HAS RESOURCE ALLOCATION POLICY CHANGE IMPROVED EQUITY? LESSONS FROM GHANA
AUGUSTINE DANSO ASANTE
A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
School of Public Health and Community Medicine Faculty of Medicine The University of New South Wales March 2006 CERTIFICATE OF ORIGINALITY
I hereby declare that this submission is my own work and to the best of my knowledge and belief, it contains no material previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis.
I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.
…………………………
Augustine Danso Asante
ii ACKNOWLEDGEMENTS
This project would not have been possible without the overwhelming support of my supervisors Professor Anthony Zwi and Dr. Maria Theresa Ho. Their guidance and encouragement coupled with critical but constructive feedbacks steered the project to its completion. Anthony and Tessa went beyond the call of supervisory duty over the course of this project. I am particularly grateful for the key role they played in ensuring that fieldwork in Ghana could be undertaken. I also thank my previous co-supervisor Dr. Anna Whelan for her support in the early stages of my candidature.
I am indebted to The University of New South Wales for the financial support it provided towards this study. Without the University International Postgraduate Award (UIPA), which I was priviledged to enjoy for two and half years, it would have been very difficult to complete this project. I am also pleased to acknowledge the financial and logistic support provided by the School of Public Health and Community Medicine to this project.
During my fieldwork in Ghana, I enjoyed the help and support of so many people. I am pleased to acknowledge the support of Dr. Elias Sory (Regional Director of Health Services – Northern Region), Dr. Kofi Asare (Regional Director of Health Services – Ashanti Region) Mr. Yakubu Zakaria (Ministry of Finance and Economic Planning), and many others who in diverse ways facilitated the data collection. My friends in Sydney have also been very supportive. I specially thank Dr. Alfred Yawson and wife, Mr. George Asabre and family and Mrs. Yolanda Fenandez and family for their support. Dr. Yawson was particularly instrumental in helping me think through the statistical analysis in this study. I also acknowledge the support of my colleagues and fellow PhD students at the SPHCM, particularly the ‘night team’.
Finally, I owe a special debt of gratitude to my family, most especially to my wife and children who have equally paid a hefty price for my long absence from home. I hope one day the kids will grow to appreciate this sacrifice and pardon me for being an “absentee” father.
This work is dedicated to the memory of my parents, particularly my father, whose advice and words of encouragement continue to guide and inspire me in this difficult pursuit of knowledge. I will forever be grateful to him for setting me on this path despite the challenges.
iii CONTENT
Page Declaration ii Acknowledgements iii List of Tables ix List of Figures xi List of Boxes xii Acronyms xiii Abstract xv
Chapter 1: Decentralisation, Resource Allocation and Equity: An Introduction to the Study Overview...... 1 1.1 Background...... 1 1.2 Factors Driving Reform in the Health Sector ...... 3 1.3 Health Sector Decentralisation...... 6 1.4 Equity and Resource Allocation ...... 8 1.4.1 Interpretation of Equity ...... 9 1.5 Equity in Resource Allocation under Decentralisation in Ghana ...... 11 1.5.1 Purpose and Main Research Questions ...... 11 1.5.2 Scope and Outline of Study...... 12
Chapter 2: Ghana and Its Health Care System Overview...... 15 2.1 Geography, Population and Administrative Divisions...... 15 2.2 Political Context...... 17 2.3 Economic Context...... 19 2.4 Structure of the Health Care System...... 21 2.5 Health Status ...... 24 2.6 Inequalities in Health ...... 25 2.6.1 Inequalities in Access to Facilities ...... 26 2.6.2 Inequalities in Child Mortality ...... 27 2.7 Health Policy Objectives...... 28 2.8 Health Sector Reforms...... 30 2.8.1 Decentralisation and Re-organisation of the Health System ...... 30 2.8.2 Decentralisation of Resource Allocation Decision-Making...... 32 2.8.3 Implementation of Budget Management Centre (BMC) Concept...... 33
iv Chapter 3: Health and Social Justice: A Search for a Theoretical Framework Overview...... 37 3.1 Introduction...... 37 3.2 Utilitarian Theory...... 41 3.2.1 Critique of Utilitarian Theory...... 41 3.3 Entitlement Theory ...... 43 3.3.1 Critique of Entitlement Theory ...... 45 3.4 Rawls’ Theory of Justice: The Framework for Current Study...... 46 3.4.1 The Rawlsian Theory and Health ...... 48 3.4.2 Equality of Opportunity and Health ...... 50 3.5 Theories of Social Justice and this Study...... 52
Chapter 4: Equity and Equitable Distribution of Health Care Resources Overview...... 59 SECTION 1 - EQUITY: CONCEPT, INTERPRETATION AND MEASUREMENT ...... 59 4.1 Equity and Equality: A Conceptual Clarification ...... 59 4.2 Defining Equity...... 61 4.2.1 Horizontal and Vertical Equity...... 61 4.2.2 Equal Treatment for Equal Need...... 63 4.2.3 Equality of Access (Equal Access for Equal Need) ...... 63 4.2.4 Equality of Health ...... 65 4.3 Measuring Equity...... 66 4.4 Defining Health Need ...... 68 4.4.1 Measurement of Need...... 69 4.4.2 Deprivation as a Proxy of Need...... 71 4.4.3 Measurement of Deprivation...... 72 4.5 Towards a Working Definition of Equity ...... 73 SECTION 2: EQUITABLE ALLOCATION OF HEALTH CARE RESOURCES...... 75 4.6 Historically-Based Funding ...... 75 4.7 Population and Needs-based Allocation ...... 76 4.7.1 The Resource Allocation Working Party (RAWP) ...... 77 4.7.2 RAWP Outside the United Kingdom ...... 81 4.8 Beyond Needs-Based Allocation ...... 85
Chapter 5: Decentralisation and Equity in Resource Allocation Overview...... 87 5.1 Definition of Decentralisation...... 87 5.2 Models of Decentralisation ...... 89 5.2.1 Devolution...... 89 5.2.2 Delegation…………………………………………………………………………….91 5.2.3 Deconcentration...... 92 5.2.4 Privatisation...... 93 5.3 Decentralisation and Health Policy in Developing Countries...... 94 5.4 Decentralisation and Donor Agencies...... 95 5.5 Conflicting Claims about Decentralisation and Equity...... 97 5.6 Evidence of Impact of Decentralisation on Equity ...... 98 5.7 Appraisal of Evidence...... 102
v Chapter 6: Research Design and Methods Overview...... 104 6.1 Research questions and study objectives ...... 104 6.2 Research Design...... 105 6.2.1 Selection of Study Regions ...... 109 6.2.2 Rationale for Selecting Ashanti and Northern Regions...... 109 6.3 ADDRESSING THE QUANTITATIVE OBJECTIVE...... 111 6.3.1 Principal Component Analysis and Relative Deprivation...... 113 6.3.2 Equity-Adjusted Share (EAS) ...... 116 6.3.3 Simple Regression...... 117 6.3.4 Comparison of Predicted and Actual Allocations ...... 117 6.3.5 Quantitative Data Collection and Management...... 118 6.3.6 Summary of Quantitative Approach...... 119 6.4 ADDRESSING THE QUALITATIVE OBJECTIVE ...... 121 6.4.1 Purposive Sampling of Districts...... 121 6.4.2 Rationale for Selection of Districts in Northern Region ...... 122 6.4.3 Rationale for Selection of Districts in Ashanti Region ...... 125 6.4.4 Qualitative Data Sources and Recruitment of Participants...... 128 6.4.5 Organisation and Analysis of Qualitative Data ...... 131 6.4.6 Interpretation of Qualitative Data...... 134 6.5 Methodological Issues ...... 135
Chapter 7: Equity in Resource Allocation in Ghana: Result of Documentary Analyses Overview...... 137 7.1 Sources of Funding for the Health System ...... 137 7.1.1 Donor Funding ...... 141 7.2 Allocation of Resources in the Health Sector ...... 143 7.2.1 Resource Allocation by Line Items ...... 143 7.2.2 Resource Allocation by Level of Health System...... 145 7.2.3 Allocation by Geographical Area...... 145 7.2.3.1 Distribution of Salary Budget ...... 146 7.2.3.2 Distribution of Internally Generated Funds (IGF) ...... 148 7.3 Mechanisms for Disbursing Funds ...... 149 7.3.1 Mechanisms for Disbursing Funds in Ashanti Region...... 150 7.3.2 Mechanisms for Disbursing Funds in Northern Region...... 151 7.4 Mechanisms for Accessing Funds...... 152 7.5 System of Financial Accountability...... 154
Chapter 8: Geographical Equity in Resource Allocation in Ghana: Quantitative Results SECTION 1: INTER-REGIONAL EQUITY IN RESOURCE ALLOCATION...... 156 8.1 Deprivation among Regions...... 156 8.1.1 Principal Component Analysis and GID ...... 157 8.1.2 The Double-Index of Deprivation (DID) ...... 160 8.1.3 Correlation between Deprivation Indices and Mortality Rates ...... 162 8.2 The Equity-Adjusted Share (EAS) ...... 163 8.3 Assessing Inter-Regional Equity in Resource Allocation...... 164 8.3.1 Actual Allocation of Government Funds...... 165 8.3.2 Regression Analysis ...... 166 8.3.3 Comparison of Actual and Predicted Allocations ...... 167 SECTION 2: INTER-DISTRICT EQUITY IN RESOURCE ALLOCATION ...... 170
vi 8.4 Deprivation among Districts in Ashanti and Northern Regions ...... 170 8.4.1 Analysis of Deprivation by Region ...... 173 8.4.2 Rurality and Deprivation ...... 174 8.5 Equity-Adjusted Share (EAS) ...... 176 8.5.1 Ashanti Region...... 176 8.5.1 Northern Region...... 177 8.6 Assessing Inter-District Equity of Resource Allocation ...... 178 8.6.1 Actual Allocation of Government of Ghana Funds...... 178 8.6.2 Actual Allocation of Donor-Pooled Funds (DPF) ...... 179 8.6.3 Regression Analysis ...... 181 8.6.4 Comparison of Actual and Predicted Allocations ...... 183 8.6.5 Actual Allocation of Government of Ghana Funds (GOG 3) ...... 188 8.6.6 Actual Allocation of Donor-Pooled Funds (DPF) ...... 189 8.6.7 Regression Analysis ...... 190 8.6.8 Comparison of Actual and Predicted Allocations ...... 192 SECTION 3: INTER- SUB-DISTRICT EQUITY IN RESOURCE ALLOCATION...... 197 8.7 Deprivation among Sub-Districts...... 197 8.7.1 West Gonja...... 197 8.7.1 Savelugu-Nanton ...... 198 8.8 Equity-Adjusted Share (EAS) ...... 199 8.8.1 West Gonja...... 200 8.8.2 Savelugu-Nanton ...... 200 8.9 Assessing Inter-Sub-District Equity in Resource Allocation...... 201 8.9.1 Actual Allocation of Funds in West Gonja ...... 201 8.9.2 Regression Analysis: West Gonja ...... 202 8.9.3 Comparison of Actual and Predicted Allocations in West Gonja ...... 203 8.9.4 Actual Allocation of Funds in Savelugu-Nanton ...... 204 8.9.5 Regression Analysis: Savelugu-Nanton ...... 204 8.9.6 Comparison of Actual and Predicted Allocations in Savelugu Nanton...... 205
Chapter 9: Behind the Algorithms: Factors Influencing Equitable Resource Allocation in Ghana Overview...... 207 9.1 Equity and Equitable Allocation of Resources ...... 208 9.1.1 Equity as a National Health Priority...... 208 9.1.2 Manpower Availability...... 209 9.1.3 Capacity to Utilise Funds Efficiently ...... 211 9.2 Level of Funding for the Health System...... 213 9.3 Mechanisms for Disbursing and Accessing Funds ...... 216 9.3.1 Timing of Release of Funds ...... 216 9.3.2 Procedural Inefficiencies...... 220 9.3.3 Government Accounting and Financial Regulations...... 221 9.4 System of Financial Accountability...... 224 9.5 Politics and Equity in Resource Allocation ...... 225 9.6 Donor Influence on Resource Allocation...... 226 9.6.1 Different Perceptions about Donor Funding ...... 227 9.6.2 Donor Influence on Policy Development ...... 229 9.6.3 Donor Use of Earmarked Funding ...... 231 9.7 Collaboration with Local Government (District Assembly) ...... 232 9.7.1 Transparency and Misconceptions about Funding Levels...... 233 9.7.2 Possible Avenues of Collaboration ...... 235 9.7.3 District Assembly’s Influence on Resource Allocation ...... 236
vii Chapter 10: Discussion Overview...... 238 10.1 To what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? ...... 238 10.1.1 Variations in deprivation levels ...... 239 10.1.2 Mechanisms for resource allocation ...... 244 10.1.3 Interpretation of equity ...... 251 10.1.4 Fragmentation of the resource allocation system...... 253 10.2 What factors influenced the equitable allocation of resources for health care in Ghana?.254 10.2.1 Low levels of funding of the health system ...... 255 10.2.2 Local capacity to utilise funds efficiently...... 257 10.2.3 Donor influence ...... 259 10.2.4 Politics ...... 260 10.2.5 Nature of collaboration with the local government (District Assembly)...... 262 10.3 Study Limitations ...... 264
Chapter 11: Conclusions and Suggestions for Further Research Overview...... 268 11.1 A Brief Review of the Study...... 268 11.2 Policy Implications of the Study...... 270 11.2.1 Examination of Equity in Resource Allocation...... 271 11.2.2 Decentralisation and Equity Policies in Ghana ...... 272 11.2.3 Inequitable Capacity Constraints...... 273 11.2.4 Effective Collaboration under Decentralisation ...... 274 11.3 Contributions to the Equity and Resource Allocation Debate ...... 275 11.3.1 Conceptual Contributions...... 275 11.3.2 Methodological Contributions...... 276 11.4 Directions for Future Research ...... 278
Appendix A: Pearson’s correlation coefficients between demographic and socio-economic variables for the ten regions in Ghana…………………………………………………………………280
Appendix B: Pearson’s correlation coefficients between demographic and socio-economic variables for Ashanti and Northern Region …………………………………………………………...281
Appendix C: Percentage Change in Actual Government of Ghana Funding (GOG 2-3) in Ashanti Region: 1998 – 2002…………………………………………………………………….282
Appendix D: Percentage Change in Actual Donor-Pooled Funds (DPF) in Ashanti Region: 1998 – 2002…………………………………………………………...283
Appendix E: Percentage Change in Actual Government of Ghana Funding (GOG 2-3) in Northern Region: 2000 – 2002…………………………………………………………………….284
Appendix F: Percentage Change in Actual Donor-Pooled Funds (DPF) in Northern Region: 1999 – 2002…………………………………………………………..285
Appendix G: List of Documents Reveived Prior to Selection of Deprivation Variables….…286
Bibliography…………………………………………………………………………………. 287
viii LIST OF TABLES Page Table 2. 1 Ghanaian Regimes Since Independence...... 18 Table 2. 2 Exchange Rate of the Ghanaian Cedi to the US Dollar...... 20 Table 2. 3 Selected Health-related Indicators...... 25 Table 2. 4 Regional Distribution of Health Facilities...... 27 Table 2. 5 Changes in Health Resource Allocation Policy in Ghana ...... 34 Table 5. 1 Types and Definition of Decentralisation...... 89 Table 6. 1 Variables for Measuring Deprivation among Jurisdictions...... 114 Table 6. 2 Selected demographic and socio-economic characteristics of districts in Northern Region ... 122 Table 6. 3 Selected demographic and socio-economic characteristics of districts in Ashanti Region ..... 125 Table 6. 4 List of Participants ...... 129 Table 7. 1 Health Expenditure and Sources of Funding: 1996-2001...... 138 Table 7. 2 Funding Allocations to Ministry of Health and Per Capital spending: 1996- 2002 ...... 140 Table 7. 3 Resource Allocation by Line Items: 2003-2004...... 144 Table 7. 4 Health Sector Non-Wage Recurrent Budget Allocation: Percentage Distribution 2002-2005 145 Table 7. 5 Regional Distribution of Ghana Health Service Doctors and Nurses 2002...... 146 Table 7. 6 Actual and Per capita Allocation of Salary Budget by Region: 2002 - 2003 ...... 147 Table 7. 7 Internally Generated Funds (IGF) by Region - 2002 ...... 148 Table 7. 8 Resource Allocation Mechanisms for the Ashanti Region 1998-2001...... 151 Table 7. 9 Resource Allocation Mechanisms for the Northern Region 1999-2002...... 152 Table 7. 10 Mode of Disbursing and Accessing of Funds...... 153 Table 8. 1 Component Matrix: All Regions ...... 158 Table 8. 2 Variables by Weight/Coefficient...... 159 Table 8. 3 Regions by General Index of Deprivation...... 159 Table 8. 4 Regions by Double-Index of Deprivation (DID)...... 160 Table 8. 5Regions by General Index of Deprivation (GID) and ...... 161 Table 8. 6 Pearson’s Correlation Coefficients of Deprivation Indices...... 162 Table 8. 7 Derivation of the Equity-Adjusted Share Index (EAS) for Inter-Regional Resource Allocation in Ghana...... 164 Table 8. 8 Region by Actual and Percentage Shares of Government Funds (GOG 2-4): 2002-2003 ...... 166 Table 8. 9 Regression of Equity-Adjusted Share and Population on Percentage Share of Government of Ghana (GOG) Funds: 2002 - 2003...... 167 Table 8. 10 Regions by Predicted (EAS-Based) Government Funding (GOG 2-4) and Percentage of Funding Difference 2002 - 2003 ...... 168 Table 8. 11Component Matrix: Northern and Ashanti...... 170 Table 8. 12 Variables by Weight/Coefficient...... 171 Table 8. 13 Districts in Ashanti (AR) and Northern Region (NR) Ranked by their General Index of Deprivation (GID)...... 172 Table 8. 14 Derivation of Equity-Adjusted Share (EAS) for Ashanti Region...... 176 ix Table 8. 15 Derivation of Equity-Adjusted Share (EAS) for Northern Region...... 177 Table 8. 16 Districts by Actual and Percentage Shares of Government Funds (GOG 2-3): Ashanti Region 1998 - 2002 ...... 179 Table 8. 17 Districts by Actual and Percentage Shares of Donor-Pooled Funds (DPF): Ashanti Region 1998 to 2002...... 180 Table 8. 18 Regression of Equity-Adjusted Share and Population on Percentage Share of Government of Ghana (GOG) Funds: 1998 – 2002 ...... 181 Table 8. 19 Regression of Equity-Adjusted Share and Population on Percentage Share of Donor-Pooled Funds (DPF): 1998 – 2002...... 182 Table 8. 20 Districts by Predicted (EAS-Based) Share of Government Funds (GOG 2-3): Ashanti Region 1998 - 2002 ...... 184 Table 8. 21 Districts by Predicted (EAS-Based) Share of Donor-Pooled Funds (DPF): Ashanti Region - 1998 to 2002...... 186 Table 8. 22 Districts by Actual Government Funds (GOG 3) in Northern Region: 2000 - 2002...... 188 Table 8. 23 Districts by Actual Donor Pooled Funds (DPF), Northern Region: 1999 – 2002 ...... 189 Table 8. 24 Regression of Equity-Adjusted Share and Population on Percentage Share of Government of Ghana (GOG) Funds: 2000 – 2002 ...... 190 Table 8. 25 Regression of Equity-Adjusted Share and Population on Percentage Share of...... 191 Table 8. 26 Districts by Predicted (EAS-Based) Government Funds (GOG-3) Northern Region: ...... 192 Table 8. 27 Districts by Predicted (EAS-Based) Donor Pooled Funds (DPF) Northern Region: ...... 194 Table 8. 28 West Gonja Sub-districts by Deprivation...... 197 Table 8. 29 Savelugu-Nanton Sub-districts by Deprivation...... 198 Table 8. 30 Derivation of Equity-Adjusted Share (EAS) for Sub-Districts in West Gonja ...... 200 Table 8. 31Derivation of Equity-Adjusted Share (EAS) for Sub-Districts in Savelugu-Nanton ...... 200 Table 8. 32 Sub-Districts by Actual Donor-Pooled Funds Allocation for West Gonja from 1999-2002..201 Table 8. 33 Regression of Equity-Adjusted Share and Population on Percentage Share of...... 202 Table 8. 34 Sub-Districts by Predicted (EAS-Based) Share of Donor-Pooled Funds in West Gonja: 2001- 2002...... 203 Table 8. 35 Actual Donor-Pooled Funds Allocation to Sub-Districts in Savelugu-Nanton 2001-2002....204 Table 8. 36 Summarised Results of Regression Analysis: Donor-Pooled Funds (DPF): Savelugu-Nanton 2001...... 205 Table 8. 37 Sub-Districts by Predicted (EAS-Based) Share of Donor-Pooled Funds in Savelugu-Nanton: 2001-2002 ...... 205
x LIST OF FIGURES Page
Figure 2. 1 A Map of Ghana Showing the 10 Administrative Regions, Total Population per Region and Number of Districts...... 16 Figure 2. 2 Organisational Structure of the Ghanaian Health Care System ...... 22 Figure 2. 3 Regions by Percentage of Poverty ...... 26 Figure 2. 4 Regional Under- 5 Mortality Profile: 1993 and 1998 ...... 28 Figure 2. 5 The Five Strategic Health Policy Objectives in Ghana...... 29 Figure 3. 1 Type of Moral Theories ...... 39 Figure 6. 1 Sequence of Mixed-Methods Approach used in the Study ...... 106 Figure 6. 2 The Four-Step Approach used to Address the Quantitative Objective ...... 119 Figure 7. 1Composition of Health Expenditure - 2001 ...... 138 Figure 7. 2 Composition of Health Expenditure - 2004 ...... 139 Figure 7. 3 Sources of Funding for the Health System 2002 - 2006 ...... 141 Figure 7. 4 Expected Total Contributions to the Health Budget by Major Donors: 2002-2006...... 142 Figure 7. 5 Allocation of Health Expenditure by Line Item - 2004...... 144 Figure 7. 6 Actual and Per capita Allocations of Salary Budget by Region: 2002-2003 ...... 148 Figure 7. 7 Funding Flow Chart...... 150 Figure 8. 1 Regions by Different Deprivation Indices ...... 162 Figure 8. 2 Percentage Difference between Actual and Predicted (EAS-Based) ...... 169 Figure 8. 3 Deprivation among Districts in Ashanti Region ...... 173 Figure 8. 4 Deprivation among Districts in Northern Region ...... 174 Figure 8. 5 Percentage Difference Between Actual and Predicted Shares of...... 185 Figure 8. 6 Percentage Difference Between Actual and Predicted Shares of...... 187 Figure 8. 7 Percentage Difference between Actual and Predicted Shares...... 193 Figure 8. 8 Percentage Difference between Actual and Predicted Shares...... 195 Figure 8. 9 Deprivation among Sub-Districts in West Gonja...... 198 Figure 8. 10 Deprivation among Sub-Districts in Savelugu- ...... 199 Figure 8. 11 Percentage Difference between Actual and Predicted ...... 203 Figure 8. 12 Percentage Difference between Actual and Predicted ...... 206
xi LIST OF BOXES
Box 4. 1 Various Definitions of Equity...... 59 Box 9. 1 Equity as a national health priority...... 208 Box 9. 2 Views on manpower shortage and incentive regimes...... 210 Box 9. 3 Account of a nurse/sub-district head about working conditions of village health workers ...... 211 Box 9. 4 Capacity to utilise funds efficiently...... 212 Box 9. 5 Efficiency concerns and IGF ...... 213 Box 9. 6 Views of policy makers and donors on level of funding ...... 214 Box 9. 7 Views of district managers on level of funding...... 215 Box 9. 8 Views on the erratic flow of funds to districts...... 217 Box 9. 9 Views on erratic funding flows at the national level ...... 218 Box 9. 10 Strategies use to cope with erratic funding flows ...... 219 Box 9. 11 A District Accountant’s Story of the Difficulties in Accessing GOG -2...... 221 Box 9. 12 Government regulations and accessing of funds by remote districts...... 223 Box 9. 13 Account of difficulties faced by remote districts in accessing cheque-disbursed funds ...... 223 Box 9. 14 Efficient use of funds and the flaws in the financial management system ...... 224 Box 9. 15 Political commitment to redistribute funds to improve equity...... 225 Box 9. 16 Perception of donor funding at the district level...... 227 Box 9. 17 Perception of donor funding at the regional level...... 228 Box 9. 18 Perception of donor funding at the national level...... 229 Box 9. 19 Donor influence on policy ...... 230 Box 9. 20 Donor earmarked funding...... 231 Box 9. 21 Complaints about behaviour of district health managers by assembly directors ...... 233 Box 9. 22 An account of a district health accountant highlighting the lack of support from the district assembly...... 234 Box 9. 23 Avenues of collaboration within the current policy framework ...... 235 Box 9. 24 Effective collaboration and resource support from the district assembly ...... 236
xii LIST OF PAPERS EMERGING FROM THIS STUDY
1. Asante AD, Zwi AB, Ho MT. Equity in Resource Allocation for Health: A Comparative Study of the Ashanti and Northern Regions of Ghana. Health Policy - Accepted
2. Asante AD, Zwi AB, Ho MT. Getting by on Credit: How District Health Managers in Ghana Cope with the Erratic Flow of Funds – In Progress
3. Asante AD. Addressing Geographical Inequities in Health through Resource Allocation: Lessons from Ghana. Fifth International Conference on Health Priorities, Wellington, November 2004. – Conference Paper
xiii ACRONYMS
AIDS Acquired Immune Deficiency Syndrome AFRC Armed Forces Revolutionary Council AHS Area Health Service ANC Ante-natal care AGC Ashanti Goldfields Company BMC Budget Management Centre CAGD Controller and Accountant General’s Department CHAG Christian Health Association of Ghana CHI Community Health Insurers CHPS Community-based Health Planning and Services CPP Convention People’s Party DA District Assembly DANIDA Danish International Development Agency DCE District Chief Executive DHA District Health Administration DHMT District Health Management Team DID Double Index of Deprivation DPF Donor Pooled Fund EAS Equity-Adjusted Share GDP Gross Domestic Product GID General Index of Deprivation GOG Government of Ghana GPRS Ghana Poverty Reduction Strategy GHS Ghana Health Service HIPC Heavily Indebted Poor Country HIV Human Immunodeficiency Virus HQ Headquarters IGF Internally generated funds (user fees/cash and carry) IMF International Monetary Fund IMR Infant Mortality Rate LGU Local Government Units MDA Ministries, Departments and Agencies MHO Mutual Health Organisations MMR Maternal Mortality Rate xiv MOH Ministry of Health MOF Ministry of Finance NDC National Democratic Congress NGO Non Governmental Organisation NHIC National Health Insurance Council NHIF National Health Insurance Fund NHIP National Health Insurance Programme NHS National Health Service NPP New Patriotic Party NSW New South Wales OECD Organisation for Economic Cooperation and Development PAHO Pan-African Health Organisation PCA Principal Component Analysis PHC Primary Health Care PNDC Provisional National Defence Council PNP Peoples National Party POW Program of Work PPME Policy, Planning, Monitoring and Evaluation PPP Private-Public Partnership PSU Private Sector Unit RAF Resource Allocation Formula RAWP Resource Allocation Working Party RDF Resource Distribution Formula RHA Regional Health Administration SAHFCF South African Health Functional Committee Formula SAP Structural Adjustment Programme SMC Supreme Military Council SMR Standardised Mortality Ratio SWAP Sector-Wide Approach UNDP United Nations Development Programme USAID United State Agency for International Development U5MR Under 5 Mortality Rate WDR World Development Report WHO World Health Organization
xv ABSTRACT
Equitable allocation of health care resources is crucial for promoting health equity. Since the emergence of the resource allocation working party (RAWP) formula nearly three decades ago, many countries have implemented resource allocation policy reforms aiming to improve equity. Little is known about whether, how and the extent to which, most of these policies have actually improved equity. This study examined whether, and the extent to which, decentralisation of health resource allocation decision-making in Ghana has improved equity in funding within regions and explored the factors that influenced the equitable allocation of resources for health care in Ghana.
The study used a mixture of quantitative and qualitative methods. Two of the ten regions in Ghana: Ashanti and Northern, covering the southern and northern sectors of the country, were purposefully selected. Principal component analysis (PCA) was used to measure levels of relative deprivation of districts applied as a proxy of need. An equity-adjusted share index (EAS) was developed and used as a yardstick against which equity in funding was assessed. Factors influencing the equitable allocation of resources were explored qualitatively through open-ended interviews with policy makers and other health sector stakeholders.
The study found that resource allocation in the Ashanti and Northern Regions were largely inequitable, in terms of differentially benefiting the most disadvantaged districts. The proportion of variance in the actual share of funds that could be explained by the predicted EAS was below 50% for all the years examined, except for the allocation of government funds to the Ashanti Region for 1999, where the proportion of variance was 56%. Resource allocation in the Northern Region favoured three urban districts over their rural counterparts. However, in the Ashanti Region, there was a significant shift in resources from richer to poorer districts from 2000 to 2002. The Kumasi Metro district, for example, saw its share of donor-pooled funds reduced drastically from 20% of the total budget in 2000 to 7.2% in 2001 and 5.6% in 2002. Key factors influencing resource allocation and equity included low funding of the health system, local capacity to utilise funds efficiently, manpower availability, politics, donor influence and the nature of collaboration with the local government.
The study concluded that intra-regional resource allocation in Ghana’s Ashanti and Northern regions was less equitable than expected, despite efforts to redistribute funds. It recommended more effective mechanisms for promoting equity through intra-regional resource allocation in Ghana.
xvi CHAPTER 1
DECENTRALISATION, RESOURCE ALLOCATION AND EQUITY: AN INTRODUCTION TO THE STUDY
“Health care reform is a worldwide phenomenon. Governments everywhere are rethinking how their systems are organised…how their systems can be reorganised to provide care more effectively and efficiently. Some countries have had more success than others, but we have found no single recipe for success” National Forum on Health- Canada 1997.
Overview
Health sector reform worldwide seeks to improve the effectiveness, efficiency and equity of health care delivery. In response to shifting priorities and challenges, including fiscal realities and changing health needs of populations, governments are redefining health policies, instituting new programmes, and changing existing legislation in an attempt to improve health system performance. Following two decades of global health sector reform that failed to address equity, many countries have introduced health policies aimed at promoting equity in health care delivery. Little is known about whether, and to what extent, these policies have actually improved equity. The purpose of this study is to provide empirical data to inform and support the process of health policy reform that seeks to achieve equity. The study is placed in the context of global health sector reform, with Section 1.1 focusing on definitional issues.
1.1 Background
Health sector reform has evolved as a central component of policy change in the past few decades. Governments worldwide are reviewing health policies, restructuring institutions, and looking for innovative ways to organise, finance and deliver health care (Sikosana et al. 1997). Policy discussions have largely centred on the development of more efficient services through initiatives such as distinguishing between the functions 1 of the purchaser and the provider, developing a mixture of public and private provision of services, reforming of financing and resource allocation mechanisms, giving greater autonomy to hospitals, developing district-based systems (Zwi and Mills 1995; Green et al. 2000).
Health sector reform is defined in various ways in the literature. Cassels (1995 p.331) saw health sector reform as concerned with “defining priorities, refining policies and reforming the institutions through which those policies are implemented”. At the Inter- country Meeting in Arusha Tanzania in 1995, health sector reform was defined as “a sustained process of fundamental change in policy and institutional arrangements, guided by government, designed to improve the functioning and performance of the health sector and ultimately, the health status of the population” (Sikosana et al. 1997 p.3). These definitions have several things in common. First, they suggest that health sector reform is about changing health policies and the institutions that facilitate implementation of those policies. Second, they emphasise that the changes are purposive, not haphazard and directed by government. Third, they indicate that health sector reform is a dynamic process (not a one-off activity) with an ultimate aim of improving people’s health.
Drawing on the above definitions, health sector reform is defined in this study, as a dynamic process, involving changes in health policies and institutions by government, that allow the health system to function at its best and, ultimately, improve people’s health. The ‘dynamic process’ highlights the fact that health sector reform is sustained and long-term rather than a one-off change. It also emphasises the fact that no particular set of prescriptions is universally accepted as constituting health sector reform. Walt and Gilson (1994) identified four key factors that should be considered in health sector reform: the context within which reform is implemented, the actors involved in or influenced by reform, the content of the reform and the process of implementation. They noted that the content of reform has largely been the focus of policy makers, ignoring the three other components. They argued that focusing on the content diverts attention from analysing the feasibility of implementing change. So what factors drive health sector reform that motivate policy makers to focus more on content of reform at the expense of context, process and actors? 2 1.2 Factors Driving Reform in the Health Sector The factors that drive health sector reform differ across countries. Mogedal et al. (1995) observed that health sector reform is usually implemented in response to a perceived dysfunction or serious constraints affecting the health system itself. While this assertion is not disputed, reform also occurs in an attempt to seek greater improvements in performance, rather than simply responding to system failure. Indeed, there is evidence that countries with well-functioning egalitarian health care systems, such as Sweden or the United Kingdom, have pursued reform that could be seen as prompted more by the pursuit of greater efficiency than the response to system failure1. Ideology has also played a major part. In many countries, the health reform process has been ideologically driven, with emphasis on market-like mechanisms, such as public/private partnerships, managed competition and integrated care (Blas and Limbambala 2001; Ham 2003). In the United Kingdom, for example, the separation of purchaser and provider was driven, in part, by an ideological preference for market-like mechanisms over bureaucracies (Ham 1997; 2003).
Between developed and developing countries, the factors that drive health sector reform differ because of disparities in economic and technological development, political traditions, demographic and epidemiological transitions, among other things. Reform in developed countries is designed to contend with sluggish economic growth, ageing populations, increasing popular expectations and rising costs associated with medical technology (OECD 1994)2. In developing countries, on the other hand, reform policies usually address or are expected to address issues of improving efficiency, extending access to underserved areas, improving service quality and addressing inequities in geographical allocation of resources (Cassels 1995 p.331).
1 Greater efficiency-oriented reforms in the Swedish health system during the last decade include reduction of health personnel, establishment of internal markets within and between hospitals, and encouragement of competition among providers (Andersen et al. 2001; Anell, 1996; WHO/EURO, 1996; Diderichsen, 1993). 2 In Canada, for example, health reform during in the 1990s was intended, among other things, to contain cost (WHO, 1996) while Australia has, since 1995, introduced several cost control and efficiency reforms including selective contracting, government subsidies for health insurance, and the move away from pure community rating (Willcox 2001). 3 The institutional context within which reform occurs also varies considerably. Developed countries largely have a functioning system of well-developed institutions as opposed to the relatively underdeveloped and weak institutions of developing countries. In spite of these dissimilarities, however, specific objectives of health sector reform are generally related, as they often fall within the five guiding principles of health reform adopted internationally as benchmarks for monitoring and measuring changes in the health sector namely; equity, efficiency, accessibility, quality, and financial sustainability (PAHO Policy Document, 1999).
In recent years, slower than expected economic growth in many parts of the world has resulted in a decline in financial resources available for health delivery (Schieber and Maeda 1999; WHO 2000; Mitton 2002). At the same time, changes in disease patterns (an increase in chronic diseases and re-emerging or new communicable diseases), changing population dynamics and rising popular expectations have resulted in a higher demand for services. This has created distributional dilemmas for policy makers in many countries. It used to be said that “health does not have a price”. However, the mismatch between resources and health needs has led many governments to a painful realisation that “health indeed does have its costs and that the inputs for producing better health are scarce economic resources that must be paid for and financed” (Vogel, 1988 p.16).
The quest for greater efficiency in the performance of health systems has also provided considerable impetus for health reforms. Efficiency-oriented reform in the health sector has been promoted by the ideological shift to neo-liberal macroeconomic policies (Gilson 1998b; Jimenez de la Jara & Bossert, 1995) and economic decline in many countries, which has reduced the ability of governments to finance health care services (Bennett, 1992; Gilson & Mills, 1995). In developed countries, efficiency objectives have been largely pursued through the introduction of market-like mechanisms into the health sector. In particular, experimentation with private sector management techniques in public health institutions in an attempt to enhance efficiency has been a typical reform measure. The move to out-source services previously provided by public facilities to private sector institutions, the policy to strengthen hospital management by
4 appointing managers to run facilities in a more professional fashion, and the use of budgetary incentives to improve performance were all adopted from successful private sector business operations (Ham 1997).
In developing countries, particularly in sub-Saharan Africa, the fall in budgetary resources due to poor economic performance has forced many governments to consider efficiency-oriented reforms in the public sector, including that of health. Managers of health care in developing countries have come under intense pressure to ensure that resources are allocated efficiently to achieve real “value for money”. The World Bank has been the main proponent of efficiency-oriented reforms in the health sector of developing countries. For example, the Bank has explicitly advocated for public sector health financing to be restricted to an essential package of services, which is cost- effective and maximises overall health gains (World Bank 1993). While there is little evidence, so far, that efficiency-oriented reforms have indeed enhanced efficiency, empirical data from several countries have suggested that such reforms may have worsened quality of services and/or equity in access to health care (Gilson 1997; Chawla and Ellis 2000).
A typical example is the user fee policy widely implemented in Africa and other developing countries. Although revenue generation is cited as the main objective for implementing the policy, it also aimed to reduce inefficiencies in health care delivery by decreasing the frivolous use of services, in particular, the use of referral facilities for first line health care services (Creese and Kutzin 1997; Breman and Shelton 2001). Nolan and Turbat (1995) observed that the standard model of user fees assumes that fees would not only produce resources, but would also offer efficiency and equity benefits. Efficiency benefits are expected to result from price signals, which offer consumers incentives for using the referral system appropriately, and facilitate the re- allocation of resources to more cost-effective primary health care. Equity benefits are expected to result from the use of resources in ways that benefit the poorest (including improving coverage and quality of PHC), and the use of exemptions or differential charges within the fee system to protect the poor from the full burden of user charges (Gilson et al. 1995).
5 Despite the potential benefits, user fees are regarded as the most regressive form of health care financing (Gilson and McIntyre 2005). Evaluative studies in many countries have shown that not only have user fees not promoted efficiency, it has also worsened equity. Studies by Kutzin (1995) on China, and McPake et al. (1992) on the Bamako Initiative, suggest that user fees have encouraged inefficient provider behaviour such as overuse of unnecessary services or poly-pharmacy, when the resulting revenue is retained at the point of collection. In Zimbabwe, for example, one third of the population in Murewa District used the hospital for first line health care services, despite a fee policy to discourage hospital use by non-referral patients (Criel et al. 1996; Criel 1998). In Ghana, Nyonator and Kutzin (1999) observed that the user fee policy failed to protect access and income for poorer members of the community, though it increased the operating revenues of health facilities. These and many other examples of negative impacts of the user fee policy in developing countries highlight the fact that efficiency-oriented reforms in the health sector may not indeed lead to efficiency and may seriously jeopardise equity3. Another key health reform policy that has been proposed to enhance efficiency and promote equity at the same time is decentralisation.
1.3 Health Sector Decentralisation Decentralisation has been a prominent policy on the health sector reform agenda in many countries. It refers to reform initiatives characterised by the transfer of fiscal, administrative, and/or political authority for planning, management, or service delivery from the central Ministry of Health (MOH) to alternative institutions (Bossert et al. 2000). Mills et al. (1990 p.11) defined decentralisation as “the transfer of authority, or a dispersal of power, in public planning, management and decision-making from the national level to sub-national levels”. The mode of the transfer depends on the type of decentralisation being pursued. The four main types of decentralisation have been described as devolution, delegation, deconcentration and privatisation (Rondinelli et al. 1983; see Chapter 5 for a detailed description of each of these types).
3 Recent health sector reforms in many African countries have sought to abolish user fees. However, Gilson and McIntyre (2005) have cautioned that removal of fees must be carefully managed; otherwise it may exacerbate the problems facing African health systems. In particular, they have called for increased funding of health systems before fees are removed. 6 Decentralisation, as a policy, is based on the assumption that decisions made at the local level better reflect the needs of local populations (Silverman 1990, Levaggi and Smith 2004). In other words, central authorities are assumed to be less informed about local needs and special circumstances and, therefore, tend to institute policies that often have little to do with realities on the ground. This notion underpins the growing shift from centralised administrative systems to increased transfer of decision-making and other responsibilities to local bodies.
Health sector decentralisation is promoted so as to achieve a variety of policy objectives including efficiency, equity, greater accountability and greater responsiveness to local needs (Mills et al. 1990, Standing 2002). It also reduces the participation of Ministry of Health (MOH) in the management and delivery of services, so that it can concentrate on policy formulation, monitoring, coordination and regulation (WHO 2000, World Bank 1993). Modifying the role of the state from one of provider to these latter roles has been a major feature of public sector reform in developing countries in recent decades (World Bank 1997).
Decentralisation policies have, in the past decade, been implemented in both developed and developing countries, but on a much broader scale in the later. The objectives of decentralisation in these countries differ depending on the health care setting. A number of recent health care reform initiatives in OECD countries have been based on decentralised systems of decision-making with respect to a wide range of planning and management activities in an attempt to increase efficiency, improve accountability, and give the public greater voice in decision-making (Evans, 1992; Saltman and Von Otter, 1992).
In the developing world, health sector decentralisation has been widely implemented during the last couple of decades as part of a broader process of political, economic and technical reform4 (Litvack et al., 1998). For many developing countries, particularly in sub-Saharan Africa, colonialism bequeathed a system of health care characterised by centralised decision-making and hospital-centred service provision. These features are
4 While developed countries have implemented decentralisation reform on their own volition, developing countries are often pressured by international agencies such as the World Bank to decentralise their health systems. 7 widely considered incompatible with the Health for All philosophy promoted through the Primary Health Care (PHC) concept (Ebrahim and Ranken 1988). Decentralisation is often proposed as an effective way to re-organise these health systems to make them more efficient and responsive to local needs (Segall 2003).
If properly implemented, as argued by the World Bank and other proponents, decentralisation can improve both technical and allocative efficiency (Hurley et al., 1995). Decentralisation has also been touted as promoting greater equity through distribution of resources towards traditionally marginalised regions and groups (Bossert et al., 2000; Bossert and Beauvais 2002). There is little empirical evidence supporting these theoretical arguments. The best current overview of the impact of decentralisation, particularly on equity, is provided by Bossert and Beauvais (2002) in their comparative analysis of health sector decentralisation in Ghana, Zambia, Uganda and the Philippines. Other works by Koivusalo (1999), Collins et al. (2000), Tang and Bloom (2000), Okuonzi (2004) and Costa-i-Font (2005) also provide useful empirical data on the impact of decentralisation on equity (see Chapter 5 for a review of these studies).
1.4 Equity and Resource Allocation Equity has been a key driver of health sector reform in recent years. There has been unceasing debate over how best to overcome inequity in health (Gwatkin 2000). One area of considerable emphasis is the equitable distribution of financial resources (Mooney 2000, McIntyre 2002). Mechanisms for allocation of public sector health resources in many countries are perceived as inequitable and blamed for contributing to maintaining existing inequities in health. Although the problems posed by resource scarcity cannot be underestimated, particularly in developing nations, there is a commonly held view that, among countries of similar socio-economic standing, it is not how much a country spends, but rather how it spends its resources, that determines the health status of its population (Yach and Harrison 1994; WHO 2000).
Evidence from both developed and developing countries suggests that inappropriate allocation of resources contributes greatly to inequities in health. In Australia, for example, although indigenous people have a life expectancy nearly 18 years shorter 8 than non-indigenous Australians (O’Donoghue 1999; Black and Mooney 2002), total expenditure per person for health services for indigenous Australians is less than the rest of the population. The Australian government, through health programmes under its direct control, spends 74 cents per capita on indigenous Australians for every $1 spent on the rest of the population (Ring and Brown 2002). In South Africa, the poorer health status of black people compared to white South African partly results from the historic imbalances and inequities in the resource allocation system. McIntyre observed that over 60% of health care spending in South Africa at the end of the 20th century was in the private sector (McIntyre 2000), the main beneficiaries of which were the minority white population and a very small black middle class. In Madagascar, Castro-Leah et al. (2000) found that the poorest 20% of the population consumes 12% of public spending on health compared to 30% share of the total enjoyed by the richest 20%.
In recent years, policymakers have come under pressure to abandon historical funding models, which have been widely perceived as inequitable (Green et al. 2000, Eager et al. 2001) and to develop explicit alternatives that would redress inequities within and between geographic regions. One issue that remains unresolved in the quest for more equitable resource distribution, however, is the appropriate principle or set of principles that should guide resource allocation in order to bridge the existing gaps in equity. One major reason for the lack of consensus among scholars on this issue is the many ways in which the term ‘equity’ is interpreted.
1.4.1 Interpretation of Equity The interpretations of equity include ‘equal expenditure per capita’, ‘equal inputs per capita’, ‘equal access for equal needs’, ‘equal utilisation for equal needs’, and ‘equal health’ (Mooney 1992). Each of these definitions has benefits and difficulties with regards to measurement and operationalisation (see Chapter 4 for discussion of interpretations of equity). Most developed countries have applied the “equal access for equal needs” interpretation in their efforts to design needs-based models for fair distribution of services and resources (Mooney 1991; Donaldson and Gerard 1993; Bradshaw and Bradshaw 1995). However, in reality, as observed by Mooney (2000), most of countries pursue ‘equal expenditure per capita’ or ‘equal expenditure for equal 9 needs’ in their resource allocation initiatives.
The best documented example of a needs-based model for resource allocation based on equality of access is the Resource Allocation Working Party’s (RAWP) model developed in England in 1976 (May and Bevan 1986; Klein 2003). The RAWP model sought to allocate National Health Service (NHS) funds between geographical areas to secure equal opportunity of access for equal needs (Carr-Hill et al. 1994). Countries such as Australia, New Zealand, Canada and South Africa, more recently, have followed the RAWP approach and have developed their own needs-based systems to improve equity.
Progress towards improving the health of disadvantaged groups has been slower than expected, even in countries with needs-based allocation models. The relatively poor health of the indigenous people in Australia where various needs-based models of funding have been widely applied attests to this fact. Consequently, there has been a growing interest in vertical equity principles as the most appropriate for public sector resource allocation in recent years (Mooney 1996; Mooney and Jan 1997; McIntyre and Gilson 2000; Wiseman and Jan 2000; McIntyre et al. 2002). Vertical equity deals with unequal but equitable treatment of unequals. Thus, treating people differently on the basis of their differences in health or whatever indicator under consideration.
Resource allocation to achieve vertical equity puts a premium on differential allocation of resources to benefit the most disadvantaged groups or the most deprived areas with the greatest health needs. Additionally, the allocation must reflect the informed values of the disadvantaged (Mooney 2000; Wiseman and Jan 2000). Proponents of vertical equity have argued that it is unfair to treat people disadvantaged as a result of previous discrimination or flaws in policy, in the same way as the general population are treated. They maintain that the disadvantaged need special attention and rather than simply accounting for differential needs as done in traditional needs-based models, there is the need to assign additional weightings to the most disadvantaged groups in society (Mooney 1996). Mooney and Jan (1997) have suggested that in determining allocation patterns that reflect vertical equity goals, there is the need for wide consultation with society to identify which groups should be prioritised in policy action and how much
10 additional weight they should receive compared to other groups.
1.5 Equity in Resource Allocation under Decentralisation in Ghana Allocation of resources within the Ghanaian health system is, in principle, designed to reduce inequalities in health between and within geographic regions. One strategy identified as crucial to achieving this equity goal is to redistribute resources in favour of the most deprived regions and districts in the country and remove financial barriers to accessing services for the most vulnerable segments of the population (MOH 2002).
The resource allocation decision-making process is decentralised. Thus, at the beginning of the planning season, the MOH assigns budgetary ceilings to the ten Regional Health Administrations (RHAs) in the country. This amount is allocated in block for all districts in the region. Each RHA uses its own region-specific resource allocation formula to re-distribute this lump sum among the districts under its jurisdiction. It is on the basis of these allocations, as determined by the RHAs, that districts plan and budget their activities. Completed activity plans and budgetary estimates of districts are collated by the RHA and returned to the MOH for approval. Once approved, the MOH disburses the funds directly to the various districts through the RHA. The RHA at this stage cannot alter what has been disbursed to individual districts.
The use of different resource allocation criteria by the ten regions means progress towards equity may vary from one region to another, depending on the allocation mechanism and the commitment to promote equity. To date, there has not been any systematic investigation assessing the allocation of resources within regions and the extent to which equity objectives are advanced (or not) through this process. This lack of empirical evidence on the extent of equity in resource allocation within regions in decentralised health systems, such as that of Ghana, motivated this study.
1.5.1 Purpose and Main Research Questions
The purpose of this study is to provide empirical data to inform and support the process 11 of health policy reform seeking to promote equity by examining the mechanisms and actual allocation of resources within regions in Ghana and investigating the factors that influence the more equitable distribution of resources at regional and other levels of the health system. It is designed to address two main questions: To what extent has decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? What factors influenced the equitable allocation of resources for health care in Ghana?
Specific objectives and hypotheses are presented in Chapter 6.
1.5.2 Scope and Outline of Study This study covers the allocation of health resources within regions and its impact on vertical equity. Specifically, the study focuses on the allocation of funds from regional to district levels. Resource allocation from national to regional levels (inter-regional funding) does not form part of the main study objective. The reason is that this study is more interested in intra-regional resource allocation, which is crucial for the promotion of equity, but largely ignored in the equity and funding debate. It is here (regional level) that greater autonomy for resource allocation decision-making has been transferred under the decentralisation policy, and therefore where attention is focused in order to examine whether and how funds are allocated to promote equity. To provide adequate background to the intra-regional analysis, however, allocation of funds from national to regional levels, and from district to sub-district levels are analysed. These assist in contextualising the main focus of the study: the allocation of funds from regions to districts and the determinants thereof. Vertical equity is defined in this study as differential allocation of resources in favour of the most deprived jurisdictions (see Chapter 5).
The analysis is largely limited to the allocation of financial resources. Allocation of other forms of health resources, such as human resources, is deliberately excluded to
12 give the study a concise focus. Allocation of salaries is also excluded from this analysis. This is because salaries are rigidly controlled by the central MOH and the Ministry of Finance. Regional and sub-regional health authorities have little influence on staff matters, and given that sub-regional analyses are the focus of this work, it is inappropriate to analyse the allocation of the salary budget in this study (this is discussed further in Chapters 10).
The study is divided into 11 chapters. Chapter 1 has introduced the rationale for the study and provided the background to resource allocation and equity in the context of global health reform including decentralisation. It specified the research topic, the purpose and the relevance of the main issues addressed by the study. Chapter 2 provides the relevant country context data, focusing largely on the political and economic context within which health policy makers in Ghana make resource allocation and other decisions. It also presents information on relevant aspects of health sector decentralisation policy and implementation in Ghana. This study builds on the existing literature, which is reviewed in Chapters 3 (health and social justice), 4 (equity and equitable distribution of health care resources) and 5 (decentralisation and equity in resource allocation). Chapter 6 describes how this study was designed and carried out, and how the data were analysed. Results of the study are presented in three chapters; Chapter 7 presents the results from textual materials, Chapter 8 the quantitative results and Chapter 9 the qualitative results. Discussion of findings and limitations of the study are presented in Chapter 10 while Chapter 11 presents the conclusions and offers suggestions for further research.
13 Chapter 1. Summary of key points
This chapter has provided the necessary background to the study and presented the research issue and the main questions addressed by the study.
The study is placed in the context of global health sector reform.
Health reform is defined here as a dynamic process involving changes in health policies and institutions by government to allow the health system to function at its best and ultimately improve people’s health.
Factors that drive health reform and the institutional context within which reform occurs, differ between developed and developing countries.
In developed countries, health sector reform is largely designed to achieve greater efficiency in the face of ageing populations, increasing popular expectations and rising costs associated with medical technology.
In developing countries, reform usually addresses issues of extending access to underserved areas, improving efficiency, service quality and addressing geographical inequities in resource allocation.
Decentralisation has been widely promoted as an effective means of promoting a variety of policy objectives including greater efficiency, equity and enhanced accountability. International institutions such as the World Bank and the IMF have been instrumental in extending health sector decentralisation in developing countries.
Inequality in access to health services is one of the main challenges facing the health system in Ghana. The country’s health policy aims to address inequality by allocating resources to benefit the most deprived regions where health needs are greatest.
The health resource allocation system in Ghana has been decentralised. The ten regions in the country receive block grants and use their own independent regional formulae to redistribute these block grants among the districts under their jurisdictions.
This study examines the mechanisms and actual allocation of resources within regions in Ghana and explores the factors that influence the equitable distribution of resources within regions.
The main purpose of the study is to provide empirical data to inform and support the process of health policy reform that seeks to achieve equity.
The two main questions addressed by the study are:
• To what extent has decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds?
• What factors influenced the equitable allocation of resources for health care in Ghana?
14 CHAPTER 2
GHANA AND ITS HEALTH CARE SYSTEM
Overview Contextual factors impact on health policy and resource allocation processes. These typically include the political system, the macro-economic context and the structure of the health care system. This chapter presents contextual information relevant to understanding equity and resource allocation issues under decentralisation: the focus of this study. The data are organised under the following headings:
Geography, population and administrative divisions Political context Economy context Structure of the health care system Health status Inequalities in health Health policy objectives Health sector reforms
2.1 Geography, Population and Administrative Divisions Ghana, a former colony of Britain, is located on the west coast of Africa, bordering Togo to the east, Cote d’Ivoire to the west, Burkina Faso to the north and to the south by the Atlantic Ocean, which washes about 560-kilometer stretch of coastline (Figure 2.1). It covers an area of about 239,000 km2, about 4,000 km2 less than the surface area of United Kingdom (243,000 km2). In 2002 the population was estimated at around 20 million with a density of 88 people per square kilometre (World Bank 2004). In common with many developing countries, the population structure is youthful, with about 40% under the age of 15. The proportion of the population aged 65 years and above is about 7% and among the highest in Africa. This is expected to rise to about 15 15% by 2050. The sex distribution is slightly in favour of females (51%). About 55% of the population resides in rural areas (United Nations 2001b; Ghana Statistical Service 2002).
Administratively, Ghana is divided into ten regions, each of which is sub-divided into a number of districts. There are a total of 110 districts in the country5. The Ashanti Region is the most populous with over 3.6 million inhabitants living in 18 districts. Northern Region is the biggest in terms of landmass covering about 70,384 km2. It has a population of about 1.8 million and 13 districts (see Figure 2.1).
Figure 2. 1 A Map of Ghana Showing the 10 Administrative Regions, Total Population per Region and Number of Districts Northern Sector
Southern Sector
Source: www.ghanaweb.com (modified)
5 New districts have recently been created, taking the number of districts to 138.
16 For analytical purposes, Ghana is usually divided into two sectors along climatic lines; the northern dry savannah grassland and southern coastal plain and rainforest. The northern belt comprises the Upper East, Upper West, Northern and parts of the Brong- Ahafo and Volta Region. This sector accounts for only about 23% of the total population of Ghana. With the exception of a few big towns, the population density is very low, with people scattered in small rural settlements averaging about 500 inhabitants. By contrast, the southern sector, which comprises the remaining regions, including the Greater Accra and Ashanti, is home to over 70% of the total population. The population density is high in key cities, particularly in Kumasi and Accra, which are the main destinations for internal migration. The next section briefly reviews the political context within which health policy makers in Ghana operate.
2.2 Political Context Ghana has a political history dominated by dictatorships and military coup d'états. The country gained independence from the British in 19576; the first British territory in Africa to become a sovereign state in its own right within the British Commonwealth (Gyimah-Boadi 2001). Three years after independence in July 1960, Ghana became a republic. Free from colonial domination, the first president - Kwame Nkrumah and his Convention People’s Party (CPP) put Ghana on the path of socialism. One of the key policy aspirations of Nkrumah’s government was to promote industrialisation in Ghana7. However, it was his policy on education (the Education Act of 1961) aimed at universal primary education for school-age children for which he is best remembered. Observers have maintained that it improved enrolment opportunities for girls and people from disadvantaged regions (Graham 1985). Nkrumah’s CPP government was removed from office in February 1966 through a military coup.
Three years after the military take-over, Ghana made a second attempt to democratise with the enactment of a new constitution. Elections were held in October 1969, and a
6 The name of the country was changed from Gold Coast to Ghana after independence. Gold Coast was the name given by the early European merchants to portray the region’s abundant gold supply. 7 Nkrumah was convinced that colonialism has subordinated Africa to the position of primary producers and without industrialisation the continent’s development goals would not be realised. 17 new civilian administration led by Prime Minister Kofi Busia was inaugurated. The Busia government pursued policies that sought to increase government investments in development programmes, especially in rural areas (Grindle and Thomas 1991). However, the economy was in serious crisis and the harsh measures adopted by the government to revive it, including devaluation of the cedi and review of the Hospital Fee Act in 1971, did not go down well with many Ghanaians, in particular, the military. Twenty-seven (27) months into office in January 1972, Busia’s government was overthrown in another coup d'état. The military intervention in politics continued with dozens of coups and attempted coups. Table 2.1 shows the different civilian and military regimes of Ghana since independence, highlighting the degree of instability present.
Table 2. 1 Ghanaian Regimes Since Independence Regime Leader Dates Description Convention People’s Nkrumah 1957-1966 Ghana achieved full independence in Party (CPP) 1957 and republican status in 1960. Regime overthrown in coup d'état National Liberation Ankrah 1966-1969 Coup d'état regime. Supervised elections Council Regime in 1969 and handed over to constitutional (NLC) government Progress Party Busia 1969-1972 Second Republic. Constitutionally Administration elected. Overthrown in coup d'état National Redemption Acheampong 1972-75 Military Regime. Composition of Council (NRC) executive and name of ruling council Supreme Military Acheampong 1975-78 changed 2 times after internal power Council (SMC) Akuffo 1978-79 struggles. Overthrown in a military uprising of young officers and other ranks Armed Forces Rawlings June to Supervised elections and handed over to a Revolutionary September constitutional government Council (AFRC) 1979 Peoples National Limann 1979-1981 Third Republic. Constitutionally elected. Party Regime (PNP) Overthrown in a coup d'état Provisional National Rawlings 1982-1992 Coup d'état regime. Called elections and Defence Council succeeded itself (PNDC) National Democratic Rawlings 1993-2000 Fourth Republic. Constitutionally elected. Congress (NDC) New Patriotic Party Kufuor 2001 to date Fourth Republic. Constitutionally elected (NPP) Adapted from Tsikata and Seini 2004
As the country moved from one political order to another, the economy became a victim of corruption and mismanagement. Stability of a sort finally came when Jerry Rawlings (an Air-force Flight-Lieutenant) ceased power for the second time, placed a ban on
18 party politics and ruled Ghana with his Provisional National Defence Council (PNDC) for ten years from 1982 to 1992. During this period, the PNDC regime was forced by a worsening economic crisis to accept an International Monetary Fund (IMF) and World Bank sponsored Structural Adjustment Programme (SAP), which became one of sub- Saharan Africa’s longest economic recovery programme designed by the IMF and World Bank.
Having committed itself to work with the international community, the PNDC regime could not resist the pressure for political reforms from the foreign donors. A new constitution was promulgated in 1992 with presidential and parliamentary elections following soon afterwards. In the end, Rawlings the military dictator became Rawlings the democrat, having formed a political party (National Democratic Congress - NDC), contested and won two elections in 1992 and 1996 (Horton 2001). He stepped down in accordance with the constitution after two terms in office in January 2001. Rawlings’ anointed successor was beaten in a new election in December 2000 by the current president - John Kufuor. It was the first time in Ghana’s political history that power was transferred peacefully from a ruling party to an opposition winner. Multi-party democracy has since been deepened with a help of a free and vibrant media. The next section looks at the Ghanaian economy, which suffered considerably from the political instability described.
2.3 Economic Context One of the major determinants of the amount of real resources available to the health sector is the strength of the underlying economy (Parkin et al. 1987). As indicated earlier, the economy bore the brunt of the political turmoil that engulfed Ghana subsequent to independence. After a period of relative prosperity in the 1960s, the economy experienced significant deterioration, with falling GDP, soaring inflation, worsening balance of payment and devastating poverty (World Bank 1995, Danida 2002). The IMF and World Bank Structural Adjustment Programme implemented by the PNDC government, resulted in greater economic stability, liberalisation of trade, falling inflation (10% in 1992) and annual GDP growth of more than 5% (World Bank 2002). 19 However, despite these achievements, the economy remained fragile. Falling prices of Ghana’s major export commodities, such as cocoa and gold, in the face of increasing prices of oil and other imports, led to persistent trade deficits. The government was compelled to finance this deficit through domestic and foreign borrowings, increasing the debt burden of the country. The total debt stock in 1999, for example, was equal in size to the country’s GDP (99.5%) and was significantly higher than the sub-Saharan African average of 76% of GDP (IMF 2000). Debt servicing in 1999 amounted to 8.5% of GDP for external debts and 5.8% of GDP for domestic debts, representing a total of about 40% of overall government expenditures. In relation to government tax revenue, the Ministry of Finance estimated that the cost of servicing external debts was equivalent to 55% of total tax revenue in 2000 (Budget Statement 2001). The high domestic borrowing by the government led to a significant rise in interest rates (over 40% in 1999/2000), which constrained the ability of the private sector to borrow and invest. This led to a fall in productivity, particularly, in the late 1990s. The local currency (the cedi) considerably depreciated and the overall poverty situation in Ghana worsened (GPRS 2002). Table 2.2 shows the value of the Ghanaian cedi to the United State dollar over a ten-year period from 1994 to 2003.
Table 2. 2 Exchange Rate of the Ghanaian Cedi to the US Dollar
Year US$ Ghanaian Cedi (⊄) 1994 1 1052.63 1995 1 1449.28 1996 1 1754.39 1997 1 2272.73 1998 1 2325.58 1999 1 3535.14 2000 1 7047.65 2001 1 7321.94 2002 1 8438.82 2003 1 8800.50 Source: IMF International Financial Statistics, 1994 to 2003
In the light of excruciating poverty and heavy debt burden, the government of the New Patriotic Party (NPP), which was voted into office in December 2000, decided to join the Heavily Indebted Poor Country (HIPC) initiative. The HIPC initiative was instituted by the World Bank and IMF in 1996 to provide more effective debt relief to a limited
20 number of countries (Labonte et al. 2004). It was anticipated that the debt-servicing payments of Ghana, under the HIPC initiative, would fall, thereby freeing up resources for spending on targeted social services in order to reduce poverty. By joining the HIPC initiative, Ghana had to develop a Poverty Reduction Strategy Paper (PRSP), which clearly sets out poverty reduction targets and measures to achieve them.
Ghana reached the HIPC decision point8 in March 2002 and has since made some savings from reduction and re-scheduling of debt-servicing payment. Total debt relief savings in 2002, for example, amounted to approximately US$275 million; $92.5 million was from the HIPC initiative (Budget Statement 2003). In the last few years, the Ghanaian economy has shown signs of improvement: inflation has dropped from the 40% levels in 2000 to about 11% in March 2005, interest rates have declined, and revenue from taxes has more than doubled. The economy, however, remains vulnerable to external shocks, and with the continued rise in global oil prices, its revival could significantly be affected.
2.4 Structure of the Health Care System The Ghanaian health care system consists of four main sectors: the public, private, traditional and other sectors. All these sectors come under the umbrella of the Ministry of Health (MOH), which currently provides stewardship to the entire health sector (MOH 2005). The public sector is made up of the Ghana Health Service (GHS), the Teaching Hospitals and quasi government institutions and statutory bodies. The private sector is dominated by private-for-profit and mission-based providers. The traditional sector comprises traditional medical practitioners, faith healers and alternative medicine providers. The other non-health sectors that are involved in health delivery include the education sector, environment science and technology, works and housing and local government and rural development (MOH 2003). Figure 2.2 shows the four sectors and the key players in each sector.
8 A decision point is reached when debt relief under enhanced HIPC is approved by the Executive Boards of the IMF and World Bank. 21 Figure 2. 2 Organisational Structure of the Ghanaian Health Care System
MINISTRY OF HEALTH
Public Sector Private Sector Traditional Other Sector
Ghana Health Private-for-profit Traditional Medicine Education Service (GHS) Providers Mission-Based Environment, Sci. Teaching Hospitals Providers Alternative Medicine and Technology Providers
Quasi Government Non-Governmental Works & Housing Institutions Organisations Faith Healers Local Government Statutory Bodies Civil Society & Rural Organisations Development
Households Source: MOH 2003
The GHS is by far the largest provider, running a total of 9 regional hospitals, some 106 district and other hospitals, and about 558 health centres and clinics based in the sub-districts (MOH 2001). The GHS is an autonomous executive agency responsible for implementation of national health policies under the control of the Ministry of Health. Before the establishment of the GHS, the MOH was both the policy maker and implementer, in charge of service delivery throughout Ghana (this issue is discussed further under decentralisation - Section 2.8.1).
Private provision of health care in Ghana is not only substantial but also recognised as an important and growing source of health care (Danida 2002). Over the past few years, the number of private facilities has expanded considerably, especially in the urban areas. The private-for-profit sector has about 71 hospitals, 910 registered clinics and 487 registered pharmacists (Obuobi et al. 1999; MOH 2001). In 1997, 63% of all private-for-profit clinics were located in the main cities in Ghana, 36% in towns and only 1% in villages. Available evidence suggests that most patients in Ghana prefer to use private-for-profit services for minor ailments because of the slowness, unavailability of drugs and poor staff attitude in the public facilities (MOH 1999).
22 However, for more serious cases, government or mission facilities are perceived to be better. While the private sector plays a significant role in health care provision in Ghana, the growth of this sector is likely to limit the scope of achieving equity through a funding formula since private sector resource allocation may not be driven directly by a funding formula.
The private not-for-profit providers are dominated by mission-based providers estimated to cover about 40% of the service delivery. This group supplies an estimated 30% of beds and 35% of outpatient care in Ghana. The Christian Health Association of Ghana (CHAG), an association of Christian health institutions, provides health care to about 30-35% of the population, mainly in the rural areas. The CHAG is the second largest provider of health care in Ghana behind the government. In 2003, it had 140 health institutions made up of 52 hospitals, 80 primary health care clinics and 4 health manpower-training centres. The hospitals and clinics provide a total of about 5600 beds (CHAG Annual Report 2004).
Traditional medicine is a major source of health care for many Ghanaians. It is currently estimated that about 70-80% of Ghanaians use traditional medicine as their front line health service (MOH 2005). Practitioners in this sector consist mainly of herbalists; Traditional Birth Attendants (TBAs) trained by the MOH, an unknown number of untrained TBAs, Spiritualists and Faith Healers. As in many developing countries, the quality of traditional medicine in Ghana has been difficult to assure. Despite some efforts to integrate traditional medicine into the formal health sector, little has been achieved in terms of bringing the two sectors together (Sena 1997). A Traditional Medicine Practice Law was passed in 2000, with the stated intention of promoting traditional medicine, making it safe and integrating the service with the allopathic system (Danida 2002). The Ghanaian health system is funded through four major funding sources: government, donors, commercial credits and internally generated funds (IGF). Data on funding sources, derived from documentary analysis, are presented in Chapter 7.
23 2.5 Health Status Ghana has made significant progress in health delivery since independence from British colonial rule in 1957. Infant mortality, for example, dropped from 84 per 1000 live births in 1988 to 57 per 1000 in 1998; a decline of about 30% in only a decade (World Bank 2001). Under-5 mortality also dropped from 154 per 1000 live births in 1988 to 108 per 1000 in 1998. Health interventions and access to key essential public health services such as immunisation has also improved in recent years (MOH 2001).
In spite of these improvements Ghana still faces serious health challenges. The health status of much of the population remains comparatively poor even among developing countries. Nutritional status has not improved, the epidemiological situation is still characterised by the predominance of poverty-related infectious diseases. Malaria and anaemia alone account for about 40% of reported deaths among children up to 15 years while pneumonia and diarrhoea remain significant causes of death in all age groups (MOH 2001). In addition to the “traditional” diseases, rising incidence of HIV/AIDS and road traffic accidents have become an important cause of adult mortality and morbidity in the country (World Bank, 2001) even as non-communicable diseases such as diabetes and strokes are on the rise (Amoah et al. 2002).
Recent evidence from the 2003 Ghana Demographic and Health Survey (GDHS) indicates a lack of improvement in child survival in Ghana. The infant mortality rate, for example, increased more recently from 57 per 1000 in 1998 to 64 per 1000 in 2003. Similarly, under-5 mortality also increased from 108 per 1000 to 111 per 1000 from 1998 to 2003 (GDHS 2003). If this trend continues, the Millennium Development Goal (MDG) of reducing childhood mortality by two-thirds of the 1990 levels by 2015 (UN 2002) will not be achieved. Table 2.3 presents selected health-related indicators for Ghana, Zambia, South Africa, Malaysia and the United Kingdom9.
As evident from the table, Ghana compares poorly with Zambia and South Africa in terms of GNI per capita. However, it has relatively better life expectancy (55 years on average) than the two countries (37 years on average for Zambia and 46 years in South
9 Malaysia and Ghana became independent from the United Kingdom in the same year, 1957. Zambia’s health system has many features in common with the health system of Ghana. South Africa was chosen for its relatively high economic status in Africa. 24 Africa). The low life expectancy of Zambia and South Africa can be partly explained by the high prevalence of HIV/AIDS in the two countries (see HIV prevalence rate among the countries).
Table 2. 3 Selected Health-related Indicators South Indicators Ghana Zambia Africa Malaysia UK Population (millions 2003) 20.4 10.4 45.3 24.8 59.3 GNI per capita (US$ 2003) 320 380 2,780 3,780 28,350 Life expectancy (2002) 55 37 46 73 77 Under-5 mortality/1000 (2002) 97 182 65 8 7 Adult literacy (% 15 and above- 2002) 74 80 86 89 - Malaria cases/100000b 15,344 34,204 143 57 Nil TB cases/100 000 (2002)b 371 588 366 120 12 HIV prevalence rate (% 2003)b 3.1 16.5 17.8-24.3 0.4 0.1 Government Health spending % GDP (2001)b 2.8 3.0 3.6 2.1 6.2 Health spending per capita (US$ 2001)b 60 52 652 345 1,989 Physicians/100,000 people (1990-2003)b 9 7 25 68 164
Source: World Development Report 2005. b Human Development Report 2004.
2.6 Inequalities in Health Inequality in health is another major challenge facing the Ghanaian health care system. This has historical antecedents. The socio-economic development agenda pursued by the British colonial administration was characterised by great geographical polarisation. Infrastructural and socio-economic development activities were concentrated in few coastal cities and other towns in the southern sector, ignoring the northern half of the country (Konadu-Agyemang 2000; Tsikata and Seini 2004). This set the course for unequal resource distribution and development.
Although the post-independence economic policies of successive governments have acknowledged the need to bridge these systemic inequalities through redistribution of resources, implementation of such policies has been fraught with difficulties, leaving large sections of the country still deprived. Figure 2.3 shows the distribution of poverty
25 among region in Ghana for 1992 and 1999. As clearly depicted, the Northern, Upper East, and Upper West regions have the highest level of poverty, with the trend worsening in Northern and Upper East Regions. Central Region is the only region in the southern sector where poverty worsened.
Figure 2. 3 Regions by Percentage of Poverty
Northern
Upper East
Upper West
Brong-Ahafo
Volta 1992 Ashanti 1999 Region
Eastern
G-Accra
Central
Western
0 102030405060708090100
% of Poverty
Source: Ghana Statistical Service, 2000. Note: Poverty is defined here as an unacceptable physiological and social exclusion (see GPRS 2002 p.6).
2.6.1 Inequalities in Access to Facilities Access to health care resources varied significantly by region (access to human resources has been covered in Chapter 7). A study by Canagarajah and Ye (2001) revealed that, while in most regions, more than 30 percent of the rural population is close to health facilities (within 30 minutes travel distance), in the three main northern regions (Upper East, Upper West and Northern), only about 15 percent of the rural population lives close to health facilities. According to the authors, investments in both public and private health facilities have tended to favour regions in the south, particularly, the Greater Accra and Ashanti regions. In 1999, for example, there were 35 privately run hospitals and 87 health centres/clinics in Ashanti Region compared to only one private hospital and 11 health centre/clinics in the Northern Region (MOH 1999).
26 Table 2.4 provides the distribution of health facilities (both public and private) among regions in Ghana. It shows that Ashanti had twice the number of health centres and four times the number of hospitals compared to the Northern region. Volta region, although relatively small in terms of population size, has many facilities than some of big regions. This may be due to the high concentration of donor/NGO activities in the region in the 1980s. Most of these NGO invested in health facilities, as the regional health infrastructure at that time was relatively under-developed (MOH 1996).
