Report No. 8382-UNI Federal Republic of Nigeria Health Care Cost, Financing and Utilization (In Two Volumes) Volume II: Annexes October 18, 1991 Western Africa Department Population and Human Resources Operations Division FOR OFFICIAL ISE ONLY Z 0~~~~~~~~~~~~~ , .-hn.A e World Bat a . o n dnthe h ' thir o ti. c m o is ; -be :~coe Wiu Wol .an authorzation CURRENCY EQUIVALENTS US$1.00 = N 8.74 (May 1991) ABBREVIATIONS AND ACRONYMS AHCS = African Health Consultancy Services AVSC Association for Voluntary Surgical Contraception BHSS = Basic Health Services Scheme CEDPA Center for Education Development and Population Activities CHW Community Health Workers CIDA = Canadian International Development Agency DAF Directorate of Adminstration and Finance CPI Consumer Price Index DHPRS = Department of National Health Planning, Research and Statistics DNHP = Directorate of National Health Pianning DPT Diphtheria. Pertussis and Tetanus DRF Drug Revolving Fund ED Essential Drug EDF = European Development Fund EDP = Essential Drugs Program EEC European Economic Community EPI Expanded Program on Immunization FHI Family Health International FMFED Federal Ministry of Finance and Economic Development FMOH = Federal Ministry of Health FP = Family Planning FPIA = Family Planning International Assistance HMD = Hospital Management Board IDRC = International Development Research Center JHUPCS = John Hopkins University Population Communication Setvices JHUPIE = John Hopkins University Program for International Education LDC = Less Developed Country LGA = Local Government Authority MCH Maternal and Child Health MFED = Ministry of Finance and Economic Development MOH Ministry of Health NEDP - Nigerian Essential Drugs Program NGO = Non-Governmental Organization NPB National Population Bureau NPC National Population Commission ODA = Overseas Development Administration PHC = Primary Health Care PHCCU = Primary Health Care Coordinating Unit RMN = World Bank Resident Mission, Nigeria SHMB = State Health Management Board SIDA Swedish International Development Authority SMOH = State Ministry of Health TFR = Total Fertility Rate UNDP = United Nations Development Program UNFPA = United Nations Fund for Population Activities UNICEF = United Nations Children's Fund USAID = U.S. Agency for International Development WHO = World Health Organization FISCAL YEAR January 1 - December 31 FOR OFFICIAL USE ONLY UIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION ANNEXES A. Federal, States and Local Government: Health Care Responsibility B. Ogun State Ministry of Health, State Health Plan: Table of Contents C. Sources of Data in Annex F D. Data Sources, Availability and Quality E. Inferring Actual State Government Expenditures from Estimates F. Background Data: Federal, Selected ftates and LGAs G. Technical Appendix for Facility Analysis H. Form A: Inventory of Health Facilities in Ogun States Form B: Information from Selected Health Facilities in Ogun State I. Financial Information System: Basic Data Requirements J. Statistical Analysis of Efficiency in Public and Private Facilities: The Frontier Production Function K. Form C: Household/Patient Questionnaire, Ogun State L. Information on Second Consultations M. Past Economic Research on Demand N. The Economic Model 0. Household Demand for Outpatient Services P. Calculations of Time Price Q. Income Variable R. Outline of Principles for a Cost Recovery Program S. Cost Recovery in Ghana and Modalities for Collecting New Hospital Fees T. Revenue and Cost Calculations for Chapter V, Part C Cases U. Estimates Administrative Cost V. External Assistance for Health and Population BIBLIOGRAPHY This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. Annex A Page 1 of 3 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Federal, State, and Local Government Health Care Responsibility Federal Ministry of Health The Federal Ministry of Health stall be responsible for health care services and for training institutions or other services of common usage among the States or of national concern or character. Such services and institutions include: 1. Special Hospitals (Orthopaedic, Eye, Neuropsychiatric) 2. Teaching Hospitals 3. National Laboratories 4. Communicable and Endemic Diseases Control (Designated as National Programme) 5. International Health and Quarantine 6. Regulation and Surveillance of standard training of health personnel 7. Regulation, Control and Surveillance of health care standards S. External Health Relations 9. Drugs and Poison Control 10. National Intersectoral Health Care Linkages 11. Primary Health Care Support (national planning, training, technical assistance, programme support) State Ministry of Health The State Ministry of Health shall be responsible for the health care system and training institutions as required for the well being of the people of the State. To avoid overlapping of responsibilities, the State Government shall provide: 1. Specialist care in wards of general hospitals especially for acute service 2. General hospitals care services including outpatient care 3. Training institutions especially for sub-professional level such as technologist, technicians, assistant and aide levels 4. Public health programmes 5. Intersectoral health care, linkages at State level; State public health laboratories 6. Any health programmes of particular relevance to the State 7. Primary health care support (State planning, training, financial, programming and operational support). -2- Annex A Page 2 of 3 Local Governments With the support of the State Ministry of Health, the Local Government shall be responsible for: 1. Community organized health and health related services 2. The provision and maintenance of infrastructure to provide health services 3. The involvement of the local co=munity in support of primary health care. NIGERIA Structur. of the Federal Ministry of Health (September I9n) Mtniser I National Council on Healt I I I I I Director Internal Lea Iformation Building General L~JAudit ;Unt Unit . _~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~IA Dope rtmsnt Department Of Dparopetmnt of Department Department of Department of eprtment Department of Finance and Planning of Dleease Population Food a Drug o Hospital of Primary Personnel Suppies Resa arch and Centrol and Activities Administration Services & Helth janagent eStatietices International and Contorl Training Care IOQ ¢ x 0 P Ft -4- Annex B Page 1 of 3 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION OGUN STATE MINISTRY OF HEALTH STATE HEALTH PLAN 1987 - 1991 PREPARED BY: OGUN STATE MINISTRY OF HEALTH ABEOKUTA, NIGERIA , - Annex B Page 2 of 3 TABLE OF CONTENTS I. INTRODUCTION AND FRAMEWORK OF THE PLANNING PROCESS . . . . . . . 1 II. GENERAL BACKGROUND . . . . . . . . . . . . . . . . 3 2.1 Geography/Climate . . . . . . . . . . . . . . . . . . . . . . 3 2.2 Government Administration Structure . . . . . . . . . . . . . 3 2.3 Demographic Characteristics . . . . . . . . . . . . . . . . . 4 2.4 Economic Activi,-3s . . . . . . . . . . . . . . . . . . . . . 5 2.5 Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . 8 2.6 Education . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.7 State and Local Government Revenue and Expenditures . . . 8 III. EXISTING HEALTH SITUATION ..12 3.1 Health Status of Ogun State at a Glance . . . . . . . . . . . 12 3.2 Main Health Problems .....12 3.3 Diseases and Conditions of Priority Concern . . . . . . . . . 14 3.4 Existing Health Resources and Facilities . . . . . . . . . . . 14 3.5 Health Programmes: Recent Performance and Coverage . . . . . . 20 3.6 Existing Health Services Delivery System.. . . . 24 3.7 Difficulties, Obstacles and Problems in Health Care Delivery 37 IV. OBJECTIVES, TARGETS AND STRATEGY.. . 40 4.1 Health Programmes . . . . . . . . . . . . . . . . . . . . 42 4.2 Primary Health Care ... . . . . . . . 42 4.3 Referred Services . . . . . . . . . . . . . . . . . 42 4.4 Health Staff Training . . . . . . . . . . . . . . . . . . . 42 4.5 Support Services . . . . . . 0 . . . . . . . . . . . . . . 50 4.6 Plan Budget, Anticipated Revenues and Projected Expenditures . 50 V. AFTERWORD: Plan Implementation, Impact and Unresolved Issues . . 51 -6- Annex B Page 3 of 3 ANNEXES 1. Organization Chart of' Ogun State Ministry of Health . . . . . . 60 2. Administrative Map of Ogun State . . . . . . . . . . . . . . . . 60 3. Map of Ogun State showing Health Facilities . . . . . . . . . . 60 4. Drug Revolving Account . . . . . . . . . . . . . . . . . . .61 ANNEX TABLES 1. Ogun State: Population Distribution by Region and Age, 1986 . . 71 2. Ogun State: Population Distribution by Region and Age, 1990 . . 72 3. Ogun State: Population Estimates, 1960, 1980, 1983, 1986, 1990 . 73 4. Ogun State: Age Specific Population Breakdown in One Rural and One Urban LGA . . . . . . . . . . . 74 S. Reported Income-Tax Payors (1985) in One Rural and One Urban Community . . . . 75 6. Ogun State Goverranent, Estimated and Actual Revenue 1981 - 1985 76 7. Ogun State Government, Actual Expenditures 1981 - 1984 . . . . . 77 8. Local Government Estimated and Actual Revenue, 1984 & 1985 . . . 78 9. Local Government Actual Expenditure by Cost Item, 1984 & 1985 . 79 10. Local Government Actual Expenditure by Section, 1984 & 1985 . . 80 7 - ~~~~~~~~~~I Annex C Page 1 of 6 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Sources of Data in State Health Finance Database 1/ Anambra 1. Estimates of Anambra State of Nigeria, 1981, 1983, 1985, 1986. (1981: 1980 and 1981 estimates for revenues, summary of recurrent expenditures, MOH recurrent: 198' health capital expenditure estimates; some 1980 actual for summary of expenditures) (1983: 1982 and 1983 estimates for revenues, summary of recurrent expenditures, MOH recurrent, health capital; actual revenue 1981) (1984: 1984 and 1985 estimates for revenues, summary of recurrent expenditures, MOH recurrent expenditure, health capital expenditure; 1985 estimates for HMB personnel and overhead) (1986: revenues, summary of recurrent and capital expenditures) 2. Preliminary Tables from Mr. Frank Akenzua of Shirley & Francis, Ltd. (actual for 1980 to 1986), submitted March 1987 Bauchi 3. Estimates of Bauchi State of Nigeria, 1981, 1982, 1983, 1985. 1986 (1981: 1980 and 1981 estimates for revenues, summary of recurrent expenditures, sector capital expenditures, MOH recurrent breakdown, HMB totals, health capital breakdown) (1982t 1981 and 1982 estimates for revenues, summary of recurrent expenditures, iector capital expenditures, MOH recurrent breakdown, HMB totals, health rapital breakdown) (1983: 1982 and 1983 estimates for revenues, summary of recurrent expenditures, sector capital expenditures, MOH recurrent breakdown, HlMB totals, health capital breakdown) (1984: 1984 and 1985 estimates for sector capital expenditures, summary of recurrent expenditures, MOH recurrent breakdown, health capital breakdown) (1986: 1985 and 1986 estimates for revenues, summary of recurrent and capital expenditures) 1/ Used in Annex F. Annex C Page 2 of 6 Bendel 4. Estimates of Bendel State of Nigeria, 1985, 1986 (1985: 1984 and 1985 estimates for revenues, summary of recurrent expenditures, sector capital expenditures, MOH recurrent breakdown, HMB totals, health capital breakdown; actual :983 revenue) (1986 Draft Estimates: 1985 and 1986 estimates for revenues, summary of recurrent expenditures, sector capital expenditures) 5. Letter from MFED, Bendel State to Resident Mission, Lagos (PMN), September 30, 1986 (Attachments reflecting actual subventions patd to parastatals by Bendel State Government, 1981-1985) Benue 6. Estimates of Benue State of Nigeria, 1981, 1983, 1985, 1986 (1981: 1980 and 1981 estimates for revenues, summary of recurrent expenditures, MOH recurrent breakdown, HMB total, health capital breakdown) (1983: 1982 and 1983 estimates for revenues, summary of recurrent expenditures, sector capital expen.itures, MOH recurrent bLeakdown, HMB total, health capital breakdown) (1985: 1984 and 1985 estimates for summary of recurrent expenditures, sector capital expenditures, MOH recurrent breakdown, HMB total, health capital breakdown; actual 1983 recurrent expenditure summary) (1986: 1985 and 1986 estimates for sector capital expenditures, health capital breakdown) 7. Bahal, J.K. Selected Appendices from 'Benue State Urban Development Project: State Government Finances Appraisal Report", 16 August 1983 (1981 Revenues by Benue LGAs by month; 1979-1982 estimated and actual selected revenues) Borno 8. Estimates of Borno State of Nigeria, 1983, 1985, 1986 (1983: 198- nd 1983 estimatas for revenues, incl. MOH and HMB revenues, summary of recurrent and capital expenditures, MOH recurrent breakdown, HMB recurrent breakdown, health capital expenditure) (19853 1984 and 1985 estimates for health revenr-e only, summary of recurrent and capital expenditures, MOH recurrent breakdown, HMB total, health capital expenditure, actual 1983 revenue) -9- Annex C Page 3 of 6 (1986: 1985 and 1986 estimates for revenues, summary of recurrent and capital expenditures) Cross River 9. Estimates of Cross River State of Nigeria, 1981, 1985, 1986 (1981: 1980 and 1981 estimates for revenues, incl. MOH and HMB revenues, summary of recurrent expenditures, MOH recurrenit breakdowns HMB recurrent breakdown) (19852 1984 and 1985 estimates for revenues, summary of recurrent and capital expenditures, MOH recurrent breakdown, HMB totals, health capital expenditures) (1986 Budget speech: estimates for summary of recurrent and capital expenditures) 10. Preliminary Tables from Mr. F,ank Akenzua of Shirley & Francis, Ltd. for 1980-1986. Gongola 11. Estimates of Gongola State of Nigeria, 1981, 1986 (1981: 1980 and 1981 estimates for revenues, summary of recurrent expenditures, MOH recurrent breakdown, HMB total) (1986 Budget Speech: estimates for revenues, summary of recurrent and capital expenditures, incl. MOH and HMB totals; 1985 recurrent and capital totals for MOH and HMB) Tmo 12. Estimates of Imo State of Nigeria, 1981, 1985, 1986 13. Letter of 11127184 from PS, Imo MFED to Res. Rep., RMN (estimates and actual for revenues and expenditures, 1980-1983) 14. Letter of 8/16/85 from Mr. John Innes, Lagos Office of the World Bank to Permanent Secretary, Imo MFED (1982 and 1983 actual for internal, external and total revenues; total recurrent expenditures) 15. "Imo States Finunces and Fiscal System 1981-1984 and Prospects for 1985- 19900 by E.C. Anusionwu (Actual recurrent and capital expenditures by sector 1980-84; actual internal and total, revenue 1981-84; estimated OH revenue 1985) 16. "Health Management Board, Owerri, 1985 Annual Accounts3 (Breakdown of actual expenditures and revenues, 1984, 1985, including of - 10 - Annex C Page 4 of 6 17. 'Ministry of Health Annual Statistical Bulletin 1984" (Actual for 1980-84 sources of MOH revenue; actual personnel and non- personnel recurrent and capital expenditures by MOH; summary of revenue and expenditures by HMB 1979-84; breakdown of HMB revenues from government hospitals by t,re of service, 1984; such other interesting information, financial, services, morbidity, etc.) 18. Spreadsheet: "Imo State Health Programsw prepared by Dr. S. Scheyer, World Bank, and Mr. J. Innes, RMN; contains actual for MOH capital and recurrent expenditures and HMB for 1982-84 obtained from Imo State. 19. Preliminary Tables from Mr. Frank Akenzua of Shirley & Francis, Ltd. Kaduna 20. Estimates of Kaduna State of Nigeria, 1980, 1981, 1982, 1986 Kano 21. Estimates of Kano State of Nigeria, 1981, 1985, 1986 (1985 estimates include 1984 actual for MOH recurrent expenditures) Kwara 22. Estimates of Kwara State of Nigeria, 1981, 1985, 1986 23. Tables (sources unknown) with 1981-1985 actual on internal (plus breakdown) and external revenues, deductions at sources of statutory allocations 1982-85, variance analysis of recurrent revenue, actual recurrent and capital expenditures 1981-85. 24. "Study of State Finances, A Report on [wara State of Nigeria", (Draft without tables) by Dept. of Economics, University of Ibadan. Laos 25. Estimates of Lagos State of Nigeria, 1985, 1986 (1985: includes 1983 actual for MOH and HMB recurrent and capital expenditures broken down, MOH and HMB revenues) 26. Lagos State Government Revenue and Expenditure Statement for the month of December , 1984, and Cumulative Figures January-December 1984; compiled by Lagos MOF 27. Lagos State Monthly Statistical Indicators, August 1986; prepared by Statistic Division; Plans, Programs and Budget Department (Has figures from 1969/70, which are higher than most other "actual' on total capital and recurrent expenditures internal and external revenues; also contains health service oriented data) - 11 - Annex C Page 5 of 6 28. Lagos, two tables on state and local governments finances; 1980-83 (Actual 1980-83 internal, external, and total revenues; recurrent and capital expenditures totals without sector breakdown) Niger 29. Estimates of Niger State of Nigeria, 1985, 1986 30. "Niger State: State Finances in Niger State, Performance, Prospects, and Policy Recommendations", prepared by RMN (1980-84 estimates and actual for internal and total revenues; 1981-85 estimates for recurrent expenditures by sector; number of health personnel 1985; 1981-85 estimates for capital expenditures; 1981-82 actual health and total capital) Ogun 31. Estimates of Ogun State of Nigeria, 1982, 1986 32. "Ogun State HOH Restructured Draft State Health Plan, 1987-1991"; prepared by Ogun HOH, December 1986 (Estimates and actual for 1981-85 total and internal, incl. MOH and HMB revenues; summary of 1981-84 actual recurrent and capital expenditures; 1984-85 recurrent, personnel, and capital expenditures by sector) 33. Tables: "Estimates and Actual Recurrent Expenditures of State Health Board from 1984-86" and "Estimates and Actual Capital Expenditures of State Health Board from 1984-86", from Ogun government (Breakdown for 1984-85 only since 1986 is half year only) Ondo 34. Estimates of Ondo State of Nigeria, 1981, 1982, 1985, 1986 35. "Ondo State: Analysis of Recurrent Expenditure, 1981-85". (Actual recurrent and capital expenditures 1981-85 by sector; actual 1981-85 internal and total revenues; actual 1984-85 MOH revenue) Oyo 36. Estimates of Oyo State of Nigeria, 1980, 1981, 1982, 1983, 1984, 1985, 1986 37. Letter of January 20, 1986, from Permanent Secretary of MFED to Central Bank of Nigeria regarding summary of 1986 budget estimates (1986 estimates by sector, recurrent and capital; revenue) - 12 - Annex C Page 6 of 6 38. MOH Estimates, 1987 (Includes some actual figures for 1985 and 1986 which may be incomplete) 39. "Oyo State - Financial Summary"; 1980-84 data. (1983 actual for internal and total revenues, total expenditures) 40. "Some Financial Aspects of Health Care Delivery System in Oyo State, Nigeria", by John F.E. Ohiohenuan, O.A. Erinosho, and B.F. lyun; a report prepared for the World Bank, September 1985. (1980-84 actual internal and total revenues; total health and total expenditures; MOH and SHC personnel, overhead, and capital expenditures; health sector revenue; recurrent breakdown of SHC; functionnal breakdown of health for 1984; some figures on state support of some voluntary agency hospitals) 41. Excerpts from Draft of 'Oyo State Five Year Plan". (1981-85 actual revenues, total recurrent and capital expenditures; 1981-85 estimates for MOH and SHC personnel, overhead, and capital expenditures; also includes projected cost of various PHC programs such as EPI, FP, etc.) 42. 1980-86 Actual Data from government bodies collected on E. English's mission of March 1987 to Oyo with Mr. Frank Akenzua of Shirley & Francis, Ltd., FMOH consultant; sources: MOH, State Health Council, MFED Plateau 43. Estimates of Plateau State of Nigeria, 1981, 1982, 1985, 1986 Rivers 44. Estimates of Rivers State of Nigeria, 1982, 1985, 1986 45. Preliminary Tables provided by Mr. Frank Akenzua of Shirley & Francis, Ltd., FMOH consultant Sokoto 46. Estimates of Sokoto State of Nigeria, 1980, 1981, 1984, 1985 47. 'Sokoto State Finances Study", First Draft; includes 1980-85. (Actual 1981-85 total recurrent and capital expenditures; internal and total revenues) 48. "Staff Appraisal Report, Sokoto Health Project", February 21, 1985 (1981-82 actual for recurrent and capital expenditures of MOH and USMB) - 13 - Annex D Page 1 of 2 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Data Sources, Availability and Quality 1. Federal Public Finance Data. The FMOH compiles its own actual expenditures regularly through the work of the Medical Statistics Division of the Directorate of National Health Planning and Research based on records of individual expenditures maintained by the Directorate of Administration and Finance (DAF). 2. State Public Finance Data. Within the published state budget estimates, budgeting and accounting procedures vary somewhat between states, and sometimes within states from year to year. For example, utilities for all ministries may be paid by "Administration" in some years, and by the individual ministries in other years. Where categories of expenses may have overlapping components, variations in classification may exist. For example, maintenance costs for training institutions might be considered either a maintenance cost or a cost of training. For this analysis, some data have been reclassified to allow consistency for comparability between states. 3. In many cases figures for actual expenditures by budget line item are not routinely compiled every year, either by the State Ministry of Finance, which releases funds to other ministries, or by the line ministries like the SMOH. If the actual have not previously been compiled, obtaining figures for past years on actual expenditures may involve extracting figures from "vote books", ledgers in which expenditures are recorded by hand as they are made. 'This extraction process is not only very time-consuming, but also subject to error, particularly when the figures are requested in categories of expenditure that differ from the unit's own categories. 4. Actual figures on state finances have been obtained from a variety of sources, although they can be assumed to have all originated either from the state Ministry of Finance and Economic Planning and/or directly from the SMOH or SHMB. The sources of actual figures include state government financial statements, documents prepared by economist consultants, reports prepared by the World Bank Resident Mission in Lagos, tables collected directly from government ministries and departments, and state health plans. 5. Where more than one source for 'actual' expenditures exist for the same state, the figures often differ suboantially. This presumably results from the problems in going back into the existing raw records to extract figures on actual expenditures in previous years. A few financial variables, such as total capital expenditures, are fairly consistent, while some variables have been seen to vary by up to a factor of twenty. When significantly different figures for the same variable from different sources are found, the analysis uses the seemingly most reliable figures. These figures were selected on a case - 14 - Annex D Page 2 of 2 by case basis considering the apparent quality of the source, comparison with other actual for the same year from other sources, and reasonableness assessed by available breakdowns of aggregate figures and comparison with the same state in other years. 