Report No. 29425-UG Uganda Improving Health Outcomes for the Poor in Uganda Current Status and Implications for Health Sector Development June 30, 2004 Human Development 1 Country Department 4 Africa Region Document of the World Bank Table of Contents Executive summary............................................................. ........................................ ... VIII Introduction..................................................................................................................... A.Background............................................................................................................ 1 B.Analytical Framework............................................................................................ 1 1 D.Policy Environment................................................................................................ C. The Context............................................................................................................ 3 4 E.DevelopmentsinOtherKeySectors .................................................................... F.New Challenges .................................................................................................... 10 G. Concluding Remarks........................................................................................... 10 12 Potential ........................................................................................................................ 2. Health, Nutrition, and Population Outcomes inUganda: OnPar but Below A. Overview.............................................................................................................. 13 13 B General Picture ofHealthandHealthMDGsinUganda..................................... . 14 D.ChildMorbidity.................................................................................................... 24 C. Child Mortality..................................................................................................... 17 E Maternal Mortality................................................................................................ F.Adult morbidity .................................................................................................... . 29 31 G.Fertility and Demographics.................................................................................. 36 3.HKnowledge, behavior andpractices affecting health............................................ .Possibilities for Meetingthe MDGs..................................................................... 38 A Introduction.......................................................................................................... . 40 40 B.The overallpicture ............................................................................................... 40 Individual Levels ...................................................................................................... C.Health-relatedKnowledge, Behavior and Practices at the Household and 41 51 4.D Utilization of . Government Response and Policy Implications .................................................. A.Introduction.......................................................................................................... Preventive and Curative Health Services ....................................... 54 54 B.The Overall Picture .............................................................................................. 54 C D Population-Based Public Health Activities........................................................ 55 67 E.Utilization of Curative Services ........................................................................... ..Utilization o f clinical preventive services ............................................................ 71 F. Conclusion and PolicyImplications ..................................................................... 74 5. A Health Systems Performance - Generating Inputsand Delivering Services ........ 76 .Introduction.......................................................................................................... 76 B.The Organization and ManagementofHealth Services....................................... C Health SystemPerformance................................................................................. 76 83 99 6.D HealthFinance andspending .............................................................................. .. Conclusion ........................................................................................................... A. Overall HealthFinance ...................................................................................... 102 B.MediumTerm ExpenditureFramework andPublic FinancingMechanisms .... 104 102 105 D.Development Expenditure andDonor Finance.................................................. C. Public Expenditureon Health ............................................................................ 112 E.Budget Process andManagement....................................................................... 113 ... 111 F.HouseholdExpenditure onHealth...................................................................... 117 7. G Impact ofthe Removal o fUser Fees................................................................... .Discussionand Recommendations..................................................................... 121 A.Evolution of User Fee Policy inUganda ........................................................... 123 123 B EmpiricalEvidence onthe Impact ofAbolition ofUserFees ........................... 124 128 D Comparison ofDistricts with andwithout User Fees........................................ C Distribution o f Benefits [Benefit Incidence] ...................................................... ... 131 E.Impact o fAbolition of User Fees on Service Provision..................................... 133 8.F Recommendations for Health Sector Development............................................ .Discussion and Recommendations ..................................................................... 134 136 A. Interventionsthat affect Infant and MaternalMortality............................ 136 B. Improve Health-PromotingPractices at the Familyand Community Level..... 137 C. Health Sector Financing.................................................................................... 138 D Health Sector Performance: A Focus onHumanResources andPrivate Public . Partnerships inHealth............................................................................................. 139 140 Bibliography ............................................................................................................... E. Strengthening Oversight and Accountability .................................................... 142 Table 2.1 Trends inhealth status indicators. Uganda 1989-2000................................. List of Tables 14 Table 2.2 Summary o f UgandanhealthMDGs ............................................................ 15 Table 2.3 Health. population and select economic indicators o f Uganda and selected Table 2.4 ARI and diarrhea prevalence inchildren inselected African countries ......28 sub-Saharan African countries. 2000.................................................................... 16 Table 2.6 Illness/injury duringlast 30 days o f survey (2002/2003) ............................ Table 2.5 Illness/injury during last 30 days o f survey (1999/2000 survey) .................. 31 32 Table 2.7 TB data for highburdenSSA countries. by incidence ................................. Table 3.1 Levels of different nutrition practices by wealth quintiles in2000..............34 44 Table 3.2 Feedingo f under-five children with diarrhea (percentage) .......................... 44 Table 3.3 Levels o f individual knowledge and practices relatedto HIV/AIDS by 48 Table 3.4 Hand washing practice by wealth quintiles in2000 .................................... wealth quintiles in2000 ........................................................................................ 49 49 Table 3.6 Tobacco and alcohol use by wealth quintiles in2000 .................................. Table 3.5 Hygiene and sanitationpractices ineight districts, 2003.............................. Table 4.1 Inequality inUgandanchildhood immunization coverage 1995 and 2000 ..50 57 Table 4.2 Inequality inANC coverage (richest vs poorest quintile) 1995 and 2000..60 Table 4.3 Access to safe drinkingwater 1995-2001..................................................... . 68 Table 4.4 Access to sanitation facilities 1995-2001 ..................................................... 69 Table 4.5 Bed net utilization inUganda 2003 .............................................................. 70 Table 4.6 ARI and diarrhea care-seeking behavior (2000) ........................................... 72 Table 4.7 Care-seeking behavior for self-reported episode o fmalaria inthe last 30 days precedingthe survey, 1999 and 2002/03 ...................................................... 74 Table 5.1 Functions and Services provided at various levels o fpublic sector health Table 5.2 Powers and Functions o f District Councils inthe Health Sector.................77 facilities................................................................................................................. 78 Table 5.3 Health Facilities by Level and Ownership ............................................... 80 iv Table 5.4 :HealthPersonnelin Ugandaas comparedwith Neighborsand SSA 85 Table 5.5 Staff Mix o f Facility, Ownership Category, and Region.............................. Average ................................................................................................................ 89 Table 5.6 Percentage Stockout Time for 10 Indicator Items by Level......................... 91 Table 5.7 Percentage Stockout Time for 10 Indicator Items by Ownership ................-9 1 Table 5.8 Availability o f Essential Drugs and Other Medical Inputsat the National Medical Stores, Feb-June 2000, n=125 days ........................................................ 93 Table 5.9 Availability o f Equipment by Ownership Category (percentage o f facilities) ............................................................................................................................... 94 Table 5.10 M i x o f services provided ingovernment, PNFP andprivate facilities in 97 Table 5.11 The status of EmOCservices inthe surveyeddistricts......................... four regions o f Uganda (percent) .......................................................................... 97 Table 6.1 Per capita expenditureon health inUganda............................................... 103 105 Table 6.3 HealthMTFEouttums (inUshBillion) ...................................................... Table 6.2 Sector shares o f GOU expenditure............................................................ 106 Table 6.4 Recurrent expenditure by level: 1999/00 -2002/03 .................................... 108 Table 6.5 Development expenditure, 1999 -2003 (Ushbillion) ............................... 112 Table 6.6 Annual per capitahousehold expenditure on health care in 1999 and 2002/03 (at fixed prices inUgandan Shillings in 1999) ................................................... 119 Table 6.7 Distribution o fPayment for the drugs amongrespondents who fell sick inthe last 30 days and sought care from a provider in2002/03 ................................... 120 Table 6.8 Payment o f drugsby type of illness............................................................ 121 Table 7.1 Percent ofwomen (15-49 years) perceiving different constraints to access health services (2000-01) ................................................................................... 124 Table 7.2 Change innonutilization o f health services by disease conditionbefore and Table 7.3 The Poor/rich ratio for utilization o f different types o fpublic facilities .... 128 after abolition o fuser fees................................................................................... . . Table 7.4. Benefit incidence o fprivate facilities by the type o f facilities ................. 130 129 Table 7.5 Utilization o f curative care inuser fee and no-user fee districts before (1999) andafter (2002/03) abolition o fuser fees........................................................... 131 Table 7.6 Change inUtilization o f curative care before and after abolition of user fees inuser fee andno-user fee districts..................................................................... 132 Table 7.7 Payment for drugs among respondents who fell sick inthe last 30 days and sought care from a healthprovider ..................................................................... 133 Listof Figures 2 Figure 1.2: District Primary Health Care Conditional Grant Funding........................... Figure 1.1: HNP and PRSP Framework ........................................................................ 5 Figure 2.1:Infant mortality and under-five mortality vis-&vis GNI o f Sub-Saharan 18 Figure2.2: Childmortality and socioeconomic development globally........................ African countries................................................................................................... 18 Figure2.3:Trend ininfant mortality inUganda, 1940-2000 ........................................ 19 Figure 2.4: Trends ininfantmortality and under-fivemortality inUganda and some 20 Figure 2.5: Overall trends inchild mortality inUganda 1980-2000............................. neighboring SSA countries ................................................................................... 20 V Figure 2.6: Inequities ininfantmortality inselected SSA countries............................ 21 Figure2.7: Regionaltrends inchildmortality, Uganda 1988, 1995 and 2000.............23 Figure2.8: Trends inchildmortality by gender inUganda 1988, 1995 and 2000.......23 Figure2.9: Trends inwealth-based variation inchildmortality inUganda.................23 Figure 2.10: Under-fivemortality trends and Millennium Development Goals ..........24 Figure 2.11:Wealth-based inequalities insevere underweight inSSA....................... Figure2.12: Trends inmalnutrition.............................................................................. 25 27 Figure2.13: Trends insevere stuntingbywealth quintiles ......................................... 27 Figure2.14: Trends insevere underweightbywealth quintiles.................................. 27 Figure 2.15: Trends inmalnutrition concentration curves............................................ 27 Figure 2.16: Estimated maternal mortality ratios for selected countries inthe SSA and World, 1995 .......................................................................................................... 30 Figure 2.17: Number o f adults infectedwith HIV/AIDS and adult prevalence rates o f 33 35 Figure2.19: Trendsinfever and diarrhea prevalence in0-4 years olds....................... Figure 2.18: Chloroquine treatment failure inAfrica .................................................. HIVIAIDSinSSA................................................................................................. 36 Figure 2.20: Trends inTFRby residence ..................................................................... 39 Figure2.21: Trends inTFR bywealth quintiles........................................................... 39 Figure2.22: Trends inTFRbywomen's education ..................................................... Figure 2.23: Trends inASFR........................................................................................ 39 39 Figure 3.1 Infant feeding practices inUganda and selected countries inthe region, latest available data............................................................................................... 42 Figure3.2 Micro-nutrition practices inUganda and selected countriesinthe region, Figure 3.3 Mothers' knowledge o f danger signs duringpregnancy anddelivery ........43 latest available data............................................................................................... 45 Figure 3.4 The gap between knowledge and practice inmalaria prevention among caregivers inUganda............................................................................................. 46 Figure3.5 Percentage o fyoung people (15-24) knowingthat condoms canprevent HIV/AIDS(1996-2002) inUgandaand selected countries inthe region.............47 Figure 3.6 Percentage of young people (15-24) usingcondoms inlast high-risksex Figure 4.1: Trends inDPT3 coverage inUganda and select neighboring countries ...47 (1996-2002) inUgandaand selected countries inthe region ............................... Figure4.2: DPT3 coverage bywealth quintiles inUgandaandselected countries......57 56 Figure4.3: Utilization o fANC providedby a trainedhealthprofessional, rankedbythe level o f use inpregnant women o fpoorest quintile inUganda and selected low andmiddleincome countries ................................................................................ 60 Figure4.4: Attended delivery by a medicallytrained person invarious countries, 62 Figure 4.5: CPR and TFR insub-Saharan African countries........................................ rankedby coverage inthe poorest quintile ........................................................... 66 Figure4.6: Use o fmodern contraceptive bywomen inrichest andpoorest quintiles, Figure 4.7: Number o f first-time testers aged 15-28 at the AIDS Information Center.66 rankedby average values ...................................................................................... 67 Figure4.8: Percentage o fhouseholds with access to water supplyandsanitationin rural areas inUganda ............................................................................................ Figure4.9: Utilization o fbednets and ITNs among children under-five inUganda ...69 70 vi Figure 4.10: Care-seeking behavior for ARI by quintile insome Sub-Saharan Countries ............................................................................................................................... Figure4.11:Utilization ofcurative services inUganda, 1999 and 2002 ..................... 72 73 Figure 5.1 District Structure for HealthCare Delivery................................................. 78 Figure 5.2 Percentage of populationwithin 5 KMradius by districts .......................... 84 Figure 5.3 Total number o fhealthworkers inthe public sector per 100,000 population ................................................................................................................................ 87 Figure 5.4 District-level breakdowns o f Healthpersonnel by Category ...................... 88 Figure6.1 Sources of HealthFinancing(%) ............................................................. 103 Figure6.2 Share o frecurrent spendingbylevel......................................................... Figure6.3 Share of the total government wage bill.................................................... 107 Figure6.4 Wage Distribution, 2002103 ..................................................................... 109 109 Figure 6.5 Health sub-district drugbudget................................................................. 111 Figure 6.6 Development expenditure.......................................................................... 112 Figure 6.7 Interactions and Timing ofthe BudgetProcess......................................... 114 Figure6.8 Monthlyout-of-pocket per capita expenditure on healthinthe past decade ............................................................................................................................. 117 Figure 7.1Concentration curves for utilization o f public health facilities .................. 130 Figure6.9 Out-of-pocket health spendingper capita................................................. 118 This report was prepared by a team consisting of Shiyan Chao (Task Team Leader. ECSHD). Son Nam Nguyen (Co-task Team Leader. AFTHI). Sarbani Chakraboty (ECSHD). Katherine A Marilyn Lauglo. ManjuRani. Sekhar Bonu. and William Ainashe (Consultants) The teambenefitedfrom guidance ..Tulenko. Peter Okwero (AFTH1). and comments from its peer reviewers Alexander S . Preker (HDNHE). Ritva S Reinikka. (DECRG). Agnes L B . . . Soucat. and KhamaOderaRogo (AFTHD) The team also receivedcomments from LouiseFox. Julie McLaughlin . and ChristopherWalker . The team has worked closely with Uganda Ministry of Health and development partners andbenefitedfrom their inputs and comments . Judy O'Connor (Country Director) and DzingaiMutumbuka(Sector Manager) provided overall guidance Financial assistance from Norwegian Trust Fund and IDA Trust Fund are . gratefully acknowledged. vii EXECUTIVESUMMARY MAIN CONCLUSIONS This study shows that Uganda has similar or somewhat better health indicators than other African countries with a similar income level. However, in 2000 those indicators showed a lack o f major improvement even with a favorable environment o f economic growth and progress in health system development. More recent health indicators would be useful but IMR, MMR and U5R are unlikely to have changed substantially in the past four years. Inequality persists in health outcomes, access to health services, and use o f health services. Further effort i s neededto address inequality inhealth outcomes and inthe health system. Inits fight against HIV/AIDS, Uganda has shown that improving knowledge, behaviors and practices for better health outcomes i s possible. It i s crucial that this success story be replicated in other individual and family health practices. Good health results in large measure from hygienic practices at the household level. Individuals, households, and communities needto be empowered and mobilized to take charge o f their health. Ugandahas instituted a number o f reforms inthe past decade: decentralization, SWAP, and shiftingresources to PHC. Yet, its health system continues to face persistent problems: physical inaccessibility o f services for a large portion o f its population; severe human resource gaps; and disrupted flows o f drugs, supplies, and equipment. These resource constraints make it imperativethat existing resources be usedefficiently. However, the health system's mechanisms for accountability appear to be weak. Technical accountability through the MOH andpolitical accountability throughthe district councils couldbe strengthened. There are a number o f interventions which could strengthen management o fthe service. These include: 0 an improved MIS system 0 a comprehensive human resource policy better bookkeeping and recordingroutines implementation o f a coordinated partnership with private providers 0 an effective system o f supportive supervision, or a strengthenedmultisectoralresponse with other sectors that affect on healthoutcomes. The health subdistricts (HSDs) are an interesting innovation in service delivery in Uganda. However, in order to get the full benefits fiom them, measures must be taken to ensure that HSDs have the management capacity to carry out their responsibilities for PHC. Along with growing autonomy and responsibility for HSDs i s the need for an identification o f constraints which need to be resolved at a higher level (e.g. human resources) and capacity building. The government has increasingly recognized the important linkages between health and poverty reduction and economic growth. However, financing health care i s still at a very low level, particularly with the largest share from private out-of-pocket expenditure. The ... Vlll government needs to further expand its current financing strategy to encourage the development o f healthinsurance schemes and explore other options for riskpooling. The removal o f user fees at public facilities has producedpositive results. As a result, service provision from public sector became more pro-poor. Even through the overall level o f out-of- pocket spendingdid not change much, the financial burden o f health care has been borne more by richer households. The free drugs providedby the government have reachedmore poor people. However, the coverage o f free services i s still limited. The increase inuse o f public providers i s less than would have been expected from the abolition o f user fees. Noticeably, the use o f both for-profit and non-profit private providers has drastically increasedrecently, not only among the rich, but also among the poor. Improving quality o f services inthe public sector i s critical to the success o fthe no-user fee policy: otherwise the public system will fail the poor. Inorder to improve and sustain a certain quality o f services, the government needs to increase its health allocationto compensate the loss o f income and increased demand due to abolition o f user fees. Given that the use the private sector has drastically increased and most health services are provided by non-public sectors, the government needs to increase its role in regulatory and oversight to ensure quality o f services and access by the poor and vulnerable groups. Overall recommendations arising from this report, which are discussed in detail inChapter 8, have center on the five areas: (a) prioritizing interventions that affect infant and maternal mortality, (b) improving healthpromotion and disease prevention practices at the family and community level through community mobilization and intersectoral collaboration, (c) mobilizing funds for the health sector including strategies that encourage riskpolling mechanisms; (d) focusing on humanresources and collaboration with the private sector to improve health service delivery, and (e) improving accountability through improving information systems and supervision. A detailed summary o fthis reportbegins immediately below. INTRODUCTION The Ugandan health sector has undergone rebuilding, reorientation, and reform since 1986. In the 1980s, health sector development focused on rebuilding the collapsed health system. This was followed in the early 1990s by a reorientation o f health services from an overemphasis on specialized curative care towards preventive and primary health care services. Since the late 1990s, the reforms have focused on improving service coverage, especially for preventive and primary health care services, under a decentralized health system. The current reform i s guided by the 10-year National Health Policy and the National Health Sector Strategic Plan (HSSP 2000/01-2004/05), as well as the Poverty Eradication Action Plan (PEN). It aims to provide all Ugandans with an essential health care package o f selected cost-effective interventions that address conditions which account for most o f the disease burden in the country. To achieve this goal, the government aims to increase public expenditure on health, shift health care resources in favor o f rural areas where the majority o f the poor live, emphasize primary health care, and bringservices closer to the people. Significant progress has been made in health systems development, particularlyin the areas of decentralization, partnershipswith donors, and the private-not-for-profithealth providers (PNFP)using a sector wide approach (SWAP).However, despite the progress on health systems development as well as positive trends in economic growth, health status i x indicators reveal a mixed picture. The country has been able to reverse the HIV/AIDS epidemic, reducing HIV prevalence from 18 percent in the early 1990s to six percent in 2003. On the other hand, while maternal mortality fell modestly from 523 to 505 per 100,000 live birthsover the past ten years, there was virtually no improvementininfant and child mortality, which stand at around 100 and 150per 1,000 live births respectively. Similarly, there has been little improvement in children's nutritional status since 1995. The total fertility rate remains highat 6.9. Uganda presents a mixed picture on the health related MillenniumDevelopment Goals (MDGs). It is expected to reach the MDGs for AIDS, TB, and malaria, but not for hunger (nutrition), child mortality, and maternal mortality. If Uganda i s to improve its health indicators and meet the MGDs, it needs to focus on improving these indicators among its poorest and most vulnerable citizens and communities. Uganda is currently at a critical stage in its health sector development. The government and development partners inthe health sector are inthe process o f preparingthe second Health Sector Strategy Plan. There i s a consensus that Uganda needs to improve health outcomes, specifically among the poor, and to combat its high infant and maternal mortalities, as well as more effectively address major diseases such as malaria, AIDS and TB. The objective of this report i s to provide a better understanding o f the connections between health outcomes and poverty inUganda and the performance o f the health system intargeting the poor. By analyzing available data from surveys and the health information system and by conducting literature reviews, the report provides an in-depth assessment o f inequalities in health outcomes, health behavior and practices, and utilization o f health services, as well as o f health system performance and financing from the equity perspective. On this basis, the report examines policy implications and sector developments related to improving health outcomes for the poor inUganda. It aims to provide a comprehensive assessment on health and poverty and consolidate the evidence for health policymaking in the country, especially for the development o f the Second Health Sector Strategy Plan UNDERSTANDINGHEALTHAND POVERTY The analysis in this report is guided by the PRSP framework on health, nutrition and population (HNP), and poverty. This framework outlines the linkages between pertinent factors inthe household, community, health systems, and policy areas that affect health-related poverty outcomes. It also acknowledges the impact that other sectors have on health outcomes. The report also uses the fkameworks for accountability and service provision in the World Development Report 2004: Making Services Workfor Poor People. Wherever possible, this report examines the relationships and interactions between citizensklients, organizational providers/frontline professionals, and policymakers in the Uganda health system using this framework. This report is organized into eight chapters. Chapter 1 reviews recent developments in the health sector the policy context, and outlines the conceptual frameworks which guide the study. Chapter 2 assesses inequalities in health outcomes and the gaps in reaching health related MDGs. Chapters 3 and 4 analyze household health-related behavior and practices as well as the utilization o f health services. Chapter 5 examines health sector performance and Chapter 6 discusses health sector financing and expenditures. Chapter 7 analyzes the impact o f the abolition o f user fees on utilization. Finally, Chapter 8 links together issues discussed in X Chapters 2 through 7 and draws policy implications and recommendations for future development. HEALTH OUTCOMESAND INEQUALITIES Although Uganda performs on par with or somewhat better than other low-income Sub- SaharanAfrican (SSA) countries, it is an underachiever in healthglobally for its income level. Furthermore,the country has not made improvementrecentlyin healthindicators commensuratewith its economic growth. Although the country has made important strides in addressing some of the most important causes ofmorbidity and mortality inthe population (e.g. HIV, and common childhood illnesses), infant and matemal mortalities remain high and life expectancy at birth remains low: in 2000 it was 42 years ranking 16* among SSA countries. Economic growth during the last ten years (estimated at over 5% per annum in GDP growth) has not translated in significant improvement in health outcomes. Health inequality persists along the lines o f economic status, urban-rural residence, region, and education levels. Uganda can do better in child and maternalmortality. The country has one of the highest levels of child mortality in the world. According to indirect estimates, Uganda's infant mortality (IMR) and under-five mortality (U5M) stood at 101 and 152 respectively per 1000 live births in 2000. The analysis o f IMR and U 5 M shows that Uganda has a relatively low level o f child mortality compared to other low income SSA countries. However, globally, Uganda i s an underperformer in child mortality for its level o f economic development. Maternal mortality, one of the most important causes of loss o f healthy life years inUganda, i s high. Even the optimistic estimate of 505 matemal deaths per 100,000 live births places Uganda among a cluster o f SSA countries with high MMR. Similar to the picture on child mortality, the MMR in Uganda i s somewhat better than other low-income SSA countries but based on a global comparison, it could do better given its income level. Adolescents contribute 44 - 48 percent o f matemal mortality. Uganda also has one o f the highest total fertility rates (TFR) o f 6.9 birthsper woman, which has remainedalmost constant between 1995 and 2000. Data are not available to assess the most recent changes inIMR, U 5 M and MMRsince 2000. InUgandaas inother developingcountries, alargepart ofthe burdenofdisease amongthe poor i s attributable to communicable diseases, especially, malaria, TB, and HIV/AIDS. Malaria i s the number one cause o f morbidity and mortality inUganda, even exceeding HN/AIDS. It accounts for 40 percent o f all outpatient visits, 25 percent o f hospitalizations, and 14 percent o f inpatient deaths. Uganda can rightfully be proud o f the way it has addressed HIV/AIDS. HIV prevalence inthe general populationwas 15 to 20 percent inthe early 1990s and was brought down to 6.5 percent in2003. Although the tide o f the epidemic has tumed, there are still an estimated 600,000 to 1.2 million people living with HIV/AIDS and prevalence i s still increasing inrural areas. The TB epidemic has increased with the A I D S epidemic and it i s estimated that 50 percent o f TB patients inUganda are HIV positive. Over 34,000 new cases are detected each year, with the highestnumber among adults intheir 20s and 30s. Inequalities in health outcomes remain significant challenges in Uganda. There are inequalities inhealth outcomes by socioeconomic status, region, and urban and rural areas. The poor and vulnerable groups bear a disproportionately heavy burden o f disease. IMR in the poorest quintile was 105.7 per 1,000, 1.8 times higher than that in the richest quintile. Children from better-off families had fewer incidences o f fever, acute respiratory infection (ARI),and diarrhea compared to the children from the poorest families. Northern and Western Regions, which are poorer areas, have worse health outcomes. The number o f internally xi displaced people has grown drastically in the past two years and was estimated at 1.4 million inDecember 2003. This means that ifanything, health status indicators are worse than those captured in the household surveys carried out before 2002. Those most affected by HIV/AIDS continue to be the poorest and most marginalized in Ugandan society: commercial sex workers, injecting drug users, men who have sex with men, internally displaced people, and young women. More than 50 percent o f all people living with HIV/AIDS as well as the newly infected are women. The TFR i s 8.5 children per woman inthe poorest quintile while the TFR for women inhighest quintile has dropped to 4.1. Uganda's adolescent fertility rate i s 178 per 1000 women, one o f the highest in SSA, again with considerable differences between the poorest and the richest. The single most important contributor to highinfant and maternal mortality i s the combination o f high fertility and short birth spacing. With regard to this problem, Uganda needs to: a) prioritize efforts to increase the use o f modem family planning methods so that women can achieve their desired family size and birth spacing; b) provide emergency obstetric care; and c) address the needs o f adolescent reproductive health. HOUSEHOLD KNOWLEDGE, PRACTICES, AND BEHAVIORAFFECTINGHEALTH Compared to other countries in Sub-Saharan Africa (SSA), Ugandans have relatively high levels of knowledge and good practice in relation to breast-feeding and HIV/AIDS prevention. However, there i s still much room for improvement with regard to nutrition, home care o f the sick child, malaria prevention, family planning, maternal health, sanitation, and alcohol consumption. There are considerable differences in health-related knowledge and behaviors between the poor and nonpoor, for example with regard to micronutrition, tobacco and alcohol consumption, sanitation practices, and HIV/AIDS. Even though Ugandans have relatively high levels of knowledge and good practice in relation to breast-feeding and HIV/AIDS prevention as compared to other countries in Sub-Saharan Africa (SSA), knowledge alone i s insufficient for producing healthy lifestyles and behavior. Individuals and families must have the means to carry out health- promoting practices. Sufficient food must be available, functioning health services accessible, clean water available, and household belongings, such as bednets, affordable. The ability to carry out care-seeking behavior is also affected by other factors: availability of information, educational level of family members, and differential decision-making within households and in communities. Considerable poor-rich disparities exist in all these areas. Poor people are also at a disadvantage in receiving information. Considerable differences exist between the wealth quintiles in radio listenership (19 percent among the poorest compared to 86 percent among the richest) and TV viewing (one percent compared to 35 percent) as well as newspaper readership (two percent compared to 42 percent) among women. Education i s a central factor that enables decision-making. In Uganda, as elsewhere, secondary school education for women i s associated with individual and household health- promoting practices. While school completion rates for Ugandan adults are relatively high within SSA, the disparities between the poorest (women - 24 percent; men - 47 percent) and xii best o f f (women - 82 percent; men- 87 percent) quintiles are great. For women, the disparities are larger than for SSA women overall. Compared to other women in SSA, Ugandan women fare relatively well in a number o f areas o f household decision-making and autonomy, except in the area of deciding whether to seek health carefor their children. Other studies also confirm the importance o f male decision- making for seeking health care outside the home for family members. Behavior change communication i s a key to changing household practices and individual behavior and must buildupon strengthening individual, family, and community capacity. This requires a focus on strategies for improving practice rather than just imparting knowledge. A successful approach will draw upon local resources, and actively involve meninhealthpromotingactivities. Sustainedandvisible nationalcommitment, a multi-sectoralapproach, and community mobilizationare neededto changepersonalandcommunity behaviorthat are necessary for improvedhealth. HEALTH SERVICE UTILIZATION Health service utilization is the result of both supply and demand for health care. Individual and household howledge, attitudes, behavior, and financial assets, along with supporting community infrastructure and networks (or the lack o f these) influence the degree to which people use health services. The availability, accessibility, affordability, and quality o f health services also influence whether people are willing to spend their time, effort, and money to ,usethem. Inequalities in the utilization of health care services persist. People who are better off, have higher levels o f education, and live in urban areas are more likely to use health services. Children in the richer families are much more likely to be immunized and more likely to receive care when sick with ARI or diarrhea. Women from better-off households are consistently more likely to utilize family planning services, receive professional antenatal care (ANC), be immunized against tetanus, and give birth with the help o f health professionals. Generally speaking, wealth-based inequalities in utilization o f health services worsened between 1995 and 2000, especially inthe case o f immunization and ANC. Mothers' level o f education i s a key factor in health service utilization. Women with secondary school level are 4.3 times more likely to receive professional A N C care and 1.5 times more likely to be immunized against tetanus compared to those without education. Education-based inequalities decreased between 1995 and 2000. Districts experiencing insecurityhave different and severe constraints on their health systems and require special arrangements for carrying out their work. Inall categories, urban residents are more likely to utilize'health services than their rural counterparts. One of the key characteristics of maternalhealth in Uganda is the discrepancy between the high utilizationof ANC services (93 percent) and the low levels of births attended by health professionals(38 percent) or in health facilities. Reasons for this are complex but the factors involve the accessibility o f services, understanding at the household and community level o f the possibility o f complications during delivery, and perceptions o f how women are xiii treated at health facilities when they deliver there. Placing women's concems at the center o f health interventions i s important ifwomen are going to use health facilities near their homes. Taking up health-promoting activities such as bed net use, use o f clean water, and sanitation practices are the means by which individuals and families can improve their health status. Community mobilization efforts inthese areas are likely to be most effective. Within the triangle of accountabilityframework of policy makers, service providers and clients, service providers outside the health sector (e.g. in water, sanitation, transport, and education) also are responsible for improving the health of the population. Collaboration and planning with these nonhealth sectors are also needed at the district level to ensure that the tools by which health can be improved are accessible to the poor. Community groups are potentially quite influential for behavior change and for maintaining infiastructure. Careful planning i s needed in order to ensure a good match between the uptake o f health- promoting activities and the availability o fthe means to do so. HEALTH SYSTEMS PERFORMANCE Although geographical access to health services has improved, there are regional- and district-level as well as urban-rural disparities.Geographic accessibility i s a key element o f improving utilization o f essential health services. The percentage o f the population living within 5 luno f a health facility has improved from 49 percent inthe late 1990s to 57 percent in 2000. Despite these improvements at the aggregate level, urban-rural, regional, district, and income disparities remain. Inthe 2002 National Household Survey (NHS), 46 percent in the poorest quintile reported distance as a constraining factor for seeking health care services as compared with only 25 percent in the richest quintile. Those living in the Eastem (25 percent) and Northern regions (2 1 percent) face greater geographical access problems compared to those living in Central (12 percent) and Western (15 percent). Geographical access to health care i s closely linkedto poverty status. The majority o f the poor live inthe rural areas where access i s limited. For example, 20 percent o f the rural population named distance as a factor in not seeking health care compared to only two percent inurban areas in2002. The availability of healthpersonnelis the most critical challengeinhealth services delivery.Attracting andretaining qualified staffinlower level health facilities inrural areas and certain parts of the country (e.g. the North which has been subject to problems o f insecurity) continues to be one o f the biggest challenges inhealth services delivery. There are severe district-level variations inthe availability o f trained staff. Katakwi district has fewer than 20 healthpersonnel per 100,000 population compared with Kampala, which has more than 200. The most recent Annual Health Sector Performance Report (AHSPR) found that in primav health care (PHC), 53 percent o f the posts were filled by trained staff when applying the 1999 staffing norms. Addressing this problem will require interventions at all levels o f the health system. There i s a need for a coherent and comprehensive human resource policy. In addition to covering issues such as the recruitmentbase, throughput, health service needs, training needs, and in-service training, such a policy should also address the issue o f the impact o f HIV/AIDS on the workforce. The timely and continuous availability of drugsand material suppliesis also a problem. The availability o f drugs and other medical equipment and supplies inhealth facilities i s a key factor affecting the perceivedand the technical quality o f care. A drug-tracking study in four districts o f Uganda found that average stockout rate for all essential drugs was 30 percent in Government facilities and 16 percent inPNFP facilities. Stockout rates were higher at H C I1 xiv and I11as comparedwith H C IV and district hospitals. Iflower level healthcenters do not have drugs, the tendency is to depend on community drug vendors or bypass H C I1to higher levels (HC IV or district hospitals). This has implications for the efficiency, quality, and accessibility o f health services for rural populations. Some o f the key factors related to the continuous availability o f drugs and supplies at the district level include: financing problems and delays with the PHCgrant, a fragmented and duplicative supply system, problems with stockouts at the National Medical Stores, limited capacity for drug supply management at the district level and health facilities, and the lack o f good information and communication systems. Ugandais investinginimprovingthe technical qualityof care butthere is considerable scope for improvedsupervisionandaccountability.The Government ofUganda (GoU) has adapted internationally accepted clinical protocols for the management o f childhoodillnesses and health personnel have been trained inIMCI, management o fmalaria, and DOTS. The critical challengefacing improvements in technical quality is the intensity and quality of supervision.An evaluation o fpublic sector healthpersonnel clinical practices found that despite training inIMCI, trained personnel didnot perform any better than untrainedpersonnel on some critical clinical guidelines. This was attributed to problems in supervision. Inboth public and private facilities, technical quality o f care with regardto drugprescribingi s poor. Antenatal care (ANC) needs to be better focused. The DHS 2000 found that only 34 percent o f women took antimalarial prophylactic treatment and 69 percent hadreceivedthe tetanus toxoid injection despite the 93 percent attendance for ANC. Another study found that only 26 percent o f pregnant women were informed about the danger signs o f pregnancy complications and only two-thirds had their bloodpressure measured despite four or five visits to providers. There i s little data on the technical quality o f care inthe private sector. The little available information indicates that community drug vendors inrural areas tend to misuse antibiotics, prescribe multiple drugs, and do not give clear instructions about the use o f drugs, including contra indications. PNFP and privateprovidersperformbetteron perceivedqualitycomparedto government clinics. Characteristics such as the cleanliness of clinic, waiting times, andclinic opening hours as well as the behavior o fhealth personnel are key to influencingthe perceived quality o f care. Typically, govemment facilities are open fewer hours than PNFP and PFP facilities. Ingeneral, waiting times are the longest ingovemment clinics. Women report the negative attitude o f healthworkers inpublic facilities i s an important deterrent for deciding to give birthin a formal healthunit. Patients report high satisfaction with the care receivedat PNFPand PFP facilities, including informal private providers such as drug vendors. Although decentralizationis intendedto promoteaccountability,the resultsinthe health sector haveyet to beseen. GoU's decentralization is to promote community participation and to make the delivery o f health services more accountable and transparent. This has yet to be demonstrated. The lack o f effective supervision i s an indication that mechanisms for technical accountability for service provision need to be strengthened. Ministries o f health in decentralized systems often face constraints incarrying out their oversight role. Ugandahas not yet shown that it has effective mechanisms inplace to carry out its oversight role on a regular basis. Similarly, more work i s needed to look at the links for accountability between the political authorities at district level (the District Council) and the health staff (the DHMT)inorder to find ways o fimprovingpublic accountability, transparency, and communityparticipation. xv The HSSP is implementedthrough a Sector-WideApproach (SWAP), which iswell establishedwith stakeholders satisfiedwith its purpose and impact. With itsjoint reviews, annual reports, healthpolicy advisory committee, and sector working groups, the MOH's capacity for monitoring the implementationhasbeen increased. MOH still needs to further enhance its oversight mechanisms (keeping track o f equity, financial management, information systems, supervision, quality assurance) to improve the service quality inpublic facilities. A multi-sectoral response is neededto improve healthoutcomes. There is little information on how the district health services coordinateand planwithin the overall district services, especially with those sectors that affect health outcomes, e.g. water and sanitation, agriculture, communication and transport, andeducation. HEALTH FINANCINGAND