Table 2. 4 Regional Distribution of Health Facilities Sub-district Total Number Region Population District & Other Health Centres/ of Health Hospitals Clinics Facilities Northern 1,820,806 13 116 129 Upper East 576,583 5 75 80 Upper West 920,089 4 51 55 Brong Ahafo 1,815,408 23 179 202 Volta 1,635,421 26 450 476 Ashanti 3,612,950 64 226 290 Eastern 2,106,696 25 128 153 Greater Accra 2,905,726 22 249 271 Central 1,593,823 14 104 118 Western 1,924,577 19 180 199 Ghana 18,912,079 215 1758 1973 Source: Ghana Living Standard Survey 4; GDHS 1998/99
2.6.2 Inequalities in Child Mortality Mortality rates also varied across regions. Although data for 1993 and 1998 suggest an improvement in under-5 mortality in all regions in Ghana, the rates remained higher in the northern sector. Figure 2.4 shows regional under-5 mortality profile for 1993 and 1998. The Northern Region witnessed a drop in under-5 mortality from 237 per 1000 in 1993 to 171 per 1000 in 1998. This was still significantly higher compared to other regions and the national average of 109 per 1000.
27 Figure 2. 4 Regional Under- 5 Mortality Profile: 1993 and 1998
No rthern
Uppe r Ea s t
Uppe r We s t
Brong-Ahafo
Vo lt a 1993 Ashanti 1998 Region Eastern
G-Accra
Central
We s t e rn
0 50 100 150 200 250 Rate per 1000 Live Births
Source: GDHS 1993 and 1998
To improve the overall health status of Ghanaians and address the inequality and other systemic problems affecting the health system, the MOH has embarked on a number of health reforms in the last decade. The rest of this chapter provides information on some of the key elements of these reforms, starting with the national health policy goals that drove the reforms
2.7 Health Policy Objectives The overall objective of national health policy in Ghana is to improve the health status of all Ghanaians. The vision of the Ministry of Health as stated in the country’s health policy document is to:
“Improve overall health status and reduce inequalities in health outcomes of people living in Ghana”(MOH 2003 p.8).
Since 1995, Ghana’s health policy has been based on a medium-term health strategy (MTHS)10, which seeks to promote greater equity in access to health and outcomes
10 The MTHS is perceived in Ghana as a local initiative to tackle the systemic health problems confronting the country. However, there was significant donor involvement in translating the ideas into the five-year programme of work (MOH 1997). 28 (MOH 2000). The main policy goal with regards to equity is to achieve equality of access to high quality care. The MOH has stated this as its mission statement:
“The Ministry of Health will work in collaboration with all partners in the health sector to ensure that every individual, household and community is adequately informed about health; and has equitable access to high quality health and related interventions” (MOH 2003 p.8).
To achieve this goal, the MTHS aims to:
strengthen district health services promote community involvement in health delivery and re-direct resources to the needy or deprived areas (MOH 2000).
The MOH has developed a conceptual framework emphasising five strategic objectives through which a package of interventions identified as crucial for improving health status to be delivered in the medium-term. The five objectives are: improving the quality of health delivery, increasing access to health services, improving the efficiency of health service delivery, fostering collaboration with other sectors, and improving financing of the health sector. Figure 2.5 shows the conceptual framework underpinning the strategic objectives.
Figure 2. 5 The Five Strategic Health Policy Objectives in Ghana
Health Status
Availability and utilisation Availability and utilisation of health interventions of health interventions
Package of Health Interventions
Quality Efficiency Access Resources Collaboration
Source: Health Sector 5-Year Programme of Work (2002-2006) 29 While equity has not been categorically stated among the five strategic objectives, the emphasis on access identifies financial barriers to accessing health services as a crucial factor contributing to the inequalities in access to services and health outcomes. Consequently, the MOH has emphasised the removal of financial barriers to accessing health care by strengthening the exemption policy to better cater for the poor and other vulnerable groups (MOH 2002). Much recently the government has introduced an ambitious National Health Insurance Scheme, which aims further to remove financial barriers to accessing health care. In the area of resources, the emphasis has been on improving the overall funding of the health system and redistributing resources in favour of the four most deprived regions.
2.8 Health Sector Reforms Reform in the Ghanaian health sector has a 20-year cycle (Akosa et al. 2003). The first cycle, which started in 1957, when the country became independent from British colonial rule, emphasises delivery of basic health services and development of basic health infrastructure and human resource. The second cycle of reform started in 1977 and had expansion of services and the implementation of the primary health care (PHC) concept as its main emphasis. The decentralisation of the health system can also be traced to the second. The third cycle began in 1997 and was based on the sector-wide approach to health delivery (SWAp). This cycle emphasised strengthening of the health system11. The concern of this study is predominantly with the decentralisation reform, which provides the basis for the current resource allocation system and the SWAp initiative.
2.8.1 Decentralisation and Re-organisation of the Health System The process of decentralising the health sector was initiated as part of a broader process to restructure public administration systems in order to increase efficiency and output. It
11 Note that the cycles are only for analytical purposes. They may overlap and are presented here for ease of description and analysis.
30 began with the enactment of the local government law in 1988 (PNDC Law 207) by the then PNDC government. The intention of the law was to establish District Assemblies (DAs) to become the highest legislative and executive body in the district (a devolution type of decentralisation). It was planned that the DAs would be the main planning authority in the district for all sectors including the health sector and that District Health Management Teams (DHMTs) would no longer receive separate allocations determined by the MOH, but would prepare and submit their own annual plans to the District Executive Committee (DEC) for approval. In other words, there would be a block allocation of funds for all sectors to the DA, which the DEC would re-allocate to various sectors based on district priorities.
In effect, none of these earlier intentions was realised until a new Local Government Act (Act 625) was passed in 1993, one year after the country had adopted a new constitution and reverted to democratic rule. Following the new Act of 1993, 110 districts were created out of the existing 65 and District Assemblies were elected. However, the transfer of district health services to the DAs never materialised. Ghana’s new Constitution (Government of Ghana 1992) mandated the establishment of an autonomous agency to take responsibility of health delivery from the central MOH. Consequently, in 1996, the Ghana Health Service and Teaching Hospitals Act (Republic of Ghana Act 525) was passed by Parliament to establish the Ghana Health Services (GHS). The GHS was given the task of running a decentralised health system with the aim of providing access to basic health services to all Ghanaians as close as possible to where they live and work (MOH, 1996; 1997).
Specific objectives of the GHS as spelt out in the legislative instrument (Act 525) include implementing approved national policies for health delivery in the country, increasing access to quality services, and managing available resources prudently to provide health services. The GHS and the Teaching Hospitals are governed by separate boards- the Ghana Health Service Council and the Teaching Hospitals Board, both of which report to the MOH. The passing of the Act 525 by Parliament in effect ushered in the process of major re-organisation of the national health system. The role of the central MOH in health care provision was reduced to one of policy oversight, overall resource allocation and financial regulation. The organisational restructuring and
31 increased decentralisation of the health system required that resource allocation be reformed. The next section describes in greater detail how this was undertaken.
2.8.2 Decentralisation of Resource Allocation Decision-Making Addressing geographical inequities was one of the main motivations behind the decentralisation reform. Under this reform, recurrent budgetary resources are allocated in block by authorities at the national level to the ten regional health administrations in the country. The RHAs in turn, re-allocate the block grant to the districts under their jurisdictions (Bossert et al. 2000). The RHAs also hold and manage funds on behalf of the districts (see Table 2.5).
The health budget is divided into four categories or line items:
Item 1- salaries – controlled by Ministry of Finance Item 2- administrative cost – under decentralised control Item 3 - service cost – under decentralised control Item 4- investment costs - controlled by Ministries of Finance and Health
The central MOH developed a population-based formula12 as a substitute for the existing historical funding method, which was used to distribute line items 2 and 3. The formula is regularly modified in consultation with regional authorities and facilities and entails a range of factors including:
Fixed cost of administration Distance from the national capital Population size and density Size of region Facility size (in beds) Infant mortality rate
12 It must be noted that the population-based allocation procedure existed more in principle than in practice. As discussed later in this study, the inter-regional resource allocation in Ghana has no well- designed and transparent formula.
32 Once the budget is allocated to regions, it is up to the RHAs to decide how to divide their share of the budget among the districts under their jurisdiction. In general, factors such as district population size, distance from regional capital (remoteness), and number of facilities are common to all regional resource allocation criteria. However, other region-specific factors are also taken into account in designing the allocation formula (Ensor et al. 2001).
2.8.3 Implementation of Budget Management Centre (BMC) Concept As mentioned earlier, this reform was implemented mainly to transfer control over, and accountability for, resource allocation to district levels, while at the same time ensuring efficiency in resource use. The budget and management centres (BMCs) are decentralised functional units or cost centres with local responsibility for planning, fund management and implementing an agreed programme of work within a given budget (World Bank 2003).
The entire MOH/GHS is divided into some 350 BMC units. These units are found at all levels of the health system and are hierarchically ordered such that national level BMCs supervise regional BMCs, which in turn supervise district level BMCs, and so forth down to the sub-district level (Bossert et al. 2000; MOH 2001). All BMCs were given training in financial management and accounting to prepare them for the task of efficient fund management. Those initially certified as having adequate capacity to manage funds were labelled Managing BMC; they receive, manage and account for funds. Those assessed as having insufficient capacity and needed additional training were classified as BMC of Record; their funds are managed for them by the regional health administration, they only keep records of the transactions.
Until 1999, all funds allocated to districts were kept and managed by the RHAs on behalf of the districts. However, following the implementation of the BMC concept, the MOH has transferred autonomy over recurrent budget from RHAs to districts. Thus, since 1999 the MOH no longer allows recurrent budget meant for district health delivery to be kept at the regional level and managed by RHAs on behalf of districts. Except for those which have not been certified as having the ability to manage funds 33 (BMCs of records), all district level BMCs receive and manage their own funds. However, the money is still channelled to the districts through the region health administration (MOH, 1996), which also determines how much funds a particular district should be allocated. Table 2.5 (next page) summarises the changes in resource allocation policy in Ghana.
Table 2. 5 Changes in Health Resource Allocation Policy in Ghana Phase 1 Phase 2 Phase 3 Initial Decentralisation with Phase Centralised Period Decentralisation Budget and Management Centre Time Period 1957 1995 1999 Role of the Spending at every level MOH issues block MOH issues block ceilings Ministry of rigidly controlled by ceilings to regions to regions Health (MOH) the MOH
MOH issues budget ceilings by line items to regions and districts
Role of Regional Budget ceilings by line RHAs design their RHAs allocate ceilings to Health items cannot be own resource districts based on region- Administrations changed by RHAs or allocation formula in specific formulae (RHAs) DHAs without consultation with approval by the MOH DHAs Funds for district health Regions allocate funds delivery are disbursed to districts based on directly to district BMCs region-specific via the region formulae
RHAs hold and manage funds meant for district health delivery on the behalf of DHAs
Role of District DHAs take direct DHAs prepare their DHAs or District BMCs Health instructions from the activity plans and manage their own funds and Administrations MOH budgets and submit to account for it (DHAs) RHAs for approval DHAs/District BMCs DHAs depend on the allocate funds to sub- RHAs to manage funds districts based on district- on their behalf, they specific criteria make no spending decisions DHAs/BMCs hold and manage funds on behalf of sub-districts. Source: Based on information from key documentary sources
34 The three phases in the table are not entirely discrete; there may be some overlaps between the phases, given the incremental nature of the reforms. In phase 3, for example, all districts BMCs did not become managing BMCs in 1999. The process started with a gradual upgrading of planning and management capacity of DHAs. Only DHAs cleared as having adequate capacity to manage funds were given the opportunity to manage their own funds. The RHA continued to manage funds on behalf of DHAs, which were not certified as having the capacity to hold and manage funds. Nonetheless, the implementation of the BMC concept, which is just an extension of the decentralisation policy, commenced in 1999.
35 Chapter 2. Summary of key points
Chapter 2 has presented some of the contextual information relevant for better understanding of this study.
The political environment within which health policy makers operate in Ghana after independence was turbulent with one military dictator after another. The country emerged from the political instability in 1992 under Jerry Rawlings. The political atmosphere has been stable since 1992. Multi-party democracy now characterises the political context.
The Ghanaian economy bore the brunt of political instability. It experienced significant deterioration with falling GDP, soaring inflation, worsening balance of payment and devastating poverty. A decade long IMF/World Bank Structural Adjustment Programme (SAP) was implemented in the 1980s with the aim of improving the economy.
The SAP resulted in greater economic stability but poverty deepened, leading to yet another IMF/World Bank rescue initiative known as the Heavily Indebted Poor Country (HIPC) initiative. Joining the HIPC scheme, Ghana had to develop a Poverty Reduction Strategy Paper (PRSP), which sets out poverty reduction targets and measures to achieve them.
Current health policy in Ghana is linked to the country’s poverty reduction strategy (GPRS), which recognises that improving the health of the poor is crucial for reducing poverty.
The health status of many Ghanaians remains poor, with significant inequalities among and within regions.
The Ghanaian health care system consists of four main sectors: public, private, traditional and other sectors. The public sector is dominated by the Ghana Health Service (GHS) and the Teaching Hospitals.
Private provision of health care in Ghana is substantial and is recognised as an important and growing source of health care. The mission-based providers constitute the second largest providers of health care after the government.
Decentralisation has been one of the key health sector reform policies implemented to improve the performance of the Ghanaian health sector, in particular, to improve equity and management of resources.
36 CHAPTER 3
HEALTH AND SOCIAL JUSTICE: A SEARCH FOR A THEORETICAL FRAMEWORK
What is particularly serious as an injustice is the lack of opportunity that some may have to achieve good health because of inadequate social arrangements.
Amartya Sen, 2002
Overview The nature of fairness in the distribution of health and health care resources continues to generate debate among health policy analysts. In recent years, this has become more passionate, as awareness of inequalities in health become pervasive and resources decline. Different theories of social justice inform approaches to distributing society’s benefits and burdens, including those associated with health. This chapter reviews the theories that shed light on the equitable distribution of health resources. The review seeks to address the following question: what principle of social justice should guide the distribution of health care resources in countries where there are marked inequalities in health? The theories reviewed include utilitarianism, entitlement theory, and Rawls’ theory of justice as fairness. The three theories are reviewed separately, and then, discussed in the context of this study. The chapter begins with a general introduction to the major theories of social justice before a detailed discussion of specific theories.
3.1 Introduction Decisions about how scarce health resources should be allocated between competing claims are exceedingly broad and disparate. However, “all are based on some moral assessment of how competing claims can be fairly adjudicated” (Gillon 1986 p. 94). Although there is no consensus on which particular moral principle should take
37 precedence in resource allocation decision-making, such decisions are explicitly or implicitly based on some theory of social justice. According to Whitehead (1994 p.1285), “social justice is said to exist when conditions allow associations and individuals to obtain what they are entitled to according to their nature and vocation”.
Theories of social justice contain “principles for evaluating laws and institutions from a moral standpoint that are independent from those laws and institutions” (Phillips 1979 p.314). Thus, they provide a standard against which existing institutions and social arrangements can be judged. Social justice must be differentiated from distributive justice. Social justice is concerned with just social arrangements that allow people to receive a fair share of societal resources and burden (Sen 2002). It is not simply about relative sharing of scarce resources. Distributive justice is mainly concerned with the relative share of various goods and services that people receive. In other words, distributive justice, as noted by Roemer, deals mainly with “how a society or group should allocate its scarce resources among individuals with competing needs or claims” (Roemer 1996 p. 1).
There is no consensus on a single theory of social justice when it comes to equitable distribution of health care resources. Culyer (2001 p.275) observed that the absence of an agreed theory of social justice is explained by the lack of a “monist theory of morality”. A monist theory of morality is the view that all ethical questions have a single correct answer and that all these answers fit together in a single coherent moral system (Crowder 2003 p.3).
Moral theories that feature prominently in the debate over health and health resource allocation can be broadly categorised into consequentialist and deontological theories. Consequentialist moral theories hold the view that an act is right or wrong according to its consequences rather than any intrinsic value it may have (Veatch 1981; Beauchamp and Childress 1989; Shaw 1999). The most well known consequentialist theory is utilitarianism, which considers a distribution as just if it maximises the greatest good for the greatest number of people (Gillon 1986; Le Grand 2001). Figure 3.1 shows the branches of moral theories that are most drawn upon in the health equity debate.
38 Figure 3. 1 Type of Moral Theories
Moral Theories
Consequentialist Deontological
Utilitarianism Libertarian Theories Social Contract Theories
(Act & Rule utilitarians) (Nozick’s Entitlement Theory) (Rawls’ Theory of Justice)
Note: This diagram is based on the researcher’s interpretation of Beauchamp and Childress’s classification (see Beauchamp and Childress 1989).
Deontological theories generally maintain that some “features of acts other than, or in addition to, their consequences make them right or wrong and that the grounds for right or obligation are not wholly independent on production of good consequences” (Beauchamp and Childress 1989 p. 36). Deontologists hold that the moral worth of an action should not be judged only in terms of its consequences. Deontological theories consist of several forms. The two prominent ones, which are examined in this study, are the social contract theory of John Rawls and the libertarian (entitlement) theory of Robert Nozick. These theories are sometimes referred to as rights-based theories because of their emphasis on basic liberties and natural rights of individuals (Gillon 1994; Gericke et al. 2005).
Egalitarian theories13 focus on equal access to the goods in life that every rational person desires (Beauchamp and Childress 1989). Generally, they disagree on the exact dimension of equality that is of fundamental importance - equality of wealth, equality of resources or equality of opportunity, but they share the conviction that equality has an intrinsic moral importance that utilitarianism neglects (Shaw 1999). While egalitarian theories propose that people should have access to an equal distribution of certain goods
13 Egalitarian theories belong to the deontological tradition. They are not shown in Figure 3.1 because this study did not examine them in similar detail as the three theories presented.
39 such as health care, “all prominent egalitarian theories of justice are cautiously formulated to avoid making equal sharing of all possible social benefits a requirement of justice” (Beauchamp and Childress 1994; Ruger 2004).
Egalitarians view access to health care as “every citizen’s right” which ought not to be influenced by income and wealth (Williams 1993). They recognise that equality of health outcomes is hardly achievable, but postulate it as a goal that health systems should strive to achieve. Egalitarians, in general, also support greater state involvement in the distribution of health care according to “need” and finance according to “ability to pay” (Williams 1993). This viewpoint contrasts with the libertarian perspective, which favours predominately private health systems and contends that people should be able to use their income and wealth to get more or better health care than their fellow citizens if they so wish. The libertarian stance limits state intervention to the provision of a minimum standard of care for the poor.
Another influential theory of justice that has informed health policy development is the Marxist theory of justice. The Marxist theory generally emphasizes that people deserve to have their needs met; thus, it makes needs the measure of right (Hoffman and Spitzer 1985). In its simplest form, the Marxist theory promotes a distributive rule based on “to each according to his needs” and “from each according to his ability”. The operation of this rule, according to Lenin, will result in actual equality (Gillon 1986, p.89). However, like other moral theories, the Marxist theory faces a conceptual problem of what exactly it refers to as needs.
In this study, Rawls’ theory of justice, in particular, his ‘maximin principle’, is adopted as the basic framework. Since no single normative theory of justice exhaustively covers the issue of equity, the study also draws on some aspects of the entitlement theory of Robert Nozick. The next section reviews the utilitarian position and highlights the limitations of applying it to this study.
40 3.2 Utilitarian Theory The utilitarian perspective dates back to philosophers such as Bentham (1780), Sidgwick (1874) and, more recently, Harsanyi (1976). It consists of several different forms. As a normative theory of justice, it has been used widely in the field of welfare economics for distributive judgement. The basic tenet of the theory is that justice prevails when a given distribution achieves “the greatest possible happiness for the greatest possible number of people” (Sidgwick and Singer 2000). Thus, the utilitarian approach considers the sum of individual utilities as a measure of social welfare.
Utility, as viewed by Bentham (1780), is an aspect of any object or event that tends to produce different pleasures in such forms as benefit, advantage, and the prevention of pain. Ebenstein (1991) noted that all utilitarians believe that sometimes “individual sacrifices are necessary to achieve collective happiness often including (but not always) the happiness of those making the sacrificing”. A better society, to a utilitarian, is one in which there is more happiness overall rather than one with greater average happiness. Their claim is that “if happiness is all that matters, then what matters is that there is the greatest quantity of it, not what amount is possessed by individuals” (Sidgwick and Singer 2000).
3.2.1 Critique of Utilitarian Theory Despite its widely acknowledged egalitarian credentials, particularly in the field of welfare economics, utilitarian theory has been criticised on many grounds. One of the major flaws is the theory’s recognition of only one fundamental imperative – namely, that welfare be promoted as much as possible; it does not matter, in itself, how that welfare is allocated (Shaw 1999 p.117). By adopting such an ethical posture, according to Shaw, utilitarianism maintains that all that matters is the maximisation of utility. Sabbagh (2001) made a similar criticism. She argued that “the aggregative character of the utilitarian principle assumes that it is possible to combine individual utilities in order to achieve a general measure of overall utility”. The assumption that “justice is determined by the beneficial outcome to all individuals in the society makes the utilitarian position “not sensitive to personal differences” (Sabbagh 2001 p. 243).
41 In On Economic Inequality, Sen and Foster (1973 p.16) observed that the trouble with the utilitarian approach is that “maximising the sum of individual utilities is supremely unconcerned with the interpersonal distribution of that sum”. This makes the utilitarian approach particularly unsuitable to measuring or judging inequality. By focusing on maximising the sum-total of individual utilities, the danger of applying the utilitarian principle as a guide for resource allocation is that people who might already be better off and as a result derive more utility from a given uniform allocation, would be allocated more resources in order to maximise the sum-total of utility. Although utilitarians attempt to counter this with the same rating of the utility function of everyone, Sen and Foster maintain that this does not exonerate utilitarians from the above criticism. By making the sum of individual utilities its primary concern, utilitarianism shows no interest in the unequal welfare levels of different individuals, and is a blunt approach to measuring and judging inequality (Sen and Foster 1973 p.16- 17).
In Inequality Re-examined, Sen (1995 p.6) attacked the utilitarian approach for “ignoring freedom and concentrating only on achievements, and also ignoring achievements other than those reflected in mental metrics such as pleasure, happiness or desire”. He noted that the way utilitarians see individual advantage as particularly limiting in the presence of entrenched inequalities. According to Sen, in “situations of persistent adversity and deprivation, the victims do not go on grieving and grumbling all the time, and may even lack the motivation to desire a radical change of circumstances” (Sen 1995 p.6). Such people, even though might be greatly deprived and confined to a very reduced life, may not appear to be quite so badly off in terms of pleasure or happiness (the mental metrics that utilitarians emphasise). Sen (1995) surmises that the utilitarian approach of maximising the sum of individual utilities would provide a misleading picture of well being.
The utilitarian approach has also been criticised by some libertarians as justifying too much redistribution of resources (Nozick 1974). Using the natural-rights theory, which claims that the individual has a right to the fruits of her labour, libertarians generally postulate a right to property and argue against excessive redistribution (Gandjour and
42 Lauterbach 2003). They also attack the utilitarian sole objective of maximising pleasure, arguing that there is more to life than experiencing pleasure (Nozick 1974).
The many criticisms against the utilitarian theory share a common point: they all consider the utility maximisation stance as tantamount to giving no independent moral weight to equality. As Shaw observed, utilitarianism is seen by critics as a “morally repugnant theory of social arrangement” that allows inequality to flourish as long as it maximises overall well being (Shaw 1999 p.122). The next section reviews the entitlement theory.
3.3 Entitlement Theory The entitlement theory of Robert Nozick14 is one of the best-known libertarian theories. It is based on the premise that the distribution of goods in a society is just if and only if all are entitled to the holdings they possess (Nozick 1974). Thus, people are entitled to their possessions as long as they have acquired them fairly, or, as long as in the process of acquiring them, they did not violate other people’s Lockean rights; the right that individuals are morally entitled to the fruits of their labour15 (Shaw 1999, 2003). Nozick observed that a valid conception of justice must consist of three principles: a principle of justice in (original) acquisition, a principle of justice in transfer, and a principle of rectification of an unjust holding (Nozick 1974 p.46-49).
The principle of justice in (original) acquisition deals with a description of how people legitimately acquire holdings; that is, the appropriation of un-held goods or the creation of new goods. The principle of justice in transfer which deals with ownership by legitimate transfer states that, if a person possesses a holding through a legitimate transfer, then he or she is entitled to that holding. Although Nozick did not specify what he meant by legitimate transfer, many consider acquisition by ‘purchasing’ or through a ‘gift’ as examples of legitimate transfers. The principle of rectification of an unjust
14 Robert Nozick, Anarchy, State and Utopia, New York, 1974 15 Nozick’s entitlement theory was inspired by the work of John Locke. For more on Locke and the Lockean right, see Locke, John. Second Treatise of Government, ed. C. B. Macpherson. Indianapolis: Hackett, 1980. 43 holding deals with rectification of past injustices, should any of the first two principles be violated. From these three principles, Nozick arrived at a principle of justice which states that “a distribution is just if it arises from another (just) distribution by legitimate means” (Nozick 1974).
Unlike the end-state theories such as utilitarianism, the entitlement and other libertarian theories of justice are ‘historical’ and deal largely with ‘procedural justice’. A historical principle of justice holds that “past circumstances or actions of people can create differential entitlement” (Westphal 1996 p.11). Thus, theories that are historical rather than end-state see the justice of a distribution as entirely dependent on the path used to reach it (Varian 1975), or, in the unhindered working of fair procedures rather than in some distributed results such as increasing public utility (Beauchamp and Childress 1989). This position makes historical/procedural theories different from the end-state theories, which largely see a just distribution in the light of the results of the distribution (who has what). With end-state theories, all one needs to look at in judging the justice of a distribution is who ends up with what. In other words, when comparing any two distributions, one needs to look only at the matrix presenting the distribution; no further information is necessary.
The entitlement theory’s position that the individual’s right to possess holdings, if legitimately acquired be respected, is often described as closely related to a free market system where people are entitled to their possessions, if genuinely acquired. The free market system forces no one is to redistribute his/her resources for the benefit of others. Justice is promoted by ensuring that procedures are fair and that there are no restrictions to the free market procedures (Varian 1975). In the context of health, the libertarian perspective in general has, indeed, been quite influential. The American health care system, for example, is rooted in libertarian tradition (Pellegrino1999; Hsiao 2003). Donabedian (1971) noted that the two theories of justice in the philosophy literature most frequently encountered in the context of medical care are the libertarian and the Marxist approaches.
44 3.3.1 Critique of Entitlement Theory The main criticism of the entitlement and libertarian theories, as far as health care is concerned, is their emphasis on private health care or the seizure of it by private health care enthusiasts. Libertarian theories are often described as supporting a mainly private health care sector, where health care is rationed primarily according to willingness (and ability) to pay. Such theories demand that state involvement in health care should be minimal and be limited to providing a minimum standard of care for the poor (Wagstaff and van Doorslaer 1998). Libertarians generally view free choice as essential to distributive justice.
Nozick (1974) observed that government action is only appropriate to protect the entitlements and rights of its citizens rather than be responsible for the distribution of the health resources. The American health system, as mentioned earlier, is perhaps the best example of a health system that operates predominantly on libertarian principles. Although not exclusively, the United States has largely accepted a free market rule that distribution of health services and goods are best left to the marketplace, which operate on the material principle of ability to pay and invokes some form of libertarian theory of justice as its justification (Beauchamp and Childress 1989; Pellegrino1999; Hsiao 2003). With regards to policymaking, however, it is important to note that different governments have different ideologies, and hence, a mixture of different conceptions of justice is usually common in many countries.
Libertarians like Nozick can also be criticised as fundamentally unconcerned about equality. Indeed, Nozick argued that while the demand for equality is widespread, there is a general absence of argument as to why equality must be built into any theory of justice (Nozick 1974 p. 234). He maintained that underpinning most arguments for economic equality is the unsupported claim that society should provide for the important needs of its members. The next section reviews Rawls’ theory of justice as fairness which is used as the main framework for this study.
45 3.4 Rawls’ Theory of Justice: The Framework for Current Study The Rawlsian theory of justice as fairness (Rawls 1972) is used as the framework for this study. The theory is widely acknowledged as one of the most salient social justice theories put forward in the 20th century. It is described as an institutional theory because of its emphasis on the basic structure of society and claims of offering a framework for structuring major societal institutions - political, economic, legal, and social. Rawls argued that the basic structure of society and the major societal institutions play a pivotal role in regulating the distribution of goods and social burdens among members of society and in determining their life chances. He wrote:
“The basic structure is the primary subject of justice because its effects are so profound and present from the start. The intuitive notion here is that this structure contains various social positions and that men born into different positions have different expectations of life determined, in part, by the political system as well as by economic and social circumstances. In this way the institutions of society favour certain starting places over others. These are especially deep inequalities. Not only are they pervasive, but they affect men’s initial chances in life; yet they cannot possibly be justified by an appeal to the notion of merit or desert. It is these inequalities, presumably inevitable in the basic structure of any society, to which the principles of social justice must in the first instance apply” (Rawls 1972 p. 54).
The foundation of Rawls’ conception of social justice is his belief that the least advantaged members of society, as measured by their possession of the primary goods, should be the yardstick against which the justness of the basic structure of society be judged (Rawls 1972; McGary 1999). Rawls argued that social arrangement is a communal endeavour to advance the progress of all who are part of the society (Rawls 1972). From a moral standpoint, therefore, inequalities of birth, natural endowment, and historical circumstance are undeserved, and thus, persons in a cooperative society should make more equal the unequal situation of naturally disadvantaged members (Rawls 1972; Beauchamp and Childress 1989). Thus, from a moral or justice standpoint, society has a duty towards improving the lives of its least advantaged members. But how should society distribute its resources (primary goods) to make more equal the situation of the naturally disadvantaged? What principle of justice should govern resource distribution?
46 Rawls’ answer to these questions lies in his ‘maximin’ principle, which is based on a ‘hypothetical social contract model’ in which valid principles of justice are chosen by rational agents acting in an ‘original position’ (Shaw 1999; Beauchamp and Childress 1989; Phillips 1979; Rawls 1972). According to Rawls, rational agents acting from behind the ‘veil of ignorance’ in an ‘original position’ where their personal circumstances and positions on the socio-economic ladder are covered would choose two justice principles for the distribution of social primary goods. The first principle is that:
“Every person is to have an equal right to the most extensive total system of equal basic liberties compatible with a similar system of liberty for all” (Rawls 1972 p. 302).
The second principle:
“Social and economic inequalities are to be arranged so that they are both (a) to the greatest benefit to the least advantaged, consistent with the just savings principle, and (b) attached offices and positions open to all under conditions of fair equality of opportunity” (Rawls 1972 p. 302).
Rawls asserted that the two principles must be lexically ordered such that the first principle takes priority over the second. In other words, society must first achieve equal and extensive distribution of basic liberties. Once this is achieved, it is possible to have unequal distribution of social and economic resources if that distribution occurred under the condition of fair equality of opportunity, and if it is to the advantage of everyone, particularly the worst-off in society. Rawls argued that these two principles of justice chosen via his “contractarian” procedure would lead to the “desire to improve the position of the worst-off” in society (Rawls 1972; Andersson and Lyttkens 1999). He derived what he calls the general conception of justice from the two principles which states that:
“All social primary goods - liberty and opportunity, income and wealth, and the bases of self-respect – are to be distributed equally unless an unequal distribution of any or all of these goods is to the advantage of the least favoured” (Rawls 1972 p. 303).
This study focuses on the second principle of the Rawlsian theory. That is, the ‘difference principle’, which deals with maximisation of the greatest benefits of the least advantaged and places emphasis on equality of opportunity. Before extending the 47 Rawlsian theory of justice to the current study, it is important to review its general applicability to health and health resource distribution.
3.4.1 The Rawlsian Theory and Health The issue of whether or not the Rawlsian theory of justice should be applied to health and health care distribution has generated debate among scholars. The theory itself was not designed to address matters of health care. Indeed, Rawls assumed a perfectly healthy state where none of his rational agents acting from behind the veil of ignorance was sick. He therefore concentrated on the distribution of social primary goods, which exclude health. The index of primary goods (the goods that every rational man would prefer more of rather than less) includes rights and liberties, powers and opportunities, income and wealth, and the bases of self-respect (Rawls 1972 p. 92). According to Rawls, “health and vigour, intelligence and imagination, are natural goods”, which “although their possession is influenced by the basic structure, they are not so directly under its control”. Thus, Rawls made it clear that health is not included in the stock of primary goods.
Despite Rawls not including health in the list of social primary goods, others have contested this and argued that health could and should be included in the index of primary goods (Green 1976; Daniels 1985; Peter 2001; Bommier and Stocklov 2002). However, while some analysts support the inclusion of health care, others have raised concerns about the potential negative effects of such inclusion, particularly, the likelihood of weighting problems between health care and the other index of primary goods. Buchanan (1984) argued that there are potential weighting problems with the inclusion of health in the list of primary goods and until a solution is found to such a problem, the theory of justice can shed only limited light upon the question of priority- relations between health care and other goods. He further argued that the “informational constraints imposed by Rawls’ veil of ignorance preclude a solution to the problem of weighting health care against other primary goods” (Buchanan 1984 p. 61). Arrow (1973) similarly argued that if health care is added to Rawls primary goods, “the force of ‘Rawls’ second principle, which requires inequalities to work to the advantage of the worst-off, would drain excessive resources into the satisfaction of the special needs of 48 persons in extreme health-care needs, perhaps to the point where the rest of society is reduced to poverty” (Arrow 1973 p. 251).
Those who argue for the extension of Rawls theory to health care share different views on how it should be extended. Some have argued for extension through inclusion of health in the index of primary goods. Others see the ‘equality of opportunity’ as the appropriate medium for extension. Green (1976) observed that “access to health care is not only a social primary good, but possibly one of the most important of such goods ‘because’ disease and ill-health interfere with our happiness and undermine our self- confidence and self-respect”. Therefore, health care distribution, he argued, must be subject to the basic principles of Rawls.
Norman Daniels’ (1981, 1985) attempt to extend the Rawlsian theory to the health field has been influential. He argued for the extension of the theory to the health field through the ‘fair equality of opportunity principle’. Daniels noted that “the most promising strategy of extending Rawls’ theory without tampering with useful assumptions about the index of primary goods” is simply by including “health care institutions and practices among the basic institutions involved in providing for fair equality of opportunity”. According to Daniels (1985 p.45), “because meeting health care needs has an important effect on the distribution of opportunity, the health care institutions are regulated by a fair equality of opportunity principle” (see Section 3.4.2).
There has been recent interest in linking Rawlsian theory to health equity debates. Peter (2001) demonstrated an indirect approach. Quoting from Rawls’ Political Liberalism (1993), he explained that the way basic social institutions work, in the view of Rawls, is more important than the resulting outcomes, as characterised by distributive patterns that are observed. Thus, in Rawlsian justice, “the focus is shifted from health outcomes to the mechanisms implicit in basic social institutions that produce social inequalities in health” (Peter 2001 p.166). Bommier and Stocklov (2002) argued that Rawls’ first principle can be used as a basis for defining health inequality once we recognise that the actual health of individuals is dependent on individuals’ health endowments as well as how health endowments are transformed into actual health through access to health resources.
49 It worth acknowledging that while Rawls’ theory of justice has been widely acclaimed as one of the most influential theories of the 20th century, it has not escaped criticisms. In particular, the maximin principle, which permits inequalities to exist if they are in favour of the least-advantaged has been criticised by many people. Some scholars have questioned the ‘fairness’ of allowing inequalities in society to persist in order to benefit the least well-off (Barry 1981). Additionally, Rawls’ concentration on primary goods and overly concern for freedoms people enjoy in pursuing their ends has also been criticised. Sen (1995) observed that equality of freedom to pursue our ends cannot be generated by achieving equality in the distribution of primary goods.
3.4.2 Equality of Opportunity and Health The concept of equality of opportunity, although controversial, is central to social and distributive justice. Liberal political philosophy has relied on this notion, which deals essentially with procedural justice to justify a system in which unequal outcomes are thought to be morally acceptable (Daniels 1985). Fair equality of opportunity, as developed by Rawls in the theory of justice as fairness, holds that society should give equal chance to life prospects. Accidents of birth such as race, sex and ethnic origin should not be the basis for distributing society’s benefits and burdens. As part of his second principle, Rawls argued that advantages arising from these accidents of birth or ‘natural lottery’ are arbitrarily distributed and hence, none of us deserves the advantages conferred by them. They must be responded to using resources of society to enhance the opportunity of those disadvantaged by such accidents of birth (Rawls 1972 p.73-74).
Rawls further argued that society must be held responsible for guaranteeing the individual a fair share of basic liberties, opportunity, and all-purpose means, like income and wealth, needed for pursuing individual conceptions of the ‘good’. Likewise, individuals should also be held accountable for choosing their ends in such a way that they have a reasonable chance of satisfying them under just arrangements (Rawls 1982b; Daniels 1985). Rawls was fundamentally concerned with opportunity to pursue careers (office and positions open to all) that have various attachments of benefits measured largely in terms of his primary goods. He emphasised the elimination of formal and legal barriers to persons seeking positions by taking positive steps to 50 enhance the opportunity of those disadvantaged by such social factors and family circumstances. He specifically stressed the role education can play in eliminating the barriers that disadvantaged people, indicating that his difference principle would allocate resources to improve the long-term expectations of the least favoured (Rawls 1982b).
Daniels (1981) developed the link between the normal functioning of individuals and the range of opportunities available to them as part of his extension of the Rawlsian theory to health sector. He argued that if it is vital to use resources available to society to counter the advantages in opportunity resulting from accidents of birth then, it is equally important to use resources to counter the natural disadvantages resulting from disease (Daniels 1981). He contended that disease and disability impair the normal functioning of individuals, and hence, affects the range of opportunities available to them. In his view, since maintaining a “normal functioning” makes a limited but significant contribution to protecting the range of opportunities open to all individuals, it is reasonable to see a principle that guarantees equality of opportunity as the appropriate principle to govern the distribution of health care. His definition of health care covers primary and secondary preventive health as well as medical services (Daniels 1993).
Amartya Sen has made valuable contributions to the health and social justice debate in general, and the equality of opportunity argument in particular. In Inequality Re- examined, he argued that a more adequate way to consider ‘real’ equality of opportunity must be through “equality of capabilities”, which include health capabilities (Sen 1995). Sen objected to Rawls focus on primary goods that individuals hold as a measure of well being. His contention was that the amount of primary goods or resources held by people does not provide adequate information on how to evaluate justice. What matters, to him, is what people can do, or be, with the resources they hold, and this is captured by their ‘capabilities’ (Sen 1990). Sen saw capability as a “set of functionings” that a person can achieve. This set of functionings mirrors the actual freedom of choice that a person has over alternative lives s/he can lead (Sen 1990).
The relevance of Sen’s notion of basic capabilities to the health and social justice debate lies in the special moral prominence it gives to health capabilities. Ruger (2004)
51 observed that Sen’s formulation generally means that society should make efforts to bring each individual’s health functioning as close as possible to certain level of minimal normal functioning (as much as an individual’s circumstances may permit). Basic capabilities, in Sen’s view, include the ability to avoid escapable morbidity and premature and preventable mortality. The direct moral importance given to health by the capability argument contrasts with Daniels’ view that health care is special because of its impact on equality of opportunity (Daniels 1981; 1985). However, they both seem to share the view that resources be allocated on the basis of medical need, not ability to pay. The next section critically appraises each of the three theories discussed above in the context of this study.
3.5 Theories of Social Justice and this Study The three theories reviewed above (utilitarian, entitlement and the theory of justice as fairness) have different levels of applicability to equitable allocation of resources in Ghana, and specifically, to this study.
Utilitarian theory The utilitarian approach has the most limited applicability as a guiding principle for health resource allocation, particularly in countries with entrenched health inequalities such as Ghana. One may argue that allocation of resources to maximise the greatest possible utility for the greatest number of people in health care settings, essentially means more resources can be efficiently allocated to maximise the health of the entire population. In Ghana and other African countries, where preventable diseases, such as malaria, are the leading causes of morbidity and mortality, allocating resources on the basis of the utilitarian principle to combat malaria could lead to the greatest possible improvement in population health (by saving many lives) something public health policies strive to achieve. However, in allocating resources to maximise aggregate population health (the sum total of individual utilities), the utilitarian approach suggests that personal circumstances do not matter. Inequalities across population groups, for example, are not significant as long as the overall utility of the general population is
52 maximised. In health care settings with entrenched inequalities like Ghana, it is difficult to justify allocating resources without due regard for such inequalities.
Another difficulty in applying the utilitarian approach to this study related to the problem of measurement. Health maximisation requires a composite measure of health gains capable of being estimated across illnesses and health care interventions (Segall 2003) as well as across different population groups. What is of great concern with regards to measurement is the utilitarian approach of counting ‘everybody for one and nobody for more than one’; that is, assigning equal weight to individual utilities. Adopting this measure in resource allocation would mean that nobody or no population group deserves any special attention, irrespective of the level of health needs and what might have contributed to that level of need. Each district in Ghana, be it rural or urban, rich or deprived, would be weighted equally when resources are distributed. One may even argue that the utilitarian approach would require that resources be allocated towards rich districts with relatively good health infrastructure to derive as greater health benefits as possible in order to maximise the sum-total of health gains.
By adopting such a moral principle, the extent of deprivation in poor and remote districts where needs may be greatest but health gains are likely to be low because of poor health infrastructure, will, in Sen’s terminology, be “substantially muffled in the utility metric’ (Sen 1995 p.7). In summary, there is no fairness in assuming that individuals or groups have the same utility function in a country where inequalities are so pervasive and everything points to the fact that without extra attention to certain groups, unjust and unfair disparities will persist. These weaknesses made the adoption of the utilitarian theory for a study of this nature questionable.
Entitlement Theory The libertarian perspective of justice has some crucial equity implications for allocation of health care resources in countries like Ghana, where current inequalities in health may be linked to unfair distribution of resources in the past. Historically, allocation of health resources in Ghana tended to favour regions in the southern part of the country. This dates back to the colonial era (see Chapters 2).
53 According to the Nozickian principle of rectification of past injustices in distribution, the current inequities in health have to be rectified. In terms of legitimacy of possessions, one may argue that regions in the southern Ghana are not legitimately entitled to receive more resources than those in the north. The colonial administration used the country’s resources it “illegitimately” acquired to develop the infrastructure in the south for their own use and at the expense of the north. There is little doubt that northern Ghana became relatively worse off than would have been if the distribution system established during the colonial era had been more just. Nozick’s third principle - rectification of past injustices - has the potential to redistribute resources in favour of the deprived regions in northern Ghana, which were unjustly treated in the past. In this regard, the entitlement theory has some potential to improve equity by shifting resources to worst off regions.
Despite the potential advantages of using the entitlement theory as the framework for the current study, there are several ambiguities that make its application to this study difficult. Nozick proposed that if an injustice has occurred, looking at the history of past distribution, we must determine what the expected outcome would have been if the injustice had not occurred and use that to rectify the situation. The difficulty here is that the current inequities in health result from a combination of factors. They are not only the result of unjust distribution of health care resources, but also unequal distribution of other resources such as those for education, housing and water and sanitation. This scenario makes it difficult to isolate the injustice that might have occurred as a result of unfair distribution of health resources from those resulting from unequal distribution of other resources in a bid to rectify past injustices. Nozick did not consider how such problems might be resolved.
Another difficulty with the application of the entitlement theory to the current study relates to how far back in history should one go in wiping clean the “historical slate” of injustices in distribution of resources. The pre-independence colonial administration of Ghana initiated the problem of unfair allocation of resources in favour of the regions in the south (Tsikata and Seini 2004). Post-independence governments, to a large extent, have continued with the lop-sided development programmes and resource allocation initiated by the colonialists. In seeking to rectify the injustices of the past, it is uncertain
54 where the cut off point should be. This issue is similarly not well addressed by the Nozickian historical theory of entitlement.
Finally, it is an oversimplification to attribute the current situation of inequalities solely to the resource allocation process. Thus, while the means by which an allocation was reached has something to do with justice, in many countries, it is not the whole story (Varian 1975). In Ghana and other African countries, corruption of public officials is endemic (Transparency International 2004). Local politicians frequently use public funds for personal gain. This might have played a significant role in depriving some regions of their fair share of resources for health care. In rectifying past injustices in distribution, it is unclear how one should separate the unfair treatment of some regions as a result of unfair state allocation and policies from a situation where the state has justly distributed the resources but corrupt local politicians have squandered the funds and left the region deprived. One might well argue that the role of the state is to prevent corruption, and hence, its inability to do so makes it culpable. The issue then is how much a price should other regions pay by losing part of their allocations in the name of rectifying past injustices in distribution, when in fact past distribution has not been ‘wholly’ unjust? The entitlement theory does not provide clear guidance on how to rectify past injustices.
Rawls’ Theory of Justice It was noted earlier that this study is predominantly interested in a principle of social justice that could guide the distribution of health resources in areas where there are already marked inequalities in health, such as in Ghana. More than any of the other theories reviewed, the Rawlsian theory, specifically his ‘difference principle’, provides the best prospects for maximising greater benefits for the least-advantaged or the most deprived. Utilitarianism, for example, as demonstrated earlier, is only concerned with maximising the sum of individual utilities and not the interpersonal distribution of that sum, as observed by Sen and Foster (1973). Such a principle will only add to the inequalities in health, if applied, rather than reducing them.
Insights from Daniels’ extension of the Rawlsian theory to the health sector demonstrate that disease and ill-health impair people’s capacity to function normally as human 55 beings. There is, therefore, the need to address health care needs if a ‘fair equality of opportunity’ is to be guaranteed. The basic issue however is - whose health needs should be addressed within the constraints of available resources? Rawls provided no direct answer to this question since he did not specifically address the issue of health. However, his second principle states that “social and economic inequalities are to be arranged so that they are to the greatest benefit to the least advantaged”. Extending this to the allocation of health resources, one could argue that resources ought to be allocated differentially to secure the greatest benefit to the worst-off groups in society. Thus, Rawls’ second principle provides the clearest justification of giving priority in resource allocation to the health needs of worst-off groups, hence, its adoption as the main framework of this study.
In terms of the volume of resources that must flow to the worst-off, Daniels et al. (1999 p.125) have noted the difference principle is not “just a mere trickle down principle but one that requires maximal flow in the direction of helping the worst-off groups”. Thus, the difference principle demands more than simply “spreading” a bit of resources to make minimal improvements in the lives of the worst-off. One might question how egalitarian is it to spend much of society’s resources to help the worst-off, most likely at the expense of the better off? Perhaps in the stricter sense of the word, this might not be “very” egalitarian. But the issue is – how egalitarian should allocation of resources be in a non-egalitarian society where historically people have not been justly treated in terms of receiving a fair share of societal resources?
Rawls has stated clearly that the difference principle was not formulated as a principle of redress. He wrote:
“Now the difference principle is not of course the principle of redress. It does not require society to try to even out handicaps as if all were expected to compete on a fair basis in the same race. But the difference principle would allocate resources in education, say, so as to improve the long-term expectation of the least favoured. If this end is attained by giving more attention to the better endowed, it is permissible; otherwise not” (Rawls 1972 p. 101).
The idea of not seeking redress is probably because Rawls saw disadvantage as resulting only from accidents of birth and not historical injustices perpetrated through bad or inequitable state policies. The reality is that, in Ghana and many other countries,
56 inequalities in health are the result of past injustices and bad state policies, including maldistribution of state resources, which have impoverished sections of the population. In such instances seeking redress becomes crucial and one can draw some insights from Robert Nozick’s principle of rectification of past injustices to support why there is the need to give the health needs of the worst-off greater priority.
Investing in education, as Rawls proposed, will no doubt improve the long-term conditions of the worst-off, but in the short and medium terms, the Ministry of Health also has a responsibility to redirect health resources to meet the health needs of the worst-off as a matter of correcting previous injustices in the allocation system. The next chapter examines how the concept of health equity has been interpreted and measured and which interpretations of equity are mostly applied to the distribution of health care financial resources.
57 Chapter 3. Summary of key points
This chapter has reviewed the key theories of social justice that shed light on how health and health care resources ought to be distributed.
Three major theories – the utilitarian perspective, Robert Nozick’s theory of entitlement and John Rawls’ theory of justice as fairness were reviewed. Norman Daniels’ extension of Rawls theory to health care and Amartya Sen’s basic capability theory were also examined.
Rawls’ Difference Principle- from his theory of justice as fairness was adopted as the basic framework for the study. This was due to its compatibility with the objectives of the study, particularly the theory’s prescription:
• That inequalities in society must be arranged to benefit worst-off members
• That societal resources be used to counter advantages conferred on some people by accidents of birth such as race, ethnicity, sex etc.
• That society should enhance equality of opportunity by improving the long-term expectations of the least favoured, particularly by investing in education.
Rawls primary goods did not include health but Norman Daniels’ extension has shown that health fits well in the Rawlsian analysis, given that disease and ill-health affects the normal functioning of human beings, and hence, impact on the range of opportunities available to individuals.
The study does not share Rawls’ view of not seeking redress for past injustices. It rather adopts a stance which combines rectification of past injustices with improving long-term expectations of the least advantaged.
Some insights were also drawn from Robert Nozick’s principle of rectification of past injustices to support the need to address the imbalances in the resource allocation system in the short and medium term.
58 CHAPTER 4
EQUITY AND EQUITABLE DISTRIBUTION OF HEALTH CARE RESOURCES
Overview Equity and, more specifically, equity in relation to health, is a controversial concept. Although scholars disagree on the methods to use to allocate resources across different population groups, there is growing consensus on the importance of a fair and just distribution of health care and health resources. In Chapter 1, key interpretations of equity were highlighted. This chapter has two objectives: first, to examine the definitions and measurements of health equity and second, to review how health care resources may be allocated to achieve the health policy objective of greater equity.
SECTION 1 - EQUITY: CONCEPT, INTERPRETATION AND MEASUREMENT
4.1 Equity and Equality: A Conceptual Clarification Conceptually, equity is rarely defined without reference to the term ‘equality’. Mooney (1992) outlined seven possible definitions of equity (Box 4.1).
Figure. 1 ns of Equity Box 4. 1 Various Definitions of Equity
Equality of expenditure per capita Equality of inputs per capita Equality of inputs for need Equality of access for equal need Equal utilisation for equal need Equality of marginal met need Equality of health (Mooney 1992 p.103)
59 The closeness of the terms “equity” and “equality” often misleads people into applying them interchangeably. However, equity and equality have some important and noteworthy distinctions in meaning. Le Grand (1987 p. 259) observed that “equality is essentially a descriptive term” while “equity is essentially a normative term”. Thus, a particular distribution can be observed and, on the basis of the observation, one can conclude whether the distribution is equal or unequal. However, the “presence or otherwise of equity cannot be established by reference solely to an observed distribution”. One has to make value judgements to couple with the observed facts about the distribution. In effect, while “equity statements are statements of value; equality statements are statements of fact” (Le Grand 1987 p. 259).
Sen (1995 p.660) had a slightly different view, stating that, “equality is an abstract idea, which has limited cutting power unless one begins to specify what is it that is to be equalised”. In other words, specification of the space in which equality is to be pursued, and the equitable accounting rules that must be followed to arrive at the aggregative concern as well as distributive ones are the central issues. In his book, entitled The Moral Sense, Wilson (1993) saw equality as related to what people are entitled to and equity as dealing with rewarding proportionately. Wilson maintained that both equality and equity are components of our ideas about fairness, and that these components may be difficult to reconcile. To sum up, while the concept of equity deals with fairness and is based on value judgement, that of equality is about entitlement and is based on facts.
Conversely, inequity and inequality have subtle but important differences. Inequity has an ethical dimension. Braveman (2003) observed that inequity could have an accusatory, judgemental or morally charged tone. Margaret Whitehead provided one of the most cited definitions of [health] inequity: “differences [in the distribution of health] that are “unnecessary and avoidable, and in addition, are also considered unfair and unjust” (Whitehead 1985 p.7). This means that, for a situation to be deemed inequitable, the cause has to be examined and judged to be unfair in the context of the social setting with which one is dealing. Whitehead argued that not all differences in health can be labelled as inequitable. She described seven health differentials, some of which are inevitable/unavoidable and others unnecessary and unfair. Differences in health that are due to natural or biological variation, including genetic make–up, were labelled as unavoidable. Gakidou et al. (2000) however, believe that with current progress in 60 technology, it is possible that such factors as genetic variations will become amenable to change in future.
Inequality describes an observed disparity in a distribution in relation to health care access. It makes no moral or intrinsic judgement about the differences. Health inequality is defined simply as the variations in health status across individuals in a population (Murray et al. 1999). This contrasts with health inequity, which emphasises the disparities that are considered avoidable, unfair and unjust. The contrast between inequality and inequity can also be seen in terms of their respective prescriptive powers. While the inequality is non-prescriptive (i.e., it makes no demands regarding addressing the observed disparity), inequity is more prescriptive, eliciting response, which seeks to promote fair distribution of resources (Whitehead 1990; Marchand et al. 1998).
4.2 Defining Equity In principle, equity objectives are key to many health systems around the world (Daniels 1985; Mooney 1986; Le Grand 1987; van Doorslaer et al. 1993; Oliver et al. 2002). In practice, they are “notoriously difficult to interpret and operationalise” (Birch and Abelson 1993), largely because equity means different things to different people. Health systems, as observed by McIntyre et al. (2000), rarely establish clearly or specify fully the equity goals they seek to achieve. One important distinction often made in defining equity is to distinguish between horizontal and vertical equity. This is examined before looking at some of the common interpretations of equity applied in the health field, including equal treatment for equal need, equality of access, and equality of health (Le Grand 1982; 1987; Mooney 1986; Culyer et al. 1992a, 1992b).
4.2.1 Horizontal and Vertical Equity In general, horizontal equity refers to equal treatment for equal need while vertical equity is defined as unequal (but equitable) treatment of unequals (Mooney 1994). Policymakers often face a dilemma choosing between horizontal or vertical equity objectives to guide health decision-making, particularly with regards to payment,
61 resource allocation and service delivery. Horizontal equity does not serve anybody’s special interest but vertical equity appropriately considers relative differences in needs of different people. These two concepts of equity have their roots in the Aristotelian formal principle, which states that “equals should be treated equally and unequals unequally in proportion to the relevant inequalities” (Gillon 1986 p.87).
In the delivery of health care, horizontal equity (i.e. equal treatment for equal need) requires that individuals who report the same degree of sickness should receive the same amount of resources (Wagstaff et al. 1991). Vertical equity, on the other hand, means individuals with unequal health needs are treated in an appropriately dissimilar manner in proportion to their level of inequality. In health care financing, horizontal equity, as interpreted in the health economics literature, means households with the same ability to pay make the same contributions. Vertical equity, by contrast, refers to a situation where households of unequal ability to pay make appropriately dissimilar payments (Wagstaff and Van Doorslaer 1993; 1998). Finally, in terms of funding of health care16, horizontal equity means allocation of equal or equivalent resources for equal needs. Vertical equity refers to allocation of different resources for different levels of need.
The distinction between horizontal and vertical equity is important as they have different implications for policy. Bambas and Casas (2001) observed that “a universal health care plan might appeal to proponents of horizontal equity on the basis that everyone needs health care at some point” while “targeted programs for the poor would appeal to vertical equity”. The choice of horizontal or vertical equity presents several issues that must be resolved. In the case of horizontal equity, one is confronted with the issue of how to determine the ‘same ability to pay’. With regards to vertical equity, Wagstaff noted that there is always the concern about the precise form that differential treatment should take. The next sub-sections examine in more detail the various interpretations of equity.
16 Funding of health care is considered here as primarily concerned with resource allocation rather than payments. The bulk of the health economics literature focuses on financing in terms of payments rather than funding in terms of resource allocation. 62 4.2.2 Equal Treatment for Equal Need This interpretation of equity means people with the same health care need should receive the same amount of treatment (both in quantity and in quality). The underlying assumption, as observed by Le Grand (1987), is that distribution of health care “should be independent of the distribution of income, wealth, or any other form of economic or political power”. Proponents of this definition of equity do not rule out differences, rather they contend that differences in treatment should be related only to differences in ‘need’.
Two practical difficulties in relation to this definition have been identified - how to measure ‘treatment’ and how to define ‘need’. Critics often argue that two persons with the same (self-reported or professional determined) sickness may receive the same treatment but one might respond well to the treatment while the other might not respond at all. They conclude that equity cannot be deemed to have been achieved in such instance. The issue is whether equal treatment should result in equal outcome. If variations in biological make-up are accepted as unavoidable, then different outcomes from equal treatment may not be inequitable. The complexities associated with the definition of ‘need’ are discussed later in this chapter (see Section 4.4).
4.2.3 Equality of Access (Equal Access for Equal Need) Norman Daniels (1985) observed that the literature on equity of access is complex and confusing, partly because “access is itself a complicated notion, a composite of many factors”. Several authors have argued for more attention to the equality of access interpretation (Le Grand 1986; Mooney 1983; Evans 1984). According to Le Grand, equality of access is “the requirement that people should face the same personal cost of receiving medical treatment” (Le Grand 1986). The proponents of this definition believe that equity is not served if some people are charged more than others, or have to travel further than others, or are required to wait longer than others in order to receive health care. The obvious difficulties with such an interpretation is the definition of access in terms of ‘cost’ of receiving treatment and how ‘personal cost’ is measured, whether it
63 should be measured in monetary terms, or in terms of utility or satisfaction forgone (Le Grand 1987).
The ‘cost-based’ interpretation of equality of access implicitly assumes that two people facing the same cost to health care would enjoy the same access. In other words, it is irrelevant whether people have money or not to pay for the “same cost they face”, or whether a given health care facility is actually being utilised or not. This definition thus emphasises the opportunity to use health care rather than actual usage. Culyer, van Doorslaer and Wagstaff have rejected the equality of access interpretation, as a whole, and the cost-based definition of access in particular. They contend that depending on income and other factors, people may have differential access even if they face the same cost (Culyer et al. 1992a; 1992b). Culyer and Wagstaff have further argued that “irrespective of how one interprets access, and irrespective of whether equality of access is applied only to those in equal need, application of the principle of equality of access to health care will not yield an equal distribution of health” (Culyer and Wagstaff 1993 p.453).
Olsen and Rodgers (1991) interpreted access differently as a “minimum attainable consumption”. To them, the meaning of different people having different degrees of access to a particular good or service is that “the upper limit on their consumption of the good or service differs”. This interpretation of access suggests that two people can face the same cost of accessing health care but still have unequal access. Thus, Olsen and Rodgers’ argument runs counter to Le Grand’s assumption that people facing the same cost to health care enjoy equal access.
In general, equality of access for equal need entails two main complications: the relationship between access and utilisation and how to define ‘need’. Mooney has distinguished between equality of access and equality of utilisation, arguing that former is wholly a supply-side phenomenon while the latter is a function of both supply and demand (Mooney 1992 p.104).
Despite the complications, equity as equality of access is generally perceived as popular among policymakers. Mooney and associates have argued that equality of access is
64 “what policymakers understand to be equity”, citing evidence from a review of policy statements in different countries including the United Kingdom, Australia, Denmark, Sweden and many others to support their case. They noted that equality of access “provides individuals with the opportunity to use health services”(Mooney et al. 1991 p. 479). Culyer et al. (1992a; 1992b) argued that many policy documents do not only contain references to ‘equality of access’ but also to ‘distribution according to need’ and ‘equality of health’.
4.2.4 Equality of Health The meaning of equality of health is self-evident; equal health status for everyone. This interpretation of equity has fewer advocates including Culyer and Wagstaff (1993). However, commentators recognise that some differences in health are simply inevitable. That is, irrespective of what is done, there would still be some variations in health status among individuals. Whitehead’s definition of inequity (Whitehead 1985 p.7) as “differences in health that are unnecessary and avoidable” and also “unfair and unjust”, supports this assumption.
There is well-documented evidence that natural and biological factors such as ageing and genetic make-up have some irreversible consequences on health, which cannot be labelled as unfair or unjust. This makes it unrealistic for anyone to interpret equity as equality of health. It is often dismissed as too expensive to be an appropriate interpretation of equity (Mooney 1986). Le Grand has observed that “since individuals’ health is located within themselves, it is impossible to take away someone’s health and give to someone else; that is, it is impossible to redistribute health”. To him, a more rational alternative is to focus on health care rather than health itself, because “health care can be redistributed by act of policy” (Le Grand 1987 p.263).
Despite the obvious difficulties with interpreting equity as equality of health, Culyer and Wagstaff (1993), have argued that equality of health should be the dominant equity principle. They used the ‘human flourishing’ debate in the philosophical literature
65 which emphasises the importance of health to human functioning or flourishing17 as the basis to argue that equity means just distribution of health. They wrote:
“It appears to be acceptable in the moral philosophy literature that a position other than one in which everyone has the same opportunity to ‘flourish’ would be hard to defend. Insofar as health is a necessary condition for ‘flourishing’, it follows that a just distribution of health is an equal distribution” (Culyer and Wagstaff 1993 p.452).
The implication here is that as much as health is a necessity for human flourishing, the appropriate equity principle should be a just distribution of health or equality of health. Culyer and Wagstaff reject need-based interpretations of equity such as equal distribution according to ‘need’, arguing that “irrespective of how one interprets ‘need’, equality of health will not be attained if persons in equal need are treated the same and persons in unequal need are treated in proportion to the relevant inequalities”. They equally reject equality of access, as earlier noted, that it would not yield an equal distribution of health. The rejection of both interpretations does not mean that the concepts of ‘need’ and ‘access’ are not important in resource allocation decisions in health. The notion of ‘need’, they claimed, has a role to play by picking out which resources are to be distributed but it does not “indicate the appropriate distribution of these resources” (Culyer and Wagstaff 1993 p.452-453).
4.3 Measuring Equity The lack of agreement on the definition of equity is manifested in how it should be measured in practice. Generally, equity can be measured from the perspectives of health care delivery and/or financing by examining payments and allocation of resources. In the delivery of health care, most of the empirical work on equity has concentrated on differences in utilisation of health care by different socio-economic groups (Le Grand 1978; O’Donnell and Propper 1991; Wagstaff et al. 1991). The empirical literature on financing has focused significantly on households’ ability to pay for treatment (Wagstaff et al. 1992; 1998; van Doorslaer et al. 1993). There is limited empirical work
17 See Norman Daniel, Just Health care, 1985. Raanan Gillon, Philosophical Medical Ethics, 1985. 66 in relation to equity in allocation of resources, particularly in developing countries. This issue is fully discussed in section 2 of this chapter.
One of the pioneering empirical works on equity in the delivery of health care was undertaken in the United Kingdom (UK) by Julian Le Grand. He used a ‘range measure of inequality’ to analyse the ratio between public expenditure per person ill for the highest socio-economic group18 and that for the lowest (Le Grand 1978; 1991). A two- level analysis was performed. First, he estimated the costs of seeking care in the NHS per person reporting illness in each socio-economic. To achieve this, each socio- economic group’s total imputed expenditure was divided by the number of persons reporting either ‘limiting long-standing illness’ (chronic) or ‘acute illness’ in the group. Second, he computed the share of expenditure received by each socio-economic group and compared this with the group’s share of ill-health (Le Grand 1978). Based on his findings, he concluded that horizontal equity (equal treatment for equal need) has not been achieved in the NHS (Le Grand 1982 p. 46).
Le Grand’s approach has attracted criticisms from O’Donnel and Propper (1991) and Wagstaff et al. (1991). One of the main weaknesses often highlighted is Le Grand’s use of an aggregate morbidity measure as a proxy of need. (This issue is discussed further below). Collins and Klein (1980) criticised his implicit assumption that only sick people receive treatment.
Wagstaff et al. (1991) used a concentration curve to measure equity. Like the other approaches, this approach focuses on horizontal equity (equal utilisation for equal need). It attempts to quantify the degree of inequity in the distribution of health care by first, categorising and ranking individuals according to income and then, constructing and comparing an ‘illness concentration curve’ with an ‘expenditure concentration curve’. The authors measured utilisation by calculating the average number of services used over time or by imputed expenditure. Expenditure comprises the imputed resource costs of utilisation of primary health care facilities as well as hospital in and outpatient
18 Socio-economic group was defined by occupation. Occupation is sometimes used as a proxy composite index of a number of other variables, such as income, education, and social culture. 67 facilities. Self-reported illness was used as an indicator of health care needs (further details are provided below).
One of the weaknesses of this approach according to Le Grand is the assumption that all persons reporting acute sickness are in equal need of health care and those reporting chronic illness also have equal need of health care (Le Grand 1991 p. 241). Other methods of measuring equity in distribution of health care, including the regression method (Puffer 1986; Mathers 1994), have been proposed and used. The next section is devoted to examining the concept of ‘need’, which lies at the heart of both the interpretation and measurement of equity.
4.4 Defining Health Need The concept of need is controversial. Bradshaw’s Taxonomy of social need (Bradshaw 1972) which started the debate over the interpretation of need, identified four dimensions; normative, felt, expressed and comparative need. Normative need exists when an expert, professional administrator or scientist defines need by laying down their desired standard and comparing it with the standard that actually exists. Felt need is identified by equating need with want and is assessed by simply asking people if they feel they need a service. Expressed need is where the individual’s felt need is turned into action. Finally, comparative need exists where characteristics of a population that receive a service are ascertained, and where people with similar characteristics that do not receive services are judged to be in need (Bradshaw 1972). The fierce debate since Bradshaw’s work has done little to narrow disagreement between scholars over the appropriate interpretation of need for health care.
Two key interpretations of need: the ‘ill-health’ and the ‘capacity to benefit’ definitions warrant closer examination. In the medical and philosophical literature, need is often equated to ill-health (Donabedian 1973; Gillon 1985). This interpretation, which deals largely with the individual’s pre-treatment health, has much appeal among clinicians. The WHO definition of health needs is also geared towards this view. According to the WHO, “health needs are to be understood as the deficiencies in health or social well- being that call for preventive, curative, rehabilitative or public welfare measures” 68 (WHO 1974 p. 14). The interpretation of need in terms of ill-health generally suggests that people who are sicker than others have greater need for health care. This has been criticised on several grounds including the fact not all health problems are amenable to health care intervention and also that such an interpretation ignores resource constraints (Mooney 2000; Acheson 1978).
The capacity to benefit notion of need has greater currency among health economists (Culyer 1995; Crampton and Laugesen 1995; Culyer and Wagstaff 1993; Normand 1991). This interpretation sees health need as determined by the amount of health care resources required to exhaust an individual’s capacity to benefit from health care (Culyer 2001; 1978). Proponents have argued that a need for health care can exist only when there is a capacity to benefit (Culyer 1976; Williams 1974). That is, individual’s need for a particular health service is related to his/her potential to benefit from it. Within this framework, “health need is not an absolute but a measure for rationing scarce resources” (Eager et al. 2001). This sharply contrasts with the ill-health notion, which ignores limitations on resources. Culyer and Wagstaff (1993) observed that a person might be ill but not need health care particularly where effective treatment of that ailment may be non-existent. An individual may also have the capacity to benefit from health care but may not be ill as in the case for preventive interventions (McIntyre et al. 2000; Culyer 1995).
Not everybody endorses the ‘capacity to benefit’ interpretation without reservation. McIntyre et al. (2000) remarked that “while the notion of ‘capacity to benefit’ is conceptually appealing, it is difficult to operationalise in some instances, particularly where health data are poor”. Crampton and Laugesen (1995) observed that there is no point in devoting resources to health care if there is little chance that people will benefit, but they also acknowledged that there are problems defining need as ‘capacity to benefit’.
4.4.1 Measurement of Need Approaches for measuring health needs vary considerably. There is no consensus on any single method. Traditionally, the use of mortality data as a proxy for health needs has 69 enjoyed wide application (May and Bevan, 1987; Birch et al. 1993; Pain et al. 1996). Mortality measures are largely based on correlations between age and sex standardised mortality ratios and various aspects of morbidity and risk of ill-health, particularly those conditions associated with the need for secondary care (Newbold 1998; Forster 1977; Bennet and Holland 1977). The mortality-based approach to measuring health needs has several limitations. One major weakness, particularly when health need is defined as the ‘capacity to benefit’, is its limited focus. Mortality measures fail to reflect the full extent of non-fatal morbidity (Gibson et al. 2002) and also ignore issues such as poverty, overcrowding and other aspects of social deprivation with respect to health care needs (Black et al. 1982; Fox 1978).
In developing countries, particularly in sub-Saharan Africa, mortality-based indicators as a proxy for health need have additional weaknesses related to the significant under- reporting of death (Kaufman et al. 1997; Murray and Lopez 1997; Cooper et al. 1998). In Ghana, although infant and child mortality rates are routinely estimated from demographic and health surveys, many deaths are still believed to go unreported, particularly in the rural areas, making mortality data in the country less reliable. McIntyre et al. (2002) made similar observations about the under-reporting of deaths in South Africa. The general weakness of mortality-based indicators as a proxy measure of health need, especially in data-poor regions, has made them a less suitable measure of need.
Most of the recent empirical work on equity uses morbidity data as a proxy of health need. Le Grand (1978), as noted earlier used aggregate morbidity measures as an indicator of need. He defined need as “the number in each socio-economic group reporting acute and/or limiting long standing illness” (O’ Donnell and Propper 1991). He calculated the age and sex standardised percentage of persons reporting ill (either chronic and/or acute illness) by occupational group to represent the proportion of the population in need of health care. Collins and Klein (1980) used specific morbidity measures as proxies of need. They divided health need into different categories including the non-sick, the acutely sick, and the chronically sick, and compared resources received by each socio-economic group within each category of health need (see Wagstaff et al. 1991).
70 O’Donnell and Propper (1991) used disaggregated morbidity measures covering non- limiting long-standing illness; limiting long standing illness; persons reporting neither restricted activity nor long standing illness; and persons reporting ‘not good health’, as an indicator of need. Similarly, Wagstaff et al. (1991) measured need for health care based on persons reporting either acute sickness or limiting long-standing illness, long standing illness alone or persons reporting ‘not good’ health. All the morbidity measures mentioned above use self-reported ill-heath as an indicator of health care need. Apart from the approach of Le Grand, the others attempt to use disaggregated morbidity data. However, none completely manages to overcome the problem of collective measure of health need. Their heavy reliance on income and occupation to define socio-economic group underestimates the link between other non-economic variables and health. It is also important to note that the use of self-reported ill-health as an indicator of health care need leads to the same problem of under-reporting of illness, particularly in low and middle-income countries. There is evidence that the poor under-report illness in household surveys leading to incorrect conclusions that the poor suffer less ill-health (McIntyre et al. 2000)
4.4.2 Deprivation as a Proxy of Need The link between deprivation and health is well established (Jarman 1984; Townsend et al. 1988; McLoone and Boddy 1994; Carstairs 1995; Eachus et al. 1996; McIntyre et al. 2002). Consequently, deprivation has been used as a proxy measure of health care need in most studies. Conceptually, no single definition of deprivation is universally endorsed. Townsend (1987 p.125) defined deprivation as a “state of observable and demonstrable disadvantage relative to the local community or wider society or nation to which an individual, family or group belongs”.
Deprivation can be material, social or both. Material deprivation encompasses lack of or inadequate food, clothing, housing, sanitation, water, household assets, poor physical and/or mental health, and poor working environment among other things. Social deprivation, on the contrary, includes no or low education, few employment opportunities and the lack of right in employment, separated families, lack of recreation,
71 lack of integration into the community possibly as a result of racial or gender discrimination, and lack of participation in social institutions (Townsend 1989; McIntyre et al. 2000). In measuring deprivation, material and social deprivation do not usually get the same level of attention. Salmond and Crampton (1999) observed that most deprivation measures focus on measuring material deprivation, mainly because there are suitable variables available for examination in routine datasets. But, as noted by Morris and Carstairs (1991), social aspects of deprivation are just as influential on health status as material deprivation.
Attempts have been made to distinguish between deprivation and poverty (McIntyre et al. 2002). Such a distinction is becoming more difficult as the definition of poverty is gradually moving away from a narrowly defined monetary approach to incorporate a range of socio-economic, cultural and other indicators to measure deprivation. The World Bank, for example, acknowledged that poverty encompasses not only material deprivation, but also low achievements in education and health (WDR 2000/2001; 1990). UNDP similarly defines poverty broadly as “deprivation in three essential elements of human life - longevity, knowledge and decent standard of living” (UNDP 1997 p.18).
4.4.3 Measurement of Deprivation Despite the extensive research done on deprivation, there is no universally acceptable method of measuring it. This may be attributed to the context specific nature of the variables used in constructing deprivation indices. There are three main sets of deprivation indices widely used around the world. These are the underprivileged area (UPA) index put forward by Jarman (Jarman 1983; 1984); the Townsend’s index of material deprivation (Townsend et al. 1988); and the Scottish deprivation index developed by Carstairs and Morris (1989). Together, these three sets of deprivation measures have 16 variables, with unemployment and overcrowding common to all the three. Apart from the United Kingdom, deprivation indices have also been developed in Sweden (the Swedish UPA – Bajekal et al. 1996) and in the United States (Taylor 1998).