6. Local Government Public Finance Data. The LGA financial data also comes from a variety of sources. The best source is annual financial statements, which have been prepared in a consistent format in some places since at least 1980. These annual financial statements include both the approved budget estimates and the actual amounts collected or spent for very specific items. They can be considered to be reasonably accurate because of both the timeliness and detcil of their preparation. If they are available, the main potential source of error becomes lack of applying consistent classification rules when extracting the data. Another source of LGA data is the state Ministries of Local Government, whose information may be less reliable and complete. Other sources include documents prepared by FMOH economic consultants. - 15 - Annex E Page 1 of 5 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Inferring Actual Health Expenditures in Nigeria 1. Chapter III of this report is based primarily on actual health expenditures rather than on approved budget estimates. However, only limited data on actual health expenditures are available. Thus, to gain an appropriate overall picture of aggregate government financing for health care in Nigeria, this annex uses budget estimates to infer actual expenditures. First, the average expenditure performance of those states for which both estimated and actual expenditures are known is analyzed, deriving average budget 'realization percentages." Second, these state- level realization percentages are combined with available information on actual expenditure at the Zederal and local government levels in order to estimate the magnitude of total actual expenditure. Finally, the estimated shares of actual health expenditure by level of government are calculated. 2. State 3ealth Expenditures. Table 1 shows the percentage of the approved budget estimates which is realized as actual expenditure in selected states over the period 1980-85.1 Realization percentages for state recurrent budgets dipped in 1981 to about 62Z of the approved budget estimates and then rose in 1983 through 1985. Total capital expenditures suffered even more, dropping from 36.7X of estimates in 1980 to only 9.6Z of estimates in 1982. Clearly the capital budgets were absorbing some of the shock of unexpectedly low revenues in these years. Subsequently, in 1983 through 1985 the capital budgeting process began to adjust to the reality of fiscal austerity. However, state capital budgeting continued to be based on overly optimistic revenue projections which encouraged states to plan on approximately twice the capital budget that they were able eventually to implement. t For example, if a state budgets 90 million Naira for its total recurrent expenditures for a year and actually spends only 45 million, its realization percentage would be 50X. - 16 - Annex E Page 2 of 5 Table 1 Average State Realization of Approved Budget Estimates of Health Expenditure (Percentage) Year 1980 1981 1982 1983 1984 1985 Avrg. N Recurrent SHMB 92.5 67.9 68.5 86.5 99.1 87.4 15 SMOH 85.6 57.8 98.2 76.6 68.0 68.8 71.2 26 Total 66.5 61.7 72.2 68.6 89.5 97.9 80.4 24 Capital SHMB 17.6 1.2 2.9 6.5 4.0 6. 5 SMOH 9.7 24.0 10.4 9.6 22.8 34.0 21.3 27 Total 36.7 28,4 9.6 17.6 25.8 45.4 28.7 25 Average: 50.3 47.5 45.0 42.3 54.5 65.2 52.6 Number of States 6 23 10 18 40 25 SHBM = State Health Management Board SMOH = State Ministry of Health Total = Sum of SHMB and SMOH N = Number of observations 3. Realization percentages for the State Ministry of Health (SMOH) and the State Health Management Board (SHMB) follow similar patterns to those followed by the entire state budget, dipping lower in the early 1980s in response to revenue shortfalls and then rising in the later years as the budgeting process adjusts to fiscal realities. 4. Estimated Magnitude of Government Health Expenditure. Table 2 assembles available information on actual 1985 government health expenditures from the three levels of government. The first row of the table reports actual federal expenditures, both capital and recurrent. The second set of rows, titled State Level, reports the known aggregate of budget estimates from both the SMOHSs and the SHMBs, equal to 583.6 million Naira. However, as demonstrated in Tahle 1 above, states have typically realized between 80X and 952 of their recurrent budget estimates and only 30? to 502 of their capital budget estimates. Applying these percentages - 17 - Annex E Page 3 of 5 and adding recurrent to capital expenditure gives a total state government health expenditure in 1985 of between 394 and 489 million Naira. 5. The last part of the puzzle is the local government. Here we adopt the assumption from available information on local government expenditure that the average LGAs budget estimate equals its actual health expenditure and consists of a half million Naira of recurrent expernditure. Multiplying by 311 LGAs gives an estimated actual LGA health expenditure of 155.5 million Naira. Table 2 Approximate Contributions of Various Governmental Levels to Financing Government Health Services in 1985 (in millions of Naira) Recurrent Capital Expenditure Expenditure Total FEDERAL LEVEL Federal MOH 177.2 56.4 233.6 STATE LEVEL State MOHs and HMBs Estimates: 436.6 151.7 583.6 Realization (Lo - Hi): (802 - 951) (301 - 5OZ) Actuals (Lo - Hi): 350-415 44-74 394-489 LOCAL GOVERNMENT LGA Actualss (311 LGAs x 0.5 Na) 155.5 0.0 155.5 TOTAL: Estimates: 769.3 388.9 1128.1 Actuals (Lo - Hi): 683-748 100-130 783-878 PERCENT FMOH OF TOTAL; Estimates: 23X 16Z 20.72 Actuals: 242-26% 432-56% 27Z-30% 6. Thus, our estimate of total actual government health expenditure in Nigeria in 1985 is between 783 and 878 millien Naira. To put this aggregate health expenditure in perspective, it is useful to compare it to various components of the Nigerian National Accounts for the same year. In comparison to the N783 to N878 million spent by the government on health services delivery and support, the entire government services sector of Nigeria is estimated to produce output valued at N2,929 million for that year. Thus government health services constitute 27% to 30% of all governmen- services. Furthermore, of the productive sectors distinguished in the national accounts, forestry (at 322 Nm), non-oil mineral extraction hotels and restaurants (at 17ONm) and other services (at 336 - 18 - Annex E Page 4 of S Nm) all consume a smaller proportion of Nigerian income and produce a smaller proportion of its GDP. 7. The total health expenditure in 1985 of between 783 and 878 million Naira amounts to between 8 and 9 naira per capita per year. In 1985 exchange rates, this amounted to about eight dollars per capita, at middle of expenditures for lower middle income countries2 8. Shares of Health Expenditure by Level of Government. According to Table 2, the federal government accounts for between 27? and 302 of all government health expenditure in Nigeria, while states account for 502 to 55Z and local government contribute the balance. These expenditure shares differ significantly from the proportions of total revenue raised at the three levels of government. The effect is for the federal government, which raises the vast majority of government revenue in Nigeria, to act as a revenue collector for the states and LGAs, which in turn are responsible for the majority of health expenditure decisions. 9. Table 3 compares the relative expenditure shares to the shares of facilities and beds owned by the three levels of government. While the federal government owns only 1? of the facilities, these contain 19? of the beds in the country, because they include the federal teaching and specialty hospitals. Since these type of hospitals are expensive to operate, a division of total federal expenditures by the total number of beds gives the high figure of 20,000 naira per year per bed. Thus the FMOH share of total government health expenditures (27Z-30Z) is much larger than its share of beds (19z) would alone predict. 10. In contrast to the federal level, the share of total expenditure by the states (at 50M-55Z) is somewhat smaller than its 672 share of total beds. This is because the stat's beds primarily consist of less expensive "general hospital," clinic, and maternity beds. The fact that the states have many facilities without beds or with only a small number of beds is reflected in the much larger proportion of facilities than of beds that are state-owned. 11. The share of the LGAs is in sharp contrast to the shares at the other two levels. Although the LGAs own fully two-thirds of all public facilities, they spend only 15? to 23Z of all public health expenditure, a proportiln which closely matches their 14X share of beds. Such a relationship .uggests an inefficient allocation of resources: Since the health care zesponsibilities of the local governments are for outpatient curative and preventive primary health care, which neither require nor use beds, it follows that the share of expenditure a'llocated to this level of care should exceed its proportion of beds. 2 Comparisons are to de Ferranti, 'Paying for Health Services in Developing Countries: An Overview, 'World Bank Staff Working Paper No. 721, 1985, Table A-1, p. 99. - 19 - Annex E Page 5 of 5 Table 3 Comparison of Expenditure Shares and Facility Shares by Level of Government Thousands of Facilities z Beds z Exp as Z Naira/Bed* Federal 81 12 11,631 19? 27Z-302 20 State 2,970 332 41,364 67? 5o2-552 9 - 12 Local 6,017 66? 8,353 14? 232-15? 18 - Total 9,068 100? 61,618 100? 100X 13 - 14 * Calculated from e=penditure totals by government level from Table 2 above. - 20 - Annex F Page 1 of 25 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Background Date: Fedral, Selected States and LOAs Table 1 Federal Revenue (Dlillono of Nairo) Petroleum Non-Petroloun Petroleum Non-Petroleum Roeeipts Receipts Recelpte Recelpt. Year (Nominal) (1981 Naire) 1972/78 0.8 0.7 8.4 8.0 1978/74 1.6 0.7 6.0 2.6 1974/75 4.2 1.0 16.0 8.6 1975/76 4.6 1.2 12.B 8.2 1976/77 5.6 1.6 12.1 8.8 197/78 6.0 2.1 10.8 8.8 1978/79 4.8 2.0 6.4 8.0 1979/80 11.4 2.; 15.2 8.8 1980 12.4 2.6 16.0 8.4 1981 10.0 8.0 .0 3.0 1982 6.8 4.5 8.0 4.2 1908 7.0 4.4 5.8 8.8 1984 9.1 8.1 4.9 1.7 1985 10.9 4.2 6.6 2.1 1986 7.7 10.8 8.? 5.0 1087 18.8 8.8 6.1 8.7 1988 20.9 7.9 8.7 2.6 1989 47.0 11.2 10.6 2.5 Source: Staff estimtes. e:tobl (hd-nir2) ACI. INIENAL, EII)ERNL, MD 161AL REVEtM FIll SElED UIfiRiM SIAIES 1,0-86 I28Zm2L, NIttLIfS OF NAIRN 2482 IS23 :11tftt 8 t2 : 2983 2 2984 ' 1t985 2 986b 2 sc..:.szcntn.s:s.n.,±n:s. I.nsu.=s;asnnzs.=nssb.. 2ssfhssusz.ssszst~: . ..:n zz.:............. ......fh=..snac ...... tItISIUIMA It21Ehb8*L: lt 28. 2281226.1t12E.. . I. 2011L t2l2E1 2 .EIIE222 : I2 22 I 2EWK 1. 2l IE ;L 161RB ME EK 2IM ERNL I ICL IIL 212118881. 16 I. L :IIIilL :EaIE)sE:E. 21I.: SlGlES I iEYE I KVCiWm1 . 2latA 6VEYEh9E miVtJ& I iEvEIAA t V182 I t 2 2EYE122 129EVE122 htVE 29141 FE : REEUEM 2. 21EYt11 I 8192.202! I 011412 2 121 Ii9Et: I EYE VY U 81E12W: F2 IEVEWMA CDSS'A1VW : 21.72 216.5 2 236.2 2 26.41 171.1: 19.S 2 27.3 t 262.2: 289.5 2 .6 1. 55.71 181.3 I 53. J2 210.1 t263.5 2 38.9 236.9 275.6 2 55.5 2 2U4.9 t 20.4 IND 1 2 3.4 218.2 2 253.5: ;'t. U 614.4 204.3 1 40.5 2 1. 7 249.22 41.2 I M2.5 2 2..? I2 10.8 2 217.32 321.1 139 266 2 49 5 it sr2 1A9.9 228.6 klYERS 2 21 2 1?.f X 57.B 2 34.4 1 t6 t 540.4 2 3..8 198.1 23u.5 2 34.6 1W92.0 2 221.4 : 31.2 t 196.6 I 233.8 2 : ; : t : 1222DE. 2 I 2 2 2 : 41.2 t 2 0.40 4 24 1.: I 2 2 : : LAGS t 197.4 1336.t 1 334.2 i 182 194.2 : 375.2 : 217.2 t 131.5 2 3t4.1 t 316.9 1 122.4 2 439.3: 358.1 1 126.2 1 484.32 420.9 12 .6 2 516.5 2 0Om : I I : t.r9 1228.32: 13.2 2 8.5 114.6 2 133.1 2 25.62 t 05.3; 230.9s 2 2 14: 154.2: 228.2: : : : am2' I : : t 23.1 2 195.3 2 219 2 24.62 179.9 t 291.52 S0 I6bL.2 2196.12 4B.12 112.6 3 225.32 I 2 : : : : atid2 44.1 2U41.5 2T1.b t 48.6 t .95.4 1 342 t 51.4 t 311.4 t 368.8 2 6.8 362.: 449.1 IP 23.2 t 240.4 2 315.6 213.17 3.8.8: 54L.5. 191.S 2 .S.' 472.4 2 OUR& 2 4.22 97.9 t U2.1 2 12.2 820.9 2 93.1t 12.2 2 105 117.2 t Wu.3 2 v.t 12 2.42 12.32 111.2 2 223.52 : 2 2 50b11 1 26.22 235.42 255.62 18.5 1 216.3: 234.2 t 14.42 2;t.6: 219: 34.41 2242 258.42 el.5 203.41 3b4.7 I : 2 " I : : : : t I S 2 2 : 2 : 30.52 149.1 1 119.6 t : : : E^ t : ; : 12.2 t 201.8 114 12.3 2 49.6: 161.92 12.92 \4.2: 257.t: 1S.42 258.92 178.32 14.22 295.I 22m.32 : 6W8 2 t : : t : 16.6 134.2 t SO.6: 18.62 2h. t 135.2 29.8 12B.9 1 150.71 : : : : 11110ESa 4,2v,13.l4.25.lS29.23 7,2.S..3i2,35,' 4,4Is42,45417 in Annex C 98t8Lt5 kw; 8i & 6424u";A A 2t 21 0 Iaa ¢ £II3. ial3ilt, 1t1is., Mi 1013. IO?L IRE FIR S&EUEUD tMU. 38W33l11I AMIOS 1 11516E 1990-M IemINAL, RILUIS IOF "1RA) Itb:19 1981 I 12 t 1993 I .934 I 1995 1936 !---- --…----.----- .1.....------------.---. …---------------I.----.----- ..---- -.--- ----- - ---- -- ----------------…--.-------…: LEA Milift tilEuw i. l t IrIam EtlEtlUt iu lil U lillElI UPtiUW ioi IlIhhEfImI EUEII. IDIRt TOTAL tMlWE1A E18t11 1011 tiIlMl EIIERIIIIL 1913. IIUIERtM. 6111.it1. 70I1t 1 I R£VEiaJ RE6aU R41E6E1 I IlEVEtl R£EVlliE C101EV£U I lI£KW REVM 1001 bEIEIU REVEIIIEV liEti IEIlE t EMB3 R£E11£ EV I I EVEIMI tM fltEtG EIXPVElM t E NVEII UE IEIEt t 1 I 2 I t 1 t 10D, 019 aJ I 82.95 469.63 552.30 I 243.s. 531.39 181.37n 371.31 1,27Y.61 1,60.43 t 529.33 1,110.29 1,63.6? 1 695.51 1,5S2.43 2,227.94 1 721.60 1,917.43 2,645.03 181a04, oVe I 124.40 53.32 623.22 I t 232.91 596.12 929.53 1 272.S4 964.1 957.63 1 230.35 623.35 904.20 2 303.35 663.58 952.4 two. 010 12,659.59 1,5$5.S9 4.255.4& :1 ,46b.55 3,747.u 5,213.97 11,595.24 2,S:.S3 4,493.54 12,449.64 2,365.03 5,314.71 12,392.61 2,816.01 ,70.63 13,913.27 3,163.44 7,156.11 13,207.60 3,06b.32 6,214.62 1K013, OvD : 210.61 257.13 418.34 1 91.10 456.70 543.9O 1 156.49 623.21 734.69 1 t t 133.43 525.06 659.534 1 IUSN I, OLL l,9s.st 35.64 2,045.61 t1,915.23 3.95 1,952.3t 43.06 2,47 2,993 .40 1192.05 2,612.14 2,154.S3 1 317.29 2,143.92 3,123.20 t 317.51 3,113.49 3,571.0 1 212.24 3,121.39 3,333.6t2 V IFEW1139 t I I 337.80 709.11 1,046.91 I 1 t . 11E1 it, 641 I I . 257.24 1,316.30 1,373.9 2 364.30 1,966.13 2,31.03 I I lw17 EID, w6Ult t I t t I1,59.69 2,602.21 4,199.91 I t I t I I I I 11 FIWRES FII 1931 Im 1992 3 uw, 91 SMIN LYr. flUtISt 010, 1Ia, I 1N" LEAs In 01o stilE - _13 FliEAI StIaIUEMIS SF LeA. wOU, oeY - so ismi 3 eVMuiuI 1611 mm - onu'S tININNII 3*1* IJEBla RE11119, 11 3 t-a mm FINESt SIAI4EBI 1263 ELAST, 0643 - tLE N143N11 MSRESS REM3, 1CEl 15, AMI 1936 * E811110 R ISIU31IS. 0 (x g ; n w~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~V - 23 - Annex F U Page 4 of 25 NIGERIA: HEALTH CARE COST, FINACING AND UTILIZATION Tabl- 4 Actual Recurreat, Capital and Total Expenditures By Federal Government of Nigerle (B Illons of Nalre) Recurrent Capital Total Total Year Expenditure Expenditure Expenditure o/ Expenditure (Nominal) (1981 Noire) 1980 5.2 7.0 12.2 14.7 1981 4.9 7.9 12.8 12.8 90S2 4.2 8.7 12.9 12.0 1988 6.8 6.8 18.6 10.2 1964 6.8 8.9 10.8 6.6 1986 6.0 5.0 11.0 6.8 1986 7.7 6.5 14.2 6.9 1987 16.4 8.8 27.0 11.9 1988 22.4 8.8 84.8 10.9 1989 80.6 18.2 60.1 11.8 1990 85.0 16.6 67.2 11.9 Source: Staff estimatee. D/ Other expenditures are Included in the total. b:ennA44 (hd-nIr2) I I I 1093.1 S OCI I a 11C391, £APII3., AND I0*L EIPEMHIMES n SELECIEO BIATES II NIGERIA ts9e0-s INMINO, ni33.26 Of 410u1 I IhtO I ISltl S I992 1 3983 1 984 1 1985 t I 30101. I2 1010L3 30L I 0T301. I 10103 I I0la I 10303 IO313 2 101*3 I 0I0L 3: [1IA0 I I010L I 10303. 2 I0I03 I 2 I 1010. 2 10103 IDIA3 I StfisE 1: £913L. :ECIM911 EIP. ; 0tI0 2tECOn (Kt. I CA£l1 29E03RENIt EU. I CWlOL. 23EC3Etnt8 EIP. I CAW114 MC 11E21 EIP. I CAPITAL 29E3Nd91: flP. . : ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~ : : : : : I 2 2 I : : :3 BE I I I : 1 2 : 3 126.17 120.31 1 147.0 I I ; t I : 1 1 135.3: A.2 K 364.92 1 154.52 2 5.5 40 49.02 1120.35 2 240 3 60.35 3 s.U62 2 288.55 2 340.171 25.9B 371.2 2 39t I S : * 0 a 90. 1 251.4 2 317.5 1 143.24 t 533.9S u.11.35 129.5 I 381J.1 511.2 1 242.06 1 292.91 1 534.8791 8.3 t 346.1 2 435 3 368.11 3 359.4 t 529.317 0633 I t 1 31.31 39.83 221.34 1 36.96 1 126.93 2 263.99 t 34.92 2 142.41 111.23 2 39.36 I 131.56 1 136.72 1 I : I GmD I t I I 50.42 20.1. 251.5 1 25.1 2 229.41 255.13 29.22 221.4 1 250.6 1 22 399. 1 221.1 : : 2 3 0O I I 2 2 90.34 2 242.93 I 323.17 1 61.68 1 284.8 1 346.60 1 62.49 : 30.6.6 363.35 t 65.15 329 m 394.35 1 80.6 1 422.29 . 502.89 1 130 I t 2 I 2 10.25t 151.711 147.962 29.82 147.51t 171.371 13.13 3178.98 I 92. 332 1 3 41.61 152.672 213O 3 I t 2 17.95 1 74. 1S 322.1 2 116.05 2 196.21 t 312.29 1223.98 2 1086.02 309.9 1 53.96 1 02.31 : 226.21 2 61.45 1 368.89 2 2.34 2 0_5. I : t 122 12 213.019 329.95 2 59.05 2390.5 I 249.55 t 28.74 u61.27 1 3 Ot 19 19.0696 155.79 1 274.65 1 11.99 1 176.61 t 2".652 CSc BRVER2 9 73.1 12K 1 I t t I t 2 2 3 3 I 2 2 I 2 S933tEC1 6,IO,13.14,35,39,23,2627,29,32,35,4342,427 in Annex C smac FO 80363.0182mm"3 m t tQt 0 tn I'. CIV22 Ka 31, Cf ItlL, OA I8I1L EWWIIES 8Y SECIIED LcM 6011 SES E1 81316E1 1"09-86 288*., iIOUSUID of U1u) 1 1980 ; 1981 2 TM12 t 1901 184 1 ii2 3 196 X,-- -.. --- -... __. -.. ------- I-- ._ --------- -. --- -_.I--__- -----__-__-_-----t_-__._----.-----I I---- - ----- ----------- - --............................................ I---- -:--------- L80 ISCE!Ul Celiat 101 MCMM CNPII*1 lOl IBEC_t I W22 lOl. ICECO CWHAP0L II2. 2IEChEI CWIIoEt. 0AL IKSMI: COITAL 1012T. I8CI I EWIJA.L 19I61 I * I : I I : : eke, oro Al I 181.09 30.22 1,081.3i1 363.59 72.15 435.35 t 461.22 111.42 784.63 11,635.37 100.25 1,736.01 11,413.90 65.12 1,639.61 11,417.07 980.05 2,291.12 1 56.65 1BM, 0 t :3,517.05 121.91 S,699. 13,950.88 90.16 3,941.03 24,27. 43.3 4,911.02 :4,726.17 12,48.78 6,213.95 :4,819.92 1,005.21 5,905.1 13,936.42 2,446.22 5,382.64 I1 , OWO I 410.90 23.33 494.2 1 1 2 851.U 40.97 86.46 880.104 1.214 881.32 1 11.05 86.931 056.14 2 818.82 122.15 9S41.39 1 88818 t !8 I I 2 I 5131.28 409.65 5,538.93 1 I IJEID a1, 0fE9 I I I :4 2,21.11 11,90.37 2,489.10 4,379.55 1 IJIU EISI,0119e 1 : : : 569.34 1 : in: EAT am I I t6.3 V 1218f11 ffE 29190l AN19 91 2218 ItS 5NY. SIOI11SI er01o 6 - LEA M1U FIlICML SIAlEEIIS 1IJESt 0510, - 0t4 . LEA 03U29T2 1193F 11 1IAI12, 2 2M1 ESIIAIES 2996 IEBD 181?, OliU - LtA MIRY SS 51EMI,0SEURE 13295 11M2118, t C** 111 -1 FD1 0119 960 OF A * L LEOEE1M IN NIG3RIA, 82IN LAL. 64131131 If LABl', Of 1 11 JA2ES119; PE1IED S1 HUM. JIS 95 m m Ii 0 e1 q - 26 - Annex F Page 7 of 25 NXIERIA: HEALTH CARE COST, FINANCING AND UTILIZAT;0N Table 7 Actual Recurrent, Capital and Total Expendituree By Federal Ministry of HbItb In Nigeria, 1978/74-87 (Nominal, Millions of Naera) Personnel as Percentage Recurrent Capital Rocurrent Capital Total of Central as (N) as (N) Total Year Expenditure Expenditure Expenditure siniotry Exp. of Total of Total (U) ___7 _/7 _ ___ 0_____0_ ________- N__________ -------A------ -----A--- NA______ _______ 197B/74 20.0 MA NA NA MA NA 1974/76 29.0 NA NA NA NA NA 1976/76 69.8 41.4 111.2 62.8 87.2 100.0 1976/77 92.8 56.8 149.1 62.2 87.8 100.0 1977/78 104.1 64.4 168.6 61.8 88.2 100.0 1978/79 80.8 58.6 184.4 60.1 89.1 100.0 1979/80 94.6 79.4 174.0 64.4 45.6 100.0 1980 a 112.6 77.6 190.4 69.1 40.9 100.0 1961 158.4 142.1 295.6 71.8 61.9 48.1 100.0 1962 158.2 120.6 278.7 69.1 66.0 44.0 100.0 1983 165.8 116.6 269.9 69.9 68.8 48.2 100.0 1984 168.1 42.8 210.9 65.0 79.7 20.8 100.0 1985 177.2 46.4 222.6 68.0 79.6 20.4 100.0 1968 245.8 181.5 877.8 65.1 84.9 100.0 1987 229.0 125.6 864.6 64.6 86.4 100.0 1988 819.6 117.8 486.9 78.2 26.8 100.0 Source: Federal Ministry of Hbelth. a Aprll-Decembor only. b:annf7 (hd-nIr2) Table 8 alliU. mEEtNl, WilX, m I Io EtlPillnMs 3 StEUCIt SIAf tIlUSIS9 OF IN 12 1601U 1990- b I aim9511., BUlIWS l m1in I 199 a 1 Il .1 ? Isa 298 1t} t3 I185 2 119 [I I I 2 I ; I ma I in I nia oi ii t Iota. L I MN uIr a I TOTALa Iua tO I. ] A t I a Ma vI MN I 1 a aI MNI M roiat. FM I oni a 10161 ownIE :fl6Cufl: cauiua MN tucmltura Cella I Ms tREcumn: Eunrn. I Ma IkCiatil CAPHAL I 211 IRECORE111 CPItAL. 2II M tBfClJWffflh CPItAL t MI IBtOSIOEU CAfl2I Ma I : : a a : 2 1 : : 2 2 I I : : I I : AIWSA t 12.35: 0.52 t 12.61 18.12 u 0.89 1 19.61 1 23.1 1 0.2B 1 23.39 2 s.571 2 .31 1 3.4 2 3.82 1 0.92 1 4.741 6.67 1 4.0o 10.12 : : ; N 1: : t I : 6.51 t 0.1I .02 2 I 1 I I t amllS SURS I 1.512 4.3?1 11.60 10 I 5.55 I 15.55 1 11.81 1 2.22 1 14.09 10.68 2 0.1 2 11.16 2 t31.33 0.53 2 13.86 2 6.22 1 3.O1 2 1.23 1 2 : l10 1 6.45: 2.:1 9.25.1 4.1: 1.41 6.1 2 5.11 5.621 10.11 6.511 11 1.5s 6.12 O.il 1 6.212 4.951 0.26 1 5.21 10.5 12 12.52 22n.n t tIES 1 6.28 6.i9 1 13.21: 2 2.44 2 14.53 1 26.9) I 25.16 1 20.52 1 35.69 t 26.5 2 1.43 21.93 1 24.19 2 2.9 I 21.0 9 t I 2 2ses I : I : I I I I I I 9.9 I .22 23.16 I 5.71 1 1.38 2 7.15 I I 1 I an I I 2 3.161 1.532 4.692 3.552s 0.521 4.07 1 3.6 I .311 .91S 2.182 0.29 2.41 1 2 : 2 2 : N I : 2 1.2 1.6 1 6.8 2 1.3 I I 8 9.3 2 7.1 2 2.47 1 9.57 5.4 1 0.32 5.1 1 1 I4 M0 t 3.713 3.952 7..68 4.17 4 .5421 8.11 3 ..75 1 2.632 6 . 39 2 1 .9 2.94 7.53 5.14 1 7.85 1 13.5t 1 5.714 2.15 2 8.49 7.135 2.421 9.71: 6MM I I 1 I I I 9.J11 1.46 1 10.77: 23.92: 2.65 1 26.41 24.12 1 1.391 25.511 1 : 2 2 s5oI0 I I I : 7.61 .67 1 16.48 I .101 6.18 1 14.35 2 : I I I I : : : 2 : NO caOM EU M IENS FM mAM a 821, 1 (1196), kIVERS (11244) M MM 41193-94) UtCME Ms RECt EtDIIBES. Na C6 RIVER DATA IS OF ISTIIIE IUITU. "agi 6ECIkE FM RIVERS 111911 KlVt ISCLOES MS ktiE9. NCElI 2,6,10l5,tS,I1,169,25.2,32,35,4.0,42,45,46in Annex C S9 E Of kSh BAIA Is 1u13 '-n Table 9 Kiwi3 U ta_, wit'a, 0 1l. EtISttlES 8 SELECIED STAIE "INl t NI 89am InNIGERIA ieee-es5 411011NA., *iU.9l OF t*IF9 S 1980 1 9989 9 9992 9 53 999 164 1 It9 9 : 1 . : I I J I 135 I 9593 59I1 I a I I fell. I 9 I a I 1059 I t9 I 9 I folk I I 8 I [DIAL I m I In t 1011. I SIlIES ItECIIREI: CAFUIL I in sfltW&U l I CAPISIA 2 3 IREt3E95 £09109. MB 95ECttR12 £019129 I i thECitltN £0111. I 8 2Bil99EO 31 14 I He S _ 9 5 I 2 I I 9 9 9 lb I t t 17.15 t I t 2 tCI SB 9til 13.9I 2 J 20.99 ; ; 20.1 9 . 9.93 1 : 23.95 t ; 30.31 19 IND 1 13.2; ; 108.39 t I 20.45 i 1 201 21 1 29.7 t 9 23.391 I Kea J ; 33.41 5 t 9 34.629 9 20.229 5 9 29.15 9 t 29.1j; S I LAGIS I I 2 I I I 9 I 20.349 I 9 32.31 t I ; t IYi lt I I 12.51 S 1.9B 2 14.03 1 12.09 S 0.57 1 12.83 t 14.05 t 0.37 1 14.42 t I5 I 0.22 1 15.22 t 96.85 t 0.19 17.04 t me I ; ; 17.42 02 17.4t 10.29 02 15.21 20.5 9 0.2 9 20.1 t 19.5 t 09 19t.99 I I 1 I 1I6.91 1 0.69 1 91.5 t 21.22 1 0.52 1 2.1.4 1 22.22I 0.21 22.43 t 22.9 J 0.06 22.96 21.16 t 0.02 2 21.18 t 29.04 1 I t su51 0 I 9 : 8.01 I 9 10.U3: : : 2 ; t ; 9 ; ; ; 0814 IS 8 lmok[E FO FISCAL tEtA 9Sf 151 9 tWI. El UMWIIUIES FOB 90lIN AM SAM t5 ItItl IME On, ttDI JE 33 ES , 2.5,9O,l4.91.9B,5IS,26232,33,3S,40,42,4D in Annex C e > (D (D 14 0 Annex F Page 10 of 25 J t r > *'7 7 ~ Sa a t SA g, l 44 '- - 1* ft X _ ... _ _ ..n e ... . .. 0___..............._. .__ 3~ 3 Wi 3. 'd W * _ f 7 f e ift . is * X a . - X U43- - _ - - - - - - - - - -------- ; aI~ ft nnrtin .a -Z I, O A. _e n r 0 r .0 0~~~~~s _ __. .. .'.e, n. _ .,a.............. 1t t | g ~ 2 - a ~ s4 S7 ~ ;* ~ 7 o7 _ t _A c 0 0.0 -- - -- - - Z ~~~~~~.d. .' . r. 0. ------- ------- -------- 33 Q 110. ft0 __ _ _ _ .0l rSS f e _ t3 v j .0 - 2 1 2 - u _ e su 4 e gI- ft S . n Y O~~ gI U - . d n o o o _ r ....-.C'e.'N ^ : uU = .- a z~_eg2 ~ *Fiwu10 t4 - -- - * Sf 7ftft t .4 < z_ f z |u >- e 0 _ ji O~~~~~~~t - - -* ^ - ' -i - I a fI I!g ' - < :- 2- i R ~ ~ ~ ~ - -w - - -> - - - -F - ---- - tsI s t o C- , -t Z -- :-sE0 --- ---------- g i " O v D a $ 0 < u M s ¢-@ i § a0 toa D U_ W~ ~~ o iit 8-8 8 _ ___----------- - i _ . .... - - .,< oW o 4 a4 a * R~~~~~~~~~i r 0~ ~ ~ ~~~. _ .__ .., ..................~a C~ ~ ~_ _~e~o v . 0 u 0 a 3- .. . ...... .... --;- a Coo 8 o8g a a n' ~ e>w|S E5 n ,{xauuv Co~~~~~~~o Table 12 0A9l E9191, E ilk, WllS 1019. (Kl8 3S FOR "1N101 911(CIED LOCAL W1 5 AM IN OI68I IW90-506Il1101, 11_79 OF N1*) I ISBo ; 995 192 I 1 19B3 t 194 i 1965 :990 …------------------------- - .5----- -------..--------.. -- ---I-------… --- -- u.5 MECII CAM19± 50501 28E130 C00k 10191 IRE(JWKIIIM (519. 10291. 5UIIfWI C09l 191t 1 1 IOEtItlS(35 I0111 1 19 EEMNI 011AL 10191. 2.KEMQI 8(1.0900 ilu 10191 I~~~~~ M10, O S P21.1S 0.00 .12.1S : 101.41 0.0N 101.4 1 M9A.9I 0.00 5M.91M 04.15 2.25 501.00 493.3 0.0 48S.S3 1 445.00 119.50 50.A0 41.15 IbOai, M 2W :5,9152.55 0.00 I 912.55 t1.929.19 0.00 1,S29.19 :I ,10.01 0.00 ,1165.bS :1IYq2.07 36.02 5,82B.S0 11,t 0.70 91.34 1,716.04 9.bb SW 0*, 0D0 : I..., . 