EXPENDITURES HealthservicesinUganda are mainly financedby three sources: privateout-of-pocket spending, government allocations, and external aid, with the private funds contributing almost 60 percent of the total health expenditure. Total per capita expenditure onhealthi s estimated as $17.1 in2002/03, of which $9.9 i s out-of-pocket spending, $3.9 (23 percent) from the government and the remaining$3.3 (19 percent) fimdedbydonors. Expenditure on health inUganda hasincreasedinrecentyears. However, its level remainsvery low, particularly interms of public spending. The government increasingly recognizes the linkages between health and development and that investing inhealth can contribute to poverty reduction and economic growth. Inthe context o f PEAP implementation, financing for the health sector has gained a more prominent position inthe government's budget allocation. As the result, public spending on health as a proportion o f the total government budget has increased from 6.5 percent to 9.9 percent between 1994/5 and 2002/3. However, such a proportion i s still a relatively small share inthe government expenditure program compared to other sectors such as public administration and security spending. The government has not provided sufficient funds to offset the loss o f income due to the abolition o f user fees or to support increased demand for free services. Without increasing fundingto the health sector, positive gains from the abolition o f user fees would be difficult to sustain. Resource allocationhasimprovedwithin the health sector, with primary health care increasinglygainingmoreweight inthe healthbudget allocation. The distribution o f recurrent expenditures shows a clear trend o f increased spending at the district level duringthe period o f 1999 to 2003. The share o f the district budget has increased from 36.8 percent o f the total recurrent expenditure in 1999/2000 to 45.9 percent in2002/03, due to the significant increase o f the Primary Health Care Conditional Grant. Government's subsidies to NGO health facilities have increased from five percent o f the total recurrent expenditure to 11percent and will continue to grow to 14percent inthe next three years basedonprojections inthe Medium- TermExpenditure Framework (MTEF). At the same time, the MOHheadquarters and two national hospitals, Butabika and Mulago hospitals, consumed 36 percent o f the total recurrent expenditure in2002/03, down from 41 percent in 1999/2000. These patterns o f the expenditure reflectedgovernment's emphasis on basic health care delivery inrecent years. The MTEFfor the next three years indicates a continuation o f the current spending pattern. Wage and personalemoluments including lunch allowances take about 40 percent of the total recurrentexpenditures in the health sector. This i s not highcompared to many other developingcountries where personal emoluments are the predominant share o f recurrent costs. xv1 Nonwage expenditure in the health sector thus accounts for more than 60 percent o f the total recurrent cost including expenses on drugs and other health commodities, as well as on maintenance o f buildings and vehicles and administrative costs. Available information points to the need to further improve allocative efficiency, particularly for expenditure for drugs and other essential services for better service delivery. Although budget planningand managementhaveimprovedunder the MTEFand SWAP process,there is still a roomfor increasedefficiency. Data on allocatingnonwage recurrent costs are limited and do not allow an in-depth assessment o f budget performance and identify actual spending on priority programs. IMPACT OF THE REMOVALOF USER FEES Abolishing user fees in public facilities has yielded positive results in utilization of curative care. Public health services inUganda have become more pro-poor because the poor benefit more from free public health services and free drugs. A benefit incidence analysis indicates that the use o f public health centers by the poorest 20 percent o f population has increased by 17 percent points as compared with before user fees were abolished, and the use o f public hospitals by the poorest 20% has increased by 10 percent points. The report also shows that the coverage o f free services i s limited. The use of the private providers-both for-profit and non-profit-has also drastically increased,not only amongthe rich, but also amongthe poor. More than 40 percent o f the poor from the lowest quintile also use private providers when they were ill.The increased use o f the private sector after the abolition o f user fees i s more than the increased use o f the public sector (11.4 percent points for the private facilities versus 8.5 percent points for public facilities). Consumers choice may indicate deficiencies in the quality o f services provided by the public sector. Households continueto contribute the largest share of healthfinancing in Uganda. Even after the removal o f user fees, the overall out-of-pocket spending has not decreased. Even though the overall level o f out-of-pocket spending has not been reduced, the financial burden o f health care has beenborne more by richer households. Whether increasedutilizationtranslatesintobetter healthoutcomesfor the poor remains to be seen. The situation raises important questions that require further investigation and policy consideration. Why has almost 70 percent o f the top quintile abandoned the public sector and why do more than 40 percent o f the poor from the lowest quintile also use the private providers when there are free services provided by the public sector? This issue i s closely related to service access and quality, which this study could not investigate further due to a lack o f data. Issues o f the quality o f health care are talung center stage and attention should be directed to these along with relevant issues o f accountability. Furthermore, even when the use o f curative health services has increased, this may not result in better health outputs and outcomes. It i s possible that people may use the fi-ee services first, and then pay for services when the free services do not resolve their problems. Such practices increase inefficiency inthe system and further stretch the very limited resources inthe households. A comparison o f the user fee andno-user fee districts shows that ifthe quality o f services does not improve in the public sector health facilities, regardless o f user fees, patients will move away from the public facilities to the private facilities. xvii Uganda needsbothpublicand privateproviders. Considering the size of disease burdenin Uganda (25 percent populationreporting sick inlast 30 days) and the health care needs of a very young population (52 percent of population i s under age 15), the current network of public health facilities i s simply inadequate to provide health care to the population. At the same time, the GoU i s not ina positionto substantially increase its investments inthe sector in the near future. The government needs to continue to buildcloser collaboration with both the profit and not-for-profit private health care sector, particularly inthe area of preventive care. The government needs to use various measures such as performance-based contracting that will enable the private sector to contribute to public policy objectives. xviii INTRODUCTION A. BACKGROUND 1.1 This report highlights the connections between health outcomes and poverty and focuses particularly on inequalities in health. It looks at how the health sector i s currently tackling these issues and examines key policy interventions for increasing access to quality health services by poor people in an equitable manner. It aims to provide a better understanding and knowledge o f health sector issues for the development o f the Health Sector Strategy Plan I1and the revised Poverty Eradication Action Plan (PEAP) currently under development. The report i s expected to informpolicy makers, and the Poverty Reduction Support Credit (PRSC) process, as well as other developmental assistance inhealth. 1.2 Chapter 1 starts with the analytical framework which guides the study. It then gives an overview o f background and policy context. Chapter 2 provides an overview o f health, nutrition, and population outcomes. Chapter 3 analyzes how knowledge, behavior, and practices at the householdlevel affect health. Chapter 4 provides information on the utilization o f health services. Chapter 5 provides a picture o f how the health system i s organized, functions, and performs. Chapter 6 reviews health financing and spending. Chapter 7 analyzes the impact o f the abolition o f user fees on the use o f services, particularly by the poor. Chapter 8 links together issues discussed inChapters 2 through 7 to provide recommendations. B.ANALYTICALFRAMEWORK 1.3 The Uganda Health Sector Report uses the health and poverty framework (Figure 2.1) for poverty reduction strategy papers to examine the linkages among determinants o f health related poverty outcomes. The framework assists in identifying pertinent factors in the household, community, health systems, and policy areas that affect health outcomes (Claeson et al. 2001). After analyzing these relationships, a picture emerges o f what the health system can do to enhance the positive effects and mitigate the negative ones. A country's commitment to eradicating poverty and improving health-related outcomes gives its central government a key role in devising interventions which will reduce the impoverishing effects o f illhealth. Interventions can be found at the macroeconomic, health system, and microeconomic levels. 1.4 Poverty i s both a consequence and a cause o f ill-health. Illhealth, malnutrition, and highfertility are often reasons why households end up in poverty, or sink krther into it if they are already poor. 1.5 At the household level, poverty can limit knowledge, use o f services, and spending choices.. Out- of-pocket payments for health services-especially for catastrophic illness-can be the factor that drives a household into poverty. Participatory poverty assessments have underlined the linkages between illhealth and poverty. 1 Figure 2.1: HNP and PRSPFramework HNP and PRSP Framew0rk:Determinants of Source: Claeson Met al., 2001 1.6 Households are important as `producers' o f health through self-care, their health promoting and damaging practices (e.g. diet, use o f water and sanitation facilities, sexual practices, tobacco consumption), and their use o f preventive and curative health services. Intra-household relationships-especially between women and their husbands-often have implications for decision-making related to the use o f household resources, o f contraceptives, or of when to seek health care. Household assets (e.g. income, land, knowledge, information and communication sources, and social networks) are important and can cushion poor families duringtimes o f crisis. Communities' values, norms, networks, ability to control their environment, and influence over decisions can also affect health outcomes. Chapters 3 and 4 discuss some o f these factors and their linkswith use o f services. 1.7 When loolung at health service provision, financing o f health services, and at the availability and quality o f other services (education, water and sanitation, agriculture, roads and transport) that affect health outcomes, potential areas for government action come into view. There are different levels o f government interventions: a) the macroeconomic level (i.e., the amount o f resources allocated to health and allocations reaching the poor); b) health systems (i.e,, the way government has put together reforms and incentives to make the system work better for poor people); c) health care provision (i-e., the extent and level o f services and whether activities are implemented inways which reachpoorpeople); and d) amounts andmechanisms o fhealthfinancing. 1.8 The household and community level i s important in the accountability triangle o f poor people, policymakers, and providers as described in the World Development Report 2004: Making Services Workfor Poor People (World Bank 2003b). Given poor people's lack o f representation and the asymmetry o f information between clients andproviders, the poor are especially voiceless in the case of health services. Ways to strengthen their participation requires efforts that will 2 empower them-for example, through the provision o f information or support to community groups. Other efforts can focus on population-oriented outreach around specific health initiatives. The take-up o f some o f these efforts i s discussed inChapter 4. C. THECONTEXT Geography, population and demography 1.9 Uganda i s a landlocked country located inEastern Africa covering an area o f 241,038 sq km. It straddles the Equator and i s bordered by Kenya in the East, Sudan in the North, Tanzania in the South and the Democratic Republic o f Congo in the West. It has a population o f about 24.8 million people (2002 census). Between 1991 and 2002 the population increased from 16.7 million to 24.8 million, a rate o f 3.4%, making Uganda one o f the fastest growing populations in the world. 51% of the populationis between 0 - 14 years and a vast majority (87%) reside inthe rural areas. Average TFR i s high at 6.9%. For the last 17 years, persistent insurgency in the north and west of the country has resulted inthe displacement o f over 1.4 million Ugandans as o f December 2003. Overview of the economy 1.10 With a GDPper capita estimated at US$320, Ugandais ranked among the poorest countries inthe world. The country has recovered from the economic decline following years o f civil strife o f the 1970s and 1980s and after a series o f Structural Adjustment Programs, it has achieved a sound macroeconomic basis for restoring growth and reducing poverty. Overall macroeconomic performance over the last decade has been remarkable; annual inflation was kept below 5% and GDP growth sustained at an average o f over 5% per annum. Between 1992 and 2000, poverty was substantially reduced and income poverty fell from 56% to 35%. The proportion o f Ugandans living below the poverty line remains unacceptably high and according to the 2003 National Household Survey, income poverty increasedto 38%. Agriculture i s the mainstay o f the economy employing a large proportion o f the population. Constrained terms o f trade and unpredictable weather have curtailed growth in the agriculture sector and negatively affected Uganda's performance. Total output from agriculture as a percentage o f GDP has declined from 51% in 1991/92 to 42% in 2001/02 (MOFPED 2003a). The services sector has increasingly become dominant in terms o f GDP, contributing about 40% with manufacturingjust under 20%. The country i s dependent on significant external donor support. InFY2002/03, 52% o f the total government budget was from external donors. Overview of health outcomes 1.11 Uganda i s known for its achievements in economic growth and poverty reduction over the last decade. However, in terms o f health status, the results are mixed. The country has been able to reverse the HIV/AIDS epidemic, reducing HIV prevalence from 18% inthe early 1990s to 6% in 2001. On the other hand, over the past ten years, while maternal mortality fell modestly from 523 to 505 per 100,000 live births, there was virtually no improvement in infant and child mortality, which stood around 100 and 150per 1,000 live birthsrespectively. Similarly, there has been little improvement in children's nutritional status since 1995 and the total fertility rate remains high at 6.9%. 1.12 Uganda's health status i s characterized by a highlevel o f disease burden. In 1995, the Burden o f Disease - Cost Effective study found that over 75% o f the life years lost from premature death were due to mainly preventable diseases. Prenatal and maternal conditions, malaria, acute 3 respiratory tract infections, AIDS, and diarrhea together account for over 60% o f the total national death burden. Apart from the heavy burden o f infectious diseases, Uganda i s also simultaneously experiencing an increase inthe occurrence o f non-communicable diseases such as hypertension, diabetes, cancer, mental illness, and chronic heart disease. Chapter 2 provides a detailed discussion o fhealth outcomes. D.POLICYENVIRONMENT 1.13 Government policies establish the context within which the health sector functions. The overall policy environment for health has been affected by policies related to poverty reduction, civil service reform, and political decentralization. These policies are around the topics o f poverty reduction, health sector development, financing, and organizational changes and are interlinkedto each other. Different policies often arise from the same pressures and address shared concerns; their effects overlap and sometimes conflict; and the relationships among them are synergistic. Government commitmentto povertyreduction 1.14 Uganda's commitment to poverty reduction was articulated inits Poverty EradicationAction Plan 2000 - 2003 (PEAP), first drafted in 1997 and revised in 2000. The summary o f the revised PEAP served as the basis for Uganda's PRSP which was the first to be approved by the WF3 and IMF in May 2000. The purpose of the PEAP is to provide an overarching framework to guide public action to eradicate poverty. The second revision o f the PEAP i s currently (in2004) being developed. 1.15 The Ugandan Participatory Poverty Assessment Reports (UPPAP) carried out in 1999 and 2002 highlighted poor health as the most frequently cited cause and consequence o f poverty. One o f the four `pillars' o f PEAP i s to increase the quality o f life o f the poor. This includes the reduction o f HIV/AIDS, reductioninfamily size, improvements inhealth care coverage, increased access to clean water, sanitation, acceptable housing, and improving access to and quality o f primary education. 1.16 Under PEAP, the government strategy for the health sector i s to shift health care resources in favor o f rural areas where the majority o f the poor live, emphasize primary health care, and bring services closer to the people. This was highlighted inthe first PEAP (1997) and later in the 2000 PEAP. Since 1997/98, a significant component o f the Poverty Action Fund (PAF), which was established in 1997 to support priority areas for poverty reduction, has been earmarked as the Primary Health Care Conditional Grant (PHC-CG). The PHC-CG aims to support and protect fundingfor the district healthsystem andPHC services. 1.17 The PAF hnds have a special status under the Medium-Term Expenditure Framework budget management. They are priorities ingovernment budget allocation and are usually protected from unplanned budget cuts. For example, in 2002/03, due to major increases indefense spending, the government cut budgets inmany ministries and sectors but the PAF fundingwas untouched. The PAF funding i s also subject to more rigorous regulations, monitoring, and evaluation. Districts are required to submit quarterly reports on the PHC-CGto the Ministry o f Finance, Planning and Economic Development (MOFPED). The MOH i s supposed to supervise the use o f the PHC grants through quarterly field visits. The MOH has carried out special studies to monitor the execution o f PHC funds (see Chapter 6 for a more detailed discussion under Tracking Studies). 1.18 The PHC-CG to public facilities consists o f three components: development, recurrent wage, and nonwage. The PHC-CG also includes a separate grant to private-not-for-profit (PNFP) facilities 4 without specifying its usage. It i s up to PNFP facility management to decide how to use the funds. However, districts and PNFP providers sign aMemorandum o fUnderstanding (MOU) that encourages PHN? facilities to provide essential primary care, particularly preventive care. After the removal o f user fees in public facilities, the government also encouraged PNFP providers to lower their fees in order to increase affordability by the poor (see Chapter 7 for a more detailed discussion). 1.19 The MOFPED and MOH have issued a range o f general and specific guidelines for PAF and specifically for PHC conditional grants. The "Guidelines on the Use o f the Primary Health Care Conditional Grants to Districts" specify the purposes and procedures on how to use the funds. The MOH also issues guidelines and instructions on the allocation and disbursement o f the PHC funds eachyear (MOH 2003b). 1.20 Despite problems and difficulties in Figure2.2 DistrictPrimaryHealthCare managing the PHC grants, in general ConditionalGrant Funding the PHC grants have proved to be better functioning funding sources than the earlier programmatic funding. Bookkeeping and reporting for the PHC grants are more accurate and timely. Most o fthe fbnds are spent on ......................................... ~~ ~~........ intended purposes (Horizons Inter- national Ltd. 2001, MOH 2002b, MOH 2003~). Therefore, the MOH believes that PHC conditional fbnding i s the most efficient way to allocate both government and donor funds for primary health care. Indeed, the allo- cation to the PHC grants has increased 1997/98 1996/99 1999/00 2000/01 2001/02 2002103 drastically since its creation (Figure 2.2) Source: MOHDatabase 1.21 The PEAP has been monitored annually. Eight o f the 30 poverty monitoringpriority indicators relate to health and an additional two monitor water and sanitation. Inrecent years, the MOH has increasingly paid attention to poverty-relatedissues in its budget allocation and execution. With the effort to better target the poor, the allocation o f the PHC-CG has become more pro-poor. The allocation formula started with a flat rate for health subdistricts, then moved to population-based allocations. In 2002/03, the MOH, in consultation with the Poverty Monitoring Unit in MOFPED, introduced "need factors" into the formula. The "need factors" reflect the situation on poverty (using expenditure data), health needs, and the availability o f health project funding fiom donors in each district. As a result, the three districts with the highest needs receivedan allocationwhich was more than 25% higher per capita than the three most privileged districts in2002/03 (MOH 2002~). 1.22 PEAP 2004 - 2007 i s currently being developed and again includes improving health outcomes and increasing people's ability to plan the size o f their families as one o f the priorities o f human development. 5 Health policy 1.23 Untilthe 1970s, Uganda's health sector was considered to be one of the best inAfrica. Steady improvements were recorded in most health indicators. However, the civil conflict o f the 1970s through 1986 severely disrupted health service delivery and eroded the functional capacity o f the health sector. As the public health sector declined, the private not-for-profit (PNFP) and for-profit (PFP) sector increased (Macrae et al. 1996, Hutchinson 1999). The health sector declined in importance as a central government priority and by 1986, health represented only 2.4% o f government expenditure (World Bank 1988). 1.24 Since 1986, the health sector has undergone reconstruction and reform. In the immediate post- conflict period, the focus was mainly on rehabilitation o f the physical infrastructure. Duringthis period, health services were fragmented and largely maintained through donor-supported vertical programs and humanitarianorganizations. The Health Policy Review Commission was formed in 1987 to review policies in the sector and identify gaps. The HIV/AIDS epidemic peaked in the early 1990s and the government embarked on a vigorous campaign to stop the its spread. The Uganda AIDS Commission was created to coordinate a multisectoral response. 1.25 The early 1990s marked the beginning o f the govemment's attempts to systematically reform the health sector within a national health policy framework. The reform agenda was outlined in the White Paper on Health (1992) and Three Year Plan Frame (1993-1995) and focused on: a) Reorienting health services away from curative care to prevention and promotion by reallocatingresources towards primary health care; b) Mobilizing additionalresources to finance the health sector; c) Strengthening planning, management, and coordination o f services at various levels; d) Renovating and consolidating existing facilities and services; e) Promoting community participation in the development and management o f health services; 0 Strengthening private/public partnership; g) Decentralizing health services delivery to the districts as part o f the govemment's overall decentralization strategy; and h) Enabling the sector to participate in ongoing govemment-wide reforms such as civil service reform and liberation o f the economy. 1.26 These strategic thrusts represented a departure from earlier developments in the health sector in the immediate post-conflict period. The underlying principles o f the White Paper and Three Year Plan Frame were continued and can be found inthe National Health Policy (NHP), finalized in 1999 and in the Health Sector Strategic Plan (HSSP) 2000/01 - 2004/05. Continuing the health reform agenda, the NHP and HSSP are clearly linkedto the PEAP. Much o f donor support to the sector i s channeled through a Sector Wide Approach which i s linked to the HSSP. 1.27 The overall objective o f NHP i s to reduce morbidity, mortality, fertility, and health-related disparities. The NHP stresses the need to improve quality and equity in access to services 6 (geographically and for all social groups, especially the poor) and to develop partnerships with donors, private-for-profit (PFP) and private-not-for-profit (PNFP) health providers and non- governmental organizations. 1.28 In 1995, a Burden of Disease and Cost-Effective Study identified the major conditions contributing to morbidity and mortality in Uganda. Its findings contributed to the definition o f the Uganda National Minimum Health Care Package (UNMHCP). Access for all to the UNMHCP i s the central focus o f the NHP. The intention o f the UNMHCP is to address priority health problems, but some would maintain that it i s too wide-ranging and has not focused sufficiently on priority areas. Over the past three years, the HSSP has prioritized malaria, immunization, HIV/AIDS, reproductive health, and sanitation. Other HSSP strategies include strengthening a number o f key policy and operational functions including: the health care delivery system; the legal and regulatory framework; the integrated support systems (human resources, quality assurance, information management systems, procurement and drugs, equipment, supplies and logistics, health care financing); the policy, planning, and information management systems; and research and development. The HSSP i s delivered through a decentralized system with districts taking primary responsibility for implementation o f the UNMHCP. 1.29 The HSSP introduced the new concept o f the Health Subdistricts (HSDs). In connection with decentralization (discussed below), the HSD i s a subunit within the district health system and i s central to the delivery o f primary health care (PHC) activities and the UNMHCP. To strengthen health care at PHC level, the `Health Infrastructure Development and Maintenance Plan' was finalized in 2001. Together with the `Guidelines for the Provision o f the Uganda National Minimum Health Care Package' (2001), the functional care levels for PHC are spelled out as discussed inChapter 5. 1.30 The Midterm Review o f the HSSP was conducted in 2003. Plans for the development o f the next HSSP are currently underway. Financing 1.31 Health financing and spending are discussed in detail in Chapter 6 but key facts are highlighted here to help provide an overall picture. There are three major sources o f financing o f health care: private spending, government allocations, and external aid. Annual expenditure on health was estimated to be U S D 17.1 per capita in 2002/03, o f which U S D 9.9 was out-of-pocket spending; USD 3.9 was from government; and USD 3.3 was from donors. The Health Financing Strategy (2002) estimated that USD 28 per capita i s required to finance the HSSP. A severe funding gap has been a challenge to Uganda since the govemment came into power in 1986. Given this, attempts have been made to find other sources o f funding. Together with decentralization moves and the Local Government Statute (1993), user fees were widely introduced around 1993. By March 2001, when user fees were abolished, all but two districts had instituted them. Details about user fees are given inChapter 7. 1.32 Government financing to the health sector and especially to PHC increased between 1997/98 - 2002/03. Between 1997/98 and 2002/03, government allocations to health increased from 6.5% to 9%, made possible through the PHC-CG discussed above. Duringthis time, public fundingto PNFPs also increased. 1.33 The Medium Term ExpenditureFramework (MTEF) i s the budget tool which provides guiding principles for the budgetary resources and sets budget ceilings for each sector. The MOH, along 7 with other spending agencies, prepares its budget through consultations and within the ceiling established by the Ministry o f Finance, Planning and Economic Development. The MTEF includes external donor funds, many o f which are targeted at the health sector. However, precise donor financing i s poorly recorded and captured. This i s partly because some donor support i s in the form o f drugs, vaccines, technical support, and training. Budget planning i s made difficult when there i s poor coordination between the development budget and its implications for subsequent recurrent budgets. Organizationalchanges 1.34 A number of organizational reforms have taken place. Chief among them is decentralization but also important are SWAP and a growing recognition o f the importance o f the private sector through the public-private partnerships for health (PPPH). Decentralization 1.35 Decentralization strengthens the democratization process and enhances participation on the local level. Since services are planned and managed closer to service users, it enables a better understanding o f local needs and priorities. It i s thought to facilitate greater inter-sectoral collaboration and i s meant to promote greater transparency and accountability. Many political, administrative, fiscal, and technical responsibilities are transferred through decentralization as a result o f pushes and pullsbetweenthe center andperiphery over time. 1.36 The foundations for decentralization in Uganda date back to 1993 with the passage o f the Decentralization Act in 1993 and the initiation o f phased decentralization. Responsibility for a number o f functions was transferred from central government to districts over a three year time frame. In 1997, decentralization was further consolidated with the enactment of the Local Government Act, which replaced the earlier decentralization act. This involved the operational responsibilities (planning, implementation, and to some extent, financing) for integrated health promotion, disease prevention, and curative services below the district level for delivery o f the UMNHCP. PNFP facilities are increasingly playing an important role in PHC in the rural areas and are included among the responsibilities o f the District Director o f Health Services (DDHS). Two hundred and sixteen existing health facilities have been upgraded or built to support the HSD. 1.37 Financial decentralization, hiringof staff, and control over the district hospital were transferred to districts over time. The National Referral Hospitals were made semiautonomous and directly accountable to the Treasury for fundingas were Regional Hospitals at a later date. Hospitals were allowed to outsource services for cleaning, repairs, and maintenance o f vehicles and equipment. Previously fragmented payrolls were later harmonized and now offer the MOH an increased possibility o f gaining an overview o fhealthpersonnel. 1.38 All these changes requirednew skills andreorganization inthe MOHandinthe local authorities. Most important was the need for financing, administrative, and technical capacity at district levels and the capacity for oversight, resource mobilization, capacity development, technical support, and coordination at the national level. The MOH underwent two reorganizations, in 1995 and in 1999. 1.39 Decentralization o f the health sector i s linked to the political administrative structure o f local government. District health services are at the level o f the district council, an elected body o f representatives. The health subdistricts established inconnection with the HSSP are at the county 8 level. They are the responsibility o f the DDHS who is, in principle, accountable to the District Council. The subcounty level, which has the ability to collect taxes, i s below the county level with the corresponding HC I11health facility. Health Unit Management Committees (HUMC) were established in many areas in connection with the user fee policy and were supposed to provide communities oversight over the collection and use o f the fees. 1.40 As with all aspects o f decentralization, financial decentralization occurred over time through a range o f mechanisms. When the block grant system of allocation was introduced in 1996/97, it was noted that allocation to PHC at the district level was one quarter the amount anticipated by central government (Jeppson 2001). Although variation existed among the districts and this amount was increased in the subsequent year, as discussed above, the government instituted the PHC-CG to ensure that more funds were directed to PHC. 1.41 Decentralizationhas yet to demonstrate that it benefits the health outcomes o f poor people. One o f the known difficulties in other countries i s that decentralization i s unable to address inequalities resulting from the inability o f poorly resourced districts to provide services equal to those o f better-off districts. Especially difficult i s the issue o f attracting and retaining qualified staff, particularly in remote areas. In terms o f public participation, there i s some evidence that even after a decade o f decentralization, public participation inUganda i s confined to usinghealth services and making contributions o f labor, land, or money (Kapiriri et al. 2003). Despite periods when HUMCs were active, marginalized groups have not been effective in influencing health service implementation, monitoring, and evaluation. SWAp 1.42 The HSSP i s implemented through a Sector-Wide Approach (SWAP) under a common development framework. Development o f the SWAP resulted from several different but converging interests. Prior to decentralization, donor participation inthe health sector was closely tied to the vertical technical programs they supported. Donors interacted with technical programs or units in the MOH or directly with districts. The internal coordination o f the MOH and its Health Planning Department were weak. Decentralizationrequired donors to change the way they interacted with the government. The Poverty Action Fund (PAF) established in 1997 provided a vehicle for donors to provide earmarked funds for priority areas. As the GoU was encouraging partners to move from project type of support to budget support, a sector-wide approach inhealth also emerged. 1.43 The GoU and development partners agreed to collaborate on implementing the NHP and HSSP through a framework detailed in a Memorandum o f Understanding signed in August 2000. A SWAP provides the mechanism for common planning, budgeting, accounting, reporting, and for joint reviews, coordinated donor inputs, and government steering o f the process. A SWAP can improve transparency and accountability o fpublic spending. 1.44 Joint review missions (JRM) have been held biannually: in October/November to review performance o f the previous financial year and in MarcWApril to agree priorities, budget allocation, and the financing plan for the coming year. Visits to districts are included inthe JRM and documentation can be quite extensive. Undertakings, which highlight a few priority actions are agreed at each JRM and reported on at the following one. Annual Health Sector Performance Reports prepared for the October meeting, include a review o f 18 monitoring indicators covering inputs, processes, and outputs. The Health Policy Advisory Committee (HPAC) oversees the implementation arising from the JRM. Committee members are from the MOH, representatives 9 o f other central ministries, the donors, and PNFP providers. Various Working Groups have been established to develop policies, guidelines and plans. The results are reportedat the JRM. 1.45 It has been reported that the SWAP has resulted in the emergence o f a much stronger planning role for the MOH. The midterm review (MTR) reports that "the SWAP provides an enabling environment for effective coordination o f efforts o f all partners, increased efficiency in resource allocation and achieves equity inthe distribution o f the resources available for health." The MTR further reports that "there i s consensus that the partnership i s sound and developing in the right direction.. .and that strengthened collaboration with the PNFP subsector has been achieved." It suggests that HPAC should play a greater role in strategic planning. What i s less clear however, i s the degree to which the findings from the JRM are followed up, especially at district level. 1.46 While stakeholders appear to be satisfied with the purpose and impact o f the SWAP, some observers question whether the SWAP and indeed the PAF undermines the government's decentralization intents inthat they diminishthe financial flexibility available to local authorities indecidingtheir ownpriorities (Kasumba and Land2003). E.DEVELOPMENTS INOTHERKEY SECTORS 1.47 Universal primary education (WE) was launched in 1997 and progress has been made in expanding the access and improving the quality o f primary education. According to the 2002 school census, net enrollment for the poorest quintile has increased to about 83% with equal numbers o f boys and girls. While gender parity in primary education appears to have been achieved, the quality o f education and dropout rates remain areas o f concern. 1.48 Water and Sanitation Uganda has made substantial progress in increasing access to safe drinking water which in 2002 was reported as 55% for the country overall. Uganda i s on track for achieving the target for safe water. Sanitation, however, i s low. Access to improved sanitation in rural areas was estimated in2000 to be 51% (MOFPED 2003b). F.NEWCHALLENGES 1.49 As the health sector continues to develop and the new PEAP and HSSP are formulated, the overall context has become more complex. The persistent financing gap continues to be a challenge. In addition, the global health initiatives, while providing much needed funds, pose additional challenges in coordination and collaboration. Similarly, the Millennium Development Goals, while giving clear targets, also need to be harnessed to move the entire development program inthe right direction. Systematic collaboration with the private sector 1.50 Private health care services have long played an important role in Uganda, especially in the provision o f PHC in rural areas. The private sector includes private-not-for-profit providers (PNFP) and private-for-profit providers (PFP) inthe health care system. A significant portion o f the population use traditional forms of care that include traditional birth attendants (TBA), traditional healers, herbalists, and spiritual healers as well as self-treatment using drugs bought from the many small drug shops and vendors found in different parts o f the country. Generally, there i s little information available on these providers other than for TBAs. 1.51 The `National Policy on Public Private Partnership in Health, Final Editorial Draft' notes that PNFP facilities account for 42 or the country's 98 hospitals and 509 (24%) o f the lower level 10 units. Inadditionthey operate 20 out of 48 training schools inthe country. Seventy-eight percent o f the PNFP facilities are faith based institutions operating under three umbrella organizations: the Uganda Catholic Medical Bureau (UCMB), the Uganda Protestant Medical Bureau (UPMB), and the Uganda MuslimMedical Bureau (UMMB). The remaining are run by humanitarian and community-based organizations. 1.52 Inrecognitionofthe importance ofprivateproviders, attempts to develop apolicythat integrated the private sector started as far back as with the Health Policy Review Commission in 1987. As noted above, this was again attempted in the 1992 govemment White Paper and it i s included in the NHP. After many years, a policy has been finalized and agreed by the MOH. Guidelines have been finalized and it now awaits fundinginorder to be operationalized. 1.53 The policy provides a framework for the integration o f PNFP and PFP providers within the existing HSSP and health care delivery system. It i s o f critical importance that the public and private subsectors are coordinated in a manner that best uses the comparative advantages o f each in each local area. Since 58% o f health care financing is directly from users, this points to the urgent need for effective mechanisms to integrate and regulate the private sector. The overall provision by the public and private sectors should be organized so that access, equity, affordability, and quality are ensured. GlobalHealthInitiatives(GHI) 1.54 Global Health Initiatives (GHI) are international alliances drawing upon the donor community who are providing massive new financial resources to specific programs.' Among them are the Global Fund to fight AIDS, Tuberculosis, and Malaria (GFATM) and the Global Alliance for Vaccines and Immunization (GAVI). Procedures in accessing the funds from these programs have been instituted so as to enhance national ownership o f activities and coordination with other parts o f the health system. Uganda has applied to and will receive U S D 76 mil from GAVI and USD 97 mil from GFATM over the next three to five years. Uganda's health system i s still fragile, attempting to provide the minimum care package with severe financial and human resource constraints and limited capacity. Use o f the GHI funds requires careful coordination so that activities are well integrated among all stakeholders and at all levels o f the health care system. Strong leadership i s also required to ensure that these initiatives do not become another type o f vertical program whose activities are not sustainable when the fundingrecedes and whose activities undermine the overall health system. MillenniumDevelopmentGoals 1.55 The Millennium Development Goals (MDG) are the result o f consensus building in the international community for more than a decade on what i s needed to for development. The eight broad goals have clear targets and indicators that countries have committed themselves to meet by 2015. If all are met, this will be an indication of the country's progress towards poverty reduction. What i s often forgotten i s that the meeting the goals i s an indication o f political commitment translated into sectoral action. The MDGs cannot be met by technical interventions alone. Moreover, the MDGsunderline that there i s a synergy resulting from the interrelationships among poverty, hunger, education, gender equality, health, environmental sustainability, and global partnerships. The greatest progress in meeting the targets will be made when sectoral efforts are especially designed to meet the needs o f poor people. Consistent and frequent efforts are required to monitor how health interventions are reaching and affecting poor people, poor communities, and services inareas which serve the poor. 11 G. CONCLUDINGREMARKS 1.56 Given Uganda's economic growth and stability, its commitment to poverty reduction, and the strategic decisions it has made in the health sector, a key question is, `why haven't some health outcomes improved?' By examining factors in the household, the community, and performance o f the health system, this report aims to help identify interventions that are important for improving health outcomes for poor people so that the next HSSPwill be more effective. 12 2. HEALTH, NUTRITION,AND POPULATION OUTCOMESINUGANDA: ON PARBUT BELOW POTENTIAL A. OVERVIEW 2.1 This chapter discusses Health, Nutrition, and Population (HNP) outcomes in Uganda, with particular attention to the Millennium Development Goals (MDGs). It examines the levels and trends o f such outcomes and compares them to those in other sub-Saharan African (SSA) countries. Wherever possible, the paper examines HNP outcomes with regardto equity, as well as its socioeconomic and demographic correlates. 2.2 Through the analysis, an overall story emerges o f a country that, although performing on par or somewhat better than Afixan countries with similar Gross National Income (GNI) per capita, has not experienced an improvement inhealthindicators equal to its economic growth. In addition, not all Ugandans have benefited from health improvements and great health disparities exist along lines o f urban-rural residence, region, income, and education level. One frequent finding i s worse health outcomes in the Northern and Westem Regions and better outcomes in the Eastem and Central Regions. Interestingly, there are "positive deviant" districts such as Moyo and Nebbi within the Northem Region and Kisoro, Kanungo, Rukunguri, and Kabarole in the Western Region which buck the trend and are among the top ten performing districts. Eastern Region maintains some o f the best health outcomes despite having the second highest poverty rate. This can be a subject o f further investigation. The linksbetweenhealth status and other factors, such as sanitation and maternal education, suggest a cross-sectoral approach i s needed. There i s strong evidence that the marginal dollar buys more health improvements in communities with poor outcomes than in communities that already have good health outcomes. If Uganda i s to improve its health indicators and meet the MGDs it now needs to focus on improving the health of its poorest and most vulnerable citizens and communities. 2.3 The primary sources o f data for analysis are the 1988, 1995, and 2000 Ugandan Demographic and Health Surveys (DHS). It i s important to note that data from the 2000 DHS may not reflect the current reality, especially given the reforms that have occurred inthe last four years. The next round o f DHS is needed to shed light on improvement in such health outcomes as child mortality. Additional information from the Ministry o f Health, UN agencies, bilateral donors, NGOs, and peer-reviewed journals i s used for discussion. InUganda, as in most other developing countries, the health care system i s constantly evolving and the challenge over the next few years will be to collect accurate, district-based data for decision making. 13 B. GENERAL PICTUREOFHEALTH HEALTH AND MDGSINUGANDA 2.4 Uganda, according to WHO classification, belongs to the "high child, very high adult" mortality strata (WHO 2000). Communicable diseases, especially HIV/AIDS, malaria, tuberculosis, vaccine-preventable diseases o f childhood, diarrheas, and Acute Respiratory Infections (ARIs) are the major causes o f mortality and morbidity. Life expectancy at birth (42.1 years) has fallen over 22% since the advent o f the AIDS epidemic and ranks 16* among SSA countries (US Census Bureau 2003). Table 2.1 contains data from 1989, 1995, and 2000 for several o f the most important health indicators inUganda. Table 2.1 Trends in healthstatus indicators,Uganda 1989-2000 1989 1995 2000 Infant Mortality 119 97 101 Under five mortality 180 147 151.5 Maternal Mortality 523 506 496 Deliveries supervisedby skilled healthproviders 38% 38% 38% Total Fertility Rate 7.3 6.9 6.9 Contraceptive Prevalence Rate 5% 15% 23% Stunted children 43% 38.8% 38% Life Expectancy HIV prevalence 15%+ 6% Source:MOH 2003 2.5 Uganda has a mixed picture on the MDGs. It i s expected to reach the MDGs for AIDS, TB, and malaria, but not for hunger (nutrition), child mortality, and maternal mortality. See Table 2.2 for a summary o f Uganda's health MDGs. Table 2.3 compares Uganda with other Sub-Saharan countries on a variety o f health and development indicators. Many o f these indicators will be analyzed in turn inthis chapter. It i s important to notice that once more Uganda presents a mixed picture, indicator by indicator (see Chapter 4 for a discussion on the service indicators presentedinthe table). 14 Table 2.2 Summary of Ugandan healthMDGs MDG Indicator 1995 level Current Goal (2015) level Eradicate Hunger Prevalence of moderate to 25.5% ' 23% (2001*) 12.75% severe malnutrition in children under 5 (reduce by %) Reduce Child Under 5 mortality rate (per 147 141(2003) 74 (reduce Mortality 1000) by 213) Improve Maternal mortality rate 1100 880 275 (reduce Maternal Health by %) Combat HIV HIVprevalencerate, among 30% 5% (2002, <(begin to pregnant 15-24yr old women UNAIDS reduce) Fact Sheet) Combat Malaria Malaria prevalence rate, unavailable 46 (2000) (Reverse general population(per upward 100,000) trend) Combat TB Cure rate 33% (1996) 63% (2000) >33% (improve) *httd/millenniumindicators.un.org/ 1UNICEF, www.childinfo.org, accessed 2/11/04 15 z m 00 x z Do C C C h 2 W .C : +L 2r. Ei 9 c 8 S n .e c ? t W 4 r b 9 w Y.f sn z azca W - - 9 c d 9 U LF C Pr. LF aa cE % 00 m b R e c U L c c .. 5CE 'iiec 2 N E cC 2 N +aC 9 C m ! I 4 7EC I. ..c E C c i C f c !i C cc c 5c a C. CHILD MORTALITY Levels of childmortality 2.6 Uganda has one o f the highest levels of child mortality in the world. According to indirect estimates, Uganda's IMR and U 5 M stood at 101 and 151 respectively per 1000 live births in2000 (UBOS and ORC 2001). Globally, Uganda ranks 36th inunder-five mortality rate and 14* inthe total number o f under-five deaths (Black et al. 2003). 