72
Deprivation can be measured either by using quantitative or qualitative approaches. In quantitative measures of deprivation, multivariate statistics are used to construct a composite, weighted index of deprivation (McIntyre et al. 2000). Qualitatively, deprivation can be measured by eliciting views from a range of stakeholders on who should be regarded as particularly disadvantaged and what sort of weight should be assigned to them in allocating resource (Mooney and Jan 1997). A mixed method approach, which combines statistical methods with stakeholder consultations, can also be employed to achieve the same goal. The numerous benefits associated with the use of a mix of quantitative and qualitative approaches to measuring deprivation often make it a more attractive option. (This is further discussed in Chapter 6 - Methodology).
4.5 Towards a Working Definition of Equity The various interpretations of equity examined so far and others in the health economics literature deal predominantly with horizontal equity. In recent years, authors such as Mooney, McIntyre, Gilson, Wiseman, Jan and others have highlighted the need for more emphasis on vertical equity in health and health resource distribution, particularly in countries where there are substantial differences in health status between different population groups (Mooney 2002; McIntyre et al. 2000). In the introduction of Mooney’s articles on vertical equity19, he observed that much has been done to stimulate debate about equity in health, but “the practical outcomes are disappointing; the poor in income remain poor in health, as do various ethnic groups and indigenous peoples” (Mooney 2000).
For equity to be served, the health needs of the most disadvantaged should be given priority over those of other people in the allocation of health care resources. This may sound like an act of ‘benevolence’ towards the needy. However, going by Robert Nozick’s principle of rectification of past injustice (Nozick 1974), this emphasis on differential allocation in favour of the worst-off could be taken as an act of redressing the past injustices in the allocation system.
19Mooney, G. (2000) Vertical Equity in Health care Resource Allocation. Health Care Analysis 8: 203-21 73 The analysis in this study is built on these pioneering works of Mooney, McIntyre and others that stress ‘vertical equity’ objectives for allocation of health care resources. The term vertical equity is defined in this study as the differential allocation of resources in favour of the most deprived jurisdictions. Two key issues need further clarification in this definition: how the terms “resources” and “deprived jurisdictions” are interpreted.
The term ‘resources’ or ‘health resources’ as used in the study refers predominantly to financial resources. Although this study notes the importance of other forms of health care resources, such as human and material, in the equity debate, it has deliberately adopted financial resources as the primary area of concentration to give the analysis a clearer and more concise focus. Equity in the allocation of human and material resources is drawn upon where necessary to explain the issue of equity in funding allocation.
The term ‘deprived jurisdiction’ is used in this study to refer to regions, districts, and sub-districts that are, on average, deprived than other areas. Since this study is largely about geographical allocation of resources, jurisdiction rather than population group is the focus. Health need in this study is measured in terms of relative deprivation to reflect the strong association between health status, socio-economic and area disadvantage in Ghana. Townsend’s definition of deprivation is adopted in this study (see Section 4.4.2). Thus, ‘deprived jurisdictions’ is defined in this study as the state of observable and demonstrable disadvantage relative to place of residence as well as demographic and socio-economic conditions of individuals, family or group (see Chapter 6 - Methodology). The next section reviews literature on resource allocation and equity.
74 SECTION 2: EQUITABLE ALLOCATION OF HEALTH CARE RESOURCES
Resource allocation, in the context of health care, is the “process of moving funds or resources from the funders of health care to the providers and consumers of health care” (Eager et al. 2001). Resource allocation decisions are made at different levels of the health system20. This study focuses predominantly on allocations made by governments at the national, regional and district levels of the health system.
Despite the disagreements over interpretation and measurement of equity, there is consensus that equitable access to health service will be difficult to achieve without some form of equity in resource allocation (Eaves 1998). The question, however, is: what is the most equitable and fair mechanism for allocating resources? In many countries, previous budget allocations, current service or facility patterns, capital development or political factors heavily influence resource allocation (Green et al. 2000). In the last half-century, health reforms in many countries have seen a shift from historically-based systems of resource allocation to a development of different funding mechanisms based on population needs. This section reviews some of these models for resource allocation in the health sector.
4.6 Historically-Based Funding Historical funding, also known as expenditure-based funding (Eager et al. 2001), is the oldest approach to funding health systems. Under this model, funds or resources are allocated on the basis of the previous year’s expenditure (Pearson 2002; Eager et al. 2001). It is the prerogative of the funder to provide a little more or a little less than the previous year’s budget. Maynard and Ludbrook encapsulated this practice in a slogan: “What you got last year, plus an allowance for growth, plus an allowance for scandals” (Maynard and Ludbrook 1980a p. 293). In the context of the British National Health
20 McKneally et al. (1997) described three levels of resource allocation: macroallocation, mesoallocation and microallocation. Macroallocation decisions are made by governments at the national, provincial and municipal level. Mesoallocations are made at the level of institutions, such as hospitals allocating their resources to programmes like cancer treatment, cardiology and dialysis. Microallocation of health care resources concerns the allocation of scarce lifesaving resources such as haemodialysis between competing claimants (see also Gillon 1985). 75 Service (NHS), May and Bevan (1986) noted that the increments that funders provide at their own discretion tend to go the “noisiest rather than the neediest”.
Historical funding has been criticised for failing to address equity and efficiency objectives (Green et al. 2000; Eager et al. 2001; Pearson 2002). Eager et al. (2001) noted that this funding system fails to match the needs of the local population and the available resources and services. The authors observed that although new population growth has taken place on the outskirts of most capital cities over the last several decades, the use of a historical funding model has ensured that funds are continually channelled to existing facilities (Eager et al. 2001 p. 73). Thus, the use of existing patterns of expenditure as the basis of current allocation ignores the degree of unmet need in more disadvantaged groups (Sheldon and Smith 2000). Another limitation of the historical funding model is that it provides incentives to health authorities to spend their allocations in order to assure that the following year’s allocation leading to inefficiency in funding use (Eager et al. 2001).
4.7 Population and Needs-based Allocation There is a small but growing international literature on the need to reorient resource allocation to make health systems more responsive to local health needs (Bourne et al. 1990; Gilbert et al. 1992; Birch and Chambers 1993; Mays 1995; Lake 2000). Needs- based models of resource allocation attempt to allocate funds on the basis of health needs of a population as opposed to historical patterns (Diderichsen et al. 1997). These funding models generally operate at the levels of the individual and the population at large. At the individual level, the funding agency provides funds to meet the identified needs of the individual consumer or patient21.
At the population level, needs-based models recognise that funding should be distributed according to size and needs of a defined population (Eager et al. 2001). Models of resource allocation based on population and needs create the opportunity for
21 There are several ways of doing this including the use of some form of ‘capitation’ funding, in which the funding agency translates the person’s particular health needs into a capitated funding allocation, which is passed on to the health care provider. The provider is expected to meet the person’s needs within the fixed capitated funding allocation (Eager et al. 2001 p. 74). 76 equalising distribution across regions and districts. They also have the advantage of devising a gradual, smooth transition process from historical allocation to a new regime where redistribution occurs over time, thereby giving the ‘losers’ in the redistribution process the opportunity to adjust and the ‘winners’ enough time to develop new planning capacity (Belli 2004). Population and needs-based allocation models have their origins in the work of the Resource Allocation Working Party (RAWP) in the United Kingdom. Since the Working Party’s report, many countries, predominantly in the industrialised world, have developed their own needs-based models for allocating funding in the health sector. Most of these formulae are designed to achieve equal access to health care for equal needs (Sax 1990).
4.7.1 The Resource Allocation Working Party (RAWP) The Resource Allocation Working Party is credited with developing needs-based allocation of health care resources. RAWP and the NHS led the world in pioneering scientific methods of equitable allocation of health resources (Carr-Hill et al. 1997; Shaw and Smith 2001). RAWP was constituted in the UK in 1975. The main reason for its constitution is not clearly established in the literature; however, a review by Mays and Bevan (1986) suggests that evidence of geographical inequalities in resource allocation was one of the key reasons. The creation of the NHS in 1948 was significantly influenced by the desire to achieve a more equitable distribution of Health and Community Health Service (HCHS) resources, which were allocated on historical basis and largely thought to be inequitable (Maynard and Ludbrook 1980a). However, some 30 years after its establishment, the NHS was still “groping its way forward” to the realisation of the equity dreams it was established to achieve, a situation that led to the continuing debate about the distribution of resources in the health system (Buxton and Klein 1978).
The RAWP was asked to recommend a system of allocating resources within the NHS which would be responsive to the health needs of the population, and to identify and correct inequalities in the existing pattern of resource distribution (Department Health and Social Services [DHSS] 1976). The equity criterion adopted by the Working Party
77 was equal opportunity of access to patients in equal needs, regardless of where they live (Paton 1985; Shaw and Smith 2001). In 1976 the Working Party reported its findings (Mays and Bevan 1986). It recommended that the distribution of NHS resources should be on the basis of population size weighted according to two fundamental criteria: difference in the need for health care and unavoidable geographical differences in the cost of providing for services (Carr-Hill et al. 1994 p.1046).
The RAWP Formula
One of the aims of the RAWP formula was to account for geographical variations in need for health care. The assessment of comparative health needs of different populations to determine their appropriate shares of resources, however, proved to be one of the major challenges faced by the Working Party. In the design of the formula, four key variables: population size, age and sex structure, morbidity, and cross- boundary flows were used. The formula also makes adjustment for service increment for teaching hospitals (SIFT).
The Working Party used population size in the calculation of allocation targets. This has attracted less attention compared to other variables in the formula (Mays and Bevan 1986), perhaps because the use of population figures in resource allocation is not uncommon. However, this is not without controversy as the choice of a basic population figure involves a choice of an appropriate base date, particularly in areas where population change is more rapid than usual (Buxton and Klein 1978). For revenue targets, the Working Party used the latest available mid-year population estimates. In practice, annual allocations were on mid-year estimates of population levels two years earlier. For capital targets, a five-year forward projection of population was recommended (DHSS 1976).
The Working Party acknowledged that, apart from population size, demographic characteristics also influence the need for health care (Mays 1986). Consequently, in developing a measure for relative need, the population of each area was weighted according to the age and sex structure. By weighting the population for age and sex structure, the RAWP took into account the variations in the use of health resources by 78 the different age and sex groups. The age and sex structure was accounted for in the form of national utilisation rates for each age and sex group.
In addition to the age and sex weightings, morbidity was also accounted for in the RAWP formula. One of the basic aims of RAWP was to ensure that resources are made available to health authorities to enable them meet the needs of the population resulting from illness. The difficulty was however, to find an appropriate measure of morbidity. In the absence of any adequate proxies of morbidity, the Working Party adopted standardised mortality ratios (SMR) as the best index of morbidity and a proxy of health needs (DHSS 1976). The SMR is a single index number which compares the mortality experience of a given region’s population to that of a reference (or base) population (Bedard et al. 1999). In 1986, the RAWP formula was revised to incorporate deprivation as an indicator of need in order to address some of the problems arising from the use of standardized mortality ratios (SMRs).
The RAWP formula was used from 1977 to 1990 (Diderichsen et al. 1997) and, according to earlier evaluations, it was successful in redistributing resources from metropolitan English regions to the poorer regions in the north (Holland 1986; Shaw and Smith 2001). Despite its success, the formula has attracted criticisms for a variety of reasons.
Criticisms of the RAWP Formula
Although each of the indicators in the RAWP formula attracted some criticisms, the use of SMR as a proxy for morbidity and an indicator for need has attracted the most extensive debate. Much of the criticisms were related to the poor correlation between morbidity variables and the SMRs. Forster (1977) examined the correlation between age/sex standardised mortality rates and morbidity rates from General Household Survey (GHS). He found that the rank order correlations between mortality and ‘acute
79 sickness’22 as well as between mortality and ‘bed sickness’ were not significant. However, there was a statistically significant correlation between mortality and chronic sickness. Forster concluded that it was doubtful if mortality could be considered a valid indicator of morbidity. Several other studies, including those by Snaith (1978) and Palmer (1978) followed Forster in casting doubt about the appropriateness of using SMR as a proxy measure for morbidity.
Apart from the poor correlation between mortality and morbidity, the SMR has also been attacked on the grounds that many illnesses are likely to create a high use of health services but would not result in many deaths and vice-versa (Goldacre 1981; Mays and Bevan 1986). Thus, in such instances, mortality statistics would not provide an adequate picture of morbidity and the need for health care. Despite the extensive criticisms of the use of SMR as a proxy of morbidity ratios, no effective measure of morbidity has been developed to date, leaving mortality still in use. Using SMR in their study on needs- based allocation of health resources in Canada, Eyles et al. (1991) argued that, in the context of their study, SMR appears to be the best available indicator of needs.
In general, the criticisms against the RAWP formula did not undermine its use outside the UK. In Australia, Canada, New Zealand and Sweden, the RAWP formula has influenced for resource allocation reform (Eyles et al. 1991; NSW Health Department 1993; Diderichsen et al. 1997; Cumming 2002). The impact of RAWP on resource allocation reforms in developing countries has been limited, however. The majority of countries in developing countries still allocate public resource across regions, districts and health plans on a historical basis, with some adjustments achieved through political negotiation (Belli 2004). South Africa and Zambia have population and needs-based allocation system modelled on the basis of the RAWP approach (McIntyre et al. 2000; Lake et al. 2002). The next section looks at the RAWP-inspired needs-based resource allocation formulae in Australia and South Africa.
22 Acute sickness was defined as restriction of normal activities because of illness/injury during a two- week period before interview. Bed sickness was defined as acute sickness requiring a stay in bed during a two-week period before interview. Chronic sickness was defined as a state of long-standing illness, disability or infirmity. 80 4.7.2 RAWP Outside the United Kingdom
Australia
The Australian state of New South Wales (NSW) is well known for its population and needs-based resource distribution formula modelled on the RAWP approach. New South Wales has a population of about 6.7 million, the most populous state in Australia (Australia Bureau of Statistics 2004). In terms of health care, the state is divided into eight Area Health Services (AHSs). Until 1989, funding allocation to these AHSs was historically based and widely considered to be unresponsive to changing population trends and health needs (NSW Department of Health 1988). Around 1980, an informal interest group began exploring issues of access and equity. This led to the establishment of a formal committee on resource allocation in 1987, and by 1989 the first NSW Health Resource Allocation Formula (RAF) was published (Eager et al. 2001). Like the RAWP model, the equity principle underpinning the RAF was equality of access to health services (NSW Department of Health 1993).
The RAF was revised in 1993 following the creation of 23 District Health Services to replace the existing six rural regions. Subsequent to the election of the NSW Labour Party into office in 1995, the name of the formula was changed to Resource Distribution Formula (RDF) to drive home the intention of the new government to specifically shift resources across the state in line with the formula estimates. The redistribution was designed to proceed gradually through the use of the RDF as occurred within the NHS with the use of the RAWP formula. However, unlike the RAF, which was designed to provide long-term targets, the enhanced RDF was used primarily to determine the annual budgets of Areas (NSW DOH 1996). This required the inclusion of as many expenditure areas as possible in the pool of funds distributed by the formula.
Essential Elements of the RDF
The RDF has undergone several revisions but remains a complex formula both conceptually and statistically. This section focuses on the key components of the formula. The NSW RDF has six major components:
81
1. Population Health
2. Non-inpatient Services comprising:
Dental Services
Primary and Community Based Services
Outpatients
Emergency Department Services
3. Acute Inpatient Services
4. Mental Health Services
5. Rehabilitation and Extended Care
6. Teaching and Research
The share of expenditure of each of these components is calculated by weighting the population by indicators of health service utilisation and health needs. The key indicators used include age and sex structure; a general needs index developed from an analysis of how standardised mortality ratio, rural-urban variations and socio-economic status explain the use of health services. A separate index of need was developed from mental health and rehabilitation and extended care. Other key indicators include a factor for utilisation of the private health services that can be substituted for public health services by the local population; a factor for a range of other unavoidable cost components such as additional transport costs in rural areas; and a factor for Aboriginality and homelessness23 (NSW DOH 1996).
Impact of the RDF
The RDF is generally accepted as a model that attempts a fair distribution of funds available to the health care system. Prior to its adoption, health services in NSW remained in historical settings reflecting population distribution of the past (Eager et al.
23 There has been recent revision of the RDF in line with changes in the health sector (see New South Wales Department of Health 2005). 82 2001). Given the changes in population dynamics, there was the need for closure and modification of some services in order to fund the development of others. The government, however, faces opposition from entrenched lobby groups determined to preserve the status quo. The RDF made it possible to quantify the maldistribution and was used by health authorities to argue for a redistribution of resources in a rational and non-politicised manner.
There are reports of progress in reducing funding disparities across NSW following the introduction of the RDF. Gibbs et al. (2002) noted that in 1989/90, approximately 16.4 percent of the health budget needed to be reallocated to achieve equity in funding. However, by 1994/95, this figure was reduced to 9.6 percent and was further down to 4.4 percent by 1998/99. This reduction is attributed to the use of the RDF. The authors further argued that while all Area Health Services in NSW have received growth in funding, a greater share has been allocated towards historically under-funded population growth AHSs such as those in Greater Western Sydney, the Central Coast, and the North Coast of NSW (Gibbs et al. 2002 p. 43).
South Africa
South Africa is one of the few countries in Africa where the RAWP model has had some influence on reform of the resource allocation policy. South Africa has a total population of about 44.8 million; nearly 75% of which is black. Public spending on health is around 3.6 percent of Gross Domestic Product (GDP). The health expenditure per capita of about US$625 is significantly high compared to those of other countries in Africa. Ghana for instance, spends about US$60 per capita on health. Life expectancy at birth is currently estimated at about 49 years, largely due to the high incidence of HIV/AIDS (Human Development Report 2004).
Although a middle-income country, South Africa had one of the most inequitable health systems in terms of resource allocation, particularly before 1994 (McIntyre et al. 2000). Health policies pursued by the apartheid government promoted racially-based access to
83 resources which largely worked against the majority black population. This resulted in substantial inequities in health status between race and socio-economic groups as well as within and between the country’s nine provinces (McIntyre and Gilson 2002). There was a strong political commitment in the immediate post-apartheid period in favour of a radical shift of resources away from the relatively wealthy provinces such as Gauteng to the poorer provinces such as Northern Province (Pearson 2002).
When the African National Congress (ANC) was elected into office in 1994, the health budget was determined by a Health Function Committee (McIntyre and Gilson 2002). The Function Committee developed a population-based formula with the goal of achieving weighted per capita equality in provincial health budgets within five years (McIntyre and Gilson 2002). The South African Health Function Committee Formula (SAHFCF) sought to promote equity through geographical resource redistribution to reflect differential and relative need. In 1995/96, the population in each province was weighted for the provincial average per capita income relative to the national average. Thus, resources were differentially allocated in favour of ‘poorer’ provinces with potential high burden of ill-health (McIntyre and Gilson 2002 p. 1645). Unlike the RAWP formula where standardised mortality ratios were used as proxies of morbidity and indicator of need, the SAHFCF only uses a crude approach to estimating health need, but as McIntyre and Gilson noted, it was the best possible given the dearth of the data available.
Impact of the SAHFCF
Application of the SAHFCF in 1995/96 saw a dramatic shift of resources across provinces. The Western Cape province was the greatest ‘loser’ experiencing a 19% real budget cut. The Northern Province, on the contrary, was the biggest ‘winner’ receiving the largest real budgetary increment of 17% (Doherty and van den Heever 1997). Despite the broad support for geographical redistribution of resources, the sudden drastic changes in allocation in line with the five-year target of redistributed set by the committee became a major concern. The South African Department of Health promoted the need for a more gradual redistribution to allow provinces adequate time to adjust to
84 the changes. The five-year target for achieving inter-provincial equality in weighted per capita spending was refined, with the Minister of Health acknowledging it would take closer to ten years for the redistribution to be achieved.
4.8 Beyond Needs-Based Allocation Recent debate in the literature about resource allocation suggests that needs-based formulae do not promote equity. This stems largely from the persistent inequities in health among different jurisdictions and population groups even in countries where needs-based models are in use.
Mooney and Houston (2004) have, more recently, argued for an alternative approach to resource allocation which embraces the concept of ‘capacity to benefit’ and MESH infrastructure. MESH stands for Management, Economic, Social and Human infrastructure. According to Mooney and Houston (2004), the capacity to benefit from resource allocation entails good management, requires availability of resources, needs a socially well functioning community, and, ideally, good human resources. Where some or all of these elements are missing, the resources might be wasted, or at best, used to lesser effects.
Mooney and Houston (2004) criticised the philosophy underpinning RAWP-type approaches, arguing that the RAWP-type of resource allocation formulae emphasise the size of the problem (greater need) rather than the ‘capacity to benefit’ from the way resources are allocated. To them, because different jurisdictions may have different capacity to manage resources well, if MESH is not included in the resource allocation formula, jurisdictions with low capacity might be inadequately resourced (Mooney and Houston 2004 p.30).
Needs-based formulae are not by themselves sufficient to address equity. Gibbs et al. (2002) noted the New South Wales RDF is only one policy for addressing the equity issue, and by itself is an insufficient mechanism. Gibbs et al. (2002) admitted that while the RDF aims to create broad resource capacity for equity to be achieved within the
85 health system, an essential ingredient in delivering on equity objectives is action at the local level within the health system. In a nutshell, the fixation on formula or “formula fever” as Trevor Sheldon (1997) put it, has distracted attention from the important issue of how allocated resources are spent. Sheldon’s suggestion that health authorities and general practitioners should focus their attention on whether current spending patterns reinforce socially produced inequalities and, if so, do something about it from the local level, needs a second look.
Chapter 4. Summary of key points
This chapter has been divided into two sections. Section one has covered the concept of equity and provided detailed review of the debate over interpretation of equity. Key issues considered were the differentiation of equity from equality and the discussion of vertical and horizontal forms of equity.
Horizontal equity refers to equal treatment for equal need while vertical equity is defined as unequal (but equitable) treatment of unequals.
A working definition of equity (i.e., equity as defined in this study) has been developed and discussed. Thus, equity is defined in this study as differential allocation of resources in favour of the most disadvantaged jurisdictions.
Section two has covered mainly the issue of equitable resource allocation.
The analysis here has focused largely on two funding approaches: the historical funding and population and needs-based funding models.
The RAWP formula and its influence on resource allocation reforms was reviewed. Resource allocations in South Africa and New South Wales state of Australia have been greatly influenced by the RWAP approach. In both countries, some re-distribution of resources across regions was achieved using their RAWP-inspired resource allocation models.
Recent debate in the literature about resource allocation suggests that the RAWP-type formulae, which emphasise the distribution of resources on the basis of need, do not adequately promote equity on their own.
86 CHAPTER 5
DECENTRALISATION AND EQUITY IN RESOURCE ALLOCATION
Overview There are conflicting claims as to whether health sector decentralisation as a policy promotes or jeopardises equity. This chapter reviews evidence of the impact of decentralisation on equity in resource allocation. It starts with a definition of decentralisation and a brief analysis of various models used to describe the process. It traces how decentralisation entered the health policy arena and the role played by international organisations such as the World Bank, International Monetary Fund and the World Health Organization (WHO). This is followed by a literature review of the impact of decentralisation on equity and resource allocation.
5.1 Definition of Decentralisation Decentralisation is a broad and somewhat ambiguous term. Its conceptual framework spans many disciplines (Prud’homme 1995), making a single universal definition difficult. Smith (1985 p.1) defined it as “the sub-division of a state’s territory into smaller areas and the creation of the political and administrative institutions in those areas”. Levaggi and Smith (2003 p.3) defined it as “the transfer of power from a central authority (typically national government) to more local institutions”. In general, decentralisation refers to the shifting of power or decision-making responsibilities from national to sub-national levels. It is a dynamic process, which describes the relationship between the central government and its peripheral institutions. Among the basic types of decentralisation often isolated for analytical purposes are geographic/spatial, political and administrative and functional decentralisation.
87 Geographic decentralisation, according to Mills et al. (1990), refers to the transfer of broad responsibilities for public functions to local organisations that have well defined geographical boundaries. Prud’homme (1995 p.2) defined it as “a process of diffusing urban population and activities geographically away from large agglomerations”. Political decentralisation describes the provision of additional public decision-making powers through a democratic process to citizens or their elected representatives (World Bank 2000). Imman and Rubinfeld (1997) adopted a more populist approach by defining it as the extent to which political institutions map the multiplicity of citizen interests onto policy decisions. Thus, political decentralisation provides a platform to local populations to make their voice heard in decision-making processes.
Administrative decentralisation is defined broadly as the “transfer of responsibility for planning, management, and the raising and allocation of resources from the central government and its agencies to field units of government agencies, subordinate units or levels of government, semi-autonomous public authorities or corporations, area-wide regional or functional authorities, or non-governmental private or voluntary organisations” (Rondinelli and Nellis 1986 p. 5). In functional decentralisation24 the responsibility and power initially under a central authority is increasingly shared or taken over by one or more sub-units with the capacity to effectively manage the agreed- upon services (PAHO 1999).
In the context of health, decentralisation describes a range of reforms typified by the transfer of fiscal, administrative and/or political authority for planning, management, or service delivery from the central Ministry of Health (MOH) to alternative institutions (Bossert et al. 2000 p.1). The sub-national institution to which authority is transferred may be the regional or local office of the same ministry, provincial or municipal governments, autonomous public service agencies, or private sector organisations. Health sector decentralisation is pursued for a variety of reasons, including improving efficiency, equity, accessibility, quality and accountability to local populations (Tang and Bloom 2000; Collins 2000; Bossert and Beauvais 2002).
24 Many studies do not differentiate between functional and administrative decentralisation, as the two are almost the same. 88 5.2 Models of Decentralisation Rondinelli et al. (1983) provided the most-cited framework for analysing decentralisation. They distinguished between four main models: deconcentration, devolution, delegation and privatisation. These typologies reflect the scope, as well as the different approaches, to decentralisation. Mills et al. (1990) observed that the distinction relates essentially to the legal context of decentralisation. However, there are many other factors that may influence the degree of autonomy enjoyed by local bodies beside the influence of the legal framework for decentralisation. The four models of decentralisation can be applied independently or in different combinations at a particular time. Rondinelli et al. (1983) observed that while some governments have applied any of the four models at the different times, others have used various combinations of the four at the same time. Table 5.1 provides the four models of decentralisation and a brief definition of each.
Table 5. 1 Types and Definition of Decentralisation Model Definition
Devolution The handing over power to sub-national levels of government (local government) that are substantially autonomous of the national government.
Delegation The transfer of managerial responsibility for specifically defined functions to autonomous (or semi-autonomous) parastatal agency outside the regular bureaucratic structure. The agency is only indirectly controlled by the central government.
Deconcentration The transfer of specific administrative functions to lower levels of central government ministries or agencies.
Privatisation The transfer of public assets and/or responsibilities to private actors, with variable degree of government regulation. The private actors may be voluntary organisations or private for-profit and not-for- profit enterprises. Source: Based on the framework of Rondinelli (1983) and Mills (1994)
5.2.1 Devolution Devolution is the most extensive model of decentralisation and involves the transfer of authority for decision-making, finance and management from the central government to 89 quasi-autonomous units of local government (Table 5.1; see also Litvack et al. 2002). The local governments normally have clearly defined geographical boundaries over which they exercise their authority and in which they perform public functions. They also have the legal backing and statutory authority to raise revenue and make independent investment decisions. In a devolved system, the local government is not merely a subordinate administrative unit (as in the case of deconcentration), but enjoys a reciprocal and mutually beneficial relationship with the central government. The central government exercises only indirect, supervisory control over a devolved local government.
Devolution is often perceived to be the most independent and flexible model of decentralisation, especially in terms of control over personnel and budget. Despite the extensive degree of autonomy it confers, Mills et al. (1990) observed that local governments are rarely completely autonomous. In developing countries, in particular, local government structures have been substantially limited. Most local governments, therefore, have weak revenue mobilisation capacity. In Ghana, for example, the District Assemblies rely heavily on the District Assemblies Common Fund, centrally allocated by Ministry of Finance and the Ceded Revenue centrally allocated by the Internal Revenue Service (Appiah et al. 2000). Dependency of local governments on central government grants is not limited to developing countries. In Finland, for instance, even though health services are provided by the municipalities, the central government supports the municipalities with block grants and also earmarks funds to cover costs of drugs and investment projects (Koivusalo 1999).
Devolution as a model of health sector decentralisation has been implemented in countries such as the Philippines, Uganda, and Norway. In these countries, the local governments or municipalities have been primarily responsible for health services delivery (Mills et al. 1990; Bossert et al. 2000). While devolution to lower levels of government can result in democratic control over services, Segall (2003) argued that it might lead to fragmentation of the health system. It runs the risk of non-compliance with national health policies, for example, prioritising primary health care over hospital development. It can also create greater inequities in the health service as richer areas could use their economic strength to absorb more of the country’s health personnel and other limited health resources (Segall 2003 p.18-19). 90 5.2.2 Delegation Delegation involves the transfer of specific responsibilities to semi-autonomous agencies that are not wholly controlled by the central government (Table 5.1). In some countries, this model of decentralisation has been implemented to circumvent the inefficient state bureaucracies and provide more management flexibility since delegated entities are less fettered by public service regulations. In others, it has been used as a means of maintaining public control over highly profitable or valuable resources (Segall 2003, Mills et al. 1990, Rondinelli et al. 1983).
The delegated organisation may, for example, be granted greater control over recruitment of staff and use of resources. However, the ultimate responsibility remains with the sovereign authority or the central government. In general, the relationship between the central government and the delegated organisation can be characterised as a principal-agent relationship, where the central government is the principal and the delegated organisation, the agent. The agent has to be induced with incentives to align itself with the expectations of the principal (Litvack et al. 2002).
The health systems of Ghana and Zambia currently operate a delegated model of decentralisation. In Ghana, health service delivery has been delegated by the Ministry of Health to the Ghana Health Service; a semi-autonomous agency (see Chapter 2). Likewise, the Zambian Ministry of Health has also delegated operational authority to the Central Board of Health (MOH 2003; Bossert and Beauvais 2002). In both countries, the central MOH exercises regulatory authority and policy oversight. Delegation to lower level quasi-independent governmental agencies such as hospitals structured as public firms can be found across northern and southern Europe (Busse et al. 2002; Bankauskaite et al. 2004).
There are a number of disadvantages associated with the adoption of delegation as a model of decentralisation. First, the central government may find it difficult to formulate effective contracts with the delegated entities and efficiently monitor them. Second, the transaction costs involved in the contractual relationship between the ‘principal’ and the ‘agent’ can be high. Finally, there is the danger that the delegated entities may act in their own interests despite signing contracts, which enjoins them to perform specific functions (Segall 2000; 2003). 91 5.2.3 Deconcentration Deconcentration is the most common and limited form of decentralisation (Bossert et al. 2000). It occurs when the central government disperses responsibilities for certain services to its sub-national branch offices (Table 5.1; see also Litvack et al. 2002). Deconcentration is usually implemented within the public service and does not require extensive structural changes (Segall 2003). The lower level agency or office, which has no independent power, could be located at the regional, provincial, state and/or district level.
A key feature of deconcentration is that it excludes the transfer of political powers alongside the administrative authority. Consequently, it is perceived as the most restricted kind of decentralisation. Prud’homme (1995) saw it as a simple redistribution of decision-making responsibilities among different levels within the central government. However, deconcentration can be more than a mere redistribution of decision-making responsibilities. Rondinelli et al. (1983 p.15) noted that it provides local authorities with some discretion to plan and implement programmes and projects or to adjust central directives to suit local conditions, within guidelines set by the central ministry or agency headquarters.
In the context of health, deconcentration describes a situation where the central Ministry of Health establishes local management teams with clearly defined administrative duties and provides them with a degree of discretion to enable them to manage without frequent interference from the headquarters (Mills et al. 1990). However, the local management team still depends on the headquarters for financial allocation and spending guidelines. This is often seen as limiting the motivation of the local agents to develop innovative programmes and maintaining viable ones (PAHO 1999).
Deconcentration has been widely applied in developing countries including Nepal, Indonesia and Cameroon (Smith 1997). In many East Asian and Eastern European countries, deconcentration was the main type of decentralisation implemented, since, as unitary states, there were no independent local governments legally accountable to local constituents until recently (Kornai 1992; Litvack 2002). The current model of decentralisation within the Ghana Health Service (GHS), where planning and management responsibilities have been decentralised from the GHS headquarters to 92 regional and district health administrations provides a typical example of deconcentration. The Caribbean islands of Montserrat and Anguilla are also examples of countries that have adopted deconcentration in health sector (PAHO 1999). The major advantages of this model of decentralisation are the greater flexibility and the discretion over resource use it provides management to respond to local conditions. In all these countries, political power remains in the central MOH.
5.2.4 Privatisation Privatisation involves the transfer of public assets and/or responsibilities to private actors, with a variable degree of government regulation. The private actors may be voluntary organisations, private for-profit and not-for-profit enterprises (Rondinelli et al. 1983; Mills et al. 1990; Bankauskaite et al. 2004). In some cases, government may also transfer responsibility to ‘parallel organisations’ such as professional groups and religious bodies. Such organisations may be given the responsibility to register and supervise their members to perform functions that were previously performed or regulated by the government (Rondinelli et al. 1983).
Privatisation has been widely used in health systems in both developed and developing countries, for efficiency and quality reasons. In some cases, governments may transfer responsibility for health service delivery, which they may be unable to mobilise funding to provide, to private actors who are considered capable of tapping additional resources. In many European countries, including Sweden and Germany, hospitals are predominantly private not-for-profit entities managed by local boards of trustees. These hospitals are supported financially by the state, which also exercises supervisory control through a wide range of statutory and procedural measures (Saltman et al. 2004; Busse and Schlette 2003).
In developing countries, governments depend heavily on private (voluntary) organisations, usually faith-based, for health service provision. In sub-Saharan Africa, in particular, there has been a noticeable trend to support private providers, many of whom are faith-based, to compensate for shortfalls in government provision of services (Sikosana et al. 1997). In Zambia, Malawi and Uganda, church-run hospitals and
93 private rural clinics receive direct government subsidies for both operations and personnel training (Sahn and Bernier 1995). In Ghana, the MOH has contractual agreements with religious health bodies such as the Christian Health Association of Ghana (CHAG) which co-ordinates mission-based health providers, for the regulation and supervision of its members. The MOH, in return, provides direct funding and other support to the CHAG (MOH 2003). Section 5.3 traces how decentralisation entered the health policy arena and the role that international organisations have played in health policy in developing countries.
5.3 Decentralisation and Health Policy in Developing Countries Decentralisation has featured prominently in the health policies of developed and developing countries. However, the context in which it entered the health policy arena differs in the two settings. In developed countries, decentralisation is largely seen as imposed on central authorities by powerful local government institutions, some of which have existed for decades. In countries such as Sweden and Canada, local government institutions have long been responsible for health delivery. Armstrong and Armstrong (1999) reported that health care in Canada has been primarily the responsibility of the provinces, since Canada became a nation in 1867. Similarly, Diderichsen (1999) noted that local government system in Sweden predates the modern welfare society: since 1860, elected representatives in municipalities and county councils have raised taxes and run public services including hospitals.
In developing countries, decentralisation is a relatively new phenomenon. It has been largely seen as initiated by the central authority rather than demanded by the local government (Mills et al. 1990). One of the cited reasons for decentralisation is the desire to achieve equitable distribution of benefits of economic growth (Zwi and Mills 1995, Mills et al. 1990). This review focuses on how decentralisation crept into the health policy of developing countries, particularly in sub-Saharan Africa.
Health policy development in many developing countries is inextricably linked to their political history. Colonial rule bequeathed centralised administrative institutions in most countries in Africa, Latin America and Asia (Rondinelli et al. 1983). Most of the 94 government departments, including health, were unresponsive to the needs of the wider population. In some settings, such as in South Africa, the centralised service provision was highly skewed towards minority segments of the population (Zwi and Mills 1995, McIntyre et al. 2000). In the late 1960s and 1970s, following independence in many countries, a shift from centralisation to decentralised planning and management was seen as crucial to broader development planning and integration of the wider population into developmental efforts (Conyers 1981, Gilson and Mills 1995). The goal of development policies in most countries was to distribute the benefits of economic growth more equitably, raise the productivity and income of all segments of the population and raise the living standard of the poor (Rondinelli et al. 1983).
The Alma Ata declaration on Primary Health Care in 1978, which expanded the health policy arena to include many groups other than medical professionals, provided additional impetus for decentralisation (Walt and Gilson 1994). Decentralisation was seen as an important ingredient in the success of the primary health care policies (Ebrahim and Ranken 1988). Neo-liberal policy advocates took advantage of the expanded health policy environment and began to dominate the health policy of developing countries through loan conditionalities. Bilateral and multilateral donors made loan agreements conditional upon a reduced public sector role, the introduction of user fees and relaxation of private sector regulations among other things (Gilson and Mills 1995, Mills et al. 1990). The World Bank, International Monetary Fund (IMF) and WHO contributed in various ways to pushing decentralisation to the top of the health policy agenda. Section 5.4 briefly reviews how decentralisation became a core health policy issue in developing countries through the policies of international donor agencies.
5.4 Decentralisation and Donor Agencies The health policy agenda of developing countries is believed to be set by bilateral and multilateral donor agencies (Buse 1994; Okuonzi and Macrae 1995). Following the adoption of the Alma-Ata declaration on Primary Health Care (PHC), the WHO advocated for some degree of decentralisation through its emphasis on strengthening district health systems (WHO 1988). However, for several reasons, the PHC initiative 95 could not gather adequate momentum to achieve its objectives. One reason was that the WHO under-estimated the cost involved for countries to re-organise their health systems in order to implement the ‘comprehensive’ PHC25 it was advocating for (Zwi and Mills 1995). The macro-economic atmosphere in the years following Alma-Ata was not conducive for effective implementation of the PHC concept.
After years of taking an ideological position to deliver free health care to their population following independence, many developing countries experienced stringent budgetary constraints in the 1980s that made it impossible to continue with free medical care (Vogel 1988). Declining commodity prices and collapsing fiscal regimes took their toll on public investment and severely affected the health sector (Standing 2002). The IMF and World Bank capitalised on the need for external funding in many developing countries to give loans on condition that structural adjustment programmes (SAP) including privatisation (a model of decentralisation) be implemented. The SAP worsened health conditions in many countries, particularly in sub-Saharan Africa (Sahn and Bernier 1995) and inadvertently increased the need for more donor funding.
Seeing the lack of leadership from the WHO in defining global health policy, the World Bank manoeuvred its way into the health policy arena and positioned itself as the major source of health funding for developing countries. The bank used its 1993 annual report: Investing in Health (World Bank 1993) to articulate strongly its policy on global health and outline a comprehensive package aimed at strengthening health systems of many countries (Zwi and Mills 1995, Ugalde and Jackson 1995). As the leading player, the World Bank, in association with several bilateral agencies, made health sector funding conditional upon implementing a range of further reforms. For example, to improve quality, governments seeking donor funding were pushed to develop a more selective approach to PHC by introducing packages of low-cost interventions as a guaranteed minimum “basic package” of services. They were also asked to develop stronger contractual arrangements with private sector providers in both non-profit and for-profit sectors (World Bank 1993; Gilson and Mills 1995; Standing 1997).
25 Comprehensive primary health care focuses on a wider range of approaches to promoting health. It emphasizes the importance of multi-sectoral and development oriented activities as well as providing a broader range of health services at the peripheral levels, instead of focusing only on a narrowly defined set of highly cost-effective interventions (Rifkin and Walt 1986; Zwi and Mills 1995).
96 Against the backdrop of the rigid bureaucracy, inflexibility and unresponsiveness of the existing centralised administrative systems, decentralisation of health services to lower tiers of government or to other agencies was promoted by the donor community as a means of increasing accountability to local populations (Standing 2002, Sikosana et al. 1997). Over 25 African countries were in the process of implementing some sort of decentralisation policy in the early 1990s, often with donor support (Conn et al. 1996; Adamolekun 1991). Despite donor enthusiasm for decentralisation, there is little evidence of a positive impact on health systems performance as well as limited understanding of the factors that influence its development within the health sector. Before reviewing the evidence of the impact of decentralisation on equity, a brief review of the claims and counter-claims that decentralisation improves or jeopardises equity is warranted.
5.5 Conflicting Claims about Decentralisation and Equity Decentralisation has been promoted as a “multi-purpose” policy capable of promoting efficiency, equity, better responsiveness to local needs and greater accountability (Mills et al 1990, Litvack et al. 1998). There are conflicting claims, however, on whether decentralisation promotes or jeopardises equity. On the one hand, it has been argued that decentralisation will promote “greater equity through distribution of resources towards traditionally marginalised regions and population groups” (Bossert and Beauvais 2002 p.14). This claim is based on the assumption that local institutions are better informed about the needs of local populations, and hence, are well positioned to address them effectively than the central government (Silverman 1990; Levaggi and Smith 2004).
On the other hand, some analysts contend that decentralisation will jeopardise equity or fuel inequities (Kleinman et al. 2002). This position is based on the assumption that richer decentralised areas have the capacity to purchase more and better quality services than their poorer counterparts, thereby triggering inter-regional inequities in access to and quality of services (Prud’homme 1994; Sikosana et al. 1997). Such analysts argue that centralised systems are more likely to redistribute resources in favour of the poorer population groups, as there is no “country with important redistribution carried out at 97 sub-national levels” (Prud’homme 1995 p.15). Another argument against decentralisation as far as equity is concerned is that local elite groups are just as likely to pursue their own interests as central level officials and politicians, thereby enabling intra-regional inequities to flourish (Collins 1989)26. The next section reviews evidence of the impact of decentralisation on equity.
5.6 Evidence of Impact of Decentralisation on Equity Empirical evidence of how decentralisation impacts on equity in the health sector is very sketchy. Most of the studies that have attempted to look at the relationship between decentralisation and equity provide no conclusive evidence of whether the policy has improved or exacerbated inequities. Often a mixed picture of positive and negative effects is present as discussed here.
In a study of decentralisation and equity of resource allocation in Colombia and Chile, Bossert et al. (2003) found that decentralisation significantly improved equity of resource allocation (per capita) across municipalities in Colombia. They compared equity in allocation of resources before decentralisation and after decentralisation and found that the pattern of resource allocation at the national level in Colombia before decentralisation was highly skewed in favour of wealthier municipalities. However, the situation improved remarkably following decentralisation and the introduction of a population-based resource allocation formula27. In Chile, there were no data prior to decentralisation for comparative analysis, but the authors found that resource allocation remained largely equitable during the period of decentralisation for which data was available (from 1991-1996).
Tang and Bloom (2000) undertook a case study of the Donglan County in rural China, where finance and management of basic health services has been devolved to townships, the lowest level of government. They found that decentralisation “led to
26 Commentators on both sides of the debate caution that centralisation and decentralisation are necessary but not in themselves sufficient for redistribution of resources. 27 This finding questions the assertion that centralised systems are more likely to redistribute resources in favour of poorer areas. The results in Colombia clearly indicate that centralisation does not automatically guarantee distribution in favour of the poor.
98 neither an increase in local government health finance, nor improvements in equity, efficiency and effectiveness” (Tang and Bloom 2000 p.198). Contrary to expectations that devolution would encourage township governments to increase funding for their health centres, they (the township governments) failed to allocate more funds to local health services (Tang and Bloom 2000). In addition, the authors reported an unequal distribution of qualified health personnel among the townships of the Donglan County28. They noted that decentralisation made it impossible for the County Health Bureau to transfer personnel between facilities, if there was any historical imbalance of staff deployment.
In Brazil, Collins et al. (2000) found that federal transfer of resources to fund the Unified Health System (Sistema Único de Saúde - SUS) was based on the location of hospital and outpatient institutions, which historically tended to be in richer and more urbanised areas. The SUS funding “tended to reinforce the unequal allocation of resources in the country” (Collins et al. 2000). They found that municipalities (municipios) in the richer state of Sao Paolo receive, on average, more than twice as much as the municipalities of the poorer state of Acre (Collins et al. 2000 p.123). In this regard, decentralisation in Brazil did not improve equity. However, the study also found that the PAB (Piso Assistencial Básic) - the periodic funding allocation from the national level to the Municipal Health Funds of those municipalities registered under the ‘basic operating rule’ (BOR) was more equitable. Under the PAB allocation, poorer municipalities saw an important increase in the amount received for health care.
In the Western Highlands Province (WHP) of Papua New Guinea (PNG), Campos- Outcalt et al. (1995) found that decentralisation of health service administration to district levels led to a decline in services to the extent that the provincial government decided to recentralise responsibility for health services to the provincial Division of Health in 1993. The authors found, in particular, that the decentralisation led to an unequal distribution of personnel29, which benefited urban districts at the expense of remote districts. They noted that “districts in more desirable locations were able to hire
28 Information presented by the authors indicates that some health centres have enough laboratory technicians but too few doctors, and others have enough physicians but no laboratory technicians. 29 Under the district system (i.e., the decentralised system), personnel decisions, which were previously made with a provincial-wide outlook were made by administrators who were concerned only about their own districts 99 and retain staff while those in remote areas were having much more difficulty and could no longer depend on depend on assistance from other districts for short term emergency needs” (Campos-Outcalt et al. 1995 p.1097).
Bossert and Beauvais (2002) undertook a comparative analysis of decision space in the decentralisation of health systems in Ghana, Zambia, Uganda and the Philippines. In Ghana, they found that wealthier regions allocate greater percentage of their resources to sub-districts than poorer regions30. No concrete evidence of decentralisation’s impact on equity was provided for Zambia due to lack of data. In Uganda, they noted that ‘own-resource’ revenue for sub-counties differs dramatically from US$4,000 to US$200,000 per year; a situation which, according to the authors, could affect horizontal equity in non-transfer resources available for the health sector.
In the Philippines, the authors reported that the benefits and costs of decentralisation had not fallen equally on all local government units (LGUs) or on all levels of government. They cited Loehr and Manasan (1991), who showed that the ‘barangays’ (villages) and cities had been fiscal net winners (of decentralisation) and provinces and municipalities net losers. While the provinces and municipalities received 57 percent of revenue transfers, they bore 92.5 percent of the costs of devolution. The cities and the barangays received 47 percent of the transfers and bore only 7.5 percent of the costs. The results generally indicated that decentralisation has not promoted equity in Ghana, Zambia, Uganda and the Philippines.
In “Learning from failed health reform in Uganda”, Okuonzi (2004) observed [based on other evidence] that “decentralisation has actually widened disparities in the nature and quality of health services” (p.1173). He noted, for example, that the availability of emergency obstetric services in Uganda varied from 4 percent to 42 percent, and this was because richer districts and those with powerful local politicians have been able to persuade non-governmental organisations to work in their districts, thereby improving service delivery in those districts.
In a study of decentralisation and equity of healthcare provision in Finland, Koivusalo observed that concerns about growing inequities between areas (in Finland) has grown
30 This and other findings of the study in respect to equity were synthesised from findings of other studies. 100 in the 1990s, but there was so far “no systematic evidence that major differentials have arisen between areas in terms of service provision and access” (Koivusalo 1999 p.1199). The author noted that, within some municipalities, user charges are used to recover cost and such cost recovery has made it harder for lower income groups to cover their hospital costs. However, in general, the Finnish population have been satisfied and strongly supported the nature and basis of health service delivery and of financing without resort to the private sector. Evidence from Koivusalo’s study suggests clearly that decentralisation in Finland has not jeopardised equity, even if it has not promoted it.
In Sweden’s devolved health system, Diderichsen (1999) noted that local governments are allowed to tailor reforms that handle the delicate balance between efficiency, equity, and political accountability, and showed that equity of access across counties can be guaranteed by the use of formulas that adjust for the variations in local tax income and health care needs. Although not categorically stated, Diderichsen’s account indicates that there has been equity in the decentralised system of Sweden. Beside the autonomy enjoyed by local governments to pursue equity, the Swedish government compensates poorer counties so that all counties have the same total revenue per head. There is also a formula that accounts for differential needs based on a model developed within Stockholm County Council, which adjusts for variations in population structure with respect to age, education, employment, and housing tenure.
In Spain, Costa-i-Font (2005) found that inequalities in health within and between Spanish health services are relatively small and sometimes not even significant, except when health services of the Catalan region is compared with that of the Navarre, Basque Country or Andalusia. However, inequalities were found within the Catalonia and the INSALUD regions, where private health care is more predominant. The author concluded that devolution in health care does not seem to lead to inter-regional inequalities, although in regions where the private sector still plays a dominant role, greater inequalities may be identified.
101 5.7 Appraisal of Evidence The current evidence on the impact of decentralisation on equity provides a mixed picture of positive and negative results. Several reasons are responsible. One is the failure to distinguish between the effects on equity of different reforms that have been implemented at the same time in the health sector. Many health systems have pursued multiple reform agendas, several of which have the potential to harm or promote equity. In Africa, for example, decentralisation has been implemented along with policies such as user fees and, in fewer cases, the introduction of population/needs-based resource allocation formula.
In some countries, such as Finland, there has been considerable redistribution of resources across decentralised regions/provinces/districts (Koivusalo 1999). Yet, it is often difficult to attribute changes in equity to decentralisation or the use of needs-based allocation models, as many studies have been ambiguous on which particular policy they are evaluating for their impact on equity. Bossert (2000) concluded that decentralisation improves some equity measures but worsens others, a finding confirmed by the evidence presented above. However, the extent to which one can attribute the changes in equity to other reform policies such as implementation of needs- based allocation models closely associated with decentralisation remains unclear.
Another reason for the mixed picture of decentralisation’s impact of equity is related to the meaning of the term ‘equity’ itself; the lack of consensus on a single definition and the level for which equity should be sought and measured (i.e., whether between regions or within regions). The controversy surrounding the appropriate definition of equity (see Chapter 4) makes evaluation of the impact of decentralisation on equity difficult. Klein (2003 p.196) advocated greater clarification of the concept, which he described as a “chameleon concept in the context of the new localism and pluralism”. He noted, for example, that equity in the context of decentralisation could mean equality in the ability to design local services, or equality in the type, level and kind of services delivered.
The level of government at which the assessment of equity under decentralisation should be focused has also been disputed. It is not clear in the literature whether the focus on equity should be on intra-regional rather than inter-regional disparities. The emphasis of equity, particularly, in the allocation of health care resources has largely 102 been on inter-regional/provincial/area equity. Equity within regions, as noted elsewhere, has been ignored. Several recent studies have stressed that intra-regional equity should be the main area of concern (Lopez-Casasnovas 2004; Costa-i-Font 2005). However, decentralisation can operate at multiple levels, and for the policy to achieve its equity goals, intra-regional equity should be of as much concern as inter-regional equity. The focus on one at the expense of the other will constrain the promotion of equity.
Chapter 5. Summary of key points
This chapter has reviewed evidence of the impact of decentralisation on equity in resource allocation.
It highlights the role played by key international agencies such as World Health Organization, World Bank and the IMF in promoting/prescribing decentralisation for developing countries.
Problems associated with centralised administrative systems inherited from colonial rule in many developing countries such as rigid bureaucracy, inflexibility and lack of policy responsiveness to local needs, facilitated the “sale” of decentralisation as a plausible policy alternative.
The ideological position taken by many developing countries to deliver free medical care after independence became difficult to sustain due to stringent fiscal constraints in the 1980s. This left many poor countries with no option but to accept neo-liberal economic and health reforms pushed by the World Bank and the IMF.
Evidence is sketchy and largely inconclusive with respect to the impact of decentralisation on equity in resource allocation. In some countries, the policy has had a favourable impact on equity while in others it has exacerbated inequities.
In Finland, Sweden and Colombia, for example, decentralisation improved equity. In Uganda and Papua New Guinea, decentralisation worsened equity. In Spain, decentralisation improved equity in some regions while in others it had no impact.
The chapter highlights the difficulty involved in isolating the impact of decentralisation on equity from other policies often implemented alongside it, such as, needs-based funding and user fees policies.
The controversial nature of the concept of equity is also contributes significantly to the lack of effective evaluation of the impact of decentralisation on equity.
103 CHAPTER 6
RESEARCH DESIGN AND METHODS
Overview This chapter presents the methods adopted for the study and the approach and tools used for data analysis. It begins by restating the research questions and the specific objectives. This is followed by a description of the research design, which details how the objectives of the study were tackled. The criteria for selection of regions and districts for the study and the techniques employed for collection of data follow. Limitations of the research methods are addressed in the final section.
6.1 Research questions and study objectives As stated in Chapter 1, this study addresses two main questions:
1) To what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? 2) What factors influenced the equitable allocation of resources for health care in Ghana?
To answer these questions, the following specific objectives were identified: To establish whether, and the extent to which, funding allocation from regions to districts has been equitable in terms of differentially benefiting those most deprived. To determine whether, and the extent to which, funding allocation from district to sub-districts differentially benefited the most deprived sub-districts. To examine the factors that influence equity in health care financial resource allocation within regions.
An auxiliary objective which was useful to address but was not a primary objective is: 104 To determine whether, and the extent to which, funding allocation from national to regional levels has been equitable in terms of differentially benefiting the most deprived regions.
On the basis of the study objectives, four working hypotheses were developed to guide the analysis:
That inter-regional resource allocation in the Ghanaian health system has been largely equitable in terms of differentially benefiting the most deprived regions That the decentralisation of resource allocation decisions to regional levels has improved equity by distributing funds in favour of the most deprived districts. That resource allocation at the district level of the health system has been equitable in terms of differentially benefiting the most deprived sub-districts That equity objectives drive resource allocation in the Ghanaian health sector.
6.2 Research Design An important aspect of any research is the design. It is the logical sequence that connects the empirical data to the study’s initial questions and, ultimately, to its conclusions (Yin 1994 p.19). It also ensures that the evidence obtained can address the initial questions as unambiguously as possible (De Vaus 2001). For many decades, two main approaches: quantitative and qualitative, have informed the design of research projects.
Quantitative research design is underpinned by the positivist philosophy, which makes knowledge claims on the basis of careful observation and measurement of the objective reality that exists “out there” in the world (Phillips and Burbules 2000). Qualitative approaches, on the contrary, are often based on social constructivism, which emphasises the subjective meanings that individuals attach to their experiences as they seek to understand the world in which they live and work. The social constructivists argue that meanings are varied and multiple and that research should rely as much as possible on the participant’s views of the situation being studied (Neuman 2000; Lincoln and Guba 2000). Although the two approaches both have strengths and weaknesses, the choice between them has been the subject of heated academic debate (Patton 2002). 105 The design of this study is based on a mixed-methods approach. This has emerged in the last ten to fifteen years and involves collecting and analysing quantitative and qualitative data in a single study (Tashakkori and Teddlie 1998). Creswell (2003) observed that a mixed-methods approach is one in which the researcher tends to base knowledge claims on pragmatic grounds. Thus, in mixed-methods, the research issue is most important, not the method; the researcher applies different methods to gain as much insight as possible into the issue under investigation. Figure 6.1 summarises the sequence of mixed-methods followed in this study.
Figure 6. 1 Sequence of Mixed-Methods Approach used in the Study
MIXED-METHODS
QUANTITATIVE QUALITATIVE
Main objective: Determine Main objective: whether, and the extent to Understand the factors that which, decentralisation of influence the equitable decision-making around allocation of resources for resource allocation has health care in Ghana. improved intra-regional equity
Data: Secondary Data: Primary / Secondary Socio-economic data Interview data Demographic data Fieldnotes Financial data Official documents
Collection Tools: Collection Tools: Document screening Interview guide Direct observation Document screening
Analysis: Analysis: Statistical analysis Content analysis
INTERPRETATION OF RESULTS
CONCLUSION
106 The rationale for using a mixed-methods approach lies in the objectives of this study. Unlike the many equity and resource allocation studies that focus only on establishing the extent of equity in allocations, this study seeks also to explore the factors that influence resource allocation. While the extent of equity in resource allocation could be established quantitatively, the factors that influence the allocation could not be identified and understood without a qualitative inquiry.
There are different models within a mixed-methods structure. In designing this study, three main strategies, namely sequential explanatory, concurrent triangulation and concurrent nested, were considered. The sequential explanatory model is characterised by a two-phased data collection and analysis procedure. The collection and analysis of quantitative data is done first, followed by collection and analysis of qualitative data. Usually priority is given to the quantitative data, but the two methods are integrated at the interpretation phase (Creswell 2003). The main rationale for this approach is to use qualitative results to assist in explaining and interpreting findings of a primary quantitative study. As observed by Morse (1991), this strategy can be particularly useful when unexpected results arise from the quantitative study. In this case, the researcher can use the subsequent qualitative data to gain insight into the unexpected results.
The sequential explanatory strategy was an attractive option, as it was intended that the qualitative results be used to make sense of whatever quantitative findings emerged. However, with the collection and analysis of the quantitative and qualitative data occurring in two distinct phases, it became apparent that the length of time and resources required were clearly beyond the reach of this study and hence, the sequential explanatory method was discarded.
The concurrent triangulation strategy, unlike the sequential explanatory, allows the researcher to use two different methods (quantitative and qualitative) within a single study in a way that makes confirmation, cross-validation, or corroboration of findings possible (Greene et al. 1998; Steckler et al. 1992). Within the scope of this method, both the quantitative and qualitative data collection occurs in one phase (i.e. data collection is concurrent). As in the sequential explanatory model, the results are integrated at the interpretation stage. The main difficulty in adopting this approach is
107 the fact that the phenomenon being studied with the two different methods must be the same. Thus, the same issue must be addressed by the two separate methods (quantitative and qualitative) so that the weaknesses inherent in one method could be offset by the strength of the other method. Since the current study addresses two somewhat separate questions (i.e. to what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? and, what factors influence the equitable allocation of resources for health care in Ghana?), it was inappropriate to use the concurrent triangulation strategy.
The final approach considered was the concurrent nested strategy, which permits both quantitative and qualitative data to be gathered simultaneously in one data collection phase (Tashakkori and Teddlie 1998). But, unlike the concurrent triangulation, where equal priority is given to both the quantitative and qualitative methods, the nested approach has a dominant method that guides the study. If, for example, a study is predominantly quantitative, the qualitative method becomes nested or embedded in the quantitative. One important feature of this approach is that it allows the embedded method to address different research questions than those addressed by the dominant method. This particular feature was found attractive for the current study in which the dominant method, quantitative, addresses a question different from the qualitative.
The main difficulty of embracing the nested approach wholly, however, was that data collected from the two methods have to be mixed during the analysis phase. Due to the significant differences in procedures for data analysis adopted in this study (principal component and regression for the quantitative and content analysis for the qualitative), it was appropriate to merge the two methods at the interpretation phase, rather than at the data analysis stage. A decision was taken to slightly modify the concurrent nested approach so that the two methods of the study (quantitative and qualitative) could be integrated at the interpretation phase. In summary, the design of this study was guided by a mixed-methods philosophy in which quantitative and qualitative methods were applied to concurrently collect and analyse data.
108 6.2.1 Selection of Study Regions Ghana is divided into ten administrative regions and 110 districts (see Chapter 2). It was impossible, given logistics and time constraints, for this study to focus on all the ten regions. Based on data about socio-economic, demographic and health profile of the regions obtained through the initial review of documents, two regions: Ashanti and Northern were deliberately selected for the study. The two regions cover the northern and southern sectors of the country. As discussed in Chapter 2, inequities in health between northern and southern Ghana are striking (Horton 2001) such that no equity- related study can lay claim to sufficient coverage without representing both sectors.
The two regions have features that make them suitable for comparative analysis within a case study framework. They highlight extremes. Ashanti has 18 districts and is the biggest region in Ghana in terms of population (3.6 million). Northern Region has 13 districts and is the biggest region in terms of landmass (70,384 sq. km), nearly one-third the size of Ghana with a population of about 1.8 million. The two regions also have different socio-economic conditions. Ashanti is among the richest regions in the country while Northern is one of the poorest. Both regions have relatively richer and poorer districts. Travel to both regions is comparatively easy, thus facilitating data collection. Section 6.2.2 discusses in more detail why the Ashanti and Northern Regions were selected for this study.
6.2.2 Rationale for Selecting Ashanti and Northern Regions Background data on inter-regional inequities in health and access to health resources have been presented in Chapter 2 (see Section 2.6). The selection of the Ashanti and Northern Regions for this study was based largely on these data.
Northern Region The Northern Region was selected for a number of reasons. There are huge disparities in health and socio-economic conditions between the northern and southern sectors of Ghana (see Chapter 2). First, there was the need for the study to cover both sectors to
109 improve the face validity of emerging insights. The two main regions in this sector beside Northern Region (i.e., Upper East and Upper West) have relatively smaller population and fewer districts. This made the Northern Region a more appropriate choice. It is the biggest region in Ghana in terms of land size and has the largest population and more districts than any other regions in the northern belt.
Second, Northern Region has a high incidence of poverty. The study was designed in such a way as to provide evidence of how poorer regions and rich ones handle the issue of intra-regional equity in resource allocation. The available socio-economic data indicate that Northern Region is one of the poorest in regions in Ghana. Seven out of ten people in the region live in poverty (GSS 2000) and about 80 percent of the population depend on subsistence farming for their livelihood (Ministry of Food and Agriculture 1995).
Finally, the Northern Region was selected specifically for the poor health status of its population. The region is among those with the worst health indicators in Ghana. It had the highest rate of child mortality, about 171 deaths per 1000 in 1998 (GDHS 1998). Infant mortality also remains among the highest in Ghana, about 70 per 1000 live births in 1998 compared to the national average of 56 per 1000 and to 41 per 1000 live births in the Greater Accra Region (MOH 2001). Apart from the high infant and childhood mortality, the region also has one of the highest incidences of Guinea-worm infection (a debilitating water-borne disease) in the world (MOH 2003).
Ashanti Region The selection of Ashanti Region was based on a combination of pragmatism and availability and access to regional data. Unlike the Northern Region, which was the best possible case in the northern sector, there were equally compelling factors for choosing a different region in the south. First, the researcher had a good knowledge of the terrain and socio-economic conditions and had earlier conducted a study in the region31. This prior knowledge of the region played a crucial role in the selection process.
31 Ashanti Region was the study region for the researcher’s master’s thesis entitled- Between Awareness and Behaviour Change: Health Education and Disease Prevention in Ghana’s Ashanti Region; The University of Stockholm 1998. 110 Second, the researcher had vital contacts in the region, specifically at the regional health administration, which was necessary for data collection. Bureaucracy may be a major obstacle in research of this nature, particularly since people in charge of resource allocation might regard information on the distribution and application of funds as politically sensitive and may be reluctant to disclose vital details. Previous contacts with key study participants helps in overcoming bureaucratic barriers and allays the fears of informants. With the limited time available for the study, such contacts who could facilitate data collection were regarded as very important.
Finally, the contrast between Ashanti and the Northern Regions in terms of socioeconomic and demographic characteristics32 was a compelling reason for choice of regions. These differences provided the opportunity to examine whether decentralisation of resource allocation decision-making would improve equity in funding under different socio-economic conditions. Additionally, understanding how poor districts in poor regions (Northern) compares with poor districts in rich regions (Ashanti) in terms of degree of deprivation and its influence on resource allocation was of immense value to the study.
Section 6.3 focuses on how the study objectives were addressed. The two primary objectives reflect the quantitative and qualitative components in the design. For purposes of clarity, the steps taken to address each component are separately discussed.
6.3 ADDRESSING THE QUANTITATIVE OBJECTIVE The quantitative component of the study addresses the following question: To what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in distribution of funds? The main objective was to examine the extent to which the change in resource allocation policy has improved equity in distribution of funds within regions. This involved examination of the extent of equity in resource allocation from regions to districts (inter-district equity) and from districts to sub-districts (inter-sub-district equity). Equity in resource
32 As indicated already, Ashanti has the largest population and is one of the richest in Ghana. The Northern Region is the biggest in terms of landmass and one of the poorest regions in Ghana. 111 allocations from the national level to regions (inter-regional equity) was assessed as an auxiliary objective as it was deemed valuable to examine but was not a primary objective of the study (see Chapter 1). The approach adopted to address these objectives is detailed below.
Approach The obvious approach to addressing the quantitative objective (i.e., to examine whether, and the extent to which, the change in resource allocation policy has improved equity in the distribution of funds) was to undertake a comparative analysis of resource allocation before and after decentralisation. This was, however, not possible for two reasons: first, expenditure data prior to decentralisation was not available. Second, this study was not only interested in establishing whether or not there was equity in the distribution of funds but also in developing an equity-focused approach that can be used as basis for a more equitable resource allocation in Ghana. In view of these, a four-step approach was adopted to assess the extent of equity in resource allocation in this study.
1) Relative deprivation across jurisdictions was measured as a proxy of need using principal component analysis (PCA). The output of the PCA was used to develop a general index of deprivation (GID) for each jurisdiction.
2) The GID was in turn used to construct an equity-adjusted share (EAS) index, which was applied as a yardstick against which equity in resource allocation was assessed.
3) Actual allocations to each jurisdiction from 1998 to 2002 were traced and regression analysis was used to determine the proportion of variance in these actual allocations that could be explained by the predicted EAS.
4) Predicted equity allocations were calculated for each jurisdiction and compared with the actual allocations to highlight the equity-gap.
Sections 6.3.1 to 6.3.4 provide detailed descriptions of each of the four approaches specified above. 112 6.3.1 Principal Component Analysis and Relative Deprivation As noted in Chapter 4, there is a growing recognition of deprivation of area of residence as an important predictor of life chances (Krieger et al. 1997). In Ghana, there is known to be a strong association between health status, socio-economic conditions and area deprivation; poor and rural communities tend to have worse health status (Canagarajah and Ye 2001; GPRS 2002). For this reason, deprivation was used as a proxy measure for health need in this study. The concept of deprivation underpinning this work has been discussed in Chapter 4 (see Section 4.4.2).
While there is general agreement in the literature that area measures of deprivation provide a powerful means of measuring variations in health status, there is no consensus on the choice of variables, their weightings and the appropriate statistical techniques that should be applied to measure deprivation (Gordon 1995; Lynch and Kaplan 2000). The choice of variables for measuring deprivation in this study was guided by careful consideration of the international and local literature on deprivation and what is known about the causes of deprivation or disadvantage in Ghana (see Appendix G for list of documents reviewed). Fifteen (15) variables were initially outlined, but it became apparent in the course of the fieldwork that information on access to radio and on urban-slum residents could not be obtained from the census dataset. The remaining 13 variables used for measuring deprivation in this study are defined in Table 6.1.
113 Table 6. 1 Variables for Measuring Deprivation among Jurisdictions Variable Definition
Female Proportion of district population who are female
Child Proportion of district population who are under five years of age
Elderly Proportion of district population who are 70 years or above
Disabled Proportion of district population suffering from any form of disability
Unemployed Proportion of district population who are between 19 and 59 years old and classified as unemployed irrespective of whether they are actively seeking for employment or not.
Overcrowding Proportion of district population that share one sleeping room with more than two other people at night.
No-Education Proportion of district population above the age of 18 who have no formal education (i.e. above primary school).
Electricity Proportion of district population without access to electricity for lighting.
No-Pipe Proportion of district population without access to piped water in the house or on site (within 500 meters radius).
Toilet Proportion of district population with toilet facilities in the house or within 100 metres radius.
Rural Proportion of district population living in a non-urban area with a population of less than 5000 inhabitants.
Mud-house Proportion of district population living in dwellings with outside walls made of mud.
Wood-fuel Proportion of district population using firewood as the main source of cooking energy. Source: Based on Ghana Statistical Service (2002) dataset.
Analysis Principal component analysis (PCA) was the main statistical technique used for the deprivation measurements in this study. PCA is a multivariate technique that identifies linear combinations of random or statistical variables that maximise variance (Kirkwood et al.1999). It linearly transforms an original set of highly correlated variables into substantially smaller subsets of variables known as components, which represents most of the dimensionality of the original data set. The components are relatively independent of each other, but they reflect the underlying relationships between all the variables. The first principal component accounts for the most variation 114 in the data, the second accounts for more of the remaining variation than any other, and so on (see Jolliffe 1986; Dunteman 1989; Tabachnick and Fedell 1996; Crampton et al.1997).
After defining the variables, they were transformed to improve normality and then correlated against each other using Pearson’s correlation before the PCA was undertaken. This was necessary, since, in PCA, variables must correlate highly with each other for results to be meaningful and interpretable (Tabachnick and Fedell 1996; Salmond and Crampton 2002; McIntyre 2002). Overcrowding was the only variable that did not exhibit high correlation with other variables (where high correlation is defined as significant at 1% level) and was excluded from the PCA. Since additivity is an important criterion that a deprivation index should meet (Gordon 1995), further analysis of the data was conducted to check if some of the variables could lead to double counting. None of the remaining 12 variables could potentially lead to double counting.
It is an important requirement in deprivation measurement that variables be assigned differential weightings to reflect their relative contribution to the overall deprivation (Gordon 1995; Salmond and Crampton 2002). In this study, the factor or component score co-efficients produced through the PCA were used as the weights of the variables. To generate the final deprivation index for jurisdiction, the weight or component score co-efficient of each variable was multiplied by its standardised scores (z-scores) and summed together. This is represented by the formula below:
GID = ∑wizi Where w = is the weight or component score coefficient; z = the z-score of each variable; and i = one to total number of variables included
The z-score is an expression of a raw score that has been standardised by subtracting the mean from the raw score and dividing the difference by the standard deviation (Walsh 1990). Jurisdictions were ranked according to their GID values; those with the highest GID values were regarded as highly deprived and vice versa. One major assumption made in the course of designing this study was that deprivation levels have not
115 significantly changed in the five-year period under investigation. There is a generally held view that deprivation levels change with time, particularly as socio-economic conditions improve or worsen. Evidence from household surveys in Ghana suggests that living standards have not improved in any noticeable proportion in the last decade (GLSS 1992/1999; GPRS 2000).
At the district level where sufficient data to undertake a PCA was lacking, a deprivation index based on two variables (double-index of deprivation - DID) was used. The use of the DID instead of GID was not unusual; several studies have suggested the use of a single-index of deprivation (SID) instead of a composite GID, which is statistically more complex to handle (Folwell 1995; McIntyre et al. 2002). Such studies have argued that the SID is simpler and equally effective in measuring deprivation across jurisdictions. In their South African study, McIntyre et al. (2002), for example, found that the SID they constructed was as effective as the composite GID in identifying deprivation among provinces. In this study, however, an assumption was made that a deprivation index based on two variables would be equally simple and much more potent in identifying deprivation among jurisdictions than a single variable index. The effectiveness of the DID in identifying relative deprivation was tested by developing a similar index at the national level and examining its correlation with the GID. Unlike the development of the GID where variables were assigned different weights, the two variable used to develop the DID were not weighted (i.e., they were given equal weights).
6.3.2 Equity-Adjusted Share (EAS) The EAS, as indicated above, was applied as a yardstick against which equity in resource allocation between and within regions was assessed. It was calculated by weighting the total population of jurisdictions by their GID score. The main assumption underpinning the EAS is that, for resource allocation to be equitable, it should reflect relative deprivation and differences in population size. All the GID scores were first normalised against the score of the least deprived jurisdiction before being applied to weight the population figures. The rationale for normalising the scores was to standardise them by converting the negatives to positive. The EAS adds up to unity and 116 gives the predicted equity allocation that should go to each jurisdiction on the basis of need, i.e., on the basis of deprivation level and population size. The higher the EAS, the greater the funding allocated to a district.
6.3.3 Simple Regression After calculating the EAS, which is used as a benchmark for accessing equity in this study, a simple regression analysis was undertaken to assess the degree of equity in resource allocation by estimating the proportion of variance in the share of funds received by jurisdictions that could be explained by their EAS. Regression analysis attempts to determine the functional relationship between two or more variables (Walsh 1990). The purpose of simple regression analysis is to evaluate the relative impact of a predictor variable on a particular outcome (Zou et al. 2003). Simple regression model contains only one independent or predictor variable (x) and the dependent variable (y).
The actual shares of funds allocated to jurisdictions from 1998 to 2002 were traced. These were transformed into percentages to represent the dependent variables. The EAS% was then ‘regressed’ as an independent variable on the percentage shares of actual funds (dependent variable). The greater the proportion of variance in the actual shares of funds accounted for by the EAS, the greater the equity in resource allocation. The influence of population size on actual share of funds was also assessed in a separate analysis. Since the EAS incorporates population, there was no need for this analysis. However, it was undertaken to be able to assess specifically the extent of per capita allocation in the health sector.
6.3.4 Comparison of Predicted and Actual Allocations The final stage of the analysis was to compare predicted and actual allocations to highlight the equity-gap, which is the gap between the actual amounts of funds allocated to jurisdictions and their predicted allocations based on deprivation and population size. This was necessary, as the regression analysis did not show whether specific jurisdictions were over or under-funded (based on the EAS) during the study period.
117 The total yearly allocations to jurisdictions were re-distributed using the proportion of EAS to obtain the predicted allocations, which were compared with the actual allocation. Section 6.3.5 focuses on the collection and management of the quantitative data.
6.3.5 Quantitative Data Collection and Management The data for the quantitative component of the study were drawn from a variety of secondary sources. The demographic and socio-economic data were extracted mainly from the 2000 Population and Housing Census dataset. This was obtained from the Ghana Statistical Service (GSS) for a fee. Since, by law, the GSS is not permitted to give out raw data to researchers, the 15 variables judged vital for measuring levels of relative deprivation across different jurisdictions (see Section 6.3.1) were submitted to the GSS, which generated the data accordingly.