0 165. :: 1 2.4s 0.00 224.4A 1 MM3.21 0.00 2S.2 1 219.31 0.00 21t.31 290.2B 1.30 295.b; USEItl, B9lGt ISI.35 .0 ISJ.S5: SL50 : 263.96 15.02 339.61 I 271.01 14.30 2S1.17 I 49.6 : 5.0 I 351.26 125.54 416.60 swig, a6. ALL I 1 410.59 : 410.59 2 : 1 1 ; "mu11, Om1 I I 1 401.002 * (18, 06: I I I 785.25 216.0 to100I.351 7tl.19 187.12 "4.301 : Ila -e0, : I I 2 649.20 11S.85 103.05 5 552.01 109.809 601.09 S [su9 sw0ig, Mm53 : : : I 209.90 0.00 69.B0 1 453.00 98.90 552.1B: I am sII, 0es. 2 : : 311.64 7.15 319.29 1 332.11 105.59 437.10: 53la9 91I50, : :: I : 23.74 06.51 S9.3 2 308.19 64.17 S13.16 2 ilOl IA, : I : 2 515.26 06.14 591.40: 314.16 219.92 594.08 I 11I0930,9GM bi6 t : : : 572.20 61.02 039.22 452.73 140.31 5'3.10: NEll, an0 : ; : I : 395.1 62.3 414.0o t 300.11 195.30 WM2.01 : 11309, am 2 2: : .9Sf 1.86 201.02 253.15 191.05 440.20 : la9 "t( , D I I I 2 214.86 50.25 333.05: 248.58 105.2S 353.06 : 1 : 5 I :2 ai 19B1 *llD 1982 f16;RES Il1Sf 0Y0 S051 Lt3 bl 1994 f15 5 t15 s s059.7 tl5s5 of A0 f Sn 5ms, otE A mm i5)9 - 1.6*021) FIHII L£ I I15U15S UltEUI, B9ElD - P06lIlIt 091f FMFM S501 CtlIE0 RV 30I59E1 F 115, tlD, COOSRIANTS 30.010 t10 OR - 601105 SMk 2121185 0420, mm - PK not 9116$I 1591105 911ISIS X Table 13 ESlIlAILD KCaa EAI, (LI, 8 lulAS EI MIIES FOR KIaT 8 fU IED LO18L tEIUI MA 1i uiuuia 19S831- OItL, NSf IIF 91 1 1991 : 1992 1 w 1994 198 I I : . I : L.64 tRf"lt tEll 101d :"tlWU WtAL 101t :SttSM CAPITA IDIIt Ma il EVIIL IDIAL MEMR tPlaE TOMt I Bass ~ "A txvsut :aa" =.azZvz t=gzaaz,eauet z : Imz.3z DVI, 030 2 213.9t 30.02 243.93J 246.98 4t02 2M.00 t 35 60.06 M 7 ^ m. i t .37 0.80 S".371 5I .43 1t8.53 5.96 IBltDQ, 010 12,641.99 25.0U 2,66.95 :2,915.06 270.05 3,099.(6 12,9.17 .06 .23 3,1. 35.01 3,523.5 12,963.b9 255.03 3,219.11: 1IMP ,10 I I1 342.44 8.02 35.46 2 332.49 30.00 362.48 1 333.96 0.00 333.96 I Sams% AM" FINWlIA 51^1ESIEiS 10 Lt8s I- 0 Xi - 33 - Annex F Page 14 of 25 NIGERIA: HEALTH CARLi COST, FINANCING AND UTILIZATION Table 14 Population Estimates, Area, Population Density, and Percent Urban for 19 Nigerian States; 1988 Tota I Area Dona Ity Population in '000 (Persons/ X state (Million) (Sq. Km.) Sq. Km.) Urban / Anambra 0.2 19.2 823 26S Bauchi 4.2 67.7 62 20% Bendel 4.8 88.6 111 81x Benue 4.2 71.6 59 481 Born. 6.1 116.1 44 22X Cross River 5.9 28.4 208 28X Gongola 4.4 89.0 49 14% ISO 6.3 10.7 589 265 Kaduna 7.0 70.2 100 291 Kano 9.9 43.1 280 80X Kware 6.9 74.3 89 265 Log" 2.9 8.6 806 Niger 2.0 74.2 27 256 Ogun 2.8 17.4 149 85X Ondo 4.7 21.1 223 47X Oyo 8.9 86.8 242 76% Plateau 8.4 29.2 116 88% Rivers 8.0 18.1 168 271 Sokoto 7.8 94.5 88 a7x Source: Staff estimates. a/ Defined as State population residing In urban areas (defined as urban areae with populations greater than i),000 inhabitant.). b:annf14 (hd-nir2) - 34 - Annex F Page 15 of 25 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION ________________________________________.__________ Table 15 Population Estimates fo- Selected Local Government Areas In Nigeria 1981-85 (Thousands) _______________________________________________________________ LOAs 1981 1982 1983 1984 1986 Oyo, Oyo 438.68 464.12 470.10 486.66 608.78 Ibadan, Oyo 1,961.18 2,080.21 2,101.66 2,176.64 2,262.21 Iberepa, Oyo 176.32 181.49 187.88 194.49 201.88 Abeokuta, Ogun 856.26 387.77 880.71 894.11 407.98 Ijebu-Ode, Ogun 298.11 380.42 814.10 826.16 83W.60 Egbado South, Ogun 832.70 844.41 366.65 869.08 882.07 Egbado North, Ogun 200.58 207.64 214.96 222.61 280.34 IJ.bu North, Ogun 177.68 188.68 190.30 197.00 203.94 Itfo/Ota, Ogun 227.29 286.29 248.67 262.14 261.01 IJbu Reom, Ogun 240.61 248.97 267.73 266.81 276.20 Obafemi/Owode, Ogun 211.05 218.47 226.16 284.12 242.36 Odeb., Ogun 178.69 184.98 191.49 198.28 206.20 Ijebu East, Ogun 178.49 184.78 191.28 198.01 204.98 Source: LOA PHC Situation Analysis, Note: Growth rote was assumed at 3.4 per year; Population estimates for lbadan range from 2.0 to 8.5 Millions. b:anntlS (hd-nir2) -35- Annex F Page 16 of 2. a- - ---- ---- -I 'I f ~Illla $ 43 n n zD~~* |:~ t; _ _ _- --------- Is! IIY - - - - -- - - - S 'la U ""< _ _… .- _ . _ *_ *- - _ ....... I It ~~ *u = I I I-eea - I ".. I ~ .i-gg I - ~~~~~~r.. £ . | l||§Si! ~~~~I | Table 17 ss;iam9 Fuau *1 inu 'tn3 3fK; , i 24 113 a s^~|ii~ g ^'3 l;3ilii503j_|3lt3 i 0 |~~~ -40- Annex F Page 21 of 25 i _ j | X §~~~R a-g | ~ ~ ~ s - - i A ...... as cS flag 2i Ot -! a '- ' ,- s ~a a- Pa 2 Ii~~~~~~i li~~~ a * -- .,, a .i S S aa S S - -4 ~~~~*In -A C"~~~~~_ l 4)si l '-4 Table 22 /Cl tUt 6 OF FER0 06 I SAl EClED STalE NINISIBIES OF HEllIH 3iD NIALIN AE6EIIKI OMDS IN NIGERIA 1981_- 3993 398& 1993 3994 3915 916 in. . VA I am m I s I MNa mm INM mS t IPE5S9181E PISOEL9M IME tL FEkSINIEL :KSUI L PWSEtL I EtML tSiML IKEh L KFEt IFIEtL PFEUSIEI I nass.fltsfl fl:. aswssnslfUffastlnis fln fl== ssnBSasnftlIcaflflflnStSsfDlfl8sRflsri sun.:sf.fas..s..3 m..nnc% a g.f,.snts.smastswtsrssc :sfhwsG LtMU I I t I',29 5,814 1 1,1t 6 I 131 I 2,019 5,10 1 2,359 6,450 2475 O I,60 2,492 1,300 1,795 5,60 1 2,348 6,040 KM : ' t 2,514 I I I SIlSo 21. 6, I; in Annex C 02I -42- Annex F Page 23 of 25 i a n~~~~~~ aa E-4 ~ ~ ~ i if~~~~~f w a - ~ ~ a if a aa X 2 i g i - Iw - C-- - 4 *-__ P - a = a IIS - 4. 8 ~ a a ' dl *w_ * -a3 -i 1, - 1 E- ~ ~ ~ ~ ~ ~ ~ ~ ~ 44 3 a.-.4 S~~~~~~~~~- a - - , a _ __...... *________ .................-...................... ^ B g 9 i I '. i~ ~~ 3II IIa - 43 - Annex F Page 24 of 25 Table 24 ACTUAL HEALTH EXPENDIIURES FOR ENVIROHENTAL SANITATION, ADMINISTRATION, PREVENTIVE, AND CURATIVE SERVICES FOR THREE LOCAL SOVERNNENT AREAS IN OYO STATE, NIGERIA, 1981-95 (NOMINAL, THOUSANDS OF NAIRA) : 1981 1982 1963 1984 1995 DYD LOA ADMINSTRATION : 26.35 0.00 25.97 39.53 31.66 ENV. SANITATION: 0.00 0.00 111.16 14.99 0.00 PREVENTIVE I 0.00 46.73 93.17 143.31 129.00 CURATIVE 1 75.12 144.67 189.03 295.53 284.33 IDARAPA LOA I ADMINSTRATION : 0.00 0.00 0.00 END. SANITATIONt 71.59 96.41 90.96 PREVENTIYE 1 0.00 0.00 CURATIYE t 150.32 193.45 202.57 IBADAN LGA : ADPINSTRATION 9 97.29 97.59 89.92 78.49 29.95 ENV. SANITATION: 659.40 641.72 612.47 590.44 539.99 PREVENTIVE 1 718.55 797.75 695.27 656.01 552.63 CURATIVE 1 437.31 402.13 369.03 4U1.73 556.24 SOURE: LEA ANNUAL FINANCIAL STATEMENTS 44 - Annex F I-age 25 of 25 Table 25 ESTIMATEO HEALTH EXPENDITURES FOR ADMINISRATION, ENVIRONENTA SANITATION, PREYENTIVE, AND CURATIVE SERVICES FOR THREE LOCAL GOVERNMENT AREAS IN OYD STATE, NIGERIA, 1981-85 (NOMINAL, THOUSANDS OF NAIRA) : 1981 1982 1983 1984 1985 OVO LGA ADNINSTRATION1 38.91 0.00 42.94 45.08 30.70 ENV. SANITATION; 51.59 0.40 126.63 0.00 5.69 PREVENTIVE : 0.00 57.95 111.20 166.60 131.90 CURATIVE t 123.41 188.63 454.64 386.69 369.15 IBARAPA L6A : ADNINSTRATION t 0.00 0.00 0.55 ENV. SANITATION! 100.98 91.94 84.32 PREYENTIYE t 0.00 0.00 0.00 CURATIVE : 241.46 240.54 249.13 IBARAN LGA ADNINSTRATION 1 162.25 162.19 193.56 87.63 74.04 ENV. SANITATION: 902.05 858.19 805.53 1,060.28 1,005.34 PREVENTIVE 1 957.50 1,160.68 730.04 1,211.63 1,091.07 CURATIVE 1 620.08 637.95 875.04 828.94 793.23 SOURCE: LGA ANNUAL FINANCIAL STATEMENTS - 45 - Annex G Page 1 of 8 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Technical Appendix for Facility Analysis This annex describes the technical aspects of estimating production ar,d cost functions as a means to examining the technical and allocation efficiency of public and private modern health care facilities. This includes discussions of the data set, previous research on production and cost analysis, and the estimation techniques. I. The Data Set Before undertakinf .he detailed statistical analysis, the general and sub-sample were compareu± to determine the representation of the study population. Table 1 shows in this annex that the sub-sample oversamples hospitals and undersamples health clinics and dispensaries. However, public and private facilities are similarly represented in both samples. Table 2 shows that with respect to types of primary and secondary facilities, the study group is representative of the general population. Finally, Table 3 shows that sub-sample is drawn more heavily from urban areas than the general group. The geographic distribution by local government area differ between the two groups. The study group greatly oversamples Ijebu Remo. II. Previous Research In the past, cost analyses of health care services in LDCs have emphasized a program approach, that is, analyzing the costs of a specific program such as immunizations or family planning. This has been encouraged by international and local organizations and governments who have an interest in funding a particular program and who need to know the total cost and cost-effectiveness of the program. One problem with such an approach is that collection of the relevant cost data requires making heroic assumptions about how to allocate the costs of inputs which perform multiple services, the issue of joint costs. In general, without careful time-motion studies, it is very difficult to get accurate data on the costs of an individual program especially given the recent emphasis by governments towards integration of primary health care activities. Another problem with the program approach is that often a single estimate of program costs is generated without an analysis of the specific technical (e.g., scale), behavioral (e.g., minimize cost, maximize profit) or institutional (e.g., ownership) factors which determine that estimate. Point estimates are meaningless without information about the cost function which generates them. In this study, an attempt is made to rectify some of these problems by estimating a multiproduct production and cost function. To date, the results of only two other studies which estimate cost functions for health care services in a developing country are available. - 46 - Annex G Page 2 of 8 Both focus on public hospitals. Bitran and Dunlop (1987) examine the determinants of hospital expenditures in Ethiopia using a multiproduct behavioral cost function following the model developed by Granneman, et al. (1986). They find that the number of inpatient days, deliveries, and laboratory exams have a positive and statistically significant effect on total cost. Calculations show that Ethiopian hospitals are operating at an optimal capacity (constant economies of scale); any expansion would lead to an increase in unit costs. There are economies of scope between inpatient and outpatient services. Capital costs (beds) are positively correlated with variable costs. Neither of the input price proxy variables were significant. A problem with this study is that the results are based on a model which assumes that hospitals cost minimize, an assumption which the authors themselves realize may not be appropriate. Anderson (1980) also estimates a behavioral cost function of public general hospitals in Kenya. The hospital is viewed as the focal point of the system with various attached satellite facilities. This means that hospital expenditures include not only that of the hospital itself but related subcenters. In this study a single output, bed days, is used to measure the services provided by the hospital although some correction for outpatient activity is included. Cost minimization is assumed. Key results show that there are economies of scale, that increasing occupancy rates will lower average costs, and that more outpatient activity will increase average costs. Both of these studies emphasize the importance of estimating cost funct±ons to make policy decisions about resource allocation in the health sector. However, the results they generate should be interpreted cautiously since they are based on a model which assumes the unlikely behavior that facilities cost minimize. This present study attempts to test the behavioral assumption of cost minimization. Similar studies for developed countries also generally assume cost minimization. However, in a more competitive market such as the United States, such an assumption may be more acceptable. The focus in these developed country studies has been to refine the conceptual and empirical models of cost. A comprehensive review of developed country studies is given in Cowing, Holtmann and Powers (1983). Further discussion is available in Cowing and Holtmann (1983); Brown, Caves and Christensen (1982); Caves, Christensen and Tretheway (1980); and Bailey and Friedlander (1982). A recent state of the art hospital cost analysis which overcomes many of the shortcomings of earlier studies is presented in two papers which examine the effect of competition on Florida hospital costs (Fournier and Mitchell, 1988; Mitchell and Fournier, 1988). The authors estimate a generalized translog multiproduct cost function which enables one to use a multiproduct approach in lieu of a scalar measure of output, choose a functional form consistent with economic theory, incorporate detailed information on input prices and include facilities which may have zero output for some services. In the translog model, only those assumptions which are necessary to satisfy basic economic theory are imposed (e.g., linear homogeneity in prices). - 47 - Annex G Page 3 of 8 Use of the generalized multiproduct translog function is, however, not appropriate for this study of health care facilities in Nigeria. The poor input price data, the only partially disaggregated autput data and the small sample size do not permit estimation of an elaborate structural form. In addition, it is clear that cost minimization needs to be tested rather than assumed. Other researchers have developed a methodology using cost functions to test whether or not firms cost minimize (Toda, 1976; Eakin and Kniesner, 1988; Atkinson and Halvorsen, 1984) which links deviations from cost minimization to differences between observed and shadow prices. Firms cost minimize with respect to their own set of shadow prices rather than with respect to observed prices. It may be, however, that firms do not cost minimize with respect to any set of prices. For this reason, it may be appropriate to use a more general test of cost minimization. The present study follows the approach proposed by Goldman and Grossman (1983) and Frank and Taube (1987) which is based on the standard economic theory that cost minimizers equate the ratio of marginal productivities with the ratio of input prices. In particular a production function is used to test for cost minimization and to calculate an index of inefficiency. Cost minimization does not have to be assumed to estimate a production function. This index of inefficiency is then used as an independent variable in the estimation of a cost function. __I. Empirical Model Specifications Production Function This study has modified the standard procedure for estimating a production function in one crucial aspect. Typically, it is assumed that input and output data reflect the maximum possible quantity of some output, given various input levels. Such a function is often called a frontier production function. Host field observers would agree that in many cases, health care facilities in LDCs are not operating on this maximum technical frontier. This also appears to be the case in Nigeria. This means that the facility data reflects both optimal and suboptimal technical operations. There is an elaborate literature on the estimation of production frontiers in the presence of real world inefficiency (Schmidt, 1986). Because of the extremely small data set available for this study (42 primary and secondary facilities with complete data), the econometric techniques suggested cannot be used since the required assumptions cannot be met. For example, most stochastic frontier models assume that the errors representing statistical noise are ideritically, independently and normally distributed. The most commonly assumed distribution for one-sided inefficiency errors is the half-normal. An alternative approach would be first to identify those facilities which appear to be technically efficient using criteria based on planning norms and field experience and then to estimate a stochastic frontier production based on this subsample of efficient facilities. The criteria used in this study focuses on the number of visits per health worker per - 48 - Annex G Page 4 of 8 year provided by a facility. This reasonable in light of the fact that the major service provided by primary and secondary facilities is the outpatient visit. If facilities produce more than 600 visits per health worker per year, they are classified as efficient. This figure was chosen in the following manner. First, personal observations of and data collection in several well-performing facilities in Nigeria by the author provided an estimate of the optimal activity levels per health worker. Then, this value was modified to be reasonable for the present data set. The final value chosen tolerates a rather low of activity. Nevertheless, many facilities fall short of this cutoff point (18 out of 42).l/ ible 5.6 in Chapter 5 presents the major characteristics of facilities which do and do not meet this efficiency criterion. Notably, almost all private facilities do not qualify for the subsample of efficient facilities.11 Since a majority of private facilities in this study do not meet the minimum efficiency criteria, the assumption of relatively efficient private producers does not appear to hold in Ogun State. Casual observations in the field support this conclusion. However, before making any final conclusion, researchers should check the reliability of these data. The differences in output level may in fact represent differences in quality, yet because the output of many private facilities is extremely low, the quality factor is unlikely to be the sole explanation. The translog form of the Cobb-Douglas function is estimated. The final form of the generalized translog Cobb-Douglas byproduct production function is shown in equation (1). Other studies have shown that the Cobb- Douglas model performs essentially as well as more comrlicated models such as the transcendental model (Frank and Taube, 1987). (1) ln V - ao + a1 (InL-l) + a2 ln Drgnum86 + L + a3ln HW + a4 ln NHW + a5 Bedsdummy a6 (BedsL-1) x Bedsdummy The Ogeneralized" function is adopted because it permits one to include outputs and inputs even if they take on a value of zero for some facilities. This specification uses the log metric for those outputs or inputs which are never zero and the Box-Cox metric for those which might be zero such as admissions and beds. A value of 0.10 was used for the /1 The production function was estimated for cutoff points ranging from 500 to 1,000 visits per health worker. The results are robust (within one standard deviation). The final choice of 600 visits per health worker allowed the largest sample size under a reasonably strict criterion of technical efficiency. - 49 - Annex G Page 5 of 8 lambda (L) in the Box-Cox metricl/. The lambda was estimated using non- linear estimation techniques and iterative OLS regressions. A dummy variable identifying those facilities with and without hospital beds is included because of the heterogeneity in the type of facilities included in the sample. In particular, it may be that dispensaries, which have not beds, have substantially different production processes. Examination of residuals suggests this may be true and should be tested. The drug input measure was approximated by the following relationship. The number of patients who received drugs in 1986 -X of patients who received in 1986 x (visits + admissions). The survey only recorded the percent of patients who received drugs; therefore, it was necessary to approximate the absolute number. It should be noted that the specification is still affected by multicollinearity of health workers with non-health workers and beds with admissions. However, the OLS estimator remains best linear unbiased estimator. The R-square is unaffected. The major undesirable property of multicollinearity is that the variances are large. Cost Function The translog form of the Cobbs-Douglas model is estimated. The final fonm of the generalized Cobb-Douglas biproduct cost function is shown in equation (2). (2) ln Rcost - bo + b, lnV + b2 (INL-1) L b3 ln Drgpct86 + b4 ln Index b5 1n WagHW + b6 ln WagNHW + b7 Bedsdummy + Be (BedsL-l) x Bedsdummy L As with the production model, the Box Cox metric is used to handle those outputs and inputs which can take on zero values. Mult!icollinearity is also a problem in the cost function between admissions and beds. However, the OLS estimator remains the best, linear and unbiased. /I As lambda approaches zero, the Box Cox metric simplifies to the log met-ic. - 50 - Annex G Page 6 of 8 Table 1 ALL TYPES OF FACILITIES: COMPARISON OF GENERAL AND STUDY GROUPS Full Sample Subsample TOTAL 445(100.0Z) 68(100.OZ) Hospitals Public 11 8 Private 27 7 Total 38 (8.5?) 15 (22.2Z) Comprehensive Health Cart Public 9 2 Private 0 1 Total 9 (2.02) 3 (4.4?) Health Clinic Public 13 0 Private 43 3 Total 56 (12.5Z) 3 (4.4?) Primary Health Care Public 17 4 Private 0 0 Total 17 (3.8?) 4 (5.9?) Maternity Public 112 16 Private 52 12 Total 163 (36.6?) 28 (41.2?) Dispensary Public 124 14 Private 4 0 Total 128 (28.8?) 14 (20.6?) Other Public 20 1 Private 13 0 Total 33 (7.4?) 1 (1.5?) Summary Public 306 (69.0?) 45 (61.2?) Private 138 (31.02) 23 (33.8?) Hospitals 38 (9.0?) 15 (22.1?) Non-hospitals 407 (91.0?) 53 (77.9?) - 51 - Annex G Page 7 of 8 Table 2 PRIMARY AND SECONDARY CARE FACILITIES: COMPARISON OF GENERAL AND STUDY GROUPS Full Sample Subsample TOTAL (excluding hospitals) 407(100.0?) 53(100.0t) Comprehensive Health Care Public 9 2 Private 0 1 Total 9 (2.21) 3 (5.7?) Health Clinic Public 13 0 Private 43 3 Total 56 (13.8?) 3 (5.7Z) Primary Health Care Public 17 4 Private 0 0 Total 17 (4.2?) 4 (7.5Z) Maternity Public 112 16 Private 52 12 Total 163 (40.0?) 28 (52.8?) Dispensary Public 124 14 Private 4 0 Total 128 (31.42) 14 (26.4Z) Other Public 20 1 Private 13 0 Total 33 (8.1?) 1 (1.9?) Summary Public 295 (72.5?) 37 (69.8?) Private 112 (27.5?) 16 (30.0Z) - 52 - Annex G Page 8 of 8 Table 3 LOCATION OF FACILITIES: COhP?"IbO,Y OF GENERAL AND STUDY GROUPS Full Sample Subsample TOTAL (excluding hospitals) 447(100.0Z) 68(100.0X) Part A - Urban vs. Rural Urban Public/Private 197 (44.3?) 40 (59.7Z) 84/113 19/21 Rural 247 (55.6Z) 27 (40.2?) Public/Private 222/25 26/1 Part Bs Local Government Area Abeokuta Total 60 (13.5?) 12 (17.6?) Public/Private 29/31 5/7 Ife-Ota Total 54 (12.2?) 0 (0.0?) Public/Private 23/31 0 Odeda Total 31 (7.0?) 8 (11.8?) Public Private 26/5 8/0 Obafemi Awode Total 26 (5.9?) 3 (4.4?) Public/Private 26/0 1/2 Egbado North Total 29 (6.5Z) 0 (0.0?) Public/Private 23/6 0 Egbado South Total 77 (17.3?) 5 (7.4?) Public/Private 69/8 3/2 Ijebu Ode Total 66 (14.9?) 5 (7.4Z( Public/Private 45/21 3/2 Ijebu North Total 30 (6.8?) 13 (19.1X) Public/Private .7/31 10/3 Ijebu East Total 30 (6.8?) 13 (19.12) Public/Private 19/3 3/0 Ijebu Remo Total 50 (11.3Z) 19 (27.92) Public Private 29/21 12/7 - 53 - NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION ANNEX H Page 1 of 19 FORM A INVENTORY OF HEALTH FACILITIES IN OGUN STATE (1) Name of health facility: (2) Location: (i) Name of settlement (ii) Type of settlement (a) Town (b) I illage (iii) Name of LGA (iv) Name and Discan*e of nearest health facility (3) Year af establishment: (4) Ownership: (a) (e) Government Private (for profit) (b) (f) Z] State Government I Private (not for profit) (c)() 2 Local Government Others (specify) (d) Mission (5) Type of health facility: (a) (e) Wa) z Hospital Maternity (b) - IZI Comp. health Centre (f) Dispensary (c) EZ] Health Clinic (g) Others (specify) (d) G Primary Health Centre (6) Classification of health care facility: (a) (d) (a) z Primary Disease Specific (b) l_____ J Secondary (e) m Vertical Dc) ~ Tertiary (f) 2 Others (specify) * by h---e -54 - Annex H Page 2 of 19 (7) Information on physical facilities: Useable (a) Physical facilities Number Yes No X-Ray Laboratory Operating Room Vehicle: 2 Wheel _ 4 Wheel _ _ (b) Number of in-patient beds: (1) Maternity (iv) Infectious Disease (it) Peadiatric _ (iii) Surgical (V) Medical TOTAL (vi) Others Cc) General physical condition of the health facility: (i, Good (ii) Fair _ (iii) Poor (8) Types of services provided: A. Preventive/Primary (i) LI Prenatal (vi) Family Planning (ii) z Postnatal (vii) Oral Rehydration (iii) o Well child (viii) Growth Monitoring (iv) Immunization (ix) [ Others (specify) .