23,301 5,521 137 1,234 1,423 31,616 2002 32,162 5,417 234 1,294 212 39,320 2003 31,599 11,069 1,549 6,593 730 51,540 2004 29,198 10,005 1,968 6,395 1,042 48,608 Ave. share 65% 19% 3% 11% 2% 100% MOH Tertiary Facilities 1999 34,690 6,287 759 5,481 873 48,089 2000 38,622 10,687 1,074 6,222 1,092 57,698 2001 35,225 11,380 1,214 10,386 3,271 61,475 2002 43,241 26,992 1,432 15,842 4,064 91,571 2003 61,435 24,725 2,005 16,275 1,347 105,787 2004 79,054 25,541 2,682 17,475 2,038 126,790 Ave. share 61% 21% 2% 14% 2% 100% Mission Facilities 1999 1,040 433 227 581 470 2,751 2000 1,580 1,337 231 1,228 153 4,529 2001 3,528 1,510 77 431 256 5,802 2002 4,912 1,752 62 2,752 28 9,507 2003 4,723 2,115 569 5,059 163 12,629 2004 4,739 1,405 453 3,994 259 10,848 Ave. share 48% 21% 4% 22% 5% 100% WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 87 Zambia Health Sector Public Expenditure Review 87 Annex Table J. Resources at the Disposal of a Typical DHMT, FY05 US$ % Share Items ZK Million (at 4,2000 per 1 USD) of Health Services Personal emoluments 32,540.2 7,747,667 -- Administration 12.6 3,000 -- Health services 165.0 39,285 100.0 I First referral 41.6 9,904 25.2 I Child health 22.9 5,452 13.9 I HIV/AIDS 17.6 4,190 10.7 I Malaria 18.7 4,452 11.3 I Maternal health 20.7 4,929 12.5 I Tuberculosis 15.5 3,690 9.4 I Water & sanitation 17.3 4,119 10.5 I Mental health 4.4 1,048 2.7 I Oral health 6.3 1,500 3.8 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 88 88 A World Bank Country Study Annex Table K. Self-Reported Versus Actual Capacity to Deliver Selected Health Services in Rural Health Centers, 2006 Assessment of Actual Capacity of RHCs Based on PET/QSDS Indicators Self-Reported Availability of Drugs and Availability of Health Services Capacity Other Consumables Capital Assets Family planning 98% 1% reported condoms not 24% reported not having available today; private counselling area; 13% reported pills not 6% reported not having available today; private examination area 9% reported injectables not available today Antenatal and 99% 13% reported Vitamin 43% reported not having postnatal care A not available today; mothers' waiting area 7% reported iron folate for anemia not available today Child 94% RHCs reported the following 7% reported refrigeration immunization vaccines were not available for cold chain not available today: 21% for BCG; 10% for polio; 24% for pentavalent; 10% percent for measles; 10% for tetanus toxoid Child health/IMCI 93% 13% reported cotrimoxazole for 7% reported not having pneumonia not available today; weighing scale 2% reported ORS for diarrhea 54% reported not having not available today; height measuring device 13% reported Vitamin A not available today Adult malaria 94% 56% reported malaria slides 70% reported not having diagnosis and not available today lab equipment treatment 76% reported not having 74% reported not having malaria smears microscopes 29% reported Coartem not available today WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 89 Zambia Health Sector Public Expenditure Review 89 Annex Table L. Self-Reported Versus Actual Capacity to Deliver Selected Health Services in Urban Health Centers, 2006 Assessment of Actual Capacity of RHCs Based on PET/QSDS Indicators Self-Reported Availability of Drugs and Availability of Health Services Capacity Other Consumables Capital Assets Family planning 100% 4% reported condoms not 16% reported not having available today; private counselling area; 2% reported pills not 11% reported not having available today; private examination area 9% reported injectables not available today Antenatal and 95% 15% reported Vitamin A 42% reported not having postnatal care not available today; mothers' waiting area 8% reported iron folate for anemia not available today Child 95% UHCs reported the following 5% reported refrigeration for immunization vaccines were not available cold chain not available today: 11% for BCG; 9% for polio; 31% for pentavalent; 9% percent for measles; 13% for tetanus toxoid Child 97 20% reported cotrimoxazole 3% reported not having health/IMCI for pneumonia not available weighing scale today; 45% reported not having 13% reported ORS for diarrhea height measuring device not available today; 15% reported Vitamin A not available today Adult malaria 100 38% reported malaria slides 58% reported not diagnosis and not available today having lab equipment treatment 58% reported not having 55% reported not malaria smears having microscopes 40% reported Coartem not available today WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 90 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 91 References Central Board of Health (CBOH). 2002. Health Institutions in Zambia: A Listing of Health Facilities According to Levels and Locations, 2002. Lusaka. Central Board of Health (CBOH), University of Zambia (UNZA), and International Health Economics (IHE). 2004. Costs of Basic Health Care Package for 1st, 2nd, and 3rd Levels of Referral in Zambia. Lusaka. Central Statistical Office. 