The financial data were collated from expenditure records of the Ministry of Health and Ghana Health Service, as well as from project documents of donors and the District Assemblies. These data provided the total funding allocation to regional, districts and sub-districts BMCs from government of Ghana and donor sources. Specifically, the data were extracted from revenue and expenditure ledgers, annual financial statements, audit reports and quarterly returns of districts and sub-district. At the national level, data were collated from the consolidated financial statements of the Ministry of Health and from annual statements of the Budget and Planning Unit of the Ghana Health Service. Funding allocations to the ten Regional Health Administrations (RHAs) from 1998 to 2002 were sought. However, both the office of the Financial Controller of the MOH and the Budget and Planning Unit of the GHS had data for only 2002 and 2003. The data collection at the national level was therefore scaled down to those two years in order to make do with what is available.
In Ashanti Region, global expenditure reports showing allocations to all BMCs were available from 1998-2002. This was not the case in the Northern Region where there were no such global expenditure reports. Funds allocated to various BMCs were extracted from revenue and expenditure reports, programme activity reconciliation statements, expenditure budget status reports, and revenue budget status reports. 118 Organisation of the quantitative data began with the entering of figures into dummy tables prepared prior to the fieldwork. Modifications in the number of districts selected for the study created the need for new dummy tables to be made in the course of fieldwork. After the quantitative data was organised, annual disbursements to different districts was compared to establish trends and changes in disbursement patterns over time. This was followed by a statistical analysis using the principal component Analysis (PCA) and simple regression. All statistical analyses were carried out using Microsoft Excel and SPSS Version 10.
6.3.6 Summary of Quantitative Approach The quantitative objective of the study was addressed using a four-step approach. This is presented in Figure 6.2.
Figure 6. 2 The Four-Step Approach used to Address the Quantitative Objective
Double-Index of Principal STEP 1 Deprivation Component Measurement of (DID) Deprivation Analysis (PCA)
General Index of STEP 2 Deprivation (GID) Equity-Adjusted Share (EAS)
% Actual STEP 3 Predicted Share of funds Simple Regression EAS%
Actual Share STEP 4 Predicted of funds Comparison Share of Funds
Step 1. Deprivation among jurisdictions was measured using a PCA. The output of the PCA was used to develop a general index of deprivation (GID). At the sub-district level 119 where data was not sufficient to undertake a PCA, a double-index of deprivation (DID) was constructed.
Step 2. The GID and DIDs were used to develop and equity-adjusted share (EAS), which was used as the equity benchmark or predictor in this study.
Step 3. A simple regression was used to assess the relationship between actual shares of funds allocated to jurisdictions and the equity predictor - EAS. The EAS% was regressed as an independent variable on the percentage of actual shares (dependent variable). The higher the proportion of variance in the actual shares accounted for by the EAS, the more equitable the allocation.
Step 4. A comparison of actual and predicted shares for each jurisdiction was undertaken to highlight the equity-gap. The predicted share of funds was calculated by re-distributing the total yearly allocations among the jurisdictions using the EAS.
6.3.7 Glossary of Terms
PCA – Principal component analysis: This is a multivariate technique that identifies linear combinations of random or statistical variables that maximises variance. It is the main method used to measure levels of relative deprivation across jurisdictions in this study.
GID – General index of deprivation: This is the sum of each variable’s z-score multiplied by its component score coefficient. It is based on the principal component analysis and shows the level of deprivation among jurisdictions. The higher the GID the greater the level of deprivation.
DID – Double index of deprivation: This is a measure of deprivation across jurisdictions based on two variables. It was constructed in this study to assess deprivation among sub-districts because there was inadequate data at that level to undertake a principal component analysis. The higher the DID the greater the level of deprivation. 120 EAS – Equity-adjusted share: This is the equity benchmark developed in this study to assess the degree of fairness in resource allocation. It gives the predicted equity share of resources that should go to each jurisdiction, given its level of deprivation and population size and was derived by weighting the population of each jurisdiction by its GID score.
6.4 ADDRESSING THE QUALITATIVE OBJECTIVE The qualitative component of the study addresses the following question: What factors influenced the equitable allocation of resources for health care in Ghana? The primary objective was to understand the factors that influence the more equitable allocation of financial resources in the Ghanaian health system. To fulfil this objective, primary data were gathered from all levels of the health system (national, regional and district/sub-district). This was complemented with secondary data gathered from administrative documents. The primary data were gathered through open-ended interviews using interview guides. Due to time and resource constraints, the qualitative component of the study could not focus on all the 31 districts in the two regions, as was the case in the quantitative component. Instead, eight districts were selected from the two case regions (four from each region) for meeting this objective. Section 6.4.1 presents the rationale for the selection of the districts.
6.4.1 Purposive Sampling of Districts The eight districts identified for the qualitative study were purposively selected. Purposive sampling is based on the assumption that the researcher knows enough about the population and its elements to identify an informative sample. Patton (1990) noted that, in a situation where a researcher has limited resources and time for the study, has a purpose in mind and recognises that proportionality (in terms of numbers) is not the primary concern, purposive sampling may be the best alternative. Merriam (1998) observed that purposive sampling emphasises a criterion-based selection of information- rich cases from which a researcher can discover, understand and gain more insight into issues crucial for the study. 121 The main weaknesses of this sampling method are the built-in selection bias and the non-statistical representativeness. The researcher must be careful not to select cases solely based on his/her own personal convenience. There is a need to distinguish ‘purposive sampling’ from ‘convenience sampling’. While the former is based on information-rich cases, the latter is based on selecting cases that are easy to access and inexpensive to study. Patton (2002 p.242) cautioned that “while convenience and cost are real considerations, they should be the last factors to be taken into account”.
Several reasons influenced the use of a purposive sampling technique to select the districts. These were largely based on the available demographic, socio-economic and locational information about the districts prior to the study. Section 6.4.2 presents the information that influenced the selection of districts in Northern Region.
6.4.2 Rationale for Selection of Districts in Northern Region The four districts selected from a total of 13 in the Northern Region were Tamale, West Gonja, West Mamprusi and Savelugu-Nanton. Table 6.2 presents some of the demographic and socio-economic factors that were considered in the selection process.
Table 6. 2 Selected demographic and socio-economic characteristics of districts in Northern Region Districts Population % % With % Un- Malaria* % of Urban No Employed Profile Malaria Education (1999) Cases Tamale Municipal 293,881 67.1 60.7 13.4 83,397 26.6 West Gonja 139,329 14.5 81.8 12.8 17,435 6.3 West Mamprusi 115,025 16.2 86.2 19.8 15,164 4.6 Savelugu-Nanton 89,968 36.2 86.0 7.4 15,876 4.9 Bole 127,147 11.9 83.3 11.8 16,279 6.0 East Gonja 174,500 13.7 79.3 13.8 12,492 3.8 Nanumba 144,278 19.8 82.7 10.2 14,650 4.5 Zabzugu-Tatale 79,201 21.2 86.4 8.2 6,691 2.1 Saboba-Chereponi 93,847 6.7 82.2 14.7 25,030 7.7 Yendi 130,504 34.9 81.7 10.2 30,651 9.4 Gushiegu-Karaga 125,430 21.1 87.1 18.9 7,841 2.4 Tolon-Kumbungu 132,833 16.7 86.4 14.4 15,236 4.7 East Mamprusi 174,863 18.1 80.2 23.4 65,608 20.1 Total 1,820,806 26.6 79.1 14.6 326,367 100 Source: Population and Housing Census of Ghana, 2000 Note: Percentage with no education is in relation to district population age 18 years and above *Figures taken from regional OPD Morbidity Returns 1999.
122 Tamale Municipal The Tamale municipal district is the most populous and urbanised of the 13 districts in the Northern Region (Table 6.2). It is the most accessible district with some of the best network of roads in the region. The district capital – Tamale, with a population of 202,317, is the political, economic, and financial capital of Northern Region. Due to its central location and good road network, Tamale has long served as the hub of commercial activities and is regarded as relatively rich compared to other districts in the region. It has 15 health facilities, including one regional hospital and a district hospital. Like many urban settlements in Ghana, sanitation is a major problem in the district, partially explaining why malaria, the leading cause of sickness and death in Northern Region, is more prevalent in the municipality than any other district. The Tamale municipal was selected for this study mainly because of its high urban and rural populations. Its status as the richest district in the region, easy accessibility and earlier contacts with key participants were the other reasons for its inclusion.
West Gonja The West Gonja district is located at the central-western part of the Northern Region. In terms of landmass, it is the biggest district in the region and in Ghana. With only about 140,000 inhabitants, the district is the most sparsely populated in the region. Damango, the district capital, has a total population of 14,442 people. Subsistence farming is the main occupation. The district is divided into six sub-districts and has a total of ten health facilities, including one district hospital. Some of the sub-districts have no health facility and receive mass immunisation services only three times in a year. Since 1995, the district has been operating a health insurance scheme which covers in-patient care, payment of prescribed drugs and payment of monthly average in-patient care cost for all referrals (Aikins 2003). The West Gonja was selected for this study for its vast landmass, sparse population and dispersed communities.
West Mamprusi The West Mamprusi district lies about 100 kilometres north of the Tamale district and covers an area of about 4,810 square kilometres. The bulk of the district’s population
123 survive on subsistence farming. Walewale, the district capital, has a population of about 13,558 people. The district is divided into five sub-districts, two of which are in areas designated as overseas (hard-to-reach). The district is serviced by a single hospital located in the capital and ten other facilities, mainly health centres and clinics. With exception of the district capital that has access to piped water, all the other communities rely on streams, rivers, dugouts and boreholes for water. Malaria, diarrhoea, upper respiratory track infections, intestinal and guinea worm diseases are the common causes of sickness and death. The West Mamprusi district was included in this study for several reasons, including the variations in socio-economic conditions among the communities, the easy accessibility from the regional capital, and reports of effective collaboration between the DHA and the District Assembly. Assessing such collaborations was important for examining the local government’s inputs in planning and funding of health delivery in Ghana.
Savelugu-Nanton The Savelugu-Nanton is a small peri-urban district located some 25 kilometres north of Tamale. The population is a little under 90,000 with about 36 percent living in urban settlements. Savelugu, the district capital, has a population of about 25,000 people. The district is divided into eight sub-districts. There is no hospital in the district, but each sub-district has a health centre. The biggest health centre is at Savelugu, where the district director of health services is the only qualified medical doctor in the district. Access to health care differs significantly in the district, with some communities having relatively easy access to health facilities in the regional capital (Tamale). Other communities, in contrast, have to travel very far through bush paths on bicycles to reach the nearest health centre. The peri-urban nature of the Savelugu-Nanton district was one reason why it was selected for this study. The easy accessibility of the district and the early progress made in establishing contacts with the key district health officials also played a significant role in selecting the district.
124 6.4.3 Rationale for Selection of Districts in Ashanti Region The four districts selected in the Ashanti Region were the Kumasi Metro, Sekyere East, Ashanti-Akim North, and Amansie East. Modifications were made to the list of districts selected prior to the fieldwork after receiving additional district specific information from key informants. The earlier list included the Asante-Akim South district, which was selected on the basis of easy accessibility to most of the communities. However, information from key informants suggested that the there were tensions between the Asante-Akim North DHA and a number of sub-districts over resource allocation. This was particularly important to the study objectives and caused a swap of the Asante- Akim North with Asante-Akim South. Table 6.3 provides some demographic, socio- economic and morbidity data used to support the selection of four districts in Ashanti.
Table 6. 3 Selected demographic and socio-economic characteristics of districts in Ashanti Region Districts Population % % with No % Un- Malaria* % of Urban Education Employed Profile Malaria (2002) (2002) Amansie East 225,309 12.2 41.4 11.2 23,664 3.8 Asante-Akim North 126,477 56.9 36.1 11.5 14,266 2.3 Sekyere East 157,396 33.7 46.8 18.1 20696 3.3 Kumasi Metro 1,170,270 100 29.3 17.9 220,025 36.5 Atwima 237,610 20.7 39.6 12.1 39,776 6.4 Amansie West 108,726 - 41.3 8.8 12,416 2.0 Adansi West 238,440 61.1 32.5 13.9 62,685 10.1 Adansi East 129,308 7.5 44.6 8.1 12,969 2.1 Ahafo-Ano South 133,632 9.2 54.8 22.6 7,810 1.3 Ejisu-Juaben 124,176 26.5 30.1 12.2 18,008 2.9 BAK 146,028 6.0 37.1 18.3 35,078 6.7 Kwabre 164,668 38.9 36.5 16.7 32,400 6.2 Afigya-Sekyere 119,093 36.6 34.9 7.8 36,176 6.8 Sekyere West 143,206 39.0 42.0 7.8 23,476 3.8 Ejura-Sekyedumase 81,115 48.8 56.5 8.1 13,490 2.2 Offinso 138,676 31.0 44.6 8.0 25,442 4.1 Asante-Akim South 96,868 16.5 37.5 14.1 11,276 1.8 Ahafo-Ano North 71,952 19.2 46.0 7.9 10,935 1.8 Total 3,612,950 51.3 36.8 14.6 620,588 100 Source: Population and Housing Census of Ghana, 2000. Note: Percentage with no education is in relation to district population age 6 years and above. *Figures taken from regional OPD Morbidity Returns 1999. The malaria data used for the Ashanti differed from that of the Northern Regions. At the time of the fieldwork, the most recent data on malaria for Ashanti were for 2002. In the Northern Region, 1999 morbidity data were the latest available.
Amansie East The Amansie East district was selected for the study mainly for its high incidence of poverty. There was the need in this study to have a mix of rich and poor districts. The 125 Amansie East district is located approximately 35 km south of Kumasi. It is one of least developed districts in the Ashanti Region with only about 12% of the population living in urban areas. The road network in the district is very poor, making access to health facilities immensely difficult in some areas. The district is divided into eight sub- districts, all of which have at least a hospital, health centre or a clinic. The government hospital at Bekwai (the district capital) is the main health facility in the district. However, the mission hospital in the Dominase sub-districts is of good standard and provides basic clinical services to the population. There are concerns about the rise of tuberculosis in the district. In 2002, the district recorded 100 TB cases, the second highest in Ashanti Region33 (Health Information Unit, Ashanti Region 2003).
Asante-Akim North The Asante-Akim North district was selected based on specific information received from key informants about resource allocation among sub-districts. The district lies on the eastern frontier of the Ashanti Region and shares a boundary northward with the Sekyere East. It is a peri-urban district with about 56% of its population living in urban areas. Konongo, the district capital, is located on the main Accra-Kumasi highway, some 40 kilometres east of Kumasi. The district is divided into five health sub-districts and is serviced by two major hospitals and several public and private facilities. Like the other districts, there is a high incidence of malaria and diarrhoeal diseases in the district.
The Presbyterian hospital at Agogo (the sub-district capital) is one of the oldest and best-equipped health institutions in Ghana. The Presbyterian mission adopted the Agogo sub-district as its ‘focal’ district. Information from a key informant suggested that there have been conflicts over resource allocation between the Agogo sub-district health administration and the DHA at Konongo. As an adopted sub-district, Agogo receives direct financial and logistic support from the Presbyterian mission. This has reportedly prompted the DHA to reduce its financial allocations to the sub-district, something the sub-district health administration was said to be unhappy with, maintaining it has many underprivileged communities and still deserves a normal allocation from the DHA. It was strategically feasible, given the emphasis of this study on inter and intra-district
33 Kumasi Metro recorded the highest TB cases of 315 in 2002. 126 equity in resource allocation, that such vital piece of information is taken into account. It was believed that more could be learnt about the “conflicts” surrounding resource allocation by looking at the Asante-Akim North. The district was also selected for its peri-urban nature; it was important to examine districts that reflect a range of possible scenarios.
Sekyere East The Sekyere East district was included in the study for several reasons including the vast inequalities in access to health care between communities, personal contacts with some members of the DHMT, and easy accessibility to the district capital. The Sekyere East district located at the north-eastern part of Ashanti Region. It is the largest district in Ashanti in terms of landmass, covering about 16% of the total area of the region. It, however, has the lowest population density of about 39 persons per square kilometre. The entire northern sector of the district is located in the Afram Plains; a vast stretch of sparsely populated land area well known for its difficult accessibility and under- development. About 120 communities make up the district, 55 of them located in the under-developed north. Almost 80% of the total population, however, live in the southern part of the district.
Effiduasi, the district capital, has a population of 18,700 and is located some 36 km from Kumasi. There is a good network of second-class roads linking towns and villages in the south. No direct route, however, exist between the north and south, the shortest route is about 173 km through other districts. Sekyere East is divided into five sub- districts and serviced by three hospitals, 10 health centres and two private maternity clinics. The bulk of these facilities are located in the southern sector. Malaria remains the major cause of morbidity in the district, but liver and cardiac failures were among the top five causes of mortality in 2000 (District Annual Report 2000).
Kumasi Metro The Kumasi metro district with a population of 1,170,270 is the second most populous district in the whole of Ghana (second only to Accra metropolitan district). It covers a
127 total land area of 2200 square kilometres and has a density of about 532 persons per square kilometre. Kumasi - the district capital, which is also the capital of the Ashanti Region, is centrally located and easily accessible by road from almost all parts of Ghana. It is a cosmopolitan city with people from different ethnic backgrounds, including migrants from other West African countries.
The Kumasi district is the most developed district in the region and has all the basic amenities. The district is divided into five sub-districts all of which have well- functioning health facilities. The district is also serviced by one of the only two teaching hospitals in Ghana and a number of private hospitals and clinics. Malaria is the major cause of ill-health and death. However, Tuberculosis and HIV/AIDS are more prevalent in the district than in other districts in the region. Non-communicable diseases, particularly, hypertension and diabetes have also increased in recent years (RHA, Ashanti Region 2002). The Kumasi district was selected for the study for two main reasons; its economic status as the richest district in Ashanti and the large heterogeneous population.
6.4.4 Qualitative Data Sources and Recruitment of Participants The qualitative objective was met by analysing both primary and secondary materials obtained from face-to-face interviews, key informants, direct observations and administrative documents (Figure 6.1). The interviews were largely open-ended, using interview guides, which specified broad domains to be covered and key issues to be explored. The six main interview domains covered were sources/level of funding for the health system, mechanisms for disbursing and accessing funds, system of financial accountability, conceptualisation of equity, donor involvement in the health sector, and collaboration between health and the local government. In all, 26 interviews were conducted, each lasting about an hour. Table 6.4 provides the list participants.
128 Table 6. 4 List of Participants
Position of Participants Number of Participants
Donor Representatives 2
Policy makers 4
Regional Health Managers 6
District Health Managers 8
Sub-district Health Managers 2
Local government officials 4
Total Participants 26
Participants were selected purposefully according to their job positions. Some participants were nominated by other participants in the course of the fieldwork. Permission was obtained first from the Director General of the Ghana Health Service as well as from heads of departments of all sectors where potential respondents had been identified. All department heads were provided with an official letter of introduction from The University of New South Wales explaining the purpose of the research including the type of information being sought and requesting their cooperation.
The interviews started at the regional levels and moved to the districts before reaching the national level34. The main reason for starting at the regional level was to assess which district financial data were available at the regional health administration so that time would not be wasted collating them from individual districts. It was also useful to gain some prior insights into the district allocation system and gather information about the various directors of health services at the regional level to make the necessary contacts at the district level. All interviews were conducted between August 2003 and February 2004 and were tape-recorded. The participants were verbally asked if they could be taped. None of them raised any concerns about participating in the study or being taped.
In addition to the formal interviews, informal conversations were held with two key
34 It must be noted that the phases overlapped due to postponements of scheduled interviews and the need to re-visit some respondents. 129 informants from both the Ashanti and Northern Regions. Information obtained from these informants was not used as part of the interview data, except for supporting the selection of districts considered to be case-rich for the study. Notes were taken in the course of the fieldwork detailing observations made during travelling across the districts. Certain aspects of the field notes that shed light on the level of deprivation in the districts and sub-districts, particularly with regards to access to good drinking water, electricity and condition of the road network, were used to complement the interview data.
The review of literature undertaken prior to the fieldwork led to the identification of a range of sources where important data could be collated for purposes of triangulation. All such sources were explored and relevant data collected at the same time as the interviews were conducted. The main sources for these secondary qualitative data were administrative and policy documents of the MOH and GHS as well as project documents from the major donors in the health sector. The documents included the following:
Health Sector 5-Year Programme of Work 1997-2001 Health Sector 5-Year Programme of Work 2002-2006 Ghana Poverty Reduction Strategy Document (Final Draft Version) 2002 Reports of External Review Team on MOH Programme of Work 2002 Reports of External Review Team on MOH Programme of Work 2001 Reports of External Review Team on MOH Programme of Work 2000 Reports of External Review Team on MOH Programme of Work 1999 Danida Health Sector Programme Support Document (final draft) 2002 Health of the Nation: Reflections on the First 5-Year Programme of Work 2001 Review of Financial Management Systems: MOH 2001 World Bank Health Sector Programme Support Project II Document, 2003 Ghana Demographic and Health Survey 1998
130 6.4.5 Organisation and Analysis of Qualitative Data The process of organising and preparing the data for analysis started in Ghana before the fieldwork was completed. This is not unusual, as several studies have drawn attention to the thin line between data collection and analysis (Patton 2002; Marshall and Rossman 1999). Organisation and analysis of the qualitative data were quite daunting, not just for the volume of data available, but also for the absence of any universally acceptable rules for analysing qualitative data. Miles and Huberman (1984 p.16) have observed that “we have few agreed-on canons for qualitative data analysis, in the sense of shared ground rules for drawing conclusions and verifying their sturdiness”.
The organisation of qualitative data started in the field with the labelling of each interview tape by level (whether national, regional, district or sub-district), name and position of the respondents as well as date of interview. Tapes were reviewed by listening immediately after each interview to make sure the recording equipment functioned well and all conversations have been duly recorded and are audible. Interviews were transcribed over six months period alongside the analysis of the quantitative data. All 26 interviews were transcribed verbatim. To ensure confidentiality of the information and protect the identity of participants, no outside help was sought for the transcription of the interview tapes. Data were analysed and presented in a de- identified form.
Transcripts of interviews were analysed for patterns, themes, and captivating account of events. Based on the main qualitative research question, (i.e., what factors influence the equitable allocation of resources for health care in Ghana?) and the interview domains covered, the following questions were posed to guide the analysis.
How is equity conceptualised by policy-makers and implementers and how does this influence resource allocation patterns at national, regional and district levels? Are the mechanisms for disbursing and accessing funds in keeping with the objectives of the resource allocation policy? What role does politics play in influencing resource allocation patterns in Ghana? What are the manifestations of this within the resources distributed?
131 What influence do donors have on the way funds are allocated through the health system? Do they use this influence? How? Why? When? What role (if any) does the local government play in shaping resource allocation at the district level?
Despite the lack of consensus on the best way to analyse qualitative data, there is no disagreement on the difficulty or inappropriateness of analysing it statistically. Interpretation is generally considered to be the core of qualitative analysis. This requires that researchers immerse themselves in the data, read and re-read transcripts, code and organise, interrogate and actively look for explanations and understandings that can only be produced from getting one’s hands dirty with the data (Gifford 1996). In this study, content analysis was the main analytical approach adopted to identify, code and categorised patterns and themes in the qualitative data. According to Patton (2002 p.463) “content analysis involves identifying, coding, categorising, classifying and labelling the primary patterns in the data”.
Analysis was, for the most part, undertaken manually, using pen and paper. First, all transcripts were thoroughly read and comments were made in the margins during reading regarding the main issues being raised by the data and how to categorise these issues. This proved to be very necessary because even though the interviews were somewhat categorised by domains covered as designed in the interview guide, new issues emerged and some of the responses were also found to be more appropriate under different headings than where they were given. Data that did not fit any of the categories developed were carefully examined to ensure that their exclusion from the study would not negatively affect the results.
Second, after getting the general feel of what respondents were largely talking about, the data were categorised into six broad headings in line with the interview domains using Microsoft Excel. The headings were:
1. Equity and equitable allocation of resources 2. Sources/level of funding for the health system 3. Mechanisms for disbursing and accessing funds 4. System of financial accountability
132 5. Donor involvement and politics 6. Collaboration with the local government
Responses from each level of the health system (national, regional, district and sub- district) were grouped together under these headings. In developing the codes and categories, attention was paid to the issue of convergence and divergence, which Guba (1978) has described as challenging to qualitative analysis35. Regular re-visiting of and interaction with the data and the classification system helped in verifying the meaningfulness and accuracy of the categories as well as the placement of data in categories. Where applicable, responses were categorised to show regional variations.
Third, detailed reading of information under each category was undertaken to identify patterns and themes emerging from each level. Themes were not imposed but allowed to emerge from the data. Themes that emerged were analysed first for different levels, showing regional variations where necessary, before cross-level analysis was undertaken. The cross-level analysis generated new insights into factors that underlie resource allocation and equity in the health system. Such insights were not glaring in the level-specific analysis. Different colours were used to highlight important words, phrases, sentences and captivating stories. Short phrases denoting the main issues being raised by respondents were given codes. All codes started with abbreviations that communicate which level a particular response comes from before the subject matter being coded. The abbreviations included the following: D=donors, PM=policymaker, RM=regional manager, DM=district manager, SDM=sub-district manager, DA=district assembly respondent, AR=Ashanti Region and NR=Northern Region.
Fourth, the codes were used to generate themes and sub-themes, which were grouped into factors that influence the resource allocation process. From this stage, meanings were developed to explain the factor that emerged as influential for resource allocation. Here, the researcher drew on his own perspective and understanding of the general socio-economic and health environment of Ghana, among other things to make sense of
35 Convergence deals with the extent to which data belonging to a certain category fit together. It involves figuring out what things fit together in the data. Divergence in contrast concerns “fleshing out” the patterns or categories. It also includes careful and thoughtful examination of data that does not seem to fit including deviant cases that do not fit the dominant identified patterns (Patton 2002). 133 the findings. Comparative analysis of the way resource allocation was perceived at different levels of the health system was undertaken in order to unravel all necessary hidden reasons. Possible explanations and counter-arguments were put forward to bring meaning to bear on the findings.
As indicated earlier, some aspects of the fieldnotes taken during data collection were used to complement the interview data. These were mainly those aspects that provide additional insights into variations in deprivation levels of districts and sub-districts. Such notes, which were taken in an unstructured manner in the field, were re-written in full each day either in the evening or immediately after the interview. They were interpreted directly and used to support arguments raised.
6.4.6 Interpretation of Qualitative Data The final stage in the data analysis was the interpretation of findings. Interpretation, as Patton (2002 p.438) observed, “involves explaining the findings, answering the why questions, attaching significance to particular results, and putting patterns into an analytic framework”. It involves going beyond the descriptive data to make sense of the results, draw conclusions and extrapolate lessons. There are several ways in which this can be done. In this study, the analytic induction approach, which involves cross-case analysis in search of explanations was utilised. In analytic induction, researchers develop hypotheses prior to entry into the field or data analysis. The hypotheses can be based on hunches, assumptions, careful examination of research and theory, or combinations (Gilgun 1995). Hypotheses are revised to fit emerging interpretations of data over the course of data collection and analysis. Analytic induction requires that enough cases be examined to assure development of a universal hypothesis. The eight districts studied for the qualitative component justified the adoption of analytic induction. Hypotheses developed prior to the fieldwork were modified to fit emerging interpretations of the data during the analysis.
Validity is a critical issue in both quantitative and qualitative research. It concerns determining whether insights derived from research are accurate from the standpoint of the researcher, the participant, or the reader of accounts (Creswell and Miller 2000). In
134 this study, the main strategy adopted in checking the accuracy of the findings was triangulation of data sources. The multiple case study design, the focus on different administrative, levels and the different stakeholders included in the study, facilitated the triangulation of data sources. Information derived from different sources at different times was compared to build a coherent justification for themes. For instance, interviews and observations were compared and interview data were cross-checked against secondary data from programme documents. Similarly, views of participants from different administrative levels about key issues such as equity, reimbursement for exemptions, and regularity of flow of funds were also compared.
6.5 Methodological Issues The limitations of this study, including the gaps in the data have been discussed in Chapter 10. However, there are several issues about the research methods that need to be highlighted here. The first relates to the mixed-method approach used for study. Despite the potential for gaining a more comprehensive understanding of the research issue using mixed-methods, it takes greater effort and expertise to adequately study a single phenomenon with two separate methods, as observed by Creswell (2003). It involves extensive data collection and time-intensive analysis of text and numeric data. These difficulties were, however, adequately overcome in this study.
The second concern relates to the external validity of the study. One may argue that the selection of only eight districts from a total of 31 in the two regions for the qualitative analysis might have some effects on the external validity of the study. While this argument may be valid, external validity is not a major concern of qualitative inquiry, as reliability and generalisability play a minor role (Creswell 2003). It would have been better if more than eight districts were studied to increase confidence and credibility in the results, but this was not possible given the constraints on resources. It must be noted, though, that in qualitative inquiry, “there are no rules for sample size” (Patton 2002 p.244). It depends on what one wants to know and the purpose of the inquiry.
Finally, the issue of selection bias associated with purposive sampling could not be completely overcome, as accessibility was a key criterion in selecting the districts.
135 However, steps were taken to ensure that information-rich cases, which supported the main purpose of the study, were selected.
Chapter 6. Summary of key points
This chapter described the various methods used in the study, including how the study was designed and research questions framed.
A mixed-methods approach was used for the design of the study.
The Ashanti and Northern Regions were selected from among the ten regions in Ghana for this study. Regional morbidity, demographic and socio-economic data were used to support the selection.
The quantitative component of the study addresses the following question: to what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? A four-step approach was used to address this question:
• Relative deprivation across jurisdictions was measured as a proxy of need using principal component analysis (PCA). The output of the PCA was used to develop a general index of deprivation (GID) for each jurisdiction.
• The GID was in turn used to construct an equity-adjusted share (EAS) index, which was applied as a yardstick against which equity in resource allocation was assessed.
• Actual allocations to each jurisdiction from 1998 to 2002 were traced and regression analysis was used to determine the proportion of variance in these actual allocations that could be explained by the predicted EAS.
• Predicted equity allocations were calculated for each jurisdiction and compared with the actual allocations to highlight the equity-gap.
The qualitative component addresses this question: what factors influenced the equitable allocation of resources for health care in Ghana? This was tackled by analysing primary and secondary data.
The primary data were gathered through open-ended interviews with 26 participants comprising health policymakers, donors, regional, district and sub-district health managers as well as local government officials. The participants were purposefully selected. The interviews were conducted between August 2003 and February 2004.
The qualitative data were analysed using content analysis.
136 CHAPTER 7
EQUITY IN RESOURCE ALLOCATION IN GHANA: RESULT OF DOCUMENTARY ANALYSES
Overview This chapter presents findings on equity in resource allocation derived from analysis of key documents. It underpins to the results presented in Chapters 8 and 9. Data are organised around the following key topics: Funding of the health system Allocation of financial resources in the health sector Mechanisms for disbursing and assessing funds Systems for promoting financial accountability.
7.1 Funding of the Health System The Ghanaian health system is funded through four major funding sources: government, donors, commercial credits and internally generated funds (IGF). The proportion of GDP allocated to health in Ghana has been one of the lowest among developing countries. Government spending on health between the early 1980s to the mid-1990s was only about 1% of GDP. However, health expenditure as a proportion of GDP has increased in recent years following the implementation of the first Health Sector 5-year Programme of Work (1997-2001), rising to the peak of 1.6% of GDP in 2000. Including donor and other forms of funding, the World Bank estimated that Ghana spent approximately 2.2% of GDP on health in 2000. Although this is still below the sub- Saharan African average of 2.5%, it compares favourably with a number of other developing countries. Nigeria, for example, spends 0.5% of GDP on health while Kenya and Thailand spend 1.8% and 2.1% respectively (World Bank 2004).
Traditionally, government allocation through the MOH has constituted the largest funding source for the health sector. Table 7.1 shows expenditure and sources of 137 funding for the health sector from 1996 to 2001. Government funding through the MOH increased from 42.4% in 1996 to about 59% in 1999 before dropping to about 51% in 2001. Donor funding as a proportion of total health expenditure dropped from 21% in 1996 to 15.3% in 1998 but increased more than two-fold to 35% in 2001 (further evidence on donor funding is presented next section).
Table 7. 1 Health Expenditure and Sources of Funding: 1996-2001 Year 1996 1997 1998 1999 2000 2001 Expenditures Billion Cedis Recurrent 146.8 182.5 258.5 334.2 532.4 792.9 % Recurrent 49.6 56.1 71.8 84.0 82.7 90.7 Capital 149.0 142.9 101.5 63.5 111.5 80.9 % Capital 50.4 43.9 28.2 16.0 17.3 9.3 Total 295.8 325.4 360.0 397.7 643.9 873.8 Funding MOH 125.5 139.5 195.0 233.2 356.3 444.6 % MOH/Total Funding 42.4 42.9 54.1 58.6 55.3 50.9 Commercial Credit 99.0 101.6 77.0 27.8 31.7 13.4 % Comm. Credit/Total 33.5 31.2 21.4 7.0 4.9 1.5 Donors 62.2 56.7 55.2 89.1 183.8 306.2 % Donors/Total Funding 21.0 17.4 15.3 22.4 28.5 35.0 IGF 9.1 27.7 33.0 47.6 72.2 109.6 % IGF/Total Funding 3.1 8.5 9.2 12.0 11.2 12.5 Total 295.8 325.5 360.2 397.7 644.0 873.8 Source: Danida Health Sector Programme Support (Phase III) Ghana 2002, Programme Support Document
The composition of health expenditure for 2001 reveals that a little over half of the total expenditure for that year came from government sources, with donor contributions and IGF constituting most of the rest (Figure 7.1).
Figure 7. 1Composition of Health Expenditure - 2001
13%
50% 35%
2%
MOH Commercial Credit Donors IGF
Source: Based on Data from Danida Health Sector Programme Support - 2002
138 IGF as a source of funding for the health sector has increased in recent years. From a meagre 3.1% in 1996, IGF revenue funded nearly 13% of health delivery in Ghana in 2001. In the current health sector programme of work (POW II 2002-2006), IGF revenue is expected to fund at least 14% of the total health budget annually over the five-year period (MOH 2002). Concerns have been expressed in recent years that the growth in IGF revenue could impede access to health care for the poor, as it may be generated from fee increases in public facilities. However, this may not necessarily be the case: better financial management rather than user fee increases may be responsible for the increase in IGF revenue (see Chapter 9 for views of participants on this issue).
There have been heightened expectations in the last few years of additional funding for the health sector from debt relief saving under the HIPC initiative. In 2002, Ghana saved about US$275 million with a significant proportion of this spending on health and education (MOH 2003). Co-ordinating Directors of District Assemblies indicated in interviews that HIPC funds have already been made available to the Assemblies for spending on health and related social services (see Chapter 9). The health expenditure composition in 2004 had a specific HIPC component. Figure 7.2 shows the composition of health expenditure for 2004. HIPC funds accounted for about ⊄104 billion (cedis) or 4% of total health expenditure and replaced commercial credits. The total donor contribution of ⊄1,035 billion (43%) was slightly higher than the total government health expenditure of ⊄1,027 billion (43%) and significantly higher than the 2001 contribution of ⊄307 billion (see Figure 7.1).
Figure 7. 2 Composition of Health Expenditure - 2004
104 250
1,027
1,035
MOH Donors IGF HIPC
Source: Based on Ministry of Health Data - 2004
139 The introduction of a National Health Insurance Scheme (NHIS) in 2005 is expected to further increase funding to the health sector. About US$147 million (26.2%) of the total health expenditure for 2005 was expected to come from a national health insurance levy (MOH 2005).
The increase in government spending on health is more in terms of an increase in MOH recurrent expenditure as a proportion of total government recurrent budget. Since the inception of the POW I, this component of funding has been increasing with modest fluctuations. Table 7.2 shows the government expenditure allocations to the MOH and the per capita spending on health from 1996 to 2002.
Table 7. 2 Funding Allocations to Ministry of Health and Per Capital spending: 1996- 2002
Indicators/Year 1996 1997 1998 1999 2000 2001 % MOH Total/GDP (excl. donor funding) 1.3 1.2 1.3 1.4 1.6 1.5 % MOH Recurrent/ GOG Recurrent 7.0 8.4 8.7 9.5 11.4 11.0 % MOH Capital/GOG Capital 4.6 4.8 3.2 2.1 1.9 0.6 % MOH Total/GOG Total 4.9 5.2 7.2 5.9 7.5 7.9 MOH Total per Capita (in real US$ terms excluding donor funding) 4.62 4.18 4.93 4.82 4.29 4.1 MOH Total per Capita (in real US$ terms including donor funding) 10.16 8.14 7.81 7.83 7.44 7.40 Source: Health Sector 5-Year Programme of Work I & II; Annual Review of the POW 1999, 2000, 2001, 2002; Danida Health Sector Programme Support (Phase III) Ghana 2002; World Bank Report No: 24842-GH January 2003
The MOH recurrent allocation as a percentage of total government recurrent budget increased steadily from 7% in 1996 to 11.4% in 2000 but slightly declined to 11% in 2001. Allocation for capital investments to the MOH as a proportion of government capital budget, by contrast, has been on the decline. Many observers have interpreted this as an increase in attention to service delivery rather than capital projects. However, the qualitative interviews revealed that the high recurrent allocation is driven by the government’s “desperate” attempt to improve salaries of health personnel as a way of addressing the increasing brain-drain problem (Chapter 9).
In per capita terms, health spending in Ghana remains one of the lowest among developing countries, despite recent improvements in government spending. Excluding donor funding, the MOH total per capita spending has been under US$5 since 1997 (Table 7.2). This is a marked decline from the US$10 per capita level in 1978. Even 140 with donor funding included, the trend has deteriorated from about US$10 in 1996 to US$7.40 in 2001 and continues to worsen. The impact of low funding on delivery of health services was explored in the interviews undertaken for this thesis (Chapter 9). Section 7.1.1 examines in more detail the composition of donor component of the health budget.
7.1.1 Donor Funding As demonstrated by Table 7.1, donor funding constitutes about 35% of the total health budget in Ghana. There are about 15 significant donor and technical agencies involved in the health sector alongside numerous international and local NGOs (PSU 2000). The rise of donor activity in the health sector dates back to the 1980s following the economic decline in the 1970s, which weakened the ability of the government to fund the health system. Donor funding rose from a tiny US$1 million in 1984 to US$12 million in 1990 and to US$25 million in 1995 (WHO 1999). Under the current health sector five-year programme of work (POW 2002-2006), donor contributions are expected to finance an average of about 36% of total health expenditures36 over a five- year period from 2002 to 2006 (MOH 2003) as depicted by Figure 7.3.
Figure 7. 3 Sources of Funding for the Health System 2002 - 2006
140
120 Government 100 Donors 80 HIPC Funds 60 IGF
Budget (Million US$) Budget 40
20
0 2002 2003 2004 2005 2006 Years
Source: Based on POW II (revised edition) 2003 figures
36 Note that in 2004, donor contributions amounted to about 43% of the total health expenditure (see Figure 7.2). The 36% is only an estimated average donor contribution for the 5-year period from 2002 to 2006. 141 At a projected US$80 million a year, donors contributed more to the health budget in 2002 than the government. Donor contribution was nearly at par with government contribution in 2003, and remained high for the remaining years. The five key donors: the World Bank, DFID, Netherlands Government, Danida and the European Union contributed almost half of all donor inflows to the health sector. Figure 7.5 shows the expected contributions to the health budget by each of the five key donors from 2002- 200637.
Figure 7. 4 Expected Total Contributions to the Health Budget by Major Donors: 2002-2006
3% 15%
38%
24%
20%
World Bank DFID Netherlands DANIDA EU
Source: Based on World Bank figures in Health Sector Programme Support Project II Document, 2003 (Report No: 24842-GH)
The Sector-Wide Approach (SWAp) is a central theme in Ghana’s health sector reform. The principle underlying SWAp forms the basis of investments and actions by the MOH and donors (Addai and Gaere 2001). Key features of SWAp in Ghana include the agreement between MOH and donors (or “development partners” as now referred to in Ghana) on annual Programme of Work (POW) and the Common Management Agreement (CMA), which covers management of the donor-pooled funds and joint MOH-Partner annual and bi-annual reviews.
Since the inception of the SWAp reforms in the mid-1990s, donor inflows to the health sector have been well co-ordinated. This effective coordination of donor funds has led many analysts to describe the Ghanaian health system as one of the few in Africa with
37 There were no figures available for the World Bank in 2002. This does not mean that the World Bank did not contribute anything in 2002. 142 the capacity to absorb high level funding (Sachs 2005, Mensah et al. 2005) and appears to have increased donor confidence in providing support to the sector.
Under the SWAp initiative, the major donors pool their funds into a ‘common basket’ and agree with the MOH on programmes to be financed by the pooled funds. For example, all funding from the World Bank for health sector support is contributed to the pooled account. Other donors retain a small percentage for earmarked funding (see Chapter 9). The pooled funds are managed jointly by the government and donors under the Common Management Arrangement (CMA) between the MOH and donors (MOH 2001). There is also an agreement between the MOH and donors regarding how funds should be allocated. Section 7.2 provides details of how funds are allocated through the health system to achieve the stated objective of bridging inter-regional inequalities.
7.2 Allocation of Resources in the Health Sector Resource allocation in the Ghanaian health system is presently driven by a policy to shift resources to address inequities in health (Chapter 2). Three types of resource allocations and their implications for equity are examined in this section:
Allocation by line items Allocation by levels and Allocation by geographical area (regions and districts).
7.2.1 Resource Allocation by Line Items Under the current health sector programme of work (POW II), allocation of financial resources to the four main expenditure items: personnel emolument (salary), administration, service, and investment, is based in pre-determined proportions (POW 2002-2006). Table 7.3 shows actual allocations by expenditure items for 2003 and 2004.
143 Table 7. 3 Resource Allocation by Line Items: 2003-2004 Item/Year 2003 2004 ⊄’000 (%) ⊄’000 (%) Item 1: Salary 735,696 45.5 857,472 35.5 Item 2: Administration 171,311 10.6 239,399 9.9 Item 3: Service 647,300 40.1 808,563 33.5 Item 4: Investment 61,587 3.8 510,407 21.1 Total 1,615,894 100.0 2,415,841 100.0 Source: Based on Ministry of Health (2004) figures
In both years, salaries accounted for the largest proportion of health expenditures (about 46% in 2003 and 36% in 2004). The fact that salaries are only paid from government sources means that, about 84% of total government health expenditure of ⊄1,027 billion in 2004 (see Figure 7.2) went into payment of salaries. This leaves the service expenditure to be covered largely with donor funding and internally generated funds. Over half of the salary budget for 2004 went to payment of Additional Duty Hours Allowance (ADHA), which compensates health workers for the extra hours they work (MOH 2005). There was a significant increase in investments from 2003 to 2004. Figure 7.5 shows the distribution of health expenditure by line item in 2004.
Figure 7. 5 Allocation of Health Expenditure by Line Item - 2004
21% 36%
33% 10%
Salar y Administration Ser v ice Investment
Source: Based on MOH (2004) Data
144 7.2.2 Resource Allocation by Level of Health System Resource targets, in terms of the proportion that should be allocated to the various levels of the health system (national, regional and district) have been determined and clearly stated in the key health policy document, the Health Sector 5-Year Programme of Work 2002-2006 (GHS 2002). There are also targets in relation to the proportion of recurrent budget that should be allocated to tertiary institutions and other statutory bodies. Table 7.4 shows non-wage recurrent budget allocations determined by the current policy.
Table 7. 4 Health Sector Non-Wage Recurrent Budget Allocation: Percentage Distribution 2002-2005
Item/Year 2002 2003 2004 2005 Headquarters (MOH, GHS & Statutory bodies) 18.5 15.0 15.0 15.0 Tertiary institutions 18.1 20.0 20.0 20.0 Regional health service 19.9 23.0 23.0 23.0 District health service 43.5 42.0 42.0 42.0 Total 100.0 100.0 100.0 100.0 Source: POWII (January 2002).
About 60% of the non-wage recurrent budget is allocated to regional and district levels, with the largest proportion (over 40%) going to districts. The focus of this study is on the allocations to regional and district health services.
7.2.3 Allocation by Geographical Area While the proportion of funds to be shifted to the various levels of the health system is clearly stated, there is ambiguity with regards to how funds should be distributed between and within regions. In other words, there are no established proportions of resources to be shifted to deprived regions and districts. Information available in official documents only indicates that a top-slice of the GHS budget is taken for targeting the most deprived regions, namely: Northern, Upper East, Upper West, and Central Region, before the other allocations are made (Report of External Review Team 2003). There is no indication as to how much is ‘top-sliced’ and how these funds are shared among the four regions. Since allocation of the non-wage recurrent budget between and within
145 regions has been discussed extensively in Chapter 8, this section focuses on the allocation of the salary budget and internally generated funds (IGF).
7.2.3.1 Distribution of Salary Budget The allocation of salary budget in the Ghanaian health system is driven by personnel at post (Ensor et al. 2001; Interview Data 2004). Human resource development and management has been a difficult challenge in the Ghanaian health sector. Apart from staff shortages, there is inequitable distribution of personnel between and within regions. The bulk of the problem is currently blamed on emigration of health professional: Ghanaian doctors and nurses are increasing leaving the country for “greener pastures”. There were 293 Ghanaian doctors and 354 nurses were registered with British National Health Service [NHS] as of January 2004 (Mensah et al. 2005). In the United States of America, Hagopian et al. (2004) reported that of the 5334 doctors from sub-Saharan Africa working in the USA, 478 were from Ghana. Those remaining in the country are reluctant to take up positions in the regions and districts they perceive as deprived and this creates inter and intra-regional inequities. Table 7.5 shows the regional distribution of health professionals by region for 2002.
Table 7. 5 Regional Distribution of Ghana Health Service Doctors and Nurses 2002 Total Doctor/ Region Population Number Number Number Number of Population 2000 of Doctors of Nurses of Dentists Pharmacists Ratio Northern 1,820,806 28 287 0 9 1:65,029 Upper East 576,583 30 202 0 6 1:19,219 Upper West 920,089 10 189 1 5 1:92,009 Brong Ahafo 1,815,408 64 221 3 12 1:28,366 Volta 1,635,421 55 305 0 21 1:29,735 Ashanti 3,612,950 79 410 3 30 1:45,734 Eastern 2,106,696 85 647 5 18 1:24,785 Greater Accra 2,905,726 179 1333 8 26 1:16,233 Central 1,593,823 41 345 3 10 1:38,874 Western 1,924,577 58 353 4 15 1:33,182
Ghana 18,912,079 629 4,292 27 152 1:30,067 Note: Only doctors and nurses working in the public sector under the Ghana Health Service are represented here. Doctors working in the two Teaching Hospitals and the large private sector in Greater Accra and Ashanti Region are not included. Data were obtained from MOH 2003 and Ghana Statistical Service 2002.
146 Table 7.5 clearly shows that regions in the northern part of the country have limited access to health personnel compared to their counterparts in the southern sector. For example, the ratio of Ghana Health Service (GHS) doctors per population is 1:16,233 in the Greater Accra Region compared to 1:65,029 in the Northern Region and 1:92,009 in Upper West region (MOH 2003). If salaries are allocated in line with personnel at post, then logically, regions with low level personnel are missing out on the salary budget. Table 7.6 shows actual and per capita allocations of salary budget for 2002 and 2003.
Table 7. 6 Actual and Per capita Allocation of Salary Budget by Region: 2002 - 2003 Actual Salary Per capita Actual Salary Per capita Region 2002 Salary 2003 Salary ⊄’000 2002 ⊄’000 2003 Northern 7,550 4,146 29,889 16,415 Upper East 8,119 14,080 13,408 23,254 Upper West 6,691 7,272 8,090 8,793 Brong Ahafo 17,741 9,772 17,980 9,904 Volta 12,969 7,930 26,944 16,476 Ashanti 31,135 8,618 33,004 9,135 Eastern 14,113 6,699 26,883 12,761 Greater Accra 24,306 8,365 36,396 12,526 Central 8,921 5,597 19,207 12,051 Western 10,710 5,565 18,237 9,476
Total 142,254 7,522 230,038 12,164 Source: Based data obtained from Ministry of Health and Ghana Health Service (2003). Note that the salary budgets are in relation to salaries of all Ghana Health Service personnel in a region including doctors and nurses. The per capita allocations are in relation to the total regional population.
In absolute terms, allocation of the salary budget for 2002 and 2003 favoured the Greater Accra and Ashanti Regions. However, there was a sharp increase in allocations to the Northern, Eastern and Volta Regions (Figure 7.6). Personnel levels were high in the Volta and Eastern Regions; this may explain the increase in salary allocations. On the contrary, Northern Region had relatively fewer personnel; the sharp increase in salary allocation to the region may be attributed to payment of Additional Duty Hour Allowance (ADHA), which was captured in the salary budget. In per capita terms, the Upper East Region, by virtue of its low population, benefited more from allocation of the salary budget in 2002 and 2003 than any other region (Figure 7.6). The Upper West and Western Regions received relatively low allocations in absolute and per capita terms.
147 Figure 7. 6 Actual and Per capita Allocations of Salary Budget by Region: 2002-2003
40,000
35,000
30,000
25,000
20,000
15,000
Amount (Millions of Cedis) 10,000
5,000
0 Northern Upper Upper Brong Volta Ashanti Eastern Greater Central Western East West Ahafo Accra Region
Actual 2002 Actual 2003 Per capita 2002 Per capita 2003
7.2.3.1 Distribution of Internally Generated Funds (IGF) As mentioned in Section 7.1, under the current health sector programme of work (POW II 2002-2006), IGF revenue is expected to fund at least 14% of the total health budget annually over the five-year period (MOH 2002). This source of funding has several equity effects. Apart from the potential to impede access to health care for the poor, IGF could create imbalance inter and intra-regional resource allocation because the MOH takes no account of the amount generated by various jurisdictions in the resource allocation process (Interview Data 2004). Table 7.7 shows IGF revenue by regions for 2002.
Table 7. 7 Internally Generated Funds (IGF) by Region - 2002 IGF Revenue 2002 Per capita IGF Region ⊄’000 2002 Northern 5,141 2,823 Upper West 3,306 3,593 Upper East 4,307 7,470 Brong Ahafo 9,412 5,185 Volta 14,597 8,926 Ashanti 21,643 5,990 Eastern 17,526 8,319 Greater Accra 34,577 11,900 Central 10,662 6,690 Western 15,463 8,034 Total 136,634 7,225 Source: MOH 2003 148 It is evident from Table 7.7 that the regions in southern Ghana, particularly, the Greater Accra, Ashanti and Eastern generated substantial amount of IGF in 2002. Greater Accra generated the highest IGF revenue of about ⊄34,577 million. Conversely, the three regions in the north: Northern, Upper West and Upper East generated little revenue internally. The Upper West Region generated the smallest amount of IGF of about ⊄3,306 million. Within individual regions, there were variations in the amount of IGF generated by districts, with poorer districts generating less revenue internally than relatively richer districts (Interview Data 2004). Having shown how resources are allocated in the health system, the next section highlights the mechanisms for disbursing funds through the health system.
7.3 Mechanisms for Disbursing Funds Information from secondary sources indicates that health care financial management has undergone considerable reform in recent years. The core of this reform, which seeks to improve the way funds are disbursed and accessed within the health system, is the introduction of the budget and management centre (BMC) concept in 1996 (see Chapter 2). As indicated earlier, the entire MOH/GHS has been divided into over 350 BMC units. Funds available to the sector are managed through these units, which have been certified as having sufficient capacity to hold up and manage funds (Danida 2002; MOH 2002). Figure 7.7 shows how funds are disbursed through the health system.
149 Figure 7. 7 Funding Flow Chart
Source: World Bank 2003 (Modified using field data)
As explained earlier in Chapter 1, allocation to regions from the national level is in the form of a block grant, which regions redistribute to district BMCs using region-specific formulae. It is important to note that, at this stage, allocation is merely on paper in the form of budgetary ceilings. Actual disbursement of funds occurs after budgets and activity plans of BMCs (which are based on the ceiling allocations) have been approved by MOH headquarters. Section 7.4 provides insights into how approved budgets are disbursed and accessed.
7.3.1 Mechanisms for Disbursing Funds in Ashanti Region A close look at the mechanisms for inter-district resource allocation from 1998 to 2002 in the Ashanti Region reveals a tendency to distribute resources equally across-the- board rather than on the basis of differential health needs. As much as 80% of the total funding for the 18 District Health Administrations (DHAs) in the region was shared equally across-the-board. That is, 80% of the budget was divided into 18 equal parts for the 18 DHAs regardless of differences in health needs. The remaining 20% was 150 distributed according to pre-determined indicators that reflected the differences of need among districts based on staff strength, rurality (distance from the regional capital), size of district and numbers of sub-districts (RHA - Kumasi, 2003). The Regional Health Administration (RHA) had no data on mortality levels for the various districts in the region; hence, there was no standard mortality ratio (SMR) in the allocation formula. For Sub-district Health Administrations (SDHAs), 50% of resources were divided equally across-the-board while 50% is shared according to the number of sub-districts (Table 7.8).
Table 7. 8 Resource Allocation Mechanisms for the Ashanti Region 1998-2001
Criteria Score % Factor District Health Administration Even 80 18 Staff strength Distance from regional capital 20 -- Size of district Number of sub-districts Sub-District Health Administration Even 50 18 Number of Sub-districts 50 89 Source: Ashanti Regional Health Administration, 2003
The study could not establish the weights or factors assigned to the various indicators making up the 20%. Officials of the RHA could not provide any information as to how this 20% was apportioned. This may be due to poor record keeping and the fact that most of the officials were new in their positions as regional managers.
7.3.2 Mechanisms for Disbursing Funds in Northern Region The Northern Region allocates resources on a relatively more refined set of criteria than the Ashanti Region. It allocates 50% of funding equally across-the-board, 20% according to the number of sub-districts, and 10% each according to the size of the district (landmass), distance from regional capital, and level of deprivation (Northern Regional Health Administration 2003, Table 7.9). Deprivation is defined in terms of the number of hard-to-reach communities, locally referred to as overseas communities, in a district. Compared to the Ashanti Region, deprivation is accounted for much better in 151 the Northern Region. Ashanti has districts such as the Sekyere East and Asante-Akim North with vast overseas areas, however, the RHA has no specified indicator for deprivation in the resource allocation mechanism.
Table 7. 9 Resource Allocation Mechanisms for the Northern Region 1999-2002 Criteria Score % Factor District Health Administration Even 50 13 Number of Sub-districts 20 70 Size of district 10 26 Distance from regional capital 10 27 Deprived/Overseas 10 11
Sub-District Health Administration Even 40 13 Number of Sub-districts 30 100 Size of district 10 26 Number of health facilities 10 108 Deprived/Overseas 10 11 Total 100 - Source: Northern Regional Health Administration, 2003
7.4 Mechanisms for Accessing Funds The way funds are assessed in the health system has significant implications for the amount of resources available to districts at any given time. Interviews with district managers revealed that the mechanism for accessing funds was a major concern (see Chapter 9). Data from key documents suggest that all BMCs operate at least three bank accounts for Government of Ghana (GOG) releases, donor-pooled funds and internally generated funds (IGF). Some district BMCs also have a special account for funds received directly from other sources, such as the District Assembly (local government). At the national level, the MOH operates a United States dollar account into which all donor contributions from those donors participating in the pooled arrangements are paid, a cedi account into which transfers from the dollar account are paid, and a second cedi account used for the funds provided by GOG to pay for service a item within a particular POW (World Bank 2003).
The mechanism for disbursing funds to BMCs from MOH headquarters depends on the expenditure item. Information from official documents indicates that for GOG item 3
152 (service), the MOH/GHS headquarters accesses the money in bulk from the government and disburses this to BMCs in the form of a cheque. A similar procedure of disbursing funds directly to BMCs in the form of cheques is in place for donor-pooled fund. Table 7.8 shows the mode of disbursement and methods of accessing the various expenditure items.
Table 7. 10 Mode of Disbursing and Accessing of Funds
Item Disbursement Form Method of Accessing by District BMCs
GOG Item 1 (Salaries) Through the Banks Staff access monthly through their banks
GOG Item 2 (Administration) Expenditure Warrants Accessed through the District Assembly
GOG Item 3 (Service) Cheques Pick up from the Regional Health Admin.
GOG Item 4 (Investments) MOH/MOF* decides Headquarters decides
Donor Pooled Funds Cheques Pick up from the Regional Health Admin. Source: Based on secondary data from the field. *MOF stands for Ministry of Finance
In most cases, the cheques are sent to BMCs from the headquarters through the Regional Health Administration (RHA), which informs the various district managers to collect them. Sometimes the region is by-passed and cheques are either sent directly to district BMCs or BMCs pick the cheques up from headquarters in Accra.
Unlike the DPF and Service (Item 3), funds for Administration (item 2) are disbursed to BMCs by the Ministry of Finance through district treasuries in the form of expenditure warrants. The procedure for accessing this fund (GOG 2) is not straightforward and involves obtaining expenditure authorisation from the Department of the Accountant General before the money can be spent. The processing of the warrants is supposed to be done through the district treasuries, but often, according to the information obtained from the interviews, accountants have to travel from their respective districts to the national capital (Accra) to follow up the process. Funding for GOG item 1 (salaries) is disbursed through a commercial banking system to all health personnel while investment funds (GOG item 4) are allocated directly by the Ministry of Finance in accordance with investment plans and procurement orders developed by the MOH. With evidence on how funds are disbursed through the health system and accessed by BMCs, the next section details how BMCs account for what they receive.
153 7.5 System of Financial Accountability Financial management and accountability have been touted as one of the most successful elements of the health sector reform in Ghana. With the implementation of the BMC concept, control of funds and accountability are no longer centralised or entrusted to a few project managers and implementing agents. All managing BMCs receive, manage and account for funds. Standardised Accounting, Treasury and Financial Rules and Instructions (ATF Rules) in a procedural manual sets out the system of accounting and reporting within the health sector.
Currently, all BMCs use the same set of books for recording accounting transactions for all sources of funds (GHS 2002). BMCs report on revenue and expenditure activity in accordance with a standard format developed by the office of the Financial Controller (MOH) on a monthly basis. A clear timetable exists for all BMCs to know when exactly their financial reports are due. Reports are submitted to the next higher level in the BMC hierarchy; for example, a sub-district BMC must submit report to a district BMC, which in turn submits a combined report, including that of the sub-district to a regional BMC.
Monitoring teams ensure that BMCs financial reports are complete, accurate and submitted on time. These teams review reports for completeness and accuracy, investigate possible causes for late submissions, and provide feedback and training to BMC staff in the report error correction process. They also verify BMCs’ reported information by tracing reported figures back to cash-books and ledgers, main ledgers back to subsidiary ledgers, main/summary cash-books to subsidiary cash-book, cash- books to the bank reconciliation, and all books to supporting documentation (Financial Management System Review 2001).
To enhance accountability at the service delivery level, an internal audit unit (IAD) has been set up within the MOH and GHS to, among other things, ensure that policy standards and procedures are complied. The task of the IAD is to oversee the adequacy of the existing control mechanisms and provide reasonable assurance that resources are being used prudently with due regard to efficiency. Information obtained from the field suggests that the introduction of the ATF Rules has considerably improved financial accountability in the health sector. However, it was observed that the way funds are
154 currently accounted for gives little indication as to whether funds are being used effectively and efficiently.
The system of accountability requires BMCs to account mainly for funds received. Thus, BMCs provide information on a pre-designed format about the total revenue and expenditure. The issue of judicious use of funds (whether the way funds are used constitute the best possible way those funds could have been used) is not considered. In effect, the accountability system makes no demands on BMCs to account for the reasonableness of the decisions they make with regards to application of funds. As evident from interviews with district and some regional officials, timely submission of reports and accuracy in reporting are the main conditions BMCs strive to meet in accounting for funds received. Chapter 9 provides further insights from the qualitative study.
Chapter 7. Summary of key points The Ghanaian health system is funded through four major funding sources: government, donors, commercial credits and internally generated funds (IGF).
Traditionally, allocation from government through the MOH has constituted the largest funding source for the health sector. In 2001, government funding through the MOH constituted about 51% of total health expenditure.
Donor funding is a significant component of the health budget in Ghana. It covers about 35% of total health expenditure under the current health sector Programme of Work (POW 2002-2006). In 2004, donor funding (⊄1,034 billion cedis) was slightly higher than funding from government sources (⊄1,027 billion).
Debt relief savings under the HIPC initiative are expected to provide extra funding for the health system. In 2004, HIPC funds accounted for about ⊄104 billion cedis, which was about 4% of total health expenditure.
The four main expenditure items in the Ghanaian health system are salary (item 1), administration (item 2), service (item 3) and investment (item 4).
Salaries account for the largest proportion of health expenditure in Ghana. In 2004, they accounted for about 36% of total health expenditure. This was about 84% of the total government health expenditure.
Services and investments are mostly paid for with donor money and internally generated funds (IGF).
IGF is not taken into account in the resource allocation process despite the significant difference in the level of these funds generated across jurisdictions.
155 CHAPTER 8
GEOGRAPHICAL EQUITY IN RESOURCE ALLOCATION IN GHANA: QUANTITATIVE RESULTS
Overview This chapter presents the quantitative analyses, which examined whether resources allocated in the Ghanaian health system were equitable in terms of differentially benefiting the worst-off regions, districts and sub-districts. The chapter is divided into three main sections. The first presents an inter-regional analysis of equity in allocation of funds from the national to regional levels. The second, and primary focus of this study, provides an analysis of equity in funding allocation from regions to districts. This focuses specifically on the allocation of funding in the Ashanti and Northern Regions. The final section deals with allocation of funding from districts to sub-districts.
SECTION 1: INTER-REGIONAL EQUITY IN RESOURCE ALLOCATION This section presents the results of the inter-regional analysis of equity in resource allocation. Deprivation among the ten regions was assessed using principal component analysis (PCA). The output of the PCA was used to construct a general index of deprivation (GID), which was used to develop an equity-adjusted share (EAS) used in this study as a benchmark for assessing inter-regional equity in funding. A simple regression analysis was undertaken to explore the relationship between the EAS and the actual share of funds received by regions. Finally, a comparison between predicted and actual allocations was undertaken to highlight the equity-gap (see Chapter 6).
8.1 Deprivation among Regions The general index of deprivation (GID) developed through the principal component analysis (PCA) was the main indicator used for assessing levels of relative deprivation among regions. An additional index of deprivation, labelled in this study as the double- index of deprivation (DID), was developed and used to shed further light on the extent
156 and variations in deprivation among regions in Ghana (see Chapter 6). A similar approach was adopted by McIntyre et al. (2002) to analyse geographic patterns of deprivation in South Africa. In that study, the authors constructed a single-index of deprivation (SID) to complement the composite and policy-perspective indices.
Although a single-index of deprivation is likely to be simpler and easier to use, an assumption was made in this study that a deprivation index based on two variables would be equally simple and much more potent in identifying deprivation among regions than a single variable index. The picture of deprivation among the regions as emerged from applying each of the two indices GID and DID is first presented separately before being analysed together.
8.1.1 Principal Component Analysis and GID Principal component analysis (PCA) is a statistical technique that identifies linear combinations of random or statistical variables that maximises variance (Chapter 6; see also Kirkwood et al. 1999). There was good correlation among the variables used in this study at 1% significant level in Pearson’s correlations (see Appendix A). The coefficients were much higher for the socio-economic variables. However, the demographic variables, in particular, the proportion of female and children under-5, correlated well with most of the socio-economic variables. The only variable that showed poor association with others was overcrowding. This may be due to the fact that overcrowding is much more of an urban phenomenon in Ghana, whereas most of the other variables in this study that reflect disadvantage were rural in orientation. Since the use of PCA requires that variables correlate strongly with each other (Kirkwood 1998), overcrowding was excluded from the analysis.
The PCA generated a three-component matrix as illustrated by Table 8. 1. Most of the variables retained more than 70% loadings in the first component, particularly the socio- economic variables. All variables with loadings of 50% or more (in bold) are considered important in PCA, however, only those in the first component matter as far as the calculation of the deprivation index is concerned. This is because they always account for the largest proportion of the variance.
157 Table 8. 1 Component Matrix: All Regions Component Variables 1 2 3 No-Electricity 0.981 -0.053 0.001 Mud-House 0.976 -0.174 -0.058 Rural 0.967 -0.001 -0.176 No-Pipe 0.932 -0.090 -0.176 Wood-Fuel 0.910 -0.133 -0.386 No-Education 0.850 -0.321 0.389 Under-5 0.766 -0.399 -0.241 Female 0.742 0.534 0.120 No-Toilet 0.671 -0.269 0.625 Elderly 0.407 0.867 0.109 Disabled 0.556 0.763 -0.171 Unemployed 0.148 0.098 0.927 % of Total Variance Explained 61.279 16.839 14.413 Note: The variables in each component with more than 50% loading together explain a particular dimension of deprivation
The first component accounted for over 61% of the total variance and explains an aspect of deprivation that is largely socio-economic. The second component accounted for about 17% of the entire variance and highlights a different facet of deprivation driven by being female, elderly and disabled (or elderly female who are disabled). Finally, the last component explains a type of deprivation driven by being unemployed in a household with no access to toilet facilities. This form of deprivation accounted for about 14% of the total variance. In this study, disability retained a loading of more than 50% in the first component and was supposed to be included in the set of variables used for the calculation of the GID. However, it was excluded because of its strong association with the variables in the second component.
The component score coefficients produced by the PCA were used as variable weights in developing the general index of deprivation (GID). Table 8.2 shows the coefficients of the important variables in the first component, which were used to develop the GID. The variables have been ranked according to their weights. The weight of a variable reflects its relative contribution to overall deprivation. Two demographic variables, children under-5 and female emerged with the highest and the lowest weights respectively.
158 Table 8. 2 Variables by Weight/Coefficient Variable Weight / Coefficient
Under-5 0.205 Wood-Fuel 0.198 Mud-House 0.159 No-Pipe 0.158 Rural 0.146 No-Electricity 0.127 No-Education 0.096 No-Toilet 0.026 Female -0.028
Key: Variable weights are component score coefficients produced by the PCA. Only coefficients of variables with 50% or more loadings in the first component were used.
Some of the variables retained unexpectedly high weights. Wood-fuel and mud-house, for example, were not expected to retain weights higher than others like no-pipe and no- electricity given the importance of access to water and electricity as markers of areal deprivation. However, they maintained the second and third highest weightings. The proportion of the population over 18 years and without education (i.e., no-education) was expected to retain a higher weighting, considering the influence of education on poverty and disadvantage but it emerged with the third lowest weight.
The GID was derived by summing up each variable’s standardised score (z-score) multiplied by its weight or component score coefficient (see Chapter 6). Table 8.3 shows the ranking of regions by their GID scores.
Table 8. 3 Regions by General Index of Deprivation Region GID Northern 1.026 Upper West 0.815 Upper East 0.795 Brong-Ahafo 0.381 Volta 0.112 Western 0.053 Eastern -0.061 Central -0.139 Ashanti -0.411 Greater Accra -2.573 Key: The higher the index value the greater the deprivation level in the region.
159 The region that emerged as the least deprived was the Greater-Accra with a high negative index of –2.573. The Ashanti and Central Regions followed with the second and third lowest indices (-0.411 and -0.139 respectively). The three regions found to have the highest level of deprivation were the Northern, Upper West and Upper East. All poverty-related studies in Ghana including the Ghana Living Standard Surveys (GLSS 1993; 1998) have consistently emphasised the high level of poverty and deprivation in these regions.
8.1.2 The Double-Index of Deprivation (DID) Bureaucrats and policy makers are more inclined to implement proposals that are self- evident and less time-consuming than complex ones. Walt (1994) observed that if a policy has relatively simple technical features, it is easier to introduce than one which is complex. The double-index of deprivation (DID) was constructed purposely to explore whether two of the variables used in this study could as effectively capture relative deprivation among regions as the general index of deprivation (GID), thereby providing a more simplified tool for identification of deprived regions. The two variables used here were no-education and no-pipe. These variables were selected not necessarily because of their weights in the PCA, but largely because of their socio-economic and health importance in Ghana. Thus, they were selected deliberately to highlight the importance of education and access to piped water in the socio-economic development of Ghana. The z-scores of these variables for each region were added to generate the DID (Table 8.4)
Table 8. 4 Regions by Double-Index of Deprivation (DID) Region DID Upper East 2.147 Upper West 2.078 Northern 1.955 Brong Ahafo 0.427 Volta 0.119 Western -0.303 Eastern -0.480 Ashanti -0.802 Central -1.139 Greater Accra -4.003 Key: The higher the index value the greater the level of deprivation
160 The results clearly show a pattern of relative deprivation similar to that obtained from the GID analysis. The three regions in northern Ghana (Upper East, Upper West and Northern) maintained the highest level of deprivation. However, unlike the GID with which the Northern Region came on top as the most deprived, the Upper East retained the highest DID score ahead of the Upper West as shown by table 8.5. Northern Region obtained the third highest index of 1.955.
On the whole, the differences between the three regions in the north in terms of DID scores were relatively small. The three least deprived regions were again found to be Greater Accra, Central and Ashanti. These same regions emerged as the least deprived in the GID analysis. The only difference here is that, while the Ashanti Region was ranked the second least deprived in terms of GID, it obtained the third lowest deprivation score of -0.802. This finding questions the prevailing view that Central Region is the fourth most deprived region in Ghana. The positions of Brong Ahafo, Volta, Western and Eastern Regions did not change between the GID and DID. Table 8.5 shows the regions and the two indices of GID and DID.
Table 8. 5Regions by General Index of Deprivation (GID) and Double-Index of Deprivation (DID) Region GID DID Northern 1.026 1.955 Upper West 0.815 2.078 Upper East 0.795 2.147 Brong-Ahafo 0.381 0.427 Volta 0.112 0.119 Western 0.053 -0.303 Eastern -0.061 -0.480 Central -0.139 -1.139 Ashanti -0.411 -0.802 Greater Accra -2.573 -4.003 Key: The higher the index value the greater the level of deprivation
Figure 8.1 shows the level of deprivation across regions in terms of the two indices constructed in this study. In general, the two deprivation indices: GID and DID, point to the same direction but differ in magnitude. The DID had higher scores than the GID.
161 Figure 8. 1 Regions by Different Deprivation Indices
5.00
4.00
3.00
2.00
1.00
0.00
-1.00
Deprivation Score -2.00
-3.00 GID -4.00 DID
-5.00
t o a i s f lt al a est o tern E W ha V hant rthern A s Accra es o per g A Eastern er Centr W N p pper U U reat Bron G Region
Key: Negative scores mean low deprivation while positive means high level of deprivation.
8.1.3 Correlation between Deprivation Indices and Mortality Rates Infant and under-5 mortality rates are two of the indicators most often used by the United Nations to assess a country’s health status (Kirkpatrick 1997). Although in developing countries, these indicators are largely viewed as unreliable due to under- reporting of deaths, they remain the best health indicators available. Correlation between the two deprivation indices and infant and under-5 mortality rates of regions was explored as a way of validating the indices. Table 8.6 shows the results obtained in Pearson’s correlation.
Table 8. 6 Pearson’s Correlation Coefficients of Deprivation Indices and Regional Morality Rates VARIABLE IMR U5MR GID DID IMR 1.000 U5MR 0.862** 1.000 GID 0.655* 0.811** 1.000 DID 0.594 0.820** 0.960** 1.000 **Correlation is significant at the 0.01 level (2-tailed). *Correlation is significant at the 0.05 level (2-tailed). IMR = Infant Mortality Rate U5MR = Under-5 Mortality Rate GID = General Index of Deprivation DID = Double-Index of Deprivation
162 There were statistically significant relationships between the mortality figure, particularly under-5 mortality and the deprivation indices (0.811** with GID and 0.820** with DID). The infant mortality rates did not correlate as highly with the deprivation indices (0.655* with GID and 0.594 with DID) as the under-5 mortality rates. In general, the statistically significant relationships between the mortality rates and the deprivation indices suggest that the indices developed in this study had some validity, particularly criterion validity38. This does not, however, suggest that there is no need for a deprivation index where mortality data is available. The key point here is that, infant and under-5 mortality are considered internationally as key indicators of health need. One would expect a good proxy measure of health need, such as the deprivation index derived in this study, to correlate well with these mortality rates.
8.2 The Equity-Adjusted Share (EAS) The equity-adjusted share (EAS) was the yardstick against which progress towards equity in resource allocation between and within regions was assessed in this study. The main assumption underpinning the EAS is that for resource allocation to be equitable, it should reflect relative deprivation and differential population sizes. Thus, the EAS is the predicted equity share of resources that should go to each region, given its level of deprivation and population size.
The EAS was developed using only the general index of deprivation (GID) for two main reasons. First, the GID was constructed mainly as a means of developing an equity benchmark against which fairness in resource allocation in the Ghanaian health system could be assessed. Second, the GID entails a range of variables that capture deprivation from different dimensions making it a more reliable index than the DID.