(v) Health Education B. Maternity C. Clinical (i) z Curative Surgery Laboratory C. Communicable Diseases TB Laprosy m Others (specify) -~~~~~~~~~~~~ - 55 - Annex H Page 3 of 19 (9) Staff Position: Type of Staff Full Time Equivalent(FTE) (i) Doctors (Li) Nurses (LIL) Nurses/aidwives (iv) Mldwives (v) Couwnliy Health Workers (vI) Phareeisets (vii) Public Health Workers (viii) Support staff (types) (ix) Others (specify) - 56 - Annex H Page 4 of 19 FORM B lItFORFI1ON FROM SELECTED HEALTH FACILITIES IN OGUN STATE* 1. Name of health facility: 2. Location: (1) Naew of settlement ._. _-- (it) Type of settlement (a) T Town (b) F-| Vlllage (iii) Name of LGA (lv) Name and distance of nearest health facility _ 3. Year of establishment: 4. Ownership: (a) Government (e) Private (for profit) (b) 2 State Government (f) J Private (not for profit) (c) 1,J ILocal Government (8) others (specify) (d) Mission I 5. Type of health facility: (a) j~j Hospital ( : Maternity (b) | Comp. health 2 Dispensary Centre (c) j Health Clinic ( Others (specify) (4) EJ Primary health centre 6. Classification of health care facility: (a) t Primary (d) j | Disease Specific (b) i: J Secondary (e) j Vertical (c) | J TertLary (f) others (specify) _ * ( - ' - by * n Health Consultancy Services Ltd., Lagos. - 57 - Annex H Page 5 of 19 7. Information on physical facillties: ____ ____ ____ ____ ____ ___ ____ ____ ____ ____ ___Useable (a) Physlcal facilities Number Yes No X-Ray Laboratory Operating Room VehLcle: 2 Wheel 4 Wheel - . (b) Number of in-patlent beds: (i) Maternlty _(iv) Infectious Disease (L1) Paediatrlc (v) Medical (111) Surgical (vi) Others Total (e) General physical condition of the health facility: (i) Good (il) Fair (lil) Poor 8. Types of services provided: A. Preventive/PrLmary (i) Prenatal (vi) r Family PlannLng (iL) J Postnatal (vii) Oral Rehydration (iiL) Well child (viii) j Growth MonitorLng (iv) 1i lImmunizat1on (ix) Others (specify) (v) z Health Education B. Maternity C. ClinLcal (i) L1J Curative (ii) [ Sursery (iii) Laboratory D. Communicable Diseases (i) TB (ii) Leprosy (iLi) = Others (specLfy) - 58 - Annex H Page 6 of 19 4 9. Staff Position: Type of Staff Full Time Equipment (FTE) 1. Doctors il. Nurses lit Nurses/midvives iv. Nidvives v. Coumunity Health Workers vi. Phermacists vii. Public Realth Workers viii. Support Staff (types) ix. Others (specify) 10. la-Patient Facility Utilization (1966) Month and Number Types Jan. Feb. Mar. A4t. May June July AUR Selpt. nct. Nov. Dec. Total A. Deliveries: * Admissions • Discharges - - - o Red days occu ancv (8D0)- - - - - - - - - - B. Surgerr: * Almiseions e Discharges - C. Infectious Diseases: I * Admissions DiLscharges - - - - - - - - * 1 _ _ __ ._ _ _ _ _ _ D. others: * Admissions- - - - - - - - - - -- o Discharges _ .. * BDO - - TOJIL * Admissions o Discharges - - * *DO Do D OQ a m m 0 I- U. Out-Patlent Feclkity Utilization (1986) - Servlces l______ Month and Number Tvpes Jan. Febr. Ar r. May June July AUF Sept. Oct. Nov. Dec. Total A. Preventive/rrjarv o Prenatal - • Postnatal . e Imunlzation . _ • Health Education _ * Family Planning - o others - - SUB -TOTAL . . _ . B. Others: * Curative _ _ o Chronic _ _ • Acute _ _ _ _ _ _ _ _ _ _ _ SUBA-TOTAL _ _______ C. DelLveries: TOTAL _~~~~~~~~~~ _ OD~ 0 ' ' x1 x~~~~~~~~~~~~~~~~~~~~~~~~~ x x *_~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- 12. Out-Patient Facility Utilization (1986) - Lits - . == = Month and Number - - == TyPes Jan. Feb. Par. A June July Aug. Sept. Oct. Nov. Dec. Total A. Preventive/primasry: e Prenatal - * Postnatal - - - * lImmnization o Health Education - o Family Planning - - a Others =_ SUB-TOTAL ' S. Others o Curative o Chronic * Acute SUB -TOTAL Deliveries TOTAL ; > qQ D R~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ m %OQ '0 - 62 - Annex H Page 10 of 19 13. What is the else of the population, if knovn, served by this facility? (i) Less then 500 (ii) 500 to under 1000 (iii) 1000 to under 5000 (iv) 5000 to under 20,000 (v) 20,000 to under 50,000 (vi) Over 50,000 14. Total Cost of operations in U in last 5 years | ~~~~~~~Y E A R | 19 a I 1 9 82 19 83 19 84 19 5 | 1 9 86 1. Personnel: ' _ R J os Salaries of health workers . . I . o Benefits of health workers j - _ i | Salaries of support staff L o Benefits of support staff - _. j_i 1. SUB-IDTA1. i ____ _ i I~ . St_ -!T-Q- ; II. Son-Personnel: t . o~~~~~~~~~~~~ -8 !- -- - a Drugs I o Other supplies ! -_, __ _ | ; o Tra_sport __ _ _ I _ ao Maintenance and repair _ I ____I o Others __._ __I- _ - 11. SUB-TOTAL _, , ,,., :,I.I _ _ Total Recurrent (I) + (11) I I I1I. Capital Expenditure (i1 + UI1) + (III) . _ I~~~~ X -~~~~~~~~~~~I -64- Annex H Page 12 of 19 1S. Percentages of total couts attributable to the two brcad categories of 'output' (1906 only). 'Output Category' Z of Total Cost Attributable to it 1. In-Patient 2. Out-Potient (TOTAL) (100) 16. Perocuteges of various coat categoties attributable to various in-patient medical specialties (1985 only): II o (iercentes o cost CLtegor/meei,:ai cazeco'e _ _ Cost Category | ledieal | Surgica l Pasediatric | Ftern-ty In!e:tious Others | Total ? . __________ .__ I_ ___________ D i ' :ses I I. Pesr.nnel: o Salaries of health I _ workers . ' 10 a Benefits of health I _ _ _ workers . __I__ _ __ _ e Salaries of suppost . __ staff . _ 100 e Benefits of support I I staff . . 1 _ 1 100 II. Non-Pe sonnel: I I o D-jgs 1 100 Other supplies I 100 _ • Transport ioo _ 100 * Maintenance and 1 I i repair _ I _ 1 __ 1 100 o others _I _ I I ! _ _100 !III. Capital Expenditure 1 I I I __ __ ___ _ I to 0 * ~~~~~- 66 - _ 00_ 88 0_ Pag 14 __ of 19 .51 ^ 0 3 - 8-k _ *.4S-__ . _ R .1 ____ _ sXo S_ ___ __ R i_ __ __ _ 3 ! , * _ e~ ~ _ ; I I 1 -g|- - - - - - - - - 0. 18. Percetasges of various cost cateoties attributable to variou out-patient services .________ 2 of COt cateRorY and tves of service _ . Cost Category Pre/Post imni- Health Faiuly 1 Well Chronic Acute Others Total Ibtal cation Education Plamling Child Curative Curative _ A. Personnel: , _ o Health salaries o fleslth benefits o100 o yon health salarles _ 0 e Non health benef its 100 S. Non-Personnel: o DZU55 __ o Others pplies = 100 o Transport __=___100 o Haintene-oce and repair to_o_ 100 o Others . loO C. Capital Expenditure U ~ ~ ~ ~ ~ ~ ~ ~ ~ ._ . . 10 0 *40 19. Sources and amount la U of revenue In laat 3 years Source __ _ YEAR AND AMOUNT 1 9 8 4 1 9 8 5 1 9 8 6 Federal Budget _ _ State 8udget _ LGA Budget _ Special Grants Affiliation Grant I__oo Prlvate Contribution . User Charges _ -__ _ _ Insurance Company Payments a L oans__ _ _ _ _ _ _ _ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Others (specify) e.g. Drugs .__ , TOAL oqtv Ir^L _ _ . ... _ 05 e g ' a s~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~39 - |_ S * " ~~~~~~~~~~~~~3 20 8ervlce Charges Servlce Type Date when cbarge hLnim chbarge IaILmm charge Categories of people started exenpted (if any) (1) Registration Card: o maternity Casea _ __ _ * Other out-patients * In-patients _____ (1i lmnization Ciii1 Pareceta o. (Iv~ Contneptive pills for I month __ ._ (v) Condoms a packet of3. __3_ _ (vI) Laboratory: * CBC (Blood Count) . _ - * BaaIc Urinal,.is _ (vii) Cheat X-Rla __ (VIii) Ante-atal Services __ six) Dellve!y: • Bas , nAcoxplicated * Ceaserean Sections . _____.____ (s) Amendectou ..._ (xi) Room/Board basic, for I day (xfil Dressins _ .. (Axiii) Full treatment for mlaria (xlv SettLng Broken Am. __._.___.____. x 0 £1. NL=ber and people charged for various services, etc. and moant realized in the last 3 years Broad Sesvice 1 9 8 4 - 198 1 9 8_ 6 Caregory No tha No not Amount Amont No that No not Amount Amount No that No not Amount Amount paid paid realised retained paid paid realized retained paid paid reaiized retained (N) (N) ____ A. Total In-Patient o Deliveries I o Poediatric . o Surgical - o lldical o In-Patient Diseases , . _ . • others a S. Total Out- Patient o Ire-Natal . . • Post-Natal o Iwmzniza- tion_ * Health Education _ ,__ o Family Planning - * Vell Child _ o Chronic Curative _ _ . * Acute Curative _ . __ _ _thes _ _ _ _ _ x 0 I.E.. iii mmli.~~~~~~~~~~~~~~~~~~~~~~~~~ - 71 - Annex 91 - 71 - ~~~~~~Page 19 ot 19 22. What proportion of your patients were you able to supply with drugs in the last 3 years? Year Proportion Supplied 1984 __ I 1985 1 98 6 - _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ _ _ _ _ _ _ _ _ 23. Indicate your sources of drug supply and theLr relative importance (ln terms of proportion of total supply) over the last 3 years. Year Sources and Pro artion : Central Zonal LG Direct from From Others Total Store Store Store Manufacturer Whole- 70 .__ ___ saler _ 1984 _ . 100 1985 __ _ _ _ _ _ __ _ _ _ _ __ _ _o_ _ _ _ _ 0 1986 too__ _ __ _ _ __ _ _ _ _ _ _ _ _ 0 24. How have you paid for your drugs over the last 3 years? Year Modes of vavment and their significance Total Cash Cheque LPO Credit Others 100 1 1985* 1 1 100 1986 100 25. Please indicate expenditure for and recelpt from drugs over the last 3 years: Year Total Payment in Of Total Drug in 14 (Expenditure) Revenue from sales 1984 1985 1986 - 72 - Annex I Page 1 of 4 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Financial Information System Data Requirements The following are the minimum data requirements necessary for monitoring the financial status and efficiency of modern health care facilities. Some of these questions may already be part of the general management information system. Special attention should be given to collection of expenditures, revenues, staffing levels and salaries and annual utilization. With this basic data in hand, a facility can track utilization levels, the cost per service provided, and the balance between revenue and expenditure. 1. Name of Health Facility 2. Name of LGA 3. Ownership (check one) a. federal e. private (for-profit) b. state f. private (non-profit) c. local g. other d. mission 4. Type of health facility (check one) a. hospital e. maternity b. comprehensive health center f. dispensary c. health clinic g. other (specify) d. primary health clinic 5. Number of usable physical facilities a. X-ray b. laboratory c. operating room d. vehicles - 73 - Annex I Page 2 of 4 6. Personnel and annual salaries (full-time equivalent: FTE) (Year ) Annual Annual Average Average Position FTE Salary Benefit (Number) (N) (N) Health Staff Doctors Nurses Nurse/midwives Comunity health worker Pharmaciests Public health worker Other health staff (specify) Non-Health Staff Administrative (managers, accountants, clerical, etc.) Maintenance staff (engineers, security, laborers, cleaners, etc.) Other (specify) - 74 - Annex I Page 3 of 4 7. Inpatient Utilization (Year ) J F H A M J J A S 0 N D Total Deliveries Admissions Discharges Bed days occ. Deaths Surgery Surgery Admissions Discharges Bed days occ. Deaths Other Admissions Discharges Bed Deaths Total Admissions Discharges Bed Days Occ. Deaths S. Outpatient Utilization (Year ) (Number of visits) J F M A M J J A S 0 N D Total Primary Health Care Family Planning ORT Immunization Other Curative Deliveries Total - 75 - Annex I Page 4 of 4 9. Total cost of operations (Year ) Personnel Health Workers Salaries Benefits Non-Health Workers Salaries Benefits Sub-total Personnel Non-Personnel Drugs Other supplies Maintenance and repair Transport Sub-total Non-Personnel Total Recurrent Total Capital 10. Sources and Amount of Revenue (Year ) Government Subvention Federal State Local Agency Affiliation Special Grants Private Constributions Drug Revolving Fund Other User Fees Loans Other (specify) - 76 - ANNEX J Page 1 of 5 NIGERIA: HEALTH CARE COST. FINANCING AND UTILIZATION Statistical Analysis of Efficiency in Public and Private Facilities: The Frontier Production Function (a) The Model 1. A short-run bi-product cost function was estimatedl/, and a Cobb-Douglas functional form was tested.2/ This simple model is homothetic: the cost-minimizing input mix remains constant as the output level changes3/ Equation 1 gives the general form for the cost function: (1) Short-run Variable Costs = f(Wage of Health Worker (HI), Wage of Non Health Worker (NHW), Admissions, Visits, Quality, Beds, Inefficiency), where HW and NHW are, respectively, health workers and non-health workers. Equation 1 includes proxies for quality and inefficiency, in addition to variables representing the traditional determinants of cost (input prices, fixed capital and output). Quality is captured by drug availability. j/ Although the following conceptual analysis borrows heavily from standard cost analysis, the results ore only suggestive of the cost structure of health core services, for two reasons: (a) the data pertain to expend!tures rather than costs. Expenditures do not Include In-kind gifts and donations, so expenditures are likely to underestimate eots; expenditures measure accounting rather than opportunity cost., and capital depreciation Is not Included; and (b) the assumption of cost minimization to not satief led. Nonethele"s, since an Inefficiency Index is included, this cost analysis provides useful information and results. V Long-run analysis cannot be conducted hero for two reasons: (a) In cross-sectional data, probably not all facilities are operating at their long-run cost-minimizing equilibrium, and (b) the survey does not provide the data on capital expenditures needed to calculate total long-run costs. !/ This short-run variablo coat function ia based mainly on traditional economic theory, which states that short-run costs are affected by Input prices, fixed capital and output levels. Several caveats to straightforward application of this theory are required: (a) In this case only personnel wages are used for Input prices but nonlabor prlces are not known. This Is a problem only If nonlabor prlces are not uniform for all providers, which may be true fol some Inputs such as fuel, vehicle maintenance, and furniture, but probobly not for drugs; (b) output and capital may not be truly exogenous variables. With respect to output, facilities are probably unable to control utiliastion levels. The variable for capital stock is exogenous because conditions In developing countries make It very difficult for facility manager. to change capitol stocks to match utilization levels quickly, If at all; and (c) short-run variable expenditures may not be truly ondogonous because political constraints probably have more to do with how general budget of health departments are set then utilization levls. This Is especially true for pubilc facilities. Nonetheless, even If budgets are fixed for overall locol government and state activities, resources can still be transferred between facilities to respond to local changes In utilization patterns and input costo; thus, expenditures can be considered the endomenous variable. - 77 - ANNEX J Page 2 of 5 Inefficiency is reflected in an index (estimated from a frontier production function) measuring deviations from cost-minimizing behavior. A frontier production function (including only technically efficient health care facilities) was estimated to calculate marginal productivities (MP) for health workers and non-health workers.4/ The efficiency index indicates whether marginal productivities per unit cost are equal across the two worker categories. The efficiency index is shown below in (2): (2) Inefficiency MP(NHW)/MP(HW) |-1 Index | Wage(NHW)/Wage(HW) 1 It was found that, on average, facilities use relatively too few heaith workers. The rel&tive productivity of non-health workers is less than their relative wage. 2. Basic economic cost theory predicts the following relationships: (a) increasing any of the input prices should increase costs; (b) as production levels increase, total variable expenditures should also increase; (c) the coefficients for admissions and visits should be positive; (d) improvements in quality are usually more costly to provide, so the quality coefficients should be positive; and (e) the larger the inefficiency index, the higher the costs, since this index reflects m deviations from cost minimization.5/ 3. Cost analysis measures of health care services focus on marginal and average costs for inpatient and outpatient services, product- specific and ray-specific economies of scale, and economies of scope. The cost estimates must be interpreted cautiously since they are based on a short-run, non-cost- minimizing cost function rather than a long-run, cost- minimizing cost function. Since only changes in variable inputs (labor) are being considered, the results are suggestive rather than conclusive. 4. Marginal costs (MC) for visits and admissions are defined in equations 3a and 3b in the usual way, namely, as the incremental cost of producing one more unit of the output. (3a) MC. = dC/dV (3b) MCI, - dC/dIn S. Average incremental cost (AIC) shows how much total costs will increase if one output (holding the others at mean values) is produced u i/ Technically efficient facilities are deined as those which service the minisu acceptable level of at least 600 visls per helth worker. 3/ The capital variable Indicates whether or not facilities are In their lon -run equilibrium. Cowing and Holtan (1968) point out that hospitals rer not In equilibrium unless the coefficient on capital Is negative and equal to Its user coat. The Institutional variables are assumed to affect th, level of expenditures but not the shape of the cost function wlth rpect to outputs or factor Input ratios (Hicks.- neutral). This assumption Is required beuse the sample sixe Is not large enough to test Interaction term. - 78 - ANNEX J Page 3 of 5 versus not being produced at all. Equations 4a and 4b give the formulas for average incremental costs. (4a) AIC, [C(V,In) - C(O,In,]/V (4b) AICj, = [C(V,In) - C(V,O)]/In 6. Product-specific economies of scale (EOS) are analogous to economies of scale in the single product firm. In this case of multiple outputs, economies of scale can be examined with respect to changes in the level of only one of the outputs while the others remain fixed at mean levels. The general formula for product-specific economies of scale for the ith output using results from a short-run total variable cost function is given in equation 5. (5) EOS; = AIC;/MC;, i=either visits or admissions 7. Ray-specific economies of scale is another version of scale economies that measures the effects of changes in total output on cost when the composition of outputs is fixed. In this case, the ratio of inpatient admissions to outpatient visits remains constant while the levels of both expand. Ray-specific economies of scale (RSE) can be calculated using equation 6. (6) RSE - C(V,In)/ (In * MC;f) + (V * MC,)] Equation 5 is the ratio of production costs to the revenues the multiproduct firm would earn if each output is priced at its marginal cost. Values greater than one indicate economies of scale. 8. Economies of Scope. One aspect of efficiency is whether facilities derive any cost savings from providing more than one type of service in a given facility. Economies of scope arise from the sharing or joint utilization of inputs. Empirically, thi3 can be determined by calculating economies of scope (SC). Economies of scope exist if the total cost of providing each service separately is more than the total cost of providing the same level of each service, but jointly. This is qhown in equation 7. (7) ([C(O,In) + C(V,O)] - C(V,In)}/C(V,In) (b) Estimation Results 9. The OLS estimations of equation 1 are given in Table 5.7. Cost measures derived from these estimations are presented in Table 5.8. Several coefficients (visits, drug availability, wage NHW, wage HW) are significant. The model has a very high overall explanatory power (the R- square value is 0.91) meaning that cost predictions will be good. This is important because calculation of the aforementioned cost measures is based on predictions. 10. As expected, visits and admissions increase expenditures, as do wages. The efficiency variable is insignificant, suggesting that - 79 - ANNEX J Page 4 of 5 deviations from cost minimization have little effect on expenditures.6/ The coefficient for beds (the proxy for capital) is positive, indicating that facilities have too many beds (the health care facilities aLe not at their long-run equilibrium). The remaining coefficient, quality, yields an unexpectedly statistically significant negative sign. Quality appears to reduce rather than increase costs. Perhaps facilities that successfully manage their drug supplies and meet the needs of their patients are more efficiently managed. In other words, the availability of drugs may be a proxy for management quality. Table 1 COBB-DOUGLAS COST FUNCTION ESTIMATIONS a/ (Dependent Variable ln Variable Costs) Variable Cobb-Douglas Constant 1.628 (2.802) ln visits 0.597 (0.113)* Admissions bI 0.011 (0.044) in percent drugs available, 1986 -1.361 (0.659)* ln wage health worker 0.586 (0.238)* in wage non-health worker 0.387 (0.186)* Beds dummy (Ono beds, 0.221 (0.356) Beds dummy x beds bl 0.093 (0.212) in efficiency index -0.160 (0.182) F-test 18.21 R-square 0.91 0 Observations 24 Source: Calculations based or, data in Ogun State Health Financing Study (AHCS). 8/ Standard errors in parentheses b/ Box-Cox Metric, L = 0.10 *Significant at 5 percent The .tietlcal Ineignificance end unexpected negative sign of the Ineffllency coefficient Is probably, Io part, o result of the =al I camplo sizo. - 80 - ANNEX .I Page 5 of 5 Table 2 COBB-DOUGLAS ESTDMATES OF COST ANALYSIS MEASURES Admissions Visits Marginal cost (Naira) 6.85 4.82 Average incremental cost (Naira) 18.05 3.73 Product-specific economies 2.64 0.77 of scale (index) Ray-specific economies 0.76 of scale (index) Economies of scope -0.10 (index) (c) Cost Analysis Results 11. Table 5.8 and Figures 5.1 and 5.2 reveal relevant features of primary and secondary health care facility service costs.7/ 8/ As expected, inpatient services are more costly to provide (by 100 percent) than outpatient visits in terms of marginal and average costs. Inpatient average costs are more than double outpatient average costs. As shown in Figures 5.1 and 5.2, marginal costs are less than average costs, indicating that facilities are generally operating on the downward portion of their average cost curves. This is especially true for admissions. Although also downward sloping, the marginal and average costs curves for visits are somewhat flatter. These results are consistent with the product-specific scale measures which indicate increasing returns to scale for admissions and nearly constant returns to scale for visits. The ray-specific scale measure, which assumes a constant mix of services, shows nearly constant returns to scale. U These esult are average.; facilities show wide veriability in results. 1e¢suse only three private facilities met the production frontier criteria, the following cOmate pertain primarily to public facilities. -.7 IFORM C* WDWJu PAnwrm QUESTIOMNIURE Settle- llousehold La ment Household Number category FTI I _ __ 1 HDUSEIJDDOLDE .CM.. 1 2 3 4 5 6 7 NOTES (1) Local Cowgnmt: AbeoIuta 0 Ife-Ota 1 Odeda 2 Obaefemi Aode 3 Egtado-Nortl 4 Egbado South .