2.7 Figure 2.1 again shows a mixed picture for Uganda with high performance compared to SSA countries o f similar income but low performance compared to non-African countries o f similar income. Figure 2.1 examines the association between IMR and U 5 M R and levels o f economic development (Gross National Income per capita) among different SSA countries in 2000. The Afiica-only scatter plot shows that compared to other SSA countries, Uganda has a lower level o f child mortality than expected for its GNI. Figure 2.2, a similar scatter plot for all countries in the world shows that globally Uganda i s an under-performer in child mortality for its level o f economic development. The data suggests that although Uganda has done well, it can do better. 2.8 Uganda i s in the middle o f the pack o f SSA countries on percent reduction o f U5MR. An analysis that grouped SSA countries inseven groups according to the range inU5MR found Uganda to be in the fourth best performing group (10 to 14% reduction) with Ethiopia, Kenya, and Malawi. Better performers were Mozambique with a 25 to 29% reduction, Swaziland with a 20 to 24% reduction, and Angola and the DRC with a 15 to 19% reduction. Worse performers were Tanzania with a 5 to 9% reduction, Madagascar with a 0 to 4% reduction, and Botswana, Namibia, Zambia, and Zimbabwe with increasedU5MR. Trendsin child mortality 2.9 The overall trend in child mortality has been stepwise periods o f improvement followed by periods o f stagnation or slight increase. According to available data from the MOH, there was a significant decline in infant mortality between 1940 and 1970, from 200 to 120 per 1,000 live births. However, this was followed by a period o f stagnation ininfant mortality when IMR stayed around 120 per 1,000 between 1970 and mid-1980s. In the next ten years (mid-1980s to 1995), IMR declined again, from 120 to 97 per 1,000. The later half o f the 199Os, however, registered yet another period with no improvement in IMR, when IMR increased from 97 to 101 (Figure 2.3). This disturbing trend i s confirmed by several sources, including the Uganda DHS which showed an increase in IMRfrom 81 to 88 from 1995 to 2000 and an increase inU5MR from 147 to 152. 17 , Figure 2.1:Infantmortalityandunder-five mortalityvis-&-visGNIof sub-SaharanAfrican countries - 240 i220 I200 g. c 180 100 -: r 5160 4 :140 4 120 100 50J 0 200 400 600 800 1000 ON Rr ciy)ila. Atlas m t h d (FYrlenl Us$) Source: World Development Indicators, World Bank Figure 2.2 Childmortalityand socioeconomicdevelopmentglobally 1 4 ee- ', ,:-.."* ' e Source: World DevelopmentIndicators,World Bank 18 Figure 2.3 Trend ininfant mortality inUganda, 1940-2000 250 0 1 , I 1940 1950 1960 1970 1980 1990 2000 2010 Year Source: MOH 2003 and DHS 2000 2.10 Despite Uganda's child mortality roller coaster ride, Uganda's trends ininfant and under- five mortality over the last forty years compares favorably those o f neighboring SSA countries (Figure 2.4). For example, with regard to decadal absolute reduction in infant mortality, Uganda ranked 5' in 1980-1990 and 2ndin 1990-2000. Changes in compositionof childmortality 2.11 The composition o f child mortality has changed over time in Uganda. Less progress has been made inreducing postnatal death (one month to just under one year o f age) relative to neonatal and age 1-4 years deaths. In the early 1980s, around 50% of child mortality was due to deaths in the 1-4 years age group, while neonatal and postnatal deaths accounted for 27% and 23% o f child mortality respectively. In2000, the share o f deaths inthe 1-4age group was down to 40% ofall childmortality, while that ofpostnatal death increased to 38%. The contribution ofneonatal causes to total child mortality stayed more or less the same (around 22%) (Figure 2.5). The most common causes of death in the postnatal period in developing countries are diarrheal diseases and acute respiratory infections. The high adolescent fertility rate i s discussed later in this chapter. Having a baby as a teen increases the infant's risk o f dying before hisher first birthday by 20% (UBCOS and ORC Macro, 2001). The most common cause o f under-five mortality i s malaria which will be discussed later in the chapter. Data are not available to assess the impact o f HIV/AIDS on child mortality. Pediatric AIDS cases are included in the STD/HIV/AIDS surveillance report. The overall mean age for children with AIDS i s 2.3 years. In addition to AIDS as direct cause o f death, HIV/AIDS also has an indirect impact with highly vulnerable children who are orphans or whose parents have AIDS. 19 Figure 2.4 Trends ininfantmortality and under-fivemortalityinUganda and some neighboring SSA countries 4CmtralAI18c.nRepublic ............... ,* Congo 0.m Rep 01 ---*i- Ethiopia K m y . +Malar# -CMoz.rrbiqu. .... .... 1830 1 no 1880 1893 m 10 Year 1650 1870 1880 1880 1885 2000 Source: DHS 2000 Figure 2.5 Overalltrendsin child mortalityinUganda1980-2000 210 --t1978 /I 180 -+-1982 4 I.. ....................................................................... 1988 .c 19901 '5 150 rIn .--8 1 1 7F -m-1995 0 I //, 0 I+2000 120 k 50" In 90 60 30 Neonatalmrtality Infantmrtality Under five mrtality Source: DHS data and authors' calculations Differentialsininfantmortality 2.12 InUganda as inmost other countries, infant and childmortality varies by family income, area or residence, and mother's education. Each o f these factors will be examined in turn. Income quintile 2.13 Child mortality levels in Uganda, as in almost all other countries, vary by income quintile, but the gaps are lower than in many SSA countries. Figure 2.6 examines inequities in infant mortality in Uganda as well as other SSA countries, using the latest available data from the DHS.Each vertical bar shows the gap in IMRbetween the richest 20 20% and the poorest 20% o f the population in each country. In Uganda, IMR in the poorest quintile was 105.7 per 1,000, or 1.8 times higher than that in the richest quintile (Gwatkin et al. 2002). This trend o f increased child mortality i s also seen in households with markers for low levels of women's empowerment, such as a woman's lack o f decision-makingpower. Both IMR and under-five mortality were higher among children of women with no decision makingpower (13 1and 196respectively) versus women with decision-making power (87 and 155). 2.14 Figure 2.9 summarizes trends inwealth differentials in child mortality in 1988, 1995, and 2000. For each survey, the wealth-based mortality differentialsbecame larger inolder age groups. For example, in 2000, mortality among children o f the poorest quintile were 25% higher than the richest quintile during the neonatal period. This increases to 30% higher for infant mortality and 73% higher for under-five mortality. N o clear general patterns emerge from the data over time, but there were some trends within certain categories. For example, for neonatal mortality, the wealth-based differentials seem to increase between 1988 and 2000. While neonatal mortality in the poorest and richest groups were around the same in 1988, the poor-rich differential increased to 15% in 1995 and 25% in 2000. Similarly, the differential in under-five mortality between the richest and poorest quintiles increased from 17% in 1988 to 51% in 1995 and 73% in 2000. Matemal malnutrition, a contributor to MMR and IMR, has been found to be two to three times more common among women in the poorest quintile versus the wealthiest quintile. This inequality grew from 1995 to 2000 (MOFPED 2002a). The increase in wealth differentialspoints to increasing inequity, at least up to 2000. Figure 2.6 InequitiesininfantmortalityinselectedSSA countries 200- A Poorest quintile Richest quintile -Average ........................................................................................................................................... 180 140 s 120 0 100 80 5v) 0"m 80 1 40 20 ............................... -----------.- ~ .............................................. ~............................ ~.~ ................... I Source: Gwatkinet al. 2002 21 Regional differentials 2.15 There are clear sub-national differences in child mortality echoed throughout all the health indicators analyzed here. The Northem region o f Uganda consistently had the worst rates o f child mortality inall categories (neonatal, postnatal, infant, and under-five mortality) inthe last two DHS surveys. In2000, with an U 5 M R o f 163 and IMR o f 110 per 1,000 live births, U 5 M and IMR inthe North were respectively 13% and 15% higher than those in the Central Region, which was the best performer in child mortality. The Western region had the second highest rates o f infant and under-five mortality, followed by the Easternregion. 2.16 Examination o f the regional trends also reveals that child mortality in all regions did not improve at an equal pace. At the same time, there were regional deteriorations in certain categories. Between 1995 and 2000, the Eastem region was the only one that showed significant reduction in child mortality (11% in infant mortality and 20% in infant mortality). Meanwhile, child mortality was stable in the central region and worsened in the North and the West. IMR increased from 92 to 110 inthe North (a 20% increase) and from 70 to 94 (a 13% increase) inthe West. The Westem region also saw a big increase in U5M, from 118 to 162, or 36% increase. In fact, this is the only region with a deterioration o f U 5 M inthe country between 1995 and 2000. Gender differentials 2.17 Unlike South Asia where there i s a enormous gender gap in child survival, there seem to be no female child disadvantage in child mortality in Uganda. On the contrary, girls consistently had lower child mortality rates than boys inall categories in 1988, 1995 and 2000. The only exception was the fact that in2000, neonatal mortality ingirls was higher than inboys, which inturnleads to a higher female IMR (Figure 2.8). 22 Figure2.7 Regionaltrendsin child Figure2.8 Trends inchild mortalityby gender mortality,Uganda 1988,1995 and 2000 inUganda1988,1995 and2000 Eastern W Norther W Western BJ Female Male I 200 NN I IMR 1USMR 1 IMR 1USMR 0 1995 2000 Figure2.9 Trends inwealth-based variationinchild mortalityinUganda 0Poorest 02nd Poorest HMiddle H2nd richest Richest r$100 2I Q 80- 60- 40 - 20 - 1 0 -- NN IMR USMR USMR USMR 1988 1995 NN ::o 1988 IMR 1995 1988 1995 2000 ~ Source: DHS and author's calculations Mother's education-based differentials 2.18 Mother's education accounts for the largest child mortality differential by far and its effect has increased over time. While differentials in child mortality by mother's education were modest in 1988 (except U 5 M where the differentials were substantial), they widened in 1995 and,2000 to the point that children o f mothers without secondary education were twice as likely to die as the children o fwomen with secondary education. 23 Millenniumdevelopmentgoalfor childmortality 2.19 The international community has set a target o f reducing under-five child mortality by two-thirds between 1990 and 2015 as part o f the Millennium Development Goal for child health. With an U 5 M R o f 175 in 1990, Uganda would need to reduce U 5 M R to 58 per 1,000 live births by 2015 in order to achieve this MDG.Inother words, this requires on average a reduction o f 4.4% per annum between 1990 and 2015. However, according to past trends, Uganda has reduced U 5 M R by 3.8% per annum only. At this rate, Uganda will miss the target in2015. To achieve MDGfor childmortality, the country will needto accelerate the pace o f reduction from 3.8% to 4.8% per annum (Figure 2.10). Figure 2.10 Under-five mortality trends and Millennium Development Goals -Uganda ----C--.-Uganda(2000 (1990 to 2015) rate of decrease of 4.4 percent required (2000 to 2015 required decreaseof 4.8 percent per pi f Uganda to 20151: at current rate of decrease,miss MDG I 180 1 ..................................... 4 ~~~............... Source: UNICEF,www.chillinfo.org D. CHILD MORBIDITY 2.20 Child morbidity is an important subject that does not receive as much attention as it should. In addition to human suffering, when children are ill,the following important consequences occur: children miss learning opportunities at school, adults have to stay home from work to care for them, money i s spent for their care, the children cannot complete household duties, children are left with life-long disabilities, and their risk o f mortality i s increased. Although there are few data on child morbidity inUganda, what i s available has been analyzed here along with disease-specific child mortality data. 24 Specificdisease contributionsto childmorbidityandmortality Child malnutrition 2.21 Child malnutrition is the component o f child morbidity that i s most studied. When compared to the levels of malnutrition among SSA countries, Uganda falls in the middle o f the spectrum. Uganda's rate o f severe underweight, 6.7%, falls between Zimbabwe's 1.7% and Ethiopia's 16%. Uganda's rate o f severe stunting o f 15% falls between Gambia's 6% and Ethiopia's 25.9%. Uganda's rate o f severe wasting o f 1% falls at the low end compared to 5% seen inNigeria. 2.22 Child malnutrition starts in the womb. InUganda, 12% o f babies are born underweight which can lead to poor cognitive function and increases risk o f mortality. About 35% o f all under-five deaths in Uganda are directly or indirectly caused by malnutrition (BASICS 1995). 2.23 Figure2.11shows the wealth-based inequalities for weight-for-age (underweight) inSSA countries. Uganda comes out well, with lower than average total percent o f children underweight and lower levels o f wealth-based weight inequalities. Zimbabwe i s one o f the few countries which seem to have not only lower levels o f malnutrition than Uganda but also lower levels of inequalities in malnutrition. On the other hand, countries like Nigeria, Niger, Mali, etc. have higher levels o f malnutrition as well as higher levels o f wealth-based inequalities. Figure2.11Wealth-basedinequalitiesinsevere underweight in SSA 25 +Poorest 20% -f- Richest20% 20 15 10 5 0 Source: Gwatkinet al. 2002, World Bank 2.24 DHS surveys from 1988, 1995 and 2000 provide an opportunity to assess the trends in malnutrition. Figure 2.12 shows the trends inmalnutrition between 1988 and 2000, which show a fall in stunting between 1988 and 1995, but stagnation between 1995 and 2000. 25 However, there was slight increase in underweight between 1988 and 1995, but subsequent drop between 1995 to 2000 to the levels seen in 1988. 2.25 Clearly, wealth-based inequalities in malnutrition are seen (Figure 2.13 and Figure 2.14). The fall inlevels of stuntingand underweight inthe poorest quintile i s less than the fall in the richest quintile of children. However, the concentration curves (Figure 2.15) reveal that the inequities in malnutrition are slowly disappearing as the curves for the 2000 survey are closer to the equality line than previous curves. 2.26 Urban-rural differentials in malnutrition are very prominent in Uganda. For most o f the malnutrition indicators, the malnutrition in rural areas i s almost double o f that seen in urban areas. For example, the percent o f children severely underweight i s 5.2 in rural areas compared to only 1.8% in urban areas. Mother's lack of education i s highly correlated with child's malnutrition. Children o f women without any education have twice the levels o f malnutrition as the children o f women with secondary education. Higher parity children have slightly higher malnutrition than children o f lower parity. Children whose preceding birth interval i s longer have lower malnutrition. Male children have higher malnutrition than female children. Malnutrition level increase from birthand peak in 12-23 months age at the time o f weaning, later dropping slightly. 26 Figure2.12 Trends inmalnutrition Figure 2.13 Trends insevere stuntingby wealth quintiles 1988 H1995 2000 1988 E1995 02000 ............................................................. ........................................................ ............................................................... Em ................................... .................................. ............................ ........... Poorest 2nd poorest Middle 2nd richest Richest Figure2.14: Trendsinsevere underweightby Figure2.15: Trends inmalnutrition wealth quintiles concentration curves El1988 H1995 2000 I Poorest 2nd middle 2nd richest Richest poorest C"ummul"ative$oport%n of fipulat%n byb?ealthyuintil& Source: DHS 1998, 1995,2000 2.27 Regional nutrition disparities are decreasing, but still a problem: the Central region has the lowest levels o f malnutrition, followed by Eastem; the highest levels o f malnutrition are seen in Northern and Westem regions. Religion-based differentials in malnutrition were not prominent. Although differences inmalnutrition by mother's age were seen, no clear pattem emerged. 27 Other illness: ARI anddiarrheal diseases 2.28 ARI and diarrheal diseases are both major causes Table 2.4 ARI and diarrheaprevalence in childrenin morbidity and mortality selectedAfrican countries among children in Uganda. It is accuratedifficultontoA Rcollect Percent of children under Percent of children under 5 5 with ARI duringtwo with diarrhea inthe two diarrhea because A R IIss and CAR data and Country weeks prior to the survey weeks prior to the survey 28.2 26.5 diarrhea represent a variety Nigeria Chad 12.5 31.2 11.3 15.3 oabout f different illness brought 24.4 23.6 by different causes Kenya 20.1 17.1 (mainly infectious) and Malawi 12.3 16.1 families can have different Mozambique 11.8 20.7 definitions o f ARI or Tanzania 13.9 12.4 diarrhea. Based on burden ZambiaUganda 27.1 23.5 12.7 23.5 ochildren f diseaseunder studies, 22% o f Zimbabwe 15.8 13.9 five years Source: UNICEF 2000 old were reported to have suffered episode infrom an ARI Figure2.16 Trendsinfever and diarrheaprevalencein the last two 0-4yearsolds weeks3. This rate AFU i s down slightly form 28.8% in 1995. 50 Fever 0Diarrhea ............................................. 2.29 The burden o f disease 45 studies revealed that 40 t? currently 40% o f children E35 under five had diarrhea In 230 within the last two weeks (Mbonye 2003). Compared 25 with 1995, diarrhea Ee,20 incidence has remained c 15 unchanged in children under six months o f age but i10 W has increased slightly in 5 children 6 to 11 months 0 from 33% in 1995 to 38%in 1988 1995 2000 2000. Source: DHS 1988, 1995 and 2000 2.30 Table 2.4 provides comparison o f ARI and diarrhea prevalence rates in different SSA countries. Uganda has a child ARI prevalence rates o f 27%, which i s one o f the highest in SSA countries. Likewise, prevalence of diarrhea i s also at the higher end at 23%. 2.3 1 According to the DHS data, the prevalence of ARI, fever, and diarrhea i s much higher in rural areas compared to the urban. However, between 1995 and 2000, there was a larger reduction inARI prevalence inrural than inurban areas, although in absolute terms, rural areas still had higher ARI prevalence in 2000. The Central region had the lowest ARI, diarrhea and second lowest fever prevalence in2000 and 1995. The Northern and Eastern http:llwww.unicef.orglinfobycountry/uganda-statistics.html 28 regions had the highest child morbidity. Children from richest quintile had lower levels o f fever, ARI and diarrhea compared to the children from the poorest quintile. No religion- based differentials in child morbidity was seen. Children o f older mothers had higher likelihood o f diarrhea and ARI in 2000. No gender-based differentials were seen in child morbidity. Younger children had higher likelihood o f ARI, fever and diarrhea. E.MATERNALMORTALITY 2.32 Matemal mortality i s one o f the most important cause o f loss o f healthy life years in Uganda4, and places surviving children at great risk, as the death o f the mother has been found in a variety o f settings to be a important predictor o f school dropout and child mortality. The Burden o f Disease Study found that maternal deaths and perinatal deaths (which are affected by delivery care) made up 20% of the life-years lost to premature death. Unfortunately, for such an important indicator, there i s conflicting data on maternal mortality ratio (MMR) in Uganda. The WHO estimate o f MMR in Uganda in 1995 i s very high-1,056 per 100,000 live births-which places Uganda in the bottom third o f SSA for MMR.According to the DHS 1995 and DHS 2000, the MMRestimates for Uganda were 527 and 504 matemal deaths per 100,000 live births respectively. This estimate would place Uganda among the top quintile of SSA for MMR. Adolescents account for 44-48% o f all maternal deaths, largely due to unsafe abortion (MOH 2003b). 2.33 Figure 2.16 shows the scatter plots of MMR by Gross Domestic Product (GDP) per capita for sub-Saharan countries and the world. Globally, Uganda i s among a cluster o f sub-Saharan countries with highMMR and low per capita GDP and i s to the right o f the trend line, showing that Uganda's MMR i s higher that other countries with comparable per capita GDP. On the SSA plot, however, Uganda is on the left side o f the trend line, showing that Uganda i s somewhat better than other SSA countries with comparable per capita GDP. Once more it shows that although Uganda has performed on par with its neighbors, it could do better. 2.34 The majority o f matemal deaths occur outside the hospitals, indicating that delay in seeking care and delay in reaching care are critical. Of the deaths that occur in the hospital, many o f the women arrived at the hospital moribund. In one study, 86.1% o f maternal deaths in 74 facilities occurred within an hour o f arrival (Mbonye 2001). Although inpatient management certainly needs to be improved, these data point more to lack o frecognition o f waming signs and lack o ftimely transport. [text for this footnote seem to bemissing.] 29 Figure 2.16 Estimated maternal mortality ratios for selected countries inthe SSA and World, 1995 2500 2.500 Rwanda SmRaLco". 2000 2,000 Ethiopia Gi c 2 0 1500 Eg - 1.5w 0 ,*B2f 2 -; p $004 PrE ; 0 1,000 2e!E e 500 500 0 C 100 1000 10000 100 1000 10000 100000 GDP per capita PPP (Internatlonalt)WDI 2002 GDP per capita, PPP (current international $) ZOO0 Source: Maternalmortality from WHO, 2000 and GDP from WDI, 2002, World Bank 2.35 The MDG on maternal health sets the target o f reducing o f MMR by three quarters between 1990 and 2015. For Uganda, this would involve reducing MMR from around 1,200 in 1990 to approximately 300 by year 2015. African countries with similar GDP per capita such as Mozambique and Kenya have widely varying MMRs with 1,100 and 590 respectively. Currently, only countries with more than twice the per capital GDP of Uganda consistently are able to achieve less than 400 MMR, so it will be quite a challenge for Uganda. Recent work in Uganda has demonstrated that clinical protocols for common life-threatening maternal conditions such as postpartum hemorrhage and pregnancy-induced hypertensive disorders can drastically reduce maternal mortality. 2.36 MMR is one of the most difficult health outcome indicators to measure because of relative rarity o f events and need for large sample sizes to calculate the point estimates with a fair degree of confidence. Hence, other output indicators can be used to monitor progress, e.g. the percentage o f deliveries attended by skilled professionals which this report examines inchapter 4. 30 F.ADULT MORBIDITY 2.37 The Uganda National Household Survey (UNHS 1999/2000 and 2002/2003) sought information from respondents about any illnesses and injuryduring the last 30 days o f the survey. Although the accuracy o f self-reporting i s often an issue for this type of questionnaire, especially with regard to the type o f illness, information from the U N H S can points to a general picture of overall morbidity in Uganda. The data on self-reported ailments in the 1999/2000 and 2002/2003 surveys are presented in Table 2.5 and Table 2.6. More than 50% o f the respondents in both surveys mentioned fevedmalaria as the main aliment, followed by respiratory problems, diarrhea and intestinal problems. Between 1999/2000 and 2002/2003, the share o f fevedmalaria in total morbidity was reduced from 56.2% to 50.9%. Fevedmalaria, measles and diarrhea were prominent in the 0-5 age group, while HIV/AIDS was muchhigher inthe 20-50 age group. Table 2.5 Illnesslinjury during last 30 days of survey (1999/2000 survey) Total 56.2 5.1 0.8 2.9 0.2 1.3 12.2 6.2 3.0 1.4 0.7 10.0 Residence Rural 55.8 5.4 0.9 2.9 0.2 1.4 11.8 6.4 3.0 1.4 0.7 10.2 Urban 58.8 3.3 0.4 2.5 0.4 0.5 15.2 4.5 3.2 1.9 0.7 8.6 Gender Male 55.9 5.3 1.3 3.2 0.3 1.1 12.0 5.6 3.5 0.9 0.6 10.4 Female 56.4 5.0 0.4 2.6 0.2 1.4 12.4 6.7 2.5 1.9 0.7 9.7 Income quintile Poorest 49.2 8.2 0.7 3.4 0.3 1.3 13.5 6.8 3.1 1.6 0.2 11.8 2"dPoorest 52.4 8.1 0.9 2.2 0.1 1.2 12.3 7.0 3.8 1.0 0.6 10.5 Middle 54.4 4.2 1.0 3.6 0.2 1.4 14.3 5.9 2.7 1.4 0.3 10.7 2"drichest 57.2 4.4 0.9 2.7 0.4 1.4 11.5 6.4 2.9 1.6 0.9 9.7 Richest 59.3 4.3 0.7 2.8 0.2 1.2 11.3 5.8 2.9 1.5 0.8 9.3 Age category 0-5 years 61.3 9.2 0.2 5.9 0.1 0.4 11.6 4.2 2.6 0.2 0.0 4.2 6-20 years 60.3 3.6 1.1 3.0 0.1 1.6 11.6 5.7 4.0 0.6 0.3 8.2 20-50 years 54.5 3.3 1.1 0.5 0.6 1.4 12.3 7.5 2.0 1.9 1.1 13.8 >50 vears 36.9 2.8 1.4 0.4 0.1 2.1 15.0 9.4 3.7 6.0 2.0 20.1 Source: Uganda National HouseholdSurvey 199912000 31 Table 2.6 Illnesshjury during last 30 days of survey (2002/2003) Total 50.9 4.0 1.1 3.0 0.2 1.1 14.2 4.4 3.2 1.0 0.9 8.8 Residence Rural 55.3 4.3 1.0 3.1 0.3 1.1 13.8 4.6 3.4 1.0 0.9 11.2 Urban 60.2 2.2 1.3 2.1 0.2 1.0 16.2 3.0 1.8 1.3 1.3 9.6 Gender Male 57.4 4.1 1.9 3.0 0.2 1.0 14.7 3.3 3.6 0.6 0.6 9.6 Female 54.7 3.9 0.4 2.9 0.2 1.2 13.7 5.3 2.9 1.4 1.2 12.3 Income quintile Poorest 47.7 7.4 0.6 3.4 0.1 0.7 14.8 5.3 4.7 0.5 0.8 14.1 2"dPoorest 55.6 4.5 1.1 3.9 0.0 1.5 12.4 5.4 4.1 0.6 0.7 10.3 Middle 58.2 3.4 0.6 3.2 0.1 1.0 14.1 4.5 2.9 1.4 0.7 9.9 2"drichest 59.6 3.2 1.5 2.3 0.8 1.2 13.0 3.8 2.7 1.4 1.1 9.6 Richest 57.7 2.1 1.5 2.0 0.2 1.1 16.3 3.2 1.9 1.3 1.2 11.5 Age category 0-5 years 63.0 7.3 0.4 6.1 0.0 0.4 13.6 2.2 2.9 0.1 0.0 4.0 6-20 years 57.0 2.3 1.2 3.3 0.1 1.0 15.4 4.5 5.2 0.4 0.5 9.2 20-50 years 56.2 2.5 1.4 0.6 0.3 1.8 12.1 5.2 2.4 0.4 1.1 15.9 >50 years 50.6 2.1 1.7 0.4 1.0 0.9 15.1 5.2 2.3 2.3 2.3 16.3 Source: Uganda National Household Survey 2002/2003 HIV/AIDS 2.38 Uganda i s the global success story in HIV/AIDS. HIV prevalence in the general population was 15 to 20% in the early 1990s and was brought down to 6.5% in 2001, which represents the most significant decline in HIV prevalence in any country worldwide (Synergy Project 2003). Although the tide o f the epidemic has turned, there are still 600,000 to 1.2 million people living with HIV/AIDS (PLWHA) in Uganda and the prevalence i s still increasing inrural areas (Steen et al. 2001, Synergy Project 2003)'. 2.39 More than half of all PLWHA as well as newly infected PLWHA are women. This observation, seen throughout SSA, has several possible causes: i)the higher transmission risk per infected sexual contact for women; ii)the observed pattern o f men having a higher number o f sexual partners, thus a single infected man can spread HIV to multiple women; iii)lower awareness among women o f how to protect themselves from HIV (78% versus 90% (STD/AIDS Control programme 2002)). This higher infection rate 5ht~://www.ae~is.codnews.ap/2002/~021115.html. Accessed 040504 32 among women requires increased behavior change communication targeted to high-risk women and gender sensitive strategies. 2.40 A recent study of the natural history o f HIV inpeople inrural Uganda revealed that the median time from seroconversion to A I D S i s 9.8 years, equivalent to what i s observed in developed countries. However, the median time from AIDS to death i s a mere nine months, much shorter than what was observed in developed countries before the introduction of antiretrovirals (ARVs). This points to the need for improved nutrition, basic primary care, and treatment o f opportunistic infections for PLWHA. 2.41 Figure 2.17 shows the relative position o f Uganda with respect to both adult prevalence rates for HIV/AIDS as well as estimated number o f adults living with HIV/AIDS. The HIV/AIDS epidemic has hit Uganda hard, but compared to some o f the other countries in SSA, Uganda has lower levels o f adult prevalence rates. The size o f Uganda's population also causes the total burden o f HIV/AIDS to be ranked as middle o f the spectrum of all SSA countries. Figure 2.17 Number of adultsinfectedwith HIV/AIDS and adult prevalenceratesof HIV/AIDSin SSA 40 , , m G3 r E ..................................................................................................... ui0 30 P $ 25 -......~ .............. ....... .......e.s~la2iland--.... .4zimeaewe-.- ~ ............... ~~~.. ~ ~.. .~.... '5E2 +Lesotho --f-s 20 amibia +Malawi P 15..................................................................................................... +Mozam$&tnya L Pm +Central African Republic z +Djibouti p 10 +Rwanda+ -............................... Cote d'lvoire + ~~~ ~~ ~~ ~~ .~....................................~. ~*E?'PP!B.. ....... .Ew" U nda +C Tanzania o ~ @ o ~ ~ ~ BurkinaFaso + +Calj(%roon ~ 5 ................................... .... +Gabon +Ghana g1g"')"..-4A~ia- ~ ~........................... ........... Leone #Gg,y&a-~issauE'itre~ $Sierra + Chad%apgola Guinea 4; &gkFIWal 0 Source: UNAIDS 2002 2.42 Despite its dramatic decline in Uganda, HIV/AIDS has resulted in immense human suffering, reduced life expectancy, deepened household poverty, and retarded development in Uganda, just like elsewhere in sub-Saharan Africa over the last two decades. Uganda's life expectancy has been reduced from 62 to 42 years due to AIDS, which i s the leading cause o f death among adults, and AIDS accounts for the majority o f Uganda's two million orphans. The remaining challenge i s to further reduce HIV 33 prevalence, especially in vulnerable groups and young people and to avoid the complacency that can follow success. Tuberculosis 2.43 The TB epidemic has increased with the A I D S epidemic in Uganda and it i s estimated that 50% o f TB patients in Uganda are HIV positive (Jareg et al. 2003). Over 34,000 new cases are detected a year, with the highest number among adults intheir 20s and 30s. Although Kampala has only 5% of the population, it reported 25.4% (9,395 o f 36,956) o f the TB cases in2001 (STD/AIDS Control Programme 2002). This high rate in Kampala may be due to the higher HIV prevalence, the crowded urban conditions that increase the spread o f TB, or better case-finding and reporting. The mortality rate i s 4.1% among identified cases. 2.44 Although the detection rate for new smear positive cases has decreased from 61% in 1997 to 52% in 2001, the cure rate for new smear positive cases has increased from 40% in 2001 to 63% in 2000 (Jareg et al. 2003). There i s wide variation in cure rate by district, from 23% inMoroto and 90% in Apao. Despite the progress in cure rates, Uganda's TB cure rate i s the worst in all SSA at 63%. Interms o f incidence, Uganda i s on par with the other high burden countries, with 324 TB cases per 100,000 population, compared to Zimbabwe with 628 and Nigeria with 235 (Table 2.7). Uganda, like most other highTB incidence countries, reports more cases in men than women in the older then 25 age groups (Stop TB 2003), which represents a combination o f a variety o f effects: men's increased susceptibility to TB, men's increased exposure, and underreporting o f cases among women. Table 2.7 TB data for highburden SSA countries, by incidence Country Incidence/100,000 DOTScure rate (%) TB Ranking * Zimbabwe 628 69 17 Kenya 515 80 11 Tanzania 344 78 14 Uganda 324 63 20 D.R. Congo 302 78 12 Etiopia 292 80 10 Mozambique 265 75 27 Nigeria 235 79 5 *(Absolute # of cases) Source:WHO Annual TB Report FevedMalaria 2.45 Ugandans suffer from 1.5 million cases o f malaria a year and 95% o f the country has endemic malaria. Because o f this, every fever must be presumed to be malaria until proven otherwise. Epidemics occur only in the high-altitude southwest where malaria transmission i s unstable and as a result, deaths from endemic malaria far exceed deaths from epidemic malaria. For example, the most recent malaria epidemic in Uganda occurred in 1998 and killed between several hundred to less then 1000people, compared to the more than 40,000 endemic malaria deaths that occurred that same year in the rest o f Uganda. Malaria i s the number one cause o f morbidity and mortality inUganda (even exceeding HIV/AIDS) and the majority o f its victims are children due to their biological vulnerability. Malaria causes 40% o f all outpatient visits, 25% o f inpatient stays, and 14% o f inpatient deaths (Root et al. 2003). There i s a disturbingtrend o f drugresistance, 34 with 81.1% resistance to chloroquine and 25% resistance to Fansidar (pyrimethamine/sulfadoxine) in the country in 2000, with some regions with Fansidar resistance levels o f up to 60% (Figure 2.18). Figure2.18 Chloroquinetreatment failure inAfrica Chloraqulnetreatment fallufe InAfrica 2.46 Because malaria i s the leading cause o f child mortality in Uganda (UNICEF 2002), the majority o f malaria deaths occur in children, Inheavily affected areas, malaria i s the main cause o f pediatric hospitalizations and the main cause o f deaths in pediatric wards. Inchildren, malaria causes death by severe anemia and cerebral malaria. It also causes stillbirth,premature birth, low birthweight, stunting, cognitive blunting(via anemia), and vulnerability to other infections. Socioeconomic data on malaria in Uganda are not available, however, it is known that inthe internally displaced, malaria cause between 20 to 30% o f all morbidity and mortality (Root et al. 2003). 35 2.47 Figure trends in2*19 and diarrhea o-4 years olds fever shows the Figure 2.19 Trends infever and diarrhea prevalencein for children in Uganda between 1988 and 5000. Prevalence o f fever seems 50 to have increased between 45 1988 and 1995 and fallen % 40 slightly between 1995 and ln K35 5 2000. Prevalence o f diarrhea seems to be slowly 230 C decreasing. fI25 5920 L :I5 h o 5 0 1988 1995 2000 Source: DHS 1988, 1995 and 2000 G.FERTILITY AND DEMOGRAPHICS Fertility 2.48 The crude birthrate o f Uganda i s 45.4 per 1000 people-one o f the highest in SSA-and only less than faster growing neighbors like Somalia (50.9) and Niger (50.6). Uganda's total fertility rate (TFR) i s 6.9 and i s higher than the SSA average o f 5.2 and much higher than world average o f 2.7. The population growth rate o f Uganda i s 2.7%, which i s higher than the sub-Saharan average o f 2.4% although less than few o f the fast growing neighbors like Somalia (3.6) and Niger (3.3). Progress has been made in reducing teen pregnancy from 43% to 31%6. 2.49 The total fertility rates (TFR) o f Uganda have remained almost constant between 1995 and 2000, with moderate decrease between 1988 and 1999. The TFR decreased from 1.3 births per women in 1988 to 6.9 births in 1995, and have remained constant since then. However, TFR i s much higher inrural areas (7.4 children per women) compared to urban areas (4 children per women) (Figure 2.20). Age specific fertility rates, in fact, show that the fertility inthe 15-19 age group has increasedbetween 1988 and 2000 ( 2.50 Figure 2.23). The TFR also differed between the women in the poorest quintile and the richest quintile. The TFR has changed little in the poorest quintile between 1988 and 2000, and has remained very high: 8.5 children per woman compared to 4.1 among women in the richest quintile, making a difference o f around four children between the poorest and richest quintile (Figure 2.21). Women who had no education had the highest A Report onthe Public Discussion onPersistently HighInfant and ChildMortality Rates inUganda-a Function of Poverty." 2002. 36 TFR. Women with secondary education showed TFR dropping dramatically from above five in 1988 to 3.9 in2000 ( 2.5 1 Figure 2 . 2 2 t a difference o fnearly four children between the uneducated and secondary education women. Demographics 2.52 Fifty-one percent o f the Ugandan population i s under 15 years o f age, making schooling and child and adolescent health vital to Uganda's success. Due to the highfertility rates, decreased life expectancy, and the loss o f many middle aged people to AIDS, Uganda will have a highdependency ratio inthe future (the ratio o f those under 15 and over 65 compared to those 15 to 65 years old). The highdependency ratio means that there will be many unsupported children and elders and that each breadwinner will have to support more people. 37 H.POSSIBILITIESFORMEETING MDGS THE 2.53 As discussed for eachindividual MDG,ifUgandacontinues with its current performance, it will meet half o f the health MDGs (AIDS, TB, and Malaria) while missing half (hunger/malnutrition, child mortality, and maternal mortality). Uganda will need to develop ways to deliver BCC, preventive, and curative services in its areas affected by conflict, particularly the Northem and Western regions. These two regions contain 11.7 million people, roughly half the entire population (UBOS 2004). These regions also have the highest population growth rate (Northem lo%, Westem 4.6%, Eastem 2.3%, Central 3.6%) which will magnify their effect on the nation's MDGs. Uganda will not meet the MDGsunless it tackles these two regions. Extra effort inCentral andEastemUgandaare unlikely to reap muchbenefit unless they are targeted toward the bottomtwo quintiles. 38 Figure 2.20 Trends in TFR by residence Figure 2.21 Trends in TFR by wealth quintiles El1988 1995 02000 E4 1988 1995 02000 Y 7 8 ....... 6 7 6 5 E 4 E 5 4 3 3 2 2 1 1 0 0 Poorest 2nd Poorest Middle 2nd Richest Richest Residence Figure 2.22: Trends inTFR by women's education Figure 2.23 Trends in ASFR 1988 1995 02000 1988 1995 02000 I 8 .................. ....................... 7 .................... 6 E 5 + 4 3 2 1 0 No education Primary Secondary 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Women's education Age category Source: DHS 1988, 1995 and2000 39 3. KNOWLEDGE,BEHAVIORAND PRACTICES AFFECTINGHEALTH A. INTRODUCTION 3.1 Health, nutrition, and population (HNP) outcomes are influenced by an array o f determinants and various interactions among them. Such determinants can be categorized as underlying (e.g. income, education, cultural and social norms), intermediate (e.g. health-related knowledge, behavior, and practices) and direct factors (e.g. health services and utilization o f health services). 3.2 Individual and household knowledge and practices can tremendously influence health outcomes, not withstanding the availability o f health services. The household i s regarded as an important 'producer' o f health. The HIV/AIDS epidemic i s a prominent example o f how knowledge, behavior, and practices can affect H" outcomes. A good understanding o fthese intermediate factors provides the basis for appropriate interventions. 3.3 This chapter examines individual and household health-related knowledge, behavior and practices inUganda. The discussion inthis chapter i s mostly based on the findings o f (i) three Demographic and Health Surveys conducted in 1988, 1995, and 2000 and (ii) the 2003 Family Care Practices Survey conducted by UNICEF. The latter i s not nationally representative, as it has a small sample size and was conducted in four districts under the UNICEF district capacity buildingsupport programand four randomly selectedones. B.THEOVERALLPICTURE 3.4 Compared to other countries insub-Saharan Africa, Uganda has relatively highlevels o f certain health-related knowledge and practices such as those relating to breast-feeding and HIV/AIDS. However, there i s still much room for improvement in other areas, namely nutrition, home care for the sick child, malaria, family planning, maternal health, sanitation, and alcohol consumption. In some cases, high levels o f health-related knowledge have not resulted ingood behavior and practices. 3.5 There are also significant disparities in health-related knowledge and practices by residence, wealth, and educational levels. Generally speaking, residents o f the Eastern and Central regions, secondary school graduates, and those in the richer wealth quintiles tend to have better knowledge, behavior, and practices inhealth and nutrition. 3.6 Inits fight against HIVIAIDS,Ugandahas demonstrated that behavior change for healthy outcomes i s possible. It i s crucial that this success story be replicated in other individual and family health practices. There i s a need for a multisectoral plan to promote healthy practices which i s backed up by adequate commitment and resources. Such a plan needs to include concrete actions which meet the needs o f the most vulnerable groups such as 40 the poor, the uneducated, and those living in the Northern region. Social mobilization and community participationi s a key to the success o f such activities. c.HEALTH-RELATED KNOWLEDGE, BEHAVIORAND PRACTICESAT THE HOUSEHOLD AND INDIVIDUAL LEVELS 3.7 Knowledge alone i s often insufficient to bringabout health promoting behavior, Material circumstances mustbe such that people can translate their knowledge into positive action. Knowledge i s acquired in many ways. Much health information i s provided by written materials, but this i s within a context where the literacy rate for females over age 15 i s only 59%: in rural areas and among older women, perhaps more than half cannot read. Other media are employed for spreading health messages but considerable wealth disparities exist. Radio listenership among the poorest quintile o f women i s 19% compared to 87% among the best off; television viewership i s 1.1% among the poorest compared to 35% among the best off. 3.8 The inverse relationship o f women's education to child survival i s well established in all regions o f the developing world. There are further linkages between women's education and preferred family size, fertility, and autonomous decision making. These are part o f a picture which includes awareness of good health practices, greater ability to deal with their environment, and increased ability to make decisions relating to their own and their children's well-being (Jejeebhoy 1995). 3.9 The ability to make decisions i s affected by differential decision-making power in the household. The GoU has demonstrated its appreciation o f this important aspect o f health outcomes by collecting data on six aspects o f household decision-making in the 2000 DHS. Ugandan women fare relatively well compared to women in other African countries and in other regions where these data have been systematically collected. Nevertheless, the levels o f decision-making are worrying low: 49% can seek health care for their children (this compares unfavorably with SSA overall where 75% o f women reported being able to make these decisions); 44% can decide to seek their own health care; 43% can make daily household purchases; 37% can make large household purchases; 67% can make meal-related decisions; and 29% can travel to visit family members. Nutrition 3.10 As discussed in Chapter 2, malnutrition i s an underlying factor o f many childhood illnesses and maternal conditions. Generally speaking, Uganda i s doing relatively well in terms o f exclusive breast-feeding and the timing o f complementary feeding for infants. However, a highlevel o f under-five malnutrition as discussed in Chapter 2 indicates less- than-optimal feeding practices for young children beyond infancy. The 2003 UNICEF survey in eight districts reveals that only 39% o f caregivers have ever received information on the nutrition requirements o f young children (UNICEF 2003). Initiation of breastfeeding 3.11 The percentage of infants breast-fed within one hour o f birth dropped from around 50% in 1995 to 30% in 2000 (Uganda Bureau of Statistics and ORC Macro 1996, Uganda Bureau o f Statistics and ORC Macro 2001). Inthe 2003 UNICEF survey, this percentage was 54%. Thus the practice o f early initiation o f breastfeeding i s still not common. Little 41 variation exists in the initiation o f breastfeeding by individual and household characteristics. Inaddition, during the first three days o f life, a substantial proportion o f newborns (43%) receive unnecessary and potentially harmful prelacteal feeding, which involves food other than breast milk and increases the risk o f diarrhea (Uganda Bureau o f Statistics and ORC Macro 2001). Exclusive breast-feeding infirst s k months 3.12 In 2000, although the percentage of infants ever breast-fed was high at 98%, the proportion o f those receiving exclusive breast-feeding under six months in accordance with WHO recommendations was significantly lower at 63% (Uganda Bureau o f Statistics and ORC Macro 2001). The level of exclusive breast-feeding among infants under six months was only 54% in the 2003 UNICEF survey. Among the internally displacedpopulations inthe North, this practice was even lower with only 39% o f infants under four months being exclusively breastfed (Oxfam 2001). However, compared to other countries in the region, Uganda i s still doing relatively well on this indicator o f child feeding (Figure 3.1). 3.13 While mother's education does not seem to be related to this practice, wealth-based differentials exist in Uganda. It was more common for infants o f poor households to be exclusivelybreast-fed under four months (Table 3.1). Complementaryfeeding among children 6-9 months 3.14 DHS 2000 data show that 70% o f the breast-feeding children age 6-9 months receive solid foods as com- plementary feeding Figure3.1 Infant feedingpracticesinUganda and selected (Uganda Bureau o f countriesinthe region, latest available data Statistics and ORC Macro 2001). Hence, Uganda's performance on -- OExclusively breastfed (0-6 months) this indicator i s 70 BBreastfed with complementary relatively good. As p6 60 food (6-9 months) 50 in the case o f 2t! 40 exclusive breast- 30 feeding, infants o f 20 poorer households 10 are more likely to 0 receive complementary feeding between six and nine months (Table 3.1). Source: UNICEF2004 Micronutrition 3.15 The use o f iodized salt is almost universalinUganda (Table 3.1). However, compared to other countries in the region, Uganda does less well in terms o f vitamin A 42 supplementation (Figure 3.2). In2000, only 43% of children under five receivedvitamin A duringthe six months before the survey, and 11%o f women with a recent birthtook vitamin A within two months of the last birth (Uganda Bureau o f Statistics and ORC Macro 2001).Wealth-based differentials for vitamin A supplementation among pregnant women were large. Those in the richest quintile were almost four times more likely to take Vitamin A within the last two months o f the last birth than those in the poorest quintile (Table 3.1). Childhooddiarrhea 3.16 Diarrhea i s one o f the most important causes o f under-five mortality and morbidity in Uganda. Most o f available studies on diarrhea-related knowledge and practices show that caretakers knew the symptoms and danger signs o f diarrhea, and over half o f them correctly identified most o fthe riskfactors associated with diarrhea (Twebaze 2001). 3.17 Appropriate child feeding practices during a diarrhea episode can expedite recovery and reduce child mortality associated with diarrhea. However, in 2000, among children with diarrhea, 41% were actually given less fluid or none at all and only 28% were given more fluids than usual. Similarly, 51% of children with diarrhea were given less food or none at all, while only 5% were given more food than usual (Uganda Bureau o f Statistics and ORC Macro 2001). Feedingpractices for children with diarrhea seem to have worsened between the 1995 and 2000 DHS. The findings o f the UNICEF 2003 survey in eight districts point to some possible recent improvements inthis regard (Table 3.2). Figure3.2 Micro-nutrition practicesinUgandaand selected countriesin the region, latest availabledata 100 1 DChildren receivingvitamin A supplementation BHousehold using iodized salt Uganda Tanzania Kenya Ethiopia Mozambique Source: UNICEF2004 43 Table 3.1 Levelsof different nutrition practicesby wealth quintilesin 2000 Indicator Description Quintiles Population Lowest Second Middle Fourth Highest average Breast feeding.. Exclusive % o finfant 72.6 88.8 73.1 78.8 58.7 74.9 breast-feeding under 4 months exclusively breast-fed Timely % o f infant from 79.2 74.1 79.3 74.6 63.9 74.6 complementary 6 to 9 months breast-feeding receiving complementary breast-feeding Micronutrient consumption Vitamin A % o funder-five 39.9 39.7 44.2 43.6 48.8 42.9 consumption in children children receiving vitamin A inthe 6 months before the survey Vitamin A % o fwomen 5.6 7.5 9.4 14.5 21.7 11.3 consumption in with a recent women birthreceiving vitamin A within 2 months o f the last birth Source: Gwatkin et al. 2004 Table 3.2 Feedingof under-fivechildrenwith diarrhea (percentage) 1995 2000 2003 @HS) (DHS) (UNICEFsurvey) Less or no fluids 30 41 26 More fluids than usual 49 28 40 Less or no food 61 51 40 More food than usual 13 5 24 Note: 2000 DHS and 2003 UNICEF survey limitedthe questionnaire to a diarrhea episode inthe two weeks preceding the survey, while the timeframe for the 1995 DHS was four weeks. Source: DHS 2000 and 2003, UNICEF2004 3.18 The level of mothers' knowledge of oral rehydration salts (ORS) i s high.Around nine out o f ten mothers know about the use o f ORS for the treatment o f diarrhea in children. Urban mothers, mothers living in the Eastern and Central regions, and those with education are more likely to h o w about ORS (Uganda Bureau o f Statistics and ORC Macro 2001). However, the use o f ORS remains low, estimated to be 11%. Such low use i s attributable to poor knowledge o f ORS preparation and lack o f access to ORS (Twebaze 2001). 44 Acute RespiratoryInfection 3.19 Acute respiratory infection (ARI).is another significant cause o f childhood mortality and morbidity. According to one qualitative study in Mpigi districts, 90% o f caretakers recognized ARI symptoms and took action, and 56% took home-based action (Nabiwemba L 1999). However, another study in the Kampala suburb shows that caretakers' knowledge of the causes o f ARI was poor and only 19% o f them were able to recognize at least one danger sign o f severe ART. For this reason, most caretakers delayed seeking appropriate care for their children with ARI (Nakazzi 1997). Reproductivehealth 3.20 Knowledge o f at least Figure3.3 Mothers' knowledgeof danger signs during one contraceptive pregnancyand delivery method i s nearly universal in Uganda Fits (UBOS and ORC Macro 2001). Disturbed vision c However, as discussed Swollen feet in Chapter 4, the use ZbP Headache o f modem L contraceptive methods 0 I-" n Fatigue i s low. One Vaginal bleeding encouraging development Abdominal pain i s the 1 I rising age o f first-time 0 10 20 30 40 50 60 70 80 mothers. According to Percent of mothers DHS data, 43% of females age 15 - 19 Source: UNICEF 2001 were mothers or pregnant with their first child in 1995 compared to 31% in2000. 