The EAS was derived by weighting the total regional populations using the GID. All the GIDs were first normalised to convert the negative values to positive following McIntyre et al. (2000, 2002; see also Chapter 6). Table 8.7 shows how the percentage equity-adjusted share (EAS) was derived for each region. The Ashanti Region retained the highest EAS of 18.3% despite having the second lowest GID. This situation resulted
38 Assessment of criterion validity involves comparing a measure against an existing benchmark measure (de Vaus 2001).
163 from the large population size of Ashanti, which has a total population of about 3.6 million people or nearly one-fifth of Ghana’s population.
Table 8. 7 Derivation of the Equity-Adjusted Share Index (EAS) for Inter-Regional Resource Allocation in Ghana Region GID NGID Total Pop W/Pop EAS% Northern 1.026 4.599 1,820,806 8,373,251 13.4 Upper East 0.795 4.368 920,089 4,018,974 6.4 Upper West 0.815 4.388 576,583 2,530,215 4.1 Brong Ahafo 0.381 3.954 1,815,408 7,178,930 11.5 Volta 0.112 3.685 1,635,421 6,026,779 9.7 Ashanti -0.411 3.162 3,612,950 11,424,067 18.3 Eastern -0.061 3.512 2,106,696 7,399,512 11.9 Greater Accra -2.573 1.000 2,905,726 2,906,768 4.7 Central -0.139 3.434 1,593,823 5,473,683 8.8 Western 0.053 3.626 1,924,577 6,979,297 11.2 Total 18,912,079 62,311,476 100.0 Key: NGID = normalised GID. W/Pop = weighted population which is the total population multiplied by the normalised deprivation index.
Northern Region, with the highest GID (the most deprived region), retained the second highest EAS of 13.4%. Three regions, Western, Brong Ahafo and Eastern had EAS ranging from 11.2 to 11.9%. These regions have population sizes within the ranges of around 1.8 to 2.1 million people. The lowest EAS was retained by the Upper West Region, which has the smallest population in Ghana. As the second most deprived region in Ghana (according to the GID), the EAS of 4.1% for the Upper West Region may on the surface appear quite low. However, this is a relatively large share compared to a region like Greater Accra, which has about five times as much population (about 2.9 million –the second highest in Ghana) but retained only 4.7% EAS. The main point here is that the EAS is driven not only by population size or GID but by a combination of both.
8.3 Assessing Inter-Regional Equity in Resource Allocation The primary objective of this section was to determine whether actual funding allocations from national to regional levels were equitable in terms of differentially benefiting the most deprived regions. This was examined using two methods. First, a simple regression analysis was used to determine the proportion of variance in the
164 actual share of funds received by regions that could be explained by the equity-adjusted share (EAS). The EAS percentage was ‘regressed’ as an independent variable on percentage of actual share of funds as a dependent variable. The greater the proportion of variance accounted for by the EAS, the greater the equity in funding allocation (see Chapter 6). Although, population is a key component of the EAS, its influence on actual share of resources was assessed separately to further assess the extent to which resource allocation has been on per capita basis. Second, the total yearly allocations to regions were re-distributed using the EAS to derive the predicted share of total resources that should go to each region given its level of deprivation and population size. These predicted shares were compared with the actual shares to determine the equity-gap, which is the difference between the two shares.
8.3.1 Actual Allocation of Government Funds Actual funding allocations to each of the ten regions over the five-year study period (1998-2002) were sought in order to assess whether allocation was equitable. However, the only data available were allocation from government sources for administration, service and investment expenditures (GOG 2-4) for 2002 and 2003. Although 2003 was not within the five-year time frame set in the study, given the unavailability of expenditure data from 1998 to 2001, the 2003 data was used. Table 8.8 shows the actual and percentage shares of government funds (GOG 2-4) allocated to each of the ten regions as well as percentage of funding difference for 2002 and 2003.
165 Table 8. 8 Region by Actual and Percentage Shares of Government Funds (GOG 2-4): 2002-2003 Actual Funding Actual Funding Percentage of 2002 2003 Funding Region ⊄Bn (%) ⊄Bn (%) Difference 2002/2003 Northern 13,297 6.9 18,655 10.0 45.1 Upper East 13,760 8.1 15,361 8.2 15.4 Upper West 11,839 6.1 12,630 6.8 10.3 Brong Ahafo 21,267 11.0 17,309 9.3 -15.8 Volta 18,134 9.4 18,158 9.7 3.5 Ashanti 32,809 18.0 28,788 15.4 -9.3 Eastern 22,187 11.5 24,370 13.1 13.6 Greater Accra 28,574 14.8 15,735 8.4 -43.1 Central 13,552 8.0 18,063 9.7 38.8 Western 17,449 9.0 17,429 9.3 3.3
Total 192,868 100.0 186,498 100.0 Source: GOG 2-4 for 2003 obtained from GHS Headquarters. Data for 2002 were obtained from the MOH Headquarters. No data were available for 1998 to 2001. ⊄Bn = Billions of Ghanaian Cedis
Table 8.8 shows that there were shifts in funding allocation in favour of the three Northern Regions (Northern, Upper East and Upper West) and the Central Region from 2002 to 2003. The big winners were the Northern and Central Regions, which witnessed about 45.1% and 38.8% rise in allocation from 2002 to 2003. The Greater Accra, Ashanti and Brong Ahafo Regions were the main losers. Greater Accra saw a substantial drop in allocation from 14.8% in 2002 to 8.4% in 2003. This represented about 43.1% fall in actual allocation in a single year.
8.3.2 Regression Analysis The regression analysis (Section 8.3) was used to determine the proportion of variance in the actual share of funds that could be explained by the EAS. Table 8.9 shows a significant correlation between the EAS and proportion of actual GOG received by regions in 2003. Nearly 75% of the variability in the share of funds could be predicted from the EAS (p<0.001). There was also a significant relationship between population size and share of funds (p = 0.013). No significant relationship existed between the EAS and actual share of funds for 2002 (p = 0.196), but population size was a crucial factor. The proportion of variance in the share of funds that could be explained by population size in 2002 was almost 85%. In summary, the results of the regression analysis indicate
166 that inter-regional resource allocation was largely equitable in 2003 but less equitable in 2002.
Table 8. 9 Regression of Equity-Adjusted Share and Population on Percentage Share of Government of Ghana (GOG) Funds: 2002 - 2003 Unstandardised Independent Variable R-Square Coefficient Sig. (P-Value) GOG: 2002 EAS 0.199 0.378 0.196 Pop% 0.848 0.736 <0. 001**
GOG: 2003 EAS 0.747 0.503 < 0.001** Pop% 0.555 0.409 0.013*
Key: *Coefficient and P-Values significant at 5%. **Coefficient and P-Values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%). Dependent Variable = Percentage Share of Actual Funds
8.3.3 Comparison of Actual and Predicted Allocations The second approach used to assess whether there was equity in the inter-regional resource allocation was to compare the actual share of funds received by each region with the predicted (EAS-based) shares for each year (Section 8.3). The actual allocations have already been presented in Table 8.9 and are not repeated here. Table 8.10 shows the predicted (EAS-based) shares and the percentage difference between the actual and predicted shares for 2002 and 2003.
167 Table 8. 10 Regions by Predicted (EAS-Based) Government Funding (GOG 2-4) and Percentage of Funding Difference 2002 - 2003 Predicted Predicted Percentage of Percentage of Funding Funding Funding Funding Region 2002 2003 Difference Difference ⊄ Bn ⊄Bn 2002 2003 Northern 25,844 24,991 -48.5 -25.4 Upper East 12,344 11,936 11.5 28.7 Upper West 7,908 7,646 49.7 65.2 Brong Ahafo 22,180 21,447 -4.1 -19.3 Volta 18,708 18,090 -3.1 0.4 Ashanti 35,295 34,129 -8.0 -15.6 Eastern 22,951 22,193 -3.3 9.8 Greater Accra 9,065 8,765 215.2 79.5 Central 16,972 16,412 -20.2 10.1 Western 21,601 20,888 -19.2 -16.6
Total 192,868 186,498 Note: ⊄Bn = Billions of Ghanaian Cedis. Percentage of funding difference is the difference between the actual and predicted allocations expressed as a percentage. It was calculated by dividing the funding difference by the predicted allocation and multiplying by 100.
In spite of the drastic cut in funding to the Greater Accra Region in 2003, it still received nearly 7 billion cedis in excess of its predicted (EAS-based) share. However, this was almost a three-fold reduction from the 19.5 billion cedis it received over the predicted shares in 2002. The Northern Region, on the other hand, was greatly under- funded to the tune of about 6.3 billion cedis in 2003 despite a 45% increase in funding. The Upper East and Upper West Regions were allocated more than their predicted EAS- based shares in both years, while allocation to Ashanti and Brong-Ahafo fell short of their predicted equity levels. In general, the comparison of actual and predicted allocations shows some redistribution of resources from richer regions to poorer regions. The 2003 allocations largely show an improvement in inter-regional equity.
The proportion of funding above and below the predicted (EAS-based) allocations for 2002 and 2003 is presented in Figure 8.2. The Central and Northern Regions benefited the most from changes in actual allocation of resources. However, in terms of equity, Greater-Accra, whose allocation was almost halved in a single year, still received more than ‘equitable’ (EAS-based) allocation, while Northern Region remained under- funded, though, to a lesser degree in 2003. Upper West Region gained excess funding both in 2002 and 2003.
168 Figure 8. 2 Percentage Difference between Actual and Predicted (EAS-Based) Allocations: 2002 and 2003
250.0
200.0
150.0
100.0 2002 2003 50.0
0.0
-50.0
-100.0
t l s fo ti n a n ern est a Ea rth W Volta aster o er er Ashan E Centra N p ter Accr Wester p ong Ah a U Upp Br Gre Region
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
Chapter 8, Section 1. Summary of key points
Deprivation is greatest in the three regions in the north of Ghana: Northern, Upper East and Upper West.
Regional under-5 mortality rates correlate highly with the general index of deprivation (GID) and the double-index of deprivation (DID). Infant mortality did not correlate well with the indices.
Funds were redistributed in 2002 to 2003 from richer regions, mainly Greater Accra, Ashanti and Brong Ahafo, to the most deprived regions in the north and the Central Region. Allocations to Northern and Central Regions increased by about 45% and 37% in 2002 and 2003 respectively. Allocation to the Greater Accra Region, on the contrary, dropped very substantially from 14.8% of the total budget in 2002 to 8.4% in 2003 representing about 43% reduction in a single year.
The shifts in funding have largely improved equity in resource allocation based on the EAS. There was significant correlation between the equity-adjusted shares and actual share of funds received by regions in 2003.
Population size was a key factor in inter-regional resource allocation in 2002, but less so in 2003.
169 SECTION 2: INTER-DISTRICT EQUITY IN RESOURCE ALLOCATION This section presents the results of the analysis of equity in resource allocation from regional to district levels. It focuses on the two case regions – Ashanti and Northern. Deprivation among the districts in the two regions was assessed together so that cross- regional comparison of deprivation among districts could be made. The results were used to develop the district equity-adjusted shares, which in turn were used as yardsticks against which equity in inter-districts resource allocation was assessed. Unlike the inter- regional analysis in Section 1, which analysed only data on government funding for 2002 and 2003, this inter-districts study was more comprehensive, analysing both government funds for administration and service expenditures and donor-pooled funds from 1998 to 2002. There were, however, some gaps in the data for Northern Region.
8.4 Deprivation among Districts in Ashanti and Northern Regions Deprivation across the 31 districts in the Ashanti and Northern Regions was assessed using the general index of deprivation (GID) developed through a combined regional principal component analysis (PCA). The same variables used for the inter-regional analysis were used for the inter-districts study. Correlation between the variables was explored using Pearson correlations prior to undertaking the PCA. There were strong correlations between most of the variables (see Appendix B).
Table 8. 11Component Matrix: Northern and Ashanti Region Component Variables 1 2 3 No-Electricity 0.932 0.234 -0.064 Under-5 0.844 0.114 0.253 Rural 0.818 0.445 -0.198 Mud-house 0.795 -0.204 0.075 No-Education 0.792 -0.496 0.203 Wood-fuel 0.770 0.416 -0.291 No-Toilet 0.753 -0.533 0.254 No-Pipe 0.749 0.519 -0.077 Female 0.714 0.023 0.022 Elderly -0.327 0.863 0.147 Disabled -0.229 0.721 0.361 Unemployed -0.008 0.106 0.925 % Total Variance Explained 49.171 21.289 10.967 Key: Important variables in the component matrix are those with loadings of 50% or more (in bold).
170 The PCA generated a three-component matrix (Table 8.11), with the socio-economic variables retaining high loadings in the first component. Only the variables in the first component, which accounted for about 49% of the total variance, were used to develop the general deprivation index (GID). They explain an aspect of deprivation driven largely by socio-economic factors as well as being a child under five years of age and a female. Variables in the second component shed light on a form of deprivation driven by being elderly, disabled and living in households with no access to piped water. This component accounted for about 21% of the total variance. The final component was almost entirely driven by unemployment and accounted for about 11% of the total variance.
The component score coefficients produced from the PCA were used as variable weights (see Table 8.12). Proportions of the population living in rural areas, using wood-fuel as the main source of cooking energy and those without access to piped water retained the highest weights. Proportions of children under-5 and of populations living in houses with outer wall made of mud, which were among the three highly weighted variables in the inter-regional analysis (Section 1), did not retain such high weightings in this inter-districts PCA. In contrast, proportion female, which obtained the overall lowest weight in the inter-regional study, was weighted higher in this inter-district study.
Table 8. 12 Variables by Weight/Coefficient Variable Weight / Coefficient Rural 0.224 Wood-fuel 0.220 No-Pipe 0.216 No-Electricity 0.187 Under-5 0.121 Female 0.107 Mud-house 0.070 No-Education 0.001 No-Toilet -0.016 Key: Variable weights are component score coefficients produced by the regression analysis. Only coefficients of variables with 50% or more loadings in the first component were used.
As in the inter-regional study (Section 8.1.2), the district GIDs were derived by multiplying the weights of the variables by their standardised scores (z-score) and
171 summing up the results. The district GID allows the ranking of the 31 districts according to relative deprivation, shown in Table 8.13. The districts were ranked, and then categorised according to high, medium and low level of deprivation.
Table 8. 13 Districts in Ashanti (AR) and Northern Region (NR) Ranked by their General Index of Deprivation (GID)
District Region GID Rank Amansie West (AWD) AR 0.894 1 Adansi East (ADED) AR 0.804 2 Gushiegu Karaga (GKD) NR 0.778 3 Ahafo-Ano South (ANSD) AR 0.737 4 Amansie East (AED) AR 0.732 5 Asante-Akim South (AASD) AR 0.694 6 Zabzugu Tatale (ZTD) NR 0.620 7 Saboba Chereponi (SCD) NR 0.487 8 Bosomtwe-Atwima Kwanwoma (BAK) AR 0.485 9 West Mamprusi (WMD) NR 0.426 10 West Gonja (WGD) NR 0.397 11 Nanumba (NMD) NR 0.374 12 Sekyere East (SED) AR 0.331 13 Bole (BD) NR 0.313 14 Ahafo-Ano North (ANND) AR 0.288 15 Afigya Sekyere (AFSD) AR 0.261 16 East Mamprusi (EMD) NR 0.214 17 Ejisu-Juaben (EJD) AR 0.180 18 Kwabre (KD) AR 0.134 19 East Gonja (EG) NR 0.108 20 Tolon-Kumbungu (TKD) NR 0.091 21 Atwima (AD) AR -0.136 22 Sekyere West (SWD) AR -0.156 23 Offinso (FD) AR -0.224 24 Savelugu-Nanton (SND) NR -0.228 25 Ejura-Sekodumasi (ESD) AR -0.424 26 Yendi (YD) NR -0.587 27 Asante-Akim North (AAND) AR -0.616 28 Adansi West (ADWD) AR -1.079 29 Kumasi Metropolitan (KMD) AR -2.906 30 Tamale Municipal (TMD) NR -2.992 31 Key: Positive GID score means high level of deprivation. Negative GID score means low level of deprivation.
In the category of high deprivation were four districts from the Ashanti Region: Amansie West, Adansi East, Ahafo-Ano South, and Amansie East were from the Ashanti Region. The Gushiegu-Karaga district was the only district from Northern Region in this category. Similarly, the low deprivation category was dominated by districts from Ashanti Region: Kumasi Metro, Adansi West and Asante-Akim North.
172 The Tamale Municipal and Yendi districts were the only districts from the Northern Region in that group. However, the Tamale Municipal was the overall lowest deprived district with a GID score of –2.992 ahead of the Kumasi Metro district (KMD) with a score of –2.906. The large medium level of deprivation category was almost evenly split between the two regions with ten districts in Northern and 11 in the Ashanti Region. However, most of the districts with low GID scores in the medium category were from Ashanti. Some peri-urban districts such as the Bosomtwe-Atwima Kwanwoma (BAK) in the Ashanti Region retained high GID scores (ranked 9th) ahead of some remote districts like the West Gonja in the Northern Region (ranked 11th).
8.4.1 Analysis of Deprivation by Region
8.4.1.1 Ashanti Region Relative deprivation as measured by GID in the Ashanti Region is shown in Figure 8.3. Seven of the 18 districts were less deprived as illustrated by their negative GID scores. Among these, the Kumasi Metro and Adansi West had the lowest GIDs.
Figure 8. 3 Deprivation among Districts in Ashanti Region
1.500
1.000
0.500
0.000
-0.500
-1.000
GID Score -1.500
-2.000
-2.500
-3.000
-3.500
st st o t st st n e st r ma e uth bre e uth Ea Ea et i Ea be inso o Wes W a BAK kyer W ie re M tw Ju e s si A ie Kwa S re Off a s u- a an kye an is kye -AkimSe North AdansiAdansi-AkimEj So Am e Kum Am e Afij Se sant sant Ejura-SekodumasiAhafo-AnoAhafo-Ano S North A A DISTRICT
Note: Positive scores depict high deprivation. Negative scores shows less deprivation
173 There were five highly deprived districts with GID scores of between 0.500 and 1.000. These included the Amansie West, Adansi East and the Ahafo-Ano South.
8.4.1.2 Northern Region Figure 8.4 shows levels of relative deprivation among the districts in Northern Region according to the GID. Three districts had negative GID scores: Tamale Municipal, Savelugu-Nanton and Yendi. There was particularly high negative score for the Tamale Municipal. In general, there was relative uniformity in deprivation among districts in the Northern Region. With exception of the Gushiegu-Karaga and Zabzugu-Tatale, which retained deprivation scores above 0.500, the remaining eight districts with positive GID scores maintained values between 0 and 0.500.
Figure 8. 4 Deprivation among Districts in Northern Region
1.500
1.000
0.500
0.000
-0.500
-1.000
GID Score -1.500
-2.000
-2.500
-3.000
-3.500
ja i le e u pal al di g Bo en ungu umba Tat Y shie b t Gon an u m es gu-Nanton N G u W t Mamprusu East Gonja a Chereponi es b on K W Zabzugu East Mamprusi Tamale Munici Savel Sabo Tol DISTRICT
Note: Positive scores depict high deprivation. Negative scores shows less deprivation
8.4.2 Rurality and Deprivation In both Ashanti and Northern Regions, the districts revealed a pattern of strong association of rurality with high deprivation, and of urbanity with low deprivation. The
174 least deprived districts with negative GID scores were largely urban. Conversely, all the districts with high GID scores were predominantly rural. On average, about 86% of the population of these districts was rural, compared to the 36% of the average rural population of districts with low GID scores. The study also found that a considerable proportion of households in these districts (with high GID scores) were without access to basic amenities such as electricity and piped water. On average, 85% and 94% of households in these districts had no access to electricity and piped water, compared to the 45% and 49% of households in districts with low GID scores. There were a number of exceptions to the strong association of urbanity with low deprivation in the Ashanti Region; some the peri-urban districts, including the Ejisu-Juaben and Kwabre, which border the Kumasi Metro, emerged more deprived than expected.
175 8.5 Equity-Adjusted Share (EAS) The equity-adjusted share was constructed separately for the Ashanti and Northern Regions because allocation decisions are made independently at the regional level. District population size was a key factor in deriving the EAS, as was the case in inter- regional EAS calculations.
8.5.1 Ashanti Region Table 8.14 shows the derivation of the EAS for districts in the Ashanti Region.
Table 8. 14 Derivation of Equity-Adjusted Share (EAS) for Ashanti Region
District GID NGID Population W/Pop EAS (%) Amansie East 0.732 4.638 225,309 1,045,028 9.5 Asante-Akim North -0.616 3.290 126,477 416,130 3.8 Sekyere East 0.331 4.237 157,396 666,880 6.1 Kumasi Metro -2.906 1.000 1,170,270 1,170,037 10.6 Atwima -0.136 3.770 237,610 895,710 8.1 Amansie West 0.894 4.800 108,726 521,900 4.7 Adansi West -1.079 2.827 238,440 674,177 6.1 Adansi East 0.804 4.710 129,308 609,076 5.5 Asante-Akim South 0.694 4.600 96,868 445,555 4.0 Ejisu-Juaben 0.180 4.086 124,176 507,373 4.6 Bosomtwe-Atwima Kwanwoma 0.485 4.391 146,028 641,209 5.8 Kwabre 0.134 4.040 164,668 665,228 6.0 Afigya-Sekyere 0.261 4.167 119,093 496,319 4.5 Sekyere West -0.156 3.750 143,206 536,995 4.9 Ejura-Sekodumasi -0.424 3.482 81,115 282,472 2.6 Offinso -0.224 3.682 138,676 510,617 4.6 Ahafo-Ano South 0.737 4.643 133,632 620,426 5.6 Ahafo-Ano North 0.288 4.194 71,952 301,800 2.7 Total 3,612,950 11,006,932 100.0 Note: NGID = normalised GID. W/Pop = weighted population. Weighted population = total population multiplied by normalised GID. The GID scores are same as those used in the inter-district deprivation analysis
Table 8.14 shows that the Kumasi Metro, by virtue of its large population (about 1.2 million, a third of Ashanti Region’s population), has the highest EAS of about 11%. The two other districts with the highest EAS behind the Kumasi Metro were the Amansie East and Atwima districts. The Amansie East had one of the highest deprivation scores (5th highest among the 31 districts) and has the fourth largest population in Ashanti. It was therefore not surprising that it scored the second highest EAS. Three of the most deprived districts in Ashanti (based on the GID scores), namely, the Amansie West,
176 Adansi East and Ahafo-Ano South only retained medium range EAS (around 5 to 6%). This is due to their relatively low populations. The Ejura-Sekodumasi and Ahafo-Ano North districts retained the lowest EAS in the Ashanti Region (about 3% in both cases). The two also have the smallest population in the region- less than 82,000 people in either case. On the whole, there were considerable variations in the EAS distribution, with some districts with relatively high GID scores retaining low EAS and vice versa.
8.5.2 Northern Region Unlike the Ashanti Region, where the EAS scores varied considerably with some districts scoring as low as 3%, the EAS scores in the Northern Region were quite high. With the exception of the Tamale Municipal district, which scored less than 5%, all the remaining 12 districts in the region retained EAS scores of at least 5% or more. Table 8.15 shows the EAS of districts in the Northern Region and how they were derived.
Table 8. 15 Derivation of Equity-Adjusted Share (EAS) for Northern Region
District GID NGID Population W/Pop EAS (%) Tamale Municipal -2.992 1.000 293,881 293,892 4.4 West Gonja 0.397 4.389 139,329 611,551 9.1 West Mamprusi 0.426 4.418 115,025 508,168 8.5 Savelugu-Nanton -0.228 3.764 89,968 338,661 5.0 Bole 0.313 4.305 127,147 547,341 8.1 East Gonja 0.108 4.100 174,500 715,367 10.6 Nanumba 0.374 4.366 144,278 629,907 9.3 Zabzugu Tatale 0.620 4.612 79,201 365,291 5.4 Saboba Chereponi 0.487 4.479 93,847 420,295 6.2 Yendi -0.587 3.405 130,504 444,336 6.6 Gushiegu Karaga 0.778 4.770 125,430 598,242 8.9 Tolon-Kumbungu 0.091 4.083 132,833 542,376 8.0 East Mamprusi 0.214 4.206 174,863 735,507 10.9 Total 1,820,806 6,750,933 100.0 Note: NGID = normalised GID. W/Pop = weighted population.
Two districts, the East Mamprusi and East Gonja, maintained the highest EAS scores of nearly 11% each. These districts were not the most deprived (ranked 8th and 9th in terms of GID), but had relatively large populations which influenced their EAS scores. By contrast, the Gushiegu-Karaga district, which had the highest GID maintained only
177 the fifth highest EAS. Similarly, the Tamale Municipal district with the highest population scored the lowest EAS. On the whole, eight districts in the Northern Region obtained EAS ranging from 8.5% to 10.9%, two between 6% and 7%, and three districts between 4.4% and 5.6%. This depicts a more uniform pattern of deprivation than the pattern that emerged from Ashanti.
8.6 Assessing Inter-District Equity of Resource Allocation The purpose of the inter-district analysis was to determine whether, and the extent to which, resource allocation from regional to district levels was equitable in terms of differentially benefiting the most deprived districts. In order to do this, data on funding allocation to the various districts from 1998 to 2002 were sought and analysed. The same approaches used for the inter-regional analysis (Section 8.4) were used. That is, a simple regression was used to determine the proportion of variance in the actual share of funds received by districts that could be explained by the equity-adjusted share (EAS). Then, the total yearly allocations to districts were redistributed using the EAS to get predicted shares, which were compared with the actual shares to determine the equity- gap. While the inter-regional analysis focused only on government funding, because no other data were available, this inter-district component of the study analysed data on both government and donor funds (see Chapter 6).
RESOURCE ALLOCATION IN ASHANTI REGION
8.6.1 Actual Allocation of Government of Ghana Funds This section analyses the pattern of government funding for administrative and service expenditures (GOG items 2-3), two main components of the public sector health budget considered in this study. The other components, salaries and investments, were not considered since regional health authorities play little role in their allocation. Table 8.16 shows the actual and percentage shares of GOG 2-3 allocations from 1998 to 2002.
178 Table 8. 16 Districts by Actual and Percentage Shares of Government Funds (GOG 2-3): Ashanti Region 1998 - 2002 Actual Actual Actual Actual Actual District 1998 1999 2000 2001 2002 ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) Amansie East 89,675 6.8 85,841 6.3 102,435 5.8 90,100 5.5 133,508 5.7 Asante-Akim North 57,780 4.4 64,093 4.7 97,273 5.5 90,275 5.5 136,928 5.9 Sekyere East 84,956 6.4 64,877 4.7 100,799 5.7 84,994 5.2 134,883 5.8 Kumasi Metro 230,923 18.5 191,884 14.0 99,390 5.6 123,514 8.5 133,430 5.7 Atwima 57,031 4.3 78,286 5.7 100,995 5.7 109,168 6.7 140,404 6.0 Amansie West 57,586 4.4 64,399 4.7 95,050 5.4 74,708 4.6 125,714 5.4 Adansi West 67,261 5.1 74,696 5.5 97,946 5.5 76,612 4.7 128,942 5.5 Adansi East 68,536 5.2 70,264 5.1 98,624 5.6 114,418 8.0 129,036 5.5 Asante-Akim South 55,794 4.2 64,267 4.7 98,070 5.5 86,902 5.3 117,380 5.0 Ejisu-Juaben 61,818 4.7 68,426 5.0 96,106 5.4 96,837 5.9 126,312 5.4 BAK 58,310 4.4 64,355 4.7 92,840 5.2 74,959 4.6 121,417 5.2 Kwabre 59,392 4.5 66,705 4.9 95,134 5.4 100,525 6.1 128,220 5.5 Afigya-Sekyere 54,764 4.2 62,332 4.6 99,574 5.6 77,336 4.7 122,155 5.2 Sekyere West 54,114 4.1 73,236 5.4 105,168 5.9 90,730 5.5 139,339 6.0 Ejura-Sekodumasi 70,300 5.3 75,674 5.5 104,899 5.9 82,323 5.0 132,614 5.7 Offinso 89,955 6.8 72,084 5.3 96,552 5.4 86,929 5.3 122,686 5.3 Ahafo-Ano South 57,188 4.3 66,009 4.8 96,535 5.4 98,534 6.0 124,504 5.3 Ahafo-Ano North 42,235 3.2 61,265 4.5 97,645 5.5 81,279 5.0 132,175 5.7 Total 1,317,617 100 1,368,694 100 1,775,037 100 1,640,142 100 2,329,647 100 Source: Ghana Health Service, Ashanti Regional Health Directorate, Kumasi Ghana. Amount is in Ghanaian Cedis. See Chapter 2 for exchange rate of the Ghanaian Cedi to the US Dollar.
Table 8.16 shows that the Kumasi Metro district received the largest proportion of the government administrative and service budgets for 1998 (18.5%) and 1999 (14%). However, it suffered a substantial lost of funding in 2000 (about 60% less than the 1999 allocation). The Asante-Akim North was the only district that did not lose any funding for the entire five years. Districts like the Ahafo-Ano North and Sekyere West benefited considerably in terms of percentage increase in funding, gaining about 40% and 30% respectively in 1988/1999 (see Appendix C). On the whole, aside the drastic reduction of funding to the Kumasi Metro district in 1999/2000 (about 60% reduction), the fluctuations in government funding in the Ashanti Region were minimal.
8.6.2 Actual Allocation of Donor-Pooled Funds (DPF) Donor funds, as indicated in Chapter 2, form a substantial proportion (about 35%) of the total health budget in Ghana. Table 8.17 provides the actual and percentage shares of donor-pooled funds (DPF) over the five-year study period from 1998 to 2002. The
179 Kumasi Metro district was allocated the highest proportion of donor-pooled funds from 1998 to 2000, receiving between 17% of the total DPF budget in 1998 and 20% in 2000.
Table 8. 17 Districts by Actual and Percentage Shares of Donor-Pooled Funds (DPF): Ashanti Region 1998 to 2002 Actual Actual Actual Actual Actual District 1998 1999 2000 2001 2002 ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) Amansie East 19,844 5.9 85,115 8.0 15,843 2.9 191,941 4.1 158,716 5.2 Asante-Akim North 22,506 6.7 54,063 4.5 14,074 2.5 274,612 5.8 158,874 5.2 Sekyere East 17,300 5.1 54,346 4.5 29,152 5.3 293,180 6.2 191,001 6.2 Kumasi Metro 56,774 16.9 189,219 15.6 110,892 20.0 306,912 6.5 169,705 5.6 Atwima 17,066 5.1 73,712 6.1 41,761 8.5 272,357 5.8 162,029 5.3 Amansie West 16,902 5.0 55,083 4.5 14,800 2.7 240,379 5.1 125,594 4.1 Adansi West 15,229 4.5 66,034 5.4 39,535 8.1 271,790 5.8 126,356 4.1 Adansi East 17,747 5.3 67,906 5.6 47,390 8.6 270,206 5.7 217,606 8.1 Asante-Akim South 18,650 5.5 53,410 4.4 18,683 3.4 266,602 5.7 191,635 6.3 Ejisu-Juaben 11,681 3.5 57,892 4.8 27,480 5.0 280,814 6.0 164,327 5.4 BAK 14,994 4.5 57,670 4.8 32,919 5.9 266,457 5.7 165,945 5.4 Kwabre 14,652 4.4 56,923 4.7 22,083 4.0 259,909 5.5 127,786 4.2 Afigya-Sekyere 16,664 5.0 49,010 4.0 23,378 4.2 214,353 4.5 223,180 8.3 Sekyere West 19,528 5.8 66,403 5.5 31,681 5.7 308,902 6.6 160,075 5.2 Ejura-Sekodumasi 11,744 3.5 61,897 5.1 13,478 2.4 293,651 6.2 182,761 6.0 Offinso 18,378 5.5 62,379 5.1 12,476 2.3 265,806 5.6 181,969 6.0 Ahafo-Ano South 15,694 4.7 55,193 4.5 25,352 4.6 258,492 5.5 208,690 6.8 Ahafo-Ano North 10,872 3.2 46,908 3.9 32,734 5.9 175,015 3.7 140,571 4.6 Total 336,226 100 1,213,163 100 553,712 100 4,711,378 100 3,056,820 100 Source: Ghana Health Service, Ashanti Regional Health Directorate, Kumasi Ghana. Amount is in Ghanaian Cedis. See Chapter 2 for exchange rate of the Ghanaian Cedi to the US Dollar.
However, as with government funds, allocation to the Kumasi Metro district dramatically fell (about 68%) in 2000/2001. Over that same period (2000/2001), districts such as Asante-Akim North, Ejura-Sekodumase and Offinso gained over 100% increase in funding (Asante-Akim North gained about 129%, Ejura-Sekodumase 156% and Offinso 150%). These huge gains were the result of the low DPF allocations to these districts in 2000. Some of the most deprived districts such as the Amansie West and Asante-Akim South also made significant gains in 2000/2001 (91% and 68% respectively). The most striking feature of DPF allocation in Ashanti was the low total allocations in 1998 and 2000 and the dramatic increase in 2001. This issue was raised during the fieldwork but no alternative data were available39.
39 The low total allocation may be due to inadequate reporting on the part of the Regional Health Administration (RHA). There were no records at the national level to validate the figures.
180 In general, the pattern of DPF distribution in Ashanti, particularly from 2000, suggests an effort to shift resources from urban districts perceived as rich to deprived rural districts. The next section assesses the extent to which actual allocations of government and donor funding were equitable, using regression analysis and the comparison of actual and predicted allocations.
8.6.3 Regression Analysis
8.6.3.1 Government of Ghana Funds (GOG) Table 8.18 provides the summarised results of the regression analysis for GOG allocation. The results show a highly significant relationship between the EAS and percentage share of GOG for 1998 (p = 0.003), 1999 (p<.001) and 2001 (p = 0.021).
Table 8. 18 Regression of Equity-Adjusted Share and Population on Percentage Share of Government of Ghana (GOG) Funds: 1998 – 2002
Independent Variable R-Square Unstandardised Sig. (P-Value) Coefficient GOG: 1998 EAS 0.470 1.082 0.002** Pop% 0.896 0.465 <.001** GOG: 1999 EAS 0.461 0.695 0.002** Pop% 0.947 0.310 <.001** GOG: 2000 EAS 0.042 0.04908 0.414 Pop% 0.046 0.01590 0.395 GOG: 2001 EAS 0.290 0.233 0.021* Pop% 0.462 0.09150 0.002** GOG: 2002 EAS 0.077 0.06645 0.264 Pop% 0.071 0.01978 0.286
Key: *Coefficient and P-Values significant at 5%. **Coefficient and P-Values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%) Dependent Variable = Percentage Share of Actual Funds (GOG)
There were, however, no significant relationships between the EAS and the percentage share of actual GOG for 2000 and 2002. The influence of population size on share of funds was very significant in 1998, 1999 and 2001. The general trend in the regression results is that in 1998 and 1999, GOG allocation in Ashanti Region was driven almost
181 entirely by population size, accounting for about 90% of the variability in the share of funds received by districts. Population size, however, was not an important determinant in 2000 and 2002.
8.6.3.2 Donor-Pooled Funds (DPF) There is similar pattern on highly significant relationships between the EAS and actual share of DPF from 1998 to 2000 (Table 8.19). The proportion of variance in the actual share of DPF (dependent variable) that could be reliably predicted from the EAS (independent variable) was about 43% in 1998 (p=0.003), 56% in 1999 (p=0.005) and 40% in 2000 (p=0.005).
Table 8. 19 Regression of Equity-Adjusted Share and Population on Percentage Share of Donor- Pooled Funds (DPF): 1998 – 2002 Independent Variable R-Square Unstandardised Sig. (P-Value) Coefficient DPF: 1998 EAS 0.433 0.930 0.003** Pop 0.912 0.420 <.001** DPF: 1999 EAS 0.561 0.935 <.001** Pop 0.955 0.380 <.001** DPF: 2000 EAS 0.402 1.226 0.005** Pop 0.819 0.545 <.001** DPF: 2001 EAS 0.033 0.170 0.472 Pop 0.068 0.077 0.295 DPF: 2002 EAS 0.006 0.361 0.757 Pop 0.003 -0.008 0.836
Key: *Coefficient and P-Values significant at 5%. **Coefficient and P-Values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%) Dependent Variable = Percentage Share of Actual Funds (DPF)
No significant relationship was found between the EAS and actual share of DPF in 2001 and 2002. Indeed, there was a dramatic decrease in the proportion of variance accounted for by the EAS from 40% in 2000 to 3% in 2001 and less than 1% in 2002. Population size, as with GOG allocations, was a major predictor of DPF allocations in 1998, 1999 and 2000, explaining about 90% on average, of the variability in the actual share of funds for those three years.
182 In general, funding allocation in Ashanti Region, according to the regression analysis, was less equitable than expected, given the small proportions of variance in the share of funds that could be explained by the EAS. With the exception of DPF allocation in 1999, where the EAS accounted for 56% of the variability in the share of funds, all the proportions of variance for the five years were below 50%. The p-values were, nonetheless, significant in 1998, 1999 and 2000/2001. However, resource allocations for those three years were almost entirely driven by population size, which, as a major component of the EAS, influenced the proportion of variance explained by the EAS. The situation changed radically in 2000/2001 and 2002, with population losing its significance as a reliable predictor of funds.
8.6.4 Comparison of Actual and Predicted Allocations While the regression analysis showed the degree of equity in funding allocation for a particular year, in terms of the proportion of variance in actual share of funds that is accounted for by the EAS, it did not show which districts received more or less than their equity-adjusted shares. To assess the situation of individual districts, the total yearly allocations were redistributed using the EAS to get the predicted ‘equitable’ allocations, which were compared to actual districts’ allocations to determine the funding difference, labelled in this study as the equity-gap (see Chapter 6).
8.6.4.1 Government of Ghana Funds (GOG) Table 8.20 shows the predicted (EAS-based) shares of government funds (GOG 2-3) and the percentage of funding difference from 1998 to 2002 (see Table 8.16 for the actual allocations). The percentage of funding difference shows clearly that districts such as Ejura-Sekodumasi, Ahafo-Ano North and Asante-Akim North, received more government funding in excess of their EAS-based shares. By contrast, districts like the Amansie East (one of the most deprived in the region), BAK and Atwima, received less than their predicted EAS-based allocations for the entire five years. The comparisons of the two shares (actual and predicted) showed that several districts were not ‘fairly’ treated, even when the allocation appeared to be largely equitable.
183 Table 8. 20 Districts by Predicted (EAS-Based) Share of Government Funds (GOG 2-3): Ashanti Region 1998 - 2002 Predicted Predicted Predicted Predicted Predicted Funding Funding Funding Funding Funding 1998 1999 2000 2001 2002 District ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 Amansie East 125,098 129,948 168,527 155,720 221,183 Asante-Akim North 49,814 51,745 67,107 62,007 88,075 Sekyere East 79,831 82,925 107,545 99,372 141,147 Kumasi Metro 140,063 145,492 188,686 174,347 247,642 Atwima 107,224 111,380 144,447 133,470 189,579 Amansie West 62,476 64,897 84,164 77,768 110,462 Adansi West 80,704 83,833 108,721 100,459 142,692 Adansi East 72,911 75,738 98,223 90,758 128,913 Asante-Akim South 53,337 55,404 71,853 66,392 94,303 Ejisu-Juaben 60,737 63,091 81,822 75,604 107,387 BAK 76,758 79,733 103,405 95,546 135,714 Kwabre 79,633 82,720 107,278 99,126 140,797 Afigya-Sekyere 59,413 61,716 80,039 73,956 105,047 Sekyere West 64,283 66,774 86,599 80,018 113,656 Ejura-Sekodumasi 33,814 35,125 45,553 42,091 59,786 Offinso 61,125 63,494 82,345 76,087 108,073 Ahafo-Ano South 74,270 77,149 100,053 92,450 131,315 Ahafo-Ano North 36,128 37,528 48,670 44,971 63,877 Total 1,317,617 1,368,694 1,775,037 1,640,142 2,329,648
Percentage of Funding Difference District 1998 1999 2000 2001 2002 Amansie East -28.3 -33.9 -39.2 -42.1 -39.6 Asante-Akim North 16.0 23.9 45.0 45.6 55.5 Sekyere East 6.4 -21.8 -6.3 -14.5 -4.4 Kumasi Metro 64.9 31.9 -48.3 -29.2 -46.1 Atwima -46.8 -29.7 -30.1 -18.2 -25.9 Amansie West -8.8 -0.8 12.9 -3.9 13.8 Adansi West -16.7 -10.9 -9.9 -23.7 -9.6 Adansi East -6.0 -8.2 0.4 26.1 0.1 Asante-Akim South 4.6 16.0 36.5 30.9 24.5 Ejisu-Juaben 1.8 8.5 18.5 28.1 18.6 BAK -24.0 -19.3 -10.2 -21.5 -10.5 Kwabre -25.4 -19.4 -11.3 1.4 -8.9 Afigya-Sekyere -8.8 1.0 24.4 4.6 16.3 Sekyere West -15.8 9.7 21.4 13.4 22.6 Ejura-Sekodumasi 108.9 115.4 130.3 95.6 121.8 Offinso 48.2 13.5 18.3 14.2 13.5 Ahafo-Ano South -23.0 -14.4 -3.5 6.6 -5.2 Ahafo-Ano North 16.9 63.2 100.6 80.7 106.9 Note: The predicted allocations were calculated by redistributing the total yearly (actual) allocations using the EAS. See Table 8.17 for the actual shares. Amount is in millions of Ghanaian Cedis. Percentage of funding difference is the difference between the actual and predicted allocations expressed as a percentage. It was calculated by dividing the funding difference by the predicted allocation and multiplying by 100.
184 Figure 8.5 shows the equity-gap in government funding from 1998 to 2002 in Ashanti Region.
Figure 8. 5 Percentage Difference Between Actual and Predicted Shares of Government Funding (GOG 2-3) in Ashanti Region: 1998 – 2002
200
) 150
100 1998 1999 50 2000 2001 2002 0
% Deviation from Equity Target (EAS -50
-100
t st st s h n K so th tro ma e e ast be A est n Ea orth e i E e N e East M tw W W i Sout ua B kyere W Nor r si s e Offi o nsi m si A Kwabre ki a nsie u-J ekye a dan is -A S um m A Adan -Akim Ej o-Ano Southo-An Ama K Afija S Sekyere f f nte A Ejura-Sekodumasi Aha Aha Asa Asante
DISTRICT
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
8.6.4.2 Donor-Pooled Funds (DPF) Table 8.21 shows the predicted donor-pooled funds (DPF) and the percentages of funding difference. The comparison of actual and predicted DPF showed a similar trend as the GOG allocations; over-funding of the Ejura-Sekodumase, Ahafo-Ano North and the Asante-Akim North districts, while the Amansie East, Atwima and Kwabre districts received less than their EAS-based allocations for the entire five years. The Sekyere West district was also allocated more than its EAS-based share for the entire five years, though not as much in excess as others like Ejura-Sekodumase. On the whole, allocations of both government and donor-pooled funds to three of the peri-urban districts bordering the Kumasi Metro: Atwima, BAK and Kwabre were below ‘equitable’ levels, but the Amansie East district was the biggest loser (Figure 8.6)
185 Table 8. 21 Districts by Predicted (EAS-Based) Share of Donor-Pooled Funds (DPF): Ashanti Region - 1998 to 2002 Predicted Predicted Predicted Predicted Predicted Funding Funding Funding Funding Funding District 1998 1999 2000 2001 2002 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 Amansie East 31,922 115,181 52,571 447,311 290,223 Asante-Akim North 12,711 45,865 20,934 178,119 115,567 Sekyere East 20,371 73,502 33,548 285,449 185,204 Kumasi Metro 35,741 128,959 58,860 500,820 324,940 Atwima 27,361 98,723 45,059 383,397 248,754 Amansie West 15,942 57,523 26,255 223,393 144,941 Adansi West 20,594 74,307 33,915 288,573 187,231 Adansi East 18,605 67,131 30,640 260,707 169,151 Asante-Akim South 13,610 49,108 22,414 190,714 123,739 Ejisu-Juaben 15,499 55,922 25,524 217,175 140,907 BAK 19,587 70,673 32,256 274,461 178,075 Kwabre 20,321 73,320 33,465 284,742 184,746 Afigya-Sekyere 15,161 54,703 24,968 212,443 137,837 Sekyere West 16,403 59,187 27,014 229,854 149,133 Ejura-Sekodumasi 8,629 31,133 14,210 120,908 78,447 Offinso 15,598 56,279 25,687 218,563 141,807 Ahafo-Ano South 18,952 68,382 31,211 265,566 172,303 Ahafo-Ano North 9,219 33,264 15,182 129,182 83,815 Total 336,226 1,213,163 553,712 4,711,378 3,056,820
Percentage of Funding Difference 1998 1999 2000 2001 2002 Amansie East -38.8 -26.1 -69.9 -58.1 -45.3 Asante-Akim North 78.1 18.9 -32.8 54.2 38.5 Sekyere East -15.1 -26.1 -13.1 2.7 3.1 Kumasi Metro 58.8 46.7 88.4 -38.7 -48.8 Atwima -38.6 -25.3 -8.3 -29.0 -34.9 Amansie West 6.0 -4.2 -43.6 8.6 -13.3 Adansi West -26.1 -11.1 16.6 -5.8 -32.5 Adansi East -4.6 1.2 54.7 3.6 28.6 Asante-Akim South 38.0 8.8 -16.6 39.8 54.9 Ejisu-Juaben -24.6 3.5 8.7 29.3 16.6 BAK -23.4 -18.4 2.1 -2.9 -6.8 Kwabre -28.9 -22.4 -34.0 -8.7 -30.8 Afigya-Sekyere 9.9 -10.4 -6.4 0.9 61.9 Sekyere West 19.0 12.2 18.3 34.4 8.3 Ejura-Sekodumasi 36.1 98.8 -5.1 142.9 133.0 Offinso 18.8 10.8 -51.4 21.6 28.3 Ahafo-Ano South -18.2 -19.3 -18.8 -2.7 21.1 Ahafo-Ano North 18.9 41.0 115.6 35.5 68.7 Note: The predicted shares were calculated by redistributing the total yearly (actual) allocations using the EAS. See Table 8.17 for the actual shares. Amount is in millions of Ghanaian Cedis. Percentage of funding difference is the difference between the actual and predicted allocations expressed as a percentage. It was calculated by dividing the funding difference by the predicted allocation and multiplying by 100.
186 Figure 8. 6 Percentage Difference Between Actual and Predicted Shares of Donor-Pooled Funds (DPF) in Ashanti Region: 1998 – 2002
200
150 )
100 1998 1999 50 2000 2001 2002 0 % Deviation from Equity Target (EAS -50
-100
t t e o h st tro est s r st s th Ea orth e Ea outh abre in N W Wes i S BAK f Sout Nor e re East M e i s Juaben e We si Atwima s n Kw Sekye r Of nsi nsi n u- kodumasi no no ekye ija kye e -Akim S Ada Ada -Akim Ejis e -S o-A o-A Ama Kuma Af S f f nte Ama nte Ejura Aha Aha Asa Asa DISTRICT
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
187 RESOURCE ALLOCATION IN NORTHERN REGION
8.6.5 Actual Allocation of Government of Ghana Funds (GOG 3) Analysis of funding allocation in the Northern Region was subject to availability of data. As indicated in Chapter 6, data on government funding were available for only the service budget (item 3) from 2000 to 2002. Table 8.22 shows actual allocations and percentage share in funds for each district from 2000-2002. The most striking feature is the low total budget for 2001 and 2002 (less than half of the 2000 budget). Possible reasons for this have been discussed in Chapter 10.
Table 8. 22 Districts by Actual Government Funds (GOG 3) in Northern Region: 2000 - 2002 Actual Actual Actual 2000 2001 2002 District ⊄ ‘000 (%) ⊄ ‘000 (%) ⊄ ‘000 (%) Tamale Municipal 84,367 8.7 34,599 8.1 39,019 8.3 West Gonja 76,421 8.9 37,863 8.8 41,930 9.0 West Mamprusi 92,298 9.6 34,365 8.0 38,001 8.1 Savelugu-Nanton 58,546 6.1 26,464 6.2 29,427 6.3 Bole 59,729 6.2 35,805 8.3 40,525 8.7 East Gonja 108,520 11.2 40,236 9.4 43,377 9.3 Nanumba 88,445 9.2 32,336 8.5 38,812 8.3 Zabzugu Tatale 60,275 6.2 26,761 6.2 12,768 2.7 Saboba Chereponi 44,359 4.6 31,295 8.3 37,857 8.1 Yendi 78,674 8.2 30,398 8.1 35,331 8.6 Gushiegu Karaga 80,245 8.3 33,194 8.7 37,710 8.1 Tolon-Kumbungu 81,423 8.4 30,879 8.2 33,994 8.3 East Mamprusi 51,722 5.4 34,634 8.1 38,574 8.3 Total 965,023 100.0 428,829 100.0 467,325 100.0 Source: Ghana Health Service, Northern Regional Health Directorate, Tamale- Ghana. Amount in millions of Ghanaian cedis. See Chapter 2 for exchange rate of the Ghanaian Cedi to the US Dollar.
The East Gonja district received the highest proportion of government funding for the entire three years (11.2% in 2000, 9.4% in 2001 and 9.3% in 2002). However, this represented a steadily decline in funding over the three-year period. The big winners were the Saboba-Chereponi and East Mamprusi districts, which witnessed over 50% increase in funding between 2000 and 2001. One of the most deprived districts in the Northern Region, the Zabzugu-Tatale districts, lost substantial amount of funding between 2001 (6.2%) and 2002 (2.7%). In general, there were relatively minor shifts in funding when compared allocations in 2000/2001 to 2001/2002.
188 8.6.6 Actual Allocation of Donor-Pooled Funds (DPF) Allocation of donor-pooled funds (DPF) in the Northern Region was analysed over a four-year period from 1999 to 2002. Table 8.23 shows the district share of DPF and the percentage shares for the four years.
Table 8. 23 Districts by Actual Donor Pooled Funds (DPF), Northern Region: 1999 – 2002 Actual Actual Actual Actual District 1999 2000 2001 2002 ⊄ ‘000 (%) ⊄‘000 (%) ⊄‘000 (%) ⊄‘000 (%) Tamale Municipal 73,498 8.5 51,396 8.5 255,859 8.8 118,176 8.6 West Gonja 70,870 8.2 60,990 8.9 288,010 8.7 136,245 8.7 West Mamprusi 67,266 8.8 58,821 8.6 263,064 8.0 124,392 8.0 Savelugu-Nanton 54,559 6.3 40,661 6.0 202,000 6.1 95,944 6.1 Bole 68,997 8.0 62,777 9.2 262,972 8.0 133,222 8.5 East Gonja 91,423 10.5 64,569 9.5 295,934 9.0 139,423 8.9 Nanumba 59,736 6.9 47,283 6.9 266,020 8.1 124,472 8.0 Zabzugu-Tatale 54,184 6.2 44,240 6.5 209,864 6.4 98,898 6.3 Saboba-Chereponi 59,441 6.9 50,635 8.4 248,536 8.5 119,930 8.7 Yendi 65,738 8.6 47,436 6.9 242,804 8.4 115,759 8.4 Gushiegu-Karaga 64,600 8.4 50,635 8.4 263,387 8.0 123,627 8.9 Tolon-Kumbungu 62,793 8.2 49,226 8.2 231,904 8.0 109,232 8.0 East Mamprusi 74,057 8.5 54,594 8.0 263,387 8.0 124,389 8.0
Total 867,163 100.0 683,263 100.0 3,293,741 100.0 1,563,709 100.0 Source: Ghana Health Service, Northern Regional Health Directorate, Tamale- Ghana. Amount in millions of Ghanaian cedis. See Chapter 2 for exchange rate of the Ghanaian Cedi to the US Dollar.
Likewise government funding, the highest proportion of donor-pooled funds (DPF) for each of the four years examined, went to the East Gonja districts. However, the district’s share of the total DPF budget reduced continuously from 10.5% in 1999 to 8.9% in 2002. The Zabzugu-Tatale district, on the other hand, was allocated relatively low share of DPF over the four-year period. No district received less than 6% share of DPF for any of the four years analysed. In general, DPF allocation in the Northern Region had one striking similarity to that of the Ashanti Region; the total allocations were exceptionally high in both regions for 2001, particularly, in the Northern Region. The qualitative study provided some insight into why donor funding was so high in 2001. This has been discussed in Chapter 10.
189 8.6.7 Regression Analysis
8.6.7.1 Government of Ghana Funds (GOG) There were significant relationships between the percentage share of government funds (the dependent variable) and the EAS (the independent variable) for 2001 and 2002. Table 8.24 shows that, the EAS, of which population is a major component accounted for the about 48% and 30% of the variability in the share of funds in 2001 and 2002 respectively. On its own, population size accounted for about 30% of the variability in the actual share of GOG-3 for 2001 (p>.05).
Table 8. 24 Regression of Equity-Adjusted Share and Population on Percentage Share of Government of Ghana (GOG) Funds: 2000 – 2002 Independent Variable R-Square Unstandardised Sig. (P-Value) Coefficient GOG: 2000 EAS 0.072 0.229 0.377 Pop% 0.155 0.244 0.183 GOG: 2001 EAS 0.468 0.297 0.010** Pop% 0.303 0.173 >0.05 GOG: 2002 EAS 0.301 0.399 >0.05 Pop% 0.200 0.235 0.126
Key: *Coefficient and p-values significant at 5%. **Coefficient and p-values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%). Dependent Variable = Percentage Share of Actual Funds (GOG-3)
8.6.7.2 Donor-Pooled Funds (DPF) Table 8.25 provides the regression results with regards to DPF allocations from 1999 to 2002. It shows that the EAS was a reliably predicts the actual share of DPF allocated to districts in three of the four years analysed - 1999, 2001 and 2002. About 41% of the variability in the share of DPF could be explained by the EAS in 2001. The p-value was 0.019. No significant relationship, however, existed between the dependent and independent variables in 2000.
Unlike in the Ashanti Region, where per capita allocation was largely responsible for the significant relationship between the EAS and share of funds, population size had little influence on funding allocation in the Northern Region. No statistically significant relationship existed between population size and share of DPF for the entire four years
190 (1999 to 2002) analysed. This was surprising as population was thought to be the main basis for inter-district resource allocation in Ghana (Bossert et al. 2000).
Table 8. 25 Regression of Equity-Adjusted Share and Population on Percentage Share of Donor-Pooled Funds (DPF): 1999 – 2002 Independent Variable R-Square Unstandardised Sig. (P-Value) Coefficient DPF: 1999 EAS 0.382 0.391 0.024* Pop% 0.284 0.229 0.061 DPF: 2000 EAS 0.249 0.269 0.082 Pop% 0.179 0.155 0.149 DPF: 2001 EAS 0.407 0.291 0.019* Pop% 0.232 0.149 0.096 DPF: 2002 EAS 0.376 0.299 0.026* Pop% 0.186 0.143 0.141
Key: *Coefficient and p-values significant at 5%. **Coefficient and p-values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%). Dependent Variable = Percentage share of actual funds (DPF)
The results of the regression analysis for Northern Region suggest that, in allocating resources among districts, the Regional Health Administration (RHA) considers other factors and seeks not to distribute funds solely on the basis of population size. On the whole, despite the significant relationships between the dependent (share of funds) and independent (EAS) variables for all the years (1999-2002) examined except for 2000, resources allocation in the region could be described as moderately equitable. This is because, like in Ashanti Region, the proportions of variance in the actual share of funds explained by the EAS were not above 50%.
The main contrasts between the two regions in terms of the regression results is that, in Ashanti, population was the key driver of resource allocation from 1998 to 2000. When this practice of per capita funding stopped after 2000, no clear pattern was observable from the allocation system again. It was unclear why allocation on the basis of population size suddenly stopped. Viewed in the context of the changes to the financial management system in 2000, which transferred control over the district budget from the RHA to DHAs, the sudden change could be as a result of district managers getting more involved in the allocation process. The results also show a limited effort by the RHA in Ashanti to assess and use health needs as the basis for resource allocation.
191 8.6.8 Comparison of Actual and Predicted Allocations
8.6.8.1 Government of Ghana (GOG) Funds Table 8.26 shows the predicted shares of government funds (GOG-3) as re-distributed using the EAS, and the percentage of funding difference between the actual and predicted shares (see Table 8.22 for the actual allocations). The re-distribution shifted resources in favour of the East Mamprusi and East Gonja districts, which had the highest EAS indices. Tamale Municipal had the least predicted allocation.
Table 8. 26 Districts by Predicted (EAS-Based) Government Funds (GOG-3) Northern Region: 2000 to 2002 Predicted Funding Predicted Funding Predicted Funding District 2000 2001 2002 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 Tamale Municipal 42,011 18,668 20,344 West Gonja 87,419 38,847 42,334 West Mamprusi 72,641 32,280 35,177 Savelugu-Nanton 48,410 21,512 23,443 Bole 78,241 34,768 37,889 East Gonja 102,259 45,441 49,520 Nanumba 90,043 40,013 43,605 Zabzugu Tatale 52,217 23,204 25,287 Saboba Chereponi 60,080 26,698 29,094 Yendi 63,516 28,225 30,759 Gushiegu Karaga 85,517 38,001 41,413 Tolon-Kumbungu 77,531 34,453 37,545 East Mamprusi 105,138 46,720 50,915 Total 965,023 428,829 467,325
Percentage of Funding Difference 2000 2001 2002 Tamale Municipal 100.8 85.3 91.8 West Gonja -12.6 -2.5 -1.0 West Mamprusi 28.1 6.5 8.0 Savelugu-Nanton 20.9 23.0 25.5 Bole -23.7 3.0 8.0 East Gonja 6.1 -11.5 -12.4 Nanumba -1.8 -19.2 -11.0 Zabzugu Tatale 15.4 15.3 -49.5 Saboba Chereponi -26.2 18.2 30.1 Yendi 23.9 8.7 14.9 Gushiegu Karaga -6.2 -12.7 -8.9 Tolon-Kumbungu 5.0 -10.4 -9.5 East Mamprusi -50.8 -25.9 -24.2 Note: Predicted shares were obtained by redistributing the total yearly (actual) allocations using the EAS. See Table 8.23 for the actual shares. Percentage of funding difference is the difference between the actual and predicted allocations expressed as a percentage. It was calculated by dividing the funding difference by the predicted allocation and multiplying by 100.
192 The percentage of funding difference, however, shows that the Tamale Municipal was highly over-funded (i.e. received funding excess of its EAS-based allocations) from 2000 to 2002. The East Mamprusi district was substantially under-funded for the entire period (2000 – 2002). Figure 8.7 illustrates the funding with regards to allocation of government funds (GOG –3). It clearly shows that in addition to the Tamale Municipal district, the two other predominantly urban districts; Savelugu-Nanton and Yendi, were also over-funded.
Figure 8. 7 Percentage Difference between Actual and Predicted Shares of Government Funds (GOG –3) in Northern Region: 2000 –2002
120
100
80
60
40 2000 2001 20 2002
0
-20
-40
-60
a ja n a u si al n si o b ag u ip o ru Bole r ug r c ant Yendi b p u N -Ka m m un amp - u u a u Nanum Chereponi M West G East Gonja - ieg n-K le elug h a v oba lo ast M West M Zabzugu-Tataleb o E am Sa Gus T T Sa
DISTRICT
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
8.6.8.2 Donor-Pooled Funds (DPF) Table 8.27 shows predicted DPF allocation and percentage of funding difference from 1999-2002. As with the allocation of government funds, the Tamale Municipal had the biggest chunk of the DPF above its EAS-based share. Savelugu-Nanton and Yendi, the two other least deprived districts in Northern Region also received more than their
193 predicted ‘equitable’ shares over the entire four years. The Nanumba and East Mamprusi districts were the most under-funded districts in terms of DPF allocations.
Table 8. 27 Districts by Predicted (EAS-Based) Donor Pooled Funds (DPF) Northern Region: 1999 to 2002 Predicted Predicted Predicted Predicted Funding Funding Funding Funding District 1999 2000 2001 2002 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 ⊄ ‘000 Tamale Municipal 37,749 29,745 143,388 68,074 West Gonja 78,552 61,895 298,372 141,653 West Mamprusi 65,272 51,432 247,932 117,706 Savelugu-Nanton 43,500 34,276 165,231 78,444 Bole 70,304 55,396 267,044 126,780 East Gonja 91,886 72,402 349,023 165,699 Nanumba 80,909 63,753 307,328 145,904 Zabzugu Tatale 46,920 36,971 178,223 84,612 Saboba Chereponi 53,985 42,538 205,060 97,352 Yendi 57,073 44,971 216,789 102,921 Gushiegu Karaga 76,842 60,548 291,879 138,570 Tolon-Kumbungu 69,666 54,894 264,622 125,630 East Mamprusi 94,473 74,441 358,850 170,364 Total 867,133 683,263 3,293,741 1,563,709
Percentage of Funding Difference 1999 2000 2001 2002 Tamale Municipal 94.7 72.8 78.4 73.6 West Gonja -9.8 -1.5 -3.5 -3.8 West Mamprusi 3.1 14.4 6.1 5.7 Savelugu-Nanton 25.4 18.6 22.3 22.3 Bole -1.9 13.3 -1.5 5.1 East Gonja -0.5 -10.8 -15.2 -15.9 Nanumba -26.2 -25.8 -13.4 -14.7 Zabzugu Tatale 15.5 19.7 18.8 16.9 Saboba Chereponi 10.1 19.0 21.2 23.2 Yendi 15.2 5.5 12.0 12.5 Gushiegu Karaga -15.9 -16.4 -9.8 -10.8 Tolon-Kumbungu -9.9 -10.3 -12.4 -13.1 East Mamprusi -21.6 -26.7 -26.6 -28.0 Note: Predicted shares were obtained by redistributing the total yearly (actual) allocations using the EAS. See Table 8.23 for the actual shares. Percentage of funding difference is the difference between the actual and predicted allocations expressed as a percentage. It was calculated by dividing the funding difference by the predicted allocation and multiplying by 100.
Figure 8.8 highlights the gap between actual and predicted allocations with respect to DPF. Two of the most deprived districts in Northern Region, Zabzugu-Tatale and Saboba-Chereponi received more DPF allocations above their predicted ‘equitable’ shares for each of the four years examined.
194 Figure 8. 8 Percentage Difference between Actual and Predicted Shares of Donor-Pooled Funds (DPF) in Northern Region: 1999 –2002
120
100 ) 80
60 1999 40 2000 20 2001 2002 0
-20 % Deviation from Equity Target (EAS -40
-60
l n e e i a u a si o l g g si p nja ru ta on ndi a ru p nt Bol p e r p ici Go Ta re Y m Na - e m st a - st Gonja u h umbu a e M u a Nanumba g K M Mun E u - W lug a-C n st est b a ve o W Zabz b Tolo E Tamale Sa Gushiegu-Ka Sa
DISTRICT
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
The only clear pattern discernible from the system of resource allocation in the Northern Region, as illustrated by Figures 8.7 and 8.8, is the over-funding of the urban districts, Tamale Municipal, Savelugu-Nanton and Yendi and under-funding of the most deprived district, Gushiegu-Karaga.
195 Chapter 8, Section 2. Summary of key points
Deprivation across districts in Ashanti Region is as high as among districts in the Northern Region. Four of the five most deprived districts in the two regions combined were in the Ashanti Region. Similarly, three of the five least deprived districts were from the Ashanti. This indicates that deprivation is not a region-specific phenomenon and may be as prevalent and severe in richer as in poorer regions.
Districts with rural characteristics are not always remotely located, according to the findings of this study. Some peri-urban districts were found to have more rural characteristics than remotely located districts. This finding highlights the need to distinguish between remoteness and rurality in an assessment of deprivation among geographical regions.
Resource allocation within regions differs between the Ashanti and Northern Regions. Allocation of funds in Northern Region benefits the most urban districts, particularly, the Tamale Municipal district. Resource allocation in Ashanti largely does not benefit the urban districts. Apart from the Kumasi Metro district, which was substantially over-funded from 1998 to 2000, there were clear patterns of resource shift from urban to rural districts.
The proportions of variance in the share of funds explained by the EAS were generally low (below 50%) except for GOG allocation in Ashanti Region for 2000 where the proportion of variance was 56%. This indicates that resource allocation in both regions were less equitable.
Population size was the main determinant of districts share of funding in Ashanti Region between 1998 and 2000. There were no traces of per capita funding in 2001 and 2002, particularly, with regards to DPF allocation.
Northern Region relies less on population size as the basis of resource allocation. It follows a more formalised resource allocation system than the Ashanti Region
Ashanti Region is more inclined to allocate resources equally across districts without considering differential health need among districts.
196 SECTION 3: INTER- SUB-DISTRICT EQUITY IN RESOURCE ALLOCATION This section examines the extent to which resource allocation within districts has been equitable in terms of differentially benefiting the most deprived sub-districts. It focuses on two districts – West Gonja and Savelugu-Nanton, both in the Northern Region (see Chapter 6 for district selection criteria).
8.7 Deprivation among Sub-Districts The inter-regional analysis showed that a deprivation measure based on two variables (DID) is largely as effective as a composite index in identifying deprivation among different regions (see Section 8.1.4). Deprivation among sub-districts in the West Gonja and Savelugu-Nanton districts was assessed using a double index of deprivation (DID). Due to lack of data, two variables: access to a hospital and access to a senior secondary school (SSS) were used to develop the DID. The number of kilometres from the main sub-district town to the nearest major hospital in the district (usually, but not always in the district capital), and to the closest senior secondary school, as provided in the summary report of the 2000 Population and Housing Census were turned into percentages. The z-scores were then calculated and summed together to obtain the DID (see Chapter 6).
8.7.1 West Gonja Table 8.28 shows the ranking of sub-districts in West Gonja by their DID scores. The Tuluwe sub-district was the most deprived with the highest DID score of 2.462.
Table 8. 28 West Gonja Sub-districts by Deprivation Scores
Sub-District DID Tuluwe 2.462 Mpaha 1.471 Buipe 0.929 Daboya -0.330 Yapei/Kusawgu -1.771 Damango/Mole -2.761 Note: The higher the DID Score the greater the deprivation level in the sub-districts
197 The Damago/Mole emerged the least deprived with the lowest DID of –2.761. Being the sub-district where the West Gonja district capital (Damango) is located, the high negative DID score was not surprising. Figure 8.9 illustrates the variations in deprivation levels.
Figure 8. 9 Deprivation among Sub-Districts in West Gonja
4.000
3.000
2.000
1.000
0.000 DID Score -1.000
-2.000
-3.000
-4.000
le e a ah ipe Mo luw / Mp Bu Daboya go Tu n ei/Kusawgu ap Y Dama SUB-DISTRICT
Note: Positive scores depict high deprivation. Negative scores shows less deprivation
8.7.1 Savelugu-Nanton Table 8.29 shows the DID scores of the sub-districts in the Savelugu-Nanton district. The most deprived sub-district was the JJK with the highest DID of 3.820 while the
Table 8. 29 Savelugu-Nanton Sub-districts by Deprivation Scores
Sub-District DID Janjon Kukuo (JJK) 3.820 Moglaa 1.569 Tampion 0.907 Zoggu 0.415 Diare -0.194 Nanton -1.090 Pong-Tamale -1.762 Savelugu -2.134 Note: The higher the DID Score the greater the deprivation level in the sub-districts
198 least deprived was Savelugu, which has the district capital and a very easy access to the regional capital (Tamale) and its health institutions. The Pong-Tamale and Nanton sub- districts were relatively less deprived compared to sub-districts like Moglaa and Tampion. Although the inter-district analysis (Section 2) revealed that less remote districts could be more deprived than remote areas, the general pattern of deprivation at the sub-district level suggests that remote sub-districts are the most deprived. Figure 8.10 illustrates the level of deprivation among sub-districts in the Savelugu-Nanton.
Figure 8. 10 Deprivation among Sub-Districts in Savelugu- Nanton
4.000
3.000
2.000
1.000
0.000
-1.000
-2.000
-3.000
-4.000
u e a g n la ale g Diar m pio Zoggu o Ta Nanton am M Savelu - T ng Po
Janjon Kukuo (JJK)
SUB-DISTRICT
Note: Positive scores depict high deprivation. Negative scores shows less deprivation
8.8 Equity-Adjusted Share (EAS) The sub-district EAS was developed using the DID scores. The DIDs were normalised to convert the negative scores into positive and used to weight the total sub-districts populations to obtain the EAS.
199 8.8.1 West Gonja Table 8.30 shows the percentage EAS derived for sub-districts in the West Gonja district.
Table 8. 30 Derivation of Equity-Adjusted Share (EAS) for Sub-Districts in West Gonja
Sub-district DID NDID Pop W/Pop EAS% Yapei/Kusawgu -1.771 1.990 31,786 63,257 14.0 Daboya -0.330 3.431 36,327 124,632 28.0 Damango/Mole -2.761 1.000 39,354 39,373 9.0 Tuluwe 2.462 6.223 10,596 65,940 14.0 Mpaha 1.471 5.232 15,386 80,496 18.0 Buipe 0.929 4.690 18,164 85,184 19.0
Total 151,613 458,882 100.0 Key: NDID = Normalised double-index of deprivation. W/Pop = Weighted Population
The Daboya sub-district retained the highest EAS (27%) among the six sub-districts. Although not heavily deprived as the three sub-districts at the bottom half of the table, Daboya has the second largest population in the district and this influenced its EAS percentage. The Damango/Mole sub-district had the lowest EAS (9%) despite having the largest population. This is due to the high negative DID score. Tuluwe, the most deprived sub-district in West Gonja, had EAS of 14% because of its small population.
8.8.2 Savelugu-Nanton Table 8.31 shows the percentage EAS of sub-districts in the Savelugu-Nanton district. The highest EAS of 18% was retained by the Diare sub-district, which had the highest weighted population.
Table 8. 31Derivation of Equity-Adjusted Share (EAS) for Sub-Districts in Savelugu-Nanton
Sub-district DID NDID Pop W/Pop EAS% Savelugu -2.134 1.000 35,111 35,105 16.0 Diare -0.194 2.940 14,016 41,204 18.0 Pong-Tamale -1.762 1.372 11,854 16,267 8.0 Nanton -1.090 2.044 15,649 31,991 14.0 Janjon Kukuo (JJK) 3.820 6.954 2,084 14,492 6.0 Tampion 0.907 4.041 7,810 31,563 14.0 Zoggu 0.415 3.549 5,471 19,419 9.0 Moglaa 1.569 4.703 7,606 35,772 16.0 Total 99,601 225,814 100.0 Key: NDID = Normalised double-index of deprivation. W/Pop = Weighted Population
200 Savelugu and Moglaa obtained 16% each despite the difference of over 27,000 in population size. The JJK, the most deprived sub-district, retained the lowest EAS of 6%, largely as a result of its low population size.
8.9 Assessing Inter-Sub-District Equity in Resource Allocation The inter-sub-district equity analysis was restricted to donor-pooled funds for 2001 and 2002 due to lack of financial data. The same approaches used in the previous sections were used here. That is, a simple regression was used to determine the proportion of variance in the share of funds that could be explained by the EAS and comparison of actual (estimated in this case) and predicted (EAS-based) funds to highlight the equity- gap. Since the same percentage share of funds was used for every year, the regression analysis was done for only 2001.
8.9.1 Actual Allocation of Funds in West Gonja Table 8.32 shows the actual DPF shares for sub-districts in the West Gonja district for 2001 and 200240.
Table 8. 32 Sub-Districts by Actual Donor-Pooled Funds Allocation for West Gonja from 1999-2002 Actual 2001 % Share Actual 2002 % Share Sub-District ⊄‘000 2001 ⊄‘000 2002
Yapei/Kusawgu 12,939 18.0 5,622 18.0 Daboya 17,505 23.0 7,606 23.0 Damango/Mole 16,744 22.0 7,276 22.0 Tuluwe 9,133 12.0 3,969 12.0 Mpaha 9,894 13.0 4,299 13.0 Buipe 9,894 13.0 4,299 13.0
Total 76,109 100.0 33,071 100.0 Note: Total yearly DPF allocations for all sub-districts were obtained from the Regional Health Administration (Tamale).
40 Noted that these figures are estimated and not hard data. The total annual allocations for all sub-districts in West Gonja obtained from the RHA were divided up using the allocation criteria (percentage share of funds) provided by the DHA. The figures may not be exactly the same as the actual amount of DPF received by sub-districts if the DHA did not apply the allocation criterion for sharing the funds.
201 The Daboya sub-district was the main beneficiary of DPF allocations for 2001 and 2002, receiving 23% of total DPF allocation each year. The Damango/Mole sub-district, with the district capital, received 22% share of funding. The district’s most deprived sub-district, Tuluwe, was the least funded with 12% share of funds. On the whole, the total DPF allocation for 2001 was more than twice the total allocation for 2002. The inter-district analysis revealed similar situation of high total DPF funding for 2001.