*5 Ijebo ode 6 Ijebu North 7 Ijebu East 8 .ljbu em9 (2) Settilm1ts 11 1 l1 ' 2 (3) Nhnsedold ber 0001 - 300 (4) oduShold teory: I - FatLent hueehold O - lNom patient household OQ~ * Survey instrument developed by African Health Consultancy Services, Ltd., Lagos. o 0 1OUSEJHLD CODE 1. SOCIO-ECofU4IC IN BOI 0LQN jZiIII>ia 9 _. it 12 13 14 15 16 17 18 Serial Wlat is the What Is the What is What is the What is What is What is What is Did member 11ow muchi did Number name of relationship eber's age of the the the the high- the work in the member earn household of me ber sex? member in marital religion est edu- major past montil? in tthe past member? to head? m - t years? status of cational occupa- Yes - I inonth? Start with F - 2 Under I of member? level of tion of No - 2 the head year is 00 member? member? member? (*) wives, chil- dren, rela- tives, wards and others 02 __ '__ _ -04.. 05. ... 09.I (10) Relationshi,: ed 'Lo wife 2, son 3, daughter 4, father 5, mother 6, domestic help 7, relative 8, others 9 0 (13) Marita l1 Status: Siagle 1, married 2, widowed 3, separated 4, divorced 5, don't know 9 (1 4) Religion: ChrLot1Un 1, umaslm 2, traditfonal 3, none 4, others 5, don't know 9 (15) Education: None o° Prim8y 1, Secondary 2, Teacher Training 3, Technical 4, NCE 5, University 6, Others 7, Don't know 9 (16) occupations Too young/old 0, Appl cant 1, Farming 2, Fishing 3, Trading 4. Artisan 5, Teaching 6, Office worker 7, profe sional 8, others 9. OQ0 0 . ,{ o 15<~~~~~~~~~~~~~~~~U -2- HOUStRIDLD ODDE:;" ! lI II III II. HOUSIIG UNtT CM0BID019 19. Type of housin1 aooerda^tion: 8__1 Snge Roa) / / Flat (3) tDuplex (5) ottilers (specifv) (7) L Roomn aN Parlour (2) /Bungalow (4) Z / Storey Building (6) 20. Is this housing unit rented or is it owned by someone In the houselhold? L / Rented (I) / 7 Owned by someone in tlle liousehiold (2) 21. What is the glor source of wrater to the houoing unit? r Piped water Into residence (I) Unpiped well/spring water / / Others (specify) ( (4) eipW water faon well or borehole (2) ! t Tanker (5) w Pip ! t^ ter-Into otteet (3) Z / Opened Canal/pond/River (6) 22~~~~~~. How. f a I9 ~ tee^9trlkn eeoutce uside but less 500 feet(2) Outside and greater than 7 X~~nside hmin8 unit (1) 500 feet (3) 23. What is the onjor Sttdm of light? N 7 WeA (1) L / Generating Plant (2) / Lantern (3) "!/ ' ! ! C~as Lamp (4s Othters (specify) (5) .tii,.,.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~O :, -3 IOUSEHDLD UODIT D DT . 24. IH7at kind of toilet faeility is there in thla houselhold? / Pit (2) /7 Bucket (3) L1Z7 Stream ( 6) Lr RBush (5) None (6) /:/ Othiers (specify) 25. Which of the follmv,z emfort/leisure items do you have in the household? /j7 Television(l) L/ Radio (2) _ Stero (3) /L 7 Video (4) 1 7 Fridge (5) LgJ..7 Others (specily) (6) 26. Does any musher of tdba household own any land hete or swhere else? L=/ Yes (I) No (2) 27. What is the total snt of land owned by all household menber? ** I Standard Plot (1) 2 to 4 standard plots (2) / At least 5 standard plots (3) 28. What is the m ltbly ret or (if owned) how much weld the housing unit rent for if rented out? L 7 Under 6 20 (I) , UZ2D to 1150 (2) 7 s151 to under 1100 (3) [| /100 to 200 (4) 29. Does anyone in the bousdaold own any other house than this one? Yes (1) | / No (2) 30. low many houses ax m by all bousehold A 'r combined? LQ Just one (I) L 2 ato 4 (2) Lr Above 4 (3) 31. Does anyone in the ld own any farm animle or birds (poultry)? L Yes (1) /Z No (2) 0 Itb b ,55 . in~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U n *rl U oo *I_ -4- IIOUSEIOW Ong: IMUSE1TOLD UNIT ONIUNSWWD 32. If yes to quesetLo 31D please Indicate what kinds and how many. Kinds of farm animaLs and birds (poultry) owned Ntumber * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~co 33. Does arone La the household own a motorized vehicle, either a motocycle or 4 -wiheel vehicle? L 7 fes (1) / 7 No (2) 3 4 . If yea to que8tion 33. 9tease Lndicate how many motorized vehicles are owned by all membe-s of the hotisehiold combined. motor-cycle(s) _ 4-weel vehicle(s) 0 I~~~~~~~~nI Mn;EIKLD CODE:J III. MEALTH STATUS ay MIEN D M@uS 35 36 37 38 39 40 41 42 43 44 .45 46 Serial What is the mme Iluld you uas If yes to For how if Has member If yes For how For how What Number of household describe member Q.38, long was member member liad to Q. 42 many days uarty ditd of member? member'a had,a whet kind member is a a.,y illness what during the days membet house- health as serious of illness unable to woman, or injury was the past four diuring dc hold excellent illness or injury? carry on is she during the nature? weeks did tlhe past first member (1) or usual current- past four member fotur abouLt Cood (2) injury activi- ly preg- weeks? suffer weeks the lair (3) within 'les nant? Yes - from was injutry. Poor (6) peat 2 (Days) Yes - I No - 2 this? member iLl- years? No - 2 uniable ness, Yes - I to etc. so - 2 carry on 01 __________________ _ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~cii 03_ _ 05 _____________________._, 06 _ _ __ ___.._._ -° -~ . 07_ _ _ 09. __ _ _ _ ._ _ _ _ .__.___ __ _ _ _ _ _ _ ___ 39. Cough (01), lever (02), Stomach (3), Diarrhoea (04 ), Weakness (05), Headaclhe (06), Guinea Worm (07), Leprosy (08), Burns (09), Bites and StLngs (10), PoLson (11), Falls (12), Others (13). 43. Cough (01), Fever (02), Stomach (03), Diarrhoea (04), Weakness (05), eadache (06), Guinea Worm (07), Leprosy (08), Burns (09). Bitm and Stings (10), Poison (11), Falls (12), Others (specify) (L3). 46. Did nothing (0), S.if medication *. Sought outside assistance (2). | o P4 6- WSELD CXDE: HEALTH STATUS 0F HoU(KJSBLD 3CXOND. 47 4gi Is 51 52 53 54 55 56 57 Serial Wihat is the ho, b9ts Whese did mhat is the Is this a How far How long llow Much flow many flow mtucit Number name of cons lted the first tome of the public or was this d.d it was tran- times was patid of household 1L%t if consulta- place where private consi lt- take to sporta- was thlis to this house- member? anever to tion take the first establish- ation get to tion person person hold was place? consulta- utent? from tilis cost to constul- for all member? outgide f tion was? Public - t here? consul- get to tedi for the con- aausltance?; (Name) Private - 2 (Kms) tation? tilis tilin sulta- Doctor I (one way) (hirs/mns) consul_ injury tions Nurse 2 (one way) tation? or iii- during ?hia. 3 ness * the -t Midwife 4 (N) during past 0x Healer 5 (one way) tIle four otherFs 6 X . past weeks? four () weeks? 01 _ _ _ _ _ _ __ _ _ _ .__ _ _ _ -__ _ _ _ _ _ _ _ 02 _ ._ 03 __-_____. 05 _ _ 06 ____,_ _ 07 ___ _ . .____ 08 . 09_ _ 10 _ ___ . Note: - 50. Pharmaey/ceh,st store 0) Traditional healer (1), Spiritualist (2), Government macernity (3), Private inteiity (4 ), Public health.centre/clinic (5), Private clinic (6), Comprehensive health ceIlLre (7 Covertneent, L pital (8), Private hospital (9). . n . R~~w - ~ ~ ~ ~ 7 KWUSKHDLD CODE: HALTH STATUS OF mSoW a. 58 = 9 , 61 62 63 64 65 66 67 68 Serial What Ls the nme Were any IL so, If answer What did Who was Where did Wbat is Is tltis a flow far Number of the household drugs where to Q. Go the consulted this con- th.e public or was the of member? prescribed? were the is yes, member do second if sultation name of private consulta- house Yes - I drugs how much next there was take the establi- tion from member No - 2 obtained? did the (c further place? place shment? here? drugs for this outside wiere Public -l (Km) cost? illness consul- the Privace-2 (one way) (U) or tation? second injuty? Doctor I consul- Nurse 2 tation Pharm. 3 took Midwife 4 place? Healer 5 Others 6 01 _ . . _._.__ 02 .. . - ! 03 .. . . i 0b __ _ _ _ _ I__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 05 _ . . . ._._ _ 06 '.,.I -0 8 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _. . _ _ _ _ _ _ _ _ _ 09 10 _ _ . Notes: 61. Chemist store (I). Pulic Health Care FaciUlty (2), private Health Care Facility (3), Others ( 4) 63. Did NothiLg (0)1, WE Medication (1), Sought outside assistance (3). 65. Pharncy/aie.iet Store (0), Traditional healer (1), Spiritualist (2), Government maternity (3), Private teatcrnity (4) Public health cetre/clinic (5), Private clinic (6), Comprehensive health centre (7) Government hospital (8) private hospitalt (). VI ~~~~~~~~~~~~~~~~~~~~~~~~~~~~'"¢l i __ l | |I m u m. * _ -8 HOUEMIOLD CODE: I j lEN.TI STATU S OF D lND-S0 oM. 69 70 1 71 72 73 74 75 76 77 Serial What is the now long did How ouch was How many tunes How muclh was Were any If yes to Q. 75, If yes to Number of the it tIake to spent for was this paid to tlhis drugs where were the Q.75, how of member? get to thiss tranport- person con- person for prescri- drugs obtained? much did j house- second con- ation to set sulted for this injury bed? Chemist store I clte drugs hold sultation? to this this injury or illness Yes - 1 Pub. Health cost? member (hrs/mns) consulte- or illness during the No - 2 Care Facility 2 (one wmy) tioni during the past four Pri. Healtlh (4) 6a4 past four weeks Care Facility 3 (one way) weeks (N) Others 4 01 '. ..... 02 03 04. . ,-__ _ 07 __ _ _ _ ._ _ __ _ ____. 08 06~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' 09 _... to - U , T~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 : -9- HOUSEHOLD CODE: MEALTH STATUS 0F HOUCSUOLD EUMS CONTD. 78 79 80 61 82 83 84 85 86 Serial What is the new Who was con- Where did the Whnat is tite Is this a llow far was How long did it H1ow utlcfh was Number of the household sulted consultation name of the public or this constil- take to get spenit on tran- of member? third if take place? place where private tation this consul- sport;ationi to house- there was the consul- eatabli- £rom here? tation get to this hold further tation took shment? (Kmn) (lhrs/mns) cons,ittation? member consulta- place? Public -I (one way) (N) tLos? Private -2 (one way) Doctor I Nlurse 2 Phazn. 3 Midwife 4 Realer 5 otheirs 6 03 _ _ _ _ _ _ _ _ _ _ 06 _ _ _ 07 = 08 __ . ._.__. 10 -- Note: 81. Phanmacy/Chemist Store (0), Traditional Healer (1), Spiritualist (2), Government Maternity (3), Private Raternity (4), Public Health Centre/Clinic (5), Private Clinic (6). Comprehensive Health Centre (7), Government Hospital (8), Private Hospital (9). x 0 o x *__~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I, - 10 - HOUSEHOLD ODE: _._. HEALTh STATUS OF BOUO R COND. 87 88 _ 89 90 91 92 93 94 Serial What is the _ any times How -wch vas Were any drugs If yes to Q.91 If yes to Q.91, Was there any Number name of the this paid to this prescribed? where were the how much was further consul - of member of P don person for drugs obtained? paid for the tation for this house- the c aulted for this injury Yes - I Chemist Store I d rugs inju.ry or hold household? tl injury or or illness Ne -2 Public Health Care illness within nmember 11 nss during during the No _ Facility 2. (N) tte past four tl past four past four Private Health Care weeks. w ^? weeks for Facility 3. Yes - I all coslUt- Others 4. No - 2 tations? ol 02 03 _ . 04 __ . 06 -- _ _ _. - 07 _ 08 09 10 4.~~~~~~~~~~~~~~~ X e4do I~~~~~~~~~~~~~~~~~~~~~~ -~~~~ - 11 - UEILD CODE: HELCTH STATIJS OF OUSE)LD MOM=S DNhTD. 95 96 97 98 99 100 101 02 Serial What is the one teber Was there any If answer to Q.98 Is it a public or What kind of Why w.as it not Number name of the recovered other health is yes, what is private facility is used? of member of umn this care source the name of the facility? this? house- the house- Olluess or that could health care Price I hold hold? Injury? have been source? Public - I Distance 2 member Tes - I used but was Private - 2 Quality of care 3 o - 2 not used? Lack oi drtugs 4 Yes - I So - 2 others 01 02 03 04 _ . .__ - 0 6 _ _ _ _ _ _ _ _ _ _ _ _ 07 - . 0 8 _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 10 1 . . _ ._.._ Note: 101. Pharacy/chamit store (0), Traditional Healer (1), SpirLtualist (2), Government maternity (3), Private nmatern"ey (I Public health centre/clLaic (5) Prlvate clinic (6)9 Comprehensive health centre (7), Gov rnment llospital (8) Private Rospital (9). OQIy I- - 12 - HOUSEHOLD CODE: I IV. INFORKATIO1 01 DE=S 103 104 105 106 107 108 Which of these drugs am Wlhat Ls an What proportion What proportion Wlat proportion What proportion Total proportim commonly used in the estimate uf of the drug do of the drug do of the drug do uf the drug do for Qtiestions household? how such is the household the household the household the household 105 to 108 spent in the members get members obtain members obtain members obtain household from public from private from chemist from otlher on each of facility? facility? store? sources? ( % ) Annual _0 (2) (2) (2) Monthly - I Paracetamol ( ____ _ __lO0 Cough Mixture (02) __ _ _ _ __ _ _ __ _ _ _ ___ _ _ _ __ _ _10 Anti Malaria drugs (03n 1007. Anti Acid (0 4 ) 1007. Hag. Trisi licate (05) . . 100 A.prin tabs (0) __0% Tetracyine cars (07) 100o Iron/Folic Acids OID ____ Multivitanmi (09) 100% Calamine Lotion- (30) __ _ _ _ __ _ _ _ _ ___ _ _ _ __ _ _ _ _ __ _ _ _1 0 Condom (11) _ 100% Contraceptive pills (1l__ _ _ _ _ _ _ _ _ _ _ _ ___ _ _ _ _ _ _ _ __ _ _ _ _ _10 other contracetives (13 100% Antlseptc (14 100 Dressings (15) . 1007 LOthers (specify ME_ 1002 ED(DO x 0 Ln - 13 1EDUSElMLD ODDS: V. UTILIZATON OF MM S3IVLS (These questions would pertain to the birth oi the youngest child in the household and his/her motIher). 109. Where was the youngest dhld born? 1 t / At home 3 !Public maternity 5 L j Private hospital 7 L 7 Others 2 P Frivate msternity * /7 Comprehensive health centre 6 L Public hospital 8 /7 Don't know 110. Did anyone assist at the birth? 1 J 7 Yes 2 L / No 3 7Dont know 111 Who assisted at birtbt 1 /7 no one, mother was alone 4 /7 Midwife or nurse/mid ifee 2 /m7 F lyly merb, relative or frieLd 5 t7 Doctor 3 L=7 Traditional birth attendnt 6 t7 Other 112. Did the mother receive my prenatal care? I £J Yes 2 L:2 no 3 L 7 Dont know 1130 If yes, from wima did mother receive prenatal care? I / = NO one, gmther wa alone 4 /7/ Midwife or nurse/midwife 2 7 Fmily inber, relativ, or friend 5 Doctor 3 /7 TradiLtiomt (non-clinician) birth attendant 6 /7 Other ow 0 t.n - ,X - HOUSEHOLD CODE: UTILIZATION OF NCR SZVZS9 WDNTD. 114 . If yes, where d&d mother receive prenatal care? 1 ! ..At home 3 lC7 Public maternity 5 Private Hosp. 7j7 Other 2 / ,?rivate maternity 4 D Cmq. health centre 617 Public llosp. 8 1 Don't know 115. Did the child bhve any iumunsiation? 1 Yes 2 !7 No 3 Don't know . Ln bsiy~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~5 2/ s~~~~~~~~~~~O _ 0 ,. . s.~~~~~~~~~~~~~~~~~~L - 96 - Annex L NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Information on Second Consultations Had a second consultation other than a pharmacy (317 out of 878) 1/ (Number) (Percentage) Public hospital 2/ 133 42 Private hospital 57 18 Public healt;h clinic 42 13 Comprehensive health clinic 30 10 Had a second consultation other than a pharmacy (poorest income group) 3/ (39 out of 349) (Number) (Percentage) Public hospital 16 41 Private hospital 7 18 Traditional healer 6 15 Had a second consultation other than a pharmacy (highest income group) 4/ (28 out of 171) (Number) (Percentage) Public hospital 8 29 Private hospital 6 21 Traditional healer 4 14 So few of any of the groups report second consultations that conclusions are ambiguous. Most of the respondents in each income class go only to one major facility and a pharmacy. Those who seek another source of care tend to be people who need hospitalization; in most cases they go to a public hospital. 1/ Since 878 of the sample group said they used a pharmacy as the second source of care, most of them probably are following up on the first visit to purchase suggested drugs. 2/ Some 57 of these said their first source also was a government hospital. 31 Among the poorest quintile, 125 of the 349 (361) who had a first consultation also have a second consultation, but 86 of the 125 (692) go to a pharmacy. 4/ Of the 285 who had a first consultation, 171 (602) go to a second source, but 146 of the 171 (851) go to a pharmacy. - 97 - Annex M Page 1 of 3 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Past Economic Research on Demand 1/ In the early 1960s, economists first became interested in estimating the demand for health services. The single-equation demand models which they used came from a simple utility-ma -imization model of consumer behavior. This model predicts that if people derive utility directly from consuming medical care: 1. The demand for medical services will depend on the price of that service, other prices, income, and tastes; and 2. Own price elasticities will be negative, cross-price elasticities will be positive for most goods, and, if medical care is a normal good, the income elasticity of demand will be positive (Joseph, 1971). Acton (1973) altered this model slightly to include combined waiting and travel time as a noncash component of the cost of purchasing medical services. This alteration left intact, however, the model's major theoretical shortcomings. The model excludes demographic variables such as age, education, and health-status except as controls for 'tastes,' when, in fact, it is well knovwn that these variables are independently important as determinants of health service use. The development of the human capital and household time allocation models enabled economists to model the effects of these noneconomic variables rigorously. Holtman (1972), and Holtman and Olsen (1976), attempted this in a one-period utility-maximization framework, but the seminal work in this area was done by Michael Grossman (1972). He modeled the demand for medical services as being derived from the desire to invest in a stock of healthy days, or health capital. Age was used as a proxy for the depreciation rate of the stock of health and education was used as a proxy for the relative efficiency of different people in combining inputs (home care, nutrition, medical services) to create added healthy days. Thus, hypotheses on two demographic variables, age and education, were added to those supplied by the simple Acton model. Phelps and Newhouse (1974), added insurance coverage to the Grossman model; Leibowitz and Friedman (1979) followed the Grossman theme in considering parental investments in child health; and Goldman and Grossmar, (1978) tried to model physician quality as a separate component in the provision of health services. Nevertheless, the Grossman investment approach still stands as the most complete theoretical treatment of the topic, at least for high income countries. 1/ This review is adapted from Akin, Guilkey and Popkin (1981). - 98 _ Annex M Page 2 of 3 Peter Heller (1976) made one of the few attempts to model health services demand in low income countries. His theoretical model was similar to Acton's 1975 version, but Heller distinguished between the necessary and discretionary components of medical service use. Akin, Guilkey and Popkin (1981); Akin, Griffin, Guilkey and Popkin (1985, 1985b, 1986, 1986b); Wong, et al (1987); Schwartz, Akin, and Popkin (1988) have done a series of health and contraceptive demand estimations in developing countries. In several of the papers money prices, at the levels observed in the sampled areas, have not been found to be statistically significant predictors of health care provi6er choice; but in others (see Schwartz, Akin, and Popkin, 1988; Akin and Schwartz, 1988), higher prices have been found to be statistically significant factors in causing reductions in specific types of health service usage. The time price variables have been similarly insignificant in several of the analyses, but, as with the money prices, there are exceptions. For example, both money and time prices are found to be negatively related to use of public prenatal care in the Bicol region of the Philippines (Akin, Guilkey, and Popkin; 1981), and to be similarly related to the use of child delivery services in urban areas (money prices) and rural areas (time price) in Cebu, Philippines (Schwartz, Akin, and Popkin, 1988). The authors speculate that many of the non-significant price results are probably caused by the inability in the estimations to sufficiently control for the quality of the care provided. Because higher prices and higher quality tend to be closely related, the marginal negative impacts of higher prices may be negated in the data by the added quality of the care of the providers who charge the higher prices.2/ Probably the most important technical improvement of the Akin et al. work was the use of exogenous prices (rather than endogenous expenditures) as the explanatory price variables. Even though basic economic theory suggests that demand relationships must be estimated with such exogenous price data, other authors have been forced by data limitations to such efforts as the use of actual expenditures of patients for actually chosen care choices as a price proxy (Gertler et al., i987), or even to estimating money price relationships from data with no money price information at all (Dor et al., forthcoming). The other important distinction of the research of this group is that the dependent variable has been carefully specified to be as close to a homogeneous good as practicable. The papers estimate separately the demand for such goods as prenatal care visits, adult outpatient services, and infant delivery, while others in the literature have tended to lump together many kinds of health care, and define the good (the dependent variable) in a very broad manner, such as private health care, public health care, or no care, without sufficiently controlling for what kind of care is needed (i.e. what the good or service "health care" demanded by 2/ The results of this Nigeria study certainly imply that added quality can overwhelm the negative impacts of the price increases instituted to raise revenues for providing quality care. - 99- Annex M Page 3 of 3 each respondent actually is). This lack of control for "need' causes the model to be for choice across options which provide very different services. It is only when the basic health care being provided is the same across observations (i.e., a homogeneous good) that correct coefficients, showing how such factors as price and quality affect the cri ices, can be estimated.3/ Other recent demand work of note has been carried out by a group including Gertler, Dor and Van der Gaag (Dor et al., forthcoming; Gertler et al., 1987). They have analyzed data from Peru and the Cote d'Ivoire and found significant negative price impacts on health care choices. The elasticity results derived from the findings are much in line with those of other micro-data demand studies in developing countries. For Peru, for example, they estimate that for up to a 10-fold increase in the 'price' of public hospital care the usage reduction response would be inelastic. Only for the poorest quintiles in the country, and then only when public hospital prices are raised 20 fold (i.e., 200%X) do the results suggest that price responses become elastic. The most important aspect of the Gertler et al. work is the emphasis on the possibility of differences in price responsiveness of households at different income levels. Other work had controlled for income levels in estimating price responses, but this work was the first in which income and the price proxy were interacted. Whether price responses will change at different income levels can only be answered empirically for each specific population, so testing for interaction effects as has been done here is important. 