3.21 The low level of good knowledge and practices related to maternal health is a great concern against the backdrop o f highmaternal mortality inthe country. Knowledge o f the danger signs o f pregnancy and delivery i s important to care-seelung behavior inmaternal health. However, according to a UNICEF survey in 2001, most women do not recognize the danger signs during pregnancy and delivery. Other than abdominal pains, all other common signs o f pregnancy and delivery complications are not well-known at all. Only four in ten mothers recognize vaginal bleeding as a danger sign. Even fewer recognize fatigue, headache, swollen feet and blurred vision as signs that an expectingmother needs to seek professional help (Figure 3.3). Malaria 3.22 The knowledge of malaria is widespread with 97% o f mothers having heard about malaria and 74% able to identify mosquitoes as the agent that spreads the disease. Regional differentials inmalaria knowledge are significant. While 93% o f mothers in the Central region are able to link mosquitoes and malaria, only 62% in the Northem region do so (UNICEF 2001). 45 3.23 Knowledge o f the use of bed nets for malaria prevention remains low. In the UNICEF 2003 survey, only 38% of caregivers were able to link the use o f bet nets to malaria prevention and 20% o f children under two slept under a bed net in the night before the survey. Among the few caregivers using bed nets, 35% were not aware of the need to treat bed nets with insecticides and 55% did not do so. There i s thus a significant gap between household knowledge and practices in malaria prevention in Uganda (Figure 3.4). The utilization o f bednets i s discussed inChapter 4. Figure 3.4 The gap betweenknowledgeand practiceinmalaria preventionamong caregiversinUganda have heard about malaria (') 97 know that mosquitoes spread malaria (') &E o know that bed nets help prevent i3 malaria L-8 use bed nets for their young children (") know that bed nets need to be treated with insecticides have treated bed nets with insectidles the last 6 months 0 20 40 60 80 100 Note: (*) These data are fromthe 2001 UNICEFsurvey. Other data are derived fiom the UNICEF 2003 survey. (**) Inthe 2003 UNICEF survey, this is actually the coverage ofbed nets among young children. Here we make the assumption that the coverage of bed nets among young children i s also the percentage o f caregivers who use bed nets for their young chldren. Source: UNICEF2001 and UNICEF2003 3.24 While most caretakers recognize malaria as a major childhood health problem, they are not aware of common malaria symptoms as well as key danger signs. Inone study among mothers whose children became sick with malaria and were confirmed with positive blood slides, 67% did not suspect that their children mighthave malaria (Lubanga 1997). Other studies show that caretakers often link convulsions, which are a symptom of severe malaria in children, with other illnesses (Twebaze 2001). Use of services for malaria treatment is discussedinChapter 4. HIVIAIDS 3.25 This area has been extensively studied in Uganda as part o f the response to the HIV/AIDS epidemic. For detailed and comprehensive discussions, readers are referred elsewhere (DHS 2000, Annual HIV/AIDS Surveillance Reports prepared by the MOH, and reports o f the repeated KABP surveys in 1995, 1998,2001,2002 and 2003). Here we summarize the findings of such reports. In general, various population-based studies in Uganda inthe last decade point to: (i) universalawarenessofHIV/AIDS(around99%); almost 46 (ii) levelsofknowledgeaboutmethodstopreventHIV/AIDS,with80%-90% ofthe high population knowing at least two important ways to avoid HIV/AIDS; (iii)remarkable increase inmean age at first sex, from 14to 16 years; (iv) significantreductioninprevalence of sex withnonregular partners, from 53% to 14%; and (v) increasing condomutilization, especially inurbanareas. 3.26 Positive changes inbehavior and practices have been the crucial factor behind Uganda's success story in HIV/AIDS control (USAID 2002). Compared to other countries in the region, the country i s doing well inthis regard (Figure 3.5 and Figure 3.6). Figure 3.5 Percentageof young people (15-24) Figure 3.6 Percentageof young people (15-24) knowingthat condoms can preventHIVlAIDS usingcondomsinlast high-risk sex (1996- (1996-2002) in Ugandaand selected countriesin 2002) inUganda and selected countriesinthe the region region I 70 I I 80 60 70 60 50 pW 50 -0 40 OMale OMale W 40 IOFemale BFemale a L !30 30 20 20 10 10 0 n Uganda Tanzania Kenya Ethiopia Uganda Tanzania Kenya Ethiopia Source: UNAIDS 2004 3.27 There are wealth-based differentials in knowledge about HIV, especially among women. While 95% and 92% o f women in the richest quintile respectively knew about sexual transmission o f HIV/AIDS and mother-to-child-transmission (MTCT), 76% and 79% in the poorest quintile did so (Gwatkin et al. 2004). 3-28 Poorer men tend to have fewer nonregular sexual partners. The percentage o f men inthe richest quintile having nonregular sexual relationships was two times higher than that o f men in the poorest quintile (Table 3.3). On the other hand, 91% o f men in the richest quintile reported condom use in their last casual sex while only 58% o f those in the poorest quintile did so (Gwatkin et al. 2004). 3.29 Notwithstanding tremendous behavior changes over the last decade, there i s still scope for further improvement, especially in sexual practices in men. In 2000, among married adults (age 15-49 years), while only 2% o f women reported having a nonregular sexual partner in the last 12 months, 12% o f men did so. Among unmarried adults, only 2% o f women reported having two or more partners inthe last 12 months while 11%o f men did so. For both married and unmarried males, the practice o f having multiple sexual partners tends to be more common among the young, richer, better educated and residents o f urban areas and the Central region (Uganda Bureau o f Statistics and ORC Macro 2001). 47 Table 3.3 Levels of individual knowledge and practicesrelatedto IV/AIDS in2000 Indicator Description Wealth Quintiles Population Lowest Second Middle Fourth Highest average Knowledge about sexual transmission of HIV Women % o fwomen age 15- 75.8 81.8 86.1 90.0 95.0 86.2 49 who know at least one way to avoid sexual transmission o f HIV/AIDS M e n % o fmen age 15-49 90.4 92.2 93.9 95.5 98.1 94.1 who know at least one way to avoid sexual transmission o f HIV/AIDS Nonregular sexpartnership Women % ofwomenage 15- 0.6 0.4 0.7 0.5 0.6 0.6 49 having sex with nonregular partners in the last 12 months M e n % o fmenage 15-49 7.0 6.4 7.0 9.5 14.1 8.9 having sex with nonregular partners in the last 12 months Source: Gwatkinet al. 2004 Sanitary practices Hand-washing 3.30 Safe water i s a critical determinant o f health status. The estimated infant mortality rate for households without access to safe water i s estimated to be twice that o f households with access to safe water. The availability o f safe water sources and sanitation facilities are discussed inChapter 4. 3.31 Although as many as 76% o f mothers reported washing their hands before food preparation in the 2000 DHS survey, only 4.5% o f households had hand washing materials and facilities, ranging from 2% in the poorest quintile to 9% in the richest (Gwatkin et al. 2004). This large discrepancy possibly indicates overreporting o f hand washing practice in the 2000 DHS. This i s confirmed by the UNICEF 2003 survey in eight districts, which shows a very low level o f hand-washing. On average, the percentage o f caregivers who reportedthat they always washed their hands with soap was only (i)30% after using the toilet, (ii) after attending to a child who has defecated 16% and (iii)9% before food preparation or feeding a child (Table 3.5). 48 Table 3.4 Handwashing practiceby wealth quintilesin2000 Indicator Description Quintiles Population Lowest Second Middle Fourth Highest average Hand- % o f mothers 92.9 69.0 79.2 87.5 89.8 76.0 washing washing their hands before food before preparing preparation food Hand- % o fhouseholds 2.2 2.4 3.9 5.3 8.8 4.5 washing with hand-washing facilities in materials and households facilities Source: Gwatkin et al. 2004 Sanitation 3.32 At the household level, around 85% of the women reported disposing o f their children's excreta properly (DHS 2000). The UNICEF 2003 survey confirms a similar level o f this practice (Table 3.5) Sanitary disposal o f excreta i s significantly less common in the northem region (52%) than in other regions (75% or higher). This practice was more common inthe higher socioeconomic groups, with mothers inthe highest wealth quintile being 1.5 times more likely to dispose their children's excreta in a sanitary fashion than those inthe poorest quintile (DHS 2000). Table 3.5 Hygieneand sanitationpracticesin eight districts,2003 East Central West North All Toroto Bugiri Masaka Mubende Kasese Kibaale Apac Yumbe Percent of caregivers always washing their hands with soap a. After using 31 33 32 41 20 24 29 26 30 toilet b. After 17 17 12 17 16 19 17 15 16 attendinga child who has defecated c. Before 8 8 10 6 10 11 12 9 9 preparingfood d. Before 9 9 10 8 13 10 8 8 9 feedingachild e. Before 15 13 14 11 17 14 15 15 eating Percentof 84 80 92 99 91 87 80 89 88 households disposing children's stool adequately Source: UNICEF2003 49 Tobacco and alcohol consumption Tobacco consumption 3.33 The direct and indirect health consequences o f smoking are well known. Globally, there are only two large growing causes o f death: HIV/AIDS and tobacco. In Uganda, the tobacco epidemic i s not as large, and therefore not at much in the spotlight as other causes o f death such as HIV/AIDS and other communicable diseases. Three percent of women (15-49 years) and 25% o f men (15-54 years) were classified as active smokers in 2000/01. However, 40% o f the men in the northem region smoke compared to only 13% inthe eastern region. Men smoke an average of four cigarettes per day (Uganda Bureau of Statistics and ORC Macro 2001). 3.34 InUganda, the socioeconomic differentials in smoking are similar to a pattem found in many other countries: smoking i s higher among the poor and less educated. Male unskilled manual laborers are two times more likely to smoke than those in technical, managerial, and clerical professions (Uganda Bureau o f Statistics and ORC Macro 2001). Similarly, men in poorest quintile are 2.5 times more likely to smoke than those inthe richest quintile. For women, thisratio is as highas 15 (Table 3.6). Alcohol consumption 3.35 Around a quarter o f the women and almost half o f the men consumed alcohol at least once in the previous 30 days. Following the same pattems in tobacco consumption, alcohol use i s highest inthe Northern region (54% among men and 35% among women). A substantial proportionof adults inthe Northern region also reported getting intoxicated in the previous 30 days (30% of men and 9% of women) (Uganda Bureau of Statistics and ORC Macro 2001). M e n in the poorest quintile were two times more likely than those inthe richest to be intoxicated(Table 3.6). Table 3.6 Tobacco and alcohol use by wealth quintiles in 2000 Indicator Description Quintiles Population Lowest Second Middle Fourth Highest average Tobacco use Women % o f women age 15- 6.0 4.9 3.9 2.0 0.4 3.3 49 who currently smoke or chew tobacco M e n % o fmenage 15-49 36.7 28.3 26.1 21.3 14.4 25.2 who currently smoke or chew tobacco Alcohol use Women % ofwomen age 15- 6.8 7.4 5.6 3.9 5.3 5.8 49 intoxicatedinthe past month M e n % o f menage 15-49 32.4 24.3 23.0 20.4 14.6 22.8 intoxicated inthe past month Source: Gwatkinat al. 2004 50 D.GOVERNMENT RESPONSEAND POLICY IMPLICATIONS 3.36 Given the low level o f good health-related knowledge, behavior, and practices at the individual and household levels, much still needs to be done to promote health at the grass-root levels. Uganda's success in improving knowledge, behavior and practices with regardto HIV/AIDS canprovide valuable lessons for other areas o fhealth, nutrition, and populationcontrol as well. 3.37 The Annual Health Sector Performance Report 2002/03 recognizes that `knowledge and attitudes o f the population are key determinants o f health-seeking behavior in communities.' At the national level, the government has a strategy for health promotion and community mobilization as spelled out in the National Health Policy as well as the HSSP, but a comprehensive operational plan has yet to be formulated. Family care practices and community mobilization had not been prominent on the agenda o f the biannual sector review until the start o f the sector-wide approach at the 9" JRM in November 2003 (see box). Although there are four HSSP indicators to measure performance in this area, no monitoring has actually taken place. Moreover, actions targeting household and communities are not backed up by adequate resources by the central government. For example, in the government's proposed budget plan for the 2004-2005 fiscal year, most o f the unfunded priorities fall into the categories o f household and community interventions. 3.38 At the district level there has been some evidence o f a shift o f priorities by local governments from provision o f public health goods-including health promotion though community mobilization-toward the provision o f private health goods (Hutchinson, Atkin and Sesengooba 2003). The central government needs to get local government participation inhealth promotion efforts though policy dialogue and advocacy. Technical support from the center to the districts inhealth promotion and communication i s critical: especially in helping the districts in planning, executing, monitoring, and evaluating a multisectoral operational plan. Some key indicators in household knowledge, behavior, and practices could be used inthe evaluation o f district performance. 3.39 At the community level, various community initiatives and structures for health promotion and community mobilization (e.g. malaria, IMCI, HIV/AIDS, river blindness, nutrition, etc.) currently exist. These fragmented channels could be consolidated and strengthened. Some community structure hitherto not much involved in health (e.g. community development workers) could be engaged. The concept o f village health teams could be further explored with a view to implementing and scaling up. Experience from the Uganda AIDS Control Project shows that technical support (either from the districts or NGOs) to communities are essential to enable them choose the right interventions and implement them in an appropriate manner. 3.40 Some projects have provided good lessons on how health promotion and community mobilization can be conducted inan integrated approach. A good example o f this was the Behavior Change Communication (BCC) component under the second Delivery o f Improved Services for Health Project (DISH II), which was implemented in 12 districts of Uganda. The overall goal o f the component was to promote positive household knowledge, behavior, and practices as well as to direct the public to matemal and child health services. With multichannel campaigns conducted inphases, the BCC component assisted the districts inpromoting family planning, good infant feeding practices, malaria 51 control, immunization, safe motherhood, prevention of mother-to-child transmission o f HIV, STD prevention and management, and voluntary counseling and testing. The component followed a standard process: each campaign included a mixture o f centrally produced print and electronic media combined with educational community-organized activities such as village meetings, video shows, soccer matches, bicycle rallies, song contests, child health fairs, kitchen garden contests and community based services. 52 Box: Communitymobilizationinthe Government'shealthagenda Community mobilization in the National Health Policy and HSSP The objectives o f the community empowerment policy within the National Health Policy are to ensure that communities take responsibility for their own health and well being, and to participate actively in the management o f their local health services. T o support this, the government plans to do the followings: a. develop guidelines for effective community participation; b. promote the establishment o f health committees with an appropriate gender balance at each level o f the local government system; and c. promote and support community-based health services. This is further elaborated inthe HSSP under "Decentralized Health Care System" where gender- sensitive community participation and empowerment are emphasized. According to the HSSP, a Village Health Committee (VHC) or similar structure will be established in every village to be responsible for a series o f tasks, including selection and oversight o f the activities o f community health workers. As part o f the district structure for health care delivery, the VHC i s to fbnction at the same level as the village council, a political institution. VHCs are to be set up in a phased manner over the plan period. Community mobilization in the agenda of Health Sector Review In the annual sector health performance reports up to the Mid-Term Review o f the HSSP, discussions on community mobilization were very limited. Since the Mid-Term Review (April 2003), there has been a more concerted effort to address community mobilization. Community ownership and participation in health service delivery was one o f the key themes for discussion at the Mid-Term Review. The MTR reiterates that community participation and empowerment i s an important strategy for enabling communities to take responsibility for their own health and well-being to participate actively in the management o f their local health services. A strategy document and plan for operationalizing VHCs was prepared and disseminated to major stakeholders. Implementation o f the strategy i s beingpiloted in some districts. The first National Health Assembly passed a resolution on community mobilization and called for using existing community management structures. The 9* JRM had a working group on `Strategies for strengthening community mobilization and empowerment.' For the first time, a logical framework for community mobilization was formulated. This log frame identified issues, actions, time frame, and units responsible for community mobilization, with a budget o f shs 5bn Community mobilization was one o f the undertakings listed for October 2004. 53 4. UTILIZATION OF PREVENTIVEAND CURATIVE HEALTHSERVICES A. INTRODUCTION 4.1 This chapter examines the utilization o f both curative and preventive health services in Uganda. Such utilization i s one o f the direct determinants o f health, nutrition, and population (3")outcomes. Other direct determinants, such as the performance o f the health system e.g. availability, affordability and quality o f health services, are discussed inChapter 5. The interactions between different determinants o fHNP outcomes are often complex, and the categorization o f determinants along the line o f supply o f and demand for health care i s not always clear-cut. 4.2 This chapter discusses the utilization o f health-related services in three categories: (i) preventive clinical services (such as immunization, antenatal care, delivery care, contraceptive services, voluntary testing and counseling for HIVIAIDS, etc.); (ii) population-based preventive activities (such as health education campaigns, promotion and distribution o f insecticide bed nets and condoms, supply o f safe drinking water and sanitation services, etc.); and (iii) curative clinical services. 4.3 Most the data presented in this chapter come from (i) Demographic and Health three Surveys conducted in 1988, 1995, and 1999 as well as (ii)two Uganda National Household Surveys conducted in 1999 and 2002. B.THEOVERALLPICTURE 4.4 Existing evidence points to a mixed picture o f health-related service use in Uganda. While there have been significant increases inthe utilization o f many clinical preventive services, most notably childhood immunization, antenatal care and VTC services; serious gaps still exist, especially in attended deliveries and family planning. Furthermore, generally speaking, the utilization o f population-based preventive services i s even lower than that o f clinical preventive services. Although the use o f curative services significantly increased after the abolition o f user fees, it i s still low, especially among rural dwellers and lower socioeconomic groups. Policy implications are thus two-fold. First much can still be done to improve the utilization o f clinical preventive services (especially those for maternal health beyond ANC) , as well as population-based preventive services. Second, mechanisms additional to the abolition o f user fees are neededto enable the poor and vulnerable to increase their use o f health services. 54 c.UTILIZATION OF CLINICAL PREVENTIVE SERVICES Childhoodimmunization Levels and trends in immunization coverage 4.5 The coverage o f routine immunization was very l o w in the early 1980s. In response, Uganda launched the Expanded Program o f Immunization (EPI) in October 1983, and polio vaccination at birth started in 1988. As a result, there was some improvement in immunization coverage inUganda duringthe 1985-1990 period. The aggregate complete immunization coverage rates inUganda increased gradually from 31% in 1988 to 48% in 1995 (MOH 2003). However, this was followed by a period o f stagnation and by the mid- 199Os, routine immunization coverage started to decline. In 1999, the national DPT3 coverage, a key HSSP indicator to monitor the utilization o f immunization services in Uganda, was only 41%. This positionedUganda as a mid-range performer among African countries (Figure 4.1). 4.6 Inthe context ofHSSPimplementationstarting in2000, the immunizationprogramwas revitalized with increased commitment, inputs, and social mobilization. Consequently, there was a dramatic reversal o f the declining trend in immunization coverage. Within three years, DPT3 coverage more than doubled, reaching 84% in2003 according to MOH data (MOH 2003). Levels and trends in immunization coverage inequality 4.7 At the household level, the two most important socioeconomic characteristics affecting childhood immunization coverage are wealth and mother's education. Such socioeconomic differentials are largest for DPT3. While mothers' education-based inequality in childhood immunization was halved between 1995 and 2000, wealth-based inequality increased by 40% during this period. Inequality was also observed by residence (urban vs. rural, regional). 55 Figure 4.1 Trends in DPT3 coverage in Uganda and select neighboringcountries CentralAfrican Republic -Chad -Congo (Brazzaville) +Congo, DR (Kinshi Wealth-based inequality 4.8 In 2000, DPT3 coverage was 35% in the poorest and 55% in the richest quintile. Compared to other low and middle-income countries, Uganda was thus a mid-range performer in terms wealth-based inequality in DPT3 immunization (Figure 4.2). Table 4.1 shows the equity trends inimmunization between 1995 and 2000. While wealth-based inequality in measles coverage improved, the situation either did not improve or worsened for other categories during this period. Moreover, ineach survey year, although the ricWpoor ratio was relatively small for DPTl or the "ever-vaccinated" category, it was much larger for the DPT3 or the "completely vaccinated" category. This means poor children were initially reached by the EPI but subsequently dropped out from the programbefore becoming fully immunized. 4.9 After adjustments for other individual and household characteristics, children in the richest quintile were twice as likely to receive DPT3 than those in the poorest quintile in 1995. In 2000, this likelihood was more than three times more, representing a 40% increase (See Appendix). 56 'igure 4.2 DPT3 coverage by wealth quintiles inUganda and selected countries --Populationaverage loo r-Poorestquintile a * Richestquintile -__._ __1___- 4 -___I_ 90 80 h70 c E $60 VI 0 52 E 50 ; 40 a" fi 30 20 10 'ource: Gwatkin et al. 2002 Table 4.1 Inequality in Ugandan childhoodimmunization coverage 1995 and 2000 (richest vs.poorest quintile; mothers with education vs. mothers without education) SecondaryINo Richestlpoorest ratio Richestlpoorest gap SecondaryINo Education gap (inpercentagepoint) Education ratio (inpercentage point) Measles 1995 1.5 25.8 1.7 34.7 2000 1.4 16.9 1.3 15.4 DPTl 1995 1.2 14.6 1.4 25.6 2000 1.2 16.0 1.2 16.5 DPT3 1995 1.5 24.8 1.8 35.8 2000 1.8 25.2 1.6 23.7 Ever vaccinated 1995 1.1 11.2 1.2 18.9 2000 1.1 5.3 1.1 6.9 Completely vaccinated 1995 1.6 23.0 1.8 29.8 2000 1.9 21.0 1.8 22.7 Source: Uganda Bureau of Statistics and ORC Macro 1996 and 2001, authors' calculations 57 Mother's education based inequalities 4.10 In2000, DPT3 coverage was 61% among children ofmotherswith secondary education but only 37% among those whose mothers had no education. Unlike wealth-based inequality, mother's education-based inequality seem to have improved in most immunization categories between 1995 and 2000 (Table 4.2). After adjusting for other household and individual factors including wealth, children o f mothers with secondary education were twice as likely to receive DPT3 than those o f mothers with no education in2000. Thisratio was four in 1995; thusmother's education-based differentials inDPT3 coverage was halvedbetween 1995 and 2000 (See Appendix). 4.11 Wealth and mother's education-based differentials in childhood immunization reflect both the level o f household's demand for such services and how well the health system i s reaching out to disadvantaged groups in the population. H o w Uganda currently fares in terms o f equity in immunization coverage i s unknown; the next DHS would shed light in this regard. Residence-based inequality 4.12 In all three DHSs, immunization rates were consistently higher in urban compared to rural areas, although the differentials seemed to decrease over time. However, after adjusting for other individual and household characteristics, urbadrural differentials become statistically insignificant (See Appendix). 4.13 Across regions in 2000, the Western region generally had the best immunization coverage, followed by the Eastern, Northern and Central regions in that order. The decrease in coverage between 1995 and 2000 was uneven among the four regions. For example, while other regions suffered a decline mostly in DPT3 coverage, it was the Central region where immunization declined across the board, especially with regard to DPT3 (by 85%), measles (by 30%) and DPTl (by 27%). After adjusting for other individual and household characteristics, children living in the West were three times more likely to receive DPT3 than those in the Central region (See Appendix for disaggregated levels by different indicators). Dropout ratesfor multidose vaccines 4.14 One o f the indicators for the quality o f an immunization program i s the dropout rates for multidose vaccines (e.g. for polio and DPT). In Uganda, the dropout rates for DPT and polio were respectively 40% and 36% in 2000 compared to 25% and 28% respectively in 1995. As a result, though the number o f children reached by the immunization program was roughly equal between 1995 and 2000 (13% and 14%), the percent o f completely immunized childrenhad declined significantly (See Appendix). Antenatal care (ANC) 4.15 Evidence in recent years has pointed to the benefits o f focused antenatal care. Key interventions include the tetanus toxoid vaccination, provision o f prophylactic antimalarial drugs, and screening for syphilis, other STDs, and hypertension. 58 Levels and trend in ANC coverage 4.16 The percentage o f pregnant women using ANC services provided by a trained health professional at least once increased fi-om 11% in 1988 to 89% in 1995, then to near universal coverage (93%) in2000. 4.17 However, the coverage of tetanus immunization among pregnant women did not go hand in hand with the utilization o f ANC services. It increased fi-om 56% in 1988 to 80% in 1995, then dropped to 70% in 2000. Thirty-four percent o f women reported taking antimalarialdrugs duringtheir last pregnancy inthe DHS 2000. Levels and trend in ANC coverage inequality 4.18 Compared to other middle and low-income countries, Uganda i s among the top performers with highcoverage and low inequality (Figure 4.3). Despite this achievement, differentials in ANC coverage still exist in Uganda. As in the case o f immunization, the two most important household socioeconomic characteristics affecting ANC utilization are wealth and mother's education. The use o f ANC services was still consistently higher in urban areas, among richer households, and among educated mothers in all three DHSs. Wealth-based inequality 4.19 Pregnant women in the richer quintiles were consistently more likely to receive professional ANC care as well as to be immunized against tetanus in all three surveys. Wealth-based inequality in ANC coverage was significantly reduced between 1988 and 1995, then seemed to increase in 2000 (Table 4.2). In 1988, after adjustments are made for other individual and householdcharacteristics, pregnant women inthe richest quintile were four times more likely to utilize ANC than those in the poorest quintile. This likelihood was reduced to two in 1995, then increasedto three in2000 (See Appendix). 4.20 Wealth-based inequality also exists in the use o f anti-malarial among pregnant women (27% inthe poorest quintile compared to 40% inthe richest). 59 Figure 4.3 Utilization of ANC provided by a trained health professional, ranked by the level of use in pregnant women of the poorest quintile inUganda and selected low and middle income countries -Esf100.0 90.0 n 80.0 P.E 70.0 =3 e I- 60.0 2B 50.0 .-Y ' 40.0 v) 30.0 9 -$ 20.0 m 2g! 10.0 Country Source: Gwatkin et al. 2002 Mother's education-based inequality 4.21 Better-educated mothers Table 4.2 Inequality in ANC coverage (richest vs. poorest are more likely to quintile) 1995 and 2000 receive ANC care Year Richest/poorest ratio Richest/poorest gap during pregnancy. At least oneANC visit to a ANCprovider Similar to the rich-poor 1988 5.8 differentials, the 1995 1.2 12.4 education-based equity 2000 1.3 14.3 gap was also Four or moreANC visits significantly reduced 1995 1.3 6.5 between 1988 and 1995. 2000 1.3 8.9 Pregnant women with ANC tetanus secondary education 1995 1.1 8.7 were 4.3 times more 2000 1.2 13.3 likely to receive Source: Uganda Bureau o f Statistics and ORC Macro 1996, professional ANC care Uganda Bureau o f Statistics and ORC Macro 2001, authors' and 1.5 times more calculations likely to be immunized against tetanus compared to those without education. Both ratios were reduced to 1.2 in 1995 and then stayed more or less at the same level in 2000. The 60 differentials in the coverage persist after controlling for other individual and household characteristics. Moreover, secondary education had a stronger association with receiving professional ANC care than primary education. (See Appendix). Residence-based inequality 4.22 After adjusting for other individual and household characteristics, living in the West was a factor that reduced the likelihood for professional A N C care in all three surveys compared to people living inthe Central region. Living inthe East was a disadvantage in 1988 but it became an advantage in 2000. Urban women were more likely to receive professional A N C care in 1988 than rural women, but this differential was not observed inthe following surveys. (See Appendix). Other individual and household characteristics associatedwith ANC attendance 4.23 Other factors affecting ANC coverage include religion, parity, and birth order. Religion- based differentials existed in 1998 and 1995, with Protestant and Muslimwomen being more likely to get professional ANC care, but not in 2000. Parity was a factor inthe 1988 and 2000 surveys with mothers o f high parity being less likely to receive professional ANC care. Higher birth order in 2000 also had lower likelihood of professional ANC. (See Appendix). Professionaldelivery care 4.24 As noted elsewhere in this report, maternal mortality in Uganda remains high. As in other developing countries, the main causes o f maternal death result from complications arising during and after delivery. Antenatal care has been shown to have limited value in identifying ahead o f time women who will experience life-threatening complications. It i s therefore essential, that all women have access to emergency obstetric care. This requires that women are delivered by a skilled attendant who has the requisite skills, drugs, and equipment necessary to deliver the baby safely. In the event of a complication, the skilled attendant needs to be able to stabilize the woman's condition before making a referral to a higher-level care facility. Levels and trends in attended delivery 4.25 One of the key characteristics o f maternal health in Uganda i s the discrepancy between the relatively high utilization o f ANC services and the relatively low levels o f births attended by health professionals or at facilities. In2000, only 36.9% o f deliveries were attended by health professionals, a marginal increase from 35.3% in 1995 according to the DHS. The percentage o f deliveries in a facility setting was even lower (19% in FY 2001/02 and 20.3% in FY 2002/03 according to government statistics) (MOH 2002, MOH2003). 4.26 Compared to other low-income countries, Uganda's performance in attended deliveries is in the mid-range for both equity and coverage (Figure 4.4) but disparities are considerable. The country average for home delivery (62%) as reported in the UDHS was highcompared to SSA overall (54%). 4.27 The 2002/03 AHSPR reports that the reasons for low use o f health facilities for delivery are: perceived poor quality o f services (lack of equipment, supplies, water, light, and 61 privacy), poor physical access to health facilities, services not being affordable, the unwelcoming attitude o f health workers,, preference for traditional birth attendants, and cultural barriers (MOH 2003b). Another reason for the low levels o fbirths at health units i s the low expectation that complications will arise during pregnancy, especially after having attended ANC and been told that the pregnancy i s proceeding normally (Onama 2001, Amoot-Kaguna and Nuwaha 2000). Figure 4.4 Attended delivery by a medically trained person invarious countries, ranked by coverage inthe poorest quintile 100 I I 90 80 70 20 10 Country Source: Gwatkinet al. 2002 Levels and trends in attended delivery inequality 4.28 Ingeneral, the percentage of attended deliveries was consistently higher inurban areas, among richer households, and by educated mothers as well as younger mothers in all three DHSs. There was also a pattern regarding religion with Muslimwomen having a higher rate o f attended deliveries, followed by Protestants and Catholics. Wealth-based inequality 4.29 Women in the richer groups were more likely to deliver with professional attendance in all three surveys. In 2000, 74% o f deliveries in the richest quintile were attended by trained professionals compared to only 21% in the poorest. The rich also had the tendency to use both public and private facilities for deliveri'es more (44% and 32% respectively) compared to those in the poorest quintile (13% and 7%). As a result, the poorest quintile was 3.4 times more likely to deliver at home than the richest. The wealth- based disparity was significantly reduced over time. In 1988, women in the richest quintile were 14 times more likely to have attended deliveries compared to those in the poorest quintile. This ratio was 3.4 in 1995 and 3.5 in 2000. After adjusting for other individual and household factors, wealth-based differentials remained prominent, especially inthe 1995 and 2000 surveys. (See Appendix). 62 Mothers' education-based inequality 4.30 In all three surveys, educated mothers had higher coverage of attended deliveries. In 2000, at 75%, attended deliveries among mothers with secondary education was 3.6 times higher than in those with no education (21%). However, this gap had already been significantly narrowed when compared to 1988 when the ratio was as highas 9.5 between the two groups. After adjusting for other factors, the relationship between secondary education and attended delivery remained strong all three surveys. Primary education as a contributor was observed in 1995 and 2000 butnot in 1988. (See Appendix). Residence-based inequality 4.3 1 Inall three DHS surveys, the percentage ofattendeddeliveries was consistently higher in urban than in rural areas. In 1988, the likelihood o f attended deliveries was 6.7 times higher among urban mothers than their rural counterparts. This urban-rural gap was narrowed to 2.6 in 1995 and 2.5 in2000. In2000, urban pregnant women were 2.5 times more likely to be delivered by a nurse, 2.5 times more likely to delivery in a public facility, and 2.4 times more likely to deliver in a private facility. On the other hand, rural pregnant women were 3.5 time more likely to deliver at home. The urbadrural differentials remain statistically significant after adjustments for other factors. (See Appendix). 4.32 With regard to regional characteristics, there was a significant decline in attended deliveries between 1995 and 2000 inthe West while modest improvements were made in the remaining areas. In this period, the rate dropped from 23% to 17% in the West, representing as much as a 24% decline. After adjusting for other individual and household characteristics, living inthe West was an factor that reducedthe likelihood for attended deliveries in all three surveys compared to people living in the Central region. This is also consistent with the fact that living in the West reduced the likelihood o f getting professional ANC care as discussed earlier. Living inthe North was an advantage in 1988 but the situation reversed in 1995 and 2000. Mothers inthe Eastern region had a lower likelihood o f attended deliveries compared to those inthe Central but this was only observed in 1995. (See Appendix). Other individual and householdfactors aflecting attended delivery 4.33 Other factors such as mothers' age, religion, partner's occupation, and birth order were found to be associated to attended delivery. 4.34 The coverage o f attended deliveries among younger mothers was also consistently higher inall three surveys. In2000, attended deliveries among mothers 20 years old and below, at 46%, was 2.2 time higher than those among mother in the 40-49 age group. In fact, between 1995 and 2000, the 40-49 age group was the one that suffered a significant reduction in attended deliveries (from 28% to 20% , or a 27% reduction), while there were some modest improvement inall other age groups. (See Appendix). 4.35 The coverage o f attended delivery was higher among mothers o f lower parity. In 2000, the rate among those with 0-2 parity was 47% or 1.5 times higher than those with a parity o f five children or more (31%). This ratio was even higher in 1988 when the 0-2 parity group was 1.8 times more likely to have attended deliveries. (See Appendix). 63 4.36 Differentials based on religious affiliation also existed in both 1995 and 2000 with Muslimwomen more likely to be attended duringdeliveries than Catholic women. There were also some associations between attended delivery and the number o f children younger than five in the family, parity, and birth order, but these associations did not appear inall surveys. (See Appendix). Familyplanning Levels and trend of family planning service utilization 4.37 As discussedinChapter 2, Uganda's TFR at 6.9 is highcompared to other SSA countries. (Figure 4.5). Although awareness o f family planning is nearly universal (93%), use o f contraception remains low in Uganda. According to the 2000 DHS , the contraceptive prevalence rate (CPR) was estimated at 23% for all methods and 18% per cent for modem methods. About 33% o f all births are not planned, and 8% are unwanted. Unmet contraceptive need i s estimated to be 35%. Twenty one percent o f women have unmet needs in birth spacing and 14% in limiting family size (Population Reference Bureau 2003). According to a study inKumi district, 25% o f women reported o f their husbands' disapproval o f contraceptive use (Mbonye 2003). Levels and trend in inequality in utilization of family planning services 4.38 When compared to other low income countries, Uganda's performance in family planning (FP) can be categorized as less than average both in coverage as well as equity (Figure 4.6). Use o f FP was consistently higher in urban areas, among women in the richer quintiles or o f higher levels of education in all three surveys. The use o f modern FP also followed the patterns o f FP use, although at lower corresponding rates. Here we focus the discussion on equity issues inuse o f modem FP. Wealth-based inequality 4.39 Women inthe richer groups were more likely to use modem FP than poorer groups in all three surveys. In2000, 25% o f women inthe richest quintile used modem FP, compared to only 10% in the poorest. The wealth-based inequality gap was significantly reduced over time. In 1995, women in the richest quintile were seven times more likely to use modem FP compared to those in the poorest quintile. This ratio was reduced to 2.5 in 2000. After adjusting for other factors, the relationshipbetween wealth and contraceptive use remained strong, especially inthe 1995 and 2000 surveys. (See Appendix). Education-based inequality 4.40 Inboth 1995 and 2000 surveys, educated mothers used modem FP methods more. In 2000, at 32%, such use among mothers with secondary education was 3.7 times higher than in those with no education (9%). However, this inequality gap had already been significantly narrowed compared to that in 1988 when the ratio was as high as 8.4 between the two groups. After adjustment for other factors, education remains a strong correlate in all three surveys, especially in the case o f secondary education. (See Appendix). 64 Residence-based inequality 4.41 In 1995 and 2000, the percentage of modem FP use was consistently higher in urban compared to rural areas. For example, at 33%, the use o f modem FP among urban women was 2.5 times higher than in 1995 (13%). This urban-rural gap was even larger in 1995, when urban women were 4.6 times more likely to use modem FP. After adjusting for other factors, the urban-rural differentials remain strong in all three surveys. (See Appendix). Other individual and household characteristics affecting utilization of FP services 4.42 Among women o f reproductive age (15-49), modern contraceptive coverage among the 20-29 and 30-39 age groups was also consistently higher in both surveys. It i s also worth noting that inabsolute terms, increases inthe use o f modem FPbetween 1995 and 2000 were greatest in these two groups (11.9% increase for the 20-29 age group and 9.1% increase for the 30-39 age group). Muslimwomen had the highest rates o f modem FPuse inboth surveys, followed by Protestants and Catholics. In2000, the use o fmodem FP among the three groups was 18.8%, 16.3% and 15.6% respectively. After adjusting for other factors in a multivariate analysis, religion i s not a determinant o f FP use. Parity was a factor with mothers having more than three children being more likely to use contraceptives. In 1995 and 2000, partner's education level had an association with a woman's use o f contraceptives. (See Appendix). 65 Figure 4.5 CPR and TFR insub-SaharanAfrican countries Tanzania 1996+ + Zambia 1996 ..... ~.................~~.............Moz_arslb.i~u.e.~gq.7~.~..............-...... . ~ ~ Ethiopia 2000 Niger 1998 + Chad 1996/97 + 0 1 2 3 4 5 6 7 a TFR 1549 Figure 4.6: Use of moderncontraceptiveby women in richest and poorestquintiles,rankedby averagevalues 80 x Poorest quintile A Richestquintile -Average 70 ................................................................................................................................ w ............. ............... ..................... f..- I.!].]. ~ ~............................ ............ .......... .......... ~ ~ ~ 50 t t 40 30 20 10 4 C Utilization of voluntary counseling and testing (VCT) services 4.43 In2000, 8% of women and 12% ofmenreportedthat they hadever beentested for HIV. Women aged 20-29 and men aged 25-39 were the most likely to get tested. Testingwas 66 more common among those living in urban areas, in the central region, and inthose with secondary education (Uganda Bureau o f Statistics and ORC Macro 2001).Wealth-based differentials in VCT are quite large. Men and women in the richest quintile were 3.6 and 9.3 times more likely to get tested (Gwatkin et al. 2004). Of those who have been tested, 59% o f women and 65% o f men had the test done in a public facility and the rest in a private facility (MOH). 4.44 Although complete Figure4.7 Numberof first-timetesters aged 15-28 at the national data on testing AIDS InformationCenter for the last two years are not available, there 18000 are indications that use 16000 o f VCT i s on the 14000 Dmle increase. For example, 12000 Bfemle the number o f first time 10000 testers among the 15-28 8000 age group at the A I D S 6000 Information Center (a 4000 key provider o f VCT 2000 services in Uganda 0 through a network o f 2000 2001 2002 branches inthe country) more than doubled Source:MoH2003 between 2000 and 2002 (Figure 4.7) D. POPULATION-BASEDPUBLICHEALTH ACTMTIES 4.45 This section focuses on (i)access to and utilization o fwater and sanitation and (ii) o f use insecticide-treated nets (ITN) inUganda. 4.46 Water and sanitation are known to be effective interventions to improve people's health, especially for their roles in reducing under-five mortality due to diarrhea. For example, improvement in the quantity and quality o f water available for household use can reduce deaths from diarrhea from by 16% to 25%, with quantity being the more important factor (Commission on Macroeconomics and Health 2001). Overall, coverage o f safe water and sanitation has increased in Uganda in the last decade. The growth in water supply has been faster than that in sanitation (Figure 4.8). Most available information i s about access, while not much exists for utilization o f water and sanitation. 4.47 The use of ITNs is one of the most effective interventions for malaria control, especially for pregnant women and under-five children. At higher levels o f use, this individual preventive measure has a protective impact on the entire community by reducing feeding opportunities o f the mosquito population. Although malaria i s the leading cause o f morbidity and mortality inUganda, bednets coverage remains unacceptably low. 67 Access and utilizationof water 4.48 While safe drinkingwater i s a major input for health, almost one in four rural households still relied on unsafe sources (such as from river, spring, etc.) for their drinking water while another 28% relied on open public wells (Table 4.3). 4.49 Around 91% o f Uganda's population live inrural areas and they are currently served by about 45,000 water point sources and small rural piped systems. Access to safe water was estimated at about 55% for rural populations, while 63% o f population in large towns have access to safe water. There are wide variations o f coverage among districts, from as low as 25% to over 75%. Over 30% of rural systems are nonfunctional. 4.50 Around 63% o f urban dwellers live within 15 minutes o f a water source. For rural dwellers, that proportion i s only 13% (MoLWE 2004). Fifty percent of the rural households have to spend more than half an hour to collect water, with little change since 1995 (Uganda Bureau o f Statistics and ORCMacro 2001). 4.5 1 Water handling and storage i s often substandard, leading to contamination o f water from a safe source before it i s consumed. A 1996 study show that among households with access to a safe water source, only 9% actually consumed water with acceptable quality as the water was contaminated between the source and point of consumption (RUWASA study, cited by the MOH EHD). The Directorate o f Water Development reports that the average per capita water used i s only half the recommended amount (Directorate of Water Development 2004) Table 4.3 Access to safe drinkingwater 1995-2001 Urban Rural Total 1995 2001 1995 2001 1995 2001 ~ ~~ Safe sources 76.3 89.9 44.3 55.3 48.7 60.5 Piped into residence 12.7 12.1 0.1 0.2 1.8 2 Public tap 31.9 51.2 1 1.5 5.2 8.9 Borehole (mainly public) 13.5 13.8 17 26.6 16.6 24.7 Springs 17.8 1.3 25.8 9.4 24.7 8.2 Rainwater 0.4 0.4 0.4 0.4 0.4 0.4 Protected well (mainly public) NIA 11.1 NIA 17.2 NIA 16.3 Unsafe sources 20.2 7.8 53.7 42.4 49.2 37.2 Open public wells 17.9 6.9 23.2 28.3 22.5 25.1 Riverlstream 1.3 0.3 15.3 8.8 13.4 7.5 Pondlake 1 0.6 15.2 5.3 13.3 4.6 Time to water source (minutes) 4 5 minutes 49.6 62.7 16.3 15.4 20.8 22.5 Mediantime to source (inminutes) 15 9.2 30.6 29.9 30.5 29.6 Source: Uganda Bureau o f Statistics and ORC Macro 2001 68 Access and utilization of sanitation 4.52 In the 1960s, Uganda had Figure 4.8 Percentageof householdswith access to high latrine coverage, around water supply and sanitationinrural areas in 95% o f the population. After Uganda two decades o f war and civil strife, the coverage dropped to 60% around 23% 1980s. With theinrecovery o f the early Rural water supply 50% the economy, sanitation 40% @Ruralsanitation gradually improved. Official statistics give a mixed picture. 30% In 2000, around 79% o f households used pit latrines. 