8.9.2 Regression Analysis: West Gonja Since the percentage share of funds for inter-sub-district allocation was the same for every year, the regression analysis was done for only 2001. The EAS and sub-districts’ share of the total population were regressed as independent variables on the percentage share of funds as the dependent variable. The summarised results (Table 8.34) show that the EAS accounted for less than 1% of the variability in the share of funds. The model was not statistically significant as evidenced by the high p-value (p=0.772).
Table 8. 33 Regression of Equity-Adjusted Share and Population on Percentage Share of Donor-Pooled Funds (DPF): 2001 Independent Variable R-Square Unstandardised Sig. (P-Value) Coefficient DPF: 2001 EAS 0.023 0.121 0.772 Pop% 0.907 0.577 0.003**
Key: *Coefficient and p-values significant at 5%. **Coefficient and p-values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%). Dependent Variable = Percentage Share of Actual Funds (GOG-3)
Population size, however, accounted for almost 91% of the variability in the share of funds. In summary, the regression analysis suggests that inter-sub-district funding in the West Gonja district is almost entirely based on population size. On the basis of the EAS, allocation in the district was largely inequitable.
202 8.9.3 Comparison of Actual and Predicted Allocations in West Gonja Table 8.34 shows the predicted (EAS-based) share of funds and the percentage of funding difference for 2001 and 2002. The actual shares of funds are already presented in Table 8.32.
Table 8. 34 Sub-Districts by Predicted (EAS-Based) Share of Donor-Pooled Funds in West Gonja: 2001-2002 Predicted Funding Predicted Funding Percentage of Percentage of Sub-District 2001 2002 Funding Funding ⊄’000 ⊄’000 Difference 2001 Difference 2002 Yapei/Kusawgu 10,492 4,559 23.3 23.3 Daboya 20,671 8,982 -15.3 -15.3 Damango/Mole 6,530 2,838 156.4 156.4 Tuluwe 10,937 4,752 -16.5 -16.5 Mpaha 13,351 5,801 -25.9 -25.9 Buipe 14,128 6,139 -30.0 -30.0
Total 76,109 33,071 Note: Predicted shares were obtained by redistributing the total yearly (actual) allocations using the EAS. See Table 8.32 for the actual shares. See Chapter 2 for exchange rate of the Ghanaian Cedi to the US Dollar.
Comparison of actual and predicted (EAS-based) shares of donor-pooled funds shows that the Damango/Mole, the most urban and least deprived sub-district, received over 156% of donor-pooled funding above its predicted (EAS-based) share.
Figure 8. 11 Percentage Difference between Actual and Predicted Shares of Donor-Pooled Funds in West Gonja District: 2001–2002
200
150
100
50 2001 2002
0
-50 %Deviation from Equity Target (EAS)
-100
a e u oy ab lew Buipe D Tu Mpaha ango/Mol Yapei/Kusagu am D SUB-DISTRICT
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
203 The Yapei/Kusawgu sub-district was allocated 23% more than its ‘equitable’ share of funds. The four remaining sub-districts were under-funded, receiving less than their EAS-based allocations.
8.9.4 Actual Allocation of Funds in Savelugu-Nanton Table 8.35 shows the actual and percentage DPF allocations to sub-districts in the Savelugu-Nanton district for 2001 and 2002. More than one-third (35.4%) of the total yearly allocation went to the Savelugu sub-district. The Zoggu sub-district received the least allocation (5.3%). The most deprived sub-district, JJK, received nearly 11% of the total funding. This was quite substantial considering that the JJK has only about 2% of the districts population. As observed throughout the study, the total DPF allocation for 2001 was strikingly high (more than double of the 2002 allocation). Possible reasons for this have been discussed in Chapter 10.
Table 8. 35 Actual Donor-Pooled Funds Allocation to Sub-Districts in Savelugu-Nanton 2001-2002 Actual % Share Actual % Share Sub-District 2001 2001 2002 2002 ⊄‘000 ⊄‘000 Savelugu 18,603 35.4 8,112 35.4 Diare 6,334 12.1 2,762 12.1 Pong-Tamale 5,577 10.6 2,432 10.6 Nanton 7,133 13.6 3,111 13.6 JJK 5,735 10.9 2,501 10.9 Tampion 3,338 6.4 1,456 6.4 Zoggu 2,791 5.3 1,217 5.3 Moglaa 3,054 5.8 1,332 5.8
Total 52,567 100.0 22,923 100.0 Note: Total yearly DPF allocations for all sub-districts were obtained from the Regional Health Administration (Tamale). Actual allocations to individual sub-district were estimated by dividing up the total allocations using the resource allocation criteria obtained from the District Health Administration.
8.9.5 Regression Analysis: Savelugu-Nanton As in the West Gonja district, the EAS accounted for less than 1% of the variability in the share of funds (Table 8.36). The model was not statistically significant (p=0.513). Population size was the main determinant of DPF allocation, accounting for nearly 87% of the variability in the share of funds.
204 Table 8. 36 Summarised Results of Regression Analysis: Donor-Pooled Funds (DPF): Savelugu- Nanton 2001 Independent Variable R-Square Unstandardised Sig. (P-Value) Coefficient DPF: 2001 EAS 0.074 0.535 0.513 Pop% 0.866 0.882 0.001**
Key: *Coefficient and p-values significant at 5%. **Coefficient and p-values significant at 1%. Independent Variables = Equity-Adjusted Share (EAS) and Population (Pop%). Dependent Variable = Percentage Share of Actual Funds (GOG-3)
8.9.6 Comparison of Actual and Predicted Allocations in Savelugu Nanton Table 8.37 shows the predicted (EAS-based) share of funds and the percentage of funding difference for 2001 and 2002 (see Table 8.35 for actual share of funds).
Table 8. 37 Sub-Districts by Predicted (EAS-Based) Share of Donor-Pooled Funds in Savelugu- Nanton: 2001-2002 Predicted Funding Predicted Funding Percentage of Percentage of Sub-District 2001 2002 Funding Funding ⊄’000 ⊄’000 Difference 2001 Difference 2002 Savelugu 8,289 3,615 124.4 124.4 Diare 9,729 4,243 -34.9 -34.9 Pong-Tamale 3,841 1,675 45.2 45.2 Nanton 7,554 3,294 -5.6 -5.6 JJK 2,668 1,163 115.0 115.0 Tampion 7,453 3,250 -55.2 -55.2 Zoggu 4,585 2,000 -39.1 -39.1 Moglaa 8,447 3,683 -63.8 -63.8
Total 52,567 22,923 Note: Predicted shares were obtained by redistributing the total yearly (actual) allocations using the EAS. See Table 8.36 for the actual shares. See Chapter 2 for exchange rate of the Ghanaian Cedi to the US Dollar.
Two sub-districts: Savelugu and JJK, were substantially over-funded, receiving more than 100% of their predicted equity-adjusted shares. The biggest loser was the Moglaa sub-district, which received about 63% below its ‘equitable’ share. The equity-gap is highlight by Figure 8.12.
205 Figure 8. 12 Percentage Difference between Actual and Predicted Shares of Donor-Pooled Funds in Savelugu-Nanton District: 2001–2002
200
150
100
50 2001 2002
0
-50 %Deviation from Equity Target (EAS) -100
n io JJK glaa elugu Daire amale o T Nanton amp Zoggu M Sav g- T n Po SUB- DI ST RI CT
Note: Positive values show the magnitude of funding gained in excess of the predicted equity-adjusted shares. Negative values depict magnitude of funding lost.
Chapter 8, Section 3. Summary of key point
Equity in resource allocation from district to sub-districts in Ghana is difficult to assess due to 1. Lack of socio-economic data at sub-district levels, and 2. Lack of financial or expenditure data at district and sub-district levels
A double-index of deprivation was used to assess levels of relative deprivation among sub-districts in the West Gonja and Savelugu-Nanton districts
In both the West Gonja and Savelugu-Nanton districts, the most urban sub-districts receive more than equitable shares of funding.
Population size is the main determinant of actual share of funds for sub-districts in the West Gonja and Savelugu-Nanton districts.
206
CHAPTER 9
BEHIND THE ALGORITHMS: FACTORS INFLUENCING EQUITABLE RESOURCE ALLOCATION IN GHANA
“Look, there are so many problems with allocation of funds…we in this district, sometimes we get less than half of what other districts get and if you ask why, they don’t give you any proper explanation…I really don’t know the basis for sharing the money, but something must be wrong with the system” (District Director of Health Services).
Overview The above statement from a displeased District Director of Health Services sums up the contentious nature of resource allocation in the Ghanaian health sector. The extent of equity in funding allocation at national, regional and district levels of the health system was explored in Chapter 8. This chapter presents the qualitative findings, which provide insight into factors that influence the distribution of resources in the Ghanaian health system. In particular, this chapter sets out to answer the following questions covered by the main interview domains:
How important is equity for policy makers and managers and how does this influence the way resources are allocated at national, regional and district levels? Has the level of funding of the health system any influence on the equitable distribution of resources? Are the mechanisms for disbursing and accessing funds in keeping with the policy to shift resources to more deprived areas? Does the system of financial accountability promote equity in resource use? What role does politics play in health resource allocation in Ghana? What are the manifestations of this in relation to the resources distributed? What influence do donors have on resource allocation? Do they use this influence? How? Why? When? What role, if any, does the local government play in shaping resource allocation at the district level?
207 9.1 Equity and Equitable Allocation of Resources Equity has dominated the health landscape of Ghana over the past decade (Chapters 2). At the policy level, the emphasis is on equitable access to services and the redirection of resources to deprived areas where health status is comparatively poor (MOH 2002). Three main themes and a number of sub-themes emerged from the interviews as influencing resource allocation under the equity domain. These are: equity as a national health priority, manpower availability, and local capacity to utilise funds efficiently. These and their sub-themes are discussed below.
9.1.1 Equity as a National Health Priority Respondents at all levels of the health system consistently noted the prominence given to equity as a national health policy goal. The majority of the respondents were aware of equity as a national health priority and some of the strategies to improve equity. At the national and regional levels, most respondents referred to the GPRS and the second 5- Year Programme of Work (POW II) as the basis for the equity strategies currently being pursued. References were also made to the specific policy of targeting extra resources at the four most deprived regions in Ghana. Box 9.1 presents some of the comments made by participants regarding equity as a national priority.
Box 9. 1 Equity as a national health priority
Extract 1. Equity as a national health priority “Equity is within the policy. The 5-Year POW II has as its main theme - bridging the gap of inequalities. So policy-wise, this is very keen. Now, a lot of the things to do with addressing equity lie very much at the lower levels. Because up here you can only work with the policy, thus, you can only shift the money but addressing the real issues with the money shifted is at the lower levels. And it is difficult to be managing it right from the top, although definitely a lot of things depend on the money” (Policy maker, GHS, 2004).
“There is no doubt about that! Equity is the number one priority issue of the MOH at the moment. Have you seen the five-year programme of work? Even the title itself - ‘bridging the inequality gap’ tells you that equity is the main issue at the moment. If you go through the document you’ll find the word equity all over; it is the most important issue at the moment” (Donor Representative, 2004).
“Access is the main problem so the policy is to improve equity in access particularly for the most vulnerable groups like children under-5, pregnant women and the elderly. In this district we implement the exemption policy to make sure that these people are exempted from paying when they seek health care, but I must add that it has not been working well in many areas” (District Director of Health Services, Ashanti Region, 2003).
208 The above observations clearly demonstrate that equity is a well-entrenched national health policy goal in Ghana. At the district level, most respondents defined equity in terms of access to services, which was consistent with the national policy. Many district officials mentioned the exemption policy that allows the most vulnerable to access health care without paying, as one of the main strategies ensuring access to services for the poor. Policy makers at the national level, were not keen to stress the role of the exemption policy in ensuring equity, most probably because of the perceived failure of the policy. They, however, mentioned the efforts being made to address geographical inequities by shifting resource to the four most deprived regions.
The informants also indicated that the policy of targeting the four deprived regions with extra funding was gradually changing. Policy makers explained that there had been increased realisation that many districts outside the four regions are also very deprived and deserved special attention. As a result, the monies set aside for targeting the deprived regions were eventually being shared among all regions to tackle inter-district deprivation. They indicated that the four most deprived regions received a greater proportion of these extra funds. While this is an important issue, given the widespread deprivation in Ghana, the policy makers provided no insights into how they do this.
9.1.2 Manpower Availability The data analysis revealed that manpower availability is a crucial factor driving resource allocation. Ghana faces a severe problem of brain drain of health professionals at all levels of service delivery (MOH 2002). In recent years, there has been pressure on government and policy makers to tackle the issue and this has led to the introduction of several incentive schemes, some of which have had significant impact on patterns of resource allocation. The general view expressed by participants at all levels was that health workforce shortages pose the biggest threat to effective implementation of the equity-oriented policies (Box 9.2). At the national level, policy makers and donors mentioned that funds could not be allocated to areas where there were no personnel to utilise it.
209 Box 9. 2 Views on manpower shortage and incentive regimes
Extract 2. Personnel availability and resource allocation “The main constraint is the human resource issue. You need to have the personnel and the activities going on in the area before you can allocate funds to that place. If you take medical doctors, for example, you can see that as you travel further north of the country the number of doctors gets thinner and thinner and as I said earlier, we cannot just allocate the resources to areas where there is no staff to use it efficiently. If you do that you run the risk of just wasting the money. I would say that currently some areas are not getting enough funding because there are simply no personnel to use the money to provide the services if you allocate the money to those areas” (Policy maker, GHS, 2004).
Extract 3. Incentive to attract personnel to under-served areas “One of the things that we have done recently is to send some of the incentive package to places we want to get staff shifted to, with the hope that this will attract people to be interested in taking up appointments there….We are trying to use all these innovative ways to see if we can attract people to the deprived areas before we give them the inputs to do the work” (Policy maker, MOH, 2004).
The observations made by respondents (Extracts 2-3), apart from highlighting how resource allocation is driven by personnel availability, also re-states the issue of inequitable distribution of health personnel in Ghana. Data from the MOH indicate that the largely urban southern regions have relatively more health personnel than the predominantly rural northern regions (see Chapter 2). With a significant proportion of the health budget going into personnel emoluments (Chapter 7), one could argue that regions with fewer health personnel are missing out on funding.
The introduction of the Additional Duty Hour Allowance (ADHA) as an incentive to augment the salaries of health workers also came out strongly as having considerable impact on resource allocation. Available evidence suggests that it has decreased real per capita allocation for non-salary recurrent expenditure. For example, for the first two quarters of 2001 and 2002, there was a 9% increase in real per capita expenditure on personnel emoluments while real per capita non-salary expenditure decreased by 4.5% (MOH 2003). There are other incentive packages besides the ADHA to motivate health workers and attract personnel to deprived areas. An official of the MOH confirmed that vehicles have been used to lure personnel to rural and deprived areas. These vehicles are sold at a highly subsidised price to doctors and are paid by instalments.
Many respondents from rural districts, however, expressed reservations about the incentive regimes, which they claimed offer no special reward for health workers in remote communities who are not doctors. In a group discussion with three sub-district heads in a remote district in Northern Region, members narrated the difficult conditions
210 in the villages they have been working in for many years without “any special reward for their sacrifice”. Box 9.3 recounts the story of one of the group members, a nurse with over 23 years experience, who had worked mostly in remote communities.
Box 9. 3 Account of a nurse/sub-district head about working conditions of village health workers
Extract 4. “There is no incentive for working in the village, especially if you are not a doctor; you don’t get promotion and your salary is very little. I completed my nursing training in 1978, twenty-five good years, I have had only two promotions – one, as a midwife after going for a midwifery course, and two, as a nursing officer, that’s all! My salary is just 704,000 cedis (about US$80) a month. You spend all your life in the bush for 704,000 a month. How do you take care of your children who must have to go to school in the city because there is no secondary school in the village? The ADHA, even our colleagues in the city get more than us. I know this because my brother is a nurse and he’s in the city. It is not fair, please tell them we all want to work in the city” (A Nurse/Sub-district Head, Northern Region, 2003).
Accounts of the harsh working conditions faced by health workers were particularly common in most rural districts in both the Ashanti and Northern regions. It contradicts the accounts of policy makers at the national level who indicated that incentive packages have been designed to motivate health workers in deprived areas. It is quite clear that medical doctors are the targets of the incentive packages for deprived areas.
9.1.3 Capacity to Utilise Funds Efficiently Capacity to utilise funds efficiently emerged as a crucial factor influencing resource allocation. This is closely related to manpower availability, though, the emphasis here is more with the financial management capacity of staff rather than availability of personnel. It was evident from interviews with policy makers that the capacity of districts to utilise funds efficiently is a matter of grave concern in the health system. All national and regional level respondents mentioned that funds could not be allocated to areas where it cannot be efficiently utilised. Box 9.4 presents some of the views expressed in relation to capacity to utilise funds.
211 Box 9. 4 Capacity to utilise funds efficiently
Extract 5. Lack of qualified personnel “I think we need to be careful about equitable allocation of funds; the thing is, you can’t just allocate resources to areas where there is no staff or people with the capacity to utilise it efficiently. We have been trying to upgrade the capacity of the district staff but - some rural areas still don’t have the staff. So these are some of the issues that we are working through, you need to have the personnel and the activities going on in the area before you can allocate funds to that place. I would say that currently some areas are not getting enough funding because there are simply no personnel to use the money, others don’t have the personnel with the requisite capacity” (Policy maker, GHS, 2004).
Extract 6. Inability to use all allocated funds “I wish you could check the financial statement of these people (managers) who are complaining, you’ll be surprised to find that there are huge balances ‘sitting’ in the accounts of these same people who are complaining that they don’t have enough money. We have the bank statements of all of them and we know the sort of balances standing in their account at the end of every year, so why complaining that they don’t have enough money to run their activities" (Policy maker, MOH, 2004).
Policy makers made references to the extent that even some districts that have been approved as having the capacity to manage funds still do not utilise funds efficiently. One policy maker revealed that, while district managers complained of inadequate funding, there were often substantial balances in their accounts at the end of the year (Extract 6). He attributed the situation to the inefficiency and limited capacity of some managers to utilise all funds available to them for the year. Several other factors contribute to the high cash balances at the end of the financial year. These include the late release funds, difficulty in accessing funds and liquidity problems faced by some of the banks (Annual Review 2002).
The importance of efficient utilisation of funds for achieving equity was also underscored at regional levels. One regional manager mentioned that they (in their region) are working hard toward ensuring that funds are equitably distributed. But their success depended on the capacity of districts to utilise the funds: “we are making sure that we share the funds equitably but sometimes the success of all this is also based on the capacity of the district to utilise these funds”. His assertion, to some extent, blurs the perceived divide between equity and efficiency. This was not the case with internally generated funds (IGF) as demonstrated below.
212 9.1.3.1 Efficiency and IGF Revenue Internally generated funds (IGF) supported about 13% of health delivery in Ghana in 2001 (Danida 2002). Due to differences in socio-economic conditions and manpower capacity, some districts generate more revenue internally than others (Chapter 7). However, variations in IGF revenue across districts were not taken into account in budgeting and resource allocation. It was apparent from the interviews that this was due to efficiency concerns. Policy makers and regional managers responsible for resource allocation indicated that, as much as they wanted to factor differential IGF revenue in the resource allocation process, they were cautious to make that move because they did not want to reward inefficiency (Extract 7).
Box 9. 5 Efficiency concerns and IGF
Extract 7. Budgeting on IGF “In fact the overall view is that IGF should play a part in determining how much resources a particular district should get. But there is also a counter view that we should not reward inefficiency. That is, we should not reward districts or facilities that are not able to generate funds internally due to their own inefficiencies or because they probably know if they generate less that will help them get more allocation” (Regional Director of Health Services, 2004).
“This has been a difficult thing to do because the moment you do that they become inefficient in generating revenue…We have tried incorporating IGF revenue into the budget before, and there were a whole lot of stories. But the fact is that they are charging for all sorts of things at the hospitals and the government is still subsidising. The Ministry of Finance is very serious about this issue now, what they are saying is that they want to give us a ceiling that includes the IGF” (Policy maker, GHS, 2004).
They believed that some districts were not generating more IGF due to their own internal inefficiencies. If districts generating more IGF were allocated fewer public funds and vice versa, for the sake of equity, it might be seen as a reward to those districts generating less due to their own inefficiencies and discourage those generating more from continuing to do so. This illustrates the dilemma faced by policy makers in choosing between equity and efficiency goals.
9.2 Level of Funding for the Health System The level of funding for the Ghanaian health system is very low even in comparison with other developing countries (see Chapter 7). Key stakeholders from all levels of the health system, including donor representatives, acknowledged that, given the enormity
213 of health problems confronting Ghana, the current level of funding is inadequate and that this affects the way monies are allocated to programmes and activities. However, there were clear differences in opinion at various levels. Generally, policy makers and regional managers admitted that the low level of funding of the health system was not good for the country, but they were very pragmatic about the level of funding that the country’s economy can support and sought to attribute the low per capita funding to the increase in Ghana’s population (Box 9.6).
Box 9. 6 Views of policy makers and donors on level of funding
Extract 8. Level of funding and the economy “It is obvious that the health system is under-funded, that is why one of the main goals of our current policy is to improve financing of the health system by pushing for more government’s allocation to the health sector. But I think we also have to understand that government’s ability to fund the health system also depends on the state of our economy. If revenue generation is not good, how much can the government do? It is not only a problem in Ghana; all developing countries are in the same situation. We have to understand that, at this stage of our development, it is not everything that we can afford” (Policy maker, GHS, 2004).
Extract 9. Per capita funding “The figures show that, in terms of per capita spending, Ghana is not doing enough, but I think since we started the 5-Year POWs, allocation from the government has increased. The issue is that per capita expenditure has been falling in real terms because of inflation and exchange rate problems (the weak cedi against the US dollar). We hope that the stability we are seeing in the cedi in the past few years will help improve the figures. But we must not also forget that the country’s population is increasing at a faster rate and this also affects our per capita spending on health…in 1984, we were about 12 million but now we are about 20 million” (Policy maker, MOH 2004).
Extract 10. Donors “The health sector is so under-funded that no matter how much you do, you might still not be able to reach everybody” (Donor Representative, 2004).
The pragmatism expressed by policy makers was lacking at the operational level where managers were critical and frustrated at times about the low level of funding, particularly from government sources. Most district directors and accountants mentioned that what they receive from government sources (GOG) was too little for the health care demands in their respective districts (Box 9.7).
214 Box 9. 7 Views of district managers on level of funding
Extract 11. Level of government funding “The money is nothing to write home about. For example, this year, our first quarter GOG administration was just a little over 2 million cedis. How can we run this office with 2 million cedis? We talk about computer age and all those things; you give 2 million cedis for a whole quarter, just one cartridge for a printer will cost you nearly 800,000 cedis, what about toner for the photocopies and others? The money is grossly inadequate!” (District Accountant, Northern Region, 2003)
Extract 12. Extra funding “…. The issue of extra funding is not even contemplated. The thing is that, even the ceiling they have allocated to you, you don’t get anything near it, let alone extra. For instance, last year we got less than 50% of our ceiling, so how can you think of extra funding? I don’t know, they say everything depends on how much revenue the government generates. So nobody has ever told us anything about extra funding, but when you don’t even get 50% of your ceiling, how can you think about putting in a request for more? It has not even occurred to me” (District Director of Health Services, Northern Region, 2003)
Extract 13. Effects on planned activities “The real problem is the amount of money they allocate to us…. You make your activity plan and budget based on the ceiling that is allocated to you, but you end up getting far less than that ceiling, so at the end of the day you have to sit down and re-plan those activities you wish to undertake and re- allocate the monies you have received to fund these activities. It’s like this…because the money is not enough you are always struggling to re-allocate the money to those activities that are really important and unfortunately there are other important activities you can’t undertake because the money is simply not enough” (District Director of Health Services, Ashanti Region, 2003).
The level of frustration by operational managers about the low level of funding is apparent. They do not seem to understand why so little is allocated to them for the huge task they face. Implicitly, they seem to put the blame on the government for not doing enough as illustrated in the comments above. Responsibility for the problem of inadequate funding was directed at GOG and not the donors. Asked how the inadequate funding affects their work, particularly the way they allocate resources, majority of operational managers explained that more often because they do not receive enough funds for their planned activities, they have to re-budget and re-allocate funds in order to make do with what is available (Box 9.7).
It was clear from the interviews that poor funding of the health system influences the resource allocation process at the district level by limiting available resources and preventing many planned activities from being carried out. While the majority of the operational managers interviewed seemed to imply that inadequate funding could be traced to government and could to be fixed, senior managers and policy makers focused their concerns elsewhere. They attributed the problem to the weak Ghanaian economy. Although, like the operational managers, many expressed concern about the potentially
215 negative impact of limited funding on health delivery, they were more appreciative of the difficulties confronting the government in the area of revenue mobilisation, which, according to them, limits the government’s ability to fund the health sector. Some indicated that there were signs of improvement in the economy and expressed the hope that such improvements would be sustained and would eventually translate into more public funding for the health system (Box 9.6).
One can argue that the policy makers were more realistic because they worked more closely with central government and had better insight into the financial constraints confronting Ghana. Conversely, the lower level managers were less realistic about government’s ability to sufficiently fund the health sector probably because they are less involved in the policymaking process and less informed about how resources were mobilised and channelled. Even though the regional officials indicated that all district directors participated in the process of designing the criteria used in allocating funds to them, many district directors interviewed were uncertain about how these criteria were developed (see opening extract, page 1), and had no idea about the basis of resource allocation at the national level. The lack of effective interaction between the national and lower levels may explain the lack of appreciation of the difficulties government face in terms of mobilising resources to fund the health sector.
9.3 Mechanisms for Disbursing and Accessing Funds Financial management has undergone considerable reform in recent years (see Chapter 7). Policy makers, managers and donors revealed that the way funds flow from the MOH to budget management centres (BMCs) had considerable influence on the pattern of resource allocation. Three crucial issues emerged: timing of release of funds, procedural inefficiency with regards to accessing GOG administration, and government accounting and financial regulations.
9.3.1 Timing of Release of Funds The MOH disburses funds at quarterly intervals based on approved POW, as well as on activity and budgetary plans of BMCs. There was a general admission by respondents at
216 all levels that the flow of funds has been erratic over the past years, particularly in relation to GOG item 2. District level managers, in particular, were critical of the irregular flow of funds and its effects on their work. All district managers mentioned that the persistent delays with disbursement of government funds made their work difficult. Box 9.8 highlights concerns with the untimely release of funds for administrative expenditure and the subsequent dependence on donor pooled funds for most of the second quarter.
Box 9. 8 Views on the erratic flow of funds to districts
Extract 14. Timing of releases “The timing is bad; I mean bad, really bad! They are not regular at all, you know, we understand that the budget has to go through some process, say from the district to the region, to national, then it will go to Ministry of Finance and probably parliament has to approve it before the money can be released. But that is no excuse, something has to be done about it; they should find a way to solve it. Sometimes you stay up to June and nothing has come, meanwhile, that is half the year gone so what services are you going to render and with what? Sometimes it is so demoralising you just don’t know what to do” (District Director of Health Services, Northern Region, 2003).
Extract 15. Poor timing and dependency on DPF “I would say that, most often if it were not DPF, health administration in this district would have come to a standstill. If we were going to rely on the GOG 2 for administration of health delivery in this district, then you would have come to meet all of us sleeping. The reason is that funding from government doesn’t come in time. It throws all our plans in disarray. We do what we have planned to do in the 1st quarter mostly in the 2nd quarter and the 2nd quarter in the 3rd quarter. As for 4th quarter activities, if we wait for GOG we won’t do it because the funds are normally not received. We depend mostly on the DPF for the 1st and parts of the 2nd quarter; thus, if only we have balance brought forward from the previous year” (District Health Accountant, Northern Region 2003).
Comments of this nature were common across districts and regions. At the time of the interviews in December 2003, many districts were still seeking to obtain their third quarter allocations for administrative expenditure and most respondents were doubtful if fourth quarter allocations would be released before the end of 2003. Most participants, however, appeared satisfied with the flow of donor funds compared to government funds. With the exception of one interviewee who said funding flows from all sources have been similarly unpredictable, all other participants expressed relative satisfaction with the release and flow of DPF (Box 9.8). It is obvious from insights they provided that much of the district health delivery was undertaken with donor-pooled funds, particularly in the first quarter, when release of government funding was delayed.
217 At the regional level, participants confirmed that there have been delays with the release of funds from the government sources, particularly funds for administrative expenditure. However, they were not as critical as their district counterparts. At the national level, the issue of delays in releasing funds was not painted as grimly as in the districts. Although no respondent denied the fact that there have been problems with the flow of funds, some played it down, depicting it as a temporary problem arising from the financial reforms being implemented in the health sector. Box 9.9 illustrates one such comment.
Box 9. 9 Views on erratic funding flows at the national level
Extract 16. “Generally…we have had some few problems over the years with the timely release of funds; it hasn’t been only for GOG but also with the donor funds. Government being government…revenue mobilisation slowing down, sometimes you get unexpected problems, etc; that has been the main issue. But at the moment, we are working with our partners to improve the situation and they are a bit flexible on that. Usually, it is at the beginning of the year that we have problems with government funding, because revenue inflows are not good at that time and this causes the release of government funds to delay. The donors have agreed that they can frontload their contributions. For example, in the first quarter, they can give us two quarters (i.e., half-year disbursements) so that when government funds delay; there will be some money available for us”(Policy maker, GHS, 2004).
In general, the reaction of national level respondents to the issue of delays was a combination of acknowledging the existence of the problem and downplaying its potential effects on service delivery. Another senior MOH official explained that, in a situation where allocation for the fourth quarter was not released, it was disbursed the following year as the first quarter allocation and the second quarter allocation was pushed back to the third quarter, and so on. He indicated that, despite the delays, BMCs always receive funds four-times in a year. He insisted that a pattern has been created whereby the cycle of government allocations usually lags one quarter behind, so it is up to managers to plan within that pattern. He also expressed the view that the complaints from district managers were sometimes an excuse for non-performance. This appeared an effort to downplay the problem of persistent delays in release of funds.
9.3.1.1 Coping with the Delays Several key strategies emerged from the data as mechanisms used by district managers to cope with the erratic funding flows. For example, all managers reported obtaining
218 supplies on credit and paying for them when funds were received. Since the DHA deals more with public health issues such as immunisation, health education and disease prevention, outreach programmes are the core of their activities. Fuel, therefore, is one of the most essential items of expenditure in their operations. All participants at the district level reported that, in order to continue operating, they obtained fuel and other supplies on credit, which they paid for when funds arrived (Extract 17).
Box 9. 10 Strategies use to cope with erratic funding flows
Extract 17. Obtaining supplies on credit “The issue is that because the situation is not new our suppliers have come to understand it and so they supply us things on credit and when the money comes we pay them. It’s a big problem because it is not easy to be crediting things all the time; there is always a limit to what somebody can supply to you on credit” (District Director of Health Service - Ashanti Region 2003).
Extract 18. Internal cash borrowing “The district hospital has problems but at least they have some IGF to fall on when things delay. Since we are here together, they understand the difficulties we face, so we borrow from them from time-to- time and pay back when we get our money” (DHA Accountant - Ashanti Region 2003).
Extract 19. Conserving donor funds “We depend mostly on our DPF balance for the first and parts of the second quarter. That’s if only we have balance brought forward from the previous year. But we try our best to, at least, save something from our fourth quarter DPF allocation otherwise we’ll be in big trouble” (DHA Accountant - Northern Region, 2003).
Apart from obtaining supplies on credit, there were reports of internal cash borrowing from the district hospitals’ internally generated fund (IGF). DHAs in Ghana generally do not themselves generate funds internally. However, the sub-district facilities and the district hospitals generate and retain all IGF revenue. Three district directors and their accountants indicated that, occasionally, the DHAs borrow money from the district hospital in order to continue with their activities (Extract 18). This practice appeared widespread in districts where the district director of health services was also the only or key medical officer at the district hospital.
Based on past experience of delays in funding, particularly in the first quarter, district managers often try to lessen the effects of the erratic flows on their activities by taking “pre-emptive” action. Over two-thirds of the respondents mentioned that government accounting regulations do not allow them to carry forward to the following year any balances in their accounts for government funds. Such balances are automatically
219 forfeited after 31 December, when the financial year closes. Consequently, whenever there are any cash balances on government funds in December, they use it to purchase materials that would be needed in the first quarter to avoid forgoing such funds. Even though in principle, this is against government regulations, in practice, managers see it as necessary to keep the health system partially running in the first quarter.
Finally, the majority of district managers confirmed that DPF was the most reliable source of funding for the district health delivery. They spoke enthusiastically about how donor funds helped them to partially operate in the first quarter. A manager in the Northern Region provided the following contrast between the DPF and government funds:
“I can tell you that without DPF the health system would have collapsed long ago. Government funds are too small and very unreliable. You never know when they are coming and cannot be sure how much you are getting”.
In coping with the delays, more than half of the respondents revealed that they do their best to conserve some of their donor funds for the fourth quarter for use in the first quarter (Extract 19).
9.3.2 Procedural Inefficiencies The second issue that arose strongly from the interviews was the cumbersome procedure that BMCs go through in accessing the government funds for administration expenditure (GOG 2). As indicated in Chapter 7, the GOG 2 is disbursed in the form of a warrant through treasuries at the District Assemblies. In principle, BMCs are supposed to access these funds through these local treasuries. However, information from interviews suggests that district accountants always travelled to the national capital (Accra) to process the warrants. The account in Box 9.11 provides insights into the difficulties of accessing government funds for administration expenditure.
220 Box 9. 11 A District Accountant’s Story of the Difficulties in Accessing GOG -2
Extract 20. “Accessing the GOG 2 involves a lot of difficulties because when you get your warrant, you have to apply for expenditure authorisation from the Controller and Accountant General Department (CAGD) in Accra through your regional director and the district and regional treasuries. At the District Assembly, it takes about 3-4 days for them to put a covering letter on the warrants, when you go to the regional treasury - the same problem…. There is no schedule-officer responsible for sending these warrants to Accra, so if you really need the money, you have to send the warrants yourself to Accra, and when you get to the CAGD in Accra, the frustrations there are just too much I can’t even describe it. They will tell you this person is not there, that person who should sign that portion is not around, a whole lot of frustrations!
The last time I went to Accra I spent over a week and couldn’t even have my documents processed. In the end, I had to leave it with a friend to continue following it up for me. When I returned, after some time they sent it to me, telling me it is ready, only for me to discover that the Controller hasn’t sign a certain portion and so we can’t access the money. I had to take it back to Accra - just imagine the transport cost from here to Accra, the risk on our roads and all other expenses! You spend a lot to chase the money, meanwhile our total allocation for the first quarter was just a little over 2 million cedis…. By the time we received it nearly one million, about half of the money is gone; so that is the problem we face - very frustrating, very, very frustrating! At times you feel like just leaving the money and save yourself the trouble of making all those desperate chase for that peanut” (District Health Accountant, Northern Region, 2003).
The above account reveals not only the cumbersome nature of the procedure for accessing GOG funds, but also the inefficiencies in the treasury system. The inability of the local treasuries to liaise with their national counterparts and process the warrants in good time partly compels district and regional BMCs to travel to Accra. National level respondents mostly highlighted the official version of the procedure for accessing funds. At the office of the Financial Controller (MOH), officials were emphatic in their assertion that district and regional BMCs were not required to travel to Accra to pursue the processing of their warrants, as the system was designed to allow them to access these funds through their local treasuries. They, however, admitted that this practice could not be stopped until the system is made more efficient and the time lag between receiving of warrants and accessing of funds through local treasuries is shortened. In general, it could be deduced from the responses by policy makers that the warrant system, though problematic, was part of government financial control mechanisms intended to ensure the prudent use of public funds.
9.3.3 Government Accounting and Financial Regulations The third factor that emerged as influencing the resource allocation process was the stringent government accounting policy that prevents the rollover of un-accessed or
221 unused funds. District officials widely reported that this policy put a lot of pressure on them to try as much as possible to access and spend the GOG funds before the year ends. Most of them indicated that funds for the last quarter are often not released or released so late that sometimes it was impossible to access before 31 December. The majority also mentioned if they were able to access such funds, they tried to spend it before the year ends. Some revealed that since they usually ran on credit, they paid their debts and used the rest to purchase materials against the first quarter when government allocations are often not received as indicated earlier.
Given the persistent delays in releasing funds, some respondents believed that the policy was a deliberate ploy by government to re-claim part of the monies it allocated to solve its own cash flow problems. One respondent described the situation as the government “giving with the right hand and taking it back with the left”. At the regional level, respondents confirmed that the volume of returns from districts increased tremendously in the fourth quarter, particularly towards the end of the year, signifying efforts to spend unused funds before the year ends. The implications of this for prudent use of public funds could be far-reaching and are discussed further in Chapter 10.
By contrast, respondents expressed satisfaction with the funding arrangement of the DPF, which can be rolled over to the following year. The majority indicated that it put less pressure on them to spend. Although no respondent admitted that the rush to spend unused funds before the end of the financial year could lead to misapplication of funds, many mentioned that spending under less pressure was desirable and could lead to more rational decision-making. Most regional officials and some policy makers at the national level also saw the policy of forfeiting unaccessed GOG funds by December 31 as a “licence” for misapplication of funds. They hinted that district managers used the ‘forfeiture policy’ as an excuse to spend the balance of government funds available before December 31. The regional authorities suspected that some managers might be misapplying funds “under the pretext of pre-purchasing materials for use in the first quarter”. They were particularly concerned about the high volume of expenditure in the monthly returns of districts towards the end of the year around November/December.
222 9.3.3.1 Government Regulations and Remote Districts The policy of paying back all GOG funds unaccessed by 31 December to government also has significant equity implications. As evident from the interviews, districts in remote areas were usually unable to access their funds on time, partly because of poor communication links with the regional health administration, which makes transmission of funding information difficult. It is likely that such districts would mostly be unable to access their last quarter allocations before the end of the year if released so late. The comments in Box 9.12 demonstrate how difficult it is for remote districts to access funds.
Box 9. 12 Government regulations and accessing of funds by remote districts
Extract 21. “For those of us in this remote district we are always at a disadvantage, sometime when the money is released, the only way you can hear about it is when you go to the regional health administration…we have no functioning telephones like others. By the time we hear about it somebody in Tamale might have already started with the processing of his warrants. By the time you get to the treasury in Accra, it is already closed for the year and that means you cannot access the money” (District Health Accountant, Northern Region, 2003).
Even with government service budget (GOG –3), which is disbursed in the form of a cheque, there was evidence from the field that sometimes the banks were unable to honour payments due to liquidity problems. This often resulted in balances still standing in the account of some BMCs by 31 December. Again, remote districts where the banking system is poor suffered the most. A narrative by an accountant who was trying to cash a cheque for a National Immunisation Day illustrates the problem (Box 9.13).
Box 9. 13 Account of difficulties faced by remote districts in accessing cheque-disbursed funds
Extract 22. “We just completed our NID last week. We got funds for the programme in the form of a cheque so I went to Tamale to cash it. I went on the 4th and the program was to start on the 5th. I sent my cheque to the bank, and stayed there from 8am to 5pm but the bank had no money to pay the cheque. I stayed overnight just to make sure I can get to the bank early the next day. Knowing the problems with the banks, my director sensed that if they wait for me probably the program wouldn’t start that day, so she had to find a solution. She went to the accountant of the hospital…and luckily, he loaned us some money to get the programme started.
Can you imagine, I still couldn’t get the money that second day; the bank had no money to pay the cheque. I returned from Tamale empty-handed. It was Friday and I had to travel back on Monday before I was able to get the money. So I finally brought the money on Monday evening when the program was supposed to have ended. Imagine what would have happened if director had not got that loan” (District Health Accountant, Northern Region, 2003).
223 The above account demonstrates some of the problems that districts in the remote parts of the country go through in accessing funds that are even disbursed in a cheque form. Having travelled to many of the districts and knowing the condition of the roads and the massive transportation problems facing some DHAs, the possibility that funds released late cannot be accessed before the year ends is high. The government’s accounting policy that disallows unaccessed funds and balances in account to be rolled-over to the following year is therefore likely influence the amount of resources allocated or available to particular districts.
9.4 System of Financial Accountability Another area where efficiency concerns manifested in the data was the system of financial accountability. Financial management and accountability is perceived as one of the successful elements of the health sector reforms in Ghana (MOH 2001). It was evident from the interviews that the system of financial accountability was largely driven by efficiency concerns. While district level respondents expressed few concerns about efficiency in the use of resources, most policy makers were concerned about this. Some policy makers linked the perceived inefficient use of resources to the continuing dependency on donor funding (Extract 23). Others complained about the weakness of the system of accountability in detecting inappropriate use of funds (Extract 24).
Box 9. 14 Efficient use of funds and the flaws in the financial management system
Extract 23. Inefficient use of funds “I believe that if we improve upon efficiency in our use of resources, over the next 10 years we can significantly reduce our dependency on donors. Currently, I think that we are not making the best judgement in the use of our resources and it cuts across all levels of the health system. Sometimes you go to health centre and you see what the money you sent there has been used for, you can’t help wondering whether that is the best way the money could have be used” (Policy maker, MOH, 2004).
Extract 24. Weakness of the accountability system “The driving force has been a really strong financial management reform centred more on accountability. Therefore, the focus has been on accounting for the funds given and not for how you spent it. I think, in a way, our audit system has also encouraged this because auditors go on duty and they just pick a sample of financial reports to see whether you have accounted for whatever you were given. How you used it, for what purpose, whether it is related to the policy or not is not looked at. Once you have all the relevant receipts to account for the amount given, that’s enough” (Policy maker, GHS, 2004).
224 Little evidence of misapplication of funds was obtained from respondents. Nevertheless, concerns about inefficient use of funds, particularly at the lower level, were widespread among policy makers and some regional managers. Potential misuse of the donor- pooled funds for which usage was more flexible was widely mentioned. Similarly, the system of accountability, in particular the auditing procedure, was described by some as not responsive to efficiency concerns because it did not look at how spending relates to policy. Despite these concerns, the system of accountability, in general, seems to favour the achievement of efficiency, rather than equity objectives.
9.5 Politics and Equity in Resource Allocation The allocation of health care resources to promote equity is a political decision (Birdsall and James 1993). The degree of political commitment to shift resources in line with established policy strategies emerged from the interviews as an important factor influencing resource allocation. One issue that strongly brought this to the fore was the continuous investment in ‘big hospitals’ while primary health care strategies such as the Community-based Health Planning Strategy (CHPS) remained largely under-funded. Box 9.15 presents the view of a participant who believed the political commitment to improve equity by shifting resources to deprived rural areas was lacking.
Box 9. 15 Political commitment to redistribute funds to improve equity
Extract 25. Level of political commitment “Personally, I think it is the commitment! There are strategies but the commitment to implement these strategies is the problem. Take CHPS, for example, you have plans to train community health nurses and place them in the community, build some small structures where people in the community can go for minor treatments, which do not need to go to the hospital. Now, you need money for all these, and all of a sudden, you shift all the money into building a ‘big hospital’ in the city. What does that mean? It means you are not committed to your own plans. Technically, the ideas are there, but the political will to release the money to implement these ideas is lacking” (A National Level Participant, 2004).
Extract 26. Limits of policy makers influence “My duty is to tell government of what I know and what I think should be done, and that is about what I can do. Government has got its own agenda and they are politicians and as far as I am concerned, I give what is a technocrat’s advice” (Policy maker, GHS, 2004).
Extract 27. Pressure on politicians from interest groups “Those outside the health sector have contributed to it. If you want serve a community, they want you to put up a big health centre or hospital as they have seen in other communities whereas maybe a small clinic could serve the same purpose. And they put pressure on politicians to provide these structures, and they succumb to these pressures sometimes for their own political survival, so we still have a long way to go” (Policy maker, MOH, 2004).
225 Views of inadequate political commitment to shift resources to where they are needed most because of politics were expressed by several policy makers who, as technocrats, felt powerless to stop the government from shifting the money in line with the political agenda, rather than with planned policy strategies. Extract 26 illustrates the limits of bureaucrats to influence political decisions when it comes to resource allocation.
The influence of politics on resource allocation was also seen as something encouraged by the various interest groups and communities, who put pressure on politicians to deviate from policy objectives and invest in big facilities rather than the small clinics or health post that the policy has envisaged. One regional manager noted that the issue of communities demanding big facilities from government rather than what was necessary to meet their health needs “partly stems from their lack of understanding of health and health care”. He observed that many people in Ghana do not believe that “effective health care can be delivered outside the hospital”. It may also be due to the long neglect of primary health care, which has led to most minor ailments being treated in big hospitals, a very inefficient situation.
9.6 Donor Influence on Resource Allocation Donors are key actors in the Ghanaian health scene (see Chapter 7). Donors contribute about 35% of the total health budget and this has increased over the past decade. The increase in donor funding is no doubt a relief to a government struggling with chronic cash flow problems. However, observers have long expressed concerns about the influence of donors on the resource allocation process and the distribution pattern (Cassels 1997; Killick 1997) given their usually selective system of funding. Perception about donor funding and how it possibly influences the resource allocation pattern was explored in the interviews with policy makers, managers and donor alike. There were different views about donor funding among policy makers and managers. However, two main themes emerged - donor influence on policy development and donor use of earmarked funding.
226 9.6.1 Different Perceptions about Donor Funding Lower level managers and senior policy makers perceived donor funding to the health sector differently. All respondents agreed that donor funding played a vital role in health care financing in Ghana. However, they differed in their perceptions on the usage of donor funds. At the district level, many respondents spoke enthusiastically about the donor-pooled funds, which they saw as crucial to maintaining the viability of health services (Box 9.16). The favourable views about DPF among district level participants were common to all regions and districts, rural and urban. A critical analysis of the views expressed at the district level shows that three issues were of prime concern to respondents and have shaped their opinions about DPF. These include the flexibility associated with the use of DPF, the comparatively regular flow of these funds (DPF), and the fact that balances could be rolled over to the following year.
Box 9. 16 Perception of donor funding at the district level
Extract 28. “In fact DPF has been one of the major sources of funding for the various investment activities; some have built with it, some have procure cars with it, some have renovated with it, the computers that you see are all here because of DPF. I ‘d say it is the lifeline of the sector at the moment. Because the administration and service money, beside the fact that they are not enough, they are very unreliable. Look, within the first 6 months of this year, out of the 4 sources of funding I indicate, DPF is the only one we received” (A District Health Accountant, Ashanti Region, 2003).
“To be very honest with you, I can tell you that without DPF the health system would have collapsed long ago. GOG funds are too small and very unreliable. You never know when they are coming and cannot be sure how much you would end up getting. Even when your budget and your activity plans have been approved, you still end up receiving far less than what they themselves have approved” (District Director of Health Services, Northern Region, 2003).
Most respondents drew a distinction between DPF and GOG with regards to the flexibility of usage, adding that the DPF gives them more room to make their own resource allocation and spending decisions. One respondent indicated that this has improved their confidence in decision-making and financial management. The relatively regular flow of DPF and the easy access to funds was another reason why district level respondents appeared enthusiastic about this source of funding. Many contrasted the ease by which DPF was accessed with the cumbersome procedure of accessing government funds. The favourable perception of DPF among lower level managers was also due to the fact that the end of year balance of DPF could be rolled over to the next
227 year. This provided a buffer for the first and second quarters when government funds were delayed.
At the regional level, respondents were not as enthusiastic about DPF as their district level counterparts, though they acknowledged the important role that donor funding plays in the country’s health delivery. Almost all regional level respondents raised two issues: the fact that DPF was part of the health budget and the flexibility regarding its usage. Many were of the view that, being part of the health budget, DPF should be subject to the same rules and regulations as those governing the use of government funds. These respondents highlighted the possibility of districts abusing the flexibility and misusing funds (Extract 29).
Box 9. 17 Perception of donor funding at the regional level
Extract 29. “People should understand that the DPF is part of the total budget, it constitutes about 30% and therefore you should follow the same rules of utilising it as other budgets. But since it is more or less cash in hand, people turn to use it differently. In a sense, the flexibility is good in that it allows the system to move, you don’t have all those bottlenecks associated with the use of GOG, at least it allows work to go on throughout the year. But I think the danger is also there and we need to find a way of balancing the flexibility so that we make sure the right thing is done before it is too late. There was a time that the system was such that we the regional directors have to counter-sign the cheques, and I was always insisting that - show me how much of it is going into programs and others before I sign. At that time at least there were some ways to check and also make sure that they use it well, but now that we don’t counter-sign the cheques, I think we need to come out with a way of checking them” (Regional Health Manager, Ashanti Region, 2003).
Given that DPF can be used for all expenditures except salaries and there is a backlog of investment activities to be undertaken in many districts, the concerns expressed by regional authorities about flexibility in usage appear to be well-founded. Although some district directors noted that they accounted for every cent of donor funds they received in order to maintain their status of a managing BMC, there were some indications of reasonably large amounts of money going into non-recurrent items such as investments and office equipment such as computers detracting from funds for direct service delivery.
At the national level, the importance of donor funding was widely acknowledged, but many respondents described the situation of having to depend so much on donors for health delivery as “unfortunate”. They largely attributed the problem to the difficult
228 economy circumstances that Ghana and other developing countries have found themselves in (Box 9.18).
Box 9. 18 Perception of donor funding at the national level Extract 30. “I think it is certainly not healthy that we depend so much on donors, but the government and people of this country should also appreciate that currently this country is not in the position to be able to manage its budget. As far as I know, if you take health sector about 40% of our budget is in the hands of donors. It’s frightening, but that is the state of play in our scheme of things now” (Policy maker, GHS, 2004).
“I don’t think any of us is happy with the situation of depending too much on donor funding, but the system as it is now; we have to depend on donors because the inefficiency is too high. And that is why it’s a dilemma as to how far we should accept donor conditionalities” (Policy maker, MOH, 2004).
Some national level respondents were, however, of the view that if the general inefficiency in the use of funds were curtailed in the health sector, Ghana would be less reliant on donors in the near future (Box 9.18).
Donor representatives generally expressed the view that given the current economic conditions in Ghana, it was likely that, in the short or medium term, donor funding would continue to play a major role in the health sector. They, however, mentioned that, under the current health sector POW, the government’s contribution has been gradually rising while donor contributions have been decreasing.
9.6.2 Donor Influence on Policy Development It emerged from the data that one way donors influenced resource allocation and national health policy was through using donor conditionalities. Killick (1997) identified two sets of policy conditionality; ‘hardcore’ and ‘pro-forma’. He defined hardcore as “policy changes stipulated as a prerequisite to the approval of, or continued access to, a grant or loan, or to subsequent assistance” (Killick 1997, p.487). In this case, conditionality becomes the use of financial leverage to promote donor objectives. International agencies, particularly the World Bank, use hardcore conditionality (Walt et al. 1999). Pro-forma conditionality refers to “mutually agreed, or non-significant, or formalistic provisions, which both parties find convenient to write into a programme” (Killick 1997, p.487).
229 The sector-wide approach (SWAp) placed donors in a partnership position with the government (MOH). Although the MOH sees the overall sector-wide approach as internally driven, with strong ministerial leadership and vision (Addai and Gaere 2001), the programme of work, which is the embodiment of national health sector policies and strategies, has to be agreed upon and approved by donors. Evidence from interviews show that the strong financial management system introduced in the health sector was a donor conditionality to ensure that funds they contribute to the ‘common basket’ are well accounted for (Interview data 2004). Donor representatives interviewed confirmed that they do not disburse their funds until they receive the quarterly financial statement, which accounts for what has been previously disbursed.
The interview data revealed that donor involvement in policy and planning was largely confined to the national level. A director at GHS explained that, due to the nature of the planning process, donors have little involvement in planning and budgeting at sub- national levels although they participate in the policy and planning processes at the national level (Box 9.19).
Box 9. 19 Donor influence on policy
Extract 31. “Because of the nature of our planning process, donors are not so much involved. At the national level, what we do is that we have our summit meetings where we present our POWs to the donors, they scrutinise it, we discuss it, and we agree on the things that have to be done. After that they (donors) are not really involved in our own regional and district level planning processes again. So I would say that they are involved in the broader national health planning process that leads to the development of the 5-year POWs, but they are not involved in our own regional and district level planning activities”. (Policy maker, GHS, 2004).
Donor involvement and influence can largely be seen in the frantic efforts made by policy makers to address donor concerns in the policy and planning process and the fact that these policies have to be agreed to and approved by the donors in order to receive funding. A director at the MOH observed that donor concerns could not be ignored because Ghana was not in the position to do away with the financial support they provided. He rationalised the donor influence on policy development indicating that one would not have any problem with donor involvement and influence on policy if they were perceived as partners in health delivery rather than donors.
It was clear from the data that the SWAp reform strengthened donors’ position as partners in health delivery in Ghana. They used their financial leverage not only to hold
230 the government accountable for the way funds are utilised, but also to have their concerns about policies addressed before pledging financial support.
9.6.3 Donor Use of Earmarked Funding The use of earmarked funding by some donors influenced the pattern of resource allocation. Under the pooled funding arrangement, donors contributed funds into a Health Fund (common basket) to finance the agreed POW (see Chapter 7). However, not all donors subscribed to this arrangement. Some donors, such as the United State Agency for International Development (USAID), continued to allocate all their funds to specific projects and programmes while some of those participating in the pooled arrangement still retained parts of their funding for earmarked purposes. The World Bank, for example, contributed 100% of all the projected US$90 million support for the POW II to the pooled account while Danida channeled 75% of its funding for the health sector through the pooled account and retained 25% for earmarked purposes (Interview with donor representatives, 2004).
Interviews with district level managers revealed that donor earmarked funds constrains re-allocation of resources in times of emergency. Whereas the pooled funds can be easily re-allocated to address emergency situations, earmarked funds do not allow that sort of flexibility. A district health accountant in the Northern region recounted how during an outbreak of cerebrospinal meningitis (CSM), the DHA struggled to secure funds from other sources while there were unutilised project funds available (Extract 32).
Box 9. 20 Donor earmarked funding
Extract 32. Limitations of using earmarked funds “In our case, if there is say 20 million cedis sitting in trachoma account and there is CSM outbreak, we cannot use the trachoma money for the CSM. Trachoma has its money for trachoma control, UNICEF has its money for breast-feeding activities; these are earmarked funds, we cannot change their use” (District Health Accountant, Northern Region, 2003).
Extract 33. Scepticism about activities of donors using earmarked funding “You will go to the same district and find 3 NGOs working there, and yet you see nothing; if UNICEF has been working in a district for long time giving ante-natal care, why is anaemia still so high in that district? It is even undermining our work to some extent, because some of them use our staff or our staff in turn go to work for them, and whether they are doing proper work or not - we don’t know because they might not be supervised. It’s a problem” (Regional Director of Health Services, Northern Region, 2003)
231 His comments provide some insight into the degree to which earmarked funding affects allocation of resources at the district level. Some regional managers also expressed reservations about the whole idea of earmarked or project specific funding indicating they have achieved little in terms of improving health (Extract 33). On the whole, in terms of equitable distribution of funds, earmarked funding has the potential of creating imbalances in the allocation pattern, as these funds are not taken into account in the budgeting process and reporting has been found to be inaccurate in some instances (MOH 2003). Notwithstanding these problems, earmarked funds can also be used to enhance equity. Danida’s earmarked funds, for example, are specifically used to fund the cost of fee exemptions in the three northern regions; this is likely to contribute to promoting equity and improving access and health status.
9.7 Collaboration with Local Government (District Assembly) The Ministry of Health has made it a policy to promote partnership for health in its second 5-Year Programme of Work. The thrust of the policy is that “good health is not just a function of health service delivery, hence other government sectors, communities and civil society must be involved” (MOH 2003 p.34). One strategic objective has been building effective partnerships with the District Assembly (DA) in pursuit of reduced health inequalities and better health for all Ghanaians. It emerged from the interviews that the level of collaboration between the health sector and the DA influenced the amount of resources available to DHAs for health activities.
In general, collaboration between DHAs and district assemblies was weak, but varied from district to district. There were accounts of more financial support and infrastructural development from the DA to facilitate health delivery in districts where the relationship between the DHA and the DA was cordial. Conversely, in districts where the relationship was frosty, there were reports of inadequate support from the DA by health managers. However, the main theme that emerged from interviews with district coordinating directors and health was the lack of transparency and misconceptions about funding levels.
232 9.7.1 Transparency and Misconceptions about Funding Levels The interviews revealed that the poor collaboration between health managers and DAs was due largely to the lack of transparency on the part of the DHA regarding how much funding it receives for health delivery in the district. Numerous donor agencies are engaged in the Ghanaian health sector (see Chapter 7). This has created the impression to many outside the sector that funding levels for the sector was significantly high. Health officials in most districts have done little to correct this impression by being transparent with the amount of funding they receive. In interviews with District Coordinating Directors (administrative heads of DAs), most referred to the issue of the health sector getting too much donor money, and hence, did not want be part of the assembly (Box 9.21).
Box 9. 21 Complaints about behaviour of district health managers by assembly directors
Extract 34. “What I have come to realise is that the donor-funded programmes within health is so much and there is a lot of money! You have DFID, you have UNICEF, you have DANIDA, you name them – they are plenty. They have even created a Financial Controller within the MOH; it is just because all these resources come to one field and they have to have a Financial Controller. They tell the regional and district directors, we have this amount of resources so when you are doing your plans, do it this or that way. They invite them to planning sessions, budgeting sessions without even including the district assemblies” (District Coordinating Director, Northern Region 2003).
Extract 35. “We have social services committee meeting which they are supposed to be present, we invite them all the time but they don’t attend, because they don’t wants us to know how much money they have. The only time you’ll see the director or accountant coming here to the assembly is when they need financial of other form of support. How can we offer you financial support if we don’t know how much funds you have. It’s very simple; tell us how much you have so that we will also know how much additional funds we can offer. Look, there are other sectors which also need financial support from us, we are not only dealing with health, we know health is important, but we have financial constraints too, and we need to be careful how we use our funds” (District Coordinating Director, Ashanti Region 2003).
Negative views about the behaviour of district health managers were widely expressed by Coordinating Directors and other district officials interviewed, particularly in districts where the collaboration between the DHA and DA was low. In particular, respondents were very critical of the way in which district health directors and their accountants only turn to the assembly when they need funding and do not want the assembly to know their financial standing (Extract 35).
233 Similarly, several health managers in districts where collaboration with the DA was found to be poor also portrayed the DA in a negative light as unsupportive of health activities. Box 9.22 presents an account of a district health accountant about the lack of support from the DA for health delivery in his district.
Box 9. 22 An account of a district health accountant highlighting the lack of support from the district assembly
Extract 36. “This is my personal opinion; to be honest with you the collaboration between the assembly and us (DHA) has not been good at all. I’ll give you an example; last year there was an outbreak of CSM in this district. The DHA was running up and down… no transport, no funds, we had nothing…It was a hell! Director approached the Assembly for help. The Assembly was just “dilly- dallying”. They were not giving her any attention. Things were getting worse…. At a point director wanted to dip her hands into certain project monies here, but I told her please, if this Assembly is not prepared to help just report the matter to the region. She was a bit concerned of the consequences of reporting the Assembly. You see, the Assembly is the people and the people are the assembly, if the people are dying the Assembly can use whatever funds available to help save lives and report to the government later, but they refused to help us...
The most annoying thing is that they prepare their annual budgets and indicate they are spending such and such amount on health! But as an accountant, I don’t see that amount they state in their budget for health. I have even heard that there is a percentage of their District Assembly Common Fund (DACF) that should be spent on health…don’t know whether it’s 1% or so…nobody sees that money. They only give us something little occasionally for our NID activities if we request” (District Health Accountant, Northern Region, 2003).
Extract 36 clearly demonstrates the tone of relationship between some DAs and health managers. There were misconceptions and mistrust on both sides regarding the level of funding for health services in the districts. From a narrow perspective, it could be argued that the unwillingness of DHAs to declare how much money they are allocated when seeking additional financial support from the assembly is the source of the misconceptions and mistrust. Respondents from the assembly speculated that DHAs did not wish to reveal their levels of funding because they were receiving too much donor money. Likewise, several health officials also mentioned that they have heard of some 1% of the District Assembly Common Fund (DACF) that should be spent on health but they see little of that money; something officials of the DA denied.
One Coordinating Director explained that there was no such mandatory spending on health from the common fund, rather, there was a percentage that was mandated to be spent on social services in the district. The assembly had the prerogative to decide which social services programmes and projects to spend the money on. From a much
234 broader perspective, however, this problem appeared to be the creation of the inconsistencies in Ghana’s decentralisation policy. Despite the deep mistrust on both sides, it is really important to improve interaction between the DHA and the DA, given that all HIPC funds for social services are being channeled via the DAs (see Chapter 10 - discussion).
9.7.2 Possible Avenues of Collaboration There are several avenues for the district health administration and the district assembly to collaborate reasonably well within the current decentralisation policy framework. In an interview with policy makers, one senior GHS official explained that the current policy, despite its weakness, has been carefully crafted to create the opportunity for health authorities and district assemblies to work together (Box 9.23).
Box 9. 23 Avenues of collaboration within the current policy framework
Extract 37. Policy makers’ view of avenues of collaboration “As a district health director, you are administratively responsible to the district chief executive and technically responsible to the regional director of health services. That has been carefully crafted to allow you the opportunity to sit down with the administrative manager and work out how best to go about the task of delivering health in the district” (Policy maker, GHS, 2004).
Extract 38. District assembly’s view of avenues of collaboration “The structures are there for effective collaboration but the DHA chooses and picks the areas it wants to collaborate with us. You see, where you don’t have the financial control, you are not consulted when it comes to decision-making; so that is where the problem is. Where the DHA sees that the assembly has a role to play financially, they want to sit and plan with us. But where they see that funding will come from outside the assembly, they want to circumvent the system and do things their own way” (District Coordinating Director, Ashanti, Region, 2004).
Many district health directors see the GHS as an autonomous entity, and are only concerned with allegiance to their regional directors even though administratively, they report to the head of the DA (Extract 37). Their position is strengthened by the direct disbursement of funds to the BMCs through the office of the regional director of health services rather than through the assembly. Except for the GOG item 2, which passes through the assembly, all health funds are disbursed to BMCs directly from the MOH headquarters. The DA, therefore, has no control over the bulk of funds flowing to the district for health delivery purposes. In the opinion of the Coordinating Directors
235 interviewed, this lack of control over health funds by the assembly was the main bone of contention because it allowed the DHA to pick and choose the areas it wished to collaborate with the assembly on (Extract 38).
9.7.3 District Assembly’s Influence on Resource Allocation The main focus of this section was to establish how the DA, through collaboration with the DHA, influenced the pattern of resource allocation for health in the district. As Extract 39-40 demonstrates, in districts where collaboration was relatively effective, there were reports of considerable financial and material support from the DA.
Box 9. 24 Effective collaboration and resource support from the district assembly
Extract 39. “The assembly has been supporting us a lot…I think you saw things for yourself when you came to the hospital this afternoon; they have invested a lot of money in putting up those staff bungalows. In some districts the support is not all that good but here they do really help. Personally, I think it all comes down to personal relationships; you met me in the District Coordinator’s office, he was also here some few minutes ago. We sit down and talk as friends; I don’t hide anything from him in terms of our funding, so they really understand our problems. There is nothing we ask, which they will deny us if they have the ability to provide. Beside the buildings and vehicles for our NID campaigns, they sometimes offer us financial support to purchase fuel and other materials. I would say our collaboration has been very effective” (District Director of Health Services, Northern Region, 2003).
Extract 40. “The assembly helps us in many different ways. For example, they support us with fuel; we have been giving them our monthly activity plan so they always know what we are doing. This little clinic here (pointing to the premises) is an assembly clinic…they gave us a revolving fund (seed money) of, I think, 2 million cedis to start with as a trail, the intention was that every quarter they’ll add 2 million so that we build up a capital. But after the first quarter I felt we’d be able to run it with the seed money so we haven’t gone for further money” (District Director of Health Services, Ashanti Region, 2003).
In the West Mamprusi district of the Northern region, the district assembly has invested significantly in construction of bungalows for health staff at the district hospital at Walewale (Personal observation 2004). The District Director of Health Services, who had a healthy relationship with the Coordinating Director of the DA, reported other financial and material support including the release of assembly vehicles for immunisation campaigns. By contrast, in districts where collaboration was not effective (Extract 36), there was reported lack of support or very limited logistic support during national immunisation days (NID). The evidence, however, suggested that the
236 effectiveness of the collaboration between the DA and the DHA has more to do with the openness and attitudes of district health managers.
The data showed that good collaboration helped a few DHAs to obtain additional resources from the assembly, which their counterparts in relatively poor collaboration with the DA did not obtain. On the whole, the government appears to be moving towards strengthening the decentralisation process by gradually increasing the financial control of the DAs. Interviews with Assembly officials revealed that much of the HIPC poverty reduction funds for 2002 were allocated through the assemblies and more of the future savings from debt-relief are expected to be allocated through the same channels. This may encourage DHAs to be more pro-active in seeking effective collaboration with the DA if they desire to get a fair share of these resources.
Chapter 9. Summary of key points
Allocation of financial resources in the health sector is more than a simple use of formulas to determine the amount of resources that should flow to different geographical areas.
This chapter has provided insights into the complexities of resource allocation, which are not apparent from quantitative data. In particular, it has shown that, in Ghana, allocation of health resources is influenced directly and indirectly by a complex set of factors including:
• The level of funding of the health system • Timing of release of funds • The way equity is perceived by policy makers and managers • Manpower availability, • Capacity to utilise funds efficiently • Government accounting and financial regulations • Political commitment • Donor activities • The level and nature of collaboration with the local government. These factors are discussed further in the next chapter.
237 CHAPTER 10
DISCUSSION
Overview This chapter discusses the key findings of the study in relation to the literature and the degree to which the study has addressed the two primary questions underpinning the quantitative and qualitative components. To what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? What factors influenced the equitable allocation of resources for health care in Ghana?
Although the chapter is structured to address these questions separately, where appropriate, findings from the quantitative and qualitative components will be discussed together. For orderly presentation, the two research questions are used as the main headings for discussion.
10.1 To what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds?
Health sector decentralisation, including decentralisation of resource allocation decision-making, has been touted as an effective policy to enhance equity in resource allocation (World Bank1994; Bossert and Beauvais 2002; see Chapter 1). While the theoretical basis for these claims is sound, empirical evidence supporting the argument is sparse. One of the two main objectives of this study was to establish whether, and the extent to which, decentralisation of resource allocation decision-making has improved equity in distribution of funds. This objective was fulfilled by analysing mechanisms and actual resource allocation from 1998 to 2002. The results showed that in the
238 Ghanaian health system, despite the decentralisation of decision-making on resource allocation, equity in terms of differential allocation of funds in favour of the most deprived districts and sub-districts, was not greatly enhanced. That is, the study found little evidence of explicit allocation of resources in favour of deprived districts and sub- districts. The equity benchmark - EAS - developed in this study to assess the degree of fairness in resource allocation, could account for only a limited proportion of variance41 in the actual share of funds received by districts. However, there was evidence, albeit limited, that at the national level, resources were deliberately allocated in favour of the most deprived regions in Ghana (see Chapter 8).
In the light of these findings, the discussion in this section draws together and debates the key findings of the study that shed light on why decentralisation of resource allocation decision-making did not result in greater equity in distribution of funds within regions in Ghana. Why, in spite of making their own independent resource allocation decisions, regional and district health authorities were unable to shift more funds to the most deprived districts and sub-districts. The discussion draws on insights from the qualitative inquiry, where necessary, and is organised around the following key topics:
Variations in deprivation levels Mechanisms for resource allocation Interpretation of equity Fragmentation of the resource allocation system
10.1.1 Variations in Deprivation Levels Deprivation was used in this study as a proxy measure for need. This section highlights the use of differential levels of deprivation across jurisdictions (regions, districts and sub-districts) as demonstrated in this study, to elucidate why decentralisation of resource allocation decision-making did not result in greater equity in distribution of funds within regions in Ghana.
41 Below 50% over the entire five-years examined, except for donor-pooled funds (DPF) allocation in the Ashanti Region for 1999, where the proportion of variance was 56%.
239 This study found that, while deprivation is widespread in Ghana, there were significant variations among regions, districts and sub-districts that need to be taken into account in allocation of public sector health resources, for the promotion of equity. Before discussing the variations in deprivation levels across jurisdictions, it will be useful to discuss briefly how the variables and approach used for measuring deprivation in this study differ from the existing measures in Ghana.
10.1.1.1 New Approach for Measuring Deprivation in Ghana Deprivation (poverty) is primarily measured at the regional level in Ghana and is based largely on data from the Ghana Living Standard Survey (GLSS). Although the GLSS collects a range of socio-economic data, its assessment of regional poverty is largely based on estimation of household income and expenditure, including income from employment, agriculture, rent, and remittances (GLSS 4, 1998). Income approach to measuring poverty or deprivation has many well-known limitations (Mack and Lansley 1985, Nolan and Whelan 1996, Layte et al. 2001). Ringen (1988), for example, has argued that low income is unreliable as an indicator of poverty, because it fails to identify households facing distinctive levels of deprivation. Thus, it is not always the case that households on low income suffer more deprivation than their counterparts on a relatively high income. The World Bank, has in recent years, emphasised the multi- dimensional character of poverty in developing countries, introducing access to education and health services and exposure to political insecurity as part of the income definition (World Bank 2000). This underscores the limitations of using income as a sole measure of deprivation or poverty.
In this study, a range of demographic, socio-economic and area variables was combined to develop the general index of deprivation (GID). They included the proportion of regional population under-5, the use of wood-fuel, stock of mud-house, lack of access to piped water, rurality, female and level of education, among others (see Chapter 6). There was a high correlation between the variables, with a number of them emerging as important in explaining deprivation within the regions and districts. The high correlation between the proportion of females and that of the population with no-education, in particular, needs special emphasis. Although this was an obvious finding, particularly in
240 a developing country such as Ghana, it is worth highlighting, given the importance of female education for health service utilisation and the limited progress made in enhancing female education in many developing countries.
There is overwhelming evidence that female education has direct benefits for improving health, particularly, the health of children (Raghupathy 1996; Caldwell 1992). In India, for example, Govindasamy and Ramesh (1997) reported that maternal education is the most powerful and significant predictor of utilisation of child health care services. However, despite its importance, female education continues to be limited and lags behind that of males in many developing countries, especially in Africa. In the West African sub-region, for example, female illiteracy is as high as 86% in Burkina Faso, 61% in Cote d’Ivoire and 58% in Togo (World Bank 2002). In the light of this grim situation of progress towards education of females, this finding needs to be emphasised for its important social policy implications. On the whole, the variables used in this study provide a broader scope of measuring deprivation more adequately than the income approach in the GLSS, which is widely used in Ghana.
10.1.1.2 Variations in Deprivation across Jurisdictions Among the ten regions in Ghana, deprivation was higher in the three regions in the north of the country: Northern, Upper East and Upper West (see Table 8.6). This finding was not unique, as several nation-wide surveys and individual studies have identified the Northern, Upper East and Upper West Regions as the most deprived regions in Ghana (GLSS 1988; 1999; CWIQ 1997, Ashiabi 2000; Konadu-Agyemang 2000; Sackey 2005). The level of deprivation found in the Central Region was not as severe as portrayed in policy documents (GPRS 2000). Central Region was better-off than most of the ten regions in Ghana (see Figure 8.1). The income-oriented deprivation measures used by policy makers may be responsible for the high ranking of Central Region as deprived.
Within the Ashanti and Northern Regions, this study found that deprivation is not significantly different among districts (Table 8.14). This was surprising, as the Ashanti Region is generally perceived to be a rich region and many would expect that districts in
241 the region would show signs of less deprivation than those in the poorer Northern Region. It suggests that deprivation is not a region-specific phenomenon and may be as prevalent and severe in rich as in poor regions. It also questions the labelling by the Ghana Health Service of all districts in the Northern Region as deprived, as opposed to only three districts: Amansie West, Ahafo-Ano South and BAK, in the Ashanti Region (GHS 2003). Clearly, some districts were better-off in the Northern Region than their counterparts in the Ashanti Region. Such ‘blanket labelling’ of all districts in a particular region as deprived because the region happens to be poor may hinder effective targeting of deprived districts.
Finally, within districts (i.e. West Gonja and Savelugu-Nanton Districts) there were variations in deprivation across sub-districts with the relatively urban sub-districts near the district capitals having better access to health and social services than remote sub- districts. The remote sub-districts had rough terrain, which presents enormous challenge to health service delivery, particularly in terms of coverage.
10.1.1.3 Common Trends in Deprivation across Jurisdictions Across all jurisdictions, there was a strong association between rurality and high deprivation. In the Ashanti and Northern Regions, the study systematically found a pattern of high deprivation among rural districts compared to urban ones. The five most deprived districts on average had about 86% of their populations living in rural areas compared the 36% on average for the five least deprived districts. This association between rurality and high deprivation in this study support the findings of several other studies (Cloke et al. 1994; Cloke and Davies 1992). In an insightful recent study in South Africa, McIntyre et al. found that magisterial districts with high levels of deprivation are largely rural (McIntyre et al. 2002). Although this finding is not unique, especially in a developing country like Ghana, it has significant implications for policy making and equitable resource allocation, particularly when it comes to assigning ‘weights’ to different geographical areas for the purpose of promoting more equitable resource allocation.