3/ In this research, we are forced by the small sample size to use a similar three outcome categorization of the dependent variable, but control for the differences in what the consumers actually demand (purchase) by including control variables for the actual symptoms that cause the need for health services. - 100 - Annex N Page 1 of 6 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION The Economic Model 1/ For this Ogun State demand analysis, we model a very specific type of behavior: the choice by households of the type of medical services to purchase for members who have a perceived health need. These individuals may choose either to consult with someone or to seek no outside assistance (i.e., use self-treatment). This is a utility-maximization model in which the type of medical service enters directly into the household utility function. While the home production of health approach might prove useful for certain cases, for this situation of actual use of medical service by the ifl, the direct approach of assuming that the health care service itself is valued is not only less complicated but also probably rore realistic. The constraint function includes time costs for using each type of medical service and excludes time costs for non-health consumption activities. The assumption is that with health care services, time spent for consumption is not a leisure activity but an actual cost of consumption. Consumers of health care are assumed to obtain welfare from the service itself but not from the process of consuming; the opposite assumption is made for consumer goods--that the consumption activity is pleasurable--and therefore the time spent in consuming other goods and services is not included as a cost of consuming them. We focus on the economic determinants of the choice of type of medical service when the population is limited to people who are ill (or report themselves to be ill). Thus it leaves out of the calculus the determinants of illness itself, on the assumption that economic variables more strongly affect the decision of what to do after the onset of illness than they affect the actual process of getting sick. The sample is limited to those families which actually report ill members. For families who report an illness during the analysis period, the choice of medical services can be modeled in the utility-maximization framework as follows: 1/ This description of the economic model is adapted from Akin, Guilkey, and Popkin (1981). - 101 - Annex N Page 2 of 6 U - U (privatet public; self-help ; other goods and services) (1) (PR) (PU) (SH) (GS) subject to the cons'raint: Y, (Pp + W,1w)PR + (Pu + WIM)PU + (P4 + WISH + PSMGS (2) m YN + WIT where: Y, income of the household unit from earned and unearned sources; Pi * money price of the it* good or service (i = PR for modern public health services, PU for modern public health services, SH for self help, TR for traditional health services, and GS for other goods and services); WI a wage rate of ill person, or, for a child, the mother, who is assumed to accompany the child; YI4-=nonlabor income of the household; - total productive time spent for the it' good or service; and T - total production time in all activities. The simplification of assuming the household is a single decision- making unit is followed. The major goal of this research is to understand the factors determining the decision, not to understand in detail how the decision is made within the household. In effect, it is assumed that the factors that affect this particular set of decisions affect different households similarly irrespective of the household decision-making process. When this model is manipulated the marginal conditions for equilibrium are essentially identical to those for other constrained maximization models. At the margin, the costs of using both time and other scarce resources (represented by money) in consumption must be equated with the average benefits of the consumption activities. If the marginal benefit per dollar of any of the activities is not equal to that for each of the others, the household will be expected to shift resources among activities. Resources (time or money) will move away from activities which provide less added benefit per unit of the constrained resource spent toward those activities which provide more added benefit per unit. In other words, cheaper activities will be substituted for more expensive ones. Assuming that (at least on the average) households will behave in such a logical maximizing fashion, a set of demand equations can be derived for the relevant medical services. The other goods and services are ignored as a dependent variable for estimation reasons because, for this model, the residual funds and time will be spent on consumption of this bundle of general consumer goods (which includes leisure). - 102 - Annex N Page 3 of 6 Manipulation of the model leads to a set of demand-for-health- services equations which can be empirically estimated: O,- FO(Ppj PFR4 POP S W- tF4, tpR* ts,4 YN, T. 2); i (PU, PR, TR) (3) wheres Qi quantity demanded of medical service i Z - a vector of socioeconomic variables (e.g., urban residence). Because of the difficulty of obtaining accurate household survey data on the extent of use of each type of service during a limited time, this demand system is modified to a system of equations with dichotomous (yes- no) dependent variables. For estimation purposes, this results in a system of possible choices of health care (one for each type of medical services and one for using no outside medical services), each of which is specified to take on only the values of 0 or 1. Because of the nature of these dependent variables it is inappropriate to use ordinary least squares regression procedures for estimation. The system is well suited, however, to estimation by the multinomial probit procedure. Multinomial Probit Estimation Method The statistical method used in this research is multinomial probit. The multinomial probit model is a random utility model where individuals make choices from a set of mutually exclusive alternatives. In this particular application, there are three choices: no consultation (self- treatment), public facility (hospital or clinic), and private facility. Individuals make a utility valuation for each choice according to the following formulas: Um - X;,1 + Zaa + CIx U12 - Xi88 + Z18 + gl2 U8- X ls + ZISI + Cg. where U1J represents the utility valuation that individual i (i-1,2,...,N) gives to choice j where J-1 is no consultation, J-2 is public facility, and J-3 is private facility. The X's are characteristics of individuals or households which do not vary by the health care choice--household income and the individua''s level of education are examples. The B's represent weights given these household characteristics in making the choice--note that they are allowed to vary across choices. The Z's represent characteristics of the choices, the price and quality of the care in this case, and the l's are the weights associated with each characteristic. - 103 - Annex N Page ' of 6 Up to this point the model described could justify the use in estimation of either a multinomial logit or a multinomial probit model. The difference in the logit and probit models relates to the assumptions made concerning the error terms. The multinomial logit model assumes that the errors are generated from a negative extreme-value distribution. This assumption restricts the correlation between the three errors in our model to zero. This assumption of the multinomial logit results in the well known Independence of Irrelevant Alternatives (IIA) problem. The IIA problem is due to the fact that the relative probabilities of any two choices are unaffected by the addition of another choice (or alternatively that the relative probability of two choices of three will not change when the third choice is no longer available). Predictions from the multinomial logit model result in an overly large joint probability for choices that are similar (see Maddala, 1983,. In our application, if there were any IIA problem, the result would probably imply too high a nredicted probability of using either public or private facilities. Because all relevant choices are explicicitly in the model, however, we doubt that any IIA problem is present. The multinomial logit model has been generalized in several directions--see Maddala (1983) for a review. McFadden (1973) discussed the usefulness of the generalized extreme-value distribution, which allows some of the errors to be correlated. For example, for the three choice situation it is possible to allow two of the choices to be correlated. In our case, if we used this estimation technique, it might be desirable to allow C2 and C3 to be correlated. Unfortunately, the other two possible correlations must be restricted to zero. A nested multinomial logit, which is equivalent to this generalized logit, has been used by some authors (Gertler et al., 1987). The multinomial probit model, on the other hand is much less restrictive than the above described models. Its generality, however, does have a high price in terms of computational cost. In fact, some type of approximation must be used if there are more than four choices for the dependent variable. Since we only have three choices in our model, we feel that the increased computational cost is well worth the added generality gained. The error assumption for this model is that the Z's are generated by a multivariate normal distribution, with mean vector zero and 3 x 3 covariance matrix C. All possible correlations among error terms are allowed and the IIA problem is completely avoided. Some normalizations still must be imposed to make the model estimable (see Terza, 1985, for a complete discussion). We follow Terza's conventions for the normalizations. They are defined as follows: - 104 - Annex N Page 5 of 6 Tit xi [12-B1) / # + (Zi2-Zi)da / l Ti2 = Xi(U3-B1) / + (Zia-Z11)(a / Io TS= V20 VO T4 =Vso / V10 where: Vl = -1l -r.2 - 4r3 + 023 V2 = OU22 + 012 - 0'13 + a23 VS = O33 + al2 - rl8 - 023. Given the normalizations, the probabilities associated with the three choices are: -~~~~~~~~~ -TVj _T2i p., = f T :i, .; } V l-T3 (l+T4) - T1; -(T2-T) 1 P;2 = i ; VL- 2V: Z-Ta4-Ta. Pis = 1 - P-l _ Pj2. If we define Y,, - 1 if individual i makes choice j and = 0 otherwise, the likelihood function for the observed random sample is: N 3 L = ir it P;; il J=1 The likelihood function can be maximized with respect to (B2-r1) / v1 (133-B) / v I/ s T and T4. Note that setting B1 = O is the standard normalization imposed in the multinomial logit model and that dividing by ^v1 is simply division b, a positive constant which is the standard normalization in probit models. Thus, the coefficients of the X variables can be interpreted as affecting the relative probability of - 105 - Annex N Pase 6 of 6 either choice 2 or 3 to choice 1. a / ,-is simply the normalized weights given to choice specific variables (the Z variables) in the utility function. T. and T4 contain all of che variances and covariances of the disturbance terms which are all allowed to be non-zero but are not separately identified. The logit specification requires the disturbances to be independently and identically distributed. Thus, a test of the IAM property would be a test that T. = -1 and T4 - 1. We perforL such a test in the empirical section of this paper. The program used for estimation was written in FORTRAN using the Lahoy P776-EM/32 compiler. The DFP algorithm in GQOPT was used for mYAiLaization of the likelihood function. Analytic first derivatives of the log likelihood function were used and their use resulted in a considerable saving of computation time. The results presented required six hours of time on a 16 MHZ Compaq 386. The corresponding logit model took 30 minutes. - 106 - ANNEX 0 Page 1 of 9 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Household Demand for Outpatient Services 1. Choice-Based Sample. Given the current situation of underutilization of facilities, usage is an issue of great concern to Nigerian policymakers. For the actual demand estimates, the choice-based methodology was used to weight the choice-based ',chose to use a facility) and community-based (neighborhood) samples to create a combined representative sarl:' of potential users with access to a facility.l/ This procedure utilized the extensive information on sick people obtained from the facility-related, choice-based survey, but allowed the total sample to be weighted in a manner that makes the results representative of the population of potential facility users. Policymakers who are concerned about current and probable usage of existing facilities can use these conclusions for guidance in facility use and pricing de^isions.2/ 2. The dependent variable consisted of individual's choices among three methods of treatment: (a) self treatment; (b) public facility; (c) private facility.3/ The frequencies for the dependent variable in the analysis sample were as follows: Sample: Choice-based Community-based Self treatment .3153 .5155 Public facility .4943 .3422 Private facility .1904 .1423 ~/ The orliginil respondents were eslected because a household member had obtained care at a sampled facility. Interviewers selected choice-based households by meeting patients at t.ealth facilities and making a follow-up Interview visit. The *election of facilities and petients wes random. Making an appointment for a follow-up visit allowed people to prepere better for the time and effort Involved in providing detailed answers to the lengthy questionnaire, guaranteeing a large number of completed interviews for health sorvico users. Neighbors wore aloo interviewed. In taking Illness histories, respondents were asked 'What did you do first--, second--, third-- ,e so that they had at least three opportunities to report using various facilities. No evidence suggeats that sampling households near a facility (who may or may not have used that facility) does not provide a random samplo of possible users. 21 However, the concluoior.s cannot bo oxtonded with full confidonce to health care usage In gennerl by the total population, including those who live in areas relatively Inccessible to the facilities. 3L/ Individuals were categorized depending on their answers to a series of questions. The fire$ question was "Have you been Ill In the poet four weeks?0 Those who were not sick were not asked follow-up questions. Those who were sick In the last four weeks were asked what they did first, macond, and third to treat the Illness. The date set contains answoer to the 'do first' and "do second' questions. The dependent varlb!e was defined as * *lf-troatmente only If they answered self-treatment to both questione. The dependont v&.,s.ble was defined as 'public facility' If they went to a public focility first or If they went to a public fcilIlty second after responding self- treatment to what they did first. The dependent variablo was defined as sprivate faellity' In a similar manner. If respondent answers Indicated a combination of public and privato facility use, what they did first vas used to define the dependent variable. About 06X of the time this algorithm for defining te.. dependent variable resulted in using the ansoer to the Odo first' question. Only 781 of the 8,186 who pursued a first source (which includes seeking self-medication) answered that they pursued a second course of actlon for treatment of the il other than a phor - 107 - ANNEX 0 Page 2 of 9 3. Estimation Model. The empirical model employed assumes that choice of health care is a function of: (a) price of the care (b) quality of the care (c) sex of the patient (d) education of the patient (e) wealth (assete) of the household (f' income of the household (lowest 20 percent) (g) urban residence (h) symptoms of the illness (i) seriousness of the illness Price and quality depend on the health care chosen (conditional variables); other variables are specific to the household (or individual patient) and do not vary across health care choices (unconditional variables). A discussion of the actual specificati'.a of these variables follows. 4. Health Care. Type oL health care chosen is classified as one of three alternatives: public care from a clinic or hospital, private care from a clinic or hospital, or self treatment.4/ Those included in the analysis suffered from cough, fever, stomach problems, diarrhea, weakness, headache, guinea worms, burns, bites and stings, poison, or falls. Table 1 lists all variables, their means and standard deviations. 5. Prices 5/ for child delivery, maternity registration, outpatient registration, and inpatient room and board were available, but outpatient registration was selected as the best proxy for relative facility prices. The assumption was that the charge for outpatient registrations tends to be high at facilities where prices for most services are high and low where they are low. Also, most ill or injured persons choose a facility expecting to be treated as an outpatient. Outpatient registration probably The care could have been either in- or outpatient. Lack of sufficient observations for each illness type prevented estimating these choices separately for groups of individuale suffering fro nrrowly defined pecific ailments. The relationship was therefore estimated for all observations In the msmple reporting illness or Injury except those reporting that they suffe'Nd from leprosy or from other' (the CIas in which the data coders seem to have Included pregnant women along with people suffering free various and sundry other ailments), who were dropped from the analysis sample. 6 This study hod lnfnrsation from sampled facilities in each gogroaphic area on the prices charged for several specific types of treatment. For empirical estimation, the beet proxy will show the price housholds perceive they will pay when they are Ill or Injured and are making their decision of what typo of facility to visit. In practice, the proxy must order the patient's choices by relative price, giving the variable a large volue for choices that iotential patients will expect to hev high prices and a Mall value fur those that they wIll expect to have low priesm. - 108 - ANNEX 0 Page 3 of 9 best reflects the price that patients consider when they are deciding whether and where to go for treatment.6/ 6. Quality. The best single indicator of relative quality is probably facility operational costs per capita.7/ Since per capita expenditure by facility is not a perfect measure of health care quality, two specific observable quality indicators were added: (a) observed physical condition of facility; and (b) percentage of the year drugs are available at the facility.8/ 7. Income 9/ enters the model as a control for tastes and preferences. However, income interactions with other variables, such as prices, are not included in the final estimated model because they did not prove to be statistically significant.9/ 9/ The prices for outpatient registration for all public and all private facilities were averaged in each geographic aroa to determine a price for privaot and public care in each health care merkot. For geographic units in which no facility of a given type is located, the prices In all neighboring areas were averaged. For "self-treatment,g we assumad a money prico of zero. The result Is a set of money prices, determined from actual charges at facilities In the market, and obviously exogenous to the health care cholcos of the users. Such an exogenous variance shoull provide a good proxy for the private and public prices that consumers face, because the facilities in their local area are available to them. It also appears from the date that this accounts for the major part of the total cost of outpatient caro, excluding drugs separately purchased. y1 Several alternative quality proxies wero available, such as functioning X-ray machines and leboratories, support personnel plus nurse per capita, and doctore per capita. However, the variables of total expenditure by the facility per person In the population oerved can be created because the facilities supplied information about what they spend on drugs, oupplies and personnel. For actual estimation, those facility- specific, expenditure per capita numbers are avoraged for all public and private facilities in the local government area to get a market-area Index for both public and privote expenditures per person served. V These measures are aleo averaged over the relevant facilities In eaeh local governmental area. V/ Most conomic models in the literature are consistent with the noncontinuous dependent variable logit and probit technique and with the basic economic assumption of utility- maximizatlon result in the Income teo from the budget constraint not being In the theoretIcal demand functions. Practically, this causes little problem for ompirical research. The Income term can appropriately be Included In the estimated model as a shift parameter Por tastes and preferences that differ by Income group. The Income teor con also be Interacted with such choice related (conditional) variables as price, If the modelor believes (or wishes to test) that responses to price differ by Income closs - one specific mannor in which tasto and preferonco dIiforencos could have their lFpacts. In #A9 moc6ld whfre the only way prices and Income enter Is through a net ' of t ~ ' var' , prico nor - 109- ANNEX- Page 4 of 9 Table 1 MEANS AND STANDARD DEVIATIONS OF VARIABLES IN LOGIT AND PROBIT MODELS Variable Mean Standard Deviation Education of ill person 1.5820 1.5933 (index) Sex of ill person 1.6188 .5131 Own a vehicle (motor) .2535 .4352 Income in lowest 202 .6007 .4899 Urban residence .5394 .4986 Days of reduced activity 4.2972 5.3736 Stomach or diarrhea symptoms .1299 .3363 Price no consultation .0000 .0000 Price at public facilities 1.1962 .2693 Price at private facilities 2.2223 1.4337 Quality (a) Expenditure on care per person in population served No consultation .0000 .0000 Public facilities 3.0416 1.4016 Private facilities 1.3571 1.0304 (b) Drug availability (percentage of time available) No consultation .0000 .0000 Public facilities 71.8735 13.7524 Private facilities 99.5824 0.6248 (c) Physical condition of facility index No consultation .0000 .0000 Public facilities 1.2867 0.3064 Private facilities 1.1634 0.3589 Source: Calculations based on data in Ogun State Health Financing Study. N = 1,763 8. Wealth. In Ogun State, vehicle ownership is a good variable for ordering households bg wealth status. Combining this variable with the lowest 20 percent income variable results in good control for income levels in the model. The incc!e variable differentiates the income poor, who appear to behave differenitly from other income groups in selecting health care, and the specific wealth variable appears to differentiate those with high incomes (that is, those who can afford vehiclas).10/ oj 1 We also Introduced several asset voriables, bosides vehicle ownership (the one asset variable that remine In tho fInal model). The other variables performed le1 woell then vehicle ownership and were not sttistically significant when entered in combination with vehicle ownership. Among the other asset variables tested ware own house, rnt house, own mnimle, own land, tol let In the reldence, and piped water In the - 110 - ANNEX 0 Page 5 of 9 9. Tvve of illness.ll The only two symptoms that were statisticalli significant predictors of health care choice in the model were stomach pain and diarrhea. Statistical tests indicated that the two could be combined into one variable because of the similarity of their impact. The remaining symptoms variable is, therefore, 'suffered stomach pains or diarrhea." 10. Severity of illness. Choices of care obviously tend to vary for those who have a minor health problem and those who are more seriously ill. The control for severity of illness is the variable "number of days unable to carry on usual activities because of the health problem." 