20% One in six households in the country has no sanitary ,o% facilities o f any kind. Rural 0% households are seven times 1990 1995 2000 more likely than urban ones not to have sanitary facilities (Uganda Bureau of Statistics Source: MOHEHD 2004 and ORC Macro 2001). However, according to the MOH, access to safe sanitation (measured by the presence o f household latrines o f acceptable quality) i s around 56%. On the other hand, the Directorate o f Water Development estimates that national household latrine coverage i s 48% with wide variation among districts (ranging from 4% to 80%) (Directorate o f Water Development 2004). Access to a sewage system i s still extremely limited. Piped sewerage services are only accessible to about 8% o f the urban centers (PRSC document). Urban Rural Total Flushtoilet (own or shared) 9.3 9.1 0.2 0.5 0.6 1.7 Source: Uganda Bureau o f Statistics and ORC Macro 1996, Uganda Bureau o f Statistics and ORC Macro 2001 69 Utilizationof insecticide-treatedbednets (ITNs) 4.53 In2000, only 13% of households in ITable 4.5 Bed net utilization inUganda 2003 Uganda had mosquito nets. Around Sleeping under bed nets 7% o f under-five children slept under a net the night previous the Yes No survey (Uganda Bureau o f Statistics Overall 10.7 89.3 and ORC Macro 2001). With regard Residence to ITN, the coverage i s even lower, Rural 7.9 92.1 around 5% among children under Urban 27.7 72.3 five (NMCP). Urban children were Wealth quintiles three times more likely to sleep Poorest 4.7 95.3 under a bed net than rural children 2ndpoorest 6.5 93.5 (Uganda Bureau o f Statistics and Middle 8.0 92.0 ORC Macro 2001)). Wealth-based 2nd richest 9.9 90.1 differentials in bed net use are Richest 24.6 75.4 significant. In the richest quintile, Region bed net coverage among under-five Central 14.0 86.0 children and pregnant women were Eastern 9.2 90.8 3.7 and 2.6 times respectively Northern 10.2 89.8 higher than among those in the Western 8.5 91.5 poorest quintile (Gwatkin et al. Gender 2004). The 2002/2003 household (of householdhead) survey does not show significant improvement in bet net utilization Male 10.1 89.9 with only around 11% o f Female 11.3 88.7 households sleeping under bed nets. Source: UNHS 2002-2003 Table 4.5 shows bet net utilizationby wealth quintiles, regions, gender, and residence. 4.54 A 2003 study conducted by UNICEF ineight districts shows that 20% o f children under five slept under a bed net, and 10% under an ITN (UNICEF 2003). Given this trend, Uganda i s likely to miss both the Abuja and HSSP targets according to which respectively 60% and 50% o f under-fives and pregnant women should sleep under I T N s by 2005 (Figure 4.9). The AHSPR 2002-2003 notes that even at the subsidized rate o f Ug.shs 5,000, the cost of I T N s remains out o freach o fthe intendedbeneficiaries. Figure4.9 Utilizationof bed nets and ITNs among children under-five in 70 Uganda 2000 (DHS) 2003 (UNICEFstudy) 2005 (Abuja target) Source: Uganda Bureau of Statistics and ORC Macro 2001 andUNICEF 2003 E.UTILIZATIONOFCURATIVESERVICES Utilization of curative services for childhood illnesses (diarrhea and A N ) Levels and trends of utilization of curative servicesfor diarrhea and A R I 4.55 Compared to other SSA countries, Uganda i s doing relatively well in terms o f care- seeking behavior for ARI and diarrhea (Table 4.6 and Figure 4.10). In 2000, around 61.4% o f children with ARI were brought to a health provider while 48.2% with diarrhea were treated with oral rehydration salts (ORs). However, there was a deterioration in care-seeking behavior for diarrhea between 1995 and 2000. The percentage o f children with diarrhea receivingORS fell from 48.2% in 1995 to 43.2% in2000, a 10% reduction. Especially, between 1995 and 2000, there was a 14% reduction o f children with diarrhea who were brought to a public facility for medical care. On the other hand, the percentage o f children with ARI who received medical care increased from 59.9% to 75.2% (a 26% increase). Most o f ARI care seemed to take place in the private sector, as only 25% o f ARI caseswere brought to a public facility, a 5% increase from 1995. (See Appendix). Levels and trends of inequality in the utilization of curative servicesfor ARI and diarrhea 4.56 In general, care-seeking behavior for children with ARI or diarrhea was better in urban areas, among educated mothers, and inricher households. Male children seem to be more likely to receive care when they have ARI anddiarrhea. 71 Table 4.6 ARI and diarrhea care- Figure4.10 Care-seeking behavior for ARI by seeking behavior (2000) quintileinsome sub-SaharanCountries Percent with Percent with diarrhea that ARI taken to a received ORS Country health provider packet CAR 41.2 24 Chad 21.7 15.6 Nigeria 49.7 34.3 Ethiopia 15.8 13.1 Kenya 57.3 36.9 Malawi 46.1 49.7 Mozambique 38.5 41.9 Tanzania 67.5 54.9 Uganda 61.4 48.2 Zambia 70.7 53.9 Source :Gwatkinet al., 2002 Wealth-based inequality 4.57 Children in the richer groups tend to be more likely to receive care when they are sick with ARI and diarrhea. In2000, 60% o f children with diarrhea inthe richest quintile were treated with ORS, compared to only 39% in the poorest. Similarly, 89% o f children with ARI inthe richest quintile received medical care compared to 72% inthe poorest. While children with diarrhea and ARI inthe richest quintile were more likely to receive medical care, they tended to receive care in a private facility setting. The percentage o f sick children inthe poorer quintiles receiving care ina public facility were consistently higher inboth surveys. However, after adjusted for other factors, inequalitybywealth inthe use o f ORS and ARI treatment i s not as prominent. There was only one association in2000 in which children o f the richest quintile were more likely to receive ORS during diarrhea compared to those of the poorest group. Similarly, in the case o f ARI treatment, the wealth factor was only observed in the multiple regression for 1995 with the richest and secondrichest more likely to get treatment (See Appendix). Mother's education-based inequality 4.58 Ingeneral, children of educated mothers are more likely to receive care when they are sick with ARI or diarrhea. In2000, the likelihood o f children o f mothers with secondary education to receive ORS, medical care for diarrhea, and medical care for ARI when they got sick was respectively 1.3, 1.6 and 1.4 times higher than those born to mothers with no education. Between 1995 and 2000, while there was a drop in care-seeking behavior for diarrhea in the country, the richer,group was not affected and there was a decline in health promoting behavior among mothers with no education. For example, duringthis period, there was no change inORS use and 6% increase inmedical care for children with diarrhea whose mothers have secondary education. On the other hand, there was a 9% reduction in ORS use and 6% in medical care for diarrhea among those of mothers without education. 4.59 After adjustments are made for other individual andhousehold factors, primary education seemed to be an advantage in ORS use in 1995 compared to no education. However, in 2000, this was replacedby secondary education. 72 Residence-basedinequality 4.60 In both the 1995 and 2000 DHS, care-seeking behavior for ARI and diarrhea was consistently higher in urban than in rural areas. The urban-rural gap actually widened between 1995 and 2000 in the case o f diarrhea. In 1995, children with diarrhea in urban areas were 16% more likely to receive ORS than those in rural areas. In 2000, this likelihood increased to 26%. While children with diarrhea and ARI in urban areas were more likely to receive medical care, they tended to receive care in a private facility setting, as the percentage o f children in the rural areas receiving care in a public facility were consistently higher. 4.61 Regional differences are also observed, especially with the 2000 DHS where there was a clear pattern. People living in the Central region have the best care-seeking behavior for ARI and diarrhea. This was followed by the Eastern, Northern and Western regions in that respective order. Moreover, between 1995 and 2000, while good care-seeking behavior for ARI and diarrhea declined in the country as a whole, Central was the only region that saw improvements. However, after adjusting for individual and household factors, regional differentials are not consistent. Other individual and householdfactors affecting the use of ORsand ARI treatment 4.62 There are some general associations with care-seelung behavior between a child's age and the occupation o f the mother's partner. However, they are neither consistent for both conditions (ARIand diarrhea) or for bothyears (1995 and 2000). There was also one age differential in 2000 with 2-year-old children more likely to receive such therapy during diarrhea compared to 1-year-olds and below. The occupation o f the mother's partner was also an factor in 2000, with children o f a mother with an unslulled partner less likely to receive ORS compared to a skilled partner. Utilization o f curative services for any sickness Figure 4.11 Utilization of curative services in Uganda, 1999 and 2002 60 , 1 50 - El1999 -130 - 40 Ea2002 - 4. 20 - 10 - 0 -r None Public Private* Other Type of facility Source: UgandaNational Household Surveys 1999 and 2002 73 4.63 For this discussion, data from the Uganda National Household Survey 1999 and 2002 were analyzed. Between 1999 and 2002, there was a significant increase in the percentage o f people seeking care for illnesses and injuries reported in the 30 days preceding the survey, from 77% to 87% (Figure 4.11). Most o f the increase i s due to the increase in use o f private facilities (by ll%), whereas the changes in usage o f public facilities i s minimal. Treatment for malaria 4.64 Malaria presents a very heavy disease burden in Uganda, with prevalence rates almost similar to that of childhood diarrhea and ARI. In the household expenditure surveys conducted in 1999 and 2002, the overall monthly self-reported prevalence rate was 15%, accounting for more than 50% o f the total disease burden reported in the previous 30 days (56.2% in 1999 and 52.5% in 2002). This i s more or less in line with the findings o f the 2000 DHS survey in which 35% o f interviewed men 15-49 years old reported at least one episode o f malaria in the three months preceding the survey. Another survey (the DISH Distribution o f Stock and Drug Use Survey) in M a y 2002 also observed that malaria represents 50% o f health center visits inUganda. Rural Urban Total 4.65 Between the two household surveys, the percentage o f people who self-reported malaria but did not seek treatment or only self-treated at home was more than halved. There was a slight increase in use o f malaria treatment inpublic facilities inrural areas. However, a major increase in such use happenedinthe private sector (Table 4.7). 'E". CONCLUSION AND POLICYIMPLICATIONS 4.66 Utilization o f curative and preventive health services i s one o f the direct determinants o f health outcomes. For Uganda, it i s a very a mixedpicture. 4.67 After a concerted effort started in 2000, the country's DPT3 coverage reached 84% in 2003, thus passing its 5-year HSSP target of 80%. ANC attendance at least once with a trained professional also was reported to be high (93%) in the 2000 DHS. However, 74 women do not appear to be getting the benefits o f focused A N C which includes tetanus toxoid, antimalarial prophylaxis, an awareness o f the danger signs o f pregnancy and delivery, and an understanding o f the need to be prepared for needing emergency obstetric care. Uganda i s also doing fairly well inutilization o f curative services for ARI, malaria, and general sickness. Use appears to have increased after user fees were abolished. 4.68 However, the indicator o f institutional deliveries (either in government or PNFP facilities) does not appear to have made sufficient progress. The HSSP target for this was originally set at 70%. The 2002103 target was revised to 28% but this was not achieved as only 20.3% deliveries took place inhealth facilities according to government statistics. Attended deliveries i s the key factor in lowering maternal mortality. As noted in Chapter 2, Uganda is far from achieving the MDG in this area and the low levels of attended deliveries need to be urgently addressed. Efforts need to directed toward eliminating the reasons women give for nonattendance: distance, cost, expectations that they will be treated in an unwelcoming manner by facility staff, and expectations that a home delivery will proceednormally. 4.69 Although Uganda appears to be on track for achieving the MDG goal in access to safe water, there are clearly obstacles inmaintaining the quality o f water at the source and at the point o f consumption. The lack o f adequate sanitation i s clearly another problem. 4.70 Available data point to increasing poor/rich gaps in utilization o f health services between 1995 and 2000 while differences inutilization based on educational background appear to have decreased. 4.71 The use of individual clinical services either for preventive or curative health care, while influenced by individual and household characteristics such as mothers education, knowledge, and attitudes, i s also influenced by the availability, accessibility, affordability, and quality o f services available. 4.72 Taking up health promoting activities such as using bed nets, using clean water, and following good sanitation practices requires not only that these things are available and affordable but also that there is consistent community mobilization to support healthy practices in homes and to maintain public facilities such as latrines in schools. Coordination among the sectors i s critical. 7s 5. HEALTHSYSTEMS PERFORMANCE- GENERATING INPUTSAND DELIVERINGSERVICES A. INTRODUCTION 5.1 The aim o f health services i s to protect or improve health. Whether health services are able to achieve this objective depends greatly on whether critical quality inputs are available, which services are provided, and how the services are organized. The objective o f this chapter i s to describe and analyze health systems performance in Uganda, especially in relationship to making health services work for poor people. The chapter begins by describing the organization and management o f health services in Uganda, including the role o f MOH and other national-level organizations in carrying out stewardship functions o f the health system. Next, health system performance in Uganda interms of geographical access, availability of trained healthpersonnel, drugs and other key inputs, technical and organizational quality, and timeliness and continuity o f care are discussed. The purpose o f this analysis i s to highlight supply-side barriers to high levels o f utilization and coverage o f essential health services. To the extent possible, under each section on performance, government polices and their impact on addressing the problems are identified. The purpose o f this analysis i s to highlight whether the current policies adopted by the government are adequate in helping the government reach its goals as identified in the Poverty Eradication Action Plan (PEAP) or whether there i s a need for new ways o f approaching the problem. B.THEORGANIZATIONAND MANAGEMENTOFHEALTH SERVICES Health services delivery: role of the public and private sectors 5.2 The health sector in Uganda i s composed o f four types o f facilities: hospitals, health centers, dispensaries (health center III), and aid posts or subdispensaries. These facilities can be government owned, private for-profit, or private not-for-profit.-Dispensaries are the most common health facility and a very important health service provider inUganda. Most dispensaries are rural (89%) (Reinikka and Svensson, 2003). Public Sector 5.3 At the district level, the public sector service delivery system consists o f a network o f health centers (11, I11and IV) and district hospitals. This district network i s mandated to deliver the Uganda National Minimum Health Care Package. In addition to the district network, there are two national referral hospitals and ten regional referral hospitals (Table 5.1). 76 Table 5.1 F nctions and Servil s provided at various levels o public sector health facilities Level Pop base/Political Function Services admin. HC I1 5000 Provide preventive, promotive Well child care, immunization Parish andoutpatient (OP) services (fixed & mobile), growth monitoring, ANC, Health Education, Family Planning, Adolescent Reproductive Health, simple treatment o f common illnesses (malaria, dressing wounds) data collection HC I11 20,000 Offer preventive, promotive, OP Very simple surgery (suturing, Sub-county zurative, maternity and in- dressing, draining o f abscesses), patient services maternity services for normal births,PNC, post abortion care, FP (more complex methods e.g. IUD, Norplant), static immunization services, minor dental treatment (mobile), laboratory services HC IV 100,000 Provide preventive, promotive, All functions ofHCI11plus: County OP, curative IP services and EmOC, selected surger;, basic emergency surgery, and blood laboratory services, blood transfusion transfusion, FP (incl. permanent Supervise lower level facilities; methods), adolescent RH, static centralized data dental care collectiodanalysis o f health trends, disease surveillance District 500,000 Laboratory and X-ray inaddition All functions o fHCIVplus: General District to services offered at HC IV. Specialized medical & surgical Hospital Other general services. In- treatment, specialized service training, consultation and investigations - lablradiography research to community based health care Dromams A number o f service standards (for Regional 2,000,000 Inaddition to services offered at eachprogram area) are given for Referral district hosp., specialist services hospitals (noting that they vary Hospital e.g. psychiatry, ENT, among district, regional, national) Ophthalmology, dentistry, intensive care, radiology, pathology, anesthesiology, neurosurgery National 22,000,000 Complicated tertiary care, referral comprehensive specialist Hospitals services, teaching and research Source: MO: 2001b 5.4 Prior to the extensive decentralization reforms undertaken in Uganda, all health facilities inthe public sector were ownedandmanagedbythe MinistryofHealth. In 1993,Uganda introduced decentralization and 56 democratically elected district councils were granted substantial authority for the management o f health care services. These district councils were also granted the authority for other key services such as education and water and sanitation. 77 ITable 5.2 Powers and Functions of District Councils inthe Health Sector Function I DescriDtion Service Organization Hospital autonomy I All hospitals expect two national referral hospitals directly administered by district councils. Contracts with private District Councils are permitted to contract out services to NGO ority over management decisions. User fees were ICommunity participation Through District Health Committees I Source: Bossert et al., 2000 5.5 These decentralization arrangements have changed the role, function and accountability arrangements o f the central government. The MOH currently plays a stewardship function inthe health sector (see description below) and i s also the owner and manager of the national referral hospitals, although there are plans to introduce hospital autonomy with hospital boards in charge of personnel and management decisions. At the district level, there are district health committees (DHC) and district health teams (DHT). The D H C i s an elected group under the district council and provides legislative oversight and support. The DHT serves as the executive arm and i s responsible for technical guidance and support as,well as for the actual delivery o f health care services. Primary health care and district hospitals are the responsibility o f district councils. A new aspect o f the decentralization process i s the development o f health subdistricts (HSD) which are centered at the hospital or health center level and have the responsibility o f managing the lower-level network o f health facilities. The decentralized arrangements aim to encourage community and civil society participation in health services delivery and to increase social accountability o f providers to communities (MOH, 2003b, Bossert et a1 2002, Reinikka and Svensson, 2003). Figure 5.1 District Structure for Health Care Delivery 78 (1 Administrative& 1 Chairperson ChiefAdrmnistrative See Health District District Officer(CAO) ~ ~ O O , O O O I DDHS & DHMT HCIV (HSD) L C N (county county ~100,000 I w Sub-countyI Lcm TownCouncil (SUb-County) 20,000 LCII Parish m (Parish) (5,000 1 v Village HealthCommittee LCI(Village) I1,000 I IFarnilTmdIndividualCommunityMembers I ~ Key: CAO =ChiefAdrmnistrativeOfficer DDHS =DirectorofDistrictHealthServices DHMT=DistrictHealthManagementTeam DHC =District Health Committee HSD =HealthSub-District HC =HealthCentre LC =LocalCouncil Source: Draft PEAP Document, Ministry of Health, 2003 Private For-profit and not-for-profit (PNFP)health facilities7 5.6 The private sector (for-profit and not-for-profit) has a long history o f providing health services in Uganda, including for the poor, and i s a very important provider of health 'The descriptionofPNFP facilities inUgandadraws heavilyfrom: ReinikkaR and SvenssonJ. 2003. Working for God?EvaluatingService Delivery of ReligiousNot-for-Profit HealthCare Providersin Uganda.DevelopmentResearchGroup. The World Bank. 79 services inUganda. The private not-for-profit (PNFP) health sector inUganda consists o f religious and nonreligious providers. A census of PNFP providers carried out in 2001 indicated that autonomous dioceses and parishes own about 70% o f all PNFP health facilities, which total 450 lower-level units (dispensaries) and 42 hospitals (MOH, 2001a). The rest are owned by NGOs, some o f which are also religious, community- based organizations (6%), and by the district council, mosques and individuals (8%). Not-for-profit health care providers are self-regulated and self-governing. This means that PNFP facilities are free to decide on the mix and price o f services provided by the facility. The Ugandan Catholic Medical Bureau has recently initiated an accreditation process and issues guidelines for lower-level health facilities that are affiliated with it. This states that diocesan health units do not aim at profit. Their central objective is to provide quality health services at low cost, be well organized, aim at integration into the district health service delivery network, and operate ina community-oriented manner. 5.7 Government subsidies for the PNFP sector have a long history in Uganda. In the years after independence, the relationship between religious providers and the government deteriorated, and public subsidies were reduced. In 1997/98, the Government reinstated financial aid to PNFP hospitals and in 1999/2000, a new program extended a similar subsidy to lower level health facilities. According to this program, each PNFP facility was to receive a fixed-amount grant for the fiscal year. Each dispensary was to receive the same amount (2.5 million Ugandan shillings or US$1,400 a year). It was expected that in subsequent years, the subsidies would become more refined, taking into account needs as well as performance criteria. Evaluation o f the impact o f this financial aid program on the mix o f services provided by religious PNFP facilities shows that this program has had a very positive effect on testing malaria and intestinal worms. For example, a PNFP provider with the median grant receipt test on average 24 more patients o f very 100 suspected malaria cases (Reinnika and Svensson, 2003). 5.8 Generally, PNFP facilities provide a mix o f public and private goods and the service mix similar to that o f government facilities. For example, PNFP facilities provide outreach, health education, immunizations, and antenatal care. Table 5.3 Health Facilitiesby Level and Ownership Ownership Levelof Government PNFP Private for- Total Facility profit Hospitals 56 (54%) 44 (42%) 4 (4%) 104 (100%0 H C JY 143 (93%) 8 (5%) 3 (2%) 154 (100%) HC I11 650 (80%) 147 (18%) 12 (2%) 809 (100%) HC I1 845 (58%) 362 (24%) 262 (18%) 1469(100%) Total 1694(67%) 558 (22%) 282 (11%) 2536 (100%) Source: Health Facility Inventory 2002 Private For-ProJit Sector 5.9 Detailed information on the size, types o f services, and quality o f health services offered inthe private for-profit (PFP) sector inUganda is limited. Overall, the private for-profit sector i s quite diverse and consists o f hospitals, clinics, nursingand maternity homes, and drug shops, as well as traditional healers and midwives (Corkery, 2000). Community 80 drugvendors are an important and unregulated component o f the for-profit private sector and in rural areas, provide about 75% o f Westem drugs at the community level. According to the limited information that i s available, community drugs vendors are the first point o f contact for the rural population for a range o f communicable diseases (childhood illnesses such as ARI and diarrhea, malaria, tuberculosis) (Twebeze, 2001). 5.10 Private for-profit PHC facilities largely focus on private goods: outpatient and medical care, and minor surgery. Generally, willingness to pay for these services i s high among patients and this generates a market demand for the growth o f private providers. Private providers are largely predominant inprimary care. For-profit facilities also provide some preventive services such as antenatal care (ANC) and family planning, although few offer immunizations. There i s a reason for this: the Expanded Program o f Immunization (EPI) delivers supplies to health facilities for immunizations and private for-profit facilities are not as yet included under the EPI program. Laboratory services are provided at the same levels as PNFP facilities (Lindelow, Reinnika and Svensson, 2003; Reinikka and Svensson, 2003). 5.11 The limited data that i s available (e.g. health facility inventories o f 1996 and 2002) indicate that the private for-profit sector i s growing very rapidly, especially in the area o f primary health care (MOH, 2001a). The results fi-om the inventory should be treated cautiously since it did not include the entire formal sector and none o f the informal sector. It i s most likely the case that the private for-profit sector i s larger than estimated by this survey. There is some data that suggests that private facilities (such as drug shops and primary health care clinics) are managed by public sector health workers (Jitta et al, 2003; McPake et a1 1999). Existence o f this practice, even on a small scale, indicates informal public sector subsidies to the private sector. Private for-profit facilities are required to register with the GoU. However, overall, there i s very limited regulation o f this sector. The unlicensedcommunity drugvendors that largely operate inruralareas are completely unregulated (Birungiet al, 2001). Role of MOH and other national entities 5.12 Indecentralized health systems such as Uganda's, central institutionsplay an important role in assuring that the stewardship functions o f the health system are conducted adequately. Stewardship functions include: health sector policy making, priority-setting, oversight o f health status, health systemplanning, quality assurance, human resource and capacity building, monitoring and evaluation, and advocacy. Although responsibility for stewardship rests primarily with the MOH, inUganda, the SWAP is the framework for implementing the Health Sector Strategic Plan (HSSP), which means that a wider range o f institutions, including donors, are involved. There are a number o f mechanisms available to the GoU to carry out its stewardship functions. These include different units inthe MOH, professional councils, the private not-for-profit bureau, SWAP processes, and the commissioning o f specific studies, e.g. the tracking studies discussed in Chapter 6. The units o f the MOH responsible for planning, standard setting, and quality assurance functions are expected to work closely with district level authorities. 5.13 The National Health Plan (NHP) lists six core functions o f the MOH: a) policy formulation, setting standards, and quality assurance, b) resource mobilization, c) capacity development and technical support, d) provision o f nationally coordinated services e.g. epidemic control, e) coordination o f health research, and f) monitoring and evaluation o f the overall sector performance (MOH 1999). Together with the special 81 attention the NHP gives to providing `an effective framework for strategic policy review and formulation, planning, budgeting, monitoring and evaluation,' these core MOH functions are consistent with stewardship responsibilities. 5.14 Some functions, including planning and budgeting, were anticipated when the NHP was finalized, and the Directorate o f Planning and Development has been generally strengthened since then (Brown 2000, MOH 2003h). Although the Health Planning Department i s charged with strategic planning in the HSSP, it appears that priority setting i s established through the SWAP. 5.15 Stewardship is, however, more than monitoring: it involves ensuring that intentions are implemented. This i s brought about by health related laws, professional councils, and regulatory agencies. Problems have been identified with health legislation and regulation: existing laws and regulations are not comprehensive and do not address key aspects o f the health sector; most are outdated; and they are not enforced. (Okuonzi 2002). Professional councils have been found not to carry out their responsibilities and pharmaceutical sector regulation appears to be ineffective (MOH 2003b). 5.16 The function o f human resources and capacity-building appears to have encountered even greater difficulties. As noted elsewhere inthis report, large numbers o f posts continue to be filled by untrained staff, especially at the lower PHC levels. A human resource plan linked with a training plan was to be developed. Coordinating all the training activities i s also a challenge. The private not-for-profit sector i s also a substantial actor in this area as they operate 20 out o f 48 o f the training schools in the country. The separation o f manpower planning and training appears to have hampered the development o f a comprehensive humanresource plan, which i s sorely needed. 5.17 Beyond oversight, stewardship also involves advocacy and accountability. Although not addressed earlier, the MOHreports that `one o f the cardinal roles and responsibilities o f the center i s to advocate for health services'(M0H 2003b). The inclusion o f advocacy and a discussion o f regularly assessing client satisfaction inthe SWAP reporting process can perhaps been viewed as an indication o f the changing role o f central authorities in a decentralized health system. 82 C.HEALTH SYSTEM PERFORMANCE Health infrastructure and impact on geographical access to health care services 5.18 Geographic proximity to health facilities i s one o f the key factors that affects utilization o f health services. Results from a survey in Uganda found that proximity to health facilities was the most important factors affecting the decision to seek care in a government facility (Reinnika and Svensson, 2003). Health facility coverage within 5 k m s i s improving in Uganda and in 2000, a inventory o f 35 districts showed that 57% o f the population was within 5 k m s o f a health facility. This survey included public and PNFP facilities (MOH, 2003b). This i s an improvement over the figure o f 49% during the late 1990s. Under the HSSP, the GoU plans to achieve a coverage rate o f 80% by 2005 (MOH, 2003b, MOH, 2003~). 5.19 Despite these improvements at the aggregate levels and the continuing efforts o f the GoU, there are urban-rural, regional-district level, and income disparities. For example, in the 2002 National Household Survey (NHS), 46% in the poorest quintile reported distance as a constraining factor for seeking health care services as compared with only 25% in the richest quintile. Those living in the Eastern (25%) and Northern regions (21%) face greater geographical access problems versus those living in Central (12%) and Western (15%). Overall, rural-urban differentials are sharp with 20% o f the rural population mentioning distance as a factor in not seeking health care as compared with only 2% in urban areas. Since living in a rural area is one of the characteristics of poverty in Uganda, there is relationship between poverty and geographical access (Uganda HouseholdBudget Survey, 2002). Poor physical access to delivery has also been noted as one o f the major factors for l o w turnout o f mothers for attended deliveries (MOH, 2003e). 5.20 Figure 5.2 shows district-level breakdowns in terms o f the percentage o f population within 5 km radius by district. This indicates that 22 districts fall below the national average o f 40% and, as expected, access i s the best in urban districts such as Kampala, Jinja, and Pallisa. 5.21 The MOH estimations o f population within 5 km does not cover private for-profit facilities. Formal private providers are largely concentrated in urban areas and therefore, geographical access in urban areas such as Kampala and Jinja are probably even higher than the numbers presented in Figure 1. In fact, there may even be an oversupply o f public and private health facilities in urban areas and in certain regions o f the country In contrast, inrural areas, there are not enough public and PNFP facilities, and people living inthese areasare largely served byinformal providers such as community drugvendors. 83 Figure5.2 Percentageofpopulationwithin5 KMradiusby districts 0 20 40 60 80 100 120 Source: WHO, 2003 5.22 There i s limited data on infrastructure and health facilities. One of the surveys among primary health care units (public and private) found that for the majority o f facilities, boreholes are the primary source of water, but other sources such as piped water, spring water, and rainwater were also important. Generally, private facilities tend to have piped water. There are regional differences and facilities in the northern region have no access to piped water and tend to depend on dumping for water disposal. Generally, government facilities are more remote than private facilities and also have limited access to telephone services. Government facilities are also further away from district and health sub-district headquarters (Lindelow, Reinikka and Svensson, 2003). Health facility infrastructure in the regions affectedbyconflict isparticularly aproblem(International RedCross, 2002). Availability of HumanResources 5-23 The availability o f skilled and motivated health personnel i s an important pre-requisite for the delivery o f quality health services. The analysis o f human resources needs to focus on the number o f staff, their qualifications, and also an appropriate mix o f different types or levels o f human resource. Table 5.4 describes the number o f key health personnel (physicians and nurses) per 100,000 population inUganda as compared with its neighbors (Kenya and Tanzania) and the SSA average. The data show that Uganda faces a severe problem with the number o f nurses in the country. In terms of physicians, Uganda i s also far below the SSA average and far below Kenya, which has a lower GDP per capita than Uganda. 84 Table 5.4 :HealthPersonnelinUgandaas comparedwith Neighborsand SSA Average Country GDP per Physicians Nursesper Midwivesper Pharmacists capita per 100,000 100,000 100,000 per 100,000 population population population population Uganda US$367 4.7 5.6 13.6 NA Kenya US$325 14.1 108.0 NA 23.O Tanzania US$204 4.1 85.2 44.8 NA sub-Saharan _ _ 15.5 73.4 30.9 1.1 Africa Average Source: For physicians and nurses, WHO 2003. For midwives and pharmacists, WHO, 1994-1998 data, varies by country. For GDP per capita, the data is from 2002. 5.24 The political problems in Uganda in the 1960s and 1970s caused many health personnel to leave the country, but there has been some recovery in the number o f healthpersonnel since 1972. However, the growth in the number o f health personnel has not kept pace with population growth (Hutchinson 2001) and Uganda continues to face a problem o f brain drain, where medical students leave the country after completing their education to work inneighboring countries or other countries where salaries are higher (Reinnika and Svensson, 2003). 5.25 In addition to an overall shortage, there are also problems with the mal-distribution o f staff by level and location o f health facility. The MOH recently completed an evaluation o f staff in health facilities as compared with staffing norms of the govemment in 42 districts. The data shows a gap between the norms and the number and designation o f staff at various levels o f health facilities. For example, HC I1i s not supposed to have clinical staff or laboratory personnel, yet the survey found these categories of personnel were present. There i s a 86% gap innursing staff and nursingaides. These are staff that are required at the HC I1level according to the govemment guidelines. At the HC 111 level, however, there are shortages in clinical (34%) and laboratory staff. The gaps in nursingstaffis also very acute at this level (MOH, 2003b). 85 Table 5.5 Actual Number of Staff andMinimumStaffingNorms All Districts, GoUandNGO - HealthFacilities HCII HC I11 HC I V I Hospital(GH) I Total Table 5.5 shows district-level breakdowns for health workers in the public sector per 100,000 population. This shows that there are considerable discrepancies across districts. The national average i s about 80 health staff per 100,000 population. Indistricts such as Katakwi, Sembabule, Ntungamo, Kibaale and Busia, there are less than 10-30health staff per 100,000 population while districts such as Kampala, Mpigi, Jinja and Rukungiri have two to three times more staff per 100,000 populationthan the national average. 86 Figure5.3 Totalnumber of healthworkers inthe publicsector per 100,000 population Katakwi Sembabule Ntungamo Kibaale Busia Kotido, Bushenyl Kamuli Apac Lira Mubende MOYO, Rakai Bugiri Hoima Kasese Mbarara Pallisa Kiboga Kapchotwa Bundibugyo Tororo Kabale Adjumani Masindi Kumi Nakasongola National average Nebi Arua Masaka lganga Luwero Mukono Kisoro Gulu Mbale Kalangala Soroti Moroto Kitgum Kabarole Rukungiri MJinja pigi Kampala I 0 50 100 150 200 250 Source: MOH 2002 5.26 Figure 5.3 presents data on the human resource mix by district and this shows that there i s also considerable variation across districts. For example, in Kampala, there are almost 45 doctors per 100,000 populationwhile i s some districts such as Kamuli and Apac, there i s less than one doctor per 100,000 population. There are similar variations for other health workers such as nurse midwives, auxiliary health personnel (AHP), and nurses. As expected, Kampala district has the largest number o f health professionals and nursing and midwifery professionals per 100,000 population. It also has the least number o f nursing and midwifery associate professionals (nursing assistants) who are the least qualified health personnel. Jinja district has the second largest number o f health professionals and also nursing and midwifery professionals per 100,000 population (WHO, 2003). 87 - 1Figure 5.4 District-level breakdowns of Health personnelby Category Tororo Soroti Sembabule Rukungiri Rakai Pallisa Ntungam Nebi Nakasongola Mukono Mubende Mpigi Moyo Moroto Mbarara Mbale Masindi Masaka Luwero Lira Kumi Kotido Kitgum Kisoro Kiboga Kibaale Katakwi Kasese Kapchowa Kamuli Kampala Kalangala Kabarole Kabale Jinja lganga Holm Gulu Busia Bushenyi Bundibugyo Bugiri Arua Apac Adjumni 0.00 50.00 100.00 150.00 200.00 250.00 Health Prof/100,000 E! N&M/100,000 HAHP/I00,000 HNurse/100,000 Source: WHO. 2003 ~~ 5.27 There i s very limited data on staffing patterns in the for-profit sector. One o f the few available surveys found that, on average, government and PNFP facilities have more staff than private for-profit facilities. In general PNFP are the most well-staffed in terms o f a doctor or clinical officer, followed by for-profit facilities (Lindelow et al., 2003). 88 Table 5.5 Staff Mix of Facility, Ownership Category, and Region Frequency No. of doctors or clinicalofficers Assistant or non-medical staff only Facility size (# o f staff) (Percent) (Percent) Small (1-5) 57 79 19 Medium (6-8) 56 46 5 Large (9-16) 42 21 0 Ownership Government 81 38 6 Private for-profit 30 58 16 Privatenonprofit 44 73 9 Source: Lindelow, Ritva and Svensson, 2003 5.28 Various studies have documented problems with public sector health personnel. Health worker absenteeism i s one o f the problems and a recent study covering 98 PHC facilities and 799 providers found that approximately 37% o f providers were absent. This rate i s higher than some comparator countries in the study such as Peru (23%) and Bangladesh (35%) but slightly lower than India and Indonesia (40%). Doctors were more likely to be absent (48% absenteeism rate) and pharmacists were least likely to be absent. Absenteeism rates were highest in government facilities and the lowest inprivate (NGO) facilities although the range across facilities was not very wide (36% in Government facilities as compared with 29% inprivate NGO facilities). The range in the absenteeism rate was wider across districts with Kisoro having a rate of 52% and Bushenyi, half o f that (26%). The most common reason for absenteeism among doctors was "official duties." The study found that strong supervision was an important correlate o f reducing absenteeism while the corelationship with community boards was weak (Chaudhury et al. 2004). 5.29 Other studies have documented problems with informal economic activities within public health facilities such as mismanagement o f drug supply, informal charging, mismanagement o f user fees (prior to their abolition), offering treatment in health workers' homes (private practice), ownership o f clinics and drug shops (managing private facilities), and part-time work in other jobs (most commonly private clinics) (McPake et al, 1999). These practices have important negative implications for quality and accessibility o f public sector health services, and have an adverse effect on the utilization ofpublic sector health services by the community. Humanresourcepoliciesof the governmentandimpact' 5.30 The government i s well aware o f the human resource problem in the health sector and various policies have been implemented, either inthe health sector or as a part o f overall civil service reforms. The objective o f the civil service reforms in the human resource sector are to ensure that public sector staff deliver timely and quality services to the population at the least cost and support national development. Key policy changes include: (i)decentralization o f management o f personnel, (ii)introduction o f Results Oriented Management (ROM), and (iii) capacity building.The process o f introducing the reforms has taken time, and initially when the decentralization reforms were introduced, 8 This section draws heavily fromthe work o f Corkery, 2000 89 there was limited capacity at the district level to carry out the new powers and functions (Corkery, 2000). The key impact o f the civil service reform on the health sector so far has been: 0 Since District Councils are now responsible for hiring and firing, staff cannot be moved between districts by central authority. This has contributed to inequality across the districts and remote and disadvantaged districts have been unable to attract sufficient personnel. 0 Although, technically a part o f the government's human resource strategy, ROM has not been introduced in the health sector and the institutional and organizational infrastructure needed to implement ROM i s not in place. Overall, accountability of public sector health personnel to clients and district level staff i s weak. 0 A positive development is that new doctors are required to serve ina rural area for a period o f two years before they are eligible for senior positions. Despite this policy, hard to reach and insecure districtshave very few staff. 0 The Government has a plan to increase the salaries o f public sector health staff, including paramedical staff, since it i s felt that this would improve motivation to work. Yet, a study o f PNFP providers in Uganda showed that PNFP facilities were able to hire qualified medical staff below market wages, and provided better quality care than government facilities. This i s partly an indication that staff hired into PNFP facilities may have altruistic motivations (Reinikka and Svensson, 2003) but could also reflect the fact that staff may also appreciate the conditions under which they work (i.e., better management, less bureaucratic controls). Nonetheless, these findings indicate that simply increasing salaries o fhealthworkers inthe public sector in the absence of other changes (inputs, organizational arrangements, prospects for career advancement, levels o f autonomy and most importantly, accountability) may have little impact on changing the behavior o f health workers inthe public sector. Availability of material resources (pharmaceuticals, supplies) 5.31 The availability o f drugs and other medical equipment and supplies inhealth facilities i s a key factor affecting perceived quality o f care and the decision to seek care. It is also important from the perspective o f technical quality, since trained staff cannot apply their skills in clinical management unless they have access to these inputs. Pharmaceuticals and vaccines 5.32 Procuring and distributing drugs and other medical supplies in Uganda was initially entirely centralized. With the adoption o f decentralization, districts and health sub- districts now receive a budget to procure their own drugs and supplies based on local needs. Centrally controlled funds are used for the procurement and distribution o f vaccines, ORs, contraceptives, TB drugs, and drugs for epidemics. Health subdistricts procure drugs from national medical stores (NSM) and distribute the drugs to the health facilities. For vertical programs, such as immunizations, procurement and distribution are centrally managed. 90 5.33 A survey o f PHC facilities inUganda found that 40% of government and PNFPfacilities reported having run out o f some or all vaccines in 1999/2000. Although facilities report being resupplied without delays, in some cases, stockouts lasted from 1-12 weeks. Stockouts o f contraceptives was also common. (Lindelow et al., 2003). A recently completed drug-tracking study in four districts o f Uganda found similar problems with stockouts. This study found that stockout o f important drugs such as antibiotics was high (35-55%). For example, cotrimoxazole, which i s commonly used to treat ARI, had a stockout rate o f 54%. (Table 5.6). This has important negative implications for reducing mortality among under-five children fi-om ARI. Interms o f ownership, public facilities had overall high stockout rates for all drugs (30%) as compared with PNFP facilities (16%). This is partly explained by the fact that PNFP facilities may have better cash flow and can buy pharmaceuticals fi-om the private market while public facilities have to tum to either the National Medical Stores or the Joint Medical Stores (MOH, 2003a). Table 5.6 PercentageStockoutTime for 10 IndicatorItemsbyLevel Item H C I I HCIII HCIV HOSP Average Stockout Ampicillin or Amoxycillin 250 63% 50% 64% 24% 55% mgcaps Chloroquine 150mgbase tabs 6% 6% 10% 0% 7yo Ciprofloxacin 250 mgtabs 1% 8% 4yo Contrimoxazole 480 mgtabs 54% 26% 39% 10% 35% Measles Vaccine 0% 2% 1% 1Yo Mebendazole 100mgtabs 40% 23% 23% 4% 26% Methylergometrine 200 mcg/ml 42% 23% 22% 0% 24% inj Paracetamol500 mgbase tabs 37% 24% 29% 0% 27% Procaine Penicillin Fortified 4 48% 22% 23% 0% 27% m.uinj Quinine 300 mg/mlinj 41% 34% 33% 1% 33% Sulphadoxine-Pyrimethamine 59% 47% 25% 0% 41yo 525 mg tabs Average Stockout 40% 26% 27% 5% 28% Source: Drug Tracking Study, MOH, 2003a Table 5.7 PercentageStockoutTime for 10IndicatorItemsby Ownership Item GoU NGO Average Stockout Ampicillin or Amoxycillin 250 mg caps 59% 28% Chloroquine 150mgbase tabs 8% 0% Ciprofloxacin 250 mgtabs 1% 8% Contrimoxazole 480 mgtabs 40% 8% Measles Vaccine 2% 0% Mebendazole 100mgtabs 28% 17% Methylergometrine 200 mcg/ml inj 28% 0% Paracetamol 500 mgbase tabs 29% 17% Procaine Penicillin Fortified 4 m.uinj 27% 32% Quinine 300 mg/mlinj 32% 41% Sulphadoxine-Pyrimethamine 525 mgtabs 44% 17% Average Stockout 30% 16% 28% Source: Drug Tracking Study, MOH, 2003a. 91 5.34 Emergency and Obstetric Care (EmOC) needs assessment conducted among public sector health facilities found that essential drugs needed to provide basic and comprehensive EmOC were lacking. For example, the study found that only two referral hospitals and four district hospitals had on stock antihypertensive drugs. Less than half o f the facilities had on stock basic oxytocins (MOH, 2003d). Antihypertensive drugs are needed to treat pregnancy-related hypertension, which if untreated can be fatal for the mother. Oxytocins are used to induce labor in case o f delayed labor, and are also needed when there are complications with the labor and delivery process. 5.35 There i s a consistent relationship between high patient volume and stockouts (MOH, 2003). Since the abolition o f user fees has increased outpatient treatments, there are concerns o f a worsening situation with regard to drug stockouts. Inthe medium term, this could affect service utilization since patients perceive the availability o f drugs to be an important indicator o f quality. There i s also a close relationshipbetween the level o f the health center and the stockout rate. Generally, the stockout rate i s higher i s lower level facilities (HC I1and 111) than in H C IV and district hospitals (MOH, 2003a). This has implications for access and efficiency since for rural populations, HC 11s are the closest health facilities. Iflower-level facilities do not have drugs, the tendency will be to depend on community drug vendors or bypass the HC I1level to higher levels (HC IV or district hospitals). Results from a 1999 survey among health workers also showed that public sector health workers leak drugs and runmany o f the drug outlets inrural areas (McPake et al, 1999). 