242 A further important dimension of the rurality and high deprivation issue is the need to distinguish between remoteness and rurality when assessing deprivation among districts. Policy makers often use these terms together or interchangeably and fail to clearly distinguish between the two as pointed out by Wakerman (2004). According to the findings of this study, depending on how they are defined, remoteness and rurality may have different implications for equitable resource allocation. Grundy (2001 p.2) observed that remoteness is commonly defined in terms of geographical or social isolation, and is associated with distance, sparsely distributed population and limited access due to road or climatic conditions. In contrast, rurality has no universally accepted definition (Humphreys 1998). After reviewing various definitions (of rurality) in the United Kingdom, Rousseau (1995) concluded that the term rural encompasses a wide range of communities: affluent, deprived, agricultural, industrial, stable, mobile and others (cited in Leduc 1997). Thus, it is difficult to pinpoint any single feature that adequately captures the essence of rurality. In Ghana, as noted in Chapter 6, rurality is defined in terms of population size, i.e., any settlement with a population of less than 5000 people (Ghana Statistical Service 2002a).
At the sub-district level, this study found almost no difference between remoteness and rurality. All sub-districts in remote locations from the district capital were more deprived than their counterparts located in close proximity to the district capital. This was significantly different from the regional levels, where some districts not far from the regional capital were more deprived than some remote ones (see Chapter 8).
10.1.1.3 Deprivation and Equity in Resource Allocation under Decentralisation In the context of this study, the most important question is: how do the variations in deprivation explain why decentralisation of resource allocation decision-making in Ghana did not result in greater equity in distribution of funds? To understand why decentralisation did not result in greater equity, it is important to focus on the limited attention paid to variations in deprivation across jurisdictions by regional and district health managers, rather than focusing on deprivation per se as restricting the promotion of equity.
243 The different levels of deprivation found in this study emphasise the fact that health needs differ across jurisdictions in Ghana, and that if resources were to be allocated more equitably, these differences in need ought to be well-understood and taken into account in the distribution process. The decentralisation of the resource allocation decision-making did not achieve the desired goal of greater equity in distribution of funds partly because regional and district health authorities have largely ignored the issue of differential need for health care. Clearly, if policy makers and managers were oblivious of the difference in need across jurisdictions, logically, it would be difficult for them to allocate resources to areas where needs are greatest, as expected under decentralisation. To a large extent, the limited attention to differential health needs reflects the lack of a meaningful equity-driven policy and mechanism for allocating resources among different jurisdictions. Section 10.1.2 discusses how the mechanisms for resource allocation in the health system have contributed to the lack of greater equity in funding under decentralisation.
10.1.2 Mechanisms for Resource Allocation The equity or inequity of health care resource allocation depends on the mechanism applied. If the mechanism is inherently inequitable, it is more likely the outcome will be inequitable. Nozick (1974) emphasised the importance of procedural justice in his entitlement theory. Although the theory itself is largely anti-egalitarian, Nozick argued, that outcomes are the result of processes and that the justice of distribution is entirely dependent on the path (mechanism) used to reach it (see Chapter 3). This study found that the mechanisms for resource allocation in Ghana have contributed significantly to the lack of greater equity in funding under the decentralised system of resource allocation.
As indicated in Chapter 7, resource allocation in the Ghanaian health sector is, in principle, driven by policy which seeks to shift more funds to the most deprived areas where health needs are greatest and to reduce the financial barriers to accessing health care for the most vulnerable population groups (MOH 2003). In practice, however, the mechanisms for resource allocation between and within regions do not reflect these strong equity goals. A close examination of the mechanisms reveals four important
244 issues, each of which contributes to understanding why decentralisation of decision- making around resource allocation did not result in greater equity in funding within regions. These are: The absence of needs-based indicators and equal sharing of funds, The use of distance from the regional capital, Lack of transparency surrounding the entire resource allocation process, and The pace of re-distribution of resources
10.1.2.1 Absence of Needs-based Indicators and Equal Sharing of Funds While there is general recognition that resource allocation formulae in the Ghanaian health sector should be revised to incorporate health needs, poverty and gender issues (MOH 2003), the MOH/GHS still has no well-established needs-based formulae for resource allocation. At the national level of the health system, Ensor et al. (2001) noted that a wide range of methods is used to allocate funding. These methods or formulae are not designed to systematically assess differential health needs. The only need indicator incorporated in the range of formulae used at the national level is infant mortality rate, which accounts for only 0.6% of the total allocation (Ensor et al. 2001).
A significant proportion of the health budget is simply divided equally across-the-board for all regions, districts and facilities. This ‘equal sharing’ accounts for over 40% of resources allocated for administrative expenditure (GOG 2). In addition to the equal sharing, 25% of resources for facilities are allocated on the basis of the number of beds. Population size accounts for only about 15% of the allocations, while around 10% is divided according to the distance of the regional capital from Accra - the national capital (Ensor et al. 2001). At the regional level, equal sharing of funds was a common practice and covered a far bigger proportion of the health budget than at the national level. In the Ashanti Region, for example, 80% of the total health budget was distributed equally across-board to districts, leaving only 20% to be distributed on the basis of need. The Northern Region distributed 50% of resource in a similar manner (see Chapter 7).
245 The lack of a needs-based model for resource allocation, coupled with the division of a substantial proportion of the health budget equally-across-the-board partly explains the limited equity in distribution of health funds within regions in Ghana despite the decentralisation policy. The equal sharing of funds across jurisdictions, in particular, did not facilitate redistribution of resources to the most deprived regions. At the national level, this study found that the MOH has in recent years, resorted to top-slicing the health budget and using this for targeting the most deprived regions with greater health needs (Chapter 7). Such monies go largely into funding extra programmes and services in the deprived regions and are captured in the annual total allocations. This may explain why inter-regional funding was relatively equitable with more funds shifted to the most deprived regions. The top-slicing for targeted allocation was devised to shift resources inter-regionally without waiting for the development of a needs-based model.
Recent debate in the literature about resource allocation suggests that the RAWP-type formulae, which emphasise the distribution of resources on the basis of need, do not promote equity. Mooney and Houston (2004) noted that such formulae emphasise the size of the problem (greater need) rather than the capacity to benefit from the way resources are allocated. They particularly criticised the philosophy underpinning RAWP-type approaches, arguing that these approaches, in essence, seek to allocate more resources to ‘bigger problems’, rather than identifying what good or benefit is sought and attempting to maximise that benefit (Mooney and Houston 2004 p.30).
Needs-based resource allocation may promote geographical equity in funding. However, if equity within regions is not taken more seriously, very little would be achieved in terms of improving the health of the most disadvantaged. The findings of this study show clearly that shifting more funds to regions and districts is not a sufficient guarantee for equity within those regions, as this largely depends on the decisions made by sub-national health authorities. What should be emphasised is better ways of re-allocating resources within regions and to effectively turn resources into services and programmes to improve the health of the most vulnerable groups at regional and sub-regional levels. Policy makers, in particular, need to be sensitised to understand that promoting inter-regional equity in funding, irrespective of whatever formula is applied, is not in itself sufficient to bridge the inequalities within regions.
246 This is particularly so as decentralisation of health systems has become common (Bossert et al. 2003) and regional and sub-regional health authorities are increasingly making resource allocation decisions with a reasonable degree of autonomy. The limited equity in funding within regions in Ghana, as found by this study, vividly highlights the need for attention to equity at regional and sub-regional levels.
There is also a need for resource allocation mechanisms to pay more attention to conditions in ‘needy’ areas that prevent the effective use of resources to maximise health gains, as equity in funding without the necessary infrastructure can hardly translate to improvements in health. Thus, in designing needs-based formulae, policy makers should consider the various factors likely to prevent regions with the greatest health needs to maximise health gains using the resources allocated to them. This study found that issues such as manpower availability and local capacity to utilise funds efficiently play a crucial role in determining the amount of funds that should be allocated to jurisdictions (discussed in Section 10.2). Allocating resources on the basis of need without taking into account the conditions of low capacity across jurisdictions will not achieve the desired results. In this regard, Mooney and Houston’s MESH42 infrastructure, which they argued should be included in resource allocation formula, is very useful. According to Mooney and Houston (2004), different jurisdictions may have differing capacity to manage resources well, if MESH is not included in the resource allocation formula, jurisdictions with low capacity might be inadequately resourced.
The results of this study emphasised that regions in the northern part of Ghana not only have the greatest health needs but also have limited manpower and lack the necessary health infrastructure for better use of resources (see Chapters 2 and 7). Without taking these into account, shifting resources to these regions on the basis of need alone will have little impact on access and health status and may in fact waste scare resources.
42 MESH stands for Management, Economic, Social and Human infrastructure. According to Mooney and Houston (2004), MESH involves good management, requires availability of resources, needs a socially well functioning community and, ideally, good human resources. The authors argued that where some or all of these elements are missing, the resources might be wasted or at best used to lesser effects.
247 10.1.2.2 Distance from Regional Capital as a Measure of Need After distributing over half of the total budget equally across-the-board, policy makers and managers of health care in Ghana use distance from the national, regional or district capital (remoteness) as a measure of rurality and need. One key finding of this study is that, while most of the districts in the Ashanti and Northern Regions that were remotely located (from the regional capitals) were rural and highly deprived, not all rural districts were remotely located.
A typical example is the Bosomtwe-Atwima Kwanwoma district (BAK) in the Ashanti Region (Chapter 8). The BAK borders the Kumasi Metro district and is only a few kilometres from the regional capital. Yet, about 95% of its population is classified as rural and showed high level of deprivation than one would expect of a district in such close proximity to the regional capital. A similar situation was found in the Northern Region where the Yendi district, located some 100 kilometres away from the regional capital, emerged significantly less deprived than, the Savelugu-Nanton district, which is just about 25 km away from the regional capital43 (see Table 8.14).
In the Ghanaian context, this finding suggests that the use of distance from the capital (national, regional or district) as a measure of rurality and indicator of relative deprivation may be inappropriate under certain conditions, and could lead to wrongful categorisation in some jurisdictions.
The implications of this for equitable resource allocation in Ghana are that some peri- urban districts that are demographically rural, but not remotely located, such as the BAK in Ashanti, are most likely to be especially disadvantaged unless other factors such as population size work in their favour. Additionally, remoteness defined in terms of distance from the capital may not adequately reflect the differences in locational problems of accessing health care by different groups. Undoubtedly, distance impacts significantly on access (Buor 2003; Lavy and Germain 1994) but residing far away from the capital does not necessarily mean one is far away from health services, and hence, has limited access to health care.
43 Note that the distance of the districts from the regional capital is in terms of the distance between the district capital and the regional capital. There may be some settlements in the districts, which are closer to the regional capital.
248 Appropriately, remoteness may be defined in terms of distance from health services rather than distance from the capital. In their recent study in South West England, Jordan et al. (2004), found that remoteness from health services is an issue that affects both urban and rural areas. From this perspective, it will be inappropriate to equate remoteness to rurality in certain situations. In Ghana, it is more useful to define remoteness in terms of geographical and financial isolation from health facilities/services rather than distance from regional capitals. This will help capture the effects of remoteness on access to health services in both urban and rural settlements.
10.1.2.3 Lack of Transparency Surrounding Inter-regional Funding Resource allocation is a process that is highly susceptible to political interference and bureaucratic abuse. However, as observed by Peacock and Segal (2000), an explicit and transparent resource allocation mechanism provides a constraint on the use of arbitrary political or bureaucratic power, and identifies key value judgements that underpin the resource allocation process. Additionally, if made transparent and publicly available, resource allocation mechanisms can generate democratic debate over their appropriateness, which can eventually add to their viability.
This study found that the process of allocating financial resources in the Ghanaian health system is anything but transparent. The current 5-Year Programme of Work (POW II) establishes broad resource targets by levels of the health system (national, regional, and districts). However, the quantum of resources to be allocated to specific regions or districts is determined by the various authorities using a range of resource allocation mechanisms.
At the national level, it was extremely difficult to establish the actual mechanisms used to allocate funds. Policy makers could not provide any concrete information on the formulae they have used in the past five years (1998-2002) for allocating funding to the ten regions. Neither could they provide expenditure data on government and donor funds allocated to the regions prior to 2002 (Chapter 6 and 8). Even when data on government funding were available (2002 and 2003), no reasonable insights into how the different shares for individual regions were computed could be provided. Regional
249 managers equally had little knowledge of how funds are distributed at the national level. The question is: in what way does the lack of transparency surrounding resource allocation in the Ghanaian health system contribute to inequitable distribution of funds? Three implications for equity of the lack of transparency surrounding the resource allocation mechanisms can be discerned. First, the capacity of funding mechanisms to contribute to equity by redistributing resources becomes difficult to establish where there is no transparency. As in other health systems, resource allocation mechanisms in Ghana, seek to overcome inequities by re-distributing funds towards the most deprived areas. The extent to which the mechanisms in place contribute to achieving this goal is difficult to ascertain because of the lack of transparency. Second, policy makers cannot be held accountable for the reasonableness of the funding decisions they make, as the appropriateness of the funding mechanisms is largely unknown. Finally, the lack of transparency provides a good opportunity for political and other interest groups to influence the allocation process often to the detriment of poorer regions with less political influence. These problems were observed in Ghana, as discussed later in Section 10.2.
10.1.2.4 The Pace of Re-distribution of Resources The pace of re-distribution of resources is crucial for equity. There is a real danger in low and middle-income countries experiencing urbanisation that resources may be shifted away from urban areas at the time when the needs of these areas are expanding. The inter-regional analysis of this study showed a substantial shift of funds towards the three most deprived northern regions from the Greater Accra region. Similarly, in the Ashanti Region, a massive re-distribution of resources away from the Kumasi metro district was found.
While re-distribution of resources may be effective in reducing geographical inequities, it could adversely affect equity and create inefficiencies if enough time is not given to regions losing and gaining resources to adjust gradually to the changes. In South Africa, for instance, the approach to re-distribute resources following the assumption of office of the Africa National Congress (ANC) government in 1994 had an initial time-frame of five years. This was too short for Provinces to adjust to the substantial shifts in funding
250 that resulted from the redistribution process in 1995/96. Eventually, the National Department of Health was forced to revise the new resource distribution approach and adopt a more gradual re-distribution system in 1996/97 (McIntyre and Gilson 2002). This underscores the need for caution in re-distribution of resources. Resources should not be shifted too rapidly not only to allow areas to develop regular capacity to absorb budget increases and cuts but also to allow for regular updating of indicators of need with the most recent data which may indicate that the relative distribution of health care needs have changed.
Mechanisms for resource allocation are largely driven by the conceptual interpretation of equity in the health sector. Section 10.1.3 discusses the findings that explain the link between interpretation of equity and the lack of greater equity in resource allocation.
10.1.3 Interpretation of Equity Equity is a value-laden concept interpreted differently by different people (see Chapter 4). In general, the way equity is interpreted has significant implications for how it is operationalised. Health systems, as observed by McIntyre et al. (2000) rarely establish clearly or specify fully the equity goals they seek to achieve. In Ghana, there seems to be no clear understanding of what equity principle is being pursued in the health sector. The Ministry of Health has embraced equality of access; a horizontal equity principle as its main equity goal but it also has a policy of addressing geographic inequities by shifting resources to the most deprived regions on the premise of achieving vertical equity (see Chapter 2).
This emphasis on different equity principles resurrects the old debate about which equity principle is most favoured by policy makers. Many commentators have argued that ‘equality of access’ is the most popular equity objectives among policy-makers (Mooney 1991; Goddard and Smith 2001). However, there is a counter-argument that policy documents of many countries contain references not only to ‘equality of access’ but also to ‘distribution according to need’ and ‘equality of health’ (Culyer et al.1992a; 1992b). The focus on different equity goals may be due partly to the difficulty in defining concepts such as access, need and health, which are used to specify how equity
251 is being pursued. Largely, policy makers have no yardstick against which to judge the consistency of their policies in terms of these concepts.
In the context of Ghana, the key issue with regard to the different equity objectives is the degree to which these objectives have been communicated within the health system and tools developed to support their operationalisation at the various levels. This study found that while the ‘equality of access’ objective is well-known among managers and other stakeholders in the health sector (Box 9.1), the policy objective to shift resources has not received any serious attention. The common approach to allocating resources in the health sector, as noted in Section 10.1.2, was either by dividing resources equally across-board or on the basis of population size. At all levels of the health system, there was evidence of substantial amounts of resources being shared equally across jurisdictions or on the basis of population without regard to variations in need (relative deprivation).
First, the equal sharing of resources among jurisdictions regardless of differential need raises concerns about the seriousness to promote equity in the Ghanaian health system. The approach does not reflect any of the two stated equity objectives of the MOH. Indeed, it is more akin to the utilitarian principle of “counting everybody for one and nobody for more than one” (Sidgwick and Singer 2000; see Chapter 3). In a country like Ghana, with marked inequalities in health, such an approach to resource allocation is likely to perpetrate the existing geographical inequities in health. It implies that no jurisdiction deserves special attention in resource allocation, regardless of its level of deprivation. While the equal sharing cannot be denounced completely as inequitable, it is simplistic and constitutes a limited way of accommodating equity within the demands of social justice, which from Rawls’ perspective, is served when resource allocation maximises the opportunities of the least privileged members of society (see Rawls 1972; and Chapter 3).
Second, the ‘equal sharing’ of resources across jurisdictions in Ghana raises concerns about whether decentralisation of resources allocation decision-making is an appropriate policy for promoting equity in funding. Proponents of decentralisation argue the policy promotes greater equity by distributing resources towards traditionally marginalised regions and population groups (Bossert and Beauvais 2002; see Chapter 5). Critics
252 maintained that centralised systems are more likely to redistribute resources in favour of poorer regions and population groups (Prud’homme 1994, Sikosana et al. 1997, Kleinman et al. 2002). The equal sharing of resources, to a large extent, suggests that the Regional Health Administrations (RHAs) in the Ashanti and Northern Regions did not use their autonomous decision-making powers to engage in any comprehensive assessment of need for purposes of resource allocation. This partly explains why decentralisation of the resource allocation system did not result in greater equity within the regions in Ghana. But is it the decentralisation policy per se that is bad or its implementation that was flawed?
The findings of this study suggest a complex combination of issues related to the decentralisation policy per se as well as its implementation, as responsible for the lack of greater equity in distribution of resources in the Ghanaian health system. First, the decentralisation policy itself has created fragmentation of resource allocation systems in the health system, which constrained equity. Second, there was evidence of implementation flaws with the MOH occasionally interfering in the resource allocation process. These two issues are discussed in Section 10.1.4.
10.1.4 Fragmentation of the Resource Allocation System Decentralisation allows regional and district health authorities in Ghana to make independent resource allocation decisions. RHAs decide the proportion of the block grant from the national level that should go to individual districts. In a similar vein, DHAs decide the amount of funds that should go to sub-districts. This study found that the decentralisation policy has created a situation where every region and district independently handles resource allocation issues, with no jurisdiction aware of what the other was doing. At the national level, the MOH did not know how resources were allocated at regional and district levels and vice versa44. The same situation existed between the DHAs and the RHAs (see Chapter 9). The question is: in what ways does the fragmented means of dealing with resource allocation contribute to explaining why the decentralisation policy did not result in greater equity in funding within regions?
44 The central MOH did not even have records of funds allocated to the regions prior to 2002 (see study limitations).
253 The answer lies in the lack of opportunity to share innovative approaches for resource allocation and deal collaboratively with solving difficult issues associated with developing effective equity-focused mechanisms. Peacock and Segal (2000) observed that the design of resource allocation mechanism entails a range of technical, bureaucratic and political issues that may be better dealt with in cooperation with others rather than every jurisdiction dealing with it separately. Although the autonomy of regional and district health authorities to develop their own funding mechanisms, may promote innovations in the development of new funding methods, it may also create funding fragmentation that could lead to inequalities. Thus, one region may develop a more equity-focused mechanism while another may not be so much concerned about equity, as in the case of the resource allocation mechanisms in the Ashanti and Northern Regions. The study found that although both Ashanti and Northern Regions did not show strong commitment to equitable funding, the Northern Region had a more equity- friendly funding mechanism than the Ashanti Region.
It also created a situation where problems that could best be dealt with co-operatively by learning from one another are being tackled in isolation. There was an instance during this study where a district health accountant became interested in the resource allocation mechanism of another district after being shown a copy by this researcher. This suggests that jurisdictions may learn from each other if they have the chance to collaborate on resource allocation matters. Policy makers need to find a way of promoting collaboration between jurisdictions without limiting the autonomy enjoyed by regions and districts in designing their own area-specific resource allocation mechanisms. To a large extent, the lack of effective collaboration across jurisdictions has restricted the sharing of innovative approaches across regions and districts.
10.2 What factors influenced the equitable allocation of resources for health care in Ghana?
This section discusses the factors that influenced the equitable resource allocation in the Ghanaian health system. Although the discussion is meant to address the second research question re-stated above, its sheds further light on why decentralisation of
254 decision-making around resource allocation did not result in greater equity in funding within regions in Ghana. Thus, it provides more insight into the reasons behind the limited intra-regional equity in funding despite the decentralisation of the resource allocation system. The major factors discussed include low levels of funding of the health system, local capacity to utilise funds efficiently, donor influence, politics and the nature of collaboration with the local government.
10.2.1 Low levels of funding of the health system Health systems are rarely adequately funded, particularly in developing countries (WHO 2000). Results of this study indicate that the low levels of funding of the Ghanaian health system significantly affected the equitable distribution of resources. Key stakeholders at all levels of the health system observed in interviews that the poor funding of the health system poses one of the difficult challenges to equity (see Box 9.1). Operational managers explained that, due to inadequate funding, health programmes were often implemented late or postponed for a long period of time. With the total MOH spending per capita (including donor funding) averaging about US$ 7.40 in 2001 (see Table 7.2), there is no doubt the Ghanaian health system is severely under- funded, even by the most modest developing countries standards45. But should the inadequate funding necessarily affect equity and equitable allocation of resources?
In all resource-poor settings, including Ghana, it will be unrealistic to expect more than the delivery of the most basic health services (Hanson et al. 2003). However, resource inadequacy per se should not been an impediment to promoting equity. Equity, as noted in Chapter 4, is about social justice or fairness. Health equity is about creating fair opportunity for everyone to attain his/her full health potential and ensuring that no one is disadvantaged from achieving this potential, if it can be avoided (Whitehead 1990; Braveman and Gruskin 2003). From this perspective, it is not the quantum of resources but its distribution that is of key concern.
45 Health spending per capita in Sub-Saharan Africa from 1997 to 2000 averages about US$29. This comprised both public and private spending and excluded donor funding (World Bank 2004).
255 There is ample evidence in the literature suggesting that some countries have succeeded in improving health status well beyond the level that would be expected given the resources available (Halstead et al. 1985; Mehrotra 2000). Similarly, there are well- resourced countries with relatively moderate health indicators and significant inequalities than one would expect. Perhaps the best examples are Cuba and the United States (US). For many years, Cuba has managed to provide free, universal and comprehensive health coverage to its citizen, even in the worst economic times (Kirkpatrick 1997). According to the United Nations Childrens Fund [UNICEF] (1996), Cuba ranks 27th in the world on infant and under-5 mortality rates, two indicators often used internationally to assess health status. The US ranks only one place above Cuba, at 26th for both indicators, despite its abundant wealth and high per capita spending on health. There are also significant inequalities in health between socio-economic groups in the US (Wilkinson 1997).
This suggests that there can be inequities and poor health status even in health systems that are adequately funded and vice versa. Without dismissing the fact that low funding is a threat to equity, it is worth noting that sometimes it is not how much a country spends as much as how it spends its resources that determines the health status of its population (Yach and Harrison 1994). Governments usually claim that they provide services for the benefit of the poor but public sector health spending in many countries tend to benefit the well-off (Castro-Leah 2000; Gwatkin et al. 2004). Australia, for example, spends about 10% of GDP on health and enjoys a high life expectancy rate of around 79 years. Yet, the health of indigenous Australians is comparable to that of any poor third world countries, with life expectancy nearly 17 years shorter than non- indigenous Australians (O’Donoghue 1999). The inequitable health spending by States and Commonwealth Governments of Australia perceived as is partly responsible. The Commonwealth Government, for example, spends 74 cents on indigenous Australians for every dollar spent on the rest of the population (Ring and Brown 2002), despite their disadvantaged position and poorer health status of the former. In Ghana, although the health system is under-funded, what appears to be the main obstacle to equitable access and resource allocation is the lack of a meaningful national equity-focused policy for allocating resources between and within regions. The mechanisms for resource allocation at various levels of the health system, as discussed
256 above, relates poorly to population needs. At the same time, there is little effort to systematically improve the health of the most disadvantaged segments of the population. These and other systemic problems such as limited capacity to manage funds efficiently, discussed next, considerably affect equity in health care funding as much as inadequate total funding.
10.2.2 Local capacity to utilise funds efficiently The capacity of local health services to manage or utilise funds efficiently is one of the important factors underpinning resource allocation, particularly at the national level. While policy makers see the need to shift more resources to deprived areas to address inequities, they revealed in the interviews that they could not allocate resources to areas where there was little or no capacity to utilise funds efficiently (see Box 9.7).
One of the key determinants of local capacity to utilise funds efficiently is health workforce availability. This was the biggest concern of policy makers. As in many developing countries, Ghana suffers from migration of health professionals to western countries (see Chapter 2). This has left the health system significantly understaffed with wide variations in capacity. In general, health workforce numbers are greater in the southern regions, where most of the few doctors left in the country prefer to work. The deprived regions in the north, which are in serious need of resources to improve health status, lack the requisite manpower46 and have low capacity to manage resources efficiently. Within regions and districts, there are rural-urban variations with most urban areas having better access to qualified personnel than their rural counterparts.
Green et al. (2000) observed that it is crucial that any allocative process be handled by staff with appropriate professional expertise who recognise the critical importance of budgets for achieving policy objectives. Policy makers interviewed in this study maintained that variations in capacity levels in the health system have left them with little option than to allocate resources in favour of jurisdictions where funds would be utilised efficiently, most of them urban. This is likely to perpetuate existing inequities.
46 Doctor-patient-ratio in the Northern Region averages 1:66000,as indicated in Table 7.5.
257 The obvious question is - should inequities in funding be maintained in the name of efficient management of funds?
Concerns of policy makers about allocating resources to areas with limited capacity to utilise funds efficiently are legitimate given the level of resource scarcity facing Ghana. However, it is inequitable to distribute resources in favour of areas with adequate capacity for efficient fund management and at the expense of deprived areas with limited capacity. Rather than skewing resources towards the areas with sufficient capacity, the “inequitable” capacity constraints facing deprived areas must be addressed. This is an important first step towards improving equity in funding allocation under decentralisation in Ghana. This will create some balance in the distribution of the salary budget (GOG 1) in particular. Though salaries were not covered in this study, it is important to note that in 2004, about 84% of the total public sector health spending went into payment of salaries and allowances (see Chapter 7). This means areas facing persistent staff shortages may not be getting their fair share of a significant proportion of the government health budget.
Addressing the inequitable capacity constraints will ensure that all jurisdictions are equitably resourced and have access to a range and quality of health care specified as necessary by policy makers. Oliver and Mossialos (2004) observed that there is the need to ensure that incentives are available for sufficient facilities and staff to locate and remain within disadvantaged areas. Such endeavour in Ghana is likely to transcend the health sector to sectors such as housing, electrification and provision of good drinking water for health staff. The lack of these facilities in many rural areas partly discourages health professional from locating and working in such areas. However, in the allocation of the MOH budget itself, it will be necessary to account for variations in capacity levels across jurisdictions. The inclusion of Mooney’s MESH infrastructure, mentioned earlier in Section 10.1.3, will be worthy of consideration here (Mooney and Houston 2004). By incorporating MESH in the resource allocation formula, the capacity of poorer jurisdictions to benefit from resource allocation will be enhanced.
258 10.2.3 Donor influence Donor funding plays a crucial role in health systems of developing countries including Ghana. There is chronic imbalance in the way public funds are allocated in many health systems (Cassels 1997). The form that donor funding takes has significant implications for easing or aggravating this imbalance. While most stakeholders in the Ghanaian health sector, particularly managers at operational levels, see donor funds as the ‘life- blood’ of the health system (see Box 9.10), the results of this study indicate that donor funding might be contributing to inequities in inter-regional and district resource distribution.
The main avenue by which donor funding might compromise equity in the health system is through the use of project funding. Project funding or project aid is a type of donor funding that is earmarked for specific purposes and often involves setting up semi-autonomous cost centres, sometimes managed by donors (Walt et al. 1999). In the last two decades, Ghana has made significant progress in coordinating donor funding in the health sector. Since the Medium-Term Health Strategy (MTHS) was embarked on in 1994, the country has been developing a comprehensive health sector 5-Year Programme of Works (POWs) and budgets that reflect articulated priorities. The donor community has been involved in the process of developing and budgeting for the POWs at the outset and has been generally supportive by agreeing to pool funds to finance the agreed POWs.
Despite good cooperation between donors and the MOH, however, some of the major donors either retained part of their funding to support specific donor projects or completely operated outside the pooling arrangements. As mentioned in Chapter 7, the United States Agency for International Development (USAID) did not join the pooling arrangement while DANIDA retained 25% of its total funding for project aid. Only the World Bank contributed 100% of its funds into the “common basket” to finance agreed POWs. There were also many non-governmental organisations (NGOs) that served as conduits of project aid implementing a range of projects in the health sector.
Cassels (1997) noted that project aid can introduce or maintain distortions in the way government funds are allocated. The effects of project aid on the pattern of resource allocation in the Ghanaian health system can be understood from the way donors pick
259 and choose regions and districts where they wish to locate their projects. The project areas, largely rural communities, are selected by donors on the basis of donor priorities rather than on the Ministry of Health’s strategic planning. Donors enter into a range of specific agreements with recipient communities, often with the MOH’s approval in principle, but without its direct involvement.
It was observed in this study that, because there is a general understanding in Ghana that the three northern regions (Northern, Upper East and Upper West) are the most deprived, many donor-led projects have been concentrated in these regions47. Although project funds were not captured in this study, indications are that they have skewed resources to the three northern regions. The key point here is that due to the limited information available on the actual amount of resources that donors are putting in the numerous projects, no account is taken of such funding when allocating resources at various levels of the health system. In terms of equity, this is a worrying situation as those regions, districts and sub-districts receiving more projects will have more funding than others receiving less project aid.
10.2.4 Politics The allocation of public resources for health has not been devoid of politics, particularly in developing countries, where the process usually lacks transparency and is free from public scrutiny (Birdsall and James 1993). While this study did not reveal any direct political influence on the amount of resources allocated to various regions and districts in Ghana, they show how successive governments, out of political expediency, adversely influence the resource allocation process, and thereby, contribute to maintaining the existing inequitable distributions. They also highlight the extent to which communities and interest groups influence resource allocation by exerting pressure on politicians to provide their areas with health facilities.
Despite the strong policy commitment to addressing inequities in health in Ghana, the government continues to invest in “big” hospitals while agreed equity-focused
47 This has done very little to improve the health status of the people in these regions, which as observed by one regional health manager, continue to deteriorate (see Box 9.12).
260 programmes suffer from poor funding (Box 9.9). A typical example is the construction of a new modern government hospital in Sunyani, the Brong Ahafo regional capital, by the current New Patriotic Party (NPP) administration. Many observers see it as an attempt by the government to boost its support base in the regions where it has marginal seats. Similar projects were undertaken by the previous government in other regions. Ironically, these projects, which have substantial recurrent cost implications, are being undertaken at a time that Ghana is struggling with a chronic shortage of health professionals to run existing facilities.
The Ghana Health Service (GHS) has an equity-oriented policy of training community health workers and placing them within their own communities to deal with minor health problems under its community-based health planning and services (CHPS) initiative (MOH 2002). Although CHPS has been hailed as significant for improving equity in access to health, implementation has been slower than expected due largely to inadequate funding. Yet, substantial amounts of money are being spent on new hospitals, which are likely to be understaffed and under-utilised (Box 9.15).
The inequitable allocation of public resources in favour of hospitals and specific geographic regions is not a new phenomenon in developing countries, particularly in Africa (Cassels 1996; Castro-Leah 2000). The practice can be understood in the context of public choice theory, which offers a coherent explanation for seemingly non-rational decision-making by governments. According to the theory, politicians and bureaucrats do not seek to optimise economic efficiency but rather to maximise their own choices of getting re-elected and staying in power. They consciously seek to provide benefits to a range of constituents they believe will help them to retain office. Similarly, individuals join with other self-seeking individuals to influence public policy to their advantage, using ‘behind-the-scene’ lobbying (Grindle and Thomas 1991; Birdsall and Hecht 1994).
Government officials and politicians in Ghana stand to gain from the construction of large hospitals and health centres. Such projects are visible, popular and serve as evidence of government commitment to health care of the people in areas where they are located. Their influence on equity of resource allocation should not only be seen in terms of their adverse effects on the level of funding available for other regions where
261 no such projects have been undertaken, but also their long-term recurrent costs. They usually skew the salary and non-salary recurrent budget allocations in favour of recipient regions. But most importantly, they shift resources towards the rich, who predominantly use hospital-based services, and away from the poor and the most disadvantaged largely living in rural areas.
10.2.5 Nature of collaboration with the local government (District Assembly) Ghana’s desire to implement a devolved system, where local government takes full responsibility for health care, is yet to become a reality. Perceived weakness in the managerial capacity of District Assemblies, the fear of local politicians not giving adequate priority to health issues, and behind-the-scene politics of competing interests within the Local Government and Health Ministries have made implementation of a true devolved system difficult. Government functions are devolved, but the health sector remains under the responsibility of the MOH and GHS. Management and resource allocation functions have been deconcentrated within the health sector (see Chapter 2).
The influence of the District Assembly (DA) on resource allocation can be understood in terms of the extra funding and logistics it provides in support of district health activities. In districts where collaboration between the District Health Administration (DHA) and DA was effective, the former received more financial and logistic support from the latter than in districts where relationships were frosty. Although the amount of funding from the DA to DHAs was not captured in this analysis, there was ample evidence from the qualitative study suggesting that effective collaboration with the DA rewards DHAs with additional funding and logistic support.
Aside the financial and logistic support, it is important to point out that effective collaboration with the DA is crucial for collaboration with key sectors such as education, agriculture and housing. The DA provides the platform for the health sector to build partnerships with other sectors at the district level.
While the existing local government and health structures set out mechanisms for effective collaboration between the DHA and the DA, this study found that collaboration between the two bodies is generally ineffective with few exceptions (Box
262 9.14). This situation has arisen largely because of the inconsistencies in the decentralisation policy. The parallel decentralisation stream (devolution of government and deconcentration within the GHS) has created confusion as to who is in charge of district health services. The DA sees itself as responsible for all services including health. It therefore feels bitter for being left out in health planning and delivery activities. The DHA, on the other hand, sees itself as the only technical expert capable of delivering health services and seldom consults the DA on health planning and delivery. Its allegiance is directed more towards regional and national health authorities. This confusion and subtle struggle for control has made effective collaboration difficult. It is worth noting, however, that such tensions are not peculiar to Ghana; it is a common phenomenon in many developing countries. Jeppsson (2001) reported similar conflicts in Uganda.
In addition, there is no statutory framework for effective inter-sectoral collaboration in the health sector. All district health managers interviewed in this study knew that they were supposed to participate in the activities of the District Assembly. However, because such participation is not a statutory obligation and there are no sanctions for non-compliance, only a few managers with good personal relationships with Assembly officials bother to participate in DA activities.
Finally, the apprehension about subsuming health under local government in Ghana appears justified. There is no conclusive evidence available from other developing countries practising complete devolution that it actually helps local health care systems. In Uganda, devolution led to a decrease in allocations for primary health care in most districts (Jeppson 2001; Okuonzi 2004). In China, Tang and Bloom (2002) found that devolution of health services to townships neither led to an increase in government health financing nor to an improvement in equity. However, in Benin, devolved decision-making contributed to an increase in equity in health care delivery and reinforced partnership with the local communities (Mogedal et al. 1995). With such mixed picture of benefits from complete devolution, the decision to subsume the health sector under the local government should be made with caution.
In the meantime, the Ghanaian government is gradually increasing funding to District Assemblies (MOH 2004). Health services have a claim to some of these funds. It
263 remains to be seen whether this will be sufficient incentive for effective collaboration between the DHA and the DA. It appears, however, that true devolution, which will see the health sector wholly placed under the local government, as initially envisaged, will not happen anytime soon, as long as local government capacity remains an issue and government continues to worry about whether spending at districts levels will be consistent with national priorities, including primary health care.
10.3 Study Limitations This study has a number of limitations. First, there were some gaps in the financial data used to assess equity between and within regions. The extent of equity in resource allocation was assessed by analysing financial data over a period of five years from 1998 to 2002. While in the Ashanti Region data were available for the entire five years, there were many gaps in the data obtained from the Northern Region. For example, data for donor-pooled funds were available for four years from 1999 to 2002 instead of five years while data for government funds (GOG 2-3) were only available for three years (2000 to 2002).
Surprisingly, at the national level, the MOH/GHS had no record of funding allocations to the ten regions in Ghana from 1998 to 2001; the only data available were for 2002 and 2003. Consequently, the analysis of inter-regional equity was limited to those two years (2002 and 2003) for which data were available. The effect of these gaps in the financial data on the overall study results was minimal. This is because the main emphasis of the study was on intra-regional equity in resource allocation (i.e. allocation from regions to districts) and the data available in the two regions were sufficient to achieve that objective.
Another problem in relation to data was the exclusion of salaries (GOG 1) and internally generated funds (IGF) from the analysis. Salaries expenditure comprises about 80% of the total government health budget (MOH 2000). Any analysis of equity in resource allocation that excludes salaries automatically overlooks a significant proportion of the health budget. In this study, however, the objective was not merely to establish the extent of equity in resource allocation but to also test the hypothesis that decentralisation of resource allocation decisions improves equity. In this regard, salaries
264 had to be excluded since decisions regarding hiring of staff and payment of salaries are not decentralised. The Ministry of Finance (not Policy makers and managers in the health sector) controls payment of all salaries.
Districts generate different amounts of IGF revenue, which needs to be taken into account in any analysis of equitable funding. The analysis in this study excluded IGF because data for such revenue was so scanty (particularly in the Northern Region) to make any meaningful analysis possible. Finally, there were limitations with the socio- economic and financial data used for measuring deprivation and assessing equity in resource allocation at the sub-district level. The census dataset used in this study was largely district level aggregates, which could not be disaggregated to sub-district levels. The lack of socio-economic data at the sub-district level significantly limited the range of variables that could be used to measure deprivation across sub-districts. Coupled with the lack of reliable financial data, these problems constrained the quantitative analysis at the sub-district level. Nonetheless, their impact on the overall study results was judged to be minimal.
In addition to the above methodology-related limitations, this study also raises some conceptual issues. Vertical equity (defined as differential allocation of resources in favour of the most deprived regions, districts and sub-districts) is the main equity principle underpinning the analysis, as observed in Chapter 4. However, in Ghana, the Ministry of Health allocates resources to achieve different equity objectives. Resources are allocated to achieve equality of access for all Ghanaians, a horizontal equity objective. They are also allocated to address existing inequities by deliberately targeting the most disadvantaged areas and population groups with more resources, a vertical equity goal. In effect, the MOH promotes both horizontal and vertical equity objectives in resource allocation.
This study, despite its emphasis on vertical equity, could not ignore the goal of horizontal equity (equality of access for equal need). Consequently, in seeking to identify the most disadvantaged regions, districts and sub-districts, differential health needs measured in terms of relative deprivation were used. This does not include the argument that special attention be paid to informed community values and preferences in an attempt to operationalise vertical equity goals (Wiseman and Jan 2000). While
265 attention to informed community values and preferences is important, it may be more essential in societies where there are considerable differences in cultural interpretation of health, as in Australia and South Africa. Although some variations may exist across regions and districts in Ghana in terms of how health is conceptually understood, such differences do not appear to be drastic enough to warrant special attention, given that culturally the country is not vastly different.
Finally, the inability to analyse the extent to which the most disadvantaged population groups in Ghana differentially benefit from resource allocation may be regarded as a further limitation. Although such ‘benefit incidence analysis’ would have provided valuable insights into equity within districts, the relevant data for such analysis were not available. It must also be noted that the primary interest of this study was in geographical equity; that is, equity in resource allocation to various jurisdictions, rather than equity in allocation among different population groups.
266 Chapter 10. Summary of key points
This chapter has discussed the key findings of the study that address the two main research questions:
• To what extent has the decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds?
• What factors influenced the equitable allocation of resources for health care in Ghana?
The discussion in relation to the first question sought to explain why decentralisation of resource allocation decision-making did not result in greater equity in distribution of funds within regions. The key discussion points were:
• Variations in deprivation levels – the limited attention paid to this issue by health authorities explains why decentralisation did not result in greater equity in distribution of funds within regions.
• Mechanisms for resource allocation – the absence of needs-based indicators in the resource allocation formulae, the use of distance from the regional capital, and the lack of transparency surrounding the resource allocation process provide further insight into why decentralisation did not result in greater equity in distribution of funds.
• Interpretation of equity – the lack of clear understanding of what equity principle is being pursued in the Ghanaian health sector also contributes to explaining the limited equity in funding under decentralisation.
• Complexities in the decentralisation policy – decentralisation created a fragmented system of resource allocation in the health sector, with no jurisdiction aware of what the other was doing. The lack of opportunity to deal collaboratively with complex issues associated with resource allocation explains why decentralisation did not result in greater equity in distribution of funds within regions.
The discussion in relation to the second question focuses on the major factors that influenced the equitable allocation of resources in the health system. The key discussion points were:
• The low levels of funding of the health system • Local capacity to utilise funds efficiently • Donor influence • Politics • The nature of collaboration with the local government These factors shed further light on why decentralisation of resource allocation decision- making did not result in greater equity in funding within regions.
This chapter has also discussed the main limitations of the study, which range from gaps in the financial data to conceptual issues.
267 CHAPTER 11
CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH
“Justice involves treating like cases alike and different cases differently” Aristotle
Overview Inequities in health persist in Ghana. The poor and the most vulnerable, who need health care the most, face severe access restrictions while the well-off receive more and better quality services. The Ghanaian Ministry of Health and its development partners (particularly donor organisations) have recognised the need to address inequities in health. However, addressing the problem within the limits of available resources has been a difficult challenge, not only because resources are scarce, but also because those available are not efficiently utilised. This concluding chapter has four components. First, it provides a brief review of the study in the context of the main research questions and study hypotheses. Second, it discusses the key policy implications of the study. Third, it reflects on the contributions made to the equity and resource allocation debate, and fourth, it offers suggestions for future research.
11.1 A Brief Review of the Study This study was motivated by the noticeable lack of empirical evidence on the extent to which resource allocation has been equitable, particularly, at sub-national levels in decentralised health systems. The initial literature survey revealed that, first, regardless of the many health sector reforms that seek to promote equity, the health status of the most disadvantaged population groups in many countries remains poor (Donaldson and Gerard 1993; Gracey et al. 2000; Gwatkin. 2000; Gwatkin et al. 2001). Second, the inequitable allocation of resources (including financial resources) has contributed
268 significantly to the overall inequities in health across jurisdictions and population groups (Deeble et al. 1998; Castro-Leah et al. 2000; McIntyre et al. 2000). Finally, while decentralisation of resource allocation decision-making in the health sector has been touted as an effective policy (World bank 1993; 1994) to enhance equity in resource allocation, little is known about whether, and to what extent the policy has improved equity, particularly at regional and sub-regional levels.
On the basis of the identified gaps in the literature, this study aimed to provide empirical data to inform and support health policy reforms seeking to address inequities in health and resource allocation. Two interrelated questions were posed: first, to what extent has decentralisation of decision-making around resource allocation in the Ghanaian health system improved equity in the distribution of funds? Second, what factors influenced the equitable allocation of resources for health in Ghana?
A set of working hypothesis was constructed to guide the analysis. It was hypothesised that: 1. Inter-regional resource allocation in the Ghanaian health system has been largely equitable in terms of differentially benefiting the most deprived regions 2. The decentralisation of resource allocation decision-making to regional levels has improved equity by distributing funds in favour of worst-off districts. 3. Resource allocation at the district level of the health system has been equitable in terms of differentially benefiting the most deprived sub-districts 4. Equity objectives drive resource allocation in the Ghanaian health sector. (see Chapter 6).
The study found that inter-regional resource allocation in the Ghanaian health system has been largely equitable in terms of differentially benefiting the most deprived regions, confirming the first hypothesis. However, this should be taken with caution given the limited data used for this level of analysis, as data were available for only 2002 and 2003.
269 There was no strong evidence supporting the second hypothesis that the decentralisation of resource allocation decision-making in the Ghanaian health system improved equity48 in distribution of funds within the Ashanti and Northern Regions. The proportion of variance in districts’ share of funds accounted for by the equity benchmark – the EAS - were generally low (below 50%) in all cases, except for GOG allocation in Ashanti Region for 2000 where the proportion of variance was 56%.
Similarly, the findings of this study did not support the third hypothesis. There was no indication that the most deprived sub-districts differentially benefited from resource allocation. Instead, this study found that district level resource allocation largely favoured urban sub-districts, which were rather less deprived than their rural counterparts. There is the need for caution here as well, because only two districts from the Northern Region were used for this level of analysis due to data problems.
Finally, there was little evidence indicating that equity objectives are key drivers of resource allocation in the Ghanaian health system. On the contrary, efficiency concerns, politics, donor influence, and the nature of collaboration with the local government were identified as among the key factors that drive resource allocation in the health system. These findings have several implications for health policy in Ghana, which are discussed in the next section.
11.2 Policy Implications of the Study Four key policy implications of the study are examined here. These are the implications for the way equity in resource allocation has been examined; second, for health sector decentralisation and equity policies in Ghana; third, for the ‘inequitable’ capacity constraints facing different jurisdictions, and fourth, for effective collaboration under decentralisation.
48 Remember that equity is defined in this study as differential allocation of funds in favour of the most deprived regions, districts and sub-districts (see Chapter 4).
270 11.2.1 Examination of Equity in Resource Allocation The findings of this study raise concerns about the way equity in resource allocation under decentralisation has been examined. In Ghana, as in many countries with decentralised health systems, equity in resource allocation has been largely examined at the national level, focusing narrowly on inter-regional equity. Equity within regions has been virtually ignored. The key issue here is that without adequate attention to intra- regional equity in resource allocation, regions with high need districts or populations may be appropriately allocated more resources at the national level, but those needy districts or populations within these regions may still not receive resources commensurate with their needs. Evidence from this study clearly indicates that intra- regional resource allocation in Ghana is not equitable, if equity is defined as differential allocation in favour of worst-off areas49.
The major policy implications of this result are that Ghana needs a well-designed and transparent equity-focused policy for inter and intra-regional resource allocation, just as it has a clear policy for allocating resources to different levels of the health system (see Chapter 7). There is no assumption here that a common resource allocation formula that will ensure equity at all levels (national, regional and districts) can be developed. This is impossible given region-specific differences. Indeed, as argued by Sutton and Lock (2000), regional heterogeneity implies that a formula which ensures equity between regions does not achieve an equitable distribution of resources within regions. However, focusing on one level at the expense of others constrains the promotion of equity, particularly in decentralised systems. There is a need for attention to intra-regional equity in resource allocation. Policy makers need to seek coherent and practical methodologies for allocating resources intra-regionally if equity is to be widely promoted.
In Ghana, this may begin with a consideration of current resource allocation policies of all the ten regions to assess areas where common tools to improve equity can be developed. Any equity-focused policy for geographical resource allocation needs to take account of internally generated funds (IGF), because some districts have better economic conditions and are able to raise more revenue locally than others. Under an
49 This study found, for example, that urban districts and sub-districts, which are less deprived, have benefited more from resource allocation than their rural and often severely deprived counterparts.
271 equity-focused system, such districts ought to be allocated fewer resources from the central source. This should, however, be done with caution, as compensatory mechanisms may discourage local revenue generation (Green 2000).
11.2.2 Decentralisation and Equity Policies in Ghana The findings of this study have implications for the benefits of health sector decentralisation and equity policies in Ghana. As indicated earlier, one of the basic assumptions underlying the decentralisation policy is the promotion of equity in distribution of resources (Bossert et al. 2000). The policy assumes inter alia that local health authorities have superior knowledge about the health needs of their populations and are therefore better positioned to allocate resources more equitably to areas where needs are greatest (Silverman 1990; Levaggi and Smith 2004).
The results of this study showed clearly that in the Ashanti and Northern Regions, resources were not distributed to areas where needs were greatest. In the Ashanti Region, for example, resources were largely allocated either on the basis of population size or were equally divided across-the-board. Although population size ought to be a key component of any mechanism for equitable resource allocation, it has to be considered together with differential needs (Green et al. 2000; Eagar et al. 2001).
These findings raise serious concerns about the benefits of the health sector decentralisation policy in Ghana as far as equity in resource allocation is concerned. In particular, they question the amount of ‘decision-space’ within which local health managers make resource allocation decisions. There was evidence from the qualitative data that regional authorities were, in practice, not given a free hand to determine how resources should be allocated in their respective regions50. These findings call for a review of the entire policy of decentralisation of resource allocation decision-making, particularly, the ‘decision-space’ granted to local decision-makers to make independent resource allocation decisions.
50 Some regional participants mentioned that sometimes they receive funds from the national level already apportioned to individual districts. They also indicated that, at times, central authorities send funds directly to districts without informing them (see Chapter 9).
272 11.2.3 Inequitable Capacity Constraints The need to address the inequitable capacity constraints faced by different health sector jurisdictions is evident in this study. Migration of health professionals abroad has left the Ghanaian health system significantly understaffed, with varying capacity across regions and districts (see Chapter 2). Green et al. (2000) observed that it is crucial for any resource allocation process to be handled by staff with appropriate professional expertise who recognise the critical importance of budgets to achieve policy goals. This study found that while policy makers and managers see the need to shift more resources to deprived areas to promote equity, they were reluctant to shift funds to areas with limited local capacity to utilise funds more efficiently. This was clear evidence of efficiency concerns superseding equity goals in allocation of resources. However, it was reasonable since there is little point in shifting scarce resources to areas where the prospect that they will be put to their intended use is questionable.
For geographical equity in resource allocation to be effectively promoted, it is imperative that steps are taken to address the variations in capacity that prevent the effective use of resources to maximise health gains. This cannot be adequately done without a shift of non-financial resources, particularly human resources, to deprived areas which have borne a disproportionate burden of the health skill shortage in Ghana. The MOH currently has a policy to entice health personnel to deprived areas (MOH 2002). However, this needs to be strengthened, as the incentive packages are currently not attractive enough to motivate staff to move to deprived areas (see Section 11.4.).
Another way to tackle the differential capacity constraints is by incorporating what Mooney and Houston have described as MESH infrastructure in the resource allocation system (Mooney and Houston 2004). MESH stands for management, economic, social and human infrastructure. Mooney and Houston (2004) argued that the capacity to benefit from resource allocation by different jurisdictions entails good management, requires availability of resources, needs a socially well functioning community and, ideally, good human resources. Where some or all of these elements are missing, the resources might be wasted or at best used to lesser effects. Although the authors proposed MESH infrastructure in addition to a weighted capacity to benefit approach (which is more complicated to deal with), if policy makers in Ghana can improve on the
273 four elements making up MESH in different jurisdictions, it may have far-reaching effects for the more efficient use of resources and ultimately contribute to improving geographical equity.
11.2.4 Effective Collaboration under Decentralisation Collaboration under the current decentralisation framework is largely ineffective. First, this study found a limited avenue for effective collaboration within regions and districts in terms of how they deal with resource allocation. Predominantly, regional and district health administrations dealt with resource allocation within their respective jurisdictions in isolation. Second, collaboration between the district health service and District Assembly (local government) in the Ashanti and Northern Regions was found to be not only ineffective but also relationships were antagonistic in some districts. In a few districts where collaboration between the two was relatively effective, the latter provided significant financial and logistic support for district health delivery, thereby increasing the overall amount of resources available to the DHAs.
The poor collaboration was largely the result of the inconsistencies in Ghana’s decentralisation policy51 (see Section 10.4.5). There is the need for policy makers to promote effective collaboration across jurisdictions within the health system and between the health sector and the District Assembly. Effective collaboration would benefit the district health service, not only in terms of gaining extra resources from the District Assembly but also providing the platform for sharing innovative approaches to resource allocation across jurisdictions within the health sector.
Effective collaboration with the DA would also help the health sector to build partnerships with other sectors such as education, housing and agriculture, whose activities have considerable implications for health delivery. The DA provides the best avenue for such inter-sectoral partnership.
51 Ghana operates a parallel decentralisation stream; devolution of government and deconcentration within the health sector (see Chapter 10).
274 Finally, with the government aiming to channel more HIPC funds52 (of which part should into district health delivery) through the DA, effective collaboration is crucial for the health sector, if it will get its fair share of these funds. Policy makers, however, need to promote collaboration without limiting the decision-space available to local decision makers in making resource allocation decisions. Transparency about funding of district health service will be a key ingredient for effective collaboration between the health sector and the District Assembly. The next section reflects on some of the contributions made by this study to the equity and resource allocation debate.
11.3 Contributions to the Equity and Resource Allocation Debate Equity is a key health policy goal in many countries. It is, however, a difficult concept to interpret and measure. This study makes a number of conceptual and methodological contributions to the equity and resource allocation debate the most important of which are highlighted below.
11.3.1 Conceptual Contributions This study lends support to vertical equity as an important conceptual principle for resource allocation. Health policy makers face the dilemma of choosing between different equity principles for distributing resources (Powell and Exworthy 2003). While there is no universally agreed principle for promoting equity, there has been an overwhelming emphasis on applying of horizontal equity principles to resource allocation in many health systems, particularly the equal access for equal needs principle (Goddard and Smith 2001; Polikowski and Santos-Eggimann 2002).
A small but growing body of literature has recently argued for more attention to vertical equity goals (Wiseman and Jan 2000; Mooney 2002; Mooney et al. 2002; Black and Mooney 2002; McIntyre and Gilson 2002). Conceptually, this study adds to these recent studies that have argued for vertical equity principle to drive resource allocation. It
52 According to the 2003 Budget statement, the Ghana government is determined to use funds saved from the highly indebted poor country (HIPC) initiative as an instrument deepen decentralisation policy and will channel more funds directly to District Assemblies. The district health service has a claim to some of these funds. Without effective collaboration it will be difficult for the health sector to access part of these funds (MOH 2003).
275 draws on different perspectives of social justice to support the contention that society has a moral duty towards improving the lives of its least-advantaged members. Employing Rawls’ theory of justice as fairness, his maximin principle in particular, this study has re-emphasised the view that it is fair to discriminate against the well-off in resource allocation, if doing so will make the least-advantaged better-off than they would been under any alternative arrangement (Rawls 1972).
The study also draws on some aspects of Nozick’s entitlement theory, specifically his ‘principle of rectification of past injustice’ to caution that obligation towards the most disadvantaged should not be misconstrued as an act of benevolence. This is important as ‘disadvantageness’ is usually less a matter of choice and more the result of unfair policies of the past (see Chapter 4). Such conceptualisation of equity, as undertaken in this study, supports the choice of a vertical equity principle, which emphasises unequal but equitable treatment of unequals, rather than promoting horizontal equity, which emphasises equal treatment for equal need.
11.3.2 Methodological Contributions This study has made a number of methodological contributions to the equity and resource allocation debate. The first is the provision of empirical data from a developing country context to show the extent of equity in funding allocation under decentralisation. The equity debate, in the context of resource allocation, has been largely theoretical, focusing more on which equity principle or set of principles should drive resource allocation. Similarly, a substantial amount of the health care literature has emphasised the role of decentralisation in reducing inequalities, improving access for the poor, and equity in resource allocation (Bossert and Beauvais 2002). In contrast, relatively few empirical studies have confirmed these theoretical arguments.
By analysing quantitatively the extent of equity in funding allocation at all the three levels of the Ghanaian health system, this study has provided valuable insights into how the issue of equity in resource allocation is handled, in practice, at different health system levels. This will contribute to filling the current gap in knowledge regarding how resources are re-distributed intra-regionally to promote equity by local managers who
276 are increasingly becoming autonomous in decision-making as a result of deepening decentralisation.
In addition to the quantitative insights, this study has also drawn on perspectives of key stakeholders involved in the resource allocation process in Ghana to reveal the factors considered important when resources are being allocated. This crucial qualitative information has shed some light on why decentralisation of resource allocation decisions in the Ghanaian health system has not resulted in the shifting of more funds to the most deprived regions and districts, as proponents of the policy have argued. In summary, the empirical quantitative and qualitative evidence provided by this study contribute significantly to inform and support health policy reforms that seek to promote equity in Ghana and other countries, in particular the decentralisation policy.
The second contribution of this study is through the use of a mixed-method approach to investigate the extent of equity in funding allocation and to explore the factors that influenced the equitable allocation of resources. The literature survey found no study that goes beyond the quantitative measure of equity in resource allocation to explore, qualitatively, the factors that underpin the allocation process in such a comprehensive manner. This distinctive use of mixed-methods to study the issue of equitable resource allocation under decentralisation at different health system levels has provided a more complete picture than that provided by previous studies which have used either quantitative or qualitative approach, but not both. In this regard, this study has extended conventional methodological boundaries to promote the use of a mixed-methods approach in addressing similar research questions.
The final contribution made by this study in terms of methodology is the measurement of relative deprivation using existing datasets. In Ghana, as in many developing countries, it is virtually impossible to undertake any comprehensive analysis of a phenomenon at sub-regional levels because of lack of (or perceived lack of) data. Currently, the MOH depends largely on regional mortality and poverty indices estimated from the Ghana Living Standard Survey 1998/99 and the Ghana Demographic and Health Survey 1998 for assessment of deprivation/poverty levels among regions to guide resource allocation. While these surveys are nationally representative, they are sample surveys from which data cannot be disaggregated to district levels. With regions
277 as the unit of analysis in most health and socio-economic studies, very little is known about the conditions in individual districts within regions.
This study overcame that barrier by using key socio-economic, demographic and area data extracted from the 2000 Ghana Population and Housing Census dataset (Ghana Health Service 2002) to measure levels of relative deprivation across regions and districts as a proxy of need. The successful use of variables in an existing dataset to adequately measure relative deprivation supports the assertion of McIntyre and others (2002) that it is feasible to undertake deprivation measurement in data poor contexts. This study demonstrated that it is possible to measure relative deprivation at sub- regional levels in Ghana using routine datasets. This may encourage analyses which focus on the district level in the future. The valuable insights into levels of relative deprivation among districts in the Ashanti and Northern Regions may also stimulate debate on the allocation of other public sector resources.
11.4 Directions for Future Research Three directions for future research are suggested. First, there is the need for further research to examine the allocation of the salary budget and its impacts on overall equity in resource allocation in the health sector. As discussed in the study limitations (Chapter 10), the analysis in this study excludes the salary budget because decisions regarding staff recruitment, transfers and salary levels are not decentralised and are still under the tight control of the central MOH and the Ministry of Finance. Regional and sub- regional health authorities have little influence on staff matters. It was therefore inappropriate to analyse the allocation of the health budget to salaries in this study.
Nonetheless, salaries and staff allowances consume the greatest proportion of the Government of Ghana health budget. For example, the proportion of the GOG budget spent on staffing increased from about 55% in 1998 to about 84% in 2004 (MOH 2005; see also Chapter 7). Elsewhere in Africa, it has been established that part of the bias in the distribution of financial resources is reflected by the inequitable distribution of human resources and salaries (Claude Bodart et al. 2001). Clearly, efforts to redistribute the non-salary recurrent budget in Ghana will have little impact on health status if
278 redistribution of staff and the salary budget is not given sufficient attention. Further research examining the allocation of the salary budget between and within regions, from an equity perspective, will be useful not only in informing and supporting human resource and equity policies in the health sector, but also in providing valuable insights into how the MOH is shifting human resources in line with financial resources to address equity.
Second, due to time and resource constraints, this study focused on two of the ten regions in Ghana (Ashanti and Northern). Further research is needed to extend the analysis in this study to the remaining eight regions so that a more comprehensive nationwide picture of how the issue of equity in resource allocation has been handled across regions and districts can be built. Such analysis, in addition to the findings of this study, will provide more robust evidence for developing a national equity-focused policy for inter and intra-regional resource allocation.
Finally, future research analysing how resources have been allocated across different population groups, as defined by their socio-economic status, will be useful in strengthening district-level resource allocation and the case for vertical equity. Such analysis could not be undertaken in the present study because the relevant population- level data were not available, and also, this study was more interested in geographical equity.
Nonetheless, the study has demonstrated that decentralisation of resource allocation decision-making is not sufficient in promoting equity in distribution of funds and that there are complex array of factors that influence the equitable allocation of resources in the health sector. Unless adequate attention is paid to these factors and the entire process of resource allocation within regions in decentralised health systems, health equity in general, and equitable funding in particular, cannot be widely promoted.
279 APPENDIX A
Pearson’s correlation coefficients between demographic and socio-economic variables for the ten regions in Ghana
Variables Rural Fem Und-5Elderly Disabled Un No- No- No- No- Mud Wood employed Edu Elect pipe Toilet House fuel Rural 1.00 Female 0.66 1.00 Under-5 0.74 0.31 1.00 Elderly 0.38 0.71 -0.03 1.00 Disabled 0.55 0.79 0.23 0.84 1.00 Unemployed 0.01 0.18 -0.12 0.28 0.00 1.00 No-Education 0.73 0.52 0.72 0.11 0.19 0.44 1.00 No-Electricity 0.97 0.71 0.70 0.35 0.46 0.13 0.83 1.00 No-pipe 0.96 0.54 0.74 0.33 0.44 0.02 0.74 0.94 1.00 No-Toilet 0.51 0.50 0.45 0.08 0.08 0.56 0.92 0.67 0.48 1.00 Mud-house 0.97 0.62 0.81 0.24 0.40 0.09 0.85 0.98 0.94 0.65 1.00 Wood-fuel 0.95 0.55 0.83 0.21 0.47 -0.23 0.67 0.91 0.94 0.40 0.93 1.00
280 APPENDIX B
Pearson’s correlation coefficients between demographic and socio-economic variables for Ashanti and Northern Region
Disabl Un No- No- No- No- MD WD Variable Rural Fem Und-5 Elderly ed empl Edu Elec Pipe Tolt hous Fuel Rural 1.00 Female 0.49 1.00 Under-5 0.63 0.67 1.00 Elderly 0.03 -0.09 -0.11 1.00 Disabled 0.04 -0.21 -0.06 0.66 1.00 Unemployed -0.07 -0.03 0.18 0.21 0.28 1.00 No-Education 0.37 0.57 0.65 -0.65 -0.36 0.06 1.00 No-Electricity 0.90 0.64 0.78 -0.12 -0.05 -0.05 0.63 1.00 No-Pipe 0.84 0.45 0.67 0.18 0.13 -0.01 0.30 0.81 1.00 No-Toilet 0.34 0.43 0.60 -0.70 -0.37 0.14 0.97 0.58 0.28 1.00 Mud-house 0.53 0.49 0.60 -0.39 -0.33 0.07 0.67 0.65 0.50 0.67 1.00 Wood-Fuel 0.93 0.47 0.55 0.01 0.04 -0.17 0.35 0.80 0.76 0.31 0.51 1.00
281 APPENDIX C
Percentage Change in Actual Government of Ghana Funding (GOG 2-3) in Ashanti Region: 1998 – 2002
60
40
20 1998/99
1999/00
0 2000/01
2001/02 -20
% Shift in Actual Funding -40
-60
-80
t BAK dumasi Atwima Kwabre o Offinso Sekyere East Adansi WestAdansi East Ejisu-Juaben Amansie East Kumasi Metro Amansie Wes Afija SekyereSekyere West Ejura-Sek Ahafo-Ano Ahafo-AnoSouth North Asante-Akim North Asante-Akim South
DISTRICT
Note: Positive values show magnitude of funding gained. Negative values depict funding lost
282 APPENDIX D
Percentage Change in Actual Donor-Pooled Funds (DPF) in Ashanti Region: 1998 – 2002
200
150
100 1998/99 1999/00 50 2000/01 2001/02 0 %Shift in Actual Funding
-50
-100
t ast th es ast so h th E or E aben yere masi in r twima West W u BAK k u f o i si Kwabre Of sie A sie an ere West od o Sout im N d y ek n A Ejisu-J Ak Sekyere East Adans Afija Se Aman Kumasi Metro Aman Sek ra-S ju E Ahafo-A Ahafo-Ano N Asante- Asante-Akim South DISTRICT
Note: Positive values show magnitude of funding gained. Negative values depict funding lost
283 APPENDIX E
Percentage Change in Actual Government of Ghana Funding (GOG -3) in Northern Region: 2000 – 2002
60
40
20
0 2000/01 2001/02 -20 %Shift in Actual Funding
-40
-60
e a e i usi on ba al us onja pr Bol araga bugu pr G Tat eponi Yendi st am um am unucipal Nanum her u-K M M East Gonj C K e We st - eg al on- We Zabzugu- East M Savelugu-Nant Gushi Tol Tam Saboba
DISTRICT
284 APPENDIX F
Percentage Change in Actual Donor-Pooled Funds (DPF) in Northern Region: 1999 – 2002
20.0
15.0
10.0
5.0 1999/00 2000/01 0.0 2001/02
-5.0 % Shift in Actual Funding
-10.0
-15.0
i a i i ja us ja le di ga us ipal on ton ta on en a ugu an Bole on umb a ep ar uc N n -T r Y K mpr un st G a - umb a est G gu- a N gu Che gu K M W E u - - le lu bz ba hie st M est Mamprve a a ma W Z bo us olon E a Sa G T T Sa
DISTRICT
285 APPENDIX G
List of Documents Reviewed Prior to Selection of Deprivation Variables
International Literature Underprivileged area index (UPA) by Jarman (Jarman 1983; 1984) Townsend index of material deprivation (Townsend et al. 1986) Scottish deprivation index (Carstairs and Morris 1989) Swedish UPA (Bajekal et al. 1996) Geographic patterns of deprivation in South Africa (McIntyre et al. 2002).
Domestic policy and other documents: Ghana Poverty Reduction Strategy document (GPRS - Final Draft Version 2002) Ghana Demographic and Health Survey (GDHS 1992/93; 1998/99) Infant, Child and Maternal Mortality Study (1994) Ghana Living Standard Survey (1992; 1998) Health Sector 5-Year Programme of Work 1997-2001 Health Sector 5-Year Programme of Work 2002-2006, and The Health of the Nation: Reflections on First 5-Year Programme of Work 1997-2001.
286 BIBLIOGRAPHY
Australian Bureau of Statistics (2004). Population Estimates by Age & Sex NSW 2001. Catalogue No.3235.0.55.001. Canberra, Australian Bureau of Statistics.
Acheson, R. (1978). "The definition and identification of need for health care." Journal of Epidemiology and Community Health 32(1): 10-15.
Adamolekun, L. (1999). Decentralisation, sub-national governments and inter-governmental relations. In. Public Administration in Africa. L. Adamolekun (ed). Boulder, Westview Press.
Addai, E. and L. Gaere (2001). Capacity-building and Systems Development for Sector-Wide Approaches (SWAp): The Experience of the Ghana Health Sector (Unpublished). Accra - Ghana, Ministry of Health.
Agyapong, I. A. (1995). "Implementing primary health care under severe economic constraints: The Dangme West District of the Greater Accra Region of Ghana." African Primary Health Care in Times of Economic Turbulence Chabot, J., Harnmeijer, J. W., Streefland, P. H. (Editors)(Amsterdam: The Netherlands).
Agyapong, I. A. (1999). "Reforming health service delivery at the district level in Ghana: the perspective of a district medical officer." Health Policy and Planning 14 (1): 59-69.
Aikins, M. (2003). Emerging community health insurance schemes/mutual health organisations in Ghana: Achievements and challenges. Accra, Danida Health Sector Support Office (HSSO).
Akosah A.B, Nyonator F. K., Phillips J.F, Jones T.C (2003). Health sector reform, field experiments and systems research for evidence-based program change and development in Ghana. Draft. Bellagio conference.
Amoah, A. G. B., Owusu, S.K., Adjei, S (2002). "Diabetes in Ghana: A Community Based Prevalence Study in Greater Accra." Diabetes Research and Clinical Practice 56: 197- 205.
Andersen, R., Smedby, B., Vagero, D. (2001). "Cost-containment, solidarity and cautious experimentation: Swedish dilemmas." Social Science and Medicine 52: 1195-1204.
Andersson, F. and C. H. Lyttkens (1999). "Preferences for equity in health behind a veil of ignorance." Health Economics 8(5): 369-378.
Andrew Green, C. C. (2003). "Health systems in developing countries: public sector managers and the management of contradictions and change." International Journal of Health Planning and Management 18(S1): S67-S78.
Anell, A. T. (1996). "The monopolistic integrated model and health care reform: the Swedish experience." Health Policy 37: 19-33.
Anne-Emanuelle, B., Zimmerman, S., Garfield, R., (2000). "To Decentralize or Not to Decentralize, It That the Question? Nicaraguan Health Policy under Structural Adjustment in the 1990s." International Journal of Health Services. 30 (1): 111 - 128.
287 Appiah, F., Ayee, J.A., Appeah, J., Baah-Wiredu, K., Martin, R., Steffensen, J., Trollegaard, S., (2000). Fiscal decentralisation and sub-national government finance in relation to infrastructure and service provision in Ghana, International Bank for Reconstruction and Development.
Armstrong, P. and H. Armstrong (1999). "Decentralised health care in Canada." British Medical Journal 318: 1201-1204.
Arrow, K. J. (1973). "Rawls’s Concept of Just Saving." Swedish Journal of Economics 75: 323- 35.
Asamoa-Baah, A. and P. Smithson (1999). "Donors and the Ministry of Health: New Partnerships in Ghana’." WHO/EIP/99.1.
Ashiabi, G. (2000). "Some correlates of childhood poverty in Ghana." Journal of Children and Poverty 6(2): 155–68.
Bajekal, M., Sundquist, J., Jarman, B. (1996.). "The Swedish UPA score: An administrative tool for identification of underprivileged areas." Scandinavian Journal of Social Medicine 24(3): 177-184.
Bambas, A. and A. Casas (2001). Assessing equity in health: conceptual criteria. In: Equity and Health: Views from the Pan American Sanitary Bureau. Pan American Health Organisation: Occasional Publication No. 8.
Bankauskaite, V., Saltman, R., Vrangbaek, K. (2004). "The role of decentralization of European health care systems." Report to Institute for Public Policy Research, London.
Barnum, H., Kutzin, J., Saxenian, H., (1995). "Incentives and Provider Payment Methods." International Journal of Health Planning and Management 10(1): 23-45.
Bartley, M., Blane, D., Smith, G. D. (1998). "Introduction: beyond the Black Report." Sociology of Health & Illness 20(5): 563-577.
Baum, F. (2001). "Health, equity, justice and globalisation: some lessons from the People's Health Assembly." Journal of Epidemiology and Community Health 55(9): 613-616.
Beauchamp, T. L. (1987). "Philosophical Problems of Allocating Medical Resources." Revue De Metaphysique Et De Morale 92(3): 293-306.
Beauchamp, T. L. (1991). "Ethical Theory and Epidemiology." Journal of Clinical Epidemiology 44: S5-S8.
Beauchamp, T. L. and J. F. Childress (1989). Principles of biomedical ethics. New York, Oxford University Press.
Beauchamp, T. L. and J. F. Childress (1994). Principles of Bioethics. (4th Edition) Oxford: Oxford University Press.
Bedard, K., Dorland, J., Gregory, A. W., Roberts, J., (2000). "Needs-based health care funding: implications for resource distribution in Ontario." Canadian Journal of Economics. 33(4): 981-1008.
288 Bedard, K., Dorland, J., Gregory, A.W., Rosenberg, M. (1999). "Standardized mortality ratios in capitation models: empirical issues from Canadian data." Canadian Public Policy 25: 47-64.
Belli, P. C. (2004). The impact of resource allocation and purchasing reforms on equity. Washington, DC, The International Bank for Reconstruction and Development/ World Bank.
Bennet, S. (1992). "Promoting the private sector: a review of developing country trends." Health Policy and Planning 7(2): 97-110.
Bennett, A. E. and W. W. Holland (1977). "Rational planning or muddling through." Lancet 1(8009): 464–466.
Bevan, G. (1991). "Equity in the use of health care resources. Current Concerns." WHO/SHS/CC/91.1 SHS Paper number 3.
Bijlmakers, L. and S. Chichanga (1996). "District health service costs, resource adequacy and efficiency: a comparison of three districts." Ministry of Health and Child Welfare, Zimbabwe.
Birch, S. and S. Chambers (1993). "To each according to need: a community-based approach to allocating health care resources." Canadian Medical Association Journal 149(5): 607- 612.
Birch, S., Eyles, J., Newbold, K.B (1993). "Equitable access to health care: methodological extensions to the analysis of physician utilization in Canada." Health Economics 2(2): 87-101.