11. Urban or Rural Residence, Education, and Sex. A variable for urban or nonurban residence is included to control for taste differences between urban and rural residents. In addition, controls for "education level of ill person" and "sex of ill person" were included to control for differences in choices of health care across education levels and sex. Education affects consumer tastes because of added knowledge about options; in many societies health care choices by households differ for male and female members. (b) Estimation Results 12. Table 2 presents results for the multinomial probit estimations. The multinomial logit results, estimated for comparison both with past research and with the probit results, are very similar to those for the profit and are in Table 3. The results strongly support the views that (a) price affects health care choice but has an impact that is small in magnitude; (b) quality as measured has a large impact on health care choices; and tc) higher prices can be offset by higher quality. 111 Becuse the dependet variable Is type of facillty chosn for helth care, coefflcionts estimated on variables such as price and quality have little or no validity unlew patients suffer from similar health problem. Because the number of observations for any given illneoo or injury tegwory was saoll, It was necessry to control for the type of health care purchase by including right-hand side control varlObles for symptom reported. All syptoms included In the survey were tested - cough fever, stmach pain, diarrhea, an, headche, guinea worm, burns, bites ond stings, polson, and fells alone and In various combinations, depeding on their frequencie. - 111 - ANNEX 0 Page 6 of 9 Table 2 MULTINOMIAL PROBIT ESTIMATION RESULTS Public Private Public versus versus versus Self Treatment Self Treatment Private Variable Coefficient fft! Coefficient "t" Coefficient "to Conditional Price -.029 -1.68 -.029 -1.68 -.029 -1.68 Quality: (a) Expenditure on care per person in population served .029 1.56 .029 1.56 .029 1.56 (b) Drug availability (percentage of time available) .015 3.95 .015 3.95 .015 3.95 (c) Physical condition of facility index 1 = good 2 = fair -.227 -4.54 -.227 -4.54 -.227 -4.54 3 = poor Unconditional Education of ill person .022 0.83 .022 1.97 -.001 -0.03 Sex of ill person (male = 1; female = 2) -.001 -0.01 .064 1.70 -.064 -0.91 Own a vehicle .197 2.14 .112 2.67 .084 1.09 Lowest income quintile .163 1.47 -.004 -0.08 .166 1.76 Urban residence .010 0.09 .183 2.90 -.174 -1.65 Days of reduced activity .028 3.33 .013 3.54 .015 1.94 Stomach or diarrhea symptoms .557 4.27 .243 4.34 .314 2.59 Constant -1.679 -4.17 -1.579 -4.38 -1.00 -0.51 Source: Calculations based on data in Ogun State Health Financing Study. - 112 - ANNEX 0 Page 7 of 9 Table 3 MULTINOMIAL LOGIT ESTIMATION RESULTS Public Private versus versus No Consultation No Consultation Variable Coefficient "to Coefficient ate Conditionals Price -.137 -2.39 -.137 -2.39 Quality: a) Expenditure per person in population served .054 1.43 .054 1.43 b) Drug availability (percentage of time) .012 3.02 .054 3.02 c) Physical condition of facility index -.262 -1.82 -.262 -1.82 Unconditional: Education of ill person .031 0.82 .081 1.66 Sex of ill person .030 0.27 .473 3.01 Own a vehicle .250 1.94 .462 2.78 Income in lowest 202 .144 1.05 -.244 -1.08 Urban residence .072 0.63 1.328 7.49 Days of reduced activity .033 3.12 .035 2.56 Stomach or diarrhea symptoms .600 3.60 .704 3.27 Constant -1.403 -3.25 -4.080 -7.45 N - 1,763 Log Likelihood - 1725.74 - 113 - ANNEX 0 Page 8 of 9 13. Raising prices with constant quality will reduce usage; but spending the increased funds to improve quality may even increase usage. Given equal prices and identical tastes, people prefer health facilities that have (a) higher per capita spending on care; (b) higher drug availability; and (c) a relatively good physical condition.12/ Efforts to include a price of time in the model are detailed in Annex 0. The price of time variables were not found to be statisticaAly significant. The strength of the results on the money price and quality variables is reassuring. In all estimates and with various combinations of variables other than quality and price included, the price and quality variables almost always had the expected signs and were highly significant statistically. In addition, the results indicate that the lowest income quintile is more likely to choose public care than other income groups but behaves much like the other income groups in choosing between private care and no use of formal care (self treatment).13/ 14. Tastes and prefer'*nces of individual clients are surprisingly strong statistically in their effect on choices. The wealthy (those who own a vehicle) seek the most medical care. Urban residents, females, and the more highly educated prefer private medical care. Those with (a) stomach or diarrh-a problems, and (b) more severe ailments, tend to seek treatment at public facilities, assuming equal prices and quality. L2 The price of tho specific typo of care Is otatistically significant at the 10% level (ta-1.68) and has the expected negative sign. As the price of a care option Increases, If the quality of care otatistically l held constant, the likolihood of that option being chosen will decrease. Thus, higher prices will reduce u8sge, as expected. But because the probit relationships aro Inherently nonlinear, th, magnitude of the effect can only be simulated by estimating tho predicted usage of each type of care for different values of price. Two of the quality of care varloble. are highly oignificeant; the third Is significant at a low level (expenditures per capita for the population served only has a t value of 1.58). All have the expected signs. The multinomial probit model effectively estimates parameters of the utility function for oach outcome. The parametors on the conditional variables are the ome for each outcomo and those on the non-conditional variables differ. Because price and facility chroctoristices (total exponditure per person served and facility physical condition) obviously are not relevant variables for the seolf treatment' choice, we use zero for tho value of each of these variables when Osolt treatment5 Is chosen. This In effect causes these factors to be omitted ("zeroed outs) in the utility equation for the "self treatment" choice (i.e., the amount of utility one receives from 'self treatment' ls not affected by tho physieal characteristics of facilities). This Is the effective woy to handle the Impacts of ouch nonexistent factors as the physical condition of tne facility which provides the Oself treatment' choice. g/ When feas are introduced, even If health service usage is not greatly reduced, financial burdens on the poor wall be increased. Practical methods to Identify the poor so as to exempt them from tfes or to provide them subsidies do oxiet. Unless such meane testo are utilized and the poor effectively protected, the social cost could be largo. Possible approaches to ldentificatlon of the poor used In developing countries include cortification by citizon committoe or local officials. Ovwr time, as tax rGeords and other Information become more reliable, more precise methods of identitfcation of the peor should become available. Another possibility which would not necessarily requiro such good information Is a self-declaration eyetem with specific incomo and wealth standards. Substidization could Include charging small amounts or nother for 'no-frille' lOvels of room and board In hospitals but full c')et for less crowded rooss, botter food, etc.; or subsidixing care for specific population groupo (for oxample, infonte and pregnant sowen) or regions of the country. - 114 - ANNEX U Page 9 of 9 15. These results suggest that if public facilities offered private sector levels of both drug availability and physical conditions (assuming constant public and private sector spending per patient), they could raise outpatient prices to the level of the private sector (an 87 percent increase) and still increase usage. Some current nonusers would be drawn to the public sector, and around half of present private care users would also switch. This interpretation of the quality/price trade-off obviously oversimplifies the problems that would be encountered in an attempt to actually make such changes. Yet the results strongly suggest that if revenues from higher prices at public facilities were used to increase cea:o quality, usage of these facilities (and of modern formal care in general) might well increase. - 115 - Annex P Page 1 of 2 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Calculations of Time Price Time Price: The time cost of traveling to a facility and receiving care is often a big part of the total cost of care. Unfortunately, its value could not be accurately verified. The model considered (a) distance; (b) travel time; (c) transport price; and (d) time cost, as components of time price. The time cost results perversely indicated that the longer it took to travel to a facility, the more ikely the individual was to choose that facility. The implication is that omitting time cost from the model will not bias the price and quality estimates, and it was omitted from the final specification. Because the data set does not contain information that allows distances to a:ternative available specific facilities to be determined, creation of a distance or time cost variable is problematic. We took the reports of distance and travel time to the facilities actually used by all individuals in each local government area, and averaged them over that geographic area for private and for public facilities to create the four variables: average distance and average travel time to public facilities and average distance and average travel time to private facilities. We then attributed values for these four variables to every individual in the sample on the basis of the local governmental area in which they reside. We also interacted each of the travel time variables with per capita household income for each individual to obtain time cost values (assuming per capita household income is a reasonable proxy for hourly wage). We also created average transportation cash price variables for each local governmental area. We alternatively estimated the full logit model with the variables: 1) distance, 2) travel time in minutes, 3) cash transportation price and 4) time cost added as explanatory variables. The cash price of transportation was insignificant (+ = -.14), while time cost, distance and travel time each were strongly significant and of the wrong sign to be picking up the impact of time costs. The time costs results, if accepted would indicate that, ceteris paribus, the longer it takes to travel to a facility, the more likely the individual is to choose that facility. The time cost variable obviously either is proxying for something other than time cost, or the endogeneity problem inherent in using distances and times reported by those who actually choose a specific facility causes the variable to be of little value as a proxy for the needed exogenous time cost variable. It should be noted that when time cost is included in its most theoretically correct form, time multiplied by per capita household income, the coefficient on price remains strongly statistically significant and negative, and its magnitude changes only marginally (from -.137 to -.154). The coefficients on the quality variables are also very stable to the addition or subtraction of the time cost variable: when the time cost variable is included the coefficient on expenditure per person in the - 116 - Annex P Page 2 of 2 population changes only from .054 to .058 and that on drug availability remains at .012. The coefficient on the physical condition variable does become Insignificant (t = -.1.01) and is reduced from .262 to .148, an indication that distance is perhaps proxying for physical condition. - 117 - Annex Q Page 1 of 2 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Income Variable Incomes UntLike the price and quality variables, the income variable is allowed to affect different choices differently, i.e., it is not a conditional variable. The reasoning is that relative tastes for public and private care and no consultation will differ by income group, even when both the prices and qualities of each health care choice are statistically controlled to be equal. As we mentioned earlier, the lowest 201 of the income distribution appears to behave differently than the others. We tested all estimations using dummies for all r-4intiles, as well as a dummy only for the lowest 60Z, but it was only with a formulation using a dummy variable for the lowest 202 that the estimated coefficient was significantly different from those for the other income groups. The lowest quintile (20Z) income group (measured as household income per capita) is found to be more likely than other income groups to choose public care but essentially equally likely to choose between the alternatives of private care or no consultation. The results indicate, therefore, that the lowest income people, when ill, are more likely to choose public care than other income groups, but behave much like the other income groups in choosing between private care or no use of formal care. The choice of public versus no consultation is statistically significant at a relatively low level, (t - 1.47), but the preference of public to private by the poor is highly significant (t = 1.76). That the t on the coefficient on the income variable for the choice between private and no consultation is almost zero (and, in fact, that the coefficient itself is very close to zero), provides strong statistical support for there being little difference in preference for the two, by different income classes, when the price, quality, etc. are controlled to be equal. The results taken as a whole suggest that if the care were perceived as more or less equivalent, lower income groups would still have a relative preference for using public facilities. These income results are certainly plausible. That poorer residents are more likely than the less poor to choose public care relative to the other two options, all other variables held constant across the choices (or that the richest 80 tend to use both self care and private care relative to public care more often than the poorest 20Z), could be due to various causes. That the poor may feel more comfortable at public facilities, and that the less poor may both have a strong belief that the private care is more effective and may be better able to treat themselves (and more confident in doing so), are some plausible reasons. An alternative could be that our model does not perfectly control for quality differences across the choices and that the less poor tend to choose private care, or even no consultation (self-treatment), because they realize that this alternative - 118 - Annex Q Page 2 of 2 (these alternatives) is (are) of higher quality than public care, in ways that we are unable explicitly to control. This quality control explanation seems less plausible, however, when it is extended to the implied relative preference of the less poor for self-care over public care. The data do indicate, however, that income differences do lead to different health care choices for the members of the sample, even when the price and quality of care (as controlled) are essentially the same. It would be incorrect to interpret these income results as indicating relative ability to afford the different choices. As mentioned above, we interacted the various alternative income variables with price and quality, both with and without the non-interacted income variable also iucluded in the equation. In no case, even when no direct income variable was included, was either the income times price interaction term or the income times a quality variable interaction term statistically significant. For this reason, we concluded that the direct lowest 20Z income term best explains the differential behavior by income classes of the Nigerian sample, and included only that variable in the final model.l/ I 1/ These findings stand in interesting contrast to those of Gertler et al., who found statistically significant income interacted with OpriceO results, in a model in which neither price nor income was included in the equLtion in a non-interacted form. - 119 - Annex R Page 1 of 3 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Outline of Principles for a Cost Recovery Program I. Preamble A. Ultimate objective of Ministry of Health: improve health status B. Means to objective: more resources per ill person C. Constraint: additional tax revenues unavailable D. Means around constraint: 1. Reduction in number of ill persons, through, a. prevention b. family planning 2. Reduction in health-care costs, through, a. improved management practices b. provide incentives for patients to seek cost-efficientt kinds and levels of care 3. User charges (cost recovery), insurance plans and community contributions II. Cost Recovery A. Fee structure and levels of fees 1. Generate revenue sufficient to improve quality 2. Provide incentives for individuals to seek an appropriate level of care, by using higher fees at higher levels of the system, but waiving the tees if a patient is referred to a higher level of the system 3. Cover an episode of illness 4. Fees to be set at zero for preventive services, and for children under some age (perhaps, under age 5) 5. Permit subsidization of the care of the poor by the rich 6. Permit the poor to obtain access to care without further impoverishment 7. Prices similar at HOH facilities and at facilities operated by other levels of government - 120 - Annex R Page 2 of 3 B. Retention of cost-recovery fees 1. A significant proportion (50 to 75 percent) of user fee revenues to be retained by the health-care facility collecting them, in order to improve the quality of care at the facility 2. Before such revenues are spent, each facility, in consultation with the MOH, will establish a list of priorities for improving quality, that are visibly linked to services being purchased by patients (e.g., drug availability, waiting time, etc.) 3. The retention of revenues by the facility will not result in any decline in MOH financial support to the facility (nor, in any decline in MOF support to the MOH), and will thus result in an increase in total funding for the facility 4. The 25 to 50 percent of revenues that are not retained by the facility will be used to increase the funding of primary and preventive programs and to improve the quality of care at facilities located in poorer areas. 5. Over time, the proportion of revenues retained by the facilities may be adjusted with regard to progress in improving quality C. Implementation 1. Increases in levels of fees to be phased over time, in step with quality improvements 2. Immediate and near-term goals for implementation of cost- recovery program a. Gain public acceptance of idea of paying fees in return for improved quality b. Establish a proper structure of incentives for patient use of the health care system and for facility collection of fees c. Put in place administrative procedures for, i. collecting and spending the fees ii. permitting follow-up visits and referrals, without allowing fraud to enter the system iii. ensure that all members of society, especially the poor. continue to have access to care d. Initial fee levels to be modest so that the public will accept them, and the near-poor will be able to afford them &7 - 121 - Annex R Page 3 of 3 e. Experience gained initially will be used in subsequent years to raise fee levels and increase cost sharing, while protecting the poor D. Monitoring of Cost-Recovery System 1. Formal monitoring system is established, in order to: a. Determine how introduction of user fees affects the poor b. Help determine the appropriate future level of fees 2. Obtain and analyze facilitv utilization information and monitor a population-based sample of households on a regular basis -122 - Annex S Page 1 of 32 NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION Cost Recovery in Ghana and Modalities for Collecting New Hospital Fees A. Cost Recovery In Ghana 1. Lessons can be learned from the inteusive study of cost recovery in Senegal, Mali, the C8te d'Ivoire, and Ghana in Vogel (1988) (Chapter VI). Perhaps the most important lesson in that study is that cost recovery seems to be progressing much better in Ghana than in the three francophone countries in the study. The key to Ghana's greater progress appears to be in the structure and application of prices that has been established, and in the administrative provisions of the cost recovery law. 2. The structure of health care prices is such that they are hierarchical and directly related to the sophistication and expense of the health care delivered. For example, the price of curative care at the hospital level is a large multiple of curative care at the health center level, and the price of curative care at the health center level is a multiple of the price of curative care at the health post level. This aspect of the pricing structure gives strong incentives for an efficient and workable referral system. Furthermore, there are no exceptions to the health facilities for which pricing is obligatory; on the other hand, the facilities themselves have the obligation to differentiate poor persons from non-poor persons, and to give free care to the poor. 3. What is also interesting about the Ghana cost recovery -xperience is that the price structure that was established in 1983 was judged to be too low, so that there was a large upward adjustment of prices in 1985. In 1986, total revenue from cost recovery increased, and by 1987, cost recovery as a percent of the MOH recurrent budget had reached 12.1 percent, whereas it had been 5.2 percent in 1985. The assertion that many of the poor were driven from the formal health care system during those years still awaits empirical verification. 4. The second importiant aspect of the Ghanaian cost recovery experience has been the admi.istrative provision that a portion of the cost recovery proceeds sho.ild remain at the site of collection, which should give an incentive to collect fees and to improve quality. The general formula used is that 50 percent of user fee revenue goes to the Ministry of Finance, 25 percent goes to the MOH, and 25 percent is retained by the facility that collects the fees. Whether the 25 percent that is retained v Gertler and van der Gaag (1988) argue that it is more administratively efficient to differentiate the poor from the non- poor by geographic area, giving free care in poor areas, and charging prices in more well-to-do areas. - 123 - Annex S Page 2 of 32 by the facility will give it a strong enough incentive to improve quality remains to be measures empirically. ?JAlso, it will be interesting to see if the 25 percent going to the MOH will be used for quality improvements. 5. What has thus far been observed, though, is that total revenue from cost recovery did increase rapidly between 1983-87, indicating either that the price-elasticity of demand for government health care is extremely low, or that there has been an outward shift of the demand function in - response to positive quality changes holding income and the level of health constant. B. Modalities for Collecting New Hospital Fees v Preamble 6. The purpose of this exercise is to maximize the fees collected without unduly inconveniencing our clients/patients. To achieve this objective all health workers without exception have to put in a little more effort than they have done so far. General 7. Pharmacists and storekeepers are to immediately notify heads of their institutions of all supplies received by their institutions. m 8. Herds of institutions are reminded of their responsibility to sign the appropriate section(s) of all receipt vouchers. 9. Officers in charge of Laboratories are expected to inform clinicians every day or week about tests that they can perform and those they cannot for lack of reagents etc. 10. At regular intervals to be agreed upon between prescribers and pharmacists depending upon the size of the institution, Pharmacists will provide each prescriber with a list of drugs available in the dispensary I with their approved prices. 