5.36 There are several factors behind problems with timely supply o f pharmaceuticals and vaccines: the financing problem and delays with the PHC grant (see Chapter 6, on health care financing), a fragmented and duplicative supply system, problems with the availability o f drugs inNational Medical Stores (NMS) (see Table 5-10), limited capacity at the district level and health facilities on drug supply management (forecasting), the lack o f good communication and information systems, and the lack o f transportation which would enable districts to distribute drugs to health facilities in a timely manner (MOH, 2003a). 92 Table 5.8 Availability of EssentialDrugs and Other MedicalInputs at the NationalMedical Stores Doxycycline caps 87% Chloroquine inj. 66% Chlorhexidine sol. 63% Plaster Adhesive zinc oxide 62% Syringes wheedle 2 ml 54% Ferrous/folate tabs 52% BandageWOW75% 42% Chlorphenamine tabs 33% Amoxycillin caps 23% Gloves examination 22% Cotrimoxale tabs 19% Oralrehydration salts 18% Mebendazole tabs 18% Procaine penicillin tabs 18% Sulphadoxine-Pyrimethaminetabs 8% Nystatin pessaries 72% Suture silk black 72% Ciprofloxacin tabs 65% Suture catgut chromic 36% Acyclovir tabs 47% Methylergometrineinj 22% Quinine inj 20% Source: Annual Health Sector Performance Report, FinancialYear 2002/03, MOH, 2003b. Medical Equipment and Supplies 5-37 Table 5.11 describes the availability o f equipment by ownership category and this shows that the availability of some equipment i s consistent with the type o f service provided. For example, 90% o f govemment facilities have cold-chain equipment (needed for immunizations) while only 44% have microscopes. This i s consistent with the fact that few government facilities provide laboratory services. However, it i s surprising that only 50% o f private and 61% o f PNFP facilities have a microscope since they claim to provide laboratory services. This has implications for quality o f care. For other medical and nonmedical consumables, facilities mentioned stockouts and in some cases there was no resupply for over 20 weeks. More than 40% o f government facilities reported buying their own medical consumables and the supply o f syringes i s particularly problematic. A study o f two health centers in Masindi district in Nothern Uganda (an area affected by conflict) found that the health centers were not able to provide good quality safe 93 motherhood services since they did not have access to basic medical equipment such as manual vacuum kits (MVA) needed in the case o f obstructed labor, and newborn resuscitation equipment (International RedCross, 2002). Table 5.9 Availability of Equipment by Ownership Category (YOof facilities) Type of Equipment Government Privatefor-profit PNFP Sterilization equipment 100 83 100 Refrigeration equipment 90 13 66 Weighing scales 94 77 93 Height measurement equipment 41 13 16 Bloodpressuremachine 90 100 86 Microscope 44 50 61 Sets o f protective clothing 43 43 45 Source: Lindelow, Reinnikka and Svensson, 2003 Technical Quality of Health Care Services 5.38 Technical quality of care i s defined as "performance according to standards that achieves the most favorable balance between health risks and benefits." (Donabedian 1986). Inferior technical quality can contribute to poor health outcomes and high health care costs from wasted expenditures. Technical quality of care can be measured in terms o f structure, process and the outcomes o f care. Timeliness and continuity o f health services are also important hallmarks o f quality and clinical guidelines typically address the linkages between the different levels o f care in terms o f timely referrals and follow-up care. 9 5.39 The GoU has adopted internationally accepted clinical guidelines for priority problems in the health sector, such as childhood communicable diseases (Integrated Management of Childhood Illnesses). The I M C I guidelines aim at improving care to prevent avoidable mortality among under-five children and the effective implementation o f these guidelines can reduce mortality from ARI and diarrhea (WHO 2001). 5.40 An evaluation o f the implementation o f IMCI in public sector health facilities in ten districts o f Uganda shows that as a result o f IMCI training, overall clinical practices o f workers trained in I M C I was much better than nontrained health workers. Nevertheless, for some key actions such as "a child needing vaccinations leaving facility with all needed vaccinations" there was little difference between the performance o f trained versus nontrained health workers and in fact duringthe first round o f the study, those in the nontrained category performed better. Some o f the key findings o f the study are that despite training in IMCIas well as the availability o f structural inputs ,process quality o f ~ _ _ Structure refers to the inputs (trained staff, drugs, medical supplies and equipment) needed to provide quality health services. For example, to adequately treat acute respiratory infections (AM) among children and adults, access to antibiotics i s needed. Process refers to clinical management practices and outcomes refer to health outcomes adjusted for other risk factors that can help determine whether facilities are providing quality services. For example, ifthere i s a great deal o f variation inmortality rates across hospitals, this could either meanthat a hospital i s treating very high-riskcases or that the quality o f care i s poor. 94 care was not adequate and there was room for further improvements. The study concludes that one o f the critical missinglinks in improving technical quality o f care for I M C I i s weak district level supervision o f healthproviders (Pariyo et al. 2003). 5.41 There are few studies in Uganda that have compared technical quality o f care across public and private facilities. One study that compared public, PNFP, and for-profit facilities found that, in general, private providers were much more likely to to prescribe and carry out the correct treatment for malaria and intestinal worms. This difference in practice i s not due to the lack o f inputs. Another important quality indicator i s rational drug prescription. In all facilities, antibiotic prescriptions are generally very high. However, government facilities are much more likely to provide antibiotics than private providers. The effect i s particularly strong in lower-level government facilities without qualified medical staff (Reinikka and Svensson, 2003). This means that people in rural areas are more likely to receive higher antibiotic prescriptions since rural facilities are generally staffed by nonmedical staff. Community drug vendors inrural areas also tend to misuse antibiotics, practice poly pharmacy (prescribing multiple pharmaceuticals), do not give clear instructions about the use o f the dispensed drugs (including possible contraindications), and almost never gave preventive health care messages (Twebaze, 2001). A study o f technical quality o f care in public and PNFP hospitals found that in PNFP hospitals, patients were more likely to receive all the drugs that were prescribed (92% o f cases). In contrast, in public hospitals only 46% o f patients received all the drugs (Ssengooba, 2002). This has implications for patient compliance since patients are more likely to comply when they are provided with all the drugs rather than having to buythe drugs themselves from pharmacies. 5.42 Data from household surveys (DHS and UNHS) show that there are gaps in the quality and comprehensiveness o f antenatal care (ANC). The study found: e Low number of visits: Less then half o f the women (46%) receive the minimum recommended four visits. e Delayedfirst visit: 44% o f the women make their first visit during the last three months o f the pregnancy when it i s too late to identify complications. Median stage o f pregnancy at first visit i s 5.9 months. e Lower coveragefor the high risk births: ANC visit i s very low in sixth and higher order births,which comprise almost 30% o fall births). e Poor technical care: Only 26% o f the women were informed about the danger signs o f pregnancy complication and only two-thirds got their blood pressure measured despite having four or more visits to the provider. The percentage i s even lower for the women who had less than four visits. 5.43 Adequate antenatal care i s a necessary but not sufficient input into good pregnancy outcomes for the mother and child. Poor technical quality partially explain the highlevels o f maternal and perinatal mortality inUganda (World Bank analysis o f DHS and UHBS, 2004). Perceivedquality of care 5.44 Characteristics such as cleanliness o f the clinic, waiting time, hours and days that the clinic i s open, and behavior o f clinic staff are key to influencing perceived quality o f care. Perceived quality o f care i s one o f the factors affecting patient decisions to seek care inone type offacility or another. 95 5.45 Typically government facilities are more restrictive in their hours o f operation as compared with PNFP and PFP clinics. For example, one survey found that 65% o f government facilities were open seven days a week as compared with 81 and 89% o f PNFP and PFP facilities respectively. Government facilities are also open fewer hours as compared with PNFP and PFP facilities. PNFP facilities had a health staff on call 24 hours a day while PF facilities were open for 10 hours every day (Akwara P et al. 2003). In general, waiting time is the longest in government health facilities (UNHCO, 2002). Women report attitude o f health workers in public facilities as a key factor influencing the decision to seek ANC and delivery care. Patients report high satisfaction o f care with PNFP and PFP facilities, including informal private providers such as drug vendors. For example, patients report the many advantages o f buying drugs from community drug vendors, including convenience, proximity to the household, easier credit facilities as well as less restrictions on the amounts o f drugs obtained as compared with public facilities. Patients recognize that community drug vendors have less technical knowledge o f treating diseases that health workers inpublic facilities. Despite this, they prefer drug vendors over other types o f public and private facilities for first contact. (Twebaze, 2001, UNCHO, 2002). The study o f public and PNFP hospitals found that patients rated PNFP facilities higher than public facilities because o f the availability o f drugs (Ssengooba, 2002). In one o f the surveys conducted among government and private facilities, patients reported that good treatment andor good staff were the most important reason for visiting private facilities (Reinnika and Svensson, 2003). Relevanceof services 5.46 The mix o f services provided by the different levels o f public sector health facilities in Uganda i s designed to enhance cost-effectiveness, quality, and continuity o f services. Nonetheless, there appears to be a gap between the services actually provided by government facilities and the norms. For example, only 50% o f facilities inthe Western Region provided delivery services as compared with 90 and 84% respectively in Central and Eastern. For immunizations, only 65% o f Northern facilities provided this service as compared with 91% inCentral (Lindelow, Reinnika and Svensson, 2003). 96 Table 5.10 M i x of services providedin government, PNFP and private facilities in four regions of Uganda (percent) Service Central Eastern Northern Western Total Curative and Diagnostic Services Outreach 82 80 74 63 77 Outpatient care 100 100 100 100 100 Inpatient care 53 77 57 66 63 Medical care 100 91 87 87 93 Eye care 53 29 61 17 41 Mental health care 16 5 13 30 15 Dental care 30 9 39 7 21 Minor surgery 81 80 74 100 83 Deliveries 90 84 61 50 76 Laboratory services 37 13 46 50 34 Preventive Care Health Education 97 93 100 90 95 Immunizations 91 77 65 75 81 Antenatal care 98 91 74 89 91 Family planning 90 86 83 76 85 Source: Lindelow, Reinnika and Svensson, 2003 5.47 Given the high levels o f maternal mortality in the country, improving access to comprehensive reproductive health care services i s a priority. Yet, as many studies have shown, there are major gaps inthis area. An evaluation o f Emergency and Obstetric Care (EmOC) inUganda across 19 districts and 197 public sector health facilities found gaps in the availability of comprehensive and basic EmOC at the hospital and health facility levels. For example, out o f the 19 districts included in the survey, only eight had some facilities that met the criteria for EmOC. Of the 197 facilities surveyed, only 28 (14%) offered EmOC services. According to international standards, for every 500,000 population, there should be at least one comprehensive EmOC and four basic EmOC facilities. Table 5.14 shows the minimum required level o f EmOC per catchment population o f the sampled districts (population: 9,392,537). This implies that in the catchment populationo f the surveyed districts, 19 comprehensive EmOC facilities and 76 EmOC facilities are needed. The current status i s that there are 23 CEmOC and 18 EmOC facilities present. There i s a big gap as far as basic EmOC services are concerned (MOH, 2003d). Table 5.11 The status of EmOC services inthe surveyed districts. Minimumrequired level Current status Gap C-EmOC 19 23 +4 B-EmOC 76 18 -56 Source: MOH, 2003d Social accountability 5.48 There i s growing evidence that one o f the core issues in ensuring quality and cost- effective health service delivery, especially for the poor, are the governance arrangements 97 that increase transparency and accountability between various actors in the health care system (policymaker and provider, policymaker and client and client and provider) (World Bank, 2004). One o f the most important tasks for political and administrative decentralization inUganda i s to bringthe public service delivery system closer to clients and to improve transparency and accountability in the delivery o f health care services. It i s important to assess the success o f these policies and gaps and to address the issue o f transparency and accountability. 5.49 As described earlier, community participation in the planning and management o f the health sector i s guaranteed through the subcountry and village-level health committees that participate in district health committees. There are only a few districts where village- level health committees exist and there i s little information on the effectiveness o f subcountry and village-level health committees. Another potential avenue o fparticipation i s the health unit management committees (HUMC). These are nine-member committees that are elected to oversee health personnel, inspections, expenditures, construction and maintenance concerns, and to decide how revenues from user fees should be used at the facility level. The HUMCs provide an accountability mechanism between provider and citizen, There i s some information on the effectiveness o f HUMCs. The information indicates that HUMC members considered themselves as financial administrators and overseers and not as representatives o f the communities. Moreover, there were examples where HUMC members did not make the best decisions and revenues from user fees were not used for improving quality o f care by purchasing drugs and supplies bur rather on increasing staff salaries and incentives. In districts where HUMC members were provided training and there was community oversight, expenditures on staff were less. Overall, the reputation o f HUMCs within the health sector i s not very good, and HUMC members have been implicated in corruption relating to drug leakages and other abuses (Hutchinson, 1998). Technical efficiency of health care services in Uganda 5.50 Technical efficiency means obtaining the greatest possible production o f goods and services from available resources. Indicators such as staff/patient ratios, the number o f outpatient visits in facilities, the admission rate in hospitals, average length o f stay (ALOS) and the occupancy rate in hospitals are often used to gauge the technical efficiency o f health services delivery. Improving technical efficiency without compromising on quality i s a critical element o f helping countries make the best use o f available resources for health care. Data on technical efficiency o f health care facilities inUganda is very limited. 5.51 Some available information i s from a study o f eight PNFP and public hospitals. This study collected unit cost data on hospitals.Unit cost i s one approach to evaluating technical efficiency, although analysis o f unit costs should take into account the risk profile o f patients in order to be interpreted accurately. Interms o f unit costs, there was no clear pattern between public and PNFP hospitals: expenditure per case treated was quite similar across the different types of hospitals. However, according to this study, clinical workers in PNFP hospital see more cases than in public hospitals. However, in both PNFP and public hospitals, workload per clinical worker was quite low, indicating efficiency concerns. The hospital with the highest workload only saw six outpatients per day and had one admission per day. Inthe hospital with the lowest workload, only 1.4 outpatients were seenper day and only one admission occurred infive days. 98 5.52 The study also evaluated bed occupancy (the number o f occupiedbeds as a proportion o f the available bed days) and bed turnover (the number o f patients occupying each bed per year). The available data shows that occupancy rates among the hospitals ranged from only 25% to a maximum o f 60%. This would not, by any standards, be considered efficient. One o f the findings o f the study i s that although no clear pattems were found between the PNFP and public hospitals in terms o f efficiency, overall PNFP hospitals made more efficiency use o f staff but were clearly not efficient in relation to total expenditure or bed use. This indicates that there are opportunities for improving efficiency in both public and PNFP hospitals. Improving efficiency measures could be a key aspect o f contracts with PNFP hospitals (Ssengooba, 2002). D. CONCLUSION 5.53 Despite concerted efforts by the GoU, disparities in geographical access by region and district, urban and rural areas, and socioeconomic status persist. Inurban areas such as Kampala and Jinja, there may even be an oversupply o f facilities given the presence o f a large the private for-profit sector, while large parts of rural areas fall below the national average o f geographical access. This has important implications for the govemment's health policy in terms o f urban and rural areas. Given the problems with attracting and retaining health personnel with formal medical training inrural areas, more thought needs to be given to other mechanisms for making health care available to disadvantaged populations, such as recruiting and training people from the community to provide health services with good supervision and referral links (mobile health services and outreach, transportation, and access to telephones). Countries such as India and Mexico that also have a problem o f improving access o f rural populations to good quality basic health services have successfully usedthe community healthworker model. 5.54 Uganda, like other countries in SSA, i s facing various problems with the quality and quantity o f human resources. Monetary incentives for public sector staff i s one o f the issues, but it i s not clear that simply increasing public sector salaries i s likely to solve the problem. Available data shows that the PNFP and FP sector staff salaries are lower than inthe public sector and yet, perceived quality is better for the private sector (PNFP and FP). This means that there are other factors in private organizations that motivate better staff performance. This could be the altruistic motivation o f staff working in PNFP facilities or a combination o f factors such as better management, organizational characteristics, and career prospects as well as accountability arrangements. More effort needs to be focused on addressing the range o f factors that affect public sector health worker performance with a key focus on increasing accountability. 5.55 Improvementshave been made indrug supply and distribution through the public system. Nonetheless, health facilities (public and PNFP) continue to experience drug stockout for essential drugs (e.g. antibiotics). Overall drug stockout rates in PNFP facilities i s less (16%) as compared with public facilities (30%). This is most likely attributable to the fact that PNFP facilities have better cash flow (less dependent on public funds and more on donations, contributions and out-of-pocket payments) and the fact that PNFP can buy from the private sector. In contrast, public sector facilities first have to depend on the National Medical Stores (NMS) and as a second-line option, they can buy drugs from Joint Medical Stores (JSM) managed by the PNFP sector. Other factors are management capacity at the district andhealth facility levels on drug supply management (forecasting) and poor communication, information, and transportation systems. Accountability o f 99 public sector health workers i s also a problem contributing to the problem o f drug leakages and selling o fpublic sector drugs inthe private sector. 5.56 A general problem with regardto the public sector seems to be the lack o f accountability o f service providers to clients. Despite the GoU's objective o f introducing social accountability through decentralization, there i s mixed evidence on the functioning of community-based organizations vis-&vis district administrative structures. 5.57 The private sector (PNFP and for-profit) plays a very important role in the delivery of health services. Subsidies to the PNFP sector i s ong-standing in Uganda, and currently the government i s providing financial subsidies to PNFP facilities. The available data show that in religious PNFP facilities, the subsidies are used to provide better treatment for malaria and intestinal worms. This i s an indication that this subsidy program has prospects for helpingthe government expand access to quality basic health services. 5.58 With regard to the for-profit sector, the issues are slightly different. The first step for the government i s to further explore the role o f the private for-profit sector in providing health services, especially for the poor, and the implications o f this for the health o f the poor (i.e., quality o f care issues). The problem o f dual-practicing health workers also needs to be addressed. This type o f a study could help the government identifyregulatory issues as well the types o f policies that would be most useful in resolving problems arising from the public-private mix. 5.59 The experience with the abolition o f user fees i s still new. There are already emerging concerns with a worsening situation with drugs and reduced incentives for health sector staff since user fees were used to provide bonuses to staff. If these problems continue, utilization o f public facilities can be expected to decline even more in the medium term. In the absence of good accountability mechanisms, this could contribute to informal payments and reduced access for the poor andor greater emphasis on self-treatment and use o f low quality private providers and community drugvendors. 5.60 District-level disparities inhealth service indicators are significant. It i s important for the government to continue evaluating the reasons for these disparities (e.g., management capacity, funding) and to identify mechanisms to help improve the performance o f poor- performing districts. This could be done through a combination o f interventions, including providing direct technical support to districts, and usingthe PHC grant to link to performance measures). 5.61 There i s limited information in Uganda on technical efficiency o f public and private facilities. The very limited available data on the hospital sector raises technical efficiency concerns. Hospitals continue to absorb a significant amount o f total public spending on health care. As Uganda continues its efforts at restructuring and reorganizing the health services delivery network for improved accessibility and quality, it should also focus on technical efficiency and how to generate incentives for public and PNFP providers to be more efficient. 5.62 The implementation o f the HSSP provides an excellent opportunity for the MOH to exercise its stewardship functions in policymalung, priority setting, and monitoring and evaluation. Important progress has been made on all fronts. These efforts will have to be continued, especially to identify the most appropriate performance indicators, developing 100 information systems that allow traclung o f this information, sharing o f this information with key stakeholders, and identifying next steps. 101 6. HEALTHFINANCEAND SPENDING 6.1 This chapter examines financing and spending pattems inthe health sector. The chapter begins with an overview on sources and level o f health financing in Uganda and discusses the current health financing strategy. Two major sources o f financing in the health sector are analyzed: public sources including donor funds, and private out-of- pocket expenditure. For public expenditures, the chapter (i) assesses recent trends, levels, and pattems o f health expenditures; (ii)reviews the budget process and management; and (iii)evaluates equity ofhealthspending usingvarious measurements. For private spending on health, the chapter provides the latest information on household expenditure on health and analyzes changes and trends on household health spending between 1999/2000 and 2002/2003. Finally, the chapter summarizes the major findings and discusses some policy implications for financing and public expenditure management. A. OVERALLHEALTH FINANCE Sources of HealthFinancing 6.2 Three major sources finance health care in Uganda: private spending, govemment budget, and extemal aid. The previous chapter showed that the government and the private-not-for-profit sector (PNFP) are major providers o f health services in Uganda. However, private out-of-pocket spending (OOPE) has been the largest financing source for health care. Household survey data from the late 1990s to the early 2000s show that household out-of-pocket spending accounted for almost 60% o f the total health expenditure". Uganda stands out among sub-Saharan countries in having one o f the highest percentages o fprivate healthexpenditures". 6.3 Extemal aid has been an important source o f health financing since the mid 1990s and contributedmore than half o f the totalpublic health expenditure. Recently, donor funding appears to be declining because more donors are channeling their funds through the government budget and their contributions are inseparable fi-om the government budget. 6.4 During 1997/98 to 2000/01, the govemment spent about 6.5% o f its total budget on health. The health share o f the total government expenditure has increased to 9% in 2002/03, which financed 54% o f the total public expenditure on health in 2002/03 (MOFPED, 2003). 6.5 Figure 6.1 indicates changes in the overall financing pattern between 1997/8 and 2002/03. Despite increases in govemment spending on health and the removal o f user loCalculations based on Uganda National Household Survey 1998/99 and 2002/03. l1See Manju Rani and Agnes Soucat, background paper for WDR 2004 on global analysis of use of public andprivate healthproviders. 102 fees at public facilities in March 2001, the total OOPE on health increased significantly between the two time periods. The share o f donor funds in health financing decreased from 43.4% to 19.2%, while the government's share increased from 19.6% to 22.9% during1997198to 2002103. 6.6 As discussed in Chapter 1, Figure 6.1 Sourcesof HealthFinancing (YO) donor funding to health via projects has decreased with the introduction o f the mr 1997198 pa2002/03 SWAP. Therefore it i s not possible to separate donor funds from government funds to quantify the total donor support to the health sector. ...~~........... However, what i s clear from Figure 6.1 i s that overall public financing for health did IO not grow. out-of- pocket spending Householdfinances Govt. Donor Private almost 60% o f the total health care costs. Sources: 1997198 data from NHA and 2002103 data from household survey and MOHreports on govemment and donor expenditures. 6.7 Annual expenditure on health i s estimated as $17.1 per capita in200210312,o f which $9.9 i s out-of-pocket spending, $3.9 i s from the govemment and $3.3 from donorsI3. Despite significant increases inhealth spending by both public and private sources in last decade, per capita expenditure on health still remains very low, particularly public expenditure on health (Table 6.1). Table 6.1 Per capita expenditure on health inUganda Per capita expenditure on health care (US Dollars) Funding Source 1990 1997198 2001102 2002103 Public sources (Government and Donors) 3.00 6.93 8.05 7.2 Private sources (households and employers) 3.18 4.07 7.00 9.9 Total 6.18 11.oo 15.05 17.1 Sources: the data for 1990-2001102from MOHNational HealthAccount; the data for 2002103based on 2002103 survey data for household expenditure and MTFE for public sources. Health Financing Strategy 6.8 Options for increasing the low level o f investment inhealth has been an important focus o f discussion on many occasions in Uganda14. The Health Financing Strategy (HFS) developed by the govemment in 2002 estimated a requirement o f Ush 1415 billion ($28 per capita) to implement the Health Sector Strategy Plan (HSSP). Clearly there i s a huge 12 This number mayunderestimate the actual spending leveldue to lack of accuratedata ondonor spending and lack of data onNGO contributions. 13 Based on the exchange rate o f 1US$=1775.04 UgandanShillings as applicable inDecember 2002 (Midpoint o fthe survey) and the population of 24.7 million. l4 See the JMR Aide-memoirs in2001and 2002 and MOH's paper on "The Case for a Bigger Budget for the HealthCare". 103 financing gap between present funding levels and the requirement for implementing the HSSP (MOH, 2002a). 6.9 In preparation for the HFS, the government commissioned a number of studies on possible financing mechanisms15. The Harvard study on the feasibility for social health insurance concludes that social health insurance i s possible only as a long-term solution, but not a feasible option for the medium term. Based on the assessment o f the existing community health insurance schemes supported by the UK Department for International Development (DfID) in Uganda, the HFS considers that community health insurance schemes are neither feasible nor efficient as a main source o f health financing. The HFS considers donor project funding a feasible source, but not an efficient source, because a large proportion o f fimding i s spent on nonessential services. Contributions by NGOs or employers are also very limited. Household payments through user fees were found to inequitable and highly unpopular politically. The government has abolished this mechanism in the public sector in 2001. Overall, the HFS concludes that the most efficient and effective mechanism for Uganda at the moment i s public financing through government budget allocations. 6.10 The MOH has used the HFS to argue for increasing funding for health. The MOH budget had a relatively larger increase at the beginning o f FY 2002103 and later, suffered less from budget cuts16. However, the government has not provided sufficient budget to cover the income loss due to the removal o f user fees for health facilities, nor to meet the increasing demand for free services. The current HFS does not suggest additional mechanisms to offset resource loss or how to better use increased out o f pocket expenditure (the next chapter further elaborates this issue). B.MEDIUM TERMEXPENDITURE FRAMEWORK PUBLICFINANCING AND MECHANISMS 6.11 The GoU finances health care mainly through its budget allocation. The health budget includes financial disbursements to health institutions to meet worker salaries, operation and maintenance expenses, and development expenditures that include counterpart funds for donor projects. Since 1994/95, the government has used a medium term expenditure framework (MTEF) to outline budget allocations for a three-year term based on the country's development strategy. As a budget instrument, the MTEF aims to increase transparency and predictability in the budget process and strengthen linkages between policy priorities and budget allocations. As the government's spending plan, the MTEF links closely to the country's overarching poverty reduction strategy, the Poverty Eradication Action Plan (PEAP). The MTEF i s the financing plan for the PEAP. It provides an overview o f the resource availability in the medium term, based on projections o f government revenues and external aid to Uganda. 6.12 The government o f Uganda has succeeded in maintaining macroeconomic stability, especially controlling inflation since the early 1990s. However, overall government spending has been growing rapidly, from 22.4% o f GDP in2000101 to 25.3% in2001102. Expansion in government spending and the associated rise in the fiscal deficit has become a major concem (MOFPED, 2003). The government depends heavily on donor 15 Magezi ,Masiko and Wheeler "Future Viability of Community Health Insurance Schemes inUganda, IHSD 2002, and Bermand Hsiao, "Feasibility Study for Social Health Insurance inUganda," Harvard School o f Public Health, 2001. l6Based onreviewing of budget allocations by the MOFED. 104 funds to finance its budget. Any significant reduction in donor fundingwill undermine the public finances, particularly in education and health programs. The government has made efforts to integrate donor funding into the MTEF in order to improve the predictability and transparency of financing. Even through the current MTEF does not include all sources of public sector financing-particularly donor project financing-it has improved the predictability o f funding from both the government and donors to the health sector. 6.13 Health takes a relatively small share in the government spending program compared to public administration and security spending. Inthe context o f implementing the PEAP, health has gained a more prominent position in the budget allocation. The government increasingly recognizes the links between health and development including that between investment in health and poverty reduction and economic growth. Provided the overall size o f the total government expenditure program i s unlikely to increase in the medium term, the strategy proposed i s to shift spending from public administrationand security to health and other sectors that are critical to poverty reduction (MOFPED, 2003). Table 6.2 Sector shares of GOUexpenditure SECTOR 94/95 95/96 96/97 97/98 98/99 99/00 00/01 01/02 02/03 03/04 04/05 05/06 06/07 Outtum Outtum Outtum Outtum Outtum Outtum Outtum Outtum Outtum Draft Est. ProjectionProjection Projection Security 19.6 18.8 18.4 14.8 19.9 15.4 13.9 12.6 14.1 10.8 10.8 10.6 10.7 Roads and Works 4.4 4.3 6.8 4.9 6.2 8.1 8.5 8.3 7.3 11.1 12.5 12.4 12.3 Agriculture 2.6 1.5 1.4 1.1 1.0 1.5 1.5 2.2 2.3 3.3 3.2 3.3 3.3 Education 19.8 18.8 22.0 26.0 26.9 26.3 24.9 24.1 23.3 19.7 20.5 20.0 20.4 Health 8.0 9.9 7.2 6.5 6.5 6.5 7.4 8.6 9.0 12.1 11.4 11.5 11.7 Water 0.0 0.0 0.0 0.5 1.2 1.5 2.4 2.6 2.6 3.6 3.4 3.4 3.5 Lawand Order 8.9 9.7 8.7 8.9 7.2 7.3 6.5 6.7 6.9 5.3 5.2 4.9 5.1 Accountability 0.0 0.0 0.0 0.5 0.6 0.8 1.1 1.1 1.2 1.0 1.o 1.o 1.o EF&SS* 8.7 6.1 6.4 4.1 2.7 4.6 5.0 6.5 7.2 12.9 11.6 11.9 12.0 Public Administration 20.1 22.3 22.0 25.0 20.7 20.3 20.2 19.3 17.4 12.7 12.6 13.5 12.8 Interest Payments Due 8.0 8.6 7.0 7.6 7.1 7.7 8.5 8.1 8.6 7.5 7.8 7.4 7.2 All Sectors 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 * Economic Functions and Social Services Source: MOFPED 2003. c.PUBLICEXPENDITURE HEALTH ON Allocationswithinthe healthsector 6.14 Table 6.4 shows actual expenditure o f the health portion inthe MTEF from 1999/2000 to 2002/03. The overall health spending by the public sector increased by more than 50% between 1999/2000 and 2002/2003 with an increase o f more than 130% in recurrent expenditure. The recent recurrent expenditures also included some donor funds channeled through the budget. The amount i s difficult to identify at the sector level. Development expenditure accounts for more than 50% o f the total health expenditure in the MTEF, though its share decreased from 67% in 1999/2000 to 51% in 2002/03. The 105 total development expenditure for health was Ush 156.34 billion in 2002/03, 12% lower than the previous year. Table 6.3 HealthMTFEoutturns (inUshBillion) 199912000 2000101 2001102 2002103 Rec. Dev. Total Rec. Dev. Total Rec. Dev. Total Rec. Dev. Total Ministry of Health 12.6 111.1 123.6 23.8 100.3 124.0 31.0 137.6 168.6 31.3 109.6 140.9 Butabike Hospital 1.8 0.1 1.9 1.8 0.1 1.8 2.4 18.5 20.8 2.4 16.3 18.7 Mulago Hospital 11.0 20.1 31.1 11.8 10.1 22.0 18.3 9.6 27.9 17.3 3.1 20.5 HealthService Commission0.7 0.1 0.8 0.7 0.0 0.8 1.1 0.0 1.2 1.3 0.0 1.3 UgandaAIDSCo"ission0.5 0.1 0.6 1.0 0.1 1.1 1.2 0.7 1.8 1.2 19.7 20.9 NGOHealthUnits 3.3 0.0 3.3 5.7 0.0 5.7 11.6 0.0 11.6 16.1 0.0 16.1 PrimaryHealthCare 7.8 0.0 7.8 14.0 9.9 23.9 49.9 11.0 60.9 59.2 7.6 66.8 District Medical Services 6.5 0.0 6.5 5.8 0.0 5.8 8.9 0.0 8.9 8.6 0.0 8.6 Referral Hospitals 11.2 0.0 11.2 13.2 0.0 13.2 14.3 0.0 14.3 12.4 0.0 12.4 District Training Schools 1.9 0.0 1.9 1.3 0.0 1.3 0.0 0.0 0.0 0.0 0.0 0.0 District Lunch Allowance 7.7 0.0 7.7 9.2 0.0 9.2 0.0 0.0 0.0 0.0 0.0 0.0 Total 64.9 131.5 196.4 88.3 120.5 208.8 138.5 177.4 316.0 149.8 156.3 306.2 Sources: Ministryo f HealthAnnual performance reports 6.15 The MTEF explicitly includes external donor funds channeled through the project mode. The expenditures from donor projects are recorded under development expenditure, which accounted for 80% o f the development expenditures in the health sector. The section on development expenditure and donor financing provides more detailed information on donor expenditure. 106 RecurrentExpenditureby Level 6.16 The distribution o f recur- Figure6.2 Share of recurrent spending by level rent expenditures shows a clear trend o f increasing spending at the district level during the period o f 1999 to 2003. The share 100% o f the district budget 90% increased from 36.8% o f 80% 70% the total recurrent 60% expenditure in 199912000 50% to 45.9% in 2002103, due 40% to significant increases in 30% the Primary Health Care 20% Conditional Grant. 10% 0% Government subsidies to 99100 00101 2001102 2WU03 2003104 2004105 2005106 NGO health facilities BMOH "ational Hosp BReferral Hosp ODistrict PHC BNGOs increased from 5% o f the total recurrent expenditure Note: Ministryof Health 17 to 11%and are expected to grow to 14% in the next Source: the MTEF from MOFPED three years based on the MTEF projections. The budget speech in June 2004 indicated the allocationo f Shs 17.7 billion for NGO facilities in2004105. 6.17 The MOH headquarters and two national hospitals (Butabika and Mulago) consumed 36% o f the total recurrent expenditure in2002/03, down from 41% in 199912000. These patterns o f the expenditure reflected government's emphasis on basic health care delivery inrecentyears. The MTEFfor the nextthree years indicates a continuationofthe current spending pattem. The district level health budget will have a modest increase o f 3 - 4% while MOH's budget will grow by 8-10% duringthe next three years. Subsidies to NGO health facilities will increase by 14%, the fastest increase among spending institutions'8. Recurrent expenditure by economic category 6.18 Wage and personal emoluments including lunch allowance account for about 40% o f the total recurrent expenditures in the health sector, which i s not high compared to many other developing nations where personal emoluments account for the majority of the recurrent cost. The share o f wage expenditure i s particularly low at the M O H headquarters, due to the inclusion o f centrally procured drugs and medical supplies inthe nonwage expenditure. 17MinistryofHealth (MOH) includes healthservice commission and Uganda AIDS Commission; National Hosp includes Butabike Hospital and Mulago Hospital; and Dist. .PHC includes primary health care grant, district hospitals and medical services as well as training schools in 199912000. 18Based onthe calculation fromthe MTEF. 107 Table 6.4 Recurrent expenditure by level: 1999/00-2002/03 1999l2000 OOIO1 01/02 02/03 03/04 04/05 05/06 Ministryof Health 13.8 25.5 33.2 33.8 35.1 38.1 42.0 Wage 3.0 3.8 3.9 3.4 4.5 4.5 4.5 Non-wage 10.8 21.7 29.3 30.4 30.6 33.6 37.5 % ofnon-wage 78.5 85.3 88.2 90.0 87.2 88.2 89.2 National Hospitals 12.8 13.6 20.7 19.7 20.3 20.8 22.2 Wage 5.8 6.5 7.9 8.2 9.0 9.0 9.0 Non-wage 7.0 7.1 12.8 11.6 11.3 11.8 13.2 % of non-wage 54.7 47.6 61.9 58.6 44.3 43.3 40.5 Referral Hospital 11.2 13.2 14.3 12.4 16.2 16.2 16.3 Wage 7.2 8.2 8.9 7.1 11.0 10.8 10.8 Non-wage 4.0 5.0 5.4 5.3 5.3 5.5 5.6 % ofnon-wage 35.6 38.1 37.9 42.4 32.4 33.8 34.1 District Health Facilities 23.9 30.3 58.8 67.8 77.5 81.0 84.0 Wage 10.5 13.8 35.0 40.1 45.3 45.4 45.4 Non-wage 13.3 16.5 23.7 27.7 32.2 35.6 38.6 % of non-wage 55.9 54.3 40.4 40.8 41.5 44.0 45.9 NGO Health Facilities 3.3 5.7 11.6 16.1 20.0 23.4 26.6 Total 64.9 88.3 138.5 149.8 169.0 179.5 191.1 Wage 26.49 32.22 46.84 80.92 69.72 69.61 69.68 Non-wage 38.42 56.06 91.71 99.65 99.24 109.85 121.37 % of non-wage 59.2 63.5 66.2 55.2 58.7 61.2 63.5 Source: Ministry of Finance,PlanningandEconomicDevelopment, 2003 Healthpersonnelpayroll 6.19 The health sector used to have three payrolls: the public health care payroll, delegated payroll, and local govemment payroll. Under that system, the MOH, local government and individual health facilities all could recruit and pay health workers. None o f these payrolls captured the complete information on total staffing (Lindelow, Reinikka and Svensson, 2003). The situation made managing the wage bills and determiningthe size o f health work force very difficult. Since July 2001, the Government has re-centralized the health worker payrolls with the exception o f referral hospitals, which continue to pay their health workers throughthe respective hospitals payrolls. 108 6.20 In past five years, the wage Figure 6.3 Share of the total government wage bill bill o f the health sector has increased. The health sector's share o f the total government wage bill more than doubled between 1997198 and 2002103. It will stabilize at around 10% o f the total wage bill in next three years according to the MTEF's projection (Figure 6.3). The discussion in Chapter 5 indicates human resources as a critical factor in health system performance. Meeting 0.04 human resource needs 199718 199819 1999100 2000101 2001102 2002103 2003104 2004105 2005106 requires more spending. Source:Ministry of Public Service However, raising the wage bill is constrainedby the sector ceilings and the total resources available to the sector. At present, all sectors budgets are set according to the sector ceilings. The social sector wage bill has increased faster than the total wage bill. In FY 2003/04, the government increased salary and lunch allowance for the health workers. However, increasing salaries may not necessarily result in improvement in productivity. An incentive system for better performance and accountability needs to be established to better use the increased finding for humanresources. 6.21 The MOH now maintains a Figure 6.4 Wage Distribution,2002/03 health worker personnel database for planning the wage bill and allocating / 25000 conditional grants to the districts. The wage distri- bution according to salary scale reflects the mal- distribution o f health per- sonnel and a distorted ratio among the various health workers. 6.22 Under the MOH payroll, there are 20,493 health UI U l U3 U4 USA U5B USC U6 U7 U8 USS workers. The salary scale Wage Scale ranges from level U1 (average annual salary Ush 14 million) to level USS Source: Ministry of Public Services, 2002/03 (annual salary Ush 847,298). The majority of health workers are at lower technical levels, therefore lower salary scales. About 35% o f the health workers are at the lowest salary scale (Level USS) and 48% are at the level U6-U8, with very limitedprofessional training (Figure 6.4). 109 6.23 The lack o f qualified staff available in the labor market and a slow and complicated recruitmentprocess makes it difficult to improve the health personnel profile. Because o f this, the health sector suffers from a severe staffing shortage and the MOHwas unable to spend its wage bill allocation o f Ush63.8 billion in2002/03. Nonwage expenditure 6.24 The nonwage expenditure in the health sector accounts for more than 60% o f the total recurrent cost and includes expenses on drugs, health supplies such as condoms and syringes, maintenance o f buildings and vehicles, and administrative costs. Whether nonwage spending i s adequate and balanced i s very critical for service provision. However, an adequate assessment o f i s not possible due to lack o f information. Expenditures on nonwage items such as those noted above are not clearly reported and recorded in health facilities. One o f the major weaknesses o f the Management and Information System inthe MOH i s absence o f any information on expenditures and costs. 6.25 Pharmaceutical expense usually i s the second largest expenditure item in the health sector, after personnel cost. However, there are no consolidated data showing how much the public sector spends on pharmaceuticals in Uganda. There are three main sources that finance drugs and health supplies in the public sector: MOH budget, PHC Conditional Grants and donor finds. The expenditures on drugs are not adequately reported and consolidated, which makes analysis on drug expenditure at the aggregate level difficult. Underfinancing o f drugs has been identified as a major problem by the health sector reviewsIg. The situation improved slightly in 2002/03. The MOH budget finances centrally-procured drugs for various levels o f care, which only accounts for a part o f the drug expenditure. On average, districts spent 66% o f the PHC grants on drugs, with a wide range o f variations, from 11% in Pader District to 205% in Moyo district. According to the guidelines for spending the PHC grants, 50% should be spent on drugs and health supplies. A Drug Tracking study indicates that not all health facilities follow the guidelines and a substantial number o f facilities did not reach the 50% target in 2001/02 (Business Synergies 2002). Donors finance a large proportion o f drugs and health supplies, in the form o f in-kindcontributions such as vaccines or in momentary forms. However, no record on the total expenditure on drugs by donors i s available. 6.26 In2002/03, the government establishedaBasket ofEssentialDrugsAccount for boththe government and donor funds. The estimated per capita funding for drugs and health supplies for the essential package increased from US$0.86 in 2001/02 to US$1.2 in 2002/03, though, it still falls far short o f what i s required for financing the basic package o f essential services which i s US$3.5, excluding ARVs and pentavalent vaccines (MOH, 2003b). Fromrecent Joint Review Mission Aide-memoirs, MOH 110 6.27 The funds for essential drugs at Figure6.5 Healthsub-district drugbudget the district level have increased recently. The government with Danish Intemational aPHC Drug Budget Development Agency support OCreidt Lines has created credit lines earmarked for essential drug k i t s for districts. The PHC grant, another main source o f funding for drugs, has also been ......................... increased (Figure 6.5). From 2001/02 to 2003/04, the overall drugbudget more than doubled. However, inreal terms, the drug budget for districts still remains 2001/02 2002/03 2003/04 very low, $0.35 per capita Source: MOH, 2003b (MOH, 2003b) and i s insufficient to meet increasing demand for care. The public's use o f health facilities i s closely related to the availability o f drugs. Therefore, adequate financing o f drugexpenditure i s critical to the improvement o f health service delivery. Government grants to the private not-for-profit health facilities 6.28 The private-not-for-profit sector has a long history in providing health care services in Uganda. Both the NHP and HSSP provide a strategic framework to increase the public sector's collaboration with the private sector, particularly the PNFP sector. PNFP facilities are allocated throughout the country, particularly inrural areas. Inmany places, PNFP i s the only health provider. The PNFP funds its services through donations and income-generating activities but mainly from user fees. Since 199798, the government resumed grants to the PNFP, which had been stopped in the 1970~~'.The grants to the PNFP increased from 1 billion shillings in 1997/98 to 17 billion shillings in 2003/04, about 11% o f the total recurrent budget. Furthermore, the grants to the PNFP sector are allocated under the PHC conditional grant, which i s protected from unplanned budget cuts. The objectives o f the govemment grants are to enable the PNFP sector improve and maintain good quality services, and to reduce or keep user fees low in order to improve access by the poor. The use o f the grants was confirmed in one study which found that the grants led to more laboratory tests and lower prices for outpatient services (Reinikka and Svensson, 2002). 