Birdsall, N. and J. Estelle (1993). "Efficiency and Equity in Social Spending: How and Why Governments Misbehave." In Michael Lipton and Jacques van der Gaag, (Eds.) Including the Poor (New York: Oxford University Press for the World Bank.).
Birdsall, N. and R. Hecht (1994). "Swimming Against the Time: Strategies for Improving Equity in Health." Working Papers 305(Inter-American Development).
Birdsall, N. and E. James (1993). Health government and the poor: the case for the private sector. In: Gribble JN and Preston SH (eds) The epidemiological transition: policy and planning implications for developing countries. Washington DC, National Academy Press.
Black, M. and G. Mooney (2002). "Equity in Health Care from a Communitarian Standpoint." Health Care Analysis 10(2).
Blas, E. and M. Limbambala (2001). "User-payment, decentralization and health service utilization in Zambia." Health Policy Planning 16(Suppl. 2): 19–28.
Blaxter, M. (1983). "Inequalities in Health - the Black Report - Townsend,P, Davidson,N." Journal of Social Policy 12(APR): 284-285.
Blecher, M. S. and D. McIntyre (1995). "Expenditure on Health Research in South-Africa, 1991/1992." South African Medical Journal 85(5): 365-370.
289 Bloom, G. (2001). "Equity in health in unequal societies: meeting health needs in Contexts of Social Change." Health Policy 57: 205-224.
Bloom, G. and D. McIntyre (1998). "Towards equity in health in an unequal society." Social Science and Medicine 47(10): 1529-1538.
Bommier, A. and G. Stocklov (2002). "Defining health inequality: why Rawls succeeds where social welfare theory fails." Journal of Health Economics 21: 497-513.
Bossert, T. (1998). "Analysing the decentralisation of health systems in developing countries." Social Science and Medicine 47(10): 1513-1527.
Bossert, T., Beauvais, J., Bower, D. (2000). "Decentralisation of health systems: preliminary review of four country case studies." Major Applied Research 6, Technical Report No. 1(Bethesda, MD: Partnership for Health Reform Project, Abt Associates Inc).
Bossert, T., Chitah, M. B., Bowser, D. (2003). "Decentralization in Zambia: resource allocation and district performance." Health Policy and Planning 18(4): 357-369.
Bossert, T., Hsiao, W., Barrera, M., Alarcon, L., Leo, M., Casares, C. (1998). "Transformation of ministries of health in the era of health reform: the case of Colombia." Health Policy and Planning 13(1): 59-77.
Bossert, T., Soebekti, R., Rai, N. K. (1991). "Bottom-up Planning in Indonesia - Decentralization in the Ministry of Health." Health Policy and Planning 6(1): 55-63.
Bossert, T. J. (1981). "Health-Policy Making in a Revolutionary Context - Nicaragua, 1979- 1981." Social Science and Medicine 15(4C): 225-231.
Bossert, T. J. and J. C. Beauvais (2002). "Decentralisation of health systems in Ghana, Zambia, Uganda and the Philippines: a comparative analysis of decision space." Health Policy and Planning 17(1): 14-31.
Bossert, T. J., Larranaga, O., Giedion, U., Arbelaez, J. J., Bowsers, D. M. (2003). "Decentralization and equity of resource allocation: evidence from Colombia and Chile." Bulletin of the World Health Organization 81(2): 95-100.
Bossert, W. (1998). "Welfarism and rationalizability in allocation problems with indivisibilities." Mathematical Social Sciences 35(2): 133-150.
Bossert, W. and M. Fleurbaey (2002). "Equitable insurance premium schemes." Social Choice and Welfare 19(1): 113-125.
Bourne, D., Pick, W., Taylor, S., McIntyre, D., Klopper, J (1990). "A Methodology for Resource Allocation in Health Care for South Africa: Part 3: A South African Health Resource Allocation Formula." South African Medical Journal 77: 456–59.
Bradshaw, G. and P. L. Bradshaw (1995). "The equity debate within the British National Health Service." Journal of Nursing Management 3: 161 168.
Bradshaw, J. S. (1972). A taxonomy of social need. In: Problems and Progress in Medical Care: Essays on Current Research. McLachlan G. (Ed.) Oxford: Oxford University Press.
290 Bratton, M., Lewis, P., Gyimah-Boadi, E. (2001). "Constituencies for reform in Ghana." Journal of Modern African Studies 39(2): 231-259.
Braveman, P. and E. Tarimo (2002). "Social inequalities in health within countries: not only an issue for affluent nations." Social Science and Medicine 54(11): 1621-1635.
Braveman, P. and S. Gruskin (2003). "Poverty, equity, human rights, and health." Bulletin World Health Organization 81: 539-545.
Breman, A. and C. Shelton (2001). "Structural Adjustment and Health: A literature review of the debate, its role players and presented empirical evidence." WHO: Commission on Macroeconomics and Health Working paper series WG6:6.
Brennan, M. and R. Lancashire (1978). "Association of childhood mortality with housing status and unemployment." Journal of Epidemiology and Community Health 32(1): 28-33.
Bryant, M. (2001). "Is it possible to plan for decentralisation: Reforming Health Systems and Programs." Management Sciences for Health.
Buchanan, A. E. (1984). "The right to a decent minimum of health care." Philosophy & Public Affairs 13(1): 55-78.
Buor, D. (2003). "Mothers' education and childhood mortality in Ghana." Health Policy 64(3): 297-309.
Buor, D. (2004). "Determinants of utilisation of health services by women in rural and urban areas in Ghana." GeoJournal 61(1): 89-102.
Buse, K. (1994). "Spotlight on international organisations: the World Bank." Health Policy and Planning 9(1): 95-99.
Busse, R., van der Grinten, T., Svensson, P.G (2002). Regulating entrepreneurial behaviour in hospitals: theory and practice. In Regulating entrepreneurial behaviour in European health care systems. R. B. Saltman, R. Busse and R. Mossialos (eds), Open University Press: 126-145.
Buxton, M. and R. Klein (1978). "Allocating Health Resources. Royal Commission on the NHS." Research Paper 3. London: HMSO.
Caldwell, J. C. (1992). "Old and new factors in health transitions." Health Transition Review 2(Supplementary Issue).
Campos-Outcalt, D., Kewa, K., Thomason, J (1995). "Decentralisation of health services in Western highland province, Papua New Guinea: an attempt to administer health services at the sub-district level." Social Science and Medicine 40: 1091-1098.
Canagarajah, S. and X. Ye (2001). "Public health and education spending in Ghana in 1992- 1998: issues of equity and efficiency." World Bank Working Paper.
Carr-Hill, R. A. (1991). "Allocating resources to health care: Is the QALY (Quality Adjusted Life Year) a technical solution to a political problem?" International Journal of Health Services 21(2): 351-363.
291 Carr-Hill, R. A., Sheldon, T. A., Smith, P., Martin, S., Peacock, S., Hardman, G. (1994). "Allocating resources to health authorities: development of method for small area analysis of use of inpatient services." British Medical Journal 309: 1046-1049.
Carstairs, V. (1995). "Deprivation indices: their interpretation and use in relation to health." Journal of Epidemiology and Community Health 49: S3–8.
Carstairs, V. and R. Moris (1989). "Deprivation: explaining differences in mortality between Scotland and England." British Medical Journal 299(886-9).
Casebeer, A. L. and K. J. Hannah (1998). "The process of change related to health policy shift: reforming a health care system." International Journal of Public Health Management 11(7): 566-582.
Cassels, A. (1995). "Health sector reform: Key issues in less developed countries." Forum on Health Sector Reform, Discussion Paper 1.
Cassels, A. and K. Janovsky (1991). "A time of change: health policy, planning and organisation in Ghana." WHO/SHS/CC 91.2.
Cassels, A. and K. Janovsky (1996). "Reform of the health sector in Ghana and Zambia: commonalities and contrasts." WHO/SHS/CC.
Castro-Leal, F., Dayton, L., Demery, L., Mehra, K. (2000). "Public spending on health care in Africa: do the poor benefit?" Bulletin of the World Health Organization 78(1).
CHAG (2000). Annual Report 1999. Accra - Ghana, Christian Health Association of Ghana.
Chawlai, M. and R. P. Ellis (2000). "The impact of financing and quality changes on health care demand in Niger." Health Policy and Planning 15(1): 76-84.
Cloke, P. and L. Davies (1992). "Deprivation and lifestyle in rural Wales - Towards a cultural dimension." Journal of Rural Studies 8: 349-358.
Cloke, P., Milbourne, P., Thomas, C (1994). Lifestyles in rural England. London, Rural Development Commission.
COHRED (2000). "Health research in Tanzania: how should public money be spent?" The Centre on Health Research for Development COHRED Document November 9.
Collins, C. (1989). "Decentralisation and the need for political and critical analysis." Health Policy and Planning 2: 168-171.
Collins, C., Araujo, J., Barbosa, J., (2000). "Decentralising the health sector: issues in Brazil." Health Policy 52(2): 113-127.
Collins, E. and K. Klein (1980). "Equity and the NHS: Self reported morbidity, access and primary care." British Medical Journal 281: 1111-5.
Congdon, P. (2001). "Health status and health life measures for population health need assessment: Modelling variability and uncertainty." Health and Place 7: 13-25.
Cookson, R. and P. Dolan (2000). "Principle of justice in health care." Journal of Medical Ethics 26: 323-329.
292 Cooper, R.S., Rotimi, C.N., Kaufman, J.S., Muna, W.F.T., Mensah, G.A., (1998). "Hypertension treatment and control in sub-Saharan Africa: the epidemiological basis for policy." British Medical Journal 316(7131): 614-617.
Costa-i-Font, J. (2005). "Inequalities in self-reported health within Spanish Regional Health Services: devolution re-examined?" International Journal of Health Planning and Management 20(1): 41-52.
Crampton, P. and M. Laugesen (1995). The use of indices of need in resource allocation formulae for primary health care. Discussion Paper No. 3. Wellington, Health Services Research Centre.
Crampton, P., Salmond, C., Sutton, F. (1997a). "NZDep91: A new index of deprivation." Social Policy Journal of New Zealand 9: 186-193.
Creese, A. (1997). "User fees." British Medical Journal 315(7102): 202-203.
Creswell, J. W. (2003). Research design: qualitative, quantitative and mixed methods approaches (2nd Edition). London, Sage Publications.
Creswell, J. W. and D. L. Miller (2000). "Determining validity in qualitative inquiry." Theory into Practice 39(3): 124-130.
Criel, B. (1998). "District-Based Health Insurance in Sub-Saharan Africa." Studies in Health Services Organisation and Policy 10(Institute of Tropical Medicine, Antwerp).
Criel, B., Macq, J., Bossyns, P., Hongoro, C.H. (1996). "A coverage plan for health centres in Murewa District in Zimbabwe: an example of action research." Tropical Medicine & International Health 1: 699-709.
Crowder, G. (2003). "Pluralism, Relativism and Liberalism in Isaiah Berlin." Refereed paper presented to the Australasian Political Studies Association Conference, University of Tasmania, Hobart 29 September – 1 October 2003. http://www.utas.edu.au/government/APSA/GCrowder.doc.pdf.
Culyer, A. J. (1976). Need and the National Health Service. London, Martin Robertson.
Culyer, A. J. (1979). "What Do Health-Services Do for People." Search 10(7-8): 262-268.
Culyer, A. J. (1983). "Economics and Social Policy." Social Policy & Administration 17(2): 168-169.
Culyer, A. J. (1995). "Need - the Idea Wont Do - but We Still Need It." Social Science and Medicine 40(6): 727-730.
Culyer, A. J. (2001). "Economics and ethics in health care." Journal of Medical Ethics 27(4): 217-222.
Culyer, A. J. (2001). "Equity - some theory and its policy implications." Journal of Medical Ethics 27(4): 275-283.
Culyer, A. J., Van Doorslaer, E., Wagstaff, A. (1992a). "Comment: Utilization as a Measure of Equity." Journal of Health Economics 11(1): 93-98.
293 Culyer, A. J., E. van Doorslaer, et al. (1992b). "Access, Utilization and Equity - a Further Comment." Journal of Health Economics 11(2): 207-210.
Culyer, A. J. and A. Wagstaff (1993). "Equity and Equality in Health and Health-Care." Journal of Health Economics 12(4): 431-457.
Culyer, A. J. and A. Wagstaff (1995). "Qalys Versus Hyes - a Reply to Gafni, Birch and Mehrez." Journal of Health Economics 14(1): 39-45.
Culyer, A. J., Wiseman, J., Drummond, M. F. West, P. A. (1982). "Revenue Allocation by Regression - a Rejoinder." Journal of the Royal Statistical Society Series a-Statistics in Society 145: 127-133.
Cumming, J. and N. Mays (2002). "Reform and counter reform: how sustainable is New Zealand's latest health system restructuring?" Journal of Health Service Research and Policy 1: 46-55.
Cummins, A. G. (2004). "North-south divide in social inequalities in Great Britain - Health inequalities in Wirral: a living Black report?" British Medical Journal 329(7456): 52-53.
Danida (2001). "Health Sector Support Phase II Ghana: Mid-term review of three components (Improving access to health care; strengthening district and sub-district capacity; improving collaboration with the private sector)." Danish International Development Agency, Accra- Ghana.
Danida (2002). Health sector programme support (Phase III): Programme support document. Accra - Ghana.
Daniels, N. (1981). "Health-Care Needs and Distributive Justice." Philosophy & Public Affairs 10(2): 146-179.
Daniels, N. (1982). "Equity of Access to Health-Care - Some Conceptual and Ethical Issues." Milbank Memorial Fund Quarterly-Health and Society 60(1): 51-81.
Daniels, N. (1985). "Fair Equality of Opportunity and Decent Minimums - a Reply to Buchanan." Philosophy & Public Affairs 14(1): 106-110.
Daniels, N. (1990). "Equality of What - Welfare, Resources, or Capabilities." Philosophy and Phenomenological Research 50: 273-296.
Daniels, N. (1992). "Liberalism and Medical Ethics -- Just Doctoring: Medical Ethics in the Liberal State by Troyen A. Brennan / The Ends of Human Life: Medical Ethics in a Liberal Polity by Ezekiel J. Emanuel." The Hastings Center Report 22(6): 41.
Daniels, N. (1996). "Justice, fair procedures, and the goals of medicine." Hastings Center Report 26(6): 10-12.
Daniels, N. (1997). "Procedural justice - Not moral principles." Behavioural Healthcare Tomorrow 6(4): 52-&.
Daniels, N. (2000). "Accountability for reasonableness." British Medical Journal 321: 1300-1.
Daniels, N., B. P. Kennedy, et al. (1999). "Why justice is good for our health: The social determinants of health inequalities." Daedalus 128(4): 215-251.
294 Daniels, N. and J. Sabin (1997). "Limits to health care: Fair procedures, democratic deliberation, and the legitimacy problem for insurers." Philosophy & Public Affairs 26(4): 303-350.
Daniels, N. and J. Sabin (1998). "The ethics of accountability in managed care reform." Health Affairs 17(5): 50.
Daniels, N. and Sabin, J.E (2002). Settings limits fairly: can we learn to share medical resources? Oxford, Oxford University Press.
Davis, M. (1977). "Necessity and Nozick theory of entitlement." Political Theory 5(2): 219-232.
De Vaus, D. A. (2001). Research design in social research. London, SAGE Publications Ltd.
Deeble, J., Mathers, C., Smith, L. (1998). "Expenditures on health services for Aboriginal and Torres Strait Islander people." Public Affairs, Commonwealth Dept. of Health and Family Services - Canberra.
Deeble, J. S. (2000). "Medicare's maturity: shaping the future from the past." Medical Journal of Australia 173: 44-47.
DHSS (1976). Sharing Resources for Health in England. London: Department of Health and Social Services, HMSO.
Diderichsen, F. (1993). "Market reform in Swedish health care: a threat to or salvation for the universalistic welfare state?" International Journal of Health Services 23: 185-188.
Diderichsen, F., Varde, E., Whitehead, M. (1997). "Resource allocation to health authorities: the quest for an equitable formula in Britain and Sweden." British Medical Journal 315(7112): 875-878.
Doherty, J., McIntyre, D., Bloom, G., Brijlal, P. (1999). "Health expenditure and finance: who gets what?" Bulletin of the World Health Organization 77(2): 156-159.
Doherty, J. and A. van den Heever (1997). "Re-distributing Health Care Resources on a Geographical Basis: Lessons Learned in South Africa." Centre for Health Policy: Community Health, University of Witwatersrand.
Doherty, J. and A. van den Heever (1997). A Resource Allocation Formula in Support of Equity and Primary Health Care, Center for Health Policy: Department of Community Health, University of Witwatersrand.
Donabedian, A. (1971). "Social responsibility for personal health services: An examination of basic values." Inquiry 8(3-19).
Donabedian, A. (1973). Aspects of medical care administration. Cambridge: MA, Harvard University Press.
Donabedian, A. (1980). The definition of quality and approaches to its assessment. MI; Health Administration Press.
Donaldson, C. and K. Gerard (1993). "Economics of Health Care Financing: The Visible Hand." New York: St. Martin's Press.
295 Dunteman, G. H. (1989). Principal components analysis. CA: Sage Publications.
Eachus, J., Williams, M., Chan, P., Smith, G. D., Grainge, M., Donovan, J., Frankel, S. (1996). "Deprivation and cause specific morbidity: evidence from the Somerset and Avon survey of health." British Medical Journal 312(7026): 287-292.
Eager, K., Garrett, P., Lin, V. (2001). "Health planning: Australia perspectives." Allen & Unwin.
Eaves, D. (1998). "An examination of the concept of equity and the implications for health policy if equity is re-asserted as one of the key government objectives for the National Health Service." Journal of Nursing Management 6(4): 125.
Ebenstein, A. O. (1991). The Greatest Happiness Principle: An Examination of Utilitarianism. Garland, New York.
Ebrahim, G. J. and J. P. Ranken (1988). Primary health care: reorienting organisational Support. London: Macmillan.
Eckstein, G. and R. Gibberd (1994). A Relative Health Need Index for New South Wales Areas and Districts. Newcastle, Health Services Research Group: University of Newcastle.
Ensor, T., Dakpallah, G., Osei, D. (2001). "Geographical resource allocation in health sector of Ghana. A (draft) paper prepared for Ministry of Health." Government of Ghana under a contract with Oxford Policy Management and Financial Support of DFID, Ghana.
Evans, R. G. (1984). Strained mercy: The economics of Canadian health care. Butterworth, Toronto.
Evans, R. G. (1992). "What seems to be the problem? The international movement to restructure health care systems." Centre for Health Services and Policy Research. University of British Columbia, Discussion Paper 92:8D.
Eyles, J., Birch, S., Chambers, S., Hurley, J., Hutchison, B. (1991). "A Needs-based Methodology for Allocating Health Care Resources in Ontario Canada: Development and an Application." Social Science and Medicine 33: 489–500.
Folwell, K. (1995). "Single measures of deprivation." Journal of Epidemiology and Community Health 49(Supp 2): S51 – S56.
Forster, D. P. (1977). "Mortality, Morbidity, and Resource Allocation." Lancet 1: 997–8.
Gakidou, E. E., Murray, C.J.L., Frenk, J (2000). "Defining and measuring health inequality." Bulletin of World Health Organization 78: 42-54.
Gandjour, A. and K. W. Lauterback (2000). "Allocation resources in health care: A comparison of cost-effectiveness analysis and evidence-based medicine." HEPAC 1: 116-121.
Gerdtham, U. G. (1997). "Equity in health care utilization: further tests based on hurdle models and Swedish micro data." Health Economics 6(3): 303-19.
Gerdtham, U. G., Johannesson, M., Lundberg, L., et al. (1998). "A note on validating Wagstaff and van Doorslaer's health measure in the analysis of inequalities in health." Journal of Health Economics forthcoming.
296 Gericke, C. A., Riesberg, A., Busse, R. (2005). "Ethical issues in funding orphan drug research and development." Journal of Medical Ethics 31(3): 164-168.
Ghana Health Service (2004). "Implementation of the year 2003 programme of work: Annual Report." Ghana Health Service. Accra - Ghana
Ghana Ministry of Finance (2002). "Ghana Poverty Reduction Strategy (GPRS)." Ministry of Finance Final Draft Version.
Ghana Ministry of Health (1995). "Medium Term Health Strategy, Towards Vision 2020." Ministry of Health Accra, Ghana.
Ghana Ministry of Health (1998). "Health Sector Five-Year Programme of Work: 1997-2001." Ministry of Health Accra, Ghana.
Ghana Ministry of Health (2001). "The health of the nation: Reflection on the first five-year health sector programme of work, 1997-2001." Ministry of Health, Accra - Ghana.
Ghana Ministry of Health (2003). "Global expenditure report: 1999-2002." Ministry of Health, Ashanti Region: Kumasi
Ghana Ministry of Health (2003). "Health Sector Five-Year Programme of Work: 2002-2006." Ministry of Health Accra, Ghana.
Ghana Ministry of Health (2003). "Regional Health Expenditure Data." Ghana Ministry of Health, Northern Region: Tamale.
Ghana Statistical Service (1994). Ghana Demographic and Health Survey (GDHS) 1993. Accra: Ghana Statistical Service.
Ghana Statistical Service (1994). Ghana Living Standard Survey (GLSS 3) 1992/93. Report of the Third Round. Accra, Ghana Statistical Service.
Ghana Statistical Service (1994). "Infant, child and maternal mortality study in Ghana." Project undertaken by the Ghana Statistical Service in collaboration with Ministry of health and UNICEF Accra: Ghana.
Ghana Statistical Service (1997). Core Welfare Indicators Questionnaire Survey (CWIQ). Accra, Ghana Statistical Service.
Ghana Statistical Service (1999). Ghana Demographic and Health Survey (GDHS) 1998. Accra, Ghana Statistical Service.
Ghana Statistical Service (2000). Ghana Living Standard Survey (GLSS 4) 1998/99. Report of the Fourth Round. Accra, Ghana Statistical Service.
Ghana Statistical Service (2000). Poverty Trends in Ghana in the 1990's. Accra - Ghana, Government of Ghana.
Ghana Statistical Service (2002a). "2000 Population and housing census: Summary of final results." Ghana Statistical Service, Accra: Ghana.
297 Ghana Statistical Service (2002). Population and housing census 2000: Summary Report. Accra - Ghana.
Ghana Statistical Service and Macro Inc. (1994). "Ghana Demographic and Health Survey 1993." Ghana Statistical Service and Macro International Inc.; Calverton, Maryland.
Gibbs, A., Sondalini, R., Pearse, J. (2002). "The NSW health resource distribution formula and health inequalities." NSW Public Health Bulletin 13(3): 42-44.
Gibson, A., Asthana, S., Brigham, P., Moon, G., Dicker, J. (2002). "Geographies of Need and the New NHS: Methodological Issues in the Definition and Measurement of the Health Needs of Local Populations." Health and Place 8(1).
Gilbert, R., Gibbeerd, B., Stewart, J. (1992). "The New South Wales Resource Allocation Formula: A Method for Equitable Health Funding." Australian Health Review 15: 6-21.
Gilgun, J. (1995). "We share something special: The moral discourse of incest perpetrators." Journal of Marriage and Family 57: 265-281.
Gillon, R. (1986). Philosophical Medical Ethics, John Wiley & Sons.
Gillon, R. (1994). "Medical ethics: four principles plus attention to scope." British Medical Journal 309(6948): 184-.
Gilson, L. (1994). "Equity in the Finance and Delivery of Health care - an International Perspective – van Doorslaer, E., Wagstaff, A., Rutten, F." Health Policy and Planning 9(3): 347-349.
Gilson, L. (1997). "The lessons of user fee experience in Africa." Health Policy and Planning 12(4): 273-285.
Gilson, L. (1998). "In defence and pursuit of equity." Social Science and Medicine 47(12): 1891-1896.
Gilson, L. (2003). "Trust and the development of health care as a social institution." Social Science and Medicine 56(7): 1453-1468.
Gilson, L. (2005). "Editorial: building trust and value in health systems in low- and middle- income countries." Social Science and Medicine 61(7): 1381-1384.
Gilson, L., Doherty, J., Lake, S., McIntyre, D., Mwikisa, C., Thomas, S. (2003). "The SAZA study: implementing health financing reform in South Africa and Zambia." Health Policy and Planning 18(1): 31-46.
Gilson, L., Kalyalya, D., Kuchler, F., Lake, S., Oranga, H., Ouendo, M. (2000). "The equity impacts of community financing activities in three African countries." International Journal of Health Planning and Management 15(4): 291-317.
Gilson, L. and D. McIntyre (2005). "Removing user fees for primary care in Africa: the need for careful action." British Medical Journal 331(7519): 762-765.
Gilson, L. and A. Mills (1995). Health Sector Reforms in Sub-Saharan Africa - Lessons of the Last 10 Years. Health Policy 32(1-3): 215-243.
298 Gilson, L., Mkanje, R., Grosskurth, H., Mosha, F., Picard, J., Gavyole, A., et al. (1997). "Cost- effectiveness of improved treatment services for sexually transmitted diseases in preventing HIV-1 infection in Mwanza Region, Tanzania." Lancet 350(9094): 1805- 1809.
Goddard, M. and P. Smith (2001). Equity of access to health care services: theory and evidence from the UK. Social Science and Medicine (53): 1149-1162.
Godvindasamy, P. and B. M. Ramesh (1997). Maternal education and the utilisation of maternal and child health services in India. Mumbia-India, International Institute for Population Services.
GOG/MOFA (1995). National plan of action on food and nutrition. Accra, Ghana Government/Ministry of Food and Agriculture.
Goldacre, M. J. (1981). "Mortality statistics as measures of need for outpatient services." British Medical Journal 283(870-871).
Gordon, D. (1995). "Census based deprivation indices: their weighting and validation." Journal of Epidemiology and Community Health 49 (Supplementary): S39-S44.
Green, A. (2001). "An introduction to health planning in developing countries." Oxford: Oxford University Press.
Green, A., Ali, B., Naeem, A., Ross, D. (2000). "Resource allocation and budgetary mechanisms for decentralised health systems: experiences from Balochistan, Pakistan." Bulletin of the World Health Organization 78(8).
Green, R. M. (1976). Healthcare and justice in contract theory perspective. Ethics and Health Policy. R. M. Veatch and R. Brenson. Cambridge, Ballinger Books.
Greene, J. C., Caracelli, V.J., Graham, W.F. (1989). "Towards a conceptual framework for mixed-method evaluation designs." Educational Evaluation and Policy Analysis 11(3): 255-274.
Grindle, M. S. and J. W. Thomas (1991). Public choices and policy change: The political economy of reform in developing countries. London, Johns Hopkins Press.
Grundy, J. H. (2001). "The impact of health system reform on remote health in Cambodia and the Philippines." Rural and Remote Health (online) 1(84): http://rrh.deakin.edu.au.
Guba, E. G. (1978). Towards a methodology of naturalistic inquiry in educational evaluation. CSE Monograph Series in Evaluation No. 8. Los Angeles:, Centre for the Study of Evaluation, University of California.
Gunning-Schepers, L. J. and K. Stronks (1999). "Inequalities in health - Future threats to equity." Acta Oncologica 38(1): 57-61.
Gupta, D. B. and A. Gumber (1999). "Decentralisation: Some initiatives in health sector." Economic and Political Weekly 34(6): 356-362.
Gwatkin, D. R. (1972). "Policies Affecting Population in West-Africa." Studies in Family Planning 3(9): 214-221.
299 Gwatkin, D. R. (1982). "Literacy, Education and Health Development - Policy Implications." Health Policy and Education 3(1): 109-112.
Gwatkin, D. R. (2000). "Health inequalities and the health of the poor: What do we know? What can we do?" Bulletin of the World Health Organization 78(1): 3-18.
Gwatkin, D. R. (2001). "The need for equity-oriented health sector reforms." International Journal of Epidemiology 30(4): 720-723.
Gwatkin, D. R. (2003). "How well do health programmes reach the poor?" Lancet 361(9357): 540-541.
Gwatkin, D. R. (2004). "Assessing inequalities in maternal mortality." Lancet 363(9402): 5-5.
Gwatkin, D. R. (2005). "How much would poor people gain from faster progress towards the Millennium Development Goals for health?" Lancet 365(9461): 813-817.
Gwatkin, D. R., Bhuiya, A., Victoria, C. G. (2004). "Making health systems more equitable." Lancet 364(9441): 1273-1280.
Gwatkin, D. R., Guillot, M., Heuveline, P. (1999). "The burden of disease among the global poor." Lancet 354(9178): 586-589.
Gwatkin, D. R. and P. Heuveline (1997). "Improving the health of the world's poor - Communicable diseases among young people remain central." British Medical Journal 315(7107): 497-498.
Gyimah-Boadi, E. (1997). "Adjusting society: The World Bank, the IMF, and Ghana." Journal of Developing Areas 32(1): 137-139.
Gyimah-Boadi, E. (2001). "A peaceful turnover in Ghana." Journal of Democracy 12(2): 103- 117.
Hagopian, A., Thompson, M.J., Fordyce, M., Johnson, K.E., Hart, G.H (2004). "The migration of physicians from sub-Saharan Africa to the United State of America: Measures of the African brain drain." Human Resources for Health 2(17).
Ham, C. (2003). "Improving the performance of health services: the role of clinical leadership." The Lancet.
Ham, C. (2005). "Lost in Translation? Health Systems in the US and the UK." Social Policy and Administration 39(2): 192-209.
Ham, C (Ed.) (1997). Health care reform: learning from international experience. Buckingham: Open University Press.
Harsanyi, J. C. (1977). "Rule Utilitarianism and Decision Theory." Erkenntnis 11: 25-53.
Hausman, D. and J. Le Grand (1999). "Incentives and health policy: primary and secondary care in the British National Health Service." Social Science and Medicine 49(10):1299-1307.
Heilig, S., Kushner, T., Thomasma, D., Agich, G. J., Atchley, W. A., Bayley, C., Bauer, K. A., et al. (2002). "Healthcare without harm: An ethical imperative." Cambridge Quarterly of Healthcare Ethics 11(2): 203-207.
300 Heuveline, P., Guillot, M., Gwatkin, D. R. (2002). "The uneven tides of the health transition." Social Science and Medicine 55(2): 313-322.
Hjortsberg, C. A. and C. N. Mwikisa (2002). "Cost of access to health services in Zambia." Health Policy and Planning 17(1): 71-77.
Holland, W. W. (1986). "The RAWP review: pious hopes. Resource Allocation Working Party." Lancet 8(2(8515)): 1087-90.
Homedesa, N. and A. Ugalde (2005). "Why neoliberal health reforms have failed in Latin America." Health Policy 71: 83-96.
Horton, R. (2001). "Ghana: defining the African challenge.[see comment]." Lancet 358 (9299): 2141-9.
Hsiao, W. (2000). "Health care finance in developing nations: A background paper prepared for the World Bank." http://www1.worldbank.org/hnp/hsd/documents/
Hsiao, W. C. (2003). What is a Health System? Why Should We Care? MA: Harvard School of Public Health.
Hudson, B. (1999). "Decentralisation and primary care groups: A paradigm shift for the National Health Service in England?" Policy and Politics 27(2): 159-172.
Humphreys, J. S. (1998). "Delimiting rural: implications of an agreed rurality index for health care planning and resource allocation." Australian Journal of Rural Health 6: 212-216.
Hurley, J., B. Birch, et al. (1995). "Geographically-decentralised planning and management in health care: some informational issues and their implications for efficiency." Social Science and Medicine 41: 3-11.
IMF (2000). Ghana: Selected Issues. Washington, D.C, International Monetary Fund.
Imman, R. and D. Rubinfield (1997). "Rethinking Federalism." Journal of Economic Perspectives 11(4): 43-64.
Janovsky, K. (1997). "Decentralization and health systems: an overview of international experience." WHO Technical Seminar: Summary of key issues.
Jarman, B. (1983). "Identification of underprivileged areas." British Medical Journal 286 (1705- 9).
Jarman, B. (1984). "Underprivileged areas: validation and distribution of scores." British Medical Journal 289: 1587-1592.
Jeppsson, A. (2001). "Financial priorities under decentralisation in Uganda." Health Policy and Planning 16(2): 187-192.
Jimenez de la Jara, J. and T. Bossert (1995). "Chile's health sector reform: lessons from four reform periods." Health Policy 32: 155–166.
Johansson, L. (1997). "Decentralisation from acute to home care settings in Sweden." Health Policy 41: S131-S143.
301 Jolliffe, I. T. (1986). "Principal component analysis." New York: Springer Verlag
Jordan, H., Roderick, P., Martin, D., Barnett, S. (2004). "Distance, rurality and the need for care: access to health services in South West England." International Journal of Health Geographics 3(1): 21.
Kaufman, J. S., Asuzu, M.C., Rotimi, C.N., Johnson, O.O., Owoaje, E.E., Cooper, R.S. (1997). "The absence of adult mortality data for sub-Saharan Africa: A practical solution." Bulletin of the World Health Organization 75: 389-395.
Killick, T. (1997). "Principals, agents and the failings of conditionality." Journal of International Development 9(4): 483-495.
Kirkpatrick, A. F. (1997). "The US attack on Cuba's health." Canadian Medical Association Journal 157: 281-284.
Kirkwood, B. (1988). Essentials of medical statistics. Oxford: Blackwell Scientific Publication
Kirkwood, B., Kirkwood, B.R., Sterne, J. (1999). Essentials of medical statistics, Oxford: Blackwell Scientific Publication
Klein, R. (2003). "Governance for NHS foundation trusts." British Medical Journal 326(7382): 174-175.
Kleinman, M., Burton, P., Croft, J (2002). "Links between finance and non-finance elements of local government: A literature review." Office of the Deputy Prime Minister London.
Knox, E. (1978). "Principles of allocation of health care resources." Journal of Epidemiology Community Health 32(1): 3-9.
Koivusalo, M. (1999). "Decentralisation and equity of healthcare provision in Finland." British Medical Journal 318: 1198-200.
Kolehmainen-Aitken, R. (1992). "The impact of decentralisation on health workforce development in Papua New Guinea." Public Administration Development 12: 175-191.
Kolehmainen-Aitken, R. (2004). "Decentralization's impact on the health workforce: Perspectives of managers, workers and national leaders." Human Resources for Health 2(1): 5.
Konadu-Agyemang, K. (2000). "The Best of Times and the Worst of Times: Structural Adjustment Programs and Uneven Development in Africa: The Case Of Ghana." The Professional Geographer 52(3): 469.
Krieger, N., William, D., Moss, N (1997). "Measuring social class in US public health research: concepts, methodologies, and guidelines." Annual Review of Public Health 18: 341- 378.
Kutzin, J. (1995). Experience with Organizational and Financing Reform of the Health Sector. Geneva: World Health Organization, Division of Analysis, Research and Assessment.
Kutzin, J. (2001). "A descriptive framework for country-level analysis of health care financing arrangements." Health Policy 56(3): 171-204.
302 Labonte, R., Schrecker, T., Sanders, D., Meeus, W (2004). Fatal indifference: The G8 and global health. Ottawa, UCT Press and International Development Research Centre.
Lake, S. (2000). Modelling ‘Need’ in Health Sector Resource Allocation Formulae: Application in a Low-income Country. Health Systems Financing in Low-Income African and Asian Countries, CERDI, Clermont Ferrand.
Lalta, S. (1991). "Improving Financial Management of Health-Care in the Caribbean - Conceptual and Empirical Issues." Social and Economic Studies 40(4): 37-57.
Lavy, V. and J. Germain (1994). Quality and Cost in Health Care Choice in Developing Countries, World Bank LSMS Working Paper No. 105.
Layte, R., Whelan, C.T., Maitre, B., Nolan, B (2001a). "Explaining deprivation in the European Union." Acta Sociologica 44(2): 105–122.
Le Grand, J. (1978). "The distribution of public expenditure: the case of health care." Economica 45: 125 - 142.
Le Grand, J. (1982). The strategy of equality: redistribution and the social services. London; Boston, G. Allen & Unwin.
Le Grand, J. (1987). "Equity, health and health care." Social Justice Research 1(3).
Le Grand, J. (1991). "The distribution of health care revisited: a commentary on Wagstaff, van Doorslaer and Paci, and O'Donnell and Propper.[comment]." Journal of Health Economics 10(2): 239-45.
Le Grand, J. (2001). What kind of health inequality? In: The issues panel for equity in health: the discussion papers, Oliver A, Cookson C, McDaid D, (eds). London: Nuffield Trust.
Leduc, E. (1997). "Defining rurality: A general practice rurality index for Canada." Canadian Journal of Rural Medicine 2: 125-131.
Lee, K., Collinson, S., Walt, G., Gilson, L. (1996). "Who should be doing what in international health: a confusion of mandates in the United Nations?" British Medical Journal 312(7026): 302-307.
Leon, D. A., Walt, G., Gilson, L. (2001). "Recent advances - International perspectives on health inequalities and policy." British Medical Journal 322(7286): 591-594.
Levaggi, R. and P. Smith (1994). "On the Intergovernmental Fiscal Game." Public Finance- Finances Publiques 49(1): 72-86.
Levaggi, R. and P. Smith (2003). Decentralization in health care: lessons from public economics. York, UK, Centre for Health Economics, University of York.
Lincoln, Y. S. and E. G. Guba (2000). Paradigmatic controversies, contradictions and emerging confluences. In. Handbook of qualitative research (2nd edition, pp. 163-188). N. K. Denzin and Y. S. Lincoln (Eds.), Thousand Oaks, CA: Sage.
Litvack, J., Ahmad, J., Bird, R. (1998). "Rethinking decentralization in developing countries." Washington (DC): World Bank Sector Studies.
303 Loewenson, R. (2000). "Public participation in health systems in Zimbabwe." IDS Bulletin- Institute of Development Studies 31(1): 14
Lynch, J. and G. Kaplan (2000). Socioeconomic position. In. Social Epidemiology. L. Berkman and I. Kawachi (eds). Oxford University Press.
Macfarlane, A. (1986). "Whatever Happened to the Black Report." British Medical Journal 293(6545): 504-504.
MacIntyre, S. (1997). "The Black Report and beyond: What are the issues?" Social Science and Medicine 44(6): 723-745.
Mack, J. and S. Lansley (1985). Poor Britain. London, George Allen and Unwin.
Main, J. A. and P. G. N. Main (1990). "The Black Report." British Medical Journal 301(6752): 608-608.
Makinen, M., Waters, H., Rauch, M., Almagambetova, N., Bitran, R., Gilson, L., McIntyre, D., Pannarunothai, S., Prieto, A. L., Ubilla, G., Ram, S. (2000). "Inequalities in health care use and expenditures: empirical data from eight developing countries and countries in transition." Bulletin of the World Health Organization 78(1): 55-65.
Marchand, S., Wikler, D., Landesmanm, B. (1998). "Class, health and justice." The Midbank Quarterly 76(3): 449-467.
Marshall, C. and G. B. Rossman (1999). Designing qualitative research (3rd ed.), Thousand Oaks, CA: Sage.
Martinez-Herrera, E. (2002). "From nation-building to building identification with political communities: Consequences of political decentralisation in Spain, the Basque Country, Catalonia and Galicia, 1978-2001." European Journal of Political Research 41(4): 421- 453.
Mathers, C. (1994). Health differentials among adult Australians aged 25 - 64 years, Australian Institute of Health and Social Welfare. Health Monitoring Series No. 1.
Matthews, C. (2003). "Caught in a vicious cycle." Australian Medicine 15(12): 16.
Maynard, A. and A. Ludbrook (1980a). "Applying resource allocation formulae to constituent parts of the UK." Lancet: 85-87.
Mays, N. (1995). "Geographical resource allocation in the English National Health Service, 1971-1994: the tension between normative and empirical approaches." International Journal of Epidemiology 24(90001): 96S-102.
Mays, N. and G. Bevan (1986). Resource allocation in the health service: A review of methods of the resource allocation working party (RAWP). Occasional Paper on Social Administration No. 81, Bedford Square Press.
McGary, H. (1999). "Distrust, social justice, and health care." The Mount Sinai Journal of Medicine 66: 236–240.
McIntyre, D. (1992). "Making Sense of the Health Budget - Reply." South African Medical Journal 82(3): 202-203.
304 McIntyre, D. (1997). "Input paper on health for the South Africa inequality report." Health Economics Unit Cape Town: University of Cape Town.
McIntyre, D. and L. Gilson (2000). "Redressing dis-advantage: Promoting vertical equity within South Africa." Health Care Analysis 8(3): 235-258.
McIntyre, D. and L. Gilson (2002). "Putting equity in health back onto the social policy agenda: experience from South Africa." Social Science and Medicine 54(11): 1637-1656.
McIntyre, D. and B. Klugman (2003). "The human face of decentralisation and integration of health services: Experience from South Africa." Reproductive Health Matters 11(21): 108-119.
McIntyre, D., Muirhead, D., Gilson, L. (2002). "Geographic patterns of deprivation in South Africa: informing health equity analyses and public resource allocation strategies." Health Policy and Planning 17: 30-39.
McIntyre, D., Ring, C., Carroll, D. (2003). "Effects of arousal and cardio-dynamics on the muscle stretch reflex in humans." Journal of Psychophysiology 17(2): 108-108.
McIntyre, D., Valentine, N., Cornell, J. (1995). "Private-Sector Health-Care Expenditure in South-Africa." South African Medical Journal 85(3): 133-135.
McKneally, M. F., Dickens, B.M., Meslin, E.M., Singer, P. A. (1997). "Bioethics for Clinicians: Resource Allocation." Canadian Medical Association Journal 157: 163-167.
McLoone, P. and F. A. Boddy (1994). "Deprivation and mortality in Scotland, 1981 and 1991." British Medical Journal 309(6967): 1465-1470.
McPake, B., Hanson, K., Mills, A. (1993). "Community Financing of Health-Care in Africa - an Evaluation of the Bamako Initiative." Social Science and Medicine 36(11): 1383-1395.
Mehrotra, S. (2000). Integrating economic and social policy: good practices from high- achieving countries. Working Papers 305. Florence, Innocenti Research Centre, UNICEF.
Mensah, K., Mackintosh, M., Henry, L. (2005). The skills drain of health professionals from the developing world: a framework for policy formulation. London, Medact.
Merriam, S. B. (1998). Qualitative research and case study applications in education. San Francisco, Jossey-Bass.
Miles, M. B. and A. M. Huberman (1994). Qualitative data analysis: A sourcebook of new methods. Newsbury Park, CA, Sage.
Mills, A., Antonius, R., Daniel, J., Gray, H., Haqq, E., Rutten, F. (2002). "The distribution of health planning and management responsibilities between centre and periphery: historical patterns and reform trends in four Caribbean territories." Health Policy 62(1): 65-84.
Mills, A., Vaughan, J.P., Smith, D.L., Tabibzadeh, I. (1990). "Health system decentralization: concepts, issues and country experience." Geneva: World Health Organization.
305 Ministry of Finance (2001). The Budget Statement and Economic Policy of the Government of Ghana for the 2001 Financial Year presented to Parliament. Accra - Ghana., Minister of Finance and Economic Planning.
Ministry of Finance (2003). The Budget Statement and Economic Policy of the Government of Ghana for the 2003 Financial Year presented to Parliament. Accra- Ghana, Minister of Finance and Economic Planning.
Mitton, C. and C. Donaldson (2003). "Resource Allocation in Health Care: Health Economics and Beyond." Health Care Analysis 11( 3): 245 - 257.
Mitton, C. R. (2002). "Priority setting for decision makers: using health economics in practice." European Journal of Health Economics 4.
Mogedal, S., Steen, S., Mpelumbe, G (1995). "Health sector reform and organizational issues at the local level: lessons from selected African countries." Journal International Development 7: 349–367.
MOH/PSU (2000). Handbook on the private health sector in Ghana. Accra - Ghana, Ministry of Health.
Mooney, G. (1986). Economics, Medicine and Health Care. Wheatsheaf, Brighton.
Mooney, G. (1992). "Economics, Medicine and Health Care (2nd Ed)." Prentice Hall Europe.
Mooney, G. (1992). Just health care: only medicine? Economics, Medicine and Health Care. Hertfordshire, Harvester Wheatsheaf: 179.
Mooney, G. (1994). Equity. Key issues in health economics. New York; London,, Harvester Wheatsheaf.
Mooney, G. (1994). Priority setting in health care. Key Issues in Health Economics. New York; London, Harvester Wheatsheaf.
Mooney, G. (1998). ""Communitarian claims" as an ethical basis for allocating health care resources." Social Science and Medicine 47(9): 1171-1180.
Mooney, G. (2000). "Judging goodness must come before judging quality--but what is the good of health care?" International Journal for Quality in Health Care 12(5): 389-94.
Mooney, G. (2000). "Vertical equity in health care resource allocation." Health Care Analysis 8(3): 203-15.
Mooney, G. (2003). "Inequity in Australian health care: how do we progress from here?" Australian & New Zealand Journal of Public Health 27(3): 267-70.
Mooney, G., Hall, J., Donaldson, C., Gerard, K. (1991). "Utilization as a Measure of Equity - Weighing Heat." Journal of Health Economics 10(4): 475-480.
Mooney, G. and S. Houston (2004). "An Alternative Approach to Resource Allocation Weighted Capacity to Benefit Plus MESH Infrastructure." Applied Health Economics and Health Policy 3(1): 29-33.
306 Mooney, G. and S. Jan (1997). "Vertical equity: weighting outcomes? or establishing procedures?" Health Policy 39(1): 79-87.
Mooney, G., Jan, S., Seymour, J. (1994). "The NSW health outcomes initiative and economic analysis." Australian Journal of Public Health 18(3): 244-8.
Mooney, G., Jan, S., Seymour, J., Wiseman, V. (2002). "Staking a claim for claims: a case study of resource allocation in Australian Aboriginal health care." Social Science and Medicine 54(11): 1657-1667.
Mooney, G. and V. Wiseman (2000). "Burden of disease and priority setting." Health Economics 9(5): 369-72.
Mooney, G. H. (1983). "Equity in health care: confronting the confusion." Effective Health Care 1(4): 179-85.
Morris, R. and V. Carstairs (1999). "Which deprivations? A comparison of selected deprivation indices." Journal of Public Health Medicine 13: 318-326.
Morse, J. M. (1991). "Approaches to qualitative-quantitative methodological triangulation." Nursing Research 40(1): 120-123.
Murray, C. J. L., Gakidou, E.E., Frenk, J. (1999). "Health inequalities and social group difference: What should we measure?" Bulletin of the World Health Organization 77(7): 537-543.
Murray, C. J. L. and A. D. Lopez (1996). Estimating causes of death: new methods and global and regional application for 1990. In. The global burden of disease. C. J. L. Murray and A. D. Lopez. Boston: MA, Harvard University Press.
Neuman, W. L. (2000). Social research methods: quantitative and qualitative approaches (4th edition). Boston, Allyn and Bacon.
Nolan, B. and V. Turbat (1995). Cost Recovery in Public Health Services in Sub-Saharan Africa. Washington, DC.
Nolan, B. and C. T. Whelan (1996a). Resources, deprivation and poverty. Oxford, Clarendon Press.
Normand, C. (1991). "Economics, health and the economics of health." British Medical Journal 303: 1572-1577.
Nozick, R. (1974). Anarchy, state, and utopia, New York: Basic Books
NSW-DOH (1993). A resource allocation formula for the NSW health system: 1993 revision. Sydney, New South Wales Department of Health: Service Development and Planning Branch.
NSW-DOH (1996). Implementation of the economic statement for health. Sydney, New South Wales Department of Health: Structural & Funding Policy Branch.
NSW-DOH (1996). NSW area strategic planning guidelines for budget allocation, achievement of benchmark cost and cross-boundary purchasing. Sydney, New South Wales Department of Health.
307 NSW-DOH (1999). Resource Distribution Formula: Technical Paper 1998/99. Sydney, New South Wales Department of Health.
NSW-DOH (2005). Resource Distribution Formula. Technical Paper: 2005 Revision. Sydney, New South Wales Department of Health.
Nyonator, F., Diamenu, S., Amedo, E., Eleeza, J. (2000). "Caring for the health of the poor: policy versus implementation. A baseline evaluation of exemption practices within health facilities in the Volta region of Ghana." Ministry of Health Unpublished paper.
Nyonator, F. and J. Kutzin (1999). "Health for some? The effects of user fees in the Volta Region of Ghana." Health Policy and Planning 14(4): 329-341.
Obuobi, A., Pappoe, M., Ofusu-Amaah, S., Boni, P. (1999). Private Health Care Provision in the Greater Accra Region of Ghana. Small Applied Research Paper 8 Partnerships for Health Reform. Bethesda MD, Abt Associates.
O'Donnell, O. and C. Propper (1991). "Equity and the distribution of U.K. national health service resources." Journal of Health Economics 10 (2): 247 - 249.
O'Donoghue, L. (1999). "Towards a culture of improving indigenous health in Australia." Australia Journal of Rural Health 7(64-69).
OECD (1994). "The reform of health care systems: A review of seventeen OECD countries." Paris: Organization for Economic Cooperation and Development.
Okello, D. O., Lubanga, R., Guwatudde, D., Sebina-Zziwa, A. (1998). "The challenge to restoring basic health care in Uganda." Social Science and Medicine 46(1): 13-21.
Okuonzi, S. A. (2004). "Learning from failed health reform in Uganda." British Medical Journal 329(7475): 1173-1175.
Okuonzi, S. A. and J. Macrae (1995). "Whose policy is it anyway? International and national influences on health policy development in Uganda." Health Policy and Planning 10: 122–132.
Oliver, A. (2003). Equity in Health and Health Care. London, The Nuffield Trust. Oliver, A., Healey, A., Le Grand, J. (2002). "Addressing health inequalities." Lancet 360(9332): 565-567.
Olsen, E. O. and D. L. Rodgers (1991). "The welfare economics of equal access." Journal of Public Economics 45: 91-106.
PAHO (1998). Health services financing and private sector participation: Developing a model incorporating American and Caribbean experiences. Technical Report Series No 65. Washington, DC, Pan American Health Organization
PAHO (1999). Annual Report of the Director 1998: Information for Health. Washington, DC, Pan American Health Organization.
Pain, C., Frankovitch, F., Cook, G. (1996). "Setting targets for CABG survey in North Western Region." Journal of Public Health Medicine 18: 449-456.
308 Palmer, N., Mueller, D., Gilson, L., Mills, A., Haines, A. (2005). "Health financing and access to services - Reply." Lancet 365(9459): 570-570.
Palmer, S. R. (1978). The use of mortality data in resource allocation. In. Morbidity and its relationship to resource allocation. J. Brotherston (ed). Cardiff, Welsh Office: 25-39.
Parkin, D., McGuire, A., Yule, B. (1987). "Aggregate health care expenditures and national income: Is health care a luxury good?" Journal of Health Economics 6: 109-127.
Paton, C. (1985). The policy of resource allocation and its ramifications: A review. London, Nuffield Provincial Hospitals Trust.
Patton, M. Q. (2002). Qualitative research and evaluation methods (3 edition), Sage Publications.
Peacock, S. and L. Segal (2000). "Capitation funding in Australia: Imperatives and impediments." Health Care Management Science 3(2): 77-88.
Pearson, M. (2002). Allocation public resources for health: developing pro-poor approaches. London, Department for International Development (DFID).
Pellegrino, E. D., Ed. (1999). The goals and ends of medicine: How are they to be defined? In: M.J. Hanson and D. Callahan (Eds.), The goals of medicine: The forgotten issue in health care reform. Washington, DC, Georgetown University Press.
Peter, F. (2001). "Health equity and social justice." Journal of Applied Philosophy 18(2).
Petticrew, M., Whitehead, M., Macintyre, S. J., Graham, H., Egan, M. (2004). "Evidence for public health policy on inequalities: 1: The reality according to policymakers." Journal of Epidemiology and Community Health 58(10): 811-816.
Phillips, D. C. and N. C. Burbules (2000). Post-positivism and educational research. Lanham, MD, Rowman and Littlefield.
Phillips, D. L. (1979). Equality, Justice and Rectification: An Exploration in Normative Sociology. London, Academic Press.
Polikowski, M. and B. Santos-Eggimann (2002) How comprehensive are the basic packages of health services? An international comparison of six health insurance systems. Journal of Health Services Research and Policy. 7 (3):133-142.
Powell, M. and M. Exworthy (2003). Equal Access to Health Care and the British National Health Service. Policy Studies 24(1).
Prud'homme, R. (1994). On dangers of decentralisation. Washington, DC, The World Bank.
Prud'homme, R. (1995). "The dangers of decentralization." The World Bank Research Observer 10: 201-20.
Puffer, F. (1986). "Access to primary care: A comparison of US and the UK." Journal of Social Policy 15: 293 - 313.
Raghupathy, S. (1996). "Education and the use of maternal health care in Thailand." Social Science and Medicine 43(4): 459-471.
309 Ranis, G. and F. Stewart (1994). "Decentralization in Indonesia." Bulletin of Indonesian Economic Studies 30(3): 41-72.
Rawls, J. (1972). A Theory of Justice. Oxford, Oxford University Press.
Rawls, J. (1982). Social unity and primary goods. In: Utilitarianism and beyond. A. Sen and B. Williams. Cambridge, Cambridge University Press: 159-185.
Rawls, J. (1993). Political Liberalism. New York, Columbia University Press.
Rice, N. and P. Smith (2001). "Ethics and geographical equity in health care." Journal of Medical Ethics 27: 256-261.
Rice, T. (2001). "Individual autonomy and state involvement in health care." Journal of Medical Ethics 27: 240-244.
Richards, T. (1996). "European health policy: must redefine its raison d'etre." British Medical Journal 312(7047): 1622-1623.
Ring, I. T. and N. Brown (2002). "Indigenous health: chronically inadequate responses to damning statistics." Medical Journal of Australia 177: 629-631.
Ringen, S (1988). "Direct and indirect measures of poverty." Journal of Social Policy 17: 351– 66.
Roemer, J. E. (1996). Theories of distributive justice. Harvard University Press.
Rondinelli, D. (1981). "Government decentralization in comparative perspective: theory and practice in developing countries." International review of administrative science 47: 133-145.
Rondinelli, D.A. (1983). "Implementing decentralization programmes in Asia: a comparative analysis." Public administration and development 3(3): 181-207.
Rondinelli, D.A., Nellis, J.R., Cheema, G.S. (1983). "Decentralization in Developing Countries: A Review of Recent Experience." World Bank Staff Working Papers 581(World Bank, Washington, D.C.).
Rubinstein, D. (1998). "The concept of Justice in Sociology." Theory and Society 17(4): 527- 550.
Ruger, J. P. (2003). "Catastrophic health expenditure." Lancet 362(9388): 996-997.
Ruger, J. P. (2003). "Health and development." Lancet 362(9385): 678-678.
Ruger, J. P. (2004). "Ethics of the social determinants of health." Lancet 364(9439): 1092-1097.
Ruger, J. P. (2004). "Health and social justice." Lancet 364(9439): 1075-1080.
Ruger, J. P. (2004). "Millennium Development Goals for health: building human capabilities." Bulletin of the World Health Organization 82(12): 951-952.
310 Ruger, J. P. (2005). "The changing role of the World Bank in global health." American Journal of Public Health 95(1): 60-70.
Ruger, J. P., Richter, C. J., Spitznagel, E. L., Lewis, L. M. (2004). "Analysis of costs, length of stay, and utilization of emergency department services by frequent users: Implications for health policy." Academic Emergency Medicine 11(12): 1311-1317.
Sabbagh, C. (2001). "A taxonomy of normative and empirical oriented theories of distributive justice." Social Justice Research 4(3).
Sackey, H. A. (2005). "Poverty in Ghana from an Assets based Perspective: An Application of Probit Technique." African Development Review 17(1): 41-69.
Sahn, D. and R. Bernier (1995). "Have structural adjustments led to health sector reform in Africa?" Health Policy 32: 193-214.
Salmond, C. and P. N. Z. Crampton (1999). Deprivation and health. In: Social inequalities in health. Howden-Chapman P, Tobias M, (Eds). Wellington: Ministry of Health, 2000:9– 63. (http://www.moh.govt.nz).
Saltman, R., Bankauskaite, V., Vrangbaek, K. (2003). "Decentralization in health care: strategies and outcomes." Madrid: EOHCS.
Saltman, R. and J. Figueras (1997). "European health care reform: analysis of current strategies." Copenhagen; WHO Regional office for Europe.
Saltman, R. and C. Von Otter (1992). "Planned Markets and Public Competition." Milton Keynes Open University Press.
Sassi, F., Archard, L., Le Grand, J. (2001). "Equity and the economic evaluation of healthcare." Health Technology Assessment (Winchester, England) 5(3): 1-138.
Sassi, F., Le Grand, J., Archard, L. (2001). "Equity versus efficiency: a dilemma for the NHS - If the NHS is serious about equity it must offer guidance when principles conflict." British Medical Journal 323(7316): 762-763.
Schieber, G. and A. Maeda (1999). "Health care financing and delivery in developing countries." Health Affairs 18(3).
Schneider, H. and L. Gilson (1999). "Small fish in a big pond? External aid and the health sector in South Africa." Health Policy and Planning 14(3): 264-272.
Segall, M. (1983). "Planning and politics of resource allocation for primary health care: promotion of meaningful national policy." Social Science and Medicine 17(24): 1947- 60.
Segall, M. (2000). "From cooperation to competition in national health systems - and back?: Impact on professional ethics and quality of care." The International Journal of Health Planning and Management 15(1): 61-79.
Segall, M. (2003). "District health systems in a neoliberal world: a review of five key policy areas." The International Journal of Health Planning and Management 18(S1): S5-S26.
Sen, A. (1990). "Justice: Means versus freedom." Philosophy & Public Affairs 19(2).
311 Sen, A. (1995). Inequality Re-examined (Oxford Scholarship Online). http://www.oxfordscholarship.com/
Sen, A. (1995). "Rationality and Social Choice." American Economic Review 85(1): 1-24.
Sen, A. (1998). "Social choice and justice." Trimestre Economico 65(260): 479-504.
Sen, A. (1999). "Economics and health." Lancet 354: 20-20.
Sen, A. (1999). "Health in development." Bulletin of the World Health Organization 77(8): 619- 623.
Sen, A. (2002). "Health: perception versus observation - Self reported morbidity has severe limitations and can be extremely misleading." British Medical Journal 324(7342): 860- 861.
Sen, A. (2002). "Why health equity?" Health Economics 11(8): 659-666.
Sen, A. (2004). "Capabilities, lists, and public reason: Continuing the conversation." Feminist Economics 10(3): 77-80.
Sen, A. and J. Foster (1973). On Economic Inequality (Oxford Scholarship Online), Oxford University Press.
Sen, A. K. (1985). Commodities and capabilities. Amsterdam: North Holland.
Senah, K. A. (1997). Money be man: The popularity of medicines in rural a rural Ghanaian community. Amsterdam, University of Amsterdam.
Senn, S. and H. Shaw (1978). "Resource allocation. Some problems in applying the national formula to area and district revenue allocations." Journal of Epidemiology and Community Health 32(1): 22-27.
Shaw, R. and P. Smith (2001). Allocating health care resources to reduce health inequalities In: Health care UK. Appelby J., Harrison A. (eds). London; King’s Fund: 7–13.
Shaw, W. H. (1999). Contemporary Ethics: Taking accounts of utilitarianism, Blackwell Publisher.
Sheldon, T. A. (1997). "Formula fever: allocating resources in the NHS." British Medical Journal 315: 964.
Sheldon, T. A. and P. C. Smith (2000). "Equity in the allocation of health care resources." Health Economics 9(7): 571-574.
Sidgwick, H. and M. G. Singer (2000). Essays on Ethics and Method (Oxford Scholarship Online). Oxford University Press.
Sikosana, P. L. N., Dlamini, Q.Q.D., Issakov, A. (1997). "Health sector reform in sub-Saharan Africa - a review of experiences, information gaps and research needs." World Health Organization Document (WHO/ARA/cc/97.2.)
312 Silverman, J. (1990). "Public Sector Decentralization: Economic Policy Reform and Sector Investment Programs." Africa Technical Department Division Study Paper, World Bank (1).
Smith, B. C. (1985). Decentralization: the territorial dimension of the state. London: Allen and Unwin.
Smith, P. C. (2002). "Performance management in British health care: will it deliver?" Health Affairs 21(3): 103-115.
Snaith, A. (1978). "Sub-regional resource allocations in the National Health Service." Journal of Epidemiology and Community Health 32(1): 16-21.
Standing, H. (1997). "Gender and equity in health sector reform programmes: A review." Health Policy and Planning 12(1): 1-18.
Standing, H. (2002). "An Overview of Changing Agendas in Health Sector Reforms." Reproductive Health Matters 10(20): 19-28.
Steckler, A., McLeroy, K.R., Goodman, R.M., Bird, S.T., McCormick, L (1992). "Towards integrating qualitative and quantitative methods: An introduction." Health Education Quarterly 19(1): 1-8.
Stefanini, A. (1999). "Editorial: Ethics in health care priority-setting: a north-south double standard?" Tropical Medicine and International Health 4(11): 709-712.
Streefland, P. (2005). "Public health care under pressure in sub-Saharan Africa." Health Policy 71(3).
Sundquist, J. (2001). "Migration, equality and access to health care services." Journal of Epidemiology and Community Health 55(10): 691-692.
Tabachnick, B. G. and L. S. Fidell (1996). "Using multivariate statistics (3rd ed)." Harper Collins (New York).
Tain, L. (2003). "Health inequality and users' risk-taking: a longitudinal analysis in a French reproductive technology centre." Social Science and Medicine 57(11): 2115-2125.
Tang, S. and G. Bloom (2000). "Decentralizing rural health services: a case study in China." International Journal of Health Planning and Management 15(3): 189 - 200.
Tashakkori, A. and C. Teddlie (1998). Mixed methodology: combining qualitative and quantitative approaches, Thousand Oaks, CA: Sage.
Taylor, T. (1998). "The natural life of policy indices: geographical problem areas in the US and UK." Social Science and Medicine 47(71325).
Townsend, P. (1987). "Deprivation." Journal of Social Policy 16 (125-46).
Townsend, P. (1991). "The Black Report - Reply." Journal of Public Health Medicine 13(3): 231-232.
Townsend, P., Phillimore, P., Beattie, A. (1998). A Health and deprivation. Inequality and the North. London, Croom Helm.
313 Transparency International (2001). Global Corruption Report 2001. Berlin - Germany, Transparency International (http://www.globalcorruptionreport.org).
Tsikata, D. and W. Seini (2004). "Identities, Inequalities and Conflicts in Ghana." CRISE Working Paper 5
Ugalde, A. and J. T. Jackson (1995). "The World Bank and international health policy: A critical review." Journal of International Development 7(3): 525-541.
UNDP (1997). Human Development Report 1997. New York, United Nations Development Programme.
UNDP (2004). Human Development Report 2004. New York: United Nations Development Programme, 2004.
United Nations (2001b). World Population Prospects, The 2000 Revision Vol. I: Comprehensive Tables. New York, United Nations: Department of Economic and Social Affairs, Population Division.
United Nations (2002). "Millennium development goals." Available at: http://www.un.org/millenniumgoals.
Upton, H. (2003). "Ethical theories and practical problems." Nursing Philosophy 4: 170-172.
Valtonen, H. and J. Laine (2003). "Study on a resource allocation formula for social services in Finland." International Journal of Social Welfare 12(4): 339-346.
Van Doorslaer, E., Wagstaff, A., Bleichrodt, H., et al. (1997). "Income-related inequalities in health: Some international comparisons." Journal of Health Economics 16: 93-112.
Van Doorslaer, E., Wagstaff, A., Calonge, S., et al. (1992). "Equity in the Delivery of Health Care: Some International Comparisons." Journal of Health Economics 11(4): 389-411.
Van Doorslaer, E., Wagstaff, A., Van der Burg, H., et al. (1998). "Equity in the delivery of health care: Further international comparisons." Erasmus University, Rotterdam.
Varian, H. R. (1975). "Distributive justice, welfare economics and the theory of fairness." Philosophy & Public Affairs 4(3): 223-247.
Veatch, R. M. (1978). "Contemporary Issues in Bioethics - Beauchamp,T.L, Walters,L." Hastings Center Report 8(3): 14-16.
Veatch, R. M. (1979). "Just Social Institutions and the Right to Health-Care." Journal of Medicine and Philosophy 4(2): 170-173.
Veatch, R. M. (1981). Theory of Medical Ethics. New York, Basic Book Inc.
Veatch, R. M. (1982). Ethical dimensions of distribution of health care. Economics of health care. J. Van der Gaag, W. B. Neenan and T. Tsukahara Jr, Praeger Publishers.
Veatch, R. M. (1994). "Health-Care Rationing through Global Budgeting - the Ethical Choices." Journal of Clinical Ethics 5(4): 291-296.
314 Veatch, R. M. (1997). "Will equity be compatible with efficiency?" World Health Forum 18(2): 153-157.
Veatch, R. M. (1998). "The place of care in ethical theory." Journal of Medicine and Philosophy 23(2): 210-224.
Veatch, R. M. (1999). "The foundations of bioethics." Bioethics 13(3-4): 206-217.
Vogel, R. J. (1988). Cost recovery in the health care sector: selected country studies in West Africa. Washington, D.C, World Bank (World Bank Technical Paper No. 82).
Vogel, R. J. (1991). "Cost Recovery in the Health-Care Sector in Sub-Saharan Africa." International Journal of Health Planning and Management 6: 167–91.
Vogel, R. J. (1993). Financing Health Care in Sub-Saharan Africa: A Policy Study. Silver Spring, Maryland, Basic Health Management.
Wagstaff, A. (1991). "QALYs and the equity-efficiency trade-off." Journal of Health Economics 10(1): 21-41.
Wagstaff, A. (2000). "Socio-economic inequalities in child mortality: Comparisons across nine developing countries." Bulletin of the World Health Organization 78(1).
Wagstaff, A., Paci, P., van Doorslaer, E. (1991a). "On the measurement of inequalities in health". Social Science and Medicine 33(5): 545-57.
Wagstaff, A. and E. Van Doorslaer (1993). Equity in the finance and delivery of healthcare: concepts and definitions. OUP, Oxford.
Wagstaff, A. and E. van Doorslaer (1994). "Measuring inequalities in health in the presence of multiple-category morbidity indicators." Health Economics 50(3): 281-291.
Wagstaff, A. and E. van Doorslaer (1998). Equity in health care finance and delivery. North Holland Handbook of Health Economics, (eds.). A. J. Culyer and J. P. Newhouse.
Wagstaff, A., van Doorslaer, E., Paci, P (1989). "Equity in the Finance and Delivery of Health Care: Some Tentative Cross-country Comparisons." Oxford Review of Economic Policy 5(1): 89-112.
Wagstaff, A., van Doorslaer, E., Paci, P. (1991b). "On the measurement of horizontal inequity in the delivery of health care." Journal of Health Economics 10(2): 169-205.
Wakerman, J. (2004). "Defining remote health." Australian Journal of Rural Health 12: 210– 214.
Walsh, A. (1990). Statistics for the social sciences with computer applications. New York, Harper and Row, Publishers.
Walt, G. (1994). "An Introduction to Process and Power." Health Policy, London, Zed Books.
Walt, G. and L. Gilson (1994). "Reforming the health sector in developing countries: the central role of policy analysis." Health Policy and Planning 9(4): 353-70.
315 Walt, G., Pavignani, E., Gilson, L., Buse, K. (1999). "Health sector development: from aid coordination to resource management." Health Policy and Planning 14(3): 207-218.
Westphal, J. e. (1996). Justice. Indianapolis/Cambridge, Hackett Publishing Company Inc.
Whitehead, M. (1989). "Health Inequalities in Britain and Sweden." Lancet 2(8658): 331-331.
Whitehead, M. (1992). "The Concepts and Principles of Equity and Health." International Journal of Health Services 22(3): 429-445.
Whitehead, M. (1994). "Who Cares About Equity in the NHS." British Medical Journal 308(6939): 1284-1287.
Whitehead, M. (1996). "Working towards social justice in health." Journal of the Royal Society of Health 116(4): 256-263.
Whitehead, M. (1985). The concepts and principles of equity and health. World Health Organization. Copenhagen.
Whitehead, M. and G. Dahlgren (1991). "What Can Be Done About Inequalities in Health." Lancet 338(8774): 1059-1063.
Whitehead, M., Dahlgren, G., Evans, T. (2001). "Equity and health sector reforms: can low- income countries escape the medical poverty trap?" Lancet 358(9284): 833-836.
Whitehead, M. and F. Diderichsen (2002). "Inequality of international public health - Reply." Lancet 359(9302): 259-259.
Whitehead, M. and F. Drever (1999). "Short version 1 - Narrowing social inequalities in health? Analysis of trends in mortality among babies of lone mothers." British Medical Journal 318(7188): 908-912.
Whitehead, M. and F. Drever (1999). "Short version 2 - Narrowing social inequalities in health? Analysis of trends in mortality among babies of lone mothers." British Medical Journal 318(7188): 912-914.
Whitehead, M., Judge, K., Hunter, D. J., Maxwell, R., Scheuer, M. A. (1993). "Tackling Inequalities in Health - the Australian Experience." British Medical Journal 306(6880): 783-787.
Whitehead, M., Petticrew, M., Graham, H., Macintyre, S. J., Bambra, C., Egan, M. (2004). "Evidence for public health policy on inequalities: 2: Assembling the evidence jigsaw." Journal of Epidemiology and Community Health 58(10): 817-821.
Whitehead, M., Scott-Samuel, A., Dahlgren, G. (1998). "Setting targets to address inequalities in health." Lancet 351(9111): 1279-1282.
Wilkinson, R. G. (1997). "Socioeconomic determinants of health: Health inequalities: relative or absolute material standards?" British Medical Journal 314(7080): 591-.
Willcox, S. (2001). "Promoting Private Health Insurance in Australia: Do Australia’s latest health insurance reforms represent a policy in search of evidence?" Health Affairs 20(3): 152-161.
316 Williams, A. (1974). Need as a demand concept (with special reference to health). In. Economic Policies and Social Goals. A. Culyer (ed.). London, Martin Robertson.
Williams, A. (1988). "Priority setting in public and private health care: A guide through the ideological jungle." Journal of Health Economics 7(2): 173-183.
Williams, A. (1997). "Intergenerational equity: an exploration of the 'fair innings' argument." Health Economics 6: 117-32.
Williams, A. (1997). "The rationing debate: Rationing health care by age: The case for." British Medical Journal 314(7083): 820-.
Williams, A. H., Ed. (1993). Equity in health care: The role of ideology. In Van Doorslaer, E., Wagstaff, A. and Rutten, F. (eds.) Equity in the finance and delivery of health care. Oxford, Oxford University Press.
Williams-Jones, B. and M. M. Burgess (2004). "Social Contract Theory and Just Decision Making: Lessons from Genetic Testing for the BRCA Mutations." Kennedy Institute of Ethics Journal 14(2): 115.
Wilson, J. Q. (1993). The Moral Sense. New York, Free Press.
Wiseman, V. and S. Jan (2000). "Resource Allocation within Australian Indigenous Communities: A Program for Implementing Vertical Equity." 8 3.
Wiseman, V., Mooney, G., Berry, G., Tang, K. C. (2003). "Involving the general public in priority setting: experiences from Australia." Social Science and Medicine. 56(5): 1001- 12.
Wistow, G. (1997). "Decentralisation from acute to home care settings in England." Health Policy 41: S91-S108.
World Bank (1995). Ghana Health sector program support II: Project document. Washington DC.
World Bank (2000). World development report 2000/2001: attacking poverty. Washington: World Bank.
World Bank (2002). A Case for Aid: Building a Consensus for Development Assistance. Washington DC, World Bank.
World Bank (2004). World development report 2004: making services work for poor people. Washington: World Bank, 2003.
World Bank, W. (1993). "World Development Report 1993: investing in health." New York: Oxford University Press.
World Bank, W. (1993a). "Better health in Africa." Washington: World Bank, Africa Technical Department: 1-218.
World Health Organization (2000). World Health Report. Geneva: WHO, 2000.
World Health Organization (2002). World Health Report. Geneva: WHO, 2002.
317 World Health Organization (1997). European health care reform : analysis of current strategies. Copenhagen, World Health Organization.
Yach, D. and D. Harrison (1994). "Inequalities in health: Determinants and status in South Africa." Kluwer Academic Publishers South Africa.
Yin, R. K. (1999). "Enhancing the quality of case studies in health services research." Health Services Research 34(5 Pt 2): 1209-24.
Yin, R. K. (2003). Applications of Case Study Research. Thousand Oaks, Sage Publications.
Yin, R. K. (2003). Case Study Research Design and Methods. Thousand Oaks, Sage Publications.
Zere, E. and D. McIntyre (2003). "Equity in self-reported adult illness and use of health service in South Africa: Inter-temporal, comparison." Journal of Health Population and Nutrition 21(3): 205-215.
Zou, K. H., Tuncali, K., Silverman, S.G (2003). "Correlation and simple linear regression." Radiology 227(3): 617-622.
318