11. Members of the Management Committee(s) and unit heads should make regular and surprise checks on revenue collectors in order to maximize the collection of hospital fees. Records 12. As usual the particulars of people reporting to the clinic are to be recorded and be issued cards with registration numbers. t Not much is known about the incentive effects of this kind of revenue sharing in development countries. v Source: Ministry of Health, Accra. - 124 - Annex S Page 3 of 32 13. At places where patients, by prior arrangement with the authorities, do take their cards home, total numbers of clinic attendance (new and old) for the day are to be obtained by checking the total numbers seen in the consulting room(s). 14. With the assistance and support of the doctor/person seeing patients, the nurse in the consulting room will enter in a note book to be provided, the following particulars of each patient seen: DATE: CHART 1 REGISTRATION SERVICES NUMBER NAME AGE SEX PROVIDED COSTS 15. At the close of each day, the doctor will close the book and submit it to the Medical Administrator who will use it in checking against monies collected by revenue collectors. 16. The cost of all services rendered to each patient up to the end of the consultations will be clearly indicated on the OPD cards by the doctor/clinician to guide revenue collectors. 17. The nucsing staff are to assist in the recording and collection of fees. 18. Patients will be expected to pay for services received up to the end of the consultations as soon as they leave the consulting room i.e. before they buy their drugs. The services to be paid for at this point include laboratory, X-ray etc. The receipt must then be presented at the dispensary etc. otherwise the prescription will not be honored. The cost of each item and their total costs should be clearly indicated on the prescription. Dispensary 19. Patients will collect drugs from dispensary only with prescription forms after payments made. The prescription forms must be properly kept at the dispensary for inspection at any time. 20. Daily issues from the stores intc the dispensary must be kept and checked against daily accounts. 21. At the close of each day revenue collectors and pharmacist/ Dispensary units will render accounts to management. - 125 - Annex S Page 4 of 32 22. The following records of each patient are to be noteds DATE: CHART 2 REGISTRATION NAME ITEMS QUANTITY COST RECEIPT NUMBER DISPENSED NUMBER 23. A weekly sin=ary of quantities of various items dispensed is to be compiled as follows: CHART 3 ITEM AMOUNT AMOUNT BALANCE RECEIVED DISPENSED Revenue Collecting 24. Health Centers will have one revenue collector each. 25. Region and District Hospitals, Polyclinics and other institutions will have at least two revenue collectors - one just outside the consulting area and the second in the dispensary. Institutions that feel that they need more revenue collectors at any of the two points will apply to their Regional Directors of Health Services for authority. - 126 - Annex S Page 5 of 32 26. Particulars to be recorded by the revenue collectors are: DATEs CHART 4 REGISTRATION NAME AMOUNT RECEIPT NUMBER PAID NUMBER Free Services 27. The particulars are to be recorded like everybody else with "free' indicated at the receipt number column of the above charts 2 & 4. Positioning of Revenue Collector 28. As much as possible, revenue collectors are to sit outside the consulting rooms to accord patients the privacy that they are entitled to. Alternate suitable arrangement can be made as deemed fit. Accounts Recordings 29. Fcr ease of checking the following recordings to be made by the accounts staff at the end of each day: DATE CHART 5 COLLECTION RECEIPT NO. OF AMOUNT REVENUE COLLECTORS POINT NUMBERS PATIENTS COLLECTOR SIGNATURE Revenue Returns to Regional Office 30. Attached is a form designed for Health Posts/Centers for revenue returns to the regional office. Other institutions may use the same form. - 127 - Annex S Page 6 of 32 (Page 9). The totals for each month from the various institutions are to be copied on to a master chart at the regional office. This will show at a glance how the various institutions are performing. 31. The forms are to be filled, in triplicate, at the end of each week and the original copy (of the weekly recordings) with the month's total indicated are to be forwarded to the regional office at the end of each month. The duplicate is sent to the district office and the triplicate kept at the institution. Ward Issues and Recordings 32. Separate books for daily or weekly ward issues are to be kept in the wards and at the point of issue. These issues are to be accounted for by the in-charge of the wards. 33. Ward recordings are to be the same as in the dispensaries. Dressings 34. Dressings are to be issued and recorded by the wards against the names of patients for whom they have been used. This is to be paid for by the patients and recorded in chart. Emergencl Units 35. Fees for services are to be collected by the staff on duty and accounted for to the officer in-charge. 36. Similar arrangements are to be made for weekends and holidays. Accounting for Regional Medical Stores 37. The Supply Officer or Pharmacist should cost and record all items supplied to each institution. 38. The Regional Director of Health Services or the Deputy Director of Pharmaceutical Services are to be satisfied with returns of revenue from each institution as a pre-condition for fresh supplies. In this respect Pay-in slips should be checked. It is recommended that a master chart be prepared at regional level to show at a glance how the various institutions are performing. Services not Specifically Mentioned 39. Serving officers will have to use their discretion in fixing charges for services not specifically mentioned in the regulations. - 128 - Annex S Page 7 of 32 Services Rendered Outside 40. Prescriptions and forms for investigations issued to patients out of consulting rooms will be paid for at the appropriate revenue collection points. In-Patient Service Accounts 41. Large institutions like Korle-Bu Teaching Hospital, Okomfo Anokye, Effia-Nkwania Hospital etc. will need to have a separate accounting system for in-paLient's fees. - 129 - Annex S Page 8 of 32 MINISTRY OF HEALTH REVENUE RETURNS FOR HEALTH CENTERS/ IOSTS AND CLINICS MONTH OF STATION DISTRICT REGION WEEK SERVICE RATE NUMBER AMOUNT 1. OUT PATIENTS, GHANAIANS 0-12 YEARS ...... ....... ...... 2. OUT PATIENTS, NON GHANAIANS 0-12 YEARS ...... ....... ...... 3. OUT PATIENTS, GHANAIANS ADULTS ...... ....... ...... 4. OUT PATIENTS, NON GHANAIANS ADULTS ...... ....... ..... 5. PRESCRIPTION ITEMS (DISPENSARY BEDS) ..... ....... ...... 6. ROUTINE LAB. INVESTIGATIONS ...... ...... 7. ROUTING X-RAY ..... ....... ..... . S. DELIVERY FEES ...... 9. (OTHERS) IN-PATIENT FEES ...... 10. MORTUARY FEES ...... TOTAL - 130 - Annex S Page 9 of 32 MINISTRY OF HEALTH REVENUE RETURNS FOR HEALTH CENTERS/ POSTS & CLINICS MONTH STATION DISTRICT WEEK SERVICE RATE NUMBER AMOUNT 1. OUT PATIENTS, GHANAIANS 0-12 YEAR3 ...... ...... ...... 2. OUT PATIENTS, NON GHANAIANS 0-12 YEARS ...... ...... ...... 3. OUT PATIENTS, GRANAIANS (ADULTS) ...... ...... ...... 4. OUT PATIENTS, NON GHANAIANS (ADULTS) ...... ...... ...... S. PRESCRIPTION ITEMS (DIS- PENSARY) ...... ...... ...... 6. ROUTINE LAB INVESTIGATIONS ...... ...... ...... 7. ROUTING X-RAY ...... ...... ...... 8. DELIVERY FEES ...... ...... ...... 9. (OTHERS) IN-PATIENT FEES .. .... ...... 10. MORTUARY FEES ...... ...... ...... TOTAL GRAND TOTAL .................. IN CHARGE: NAME .............. SIGNATURE: ................ -131 - ANNEX S Paee 10 of 32 GHANA L.I. 1277. HOSPI[AL FEES REGULAIfONS, 1983 IN oeorcise of the powers conferred on the Secretai, responsible for Health by section II of the Hospital Fees Act, 1971 (Act 387), and with the approval of the Council, these Regulations are made this 7th day of July, 1983. 1. Subject to the provisions *f the Act and thcso Regultion4o Feei pa.Rablo. the basic fees specifi-I 4'i the First Scl.. !ile to those Regulations shall be payable in respect of sci ;.cs rendered in a hospital and specified in relation to those fees 2. (1) Patients suffering from. Leprosy or Tuberculosis shall be Exemptions exempted from payment of all fees. uetion de (2) No fees other than cost of prescribed drugs shall be paid in respect of services rend'!ted in a hospital to any persons suffering fro:- (a) Meningitis (b) Chicken-pox (c) Cholera (d) Diphtheria (e) Malnutrition (Protein - Calorie - Malnutrition and Maraswits) (f) Measles (g) Onchocerciasis (h). Poliomyelitis (acute) (i) Relapsing fe lr (j) Schistosouiiasis (k) Smallpox (I) Tetanus ('n) Trachoma (n) Trypanosomiasis (o) Typhus (p) Typhoid (q) Venereal disease (r) Whooping-cough (s) Yaws (t) Yellow fever (u) Sickle cell disease (v) Vital Hepatitis (w) Haemorrhagic fevers (x) Ribies. (3) No fees other than hospital accommodation and catering services shall be paid ill any Governmenithospital or cl;rtc it respect of: (i) Ante-natal and Post-natal services; (ii) Treatment at Child Welfare ClinicL -132 - ANNEX S Page 11 of 32 LI. 1277. 2 HOSPITAL FEES RECULATIONS, 1983 (4) No fees shall be paid 'r- any immunisation against any diseases except for vaccination ce .ificates for international travel. Exemption 3. No fees (othlr than fees for special amenit.es) shall be paid in 2n(df Act.on rc pect of services rci.dered in a hospital to nv health services personnel including trainees. Storage o. 4. (1) The person liable to pay for the cold storage of a dead dead dies. body shall be the person requesting the service. (2) No fees shall be paid for the cold storage of a dead body at the request of any department of State. (3) The Medical Officer concerned may in his discretion waive the fees payable for cold storage of a dead body where there is delay in releasing the body to relatives due to post-mortem examinati mn, the coroner's teport or difficufty :n tracing the relatives of the dead person. Revocathn S. The Hospital Fees Regulations, 1971 (L.1. 706) are hereby of L.I. 706. revoked. commence- 6. These Regulations shall be deemed to have come into force on emet. the 21st day of Aptil, 1983. FIRST SCHEDULE PART 1-OUT-PATIENTS Service Fee 1. Casualty treatment in all hospitals, Polyclinics and general out-patients attendance .. .. .. .. .. ¢5.00 adults ¢2.50 children up to 12 years. Non-O.hanaian 00.ad.. .. .. .. .. .. aults. 10.00 children. 2. SpeciaUst out-patient clinics .. .. .. .. .. ¢25.0 1st visit 5.OO subsequent visits a05.00 Ist visit 3. Investigations:- ClO.00 subsequent visits (a) Routine laboratory services-lmmunological investiga- tions, blood films, blood grouping, and cross matching, direct sputum examinations, direct urine examinations, senotyping, haemoglobir ;timation, sickiing tests, etc. 02.00 per department (b) Speal inwetition, e.g. s.C.G. .. .. .. ¢20.00 (c) Speciel Ray xamination, e.g. contras modxa examina- tion .. .. .. .. .. .. .. ¢20.00 (d)Routite X-Ray examination, eg. chost, abdonxen skeletl system .. .. .. .. .. .. ¢10.00 - 133 - ANNEX S Page 12 of 32 3 Ll. 1277. HOSPITAL FEES REGULATIONS, 1983 PART 11-MEDICAL AND XRGICAL TREFTMENT Service Fee 1. Minor operative procedure as set out in Pals A of the Second Schedule: Ghanaian .. .. .. .. .. .. .. 010.00 Non-Ghanalan .. .. .. .. .. -0. 50.00 2. Minor surgical operation as set out in Part B of the Second Schedule; G;hanaian . .. .. .. 100.00 Non-Ghanalan .. .. .. .. .. .. £200.00 3. Major surgical procedure as set out in Part C of the Second Schedule: Ghanaian .. .. . - .. .. .. ¢200.00 Non-Ghanaiw% ....... . . . .* **4r 4. Delivery:- Regional/District hospitals G*hanaian 0.... . . . . . . 50.00 Non-Ghanaian .. .. . .. .. .. ¢200.00 Polyclinic,/Health Centres, H/Posts Ghanaian . . .. .. .. .. I .. .. ¢30.00 Non-Ghanaian .. .. .. .. .. .. £AO.00 PART Ill-DENTAL CHARGES Service Fee 1. Conutatiom n samtlon .. .. .. .. .. ¢20.00 Non-Ohanalan .. .. .. .. .. .. ¢60.00 2. Opemive procedure:- Non-Ghanaien (a) Dressing .. .. .. .. .. .. .. £5.00 C15.00 (b) Extraction .. .. .. .. .. .. .. ¢10.00 £¢30.00 (e) Filling *.. * * * .. .. .. .. ¢20.00 ¢60.00 (di Root Canal Therapy .. .. .. .. .. £50.00 Z150.00 3. Dentures . .. .. .. .. .. . £0.00 £Z150.00 () ¢tO.00 ¢60.00 (for additionsa tooth) (a) Dentur bearing 1-3 teeth .. .. .. .. do. (b) Denture bearing 4-8 teet .. .. .. .. o. (c) Denture bearng 14 teeth .. .. .. .. do. 4. Minor oral surgery. e.g. cyst root .. .. .. .. 50.00 £Z100.00 S. Major oral surgey .. .. .. .. .. .. ¢IOD £30000 6. (old Filling special Dental procedure: Gold (Cap or Fillig) o. .. .. ...... .. pet otst maneI - 134 - ANNEX S Page 13 of 32 LL 77. 4 HOSPITAL FEES REGULATIONS, 1983 PART IV-MEDICAL EXAMINATIONS Service Fee 1. For admission to schools and colleges: a.anaa . . .. .. .. ¢lo.oo NonaGhanaian ,. . .. .. ¢30.0 2. For employment-All cases .. .. .. .. S50.00 3. For Isumu and other claims: Ohbasl ,. ,. ,. .. .. . 2OO.O Non-Ghanala ..Z 4 .. .. .. .. .. i4OO. 4. Autopsy report 0... .. .. .. .. .. .. V5.00 S. Polis form ..0 .. .. .. .. .. SO.OO 6. Medical loards on public servants by Ministry of Health .. Free 7. Medical Boards on all other persons .. .. .. .. (Z0l.00 PART V-PRESCRIPTIONS . onts and ou wiientswUllpavCIX ietionrs,bt ob and othEr beanli nsti.ituuons. PART VI-HOSPITAL ACCOMMODATION AND CATERING SERVICES Service Fee per day ad.lt .. .. .. .. ........7.. 2. OAnoft a V.wr P. wards with catering svic: * iIGhm ..... .. .. ......... .. rZ7O.OO acouid 0. . . . . . . 50.00 * . t...c dN p. ..a f ty wards .. t - . ciBcalbv nquste this faity. eWhe food Is not provded by the hospita (with the apoval of the medical ofi * on.d)* o whM ho bosit do not bawv atern failitim the fer shall be edUcd '.byms.aI -135 - ANNEX S Page 14 of 32 5 L.I. 1277. HOSPIT4L FEES REGULATIONS, 1983 PARt V11-MISCELLANEOUS Service F'ee 1. Cold storago of persor., who die it, lhospital: (a) First 2 days .. Free- (b) Next S days ... . . SO pc day '.') Thereafter Up to a niaxitiiui otf 7 days .. .95.O) pes 4lay 2. Cold btnrage I jerwrons who die out of hospitals will be at V95.00 a day for a niaximtim of 14 days. 1. Coroner's Cases (a) Until body is identified . . .. FreeWo (b) First 3 days after post-mortem .50.00 per day (c) Therea.ter up to a maximum of II days .. . 95.CN per day 4. Unidentified bodies will be intcrred by the government after 14 days. 5. Ordhopacdic urrliaiIixs aasd prOmhilis .. . .. At cclt 6 Physiotherapy .. .. .. .. .. ¢10.00 per course 7. Ambulance: (a) A ride within town/city .. . . ...00 (b) A ride outside city .. To be based on milag at Government rate. -*The two free days shall be the first two week-days (exclusive of public holidays) upon wbich the body is stored. Where any Saturday, Sunday or PubliL Holiday imredfatly precedes, falls betueen or imn ediately follows those two week-days, that Saturday. Sundsy ;r Ptiblic lHoliday shall alsv be counted as a free day in the calcs:dguon of fes payable. SECOND SCHEDULE PART A MINOR OPERATIVE PROCEDURES OBUSETRIC AND GYNAECOLOGICAL 1. Paracentesis Abdominis 2. Lumbar Puncture OPIITHALMIC SURGERY I. Irrigation of Lachrymal drainage system 2. Fluoresceine angiography 3. Fundus photography 4. Electroretinogram S. Retrobulbar injections (alcohol, xylocaine) 6. Visual Fields 7. VER S. Eeactroowlogram - 136 - ANNEX S LI. 1277. 6 Page 15 o432 HOSPIT.4L FEES REGULA170,%.5, 1983 EAR, NOSE ANt) TfiROA I Ear 1. Syringing on account of wax. ptis or foreigia bhdfy 2. Removing of poilyps 3. Cualric 1'e,t 4. Audionicriic lest Nose I Incisioni and 1ackhting f nasal seplal ah.ccss 2. Nwud l Iol, Itectoni% 3. Antrunil lavage 4. Cauterization with electric-cautery or silver nitrat2 tfor epistaxis Throat A (3) Remnoval ol ft. in the pharyinx 6iEN.1 0 t. SURJLfRi I. Incision and i)rair-ige o* Abscessc 2. Paracentosis 3. Urethral dilatation 4. Gastroscopy 5. Coloaaoscopy 6. Closed reduction of fractures 7. Aspiration of knee and other joints PART B MINOR SURGICAL OPERATIONS OBtiETRIC ANL GYNAICo(t ofCAL".V 1. Dilatation and Curettage 2. E tcuation of Uterus 3. F.xcisioniincisionl 4. CautLrisation of Lower genital tract 5. Laparoscopy 6. Sterilisation 7. Episiotomy and Repair 8. Forceps Delivery 9. Vacuum Extraction OPHTHALMIC SURGIERY I. Excision of pterygium 2. Removal of Corneal Foreign Bodies 3. Removal of intra-Corneal Foreign Rodies 4. Repair of lid lacerations not involving puncta S. Repair of Conjunctival tears 6. 1 and C of Chalazion -137- ANNEX S Page 16 of 32 7 L.I. 1277. HOSPIl'AL FFEES REGUILATIONS, 198.? 7. Excision of small lesions of lid such as Warts, Xanthe- lasma, moles 8. Tarsorrhaphy 9. Lacrimal system Probing 10. One-snip. th.;e-snip operations of Canaliculus II. Rzmoval Lid Conerations i2. Cauterisation of Corneal lesion such as r)endritia. (ilIcer 13. Rcmoval of Bauid Keratopath% 14. Anterior t hamber washout for Hyphema 15. Removal of Conjunctival lumours 16. Conjunctivectomy 17. ConJunctival flaps EAR, NOSE AND TiHROAT Ear I . I ncision of maastoid abscess 2. Meatal Rcpair 3. Removal of embedded foreign body 4. Suturing of lacerated pinna 5. Suturing of torn ear lobe 6. Aural polypectomy 7. Lxci.ion otf Keloids. Cysts, Lipomata Nose 1. Treatment of fracture of the nose 2. Nasal I 'Iypectomy Throat 1. Incision of Peritonsillar Abscess 2. Elective and emergency Tracheotemy 3. Removal of Thyroglossal cyst General Surgery 1. Hernioarhaphy (d) Inguinal (b) Femoral (c) Umbilical 2. Varicoso veins operations 3. Haemorrhoidectomy 4. Minor wounds, grafts S. Hydrocelectomy 6. Circumcision 7. Fscision of cutaneous and suh utanoous lumps 8. biopsies--Breast 9. Laparoscopy 10. Minor amputations-fingers and toes II. Cystostomy - 138 - ANNEX S Page 17 of 32 L,L 12770 1l)'PIAL FEES REGULAIIONS, 1983 PART C MAJOR SURGICAL OPERATIONS OBSrETRIC AND GYN/XC0I.QJICAI I. ( aesarean Sectioisn (Central Prolapse repairl '. Hysterectomy ;. Myomectomy 4 Fistula repair s. Salpingectomy a. Oophorectomy(Ovariectomy 7. Tubal Surgery 8. Cancer Surgery 9. Colporrhaphy OPHTHALMIC SURGERY 1. Repair of Corneal Lacerations 2. Strabismus Surgery of all kinds I. Cataract Surgery of all kinds 4. Optical Iridectomy 5. Retinal Detachment 6. Enucleation 7. Evisceration 8. Orbital Snrgery (Exploration, biopsy, excision) 9. Exenteration 10. Plastic repair of Ectropion and Entropion 11. Dacryocystorrhinostomy 12. Peripher ' iridectomy 13. Glaucoma Surgery of all kinds 14. Exan,ination under General Auaesthesia 15. Lid repairs involving Puncta 16. Corneal Grafting 17. Scleral transplant 18. (Ptosis) Surgery of all kinds 19. Removai of Intraocular Foreign Body 20. CaIdwell Luc 21. Resection of Lid Tumors 22. Repair of Ruptured Globe EAR, NosE AND THROAT 1. Tonsillectomy and Adenoidectomy 2. Caldwell-Luc 3. Exploration of Frontal Sinus 4. Mastoidectomy-Radical and Simple 5. M:ixillectomv-Total and pertial 6. Laryngectomy-Total 7. Repair of cleft palate -139 - ANNEX S Page 18 of 32 9 LL 1277. HOSPITAL FEES REGULATYONS, 1983 8. Repair of cleft lip 9. Endoscopies for removal of foreign body from bronchus, oesophagus, larynx (a) Dilation of .ricture of oesophagus (b) Biopsies of Larynx, Bronchus, Oesophagus 10. Submucous resection of Nose II. Tyn?panoplasties 12. Excision of Preauricular Sinus Tracts. GENERAL SURGERY Head and Neck Burr holes for trauma Craniotomy, for trauma or tumour Glossectomy Parotidectomy (or excision of tumour) Radical neck dissection Resection man-ible or maxilla Submandibulectomy Thyroid, operations on Extended simple mastectomy Radical mastectomy Simple mastectomy T1horax Diaphragmatic hernia, repair of Oesophagectomy, all types Excision mediastinal tumour Open cardiotomy Pneumonectomy and lobectomy Segmen.. .,or wedge excision on lung Thoracoplasty, all types Valvulotomy, closed Muwculoskeleval Arthrodesis Arthroplasty Fractures, open reduction and fixation -140 - ANNEX S Page 19 of 3 LUL 1277. to HOSPITAL FF.ES REGULATIONS, 19e3 Abdomen Abdonino-perineal resection Adrenalectomy Appendicectomy cold and complicated Cholecystectomy Cholecystostomy Common duct, operations on Colectomy, pardal or total Colostomy Enterostomy, all types Exploratory laparotomy Exploratory and 1) sis of adhesions Gastro-enterostomiiy Gastrectomy, partial or total Repair of Incsional Hemiae Repair of para-umbilical herniae 111Wssuscept'fn, operations for Pancreas, operations on Perforations, closure of Pyloromyotomy Resection, small intestine Splenorenal or P.C. Anastomosis Splenectomy, for trauma, Splenectomy, for disease Vagotomy Omphalocele Genito-LS rinlary Cystectomy, p.:r'a:%1 or total Nephrectomy Pyelotomy Prostatectomy, transurethual Prostatectomy, all other types Urethral dilatation Extremities and Vasular Amputations, major Aneurysmectomy Embolectomy Neurorrhaphy Sympathectomy dorsal Sympathectomy, lumbar Tenorthaphy Vascular gfor prothsi -141- ANNEX S Page 20 of 32 tI LL 1277. fIOP ITAL FEES REGULATIONS, 1983 Mfusculoskeletal ArLwroa.sis Artbropiasty Fractures, open reduction and fixation Dislocations-open reduction Laminec;t, ay Other major reconstructive surgery Cleft palate or hafelip repair Pedicled and other major graft Other fonns of reconstructive surgery. Secretary repomsiblefo. Heath Date of Gazette notificatw: 8Sth JulY, 1983. -142- ANNEX S Page 21 of 32 GHANA IL 1313. HOSPITAL FEES REGULATONS, 1M85 IN vercise of the powers conferred on the Secretary responsible for Health by seCtiOn II of the Hospital Fees Act, 1971 (Act 387), and with the approval of the Provisional National Defence Council thes Regulatios are niade this 19th day of July, 198S. 1. Subject to the provisions of the H'spital Fees Act, 1971 Flowpayaw 387) and these Rezulations, lhe basic rees specified in the First uiaedule to these Regulations shall be p;,yable in respect of scrvices tzndered in a hospital a :specihed in reUion to those fees 2 (1) Patients suffering from Leprosy or Tub,iculosis are exemp- hexmou ted from payment of all fees. unde (2) No fees other than cost of prescrlbed drugs shall be paid in At 387. respect of sorvices rendered in a hospital to any perso suffering from` (a) Meningitis (b) Chicken-pox (c) Cholera (d Diphtheria (e) Malnutrition (Protein - Calorie - Malnutrition and Marasmus). ff) Measles () Onchocerciasis Poliomyclitis (acutd) fl Relapsing fever Sl) chistosoniasis (k) Smallpox (1) Tetanus (m) Trachoma (n) Trypanosomias (o) Typhus (p) Typhoid (q) Venereal disease Whooping-cough Yaws . Yellow fevef Sickle cell disase Vital Hepatitis ( Hasmorrhagic feves (x) Rabies. (3) No feos other than hospital acwommodation and catenag asgces shall be paid in any Government hospial or diic 2 repect (i) Ante-natal and Post-natal serics; (11) Tratmet at Child Welfas Clinics. - 143 - ANNEX S Page 22 of 32 Ll) 13113. 2 NOSPffAL Fr"' REGULA7TONS, 1985 (4) No fees shall be paid for any immunisation ag8ast any diseases excep, for vaccination certificates for international travel. =zomp=to 3. No fees (other than fees fcr specia1 amenities) shall be pal I in 2r)espec of services rendered iu a hospitl to any heal services 387. persoonel including trainees. S of 4. (1) The person liable to pay for the cold storago of a dead do" b*dm body shall be the person requesting the service. (2) No fees shall be paid for the cold storage of a dead body at tho request of any department of State. (3) The Medical Offmcer concerned may in his discrtion waive the fees payabt for cold storage of a dead body where there is delay in releasing t-, body to relatives due to post-mortem examination, the caroner's eport or difficulty in tracing the relatves of tfe dead person. RvL4mdo S. The hospital Fees Regulations, 1983 (L.I. 1277) ae hreby of"12n. xrevoked. C6 These Regulauuns .:z11 be deemed to have comPe inlo force in the Ist day of June, 1985. FIRST SCHEDULE PART l-OUT-PATIEN1S FM 8ae Adults I Cblltra (Than Au (a)