6.29 After abolishing user fees in public facilities, the government also encouraged PNFP facilities to reduce their service charges. A recent study has shown that almost 70% o f PNFPhospitals have reduced fees and most o f them have implementeda flat fee structure (John and Maniple, 2003). The reduction o f fees is mainly towards services more likely to benefit vulnerable groups such as mothers and children. Services with large externalities such as TB treatment have been offered free o f charge in most o f the PNFP hospitals, because the govemment has provided subsidies for such services (there i s more discussion on the impact o f fee reduction inthe next chapter). 2oBased on the information from draft health sector PEAP ReviewPaper, Ministry of HealthAugust 2003. 111 D. DEVELOPMENTEXPENDITURE DONOR AND FINANCE 6.30 As noted earlier, development expenditure accounts for more than half o f the total health expenditure in the MTEF (see Table 6.3). It finances construction and renovation o f buildings, medical and office equipment, communication facilities, ambulance services, and other transport requirements with the objective o f improving accessibility o f health facilities within 5 kmfrom 49% in 1999 to 80% by 2005 (MOH 2003b). 6.3 1 External donor funding Figure6.6 Developmentexpenditure constitutes a large part o f development expenditure as shown in Figure 6.6. Out o f the total development expen- 200 diture o f Ush 156.34 billion 180 in 2002/03, 75% was funded 160 ............... by donor projects. That i s what has beenrecorded inthe MTEF and it does not -3 e 140 .r 120 HDonor include donor funds made E, 0Local 100 through central budgetary -2 support. There are also 80 expenditures financed by 60 donors that have not been captured by the MTEF. 40 Therefore, the actual spend- 20 ing could be higher than that 0 presented in Table 6.6. 1999/2000 2000/01 2OOVO2 2002/03 However, donor projects also Source: Ministry of Health, Annual review reports finance recurrent cost, which was not clearly separated under the MTEF.Hence, the total development expenditure could be lower than what showed inTable 6.6. 6.32 Inrecent years the sector has been focused on construction and rehabilitation of health centers at the district level. Development expenditure allocated at the MOH level, particularly under donor financing, i s for health center IVs, 111s and 11s. Rehabilitation o f two nationalhospitals i s also mainly financed by donor projects. Table6.5 Developmentexpenditure, 1999-2003 (Ushbillion) 199912000 2000/2001 200112002 2002/2003 Local Donor Total Local Donor Total Local Donor Total Local Donor Total Ministryof Health 11.46 99.62 111.08 7.39 92.89 100.28 14.92 122.72 137.64 26.98 82.6 109.58 Butabike Hospital 0.12 0.12 0.05 0.05 1.54 16.92 18.46 1.75 14.55 16.3 Mulago Hospital 5.07 15.05 20.12 1.73 8.41 10.14 5.18 4.44 9.62 1.58 1.56 3.14 Health Service Commission 0.08 0.08 0.04 0.04 0.03 0.03 0.04 0.04 UgandaAIDS Commission 0.08 0.08 0.09 0.09 0.68 0.68 1.2 18.5 19.7 Primary Health Care 0 9.94 9.94 10.98 10.98 7.58 7.58 Total 16.81 114.67 131.48 19.24 101.3 120.54 33.33 144.08 177.41 39.13 117.21 156.34 Source: Ministryo f Health 112 6.33 Based on a study o f expenditure o f five major donor projects21, the MOH reported that 68% o f Ush 66 billion donor project funds were spent as project overheads and technical assistance, only 10% on drugs and supplies, 9% on human resources and other recurrent costs, and 13% on development expenditures directly related to the support o f HSSP. Therefore, the government Health Financing Strategy concludes that donor funding via projects i s inefficient. 6.34 Although donor project fundingdecreased in 2002/2003 and more donors have channeled their funds through budgetary support, more resources will be provided through different international initiatives such as the Global Fund for HIV/AIDS, Malaria and TB, and the Global Vaccine Initiative. The government has submitted three proposals to the Global Fundandthe total approved funds for nextthree years is $97.68 million(MOH2003b). E.BUDGETPROCESSANDMANAGEMENT 6.35 This section reviews aspects ofbudget performance inthe health sector from two aspects: (i) budgeting and priority setting and (ii) monitoring and evaluation o fbudget execution. 6.36 Budgeting in Uganda is a participatory process involving various stakeholders such as government line ministries, local governments, donors, and civil society organizations. Figure 5.7 indicates the budgeting process, which usually begins with the annual consultative budget workshop in the second quarter o f each fiscal year (October to December) and concludes with the budget speech at the end o f fiscal year (June). 6.37 Sector working groups (SWGs), which consist o f line ministries, local governments, and other government agencies, play an important role in the budget process. They prepare budget framework papers based on the budget ceilings indicated by the MOFPED. Based on the sector budget framework papers, the MOFPED completes the national budget framework paper, updates the MTEF, and submits it to the cabinet for approval. Before the budget i s finalized, there i s consultation process between the MOFPED and various budgetingentities. 6.38 Inrecent years, the MOH has actively participatedinthe budget process and used its budget framework paper and the Health Financing Strategy paper to demonstrate that the health sector i s underfunded and emphasize the important role that the health sector plays inpoverty reduction. As the government is moving to the Result OrientedManagement and Output Oriented Budgeting system, budgeting i s linked to results and performance. In its annual performance reports as well as at the PEAP review and by-annual sector review meetings, the health sector has successfully argued for increased budget allocations by showing that this has resulted in more health services and higher immunization coverage (See Chapter 5). '' The five donors are United States Agency for International Development, Swedish International Development Agency; German Technical CooperationAgency, UK Department for International Development and Danish International Development Agency. 113 Figure 6.7 Interactions and Timing of the Budget Process The NationalBudgetProcess Cabinet Approval Cabinet o f BFPIMTEF Approval -Indicative Sector Ceilings MFPED I Consultations I -----___-------____ Line Ministried SpendingAgencies/ MTEF Allocations SWGs/Donors within the Ceiling Source: Beynon, J (2003). 6.39 The Health Sector working group consists o f the Ministry staff from various departments, and representatives from donors, the PNFP sector, and districts. The working group drafts the Budget Framework Paper (BFP), then circulates it to ministry departments, agencies, and donor community for comments. The preparation for the BFP becomes a more organized consultative process. In the BFP, the health sector has to show its real commitment to the sector priorities and often has to make hard choices in budget allocation. BudgetSetting 6.40 The MOH has estimated the total cost for implementing the Health Sector Strategy Plan (2000/01-2004/05), which requires a total of US$ 954 million. The current fundinginthe health sector is about 21% of the requirement (Ministry o f Health, 2000). However, the real financing gap i s unclear. The health sector i s one o f the sectors that rely heavily on external aid. Duringthe late 90s, external donor hnds accounted for more than half of the total public health expenditure. At that time, donor support to the health sector was mainly through donor projects and in-kind supplies such as drugs and technical assistance. Such donor expenditures are poorly recorded in the health system so the total amounts are unclear. The attempt made by the MOH to capture donor expenditures inthe 114 health sector has not been very successful. Therefore, the total resource envelope for the health sector i s incomplete. 6.41 The SWAP process and the improvement o f budget management increasingly encourage donors to channel their funds through budget support. The World Bank provides almost half o f its assistance to Uganda's general budget, and United Kingdom's Department for International Development (DflD) has moved almost exclusively to budget support (World Bank, 2003). Since 2000/0 1, the government health budget including donor budgetary support has exceeded total donor project funds. To improve budget predictability and transparency, the government has called upon donors for budgetary support. Although it i s not possible for all donors to put funds into the general budget, the GoU i s making an effort to include all external funding into the MTEF in next few years. Starting in 2004/05, the MOFPED will impose sector ceilings, which cover both the GoU budget and the extemally funded projects in order to exert better control over the growth o f the government expenditure (MOFPED, 2003). Fundingprioritiesinthe sector 6.42 The MOH's budget framework paper in 2002/03 outlined sector priority areas for public financing. Reducing infant and maternal mortality has been identified as the top priority inthe healthsector inpast two years. A major effort has beenmade to improve coverage o f immunization and emergency obstetric care. Funding was earmarked for equipment, transport, and drugs for HC I V s to be able to provide such services. 6.43 The government has achieved considerable improvement in budget planning, allocation, and managing aggregate spending levels to be consistent with the PEAP and the sector strategy outlined in the HSSP. However, the sector's institutional arrangements and incentives have not succeeded in ensuring that budgeted expenditures translated into actual service delivery, particularly in the 1990s. The weak link between spending and service delivery may be partly reflectedby the fact that health outcomes stagnated despite increasing public spending inthe 1990s. MonitoringandEvaluationof BudgetExecution 6.44 District health management teams report to the MOH on a quarterly basis in a standard format regarding its performance. Although the reporting on health service provision and health statistics has improved greatly since 2000, there i s still no reporting that includes financial information in the MOH's M I S system." Routine monitoring and evaluation of expenditure and budget execution in the health sector is still weak. The MOH has depended on other means to monitor the expenditure information. TrackingStudies 6.45 Expenditure tracking surveys were conducted to trace public spending inthe social sector inUgandaduringthe mid-1990s andshowedthat transfers offunds from the central level to the frontline providers suffered from leakage (Reinikka, 2001, Reinikka and Svensson. 2001). The results from the early tracking study on education have raised awareness o f such leakages among the public as well as policymakers and measures were taken to change the situation. A later survey indicated a drastic improvement in spending funds 22Based onthe discussionwith MOHofficials andreview o f MISdatabase. 115 for the intended purposes (Ablo and Reinikka, 1998, Reinikka, 2001). The expenditure tracking surveys have been used as a powerful tool for monitoring and evaluating budget execution inthe health sector. The following are three examples. 6.46 To better understand the tremendous delays and usage inefficiencies o f the Primary Health Care Conditional Grants, the MOH conducted a tracking study to trace the flow o f funds from the PHC grants in 2001. The study results indicated that the problems o f underutilization o f PHC grants were mainly due to delays in submission o f funding request from districts to the MOH headquarters (at least half o f the districts studied) and delays in releasing funds from MOFPED. The study identified bottlenecks in management and the use o f PHC grants and made recommendations for improvement (Horizons International ltd. 2001). The following year, the flow o f funds in the health sector was substantially improved23. 6.47 In2002, the MOHcommissioned the DrugTracking Study, to assess the use and impact o f the substantially increased financial support and allocations o f drugs to the district health system. The study results showed that spending on drugs in many districts was below the target level set by the PHC guidelines. The most important factor for drug availability in the district health system was the irregularity o f funding from the PHC grants. Stockouts were more likely to be associated with irregular and delayed release o f funding. There were also delays in drug procurement and distribution due to poor management. The study showed that patient attendance increased with the availability o f drugs. Health facilities that had better procurement systems and better funding performed better. The study provided evidence to support separating procurement from cash supply24and supported the establishment o f credit facilities or a prepayment system at National Medical Stores to improve drug availability. An improvement in managing funds for drugs was reportedat the sector review conducted inlate 2003 (MOH 2003b). 6.48 The MOH also commissioned a funds tracking study on the National Service Delivery Program in2003/04. These programs were established in2000/0 1 as areas requiring PAF funds for district PHC activities but which also required management from the central level. The study found that delays were experienced at the MOH due to the long approval process and to lack o f accounting by the subcomponents for earlier advances. Despite guidelines to the contrary, funding advances were made although previous advances had not been accounted for. It identified a number o f weaknesses in the registration, reporting, and accountability mechanisms innearly all the subcomponents so that it was not possible to see whether activities and purchases were carried out as planned. Monitoring on the use o f funds was often left to MOH internal auditors who are financially oriented and who didnot carry out technical monitoring or supervision. When technical monitoring was carried out, it was handled by junior staff with very little involvement o f management staff. Non-facility NGOs (the Uganda Protestant Bureau, Uganda Catholic Bureau, Family Planning Association o f Uganda) were found to have the best accountability routines with all o f the transactions sampled, showing 100% on implementation and accountabilityreturns. 23 Reportsmade at the Joint Review Mission in2002. 24 See Consultantreport: Business Synergies "Drug Tracking Study-FinalReport" August 22,2002. 116 F.HOUSEHOLD EXPENDITURE HEALTH ON 6.49 The previous section suggests that households spend a significant amount o f their resources on health care and this accounts for almost 60% o f the total health financing in Uganda. Overall, health expenditure as a percentage o f total household expenditure did not change much between 1999 and 2002. Households spent on average 4.4% o f total household expenditure on health in 2002, slightly higher than they did (3.3) in 1999. Better-off households spent more on health care than did poor households. Health expenditure varies from 3.8% o f total household expenditure in the poorest quintile to 4.9% in the richest quintile. Rural residents spent a higher proportion o f the total household expenditure on health than urbanpopulation. Out-of-pocket-expenditure (OOPE) 6.50 Out-of-pocket expenditure (OOPE) on health has shown an increase in the past decade. The main change i s due to increased spending by urban residents. In general, the urban population spent almost twice as much as rural residents on health care. A rural-urban gap i s increasing as shown in (Figure 6.8). 6.51 Prior to abolishing user Figure 6.8 Monthly out-of-pocket per capita expenditure fees in 1999/00, on health inthe past decade Ugandan households spent on average Ush 15548.2 (about $10.7*') 140'0 T----......................................... -Total per capita per year on - - -X - -.Urban health care including consultation, medicines and hospitalization. ~ 0 0 . 0 i .................................................. / .x---- Aggregate average household expenditure on health increased to Ush 17606 in 2002/03 after user fees were x . . -.. x - - - . . x - abolished. After 40.0 ........................................................... II t adjusting for inflation, average per capita ex- 20.0 ............................. ___._._. ........-.:---:.-. -+---- .e. penditure at 1999 prices increased slightly to Ush 0.0 ! 16244.2 (or $9.926), 93/94 94/95 95/96 97 99/00 02/03 However, the aggregate Sources: Uganda National HouseholdSurvey 93-03 data hides important changes inthe distributionalpattems o f the expenditure (Table 6.6). ~~~ 25At the 1999exchange rate ofUS$1=1454.83 Ugandan shilling. 26At the December 2002exchange rate ofUS$1=1775.04 Ugandanshilling. 117 6.52 Poor people (the poorest Figure 6.9 Out-of-pocket health spending per capita two quintiles) have reduced their OOPE by almost 50% while the rich (the highest 1 El1999100 quintile) have almost 30 25 2002/03 ...........~...........................~. doubled their health spending. The .reductiono f OOPE while use o f services by the poor increased could due to a number of reasons: (i) increased access to free services from public 5 provider; (ii)reduction o f 0 fees by PNFP providers; Poorest 2nd Middle 4th Richest and (iii)using more low- 20% Expenditurequintile 20% cost private providers such Sources: UgandaHouseholdExpenditure Survey, 1999 and 200212003 as traditional doctors, and (iv) in-kind payment which was not captured by the household survey. The links between OOPE and utilization i s discussed indetail inChapter 7. 6.53 Most o f the OOPE in both 1998 and 2002103 was for medicines and hospital/clinic charges. The highest reduction in spending for the poor was in hospital/clinic charges (60%) followed by a reduction in drug expenditure (40%). The rich have considerably increased their spending on medicines, hospital and clinic charges, and traditional doctors, which may due to a shift to private facilities. 6.54 Compared with before user fees were abolished, the poor now spend much less than the rich for accessing health care. However, one also needs to be cautious indrawing further conclusions, because the household expenditure on health was surveyed at the household level. Different members may go to different types o f health providers. It i s not possible to separate the expenditure data into on public facilities and private facilities. Therefore, it is not possible to assess whetherinformal payments are beingmade at public facilities. 118 Table 6.6 Annual per capita householdexpenditureon health care in 1999 and 2002/03 (at fixed pricesinUgandan Shillingsin 1999) Consultation Hospital/clinic Traditional Variable fees Medicines charges doctors Others Total Year 1999 Expenditurequintiles Poorest 20% 177.5 3364.0 3094.8 111.6 10.5 6758.4 2nd quintile 470.9 3674.7 6013.3 262.4 29.3 10450.6 Middle 255.9 4360.2 6640.1 326.5 14.5 11597.1 4th quintile 314.6 5705.4 8394.7 644.0 218.3 15277.0 Richest 20% 2057.2 8368.6 11470.6 1178.8 210.7 23285.9 Residence Rural 467.0 5117.5 7664.3 718.0 107.5 14074.2 Urban 3 194.0 9283.6 10479.5 256.5 227.0 23440.6 Average 896.1 5773.1 8107.3 645.4 126.3 15548.2 Year 2002103 Expenditurequintiles Poorest 20% 183.8 2020.9 1246.8 92.7 40.7 3585.0 2nd quintile 276.3 2949.8 2127.1 181.3 13.1 5547.6 Middle 624.7 5227.5 3615.2 394.7 36.0 9898.0 4th quintile 785.7 7405.8 7339.7 394.5 166.0 16091.7 Richest 20% 2490.6 19510.5 9188.0 4868.8 67.5 46125.4 Residence Rural 849.1 6550.1 5637.0 1284.3 60.0 14380.6 Urban 982.7 11646.8 1866.5 707.3 87.4 25290.8 Average 871.9 7420.7 6701.2 1185.8 64.7 16244.2 Percent changebetween 1999 and 2002103 Expenditurequintiles Poorest 20% 3.6 -39.9 -59.7 -16.9 286.2 -47.0 2nd quintile -41.3 -19.7 -64.6 -30.9 -55.4 -46.9 Middle 144.1 19.9 -45.6 20.9 149.0 -14.7 4th quintile 149.7 29.8 -12.6 -38.7 -24.0 5.3 Richest 20% 21.1 133.1 67.3 313.0 -67.9 98.1 Residence Rural 81.8 28.0 -26.5 78.9 -44.2 2.2 Urban -69.2 25.5 13.2 175.8 -61.5 7.9 Average -2.7 28.5 -17.3 83.7 -48.8 4.5 Sources: UgandaHouseholdExpenditure Survey, 1999 and 200212003 HouseholdExpenditureon Drugs 6.55 More than 45% o f health spending i s on medicines. For people who were sick and sought care inpublic facilities, 65% o f them obtained drugs for free. For those who used PNFP providers, only 8% received free drugs and about 80% bought drugs. The PFP providers charge for almost all the drugs. 6.56 The good news is that government's free drugs have reached more poor people. About 73% o f people from the poorest quintile obtained free drugs inpublic facilities compared 119 to 46% from the richest quintile. However, the bad news i s that there were clearly shortages o f drugs in public health facilities and not all the drugs are free in the public health facilities. On average, more than 20% o f the population who used public facilities purchased some drugs and 12% paid for all their drugs. Combining purchases in the public and private subsectors, one finds that overall, more than 50% o f the people from the poorest quintilepay for drugs. Table 6.7 Distribution of Paymentfor the drugs amongrespondentswho fell sick inthe last 30 days and sought care from a provider in 2002/03 Poorest 20% 2ndpoorest Middle 2ndrichest Richest20% Average To any provider No drugs required 0.5 0.8 0.4 0.5 0.6 0.6 Obtainedfree 32.3 25.4 21.0 17.3 10.3 20.5 Some drugs purchased 11.8 8.1 8.5 9.9 7.5 9.0 All drugs purchased 55.4 65.8 70.1 72.4 81.7 70.0 Among respondentswho visited apublic sector provider No drugs required 1.1 1.2 0.6 1.1 1.7 1.1 Obtainedfree 72.7 69.9 67.1 63.4 46.3 65.3 Some drugs purchased 19.0 16.8 20.9 24.7 25.7 20.9 All drugspurchased 7.2 12.1 11.4 10.8 26.4 12.6 Among respondentswho visiteda NGO/religiousprovider No drugs required 0.7 0 0 0 0 0.1 Obtainedfree 9.2 9.0 9.0 8.4 7.1 8.3 Some drugs purchased 21.4 6.8 17.4 14.8 5.7 12.4 All drugspurchased 68.8 84.2 73.6 76.9 87.2 79.2 Among respondentswho visitedaprivateprovider No drugs required 0 0.3 0 0.1 0.3 0.2 Obtainedfree 0.5 0.1 0.6 0.5 0.8 0.5 Some drugs purchased 4.5 3.1 2.2 4.1 2.9 3.3 All drugspurchased 95.0 96.5 97.2 95.3 96.1 96.0 Source: UgandaNational Household Survey 2002/03 6.57 The household survey data also show regional variation interms o f the availability o f free drugs. People from the Central region (13%) are least likely to obtain free drugs (13%) while those in the Eastern region are most likely to obtain free drugs (32%). Overall, only 20% o f the population received free drugs. Some evidence i s available indicating that the use o f community drug vendors i s common because o f their accessibility, availability o f drugs, and credit services (Twebaze 2001). 6.58 Table 6.11provides information on the payment o f drugs by type o f illness. Except for AIDS treatment, where 44% obtained free drugs, drugs for most o f illness have to be bought. For nearly all illnesses, well over half paid for all their drugs when ill.Such information clearly indicates that treatment o f the majority o f illness inUganda i s not free and the provision of free drugs i s still very limited. 120 Table 6.8 Paymentof drugsby type ofillness Obtained Purchased Purchasedall Sufferedfrom No drugs needed free some drugs drugs Total Malaria 0.37 18.52 7.97 73.15 100 Respiratory 0.3 1 23.37 7.42 68.9 100 Measles 0.04 16.89 12.88 70.19 100 Diarrhoea 0.37 26.17 8.1 65.36 100 AIDS 0 43.61 32.19 24.2 100 Pregnancyrelated 3.78 18.81 23.97 53.43 100 Dental 1.05 18.09 19.52 61.34 100 Accident 0.75 15.1 12.38 71.77 100 Intestinal 0.19 22.35 10.79 66.66 100 Skin 0.91 31.03 11.55 56.51 100 Hypertension 0.18 16.87 13.03 69.92 100 Ulcers 2.5 17.1 6.42 73.97 100 Mental 0.28 11.74 14.31 73.68 100 Other fevers 1.17 21.42 8.99 68.42 100 Others 1.29 24.49 9.79 64.42 100 Total 0.55 20.46 9.03 69.96 100 Source: Uganda National Household Survey 2002103 G.DISCUSSION RECOMMENDATIONS AND 6.59 Financing health care in Uganda i s still at a very low level. The most challenging question i s how to mobilize more financial resources for health. The MOH needs to continue to pursue bigger budgets for health to further improve both coverage and quality o f service provision. The free health services from public providers has reached more poor population but its coverage i s very limited. To increase provision o f free services and improve its quality, the government has to invest more in the health sector. The success o f the abolition o f user fees will depend on whether the government could increase resources to the health sector to compensate the loss o f income and to meet increasing demand. 6.60 Furthermore, based on the MTEFprojection, the government budget on health i s unlikely to grow substantially in the medium term. In order to improve financing in the health sector, the government has to explore other financing mechanisms and continue to improve efficiency in the use o f scarce resources. The current health financing strategy needs further options for increasing resources. 6.61 Despite the abolition of user fees at public health facilities, the out-of-pocket expenditure i s the largest source of financing for health in Uganda. Even though the financial burden o f health care has been increasingly borne by better-off households, the user fees are neither equitable nor efficient to finance health care. The government has to foster risk- pooling mechanisms in health financing and encourage the growth o f health insurance. Secondly, donor funding will play increasingly a dominant role in financing health care under the various internationalinitiatives inthe near future. The government has to work with donor agencies to improve efficiency and effectiveness o f donor financing. Monitoring donor funding i s critical, especially for donor funds that are not channeled throughthe regular budgetary framework. 121 6.62 Monitoring links between expenditures and outputs and outcomes i s a critical step for the government. Allocation o f public funds has improved in the health sector. More resources are channeled towards primary health care at the district level through the PHC conditional grants. However, the link between increased resources and service delivery i s still weak. The public uses health services more but little i s know about the quality o f services and eventual health outcomes. The government needs to strengthen mechanisms for accountability and to target funding so that it improves service quality and outcomes. 6.63 Budget planning and management processes have improved under the MTEFand SWAP process. However, monitoring and evaluating expenditures i s still very weak. The MOH needs to improve information systems on health spending, particularly bookkeeping and reporting on expenditures at the each level. Improvements in transparency o f expenditures are needed, especially nonsalary recurrent expenditures and program expenditures. Better balance and coordination between the development and recurrent budgets i s needed. 6.64 Provision o f health services by the private sector both for-profit and not-for-profit has been increasing in Uganda. The government needs to strengthen its regulatory role to ensure the population i s getting value for money. The government also needs to use various measures such as performance-based contracting to enable the private sector both-for-profit and not-for-profit-to contribute to public policy objectives. 122 7. IMPACT OFTHE REMOVAL OFUSERFEES 7.1 One o f the recent major health policy changes in Uganda was the removal o f user fees from government health facilities inMarch2001. This policy change has had a profound impact on both the supply and use o f services, particularly for the poor. This chapter builds upon previous chapters on utilization o f health services and service provision and provides an in-depth assessment o f the connections between demand and supply o f health services with and without fees. A. EVOLUTION USERFEEPOLICY INUGANDA OF 7.2 In the 1960s, health care inUganda was free at point of use. However, with economic deterioration and the political and social turmoil experienced inthe 1970s and 8Os, health services became underfunded and less accessible. Health staff was under- and irregularly paid (Kipp et al., 1991), which together with increasing corruption contributed to the practice o f informal charges among public health providers. 7.3 Formal user fees at public facilities were not introduced systematically. Although the Health Policy Review Commission in 1987 recommended introducing user fees at the national level, political considerations led the national parliament to reject the proposal for the introduction o f user fees in 1993 and no national policy document was developed for introduction o f user fees (Owor R. et. al, 1987). Finally, the Local Government Act in 1993 gave locally elected district councils the authority to charge for health services in public facilities along with decentralizing the delivery o f public health services to them. The MOH, with donor support, encouraged districts to charge for health services. However, giventhe decentralized system inUganda, the implementation of user fees was a local decision, and the amount and other user-fees policies (such as exemption policies, etc.) were decided locally. On average, the fees introduced were quite modest-Ush 500 for adults and 300 for children, and were consistently lower than those charged by nongovernment health facilities (Deininger and Mpuga, 2004). By 1999, 42 out o f 45 districts had introduced user fees (Okuonzi, 2001). In general, user fees were introduced quite late in the northern region-the poorest region in Uganda. The user fees were mainly used to supplement the health workers' salaries. 7.4 The government directed an Interministerial Task Force to carry out an assessment on user fees inthe country. The report from that assessment indicatedmixedresults. There were indications o f improved services on the one hand, and public outcry over the inability to pay on the other (Inter-ministerial Task Force, 1999). The household survey data from the DHS conducted in 2000-01 just before the abolition o f user fees showed that women considered the cost of services to be a significant problem in accessing the health services when sick along with other factors such as distance to the health facility, having to take transport, and negative attitudes o f the providers (Table 7.1). 123 Table 7.1 Percent of women (15-49 years) perceiving different constraints to access health services (2000-01) Obstacle Urban Rural Total Knowing where to go 4.5 7.3 6.8 Getting permissionto go 5.3 8.6 8.1 Gettingmoney needed for treatment 45.5 66.7 63.2 The distance to the health facility 14.3 49.8 43.9 Having to take transport 18.2 48.5 43.5 Not wanting to go alone 14.4 23.0 21.5 Concernthat there may not be a female healthprovider 16.3 16.7 16.6 Negative attitude o f healthprovider 53.1 39.8 42.0 Source: Authors' calculations from Demographic and Health survey, 2000-01 7.5 The Uganda Poverty Participatory Assessment also indicated that user charges were a major obstacle for the poor in accessing health care (MOFPED, 1999). Similarly, the data from the 1999 household survey also showed that inability to pay was one o f the major reasons for not using curative health care, particularly among the poor. A policy paper on user fees submitted by the MOH to the Cabinet in 1999 was approved with the condition o f formulation o f a broad national policy on user fees and exemption o f certain categories o f patients and diseases (Okuonzi, 2001). However, user fees became one o f the most important campaign issues in the 2001 presidential election, and the policy decision taken in 1999 was overriden by the new policy in March 2001, which abolished user fees inpublic facilities (Kiyonga, 2001). B. EMPIRICAL EVIDENCE ONTHE IMPACT OFABOLITION OFUSERFEES 7.6 A number o f small-scale studies, district visits, and MOH documents have reported significant increases in the use o f public health facilities after the abolition o f user fees (Mwesigye, 2002; Lake, 2002; and MOH 2002). 7.7 The study carried out by the MOH in collaboration with WHO also showed an increase in utilization o f health services in public facilities after the abolition o f user fees27. This longitudinal study was conducted immediately after the abolition o f user fees to assess the impact o f their removal on people's care-seeking behavior as well as their effects on the health systemperformance, Qualitative and quantitative data were collected from 106 health facilities including public hospitals and health centers, as well as PNFP facilities in six districts from 2001 to the end o f 2002. The key findings o f the study are summarized inthe Box 7.1, *'See MOHAnnual Performance Review 2003. 124 Box 7.1 Key findings fromthe study on the removalof user fees conducted by the MOH and WHO, 2002. Utilizationof services The study findings show that there has been a significant increase in utilization o f health services after March 2001. The increase was mainly for outpatient services and was significantly higher in public facilities compared to PNFP facilities and higher in lower level public facilities than in the referral facilities. All socioeconomic categories benefited from the policy change, with the highest benefits flowing to the poorest. Care-seekingbehavior Public facilities became a more popular health care option. The mainhindrances to public services were: the absence o f drugs, poor staff attitudes and long waiting times at the health facilities. However, inability to care for clients holistically is a characteristic o f all health care providers, withboth users and nonusers o f public facilities found to be using more than one health care provider to achieve their desired health outcomes. On average, 80% were using two providers, irrespective o f who the providers were. Access to services i s not yet adequate for both the public and PNFP units. While regular availability o f supplies was the mainhindrance inthe public units, cost o f supplies was a hindrance at the PNFPunits. Drugsand supplies On average, supplies provided to the health facilities per patient have remained the same as when patients were paying user fees. In addition, stockouts are presently at similar levels as with user fees. Drug stockouts remain the main complaint by the communities and the health workers inthe absence o f the user fees. Supplies to the PNFP facilities per client are significantly lower than those at the public facilities. However, the public facilities are having significantly higher stockout days than the PNFP units. Supplies are more erratic at the referral facilities than at the lower level facilities. Source: WHO, SIDA, and MOH, M a y 2003, "The effects o f abolition o f cost-sharing in Uganda," Report on the findings for 2002. 7.8 However, this study and other small-scale studies suffer fiom limitations o f a relatively small sample size and lack of adequate measures for the socioeconomic status of users. Hence, a more robust analysis i s neededto come to more solid conclusions regarding the impact o f user fees. Trends inUtilizationof CurativeHealthCare 7.9 The Uganda National Household Surveys (UNHS), 1999/2000 and 2002/03 provide, for the first time, nationally representative sample data to assess behavior change in seeking curative health care before and after abolishing user fees. The data allows disaggregated analysis by socioeconomic status o f the population (measured in terms o f household expenditure). Inaddition, since the user fees were not implementedat the nationallevel, the districts where no user fees were implemented provide a control group and a quasi- experimental studydesign rarely available. 7.10 The results from U N H S conducted before and after the abolition o f user fees reveal an interesting and complicated story about impact o f abolition o f user fees in Uganda. Overall, the nonutilization o f curative health services among those reporting sick in 30 days prior to the survey declined from 31% to 19% between 1999 and 2002/03. The aggregate increase in utilization, however, was mainly due to the 11.5 percentage points increase in use o f private sector facilities while the aggregate use o f public facilities remained more or less the same at approximately 24%. The main reasons given for 125 nonuse were: a) the illness was too mild, b) the services were too costly, and c) services were to far away. 7.11 The disaggregated data by expenditure quintiles and sex show that nonutilization declined much more by 18 percentage points in the poorest quintile compared to 9.5 percentage points in the richest quintiles, narrowing the poor-rich differences in the utilization after abolishing the user fees. While the use o f public facilities increased by 8.6 percentage points in the poorest quintile, the same declined by 7 percentage points in the richest quintile. From a welfare economics point o f view, it may be said that the abolishment o f user fees has achieved a desired pattern o f utilization: public facilities reachedmore poor people while the better-off groups who can afford to pay shifted to the private facilities. However, even in the poorest quintile, the increase in utilization o f private facilities2* exceeded the increase inutilization o f public facilities for curative care (11.4 percentage points versus 8.6 percentage points for public facilities). Hence, the private sector became the major provider even for the poor. 7.12 The household survey did not provide enough data to assess the reasons behind substantial decrease o f nonusers and the shift from public providers to private providers between the two surveys. There are two possible interpretations: a) the public facilities did not substitute for private facilities among the poor after user fees were abolished. Additional people (who would not have otherwise used health services) used public facilities, and this resulted in an overall decline in the nonutilization rate; and b) the private facilities provided a substitute for public facilities for richest quintile. The probable increaseduse ofpublic facilities by the poor due to the removal o f user fees may have led to the rich moving out o f public facilities. The critical information to explain why the rich opt out the free services i s not available. However, it i s possible that the quality o f services plays a significant role inpeople's choice o fproviders. 7.13 When the analysis i s disaggregated by gender, one finds that both poor men and women increased their use o f public facilities. However, the increased use o f public facilities did not replace their use o f private facilities. M e n from the poorest quintile as well as men from the richest quintile increasedtheir use o f private facilities by almost 12.6% and 10% points between 1999 and 2002/03. Similarly, the use o f private facilities increased among women in the lowest quintile by 10% points and by 14% in the richest quintile. Further analysis inthe next sections attempts to explain the situation inmore detail. 28 The private facilities include bothprivate not-for-profit and private for-profit facilities. In 1999,no ownership question was asked to respondents who reporting visiting health centers, dispensaries, and health clinics. It i s assumed that all health centers and dispensaries are inthe public sector and all clinics are in private sector. Furthermore, no distinction could be made between private for-profit andprivate not-for- profit facilities, though this distinction was made in2002103 data. 126 Table 7. 1Trendsinutilizationof publicand privatefacilitiesin 1999 and 2002 None Public Private Other Exp. quintiles 1999 2002103 1999 2002/03 1999 2002103 1999 2002103 Poorest 20% 45.1 27.2 23.4 32.0 29.3 40.7 2.1 0.1 2nd quintile 38.6 21.4 21.5 28.0 38.3 50.2 1.6 0.5 Middle 31.0 18.0 23.8 24.7 43.8 57.3 1.4 0.1 4th quintile 28.2 15.7 24.2 21.1 45.9 62.6 1.8 0.6 Richest 20% 25.5 16.0 23.3 16.3 50.0 67.0 1.3 0.7 Residence Rural 31.4 19.9 24.4 25.4 42.6 54.2 1.6 0.4 Urban 24.0 16.4 16.3 15.3 58.8 67.9 0.9 0.4 Average 30.5 19.4 23.4 24.1 44.6 56.1 1.5 0.4 Men Poorest 20% 42.4 27.2 24.7 30.0 30.1 42.7 2.8 0.1 2nd quintile 39.1 21.6 22.6 27.5 37.0 50.4 1.3 0.5 Middle 31.4 17.6 23.1 23.8 44.1 58.5 1.5 0.1 4th quintile 27.1 15.8 23.2 21.7 47.4 62.3 2.3 0.2 Richest 20% 25.4 15.3 24.5 15.3 48.5 68.7 1.6 0.7 Residence Rural 30.9 19.8 24.7 24.5 42.5 55.4 1.9 0.3 Urban 23.4 14.9 17.1 15.1 58.7 69.5 0.8 0.5 Average 29.9 19.2 23.7 23.3 44.6 57.2 1.8 0.3 Women Poorest20% 47.2 27.3 22.5 33.6 28.7 39.0 1.6 0.1 2nd quintile 38.2 21.3 20.7 28.3 39.3 50.0 1.9 0.4 Middle 30.7 18.3 24.6 25.5 43.4 56.2 1.3 0.1 4th quintile 29.0 15.5 25.0 20.5 44.6 62.9 1.4 1.1 Richest 20% 25.5 16.7 22.1 17.3 51.4 65.4 1.o 0.6 Residence Rural 31.9 20.0 24.1 26.3 42.6 53.2 1.4 0.5 Urban 24.6 17.5 15.5 15.5 58.9 66.6 1.o 0.4 Average 31.0 19.7 23.0 24.8 44.1 55.1 1.32 0.5 Source: UgandaNational Household Survey 199912000and 2002103 Trends in disease-specific utilization of services 7.14 The results shown intable 7.3 reveal that highest increase inutilization was observed for some o f the most important public health problems such as malaria, measles, and diarrhea, while nonutilization for conditions such as dental problems, intestinal infections, and skin conditions remainmore or less the same level. 127 Table 7.2 Change innonutilizationof health services by disease condition before and after abolition of user fees Decease in non- utilization Disease 1999 2002103 in2002103 Total Non- Total N o Treated utilizati Treated at Non- attendance at home on N o attendance Home tuilization Malaria 5.3 24.9 30.2 2.5 11.4 13.9 16.3 Respiratory 11.9 23.5 35.4 13.1 16.4 29.6 5.8 Measles 1.9 23.3 25.1 2.0 5.8 7.9 17.3 Diarrhea 7.8 24.1 31.9 7.5 9.7 17.1 14.7 AIDS 5.O 11.7 16.7 2.4 0.2 2.5 14.2 Dental problems 16.2 12.3 28.6 15.6 12.5 28.1 0.4 Accident 5.4 21.2 26.6 9.6 9.4 19.0 7.7 Intestinal infections 10.0 15.7 25.7 14.7 11.6 26.4 -0.6 Skin conditions 16.6 15.3 31.9 18.1 10.7 28.8 3.2 Hypertension 12.7 16.3 28.9 9.9 9.2 19.0 9.9 Ulcers 8.4 15.3 23.8 11.3 11.6 22.9 0.9 Others 13.0 18.2 31.3 14.9 14.0 28.9 2.3 Total 7.8 22.7 30.5 7.4 12.0 19.4 11.1 ~~ ~ Source: Uganda NationalHousehold Survey 199912000and 2002103 Reasons for nonuse of curative health services 7.15 As mentioned above, cost was perceived as one o f the major reasons for nonutilizationo f services from the Uganda National household surveys (1999/00 and 2002103) conducted before and after the abolition of user fees. However, in these surveys, the question was posed only for the individuals who reported to be sick but chose not to seek care. Mild illness and cost o f the services were the most commonly reported reasons in both the surveys, though the reporting o f cost o f the services declined moderately in 2002/03 survey. Overall (including even the individuals who used services) 16% mentioned cost o f the services as a reason for not seehng services in99/00, only 7% said so in2002/03. However, these figures should be interpreted cautiously, as many individuals who sought care may have found cost a significant problem while the gravity o f the problem forced themto seek care even at the risko f future impoverishment. c.DISTRIBUTION OF BENEFITS[BENEFITINCIDENCE] 7.16 A standard incidence analysis using information on unit cost of health services and utilization o f services by different socioeconomic categories has been commonly used to estimate the distribution o f benefits o f public expenditures in the social sector (Demery 1997). More recent research has shown that the standard benefit incidence analysis has many limitations (van de Walle, 1998, and Lanjouw and Ravallion 1999). Due to lack o f detailed costing data at the facilities level, benefit incidence analysis often makes strong 128 assumptions on subsidy and unit costs and assumes that cost o f provision reflects the benefit to users and does not adequately reflect quality o f services, etc. 7.17 Cost information on service provision as well as out-of-pocket expenditure at each level o f public facilities i s not available. Rather than malung strong assumption on unit costs for various levels o f health facilities, this study uses the distribution o f utilization of public facilities by expenditure quintiles to illustrate the distribution o f benefit incidence o f usingservices. 7.18 The benefit incidence analysis indicates that after abolition o f user fees, the public expenditure to public health facilities became more pro-poor, especially at the lower level health facilities. While only 11.4% o f the total users belonged to the poorest quintile and 26.4% to the richest in 1999, the distribution reversed in2002103 (Table 7. 2) with almost 50% o f the total users o f health centers belonging to the poorest 40% o f the population. The benefit-incidence o f the public hospitals also became more equally distributed in 2002/03 with the percentage o f users from the poorest two quintiles increasing from 18.6% in 1999 to 38% in2002103. Table 7.3 Benefit incidence of public healthfacilities in 1999 and 2002/03 Healthcenter Public hospital 1999 2002103 1999 2002/03 Expenditure quintiles % Cumulative % Cumulative % Cumulative % Cumulative Poorest 20% 11.4 11.4 28.6 28.6 7.6 7.6 17.5 17.5 2ndquintile 13.9 25.3 21.9 50.5 11.0 18.6 20.3 37.8 Middle 26.6 51.9 22.7 73.1 16.5 35.1 18.8 56.6 4th quintile 21.7 73.6 15.2 88.3 22.7 57.8 22.5 79.2 Richest 20% 26.4 00.0 11.7 100.0 42.2 100.0 20.9 00.0 Residence Rural 96.1 96.1 95.2 95.2 87.6 87.6 84.3 84.3 Urban 3.9 100.0 4.8 100.0 12.4 100.0 15.7 100.0 Source: UgandaNational Household Survey 199912000and 2002103 7.19 The concentration indices and the concentration curves, which are graphic illustrations o f the distributive effect o f utilization o f public facilities (Figure 7.1), support the conclusions derived from Table 7.3 that utilization o f curative services inpublic facilities became more pro-poor after abolishing user fees in 2003. Given such a utilization pattem, public spending on health facilities therefore became more progressive, i.e., public fundingthrough health centers reached more poor population. Table 7.3 The Poor/rich ratio for utilization of different types of public facilities Public hospital outpatient Public hospital inpatient Health center Dispensary 1999 0.77 0.56 1.83 1.31 2002 1.08 0.61 2.45 2.19 Source: Authors calculations from Uganda National Household Survey 1999/2000and 2002103 129 7.20 The Poodrich ratios inutilization o f Figure 7.1 Concentrationcurves for utilization various public health facilities also of publichealthfacilities show improvement in reaching the poor. hospitals,Except i s still used more inpatient care at +Equality line which 100% by the rich, the use o fhealthcenters and dispensaries by the poor has increased significantly. A higher percentage o f the poor also made public hospital outpatient visits than did the rich in 2002/03, reversing ~ . . . ~ ~ ~ ~ ~ ~ ~ ~ . ~ ~ ~ ~ the pattern observed in 1999 (Table 7.3). 7.21 The overall usage o f private facilities i s pro-rich but the pattern improved between 1999 and 2003 0% 20% 40% 60% 80% 100% (Table 7.4). The available household data does not permit the separation o f use o f PNFP facilities Source: UgandaNational Household Survey 2002103 from that o f the private-for-profit sector facilities. Also there i s no data that would allow an estimation o f share o f government subsidy in the total cost o f services provided by PNFP. It is not clear whether increased use of private facilities is due to provision o f more affordable care to the poor (e.g. more exemptions and reduction o f user fees, etc.) in PNFP facilities and hence the attraction o f more poor patients to these facilities, or merely due to the lack o f other good quality options for the poor who are forced to use these facilities. Table 7.4. Benefitincidenceof private facilitiesby the type of facilities I Outpatient hospital IInpatient Hospital1 Clinics I Private total* Expenditurequintiles 1999 2002 1999 2002 1999 2002 1999 2002 Poorest 20% 5.1 9.9 3.5 12.5 5.5 13.3 6.0 13.5 2nd quintile 9.2 13.8 14.8 16.5 11.9 15.2 11.1 17.1 Middle 16.1 16.6 10.1 10.5 17.0 19.4 16.7 20.0 4th quintile 18.3 25.9 24.1 19.6 22.3 23.9 23.0 23.4 Richest 20% 51.3 33.7 47.5 40.9 43.3 28.2 43.3 26.1 Residence Rural 86.3 81.0 91.4 88.9 81.3 81.2 83.4 83.8 Urban 13.7 19.0 8.6 11.1 18.7 18.8 16.6 16.2 *Including pharmacies and shops Source: UgandaNational HouseholdSurvey 199912000and 2002103 7.22 More in-depth analysis usingmultinomial regressions for the year 1999 and 2002/03 also confirms the findings seen in the bivariate analysis that the utilization o f both public and private facilities increased inthe year 2002/03 and the reduction o f the nonutilization rate i s more significant among the poor than among the rich. The use o f public facilities increased among the poor, malung the public facilities more pro-poor (see Annex 7 for more details). 130 D.COMPARISONOFDISTRICTS WITHAND WITHOUT USERFEES 7.23 Kalanagala district in the Central region and Lira district in the Northem region never implementeduser fees. These two districts provide a useful control group, and a very rare opportunity to study the impact o f user fees in a quasi-experimental setting. Ifchanges in the utilization o f health services are due to the abolition o f user fees, then no changes should have been observed in Kalanagala and Lira. If similar changes in utilization patterns are observed in these two districts also, then the changes are due to some other factors. 