1998. Living Conditions in Zambia. Lusaka. ------. 2002. Zambia Demographic and Health Survey 2001/02. Done in cooperation with the Central Board of Health and ORC/Macro. ------. 2003. Zambia 2000 Census of Population and Housing. Summary Report. Lusaka. Das Gupta, Monica, Varun Gauri, and Stuti Khemani. 2003. "Decentralized Delivery of Primary Health Services in Nigeria: Survey Evidence from the States of Lagos and Kogi." Africa Region Human Development Working Paper Series, The World Bank, Washington, D.C. International Monetary Fund (IMF). 2007. "Zambia Case Study." Preliminary Draft, February 20. Kombe, G., and O. Smith. 2003. "The Costs of Anti-Retroviral Treatment in Zambia." Partnership for Health Reform Project, Abt Associates, Bethesda, Maryland, U.S.A. Lindelow, Magnus. 2006. Tracking Public Money in the Health Sector in Mozambique: Con- ceptual and Practical Challenges. Washington, D.C.: The World Bank. Ministry of Health. 2005a. 2005 Annual Health Statistical Bulletin. Directorate of Planning and Development, Lusaka. ------. 2005b. Human Resources for Health Strategic Plan (2006­2011). Lusaka. ------. 2006a. 2005 Budget Status by Activity With Supplementary Funding. Lusaka. ------. 2006b. Zambia National Malaria Indicator Survey. Lusaka. Mphuka, P. 2005. "The Costs of Meeting the MDGs in Zambia." Research commissioned by the CSPR, JCTR and CCJDP. Lusaka. Mumbwali, S. 2006. "Funding Analysis, 2005 and 2006." DfID/Zambia. Mwikisa, C.N., L. Mwansa, P. Nankamba, D. Chimfwembwe, M. Goma, B. Chitah, and R. Maswenyeho. 2000. Zambia National Health Accounts 1995­1998. Ministry of Health and Central Board of Health. National AIDS Council. 2006. The Zambia HIV/AIDS Strategic Framework 2006­2010. Lusaka. Phiri, Felix, and Marie Tien. 2004. Zambia National Health Accounts 2002: Main Findings. Ministry of Health and PHRplus Project. Reinikka, Ritva, and Jakob Svensson. 2004. "Local Capture: Evidence from a Central Gov- ernment Transfer Program in Uganda." Quarterly Journal of Economics. ------. 2006. "Using Micro Surveys to Measure and Explain Corruption." World Devel- opment 34(2). 91 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 92 92 A World Bank Country Study University of Zambia (UNZA), Ministry of Health, and Institute of Health Economics (IHE). National Health Accounts for Zambia, 2002­2004. Institutional Collaboration - Health Economics Policy analysis and Health Economics Project. World Bank. 2001. "Zambia Public Expenditure Review: Public Expenditure, Growth, and Poverty--A Synthesis." Southern Africa, Africa Region, The World Bank. ------. 2004. World Development Report 2004: Making Services Work for the Poor. Wash- ington, D.C. Yates, Rob. 2006. "Ensuring Efficient and Equitable Allocation of the 2006 Flexible Health Fund." DfID/Zambia. WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 93 WB Zambia.qxd:WB Zambia.qxd 11/27/08 2:44 PM Page 94 Eco-Audit Environmental Benefits Statement The World Bank is committed to preserving Endangered Forests and natural resources. We print World Bank Working Papers and Country Studies on 100 percent postconsumer recy- cled paper, processed chlorine free. The World Bank has formally agreed to follow the rec- ommended standards for paper usage set by Green Press Initiative--a nonprofit program supporting publishers in using fiber that is not sourced from Endangered Forests. For more information, visit www.greenpressinitiative.org. In 2007, the printing of these books on recycled paper saved the following: Trees* Solid Waste Water Net Greenhouse Gases Total Energy 264 12,419 96,126 23,289 184 mil. * Pounds Gallons Pounds CO2Equivalent BTUs 40' in height and 6­8" in diameter Zambia Health Sector Public Expenditure Review is part of the World Bank Country Study series. These reports are published with the approval of the subject government tocommunicate the results of the Bank's work on the economic and related conditions of memercountries to governments and to the development community. This book portrays the performance of the health sector in Zambia using quantitative techniques. While there have been a number of health sector assessments in the country, they have relied on qualitative and anecdotal evidence for the most part. For the first time, this pubic expenditure review of the health sector brings together the results of three separate but related analytical efforts: multi-year national health accounts, a public expenditure tracking and quality of service delivery survey, and resource and impact modeling using the Marginal Budgeting for Bottlenecks software. These exercises combine to yield more powerful findings on the weaknesses and prospects of the Zambian health system. World Bank Country Studies are available individually or by standing order. Also available online through the World Bank e-Library (www.worldbank.org/elibrary). ISBN 978-0-8213-7804-5 THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org SKU 17804 E-mail: feedback@worldbank.org