7.24 Table 7.5 shows that before the abolition o f user fees, the utilization o f curative care was almost 6% higher in the no-user fee districts. After the abolition o f fees, the utilization improved in both groups o f districts, though much more in the userfee districts, thus closing the gap observed in utilization before introduction o f user fees. However, while the use o f public facilities decreased in the no-user fee districts during this period, it remained almost the same at the aggregate level. In both groups o f districts, the utilization o f private facilities increased significantly and was mainly responsible for decline innonutilizationrates. Table 7.5 Utilization of curative care inuser fee and no-user fee districts before (1999) and after (2002/03) abolition of user fees None Public Private Other After Before After Before After Before After User fee districts 30.7 19.4 23.0 24.2 44.8 56.0 1.6 0.4 No-user fee districts 24.3 19.7 33.1 21.4 41.7 58.2 0.9 0.6 Source: UgandaNational Household Survey 199912000 and 2002103 7.25 However, while the nonutilization rate reduced in the user fee districts in all the five quintiles (more so inthe poorer quintiles), the reduction innonutilizationrates were more or less confined to the poorest quintile and the 2ndrichest quintile in the no-user fee districts: the nonutilization in the poorest quintile decreased in both use-fee and no-user fee districts, though more so inthe user fees districts (18.3 percentage points compared to 14.6 percentage points). The reduction in nonutilization rate in the no-user fee districts was mainly due to an increase in the use o f private facilities, while in the user fee districts, both the increase in use o f public facilities (9 percentage points) and private facilities (11.3 percentage points) contributed to the decline in the nonutilization rates in the poorest quintile. 7.26 Inthe user fee districts, the utilization ofpublic healthfacilities increasedconsiderably in the poorest two quintiles, remained almost the same in the middle quintile, and declined inthe richest two quintiles after the abolition of user fees. However, inthe no-user fee districts, the utilization o f public facilities remained almost the same in the poorest quintile and deteriorated in the other four quintiles. The utilization o f private facilities increased in all the five quintiles in both user fee and no-user fee districts after abolition o f the user fees. 7.27 The above analysis comparing the user fee (intervention) districts and no-user fee (control) districts supports the argument that abolition o f user fees indeed contributed partially to the reduction of health service nonuse in the user fee districts, especially 131 among the poor. The poor in the user fee districts increased their use o f public facilities after abolition o f user fees. However, the data also show that regardless o f user fee policies in public facilities, both poor and rich people increased significantly the use o f private facilities. The overall use o f public facilities inno-user fee districts has drastically decreased. Table 7.6 Change in Utilization of curative care before and after abolition of user fees inuser fee and no-user fee districts Before abolition of user fees User fee districts No-user fee districts 2nd 2nd 2nd 2nd Poorest poorest Middle richest Richest Poorest poorest Middle richest Richest None 45.8 39.5 31.7 28.2 25.6 38.0 18.2 18.2 25.6 19.8 Public 23.0 21.2 23.1 23.6 23.1 28.3 28.4 38.3 39.0 30.4 Private 29.0 37.6 43.8 46.3 50.0 32.3 53.5 42.6 34.6 48.9 Other 2.2 1.7 1.4 1.8 1.3 1.5 0.0 0.8 0.9 0.9 After abolition of user fees None 27.5 21.6 17.9 15.7 16.0 23.4 17.9 19.0 16.2 15.4 Public 32.1 28.3 24.9 21.3 16.4 30.0 18.9 18.7 8.8 13.3 Private 40.3 49.6 57.1 62.5 66.9 46.7 63.2 62.3 68.9 71.3 Other 0.1 0.5 0.1 0.5 0.7 0.0 0.0 0.0 6.1 0.0 Source: UgandaNational HouseholdSurvey 199912000and 2002103 Paymentfor drugs 7.28 Although no data for the payment for drugs are available in the 1999 U N H S survey, the data from the 2002103 survey show that less than 35% o f the people paid for drugs inthe public health facilities after abolition o f the user fees. One encouraging finding was that the percentage o f the people who paid for drugs in the public health facilities i s significantly lower inthe poorest two expenditure quintiles (Table 7.7). 132 Table 7.7 Payment for drugs among respondentswho fell sick inthe last 30 days and sought care from a healthprovider 2nd 2nd Poorest poorest Middle richest Richest Total Among respondents who sought care from a public sector provider No drugs required 1.1 1.2 0.6 1.1 1.7 1.1 Obtained free 72.7 69.9 67.1 63.4 46.3 65.3 Some drugs were purchased 19.0 16.8 20.9 24.7 25.7 20.9 All drugpurchased 7.2 12.1 11.4 10.8 26.4 12.6 Among respondents who sought care from a NGOheligious provider No drugs required 0.7 0.0 0.0 0.0 0.0 0.1 Obtained free 9.2 9.0 9.0 8.4 7.1 8.3 Some drugs were purchased 21.4 6.8 17.4 14.8 5.7 12.4 All drugpurchased 68.8 84.2 73.6 76.9 87.2 79.2 Among respondents who sought care froma private provider No drugs required 0.0 0.3 0.0 0.1 0.3 0.2 Obtained free 0.5 0.1 0.6 0.5 0.8 0.5 Some drugs were purchased 4.5 3.1 2.2 4.1 2.9 3.3 All drugpurchased 95.0 96.5 97.2 95.3 96.1 96.0 Source: UgandaNationalHousehold Survey 2002103 E.IMPACTOFABOLITION OFUSERFEES SERVICE PROVISION ON The impact on healthworker incentives 7.29 A small study conducted in three districts noted that health providers favored user fees significantly more than did users, and that fees collected were used to supplement health worker salaries (Kondo-Lule and Akelo, 1998). They also noted that in most cases, the stated exemption policies were not working as health workers had an incentive in not informing patients about them (Kivumbiand Kintu, 2002). Immediately after user fees were abolished, the government increased the allocation to the health sector to compensate for the income lost from user fees. However, it i s clear from the recent government budget allocation that increased budget in the health sector i s not sufficient to offset the income loss due to the abolition o f user fees and nor sufficient to meet increased utilization. Therefore, more pressure i s placed on health workers to provide more services with less. Without an adequate incentive system inplace, there i s a danger that informal charges will increase or more providers will spend more time away from their duties either in private practice or c a v i n g out other business to supplement their income. The impacton the quality of services 7.30 As discussed in Chapter 5, information on service quality is very limited. A health facility survey conducted in June 2002 showed widespread drug stockouts in the health facilities as well as a large number o f vacancies in the public health facilities (ORC, Macro et al., 2003). Deininger and Mpuga (2004), using the data from the community 133 questionnaire inthe UDHS survey, show that quality inthe form o f staff presence in the facility, availability o f drugs, and informal charges did not change before and after the abolition o f user fees. However, their information i s based on health providers' responses rather than on the responses o f the community members. The MOH and WHO'Suser fees study also shows that drug stockouts remain the main complaint by the communities and health workers. 7.3 1 Closely monitoring quality o f service provision and users' perceptions and experiences after the abolition o f user fees i s much needed because the quality o f services i s a critical factor for continuing to attract users. F.DISCUSSION RECOMMENDATIONS AND 7.32 This chapter assessedthe impact o f the abolition of user fees inUganda using available information. Abolishing user fees inpublic facilities has yielded positive results: the poor use public services more and get more free drugs. The public health services inUganda became more pro-poor after user fees were removed. However, the coverage o f flee services i s still limited and few people benefit. The magnitude of increasing use o f public providers i s less than one would have expected from the abolition of user fees. Noticeably, the use o f the private providers both for-profit and non-profit has drastically increasedrecently, not only among the rich, but also among the poor. 7.33 The situation raises important questions that need further investigation and policy consideration: 7.34 Firstly, why has almost 70% o f the top quintile abandoned the public sector and why do more than 40% o f the poor from the lowest quintile also use private providers when free services are provided by the public sector? The issue i s related to access and quality o f services provided by the public sector. 7.35 Secondly, even while the use o f curative health services inpublic facilities has increased, it may not produce better health outputs and outcomes. It is possible that people may use the free services first, and then pay for services when free services do not resolve their problems. Such practices can increase inefficiency and further stretch the very limited resources inthe households. 7.36 The comparison of the user fee and no-user fee districts shows that regardless o f user fees, patients increasingly use private facilities. The government needs to ensure the quality o f services provided by private providers so that the population gets value for money. 7.37 Considering the size o f the burden o f disease in Uganda (25% population reporting sick in last 30 days) and other heath care needs of a very young population, the current network o f public health facilities i s not adequate for providing health care to the entire population. At the same time, the government i s not ina position to substantially increase its investments in the sector inthe near future. Therefore, the private sector i s required to play an important role in service delivery. The government needs to encourage the private sector, particularly the PNFP sector, through various measures to contribute to preventive care and public health objectives. 134 7.38 Overall, inorder to maintain and further improve pro-poor service provision inthe public sector, the government has to pay close attention to the quality o f services and increase fundingallocationto ensure quality does not suffer from a lack of fundingat the facility level. 135 8. RECOMMENDATIONS FORHEALTHSECTOR DEVELOPMENT 8.1 The previous chapters o f this report discuss determinants o f health-related poverty outcomes and the strengths and challenges facing the Ugandan health sector. This chapter links issues arising from them and highlights five key areas so as to provide recommendations for the short to medium term for Uganda's progress towards achieving the MDGtargets and contribution to nationalpoverty reduction efforts. 8.2 While the key health policies and reforms appear to be in place, some health indicators have remained persistently resistant to improvement. Uganda needs to refocus its efforts on reachingits poorest and most vulnerable citizens and communities. This requires: A. Prioritizing interventions which will affect infant andmaternalmortality B. Improving healthpromotion anddiseasepreventionpractices at the family and community level through community mobilization and intersectoral collaboration C. Mobilizing funds for the health sector including strategies that encourage risk-pooling mechanisms D. Focusingonhumanresources andcollaborationandcoordinationwiththe private sector to improve health sector performance E. Focusingon healthmanagement information systems and strengthening capacity for supervision inorder to carry out oversight and accountability A. Interventionsthat affectinfant and maternalmortality 8.3 The HSSP i s ambitious and comprehensive. Because financing for it falls short, difficult decisions have to be made to adapt to this lack o f financing. The benefits o f prioritizing within the HSSP were demonstrated when immunization efforts were revitalized in2000. Reproductive health has been a priority area for two years; a reproductive health strategy has been finalized and costed but has yet to be financed and implemented. Interventions that improve maternal mortality will also improve neonatal mortality-two areas which would contribute to achieving the MDGtargets. Recommendations: 0 Since the single most important contributor to highinfant and maternalmortality i s the combination o fhighfertility and short birthspacing, Uganda shouldprioritize efforts to improve the use o f modern family planning methods so that women can achieve their desired family size and birthspacing. Special attention should be given to families inthe lowest wealth quintiles, to people livinginthe Northern region, and to women with low educational levels. 136 Increasingthe numbers o f institutional deliveries requires increasing awareness among women and their families o f the danger signs and complications o fpregnancy and delivery. Moreover, women needto be alerted to the possibility that complications may arise duringchildbirth even though the pregnancy seems normal. Women should be assisted indeveloping a birthplan for delivery at a health facility. Together with this: The provision of basic and comprehensive obstetric care is essentialfor reducing maternal and infant mortality. Currently there i s a lack o f emergency obstetric care facilities, lack o f appropriately trained staff, and shortages o f drugs, equipment, and supplies. Providing adequate numbers o f facilities throughout the country and human resources will take years to achieve. Inthe short-run, the provision o f communication (radio communication, mobile telephones), transportation (ambulances, community transport), and a good referral system are critical and these interventions should be feasible. Antenatal care should focus on interventions which have been found to improve maternal health outcomes: use of tetanus toxoid, use o f antimalaria prophylaxis, screening for hypertension, and screening for and treating STDs 0 Loweringteenage pregnancy, raising the age o f sexual debut, providing youth-friendly services including sympathetic post-abortion care, and assisting youth intheir sexual and reproductivehealthrequires the concerted and sustained efforts o f families, schools, and communities. B. IMPROVEHEALTH-PROMOTING PRACTICES AT THE FAMILYAND COMMUNITYLEVEL 4. The main c a u w o f child mortality-malaria, diarrheal disease, ARI, and malnutrition- require interventions at the family and community level and cannot be implementedby the health sector alone. Women's education, especially completion of secondary school, i s a very important determinant of disparities inthe health-related MDGs. Secondary education o f women affects child survival, TFR, attended deliveries, and other aspects o f care-seeking behavior. Intersectoral collaboration i s essential among the water, sanitation, and education sectors. Decentralization can help this process through integrated district planning. Community mobilization i s the key to improvedpractices at the household level. Recommendations: The MOHshouldconsider different options for community mobilization and assist each district indevising its own model. While each district can decide on the structure and process for achieving its results, a mechanismfor holding districts accountable for their performance should be developed. Further work i s neededto understand the potential and requirements o f an effective and sustainedprogram that involves community health workers. The MOH and districts can mobilize NGOs and community-based organizations to carry out work for behavior change. The emphasis on BCC should be to assist individuals and families improve their decision-making capacity and their practices for healthpromotion and disease prevention. Caretakers and families needto be able to recognize the danger signs o f illness, understand the consequences o f action and non-action, understand the 137 basics o f home-basedmanagement o f diseases, know where and how to seek care outside the home, and be able to take positive action. 0 Sustained emphasis on increasing educational levels to secondary school with gender parity must be continued; informal education and adult literacy initiatives should be explored with other sectors; and the continueduse o f indicators to monitor women's empowerment i s recommended. 0 Some key indicators inhousehold knowledge, behavior, and practice shouldbe used in the evaluation o f district performance and inthe formulation o f the district league table. C. HEALTH SECTOR FINANCING 8.4 As has been pointedout throughout this report, the most pressing and greatest challenge for Uganda i s how to increase financing to the health sector. Although government spending on health has increased in recent years, total expenditure including out-of- pocket expenditure falls short o f what i s needed to finance the HSSP which.is estimated to be $28 per capita. The current MTEFprojections do not indicate substantial growth in the health budget in the medium term. Important progress has been made in improving the allocative efficiency o f public financing by channeling increasing amounts o f funds for primary care. There i s a need to further enhance efficiency in allocations by targeting public health. 8.5 The data on health care expenditure and utilization before and after user fees were abolished were discussed in Chapters 6 and 7. Since the abolition o f user fees, public sector curative care has become more pro-poor. While use o f public facilities has increased among the poorest quintile, the greatest increase has been seen in the private sector (PNFP and PFP) by both the better-off and the poor. The key question is: why do thepoor use more of theprivate sector whenfree services are available? Recommendations: 0 The GoU needs to continue to increasebudget allocationto health inorder to further improve both coverage and quality o f service provision. The success o f the abolition o f user fees will depend on whether the government can continue to provide resources to the health sector that will compensate for the loss o f income and to meet increasing demand. 0 The government needs to improve its information systems so that it i s able to track spending. Studies like the expenditure tracking surveys for primary care shouldbe continuedand expanded to other areas such as public health. The government needs to collect more current information on utilization andprivate expenditure on health care. Information i s needed on the patterns o f attendance, the paths to different types o fproviders, reasons for selecting certain providers, and expenditures on treatment. 0 Uganda shouldharness the considerable out-of-pocket spending on health. The government needs to encourage the development o f social insurance schemes and alternative riskpooling mechanisms. 138 D. HEALTH SECTOR PERFORMANCE: A FOCUSON HUMAN RESOURCESAND PRIVATE PUBLICPARTNERSHIPSINHEALTH D.1 Human Resources 8.6 One o f the main obstacles to service delivery i s the scarcity and distribution o f appropriately trained health personnel. As discussed in Chapter 6, the construction o f health facilities inrural areas i s insufficient in itself; health facilities must be functional, i.e., adequately and appropriately staffed and receiving the necessary inputs to provide quality services. This requires closer coordination between the development and recurrent budgets. 8.7 Although low salaries are often highlighted as a problem in attracting and retaining staff, there i s evidence that other factors are equally important. The PNFP sector pays lower salaries and yet i s able to retain staff in remote areas and i s often perceived to provide better quality services. This points to other factors being important in health worker motivation. The problems o f human resources are urgent and should be addressed in the following ways: Recommendations: Agree a comprehensive humanresource development plan with all the major stakeholders: Ministries o f Health, Education and Sports, Local Government, Public Service and Finance, Planningand Economic Development, the private not-for-profit sector which runs20 out o f 48 o fthe training schools, and professional councils. The planneeds to consider health service needs, training capacity, the recruitmentbase, in- service training, the impact o f HIV/AIDS on the workforce, and how terms and conditions o f service affect healthworkers inthe public and private sectors. Since the inter-ministerial mechanism has not worked, another means o f coordinatingthe plan should be found. e Develop a human resource management policy that addresses the terms and conditions o f service, including incentives, promotions, and career development; recruitment; deployment and retentionmechanisms, especially for achieving equity; and the impact o f HIV/AIDS inthe workplace. The foundation for developing a performance-based incentive system needs to be established; this requires a robust management information system, well-established managementroutines, and a highdegree o f transparency. e The Ministries o fHealth, Local Government and Public Service should increase capacity inhumanresourcemanagement at bothcentral anddistrict levels. Improving the living conditions o f staff i s one o f the ways to retainhealth workers in remote areas. Decentralizationprovides the opportunity for districts to devise strategies to improve the living conditions o ftheir health workers. 0.2 Public-privatepartnershipsfor health 8.8 The PNFP sector provides good services in areas where public services are sparse, provides quality services aimed at benefiting poor people, and has a good history o f transparency infinancial transactions. The National Policy on Public-PrivatePartnership inHealth provides guidance on collaborationwith the private sector. It is especially at 139 the district level that integrated planning will help to ensure that public and PNFP services complement each other. 8.9 Less i s known about the private for-profit sector, which i s growing rapidly and increasingly playing a role inthe provision o f services. Given the importance o f boththe PNFP and PFP sectors, it i s important that central government carry out its oversight responsibilities to ensure that people are getting value for money. Recommendations: 0 The financial subsidy to PNFPproviders shouldbe continued and expanded with concrete guidance for directing the activities o f the PNFPto public health activities. 0 The MOHand local governments should strengthen their capacity to implement the National Public-Private Partnership inHealth Policy. With regardto the PFP sector, the first step is to undertake an assessment to gain an overview o f the scope o f services offered and what links exist with the public sector. 0 Central government should strengthen its capacity to address issues o fregistrationand regulation. E. STRENGTHENINGOVERSIGHTANDACCOUNTABILITY 8.10 The lack o f accountability in the health system has been noted throughout this report. Enforcement o f accountability in a decentralized setting i s a major challenge. Weaknesses in the health information system make monitoring and evaluation more difficult. Weahesses in supervision have been noted in Annual Health Sector Performance Report, special program reviews, and the aide-memoire from the Joint Review Missions. Although units exist at the MOH to carry out this function, clearly more needs to be done. Supportive supervision i s known to improve worker motivation and i s one o f the components in a supportive work environment. Improving services for the poor can be strengthened through increasing accountability innumber o f ways: Recommendations: 0 Continue to carry out participatory poverty assessments and develop other means to ensure feedback from beneficiaries to increase the voice o f the poor. More important i s the needto follow up on the findings from the poverty assessment when formulating new policies. 0 Monitor client satisfaction with service delivery, not only curative services, but also preventive services and health promotion. Baseline data needto be collected for the HSSP indicator and targets for improvement needto be set inthe HSSP. Follow the work o f the Uganda National Health Users/Consumers Organizationto see whether it provides useful mechanisms for districts and facilities to obtain feedback on user satisfaction. 0 Further improve the health information system and the use o f information for monitoring and evaluation by further strengthening the capacity o fthe MOH in identifying performance indicators. Ensurethat data are accurate, valid, timely, and 140 include the poor. The system should also strengthen mechanisms that ensure managers use information gained for service delivery. 0 Continue to publishthe district league tables and include other key indicators such as costing and expenditure as well as accountability measures for district performance in the league table. Reporting on district performance should be continued on a regular basis andbe made known to the public. Worlung with districts to investigate the factors behind weak performance and findpossible solutions will be needed to assist poorly performing districts. As the provision o f drugs i s one o f the main indicators of perceived service quality, monitoring drug spending by the public sector should be a regularly reportedindicator. 0 The MOHneeds to a) strengthen its capacity for providing technical leadership for the districts andb) assist districts indeveloping their capacity to provide supportive supervision to Health Subdistricts. 141 BIBLIOGRAPHY Ablo, E. and R.Reinikka, 1998. "Do budgets really matter?Evidence from public spendingon education and health inUganda," World Bank Policy, ResearchWorking Paper, No. WAS 1926. Ahn, J. ,P.Hutchinson, andK.Strumpf, , "Decentralization and Government Provision of Public Goods: The Public Health Sector inUganda," Measure Evaluation, Chapel Hill, North Carolina Akwara, P., S. Alayon et al, 2003. Uganda Delivery ofImproved Servicesfor Health (DISH) Facility Survey 2002, Measure EvaluationTechnical Report Series, No. 14. Chapel Hill, North Carolina BASICS , 1995. "Malnutrition and ChildMortality: Program Implications ofNew Evidence" Washington Birungi,H.,F.Mugisha, X.Nsabagasani, S.Okuonzi, andA. Jeppson, 2001. "The policy on public-private mix inthe Ugandan health sector: catching up with reality," Health Policy and Planning, 16 (Suppl2) 80-87 Black, R., S Morris, and J. Bryce, 2003. Where and why are 10 millionchildren dying every year?", Lancet 2003 July 26: 362(9380): 262 Bossert, T, J. Beauvais andD.Bowser, 2000. Decentralization of Health Systems: Preliminary Review of Four Country Case Studies,Major Applied ResearchNo. 6, Technical Report No. 1, Bethesda, MD: Partnership for HealthReformProject, Abt Associates Inc. Business Synergies, 2004. "Ministry of Health FundsTraclung Study Draft Report" Business Synergies, 2002. "Drug Tracking Study -FinalReport" Chaudhury, N.,2004. "Absenteeism among health workers inUganda," Draft inprogress, World Bank, Washington DC. Claeson, M.,C.C. Griffin, T.A. Johnston, M.McLachlan, A. Soucat, A. Wagstaff, and A. Yazbeck, ,2001, "Health, Nutrition, and Population" inthe PRSP Sourcebook, World Bank Cong, R., 2004. Water and Sanitation Sector Moves to SWAP: The Uganda Experience (presented at the Water Week, World Bank, Washington DC, Feb 2004) Corkery J., 2000. "Public sector reforms and their impact on health sector personnel inUganda," International Labor Office and World Health Organization, Geneva Deininger, W., and P. Mpuga, 2003. "Has the Abolition ofUser Fees improvedthe Welfare of Vulnerable Groups? Evidence from Uganda," World Bank, Washington DC 142 Demery, L., 1997. "Benefit Incidence Analysis," Poverty Reduction and Economic Management Network, World Bank (unprocessed mimeo) DFiD,2003,A Review of theHuman Resource Content of PRSP and HIPC documentation in 6 selectedAfrican Countries DflDHealth Systems Resource Center, London, UK Directorate of Water Development 2004, Republicof Uganda, 2004. Rural Water Supply and Sanitation, http:llwww.dwd,co.udrural-background.php,assessed3115/04 Donabedian A, 1986. "Criteria and Standards for Quality Assessment and Monitoring," Quality Review Bulletin, 99-108 Gwatkin, D., S. Rutstein, K.Johnson, E.A.Suliman, and A. Wagstaff, 2004. Socio-economic Differences in Health, Nutrition and Population, World Bank, WashingtonDC. Gwatkin, D., S. Rutstein,K.Johnson, R.Pande, and A. Wagstaff, 2002. Socio-Economic Differences In Health, Nutrition, and Population in Uganda, World Bank Horizons InternationalLtd.(HIL), 2001. "PHC Conditional Grant Trackmg Study" Hutchinson, P., J. Akin andF. Ssengooba, 2003. TheImpacts of Decentralization on Health Behaviors in Uganda,Measure Evaluation, Carolina Population Center, University of North Carolina at Chapel Hill Hutchinson, P., 1999. "Decentralization inUganda's Health Sector,'' inHealth Care in Uganda: SelectedIssues, World Bank, Washington DC Hutchinson, P., 1998. "Equity of Access to Health Services inUganda: The Effects of Income, Gender,Proximity to Services and Quality o f Care" inHealth Care in Uganda: Selected Issues,World Bank DiscussionPaperNo. 404, World Bank, WashingtonD C Hutchinson, P., 1999.Health Care in Uganda: SelectedIssues, World Bank Discussion Paper No. 404 Jareg, P., S. Baine, J. Imoko, and H.Luwaga, 2003. "Tuberculosis Control Programmes and their ImpactonHealth Systemsand Service Development," World HealthOrganization /HeSo, Norway Jeppson, A,, 2002. "SWAP dynamics ina decentralized context: experiences from Uganda," Soc Sci Med 55 2053-2060 Jeppson, A., 2001. "Financial priorities under decentralization inUganda," Health Policy and Planning 16 (2) 187- 192 Jeppson, A., P-0. Ostergren, and B.Hagerstrom, 2003. "Restructuring a Ministry of Health-an issue of structure and process: a case study from Uganda," Health Policy and Planning 18 (1) 68-73 John, O., and E.Maniple, 2003. "Faithfulness to the mission: effect ofreducing user fees on 143 access to PNFP health services", Study Report, Uganda Martyrs' University. October. Kapiriri, L.,O.F. Norheim, K.Heggenhougen, 2003. "Public participation inhealth planning and priority setting at the district level inUganda," Health Policy and Planning 18 (2) 205-213 Kasumba, G. A. Land, 2003. "Sector-Wide Approaches and Decentralisation Strategies pulling inOpposite Directions?A Case Study fromUganda." European Center for Development Policy Management Maastricht, The Netherlands Kerstiens, B., A. Akii, N.Mbona, A. Zziwwa, and W.Edson, 2004. "Improving the management o f obstetric emergencies inUganda through case management maps," USAID/Quality Assurance Project Kipp,W, A. Kielmann, andE.Kwered, 1991. "Postwar Uganda: a review ofhealth services in Kabarole District," International Journalfor Health Planning and Management 6: 221- 228. Kivumba, G. and F.Kintu,2002. "Exemptions and waivers from cost sharing: ineffective safety nets indecentralized districts inUganda," Health Policy and Planning 17 (SUppl2)) 64-71 Kiyonga, C. 2001. " Policy Statement: Abolition o f User-charges and Introduction o f a Dual System" Uganda Health Bulletin, 7 (2) Konde-Lule, J. And D.Okello, 1998. User Fees in Government Health Units in Uganda: Implementation, Impact and Scope, Small Applied ResearchPaper No. 2. Bethesda, MD, Partnership for Health ReformProject, Abt Associates Lake, S. 2002. "Financing and delivering the goals o f the health sector strategic plan: Selected issues arising duringthe Joint Review o f 8-19, April 2002" Report prepared for Ireland Aid, Embassy of Ireland, Uganda. Lanjouw, P. and M.Ravallion, 1999. "Benefit Incidence, Public Spending Reforms and the Timingo fProgramCapture," WorldBankEconomic Review 13. Lindelow, M.,R. Reinikka and J. Svensson, ,2003, "Health Care on the Frontlines, Survey Evidence on Public and Private Providers inUganda," Africa Region Human Development Worlung Paper Series, World Bank Lubanga R, , 1997, "Maternal diagnosis and treatment o f fever inan endemic malaria zone of Uganda: implications for malaria control program," Acta Tropica Macrae, J., A Zwi, L.Gilson, 1996. "A Triple Burden for Health Sector Reform: Post Conflict Rehabilitation inUganda," Soc Sci Med 42(7) 1095-1108 Mbonye, A.K. ,2003. "Prevalence o f childhood illnesses and care-seelung practices inrural Uganda," Scientijk WorldJournal (3)72 1-30 144 Mbonye, A.K. ,2001. "Risk Factors Associated With MaternalDeaths inHealthUnitsin Uganda," African Journal of Reproductive Health Dec 5(3): 47-53. Mbonye, A.K., 2003. "Delivering Family PlanningMessagesthrough Prenatal Care Clinics in KumiDistrict, Uganda, 2003. ``International ElectronicJournal ofHealthEducation, 2003: 6 (34 -40). MinistryofFinance, PlanningandEconomic Development, Republic ofUganda., 2003a. "Background to the Budget Financial Year 2003/04" MinistryofFinance, PlanningandEconomic Development, Republic ofUganda, 2003b. "Uganda's ProgressinAttaining the PEAP targets and MillenniumDevelopment Goals'' MinistryofFinance, PlanningandEconomic Development, Republic of Uganda, 1999."Uganda ParticipatoryPoverty Assessment" Ministry of Finance, Planning, and Economic Development, Republic of Uganda, 2002a. "Infant Mortality inUganda 1995-2002: Why the Non-Improvement?", Discussion paper 6 MinistryofHealth, 1999-2003. Annual "HIV/AIDS Surveillance Reports,' STD/AIDS Control Program Ministry o f Health, Republic of Uganda, 2003. Mid-TermReviewReport: Health Sector Strategic Plan 2000/01-2004/05 Ministry o f Health, Republicof Uganda, 2003a. "Health Sector PEAP ReviewPaper-Draft" Ministry of Health, Republic of Uganda, 2003b. "Annual Health Sector Performance Report, Financial Year 2002103" MinistryofHealth, Republic ofUganda, 2003c. "Finance year 2003104 district transfers for health services" Ministry o f Health, Republic of Uganda, 2003d. "Status o f Emergency Obstetric Care inUganda: A NationalNeeds Assessment ofEmOC ProcessIndicators" MinistryofHealth, RepublicofUganda, 2003e.Presentation onthe Results fromthe Evaluation of Abolishing User Fees inHealthFacilities inUganda Ministryof Health, Republic ofUganda, 2003f. "Sub-National (District) HealthSystems Performance Assessment Report" MinistryofHealth, RepublicofUganda, 2003g. "Policy priorities for improvingRHservices including Emergency Obstetric care" Ministry of Health, Republic of Uganda, 2003h. "Midterm ReviewReport of the Health Sector Strategic Plan 2000-2005" MinistryofHealth, Republic ofUganda, 2003i. Strategy for PromotingHouseholdSanitation and Hygiene. Draft. November 2003. 145 MinistryofHealth, Republic ofUganda, 2002a. "Health Financing Strategy" MinistryofHealth, Republic ofUganda, 2002b. "Annual Health SectorPerformance Report, FY200112002" Ministry of Health, Republic of Uganda, 2002c. "Budget Framework Paper for the Health Sector 2002103 to 2004105" MinistryofHealth, Republic ofUganda, 2001a. "Guidelinesfor the ProvisionoftheUganda National MinimumHealth Care Package" MinistryofHealth, Republic ofUganda, 2001b. "HealthInfrastructureDevelopmentand Maintenance Plan" MinistryofHealth, RepublicofUganda, 2001c. "Annual Health SectorPerformanceReport, FY 2000l2001" MinistryofHealth, Republic ofUganda, 2000. "Health Sector Strategic Plan, 2000101 -2004105". MinistryofHealth, RepublicofUganda, 1999."National HealthPolicy" Ministry of Public Service, Republicof Uganda, 2003, "Study of Cost Efficiency and Effectiveness inHumanResourcesDevelopment inSocial Sectors" Okuonzi, S.A., 2002. "Improving Health Sector Performance: How Can Legislation Help?" Uganda HealthBulletin, 8 (1) Okuonzi, S.A.2001. "Your Money or Your Life: Myths and Realities of Cost-Sharing" Uganda HealthBulletin, 7 (2) Ower, R. 1987. "Report and Recommendations of the Health Policy Review Commission", MinistryofHealth. Oxfam, 2001. Food andNutrition Survey among the Internally DisplacedPopulation inKitakwi District, Uganda Pariyo, G.et al, 2003. "Interim Conclusions Based on Findingsfrom the Uganda IMCIImpact Study," Makerere Universityand Johns HopkinsUniversity PopulationReference Bureau, 2003. Country Profiles for Populationand ReproductiveHealth: Policy Development and Indicators, h-, accessed 2/20/04 Reinikka R and J. Svensson, 2002. "Working for God?Evaluating Service Delivery of Religious Not-for-Profit Providers inUganda," World Bank, WashingtonDC Reinikka, R.,2001. "Recovery in Service Delivery: Evidence from Schools and Health Centers" inRitiva Reinikka and Paul Collier, (eds.) Uganda's Recovery: TheRole of Farms, Firms, and Government,Regional and Sectoral Studies, World Bank RepublicofUganda andDevelopmentPartners, 2001-2003. Bi-annual Aide-Memoire from the Joint ReviewMissions 146 Root, G., A. Collins, K.Mungutiand K.Sargent, 2003. "Uganda RollBack Malaria Scoping Study," Malaria Consortium Ssengooba, F.,L.Atuyambe, B.McPake, K.Hanson, and S Okuonzi, 2002. "What canbe Achieved with greater Hospital Autonomy? Comparison of Public and PNFP Hospitals in Uganda," Public Administration and Development 22: 415-428 Steen, S., P. Jareg, I.T.Olsen, X. Mugisha, M.Lubaale and S. Baine, 2001. "Providing a Core Set of Health Interventionsfor the poor-a Systematic Approach" HeSo, Norway Synergy Project, 2003, "Uganda Country profile: HIVIAIDS," USAID Twebaze D, 2001. "A Literature Review of Care Seeking Practices for Major Childhood Illnesses inUganda," Basic Support for InstitutionalizingChild Survival (BASICS), Arlington, VA Uganda Bureau of Statistics, Republicof Uganda, 2004. htt~:l/www.ubos.ora/2OO2census.html, accessed 3/17/04 Uganda Bureauof Statistics, Republic of Uganda, 2003. UgandaNationalHousehold Survey 2002. Uganda Bureau of Statistics, Republic of Uganda, 2000. UgandaNationalHousehold Survey 1999. Uganda Bureauof Statistics and ORC Macro, 2001. Uganda Demographic andHealth Survey 2000-2001. Uganda Bureauof Statistics and ORC Macro, 1996. Uganda Demographic and Health Survey 1995-1996. Uganda Bureauof Statistics and ORC Macro, 1989.Uganda Demographic and Health Survey 1988-1999 UNAIDS,2004. UNAIDSonline database www.unaids.org, accessed February 2004 UNHCO, 2002. "Baseline Survey on Patient's Rights," Uganda NationalHealth UserdConsumers' Association, Kampala, Uganda UNICEF, 2003. Family Care Practices inUganda: Summary Report UNICEF, 2002. "UNICEF HumanitarianAction, Uganda, DonorUpdate" UNICEF, 2001."The 2001Baseline Study for the 2001-2005 Government of Uganda-UNICEF Country Programme: NationalReport USCensusBureau, 2003. http://www.census.gov/ipc/hiv/uaanda.pdf, accessed3/23/04 USAID,2002. What happenedinUganda?DecliningHIVPrevalence, Behavior Change, and the NationalResponse 147 van de Walle, Deminique, 1998. "Assessing the Welfare Impacts of Public Spending," World Development 26 (3) World Bank, 2003a. "The Republic of Uganda Public Expenditure Review 2003" World Bank, 2003b. WorldDevelopment Report 2004: Making Services Workfor the Poor, World Bank, Washington DC World Bank, 2002. WorldDevelopment Indicators World Bank, 1988. WorldDevelopment Report World Health Organization, 2003. Stop TB,"Global Tuberculosis Control" World Health Organization, 1999. "Health Reforms and Priority Health Interventions: the Case of Immunization Services," Department of Vaccines and Biologicals. World Health Organization, 2002. World Health Indicators World Health Organization, 2000. The WorldHealth Report. Health Systems. Improving Performance World Health Organization, SIDA, and Ministryof Health, 2003. "The effects of abolition of Cost-sharing inUganda," Report on findings for 2002 148 =Ia P1!93 zw o- w m 1 ' 4 r - b b 0w -o 0; 9m m * m "Z?Z e w r - r - m O 2 2 0 * * * * * 0 0 6 ? - ? 2 2 2 O d h l 0 0 2 2 3 0 0 0 * w o w Y Y ? 2 2 w o 2::* m w m 0 0 0 y m o c ? ? 3 2 0 0 0 0 0 2 5 0 0 0 2 ** * * 2 2 w l + - 0 - m 9 9 9 Y 0 0 0 0 ** s g s a 0 0 0 0 * * h$E 2w 2 Z?Z m - a 2 3 0 0 * N a m 0999 N c? 2 2 N - **z ** 2 0 - 0 N 0 0 P L " cO? cW ? 0L" m - 8 8 Q Q N ? 9 p ! 0 0 0 0 0 0 0 ** ** ** d. c? m o * z i c q 2m 8N 2o 2 2 O W N 0 2 2 4 2o 2m;w 8 8 a m z z b , 0 0 0 0 0 * * * * N 2 z. 2. ** ** N - P --him \ 9 y 3 2 0 0 N v, 0 2 2 2 N Q P - m - 9 2 2 2 0 0 c0? 0 L0" t 0 0 0 0 0 * ** * * ** * m m Y L " ? b b - P 9 d. 2 2 Q b T p ! ? ~ m 2m 2 - 0 0 2 3 - 0 n ** I 3 0 3 3 3 s m 8 I 0 2 W 2 * * * * 0 2 0 0 0 2 8 - W 2 M d 2 2 ; 2 2 2 0 0 0 0 0 * ** * m 2 * m d W '"099 - - m 2 00 2 2 q - m 0 0 0 0 0 c 8 ** b m m M e 2 2 z h e, g m 2 - o r 4 9 9 9 0 0 0 00 - 9 2m 9o 2a - * x d 8 8 % - o m * -? *m 0 \o9 9 m o - - Ifi 8 8 8 0 0 - 0 VI **m o o 3 9 9 9 3 0 - - wl .wl 3 Appendix 5: Trends insocioeconomicand demographic characteristics of under-fivestuntingand underweight 1988DHS 1995DHS 2000DHS Overall 44.4 19 23 4.814 38.3 15 25.5 6.681 38.6 14.8 22.5 4.891 Residence Rural 46.3 19.9 24 5.189 40.3 15.9 26.8 7.146 39.9 15.6 23.6 5.217 Urban 25.3 9.4 12.6 1.618 22.5 7.9 15.3 3.014 26.5 7.1 12.4 1.804 Region Central 33.5 11.1 21.1 4.829 34.6 12.6 19.9 4.414 Eastem 35.6 14.3 27.3 1.014 35.4 12 22.5 4.121 Northem 41.9 17.8 31.6 9.099 36.9 14.6 25 6.497 Westem 42.8 11.4 23.8 6.321 47.8 20.7 23.7 5.226 Wealth quintile Poorest 48.4 22.3 25.5 6.491 45.8 19.9 32.1 8.093 41.4 11.7 27.5 6.125 2nd poorest 51.1 24.2 29.2 6.807 42.5 16.3 26.6 8.103 39.6 15.5 24.6 5.922 Middle 41.2 11 23.5 4.314 38.9 13.6 27.1 1.511 43.4 11.3 25.9 5.264 2nd richest 40.1 16.4 18.9 2.629 38.5 15.7 23.6 5.658 38.8 14.2 19.9 3.998 Richest 30.9 12.9 15.6 3.203 25 9.2 16.5 3.841 26.6 1.2 11.2 2.088 Religion Catholic 43 17.1 21.5 4.774 38.9 15.6 21.4 1.653 39.9 15.8 22.2 5.409 Protestants 46.8 20.8 23.8 4.715 38.4 14.9 24.5 5.426 37.2 13.5 21.3 4.4 Muslidothers 41.9 19.3 25.2 5.464 36.1 13.9 23.6 7.217 38.8 15.5 25.7 4.801 Age category e20 44.8 19 22.3 4.133 41.3 15.5 25.3 5.72 43.6 17 23 3.981 20-29 45.1 20.2 24.2 5.441 38.3 14.1 24.1 6.396 38 14.2 21.5 4.419 30-39 42.6 16.6 21.3 4.338 35.2 15 21.7 8.218 31 14.2 24.1 6.213 40-49 35.7 16.9 21.1 3.834 41.8 18.3 26.4 6.819 32.1 11.9 24.8 8.124 Mother's education None 41.5 22.1 26 4.101 43.1 18.4 29.1 8.416 45.5 18.5 28.6 7.721 Primary 44.1 11.9 22 5.346 38.2 14.1 25 6.492 31.1 14.3 21.5 4.263 Secondary or higher 26 9.4 13.9 2.784 25.7 10.1 16.6 2.622 28.9 9.7 15.2 2.312 Parity 0-2 42.6 18 22.2 4.662 31.5 14.5 26 6.542 31 14 20.2 4.227 3 to 4 47.4 18.5 23.2 4.598 39.2 15.1 24.4 5.922 39.7 14.8 22.8 4.214 5+ 43.3 20.2 23.5 5.28 38.4 15.7 26 1.553 38.9 15.5 24.2 6.09 Birthinterval 1sti<2yr 46.1 20.9 25.3 5.096 39.3 15.5 25.8 6.165 40.8 16.4 22.6 4.982 2-3 YS 43.9 18 21.9 5.08 31.7 14.6 25.3 7.148 38.4 14.5 23.3 4.899 4+years 31.4 16.1 19.1 2.862 31.1 15 25.1 6.512 31.2 10.3 18.8 4.553 Sex of the child Female 41.6 18 23.1 4.188 36.1 14.1 24.1 5.848 36.9 13.8 21.4 4.32 Male 47.3 20 22.9 4.963 40 16 21.1 1.553 40.4 15.8 23.1 5.469 Age of the kid 0-11 months 21.3 5 13.4 1.838 11.7 4.2 11.6 5.337 18.4 4.6 16.6 3.417 1 years 52.9 20.2 31.5 8.586 45.4 18 35.4 9.431 46.3 19 32.1 7.56 2 years 52.2 25.4 26.2 7.057 44.9 20.1 26.6 1233 41.6 17.1 24.1 6.114 3 years 50.1 24 21.8 2.908 49.8 20.5 20.9 3.843 44 15.1 18.2 3.875 4 years 52.6 26.1 23.2 3.133- __ _- 44.2 18.5 19.3 2.768 Source: DHS 1988, 1995 and 2000, Authors' calculations 156 3 2 2 2 2 s 2 2 2 09' g z m a 0 0 0 0 0 0 0 0 0 0 3 2 *w 2 2 m 2 2 2 2 $ 2 0 0 0 0 0 0 * * * ** * 3 2 Z Pc W ? " 2 z q 0 0 0 0 0 0 z * * 3 m m Y t - 3 3 m h l m m - w 3 3 -! 2 6 : 3 3 3 0 0 0 9 0 0 0 0 0 0 0 ** * m m m % ? v i 0 0 - Z f f O0 0 0 0 8 $ 2 Z 2 2 G 0 0 0 0 0 0 * * ** * .- : I Y r 0 .I m * ** * 8 * e h { L U .I M 0 I h 0 3 3 2 2 53 2 F 2 m O h 1 Q O 0 0 0 0 0 0 0 8 c ? c ! ? c ! 0 0 0 0 *d * * ** * * e * ** ** 0 c! 22: m w 2 3 3 2 z mc! d- "vr m O v r Q 0 0 0 +. 3 3 c! 0 0 0 0 8 c0? ?0" 0 * * * * * * Q c? hl c! 4 3 z0 z0 m c! hl 2 2 2 O P P 3 0 "I 3 0 0 0 0 c0 ?0" t0 0 * ** 3 d + 9 3 -7 W 0 0 0 0 8 * * a * * * * ** *v?r*?w *9 *vr * * ** * * * * % * vlhlb \4'?"! Q\ r! W Q W 2 2 2 - 3 0 0 0 ri 3 3 3 3 2 z 2 d 2 2 ; r - w m Q\ W Q Q 0 0 0 "I c ! c ! c ? " ' ? ? 0 0 0 0 0 0 0 8 0 0 0 *P ** U v r -7 ? ? F c w m m **m 09099 h l P m c ! c ? c ? c! 0 0 0 ri 0 0 3 3 3 3 i m m P m e - 3 2 2 2 2 0 0 0 c 0 0 0 0 0 * *a * * * * * * hl *m c O v l d + ' " t F c! - 3 3 I k hv1 e4 Appendix 7: Socioeconomicand demographic characteristics of childhood fever, ART, and diarrhea DHS 1988 DHS 1995 DHS 2000 Fever Diarrhea ARI Fever Diarrhea ARI Fever Diarrhea ARI Overall 41.2 24.2 na 46.5 23.6 27.2 44 19.6 22.5 Residence Rural 42.2 24.7 na 48 24.1 28.3 45 4 20.1 23 Urban 31.5 19.9 na 35.4 19.6 19.1 33 15.6 18.6 Region central 39.1 19.8 na 39.4 16.4 21.5 38 14.5 19.5 eastem 66.5 33.3 na 59.4 26.3 23.2 54.1 23.4 23.4 northem 31.9 21.4 na 58.4 34.3 30.8 50.5 26.8 23.2 westem 25.5 21.5 na 32.8 20.1 33.9 33.8 16 24.6 Wealth quintile Poorest 38.7 26 na 50.3 30 33.4 52.3 27.1 26 2nd poorest 49 25.7 na 43.7 22.5 28.1 46.6 23.3 24.8 Middle 39.4 23.4 na 51.3 26.8 31 40.2 14.4 21.7 2nd richest 41.4 22.9 na 47.3 20.6 24.9 42.4 16.5 20 Richest 36.6 22.2 na 39.7 17.7 17.7 35.1 13.9 18.5 Religion Catholic 40.8 23 na 45.9 22.1 27 42.6 19.4 22.5 Protestants 40.7 24.6 na 45.3 22.9 28.4 44.3 19.9 22 Muslidothers 43.8 26.8 na 51.3 28.8 24.8 45.9 19.6 23.5 Age category <20 37.8 23 na 48.5 24.1 26.9 41.9 18.8 20.4 20-29 41.8 23.9 na 45.8 23.7 26.9 43.2 19.5 22.1 30-39 42.7 25.5 na 46.4 23.2 28.4 47.8 20.2 23.5 40-49 41.5 26.5 na 47 19.3 26 42.1 23.4 35.8 Mother's education None 39.4 23.3 na 48.6 26.6 21.8 43 21.1 23.3 Pnmary 44.5 25.8 na 46.5 23.3 28.7 45.1 20.3 22.7 Secondaryor higher 30.4 19 na 40.7 16.6 17.1 36.4 12.8 20.1 Parity 0-2 41.4 26.2 na 46.7 25.6 21.9 43.2 21.8 22.5 3 to 4 41.1 24.6 na 46.8 23.5 26.2 42.3 19.6 21.2 5+ 41.1 22.2 na 46.1 21.3 27.3 46.5 17.5 23.9 Birthinterval IStk2yr 37.7 22.4 na 46.2 23.2 28.2 42 19.4 20.5 2-3 y r ~ 43.7 25.5 na 46.6 22.9 25.9 45.6 19 23.7 4+years 43.3 24.5 na 41.9 27.9 28.8 43.8 23.7 24.8 Sex of the child female 41.1 23.1 na 44.8 21.9 26.3 43.2 18.8 22.2 male 41.2 25.3 na 48.4 25.3 28.2 44.8 20.5 22.8 Age of the kid 0 44 35.2 na 47.3 26.1 31.6 44.9 28.3 27.6 1 49.2 39 na 55.6 34.4 30.8 57.5 29.4 28.5 2 40.2 19.9 na 42.3 17.2 23.8 44.6 18 21.8 3 39.7 13.3 na 37.7 12.3 20.2 40 11.1 18.6 4 29 7 . 7.2 na -- __ -- 29.9 7 9 .._ 13.9 Source: DHS 1988, 1995 and 2000., authors' calculations Note: na=not available 159 E: s 2 G E: 2 0 0 0 2 2 2 0 0 0 0 0 0 * * * t- 2 2w 2 s o o \ 2 2 z 2 2 2 0 0 0 0 0 0 z0 20 q0 2 * * 0 2 3 P 3 2 2 2 2 0Np! m 0 0 0 0 0 0 0 0 0 2 st- 2. m b N 9 - 0 0 o ? ? 3 2 % q 9 20 20 2 2 2 % 2 s 0 0 0 0 0 0 0 0 0 0 2 2 % 0 0 0 o\ 8 2 ? % 0 0 0 0 2 *:: * : * h Q) : *: * * * * Appendix 9: Childhoodimmunizationinchildrenaged12-23 monthsby various so-cioeconomicand demographiccharacteris :s in 1988,1995 and 20001 [S surveys 1988DHS 1995 DHS 2000 DHS Ever AllEPl Ever AllEPI Ever AllEPI DPT DPT vacc- vaccin- DPT DPT vaccin- vaccin- DPT DPT vaccin- vaccin- Measles 1 3 inated ations Ieasles 1 3 ated ations seasles 1 3 ated ations Overall 11.5 95.2 51 11.5 41.1 59.1 81.3 61.4 85.9 41.4 51 16.9 46 3 90.6 36 I Residence Rural I O 95 53.1 69.8 44.5 51.8 80.1 59.1 84.8 46.3 55.5 15.4 44.1 90.1 36.1 Urban 83.9 96.6 85.1 81.7 15.9 14.6 91.1 15.4 94.7 56.1 68.4 88.4 59.4 94.1 42.1 Wealth quintiles Poorest 63.4 91 45.5 59 35 49.3 15.2 51.9 81.8 31.4 45.4 69.3 31.4 89 23.2 2ndpoorest 68.5 94.6 41.9 66.1 39.9 51.9 81 59.2 83.8 48.1 59.3 17.1 41.3 90.7 39.2 Middle 15.1 89.5 52.1 12.4 48 54.3 80 56.2 84.9 42.5 59.6 18.3 50.2 90.1 38.6 2nd richest 15.4 96.8 61.3 81.6 54 63.1 81.3 64.2 86.4 49.9 61.3 11.2 50.1 89.3 42 Rxhest 15.8 91.4 80.4 83.4 63.1 15.1 89.8 16.1 93 60.4 62.3 85.3 56.6 94.3 44.2 Region Central 66 86.3 70.8 87.5 53.4 50.9 68.2 38.3 82.2 29 Eastern 48 16.9 49.1 85.2 34.4 53.2 18.3 44.8 95.5 31.8 Northern 51.5 16.3 48.4 84.1 34.1 51.9 18.3 44.9 89.7 33.2 Western 12 85.2 14.1 86 65.1 61.6 83.8 51.9 94.9 46.4 Religion Catholic 14.9 96.9 50.4 14.2 42.2 59.6 81.6 60.8 86.2 46.7 56.9 14.9 44.5 89.5 35.5 Protestants I O 91.1 62.5 I O 53.7 61.5 82.2 62.9 86.1 50.5 51.9 78.8 48.4 91.5 38.2 Muslimdothers 62.4 100 64.9 61.3 50.1 55.6 18.9 59.4 84.6 41.9 55 11.2 45.1 91.2 36 Mother's education None 66.5 93.9 42.9 62.7 35.1 49.1 11.9 41.1 18.9 38.3 54.3 10.8 31.1 88.3 28.4 Primary 11.9 95.2 60.2 15.3 49.1 60.8 83.4 64.6 81.4 48.4 55.6 11.2 41.1 90.6 31.2 Secondary 88.4 100 92.9 91.5 83.2 83.8 91.5 83.5 91.8 68.1 69.1 81.3 60.8 95.2 51.1 Mother's age <20 69.1 95.5 53.8 10.6 45.8 55 80.2 60.2 84.4 42.8 51.4 14.8 45.8 92.3 35.4 20-29 12.8 95 62.3 14.4 51.1 63.5 84.3 63.1 88.3 50.4 56.6 11.6 46.5 89.1 31.4 30-39 69.2 96 46 63.1 38.8 51.5 17.1 59.5 83.1 46.1 65.3 79.9 49.8 91.6 39.2 40-49 13.4 91.2 58.6 80.2 58.6 43 10.7 44.3 10.7 34.2 31.4 53.8 18.8 89.3 15.4 Parityof the child first 10.3 93.5 53.8 14 45.3 64 83.1 63.5 81.4 46.3 59.2 I1 53 93.1 42.4 2 or 3 16.4 96.3 63.2 15.1 53.6 62.3 85.6 66.3 88.1 50.9 56.4 19 46.1 91.2 36.1 4 or 6 68.6 95.2 59.5 10.8 48.6 58.6 80.2 60.3 86.1 41.8 54.2 15.3 41.6 88.3 32.8 I+ 68.4 94.1 46 64.2 38.9 52 13.1 52.1 19.1 41.1 60.9 15.5 49 91.3 39.8 Sex of the child Female 69.5 96.8 55 69.1 45.1 59 79.2 59.5 84.8 46.6 51.5 16.3 41.9 89.4 3 1 Male 13.3 93.6 59 13.4 50.3 60.4 83.6 63.4 81.1 48.3 56.5 11.4 44.1 91.8 36.5 Source: DHS 1988, 1995,2000 and authors' calculations 162 Appendix 10: Resultsof multivariatelogisticregressionfor DPT-1andDPT-3 immunizationin 1995 and2000 DPT 1 DPT3 1995 2000 1995 2000 OR 95% C I OR 95% CI OR 95%CI OR 95YuCI Urban(rural) 1.52 0.85 2.72 1.52 0.85 2.72 1.00 0.62 1.61 1.53 0.94 2.50 Region(Central) Eastem 0.65 0.40 1.07 0.65 0.40 1.07 0.49*** 0.32 0.73 2.11*** 1.30 3.43 Northem 0.75 0.41 1.36 0.75 0.41 1.36 0.54** 0.31 0.94 3.12*** 1.87 5.20 Westem 1.30 0.71 2.40 1.30 0.71 2.40 1.87** 1.12 3.14 3.36*** 2.21 5.10 Wealth quintile (poorest) 2nd poorest 1.23 0.73 2.06 1.23 0.73 2.06 1.22 0.78 1.91 1.99*** 1.30 3.06 Middle 1.24 0.73 2.12 1.24 0.73 2.12 1.27 0.82 1.97 2.53*** 1.51 4.25 2nd richest 1.11 0.65 1.89 1.11 0.65 1.89 1.54* 0.98 2.42 2.33*** 1.43 3.80 kchest 1.49 0.83 2.67 1.49 0.83 2.67 2.19*** 1.33 3.59 3.11*** 1.71 5.66 Religion (catholic) Protestants 0.92 0.64 1.33 0.92 0.64 1.33 0.89 0.66 1.21 1.16 0.88 1.53 Muslims 0.88 0.55 1.41 0.88 0.55 1.41 1.01 0.64 1.60 1.15 0.78 1.68 Education (None) Primary 1.77*** 1.28 2.45 1.77 *** 1.28 2.45 1.86*** 1.32 2.60 1.57** 1.12 2.18 Secondary 11.42 *** 4.04 32.26 11.42*** 4.04 32.26 4.32*** 2.48 7.53 2.20** 1.32 3.64 Mother age group (15-19 years) 20-29 years 1.56 0.92 2.64 1.56 0.92 2.64 1.31 0.88 1.94 1.14 0.73 1.77 30-39 years 1.75 0.86 3.57 1.75 0.86 3.57 2.16** 1.19 3.90 1.03 0.56 1.88 40-49 years 1.70 0.56 5.12 1.70 0.56 5.12 1.48 0.54 4.02 0.28** 0.09 0.89 Birthorder (first) 2-3 order 0.83 0.51 1.33 0.83 0.51 1.33 0.81 0.52 1.26 0.76 0.49 1.17 4-6 order 0.53 0.28 1.03 0.53 0.28 1.03 0.54** 0.33 0.88 0.66 0.40 1.09 +7 order 0.37** 0.17 0.80 0.37 ** 0.17 0.80 0.32*** 0.17 0.62 1.17 0.63 2.19 Birthinterval (first or <2year) 2-3 years 1.47 1.01 2.14 1.47 1.01 2.14 1.48** 1.09 2.01 0.96 0.72 1.28 4+ years 1.15 0.65 2.04 1.15 0.65 2.04 1.41 0.87 2.27 1.13 0.74 1.72 F( 20, 263) 3 4 6 5 Prob > F 0 ' 0 0 0 ** p< 0.05; *** p chi2 0.00 0.00 0.00 PseudoR2 0.06 0.04 0.06 Logpseudo-likelihood -2632.37 -2226.24 -1953.13 *p<0.05 **p chi2 0 0 0 PseudoR2 0.1616 0.065 0.0888 Logpseudo-likelihood -1158.48 -1502.79 -1 106.87 *p<0.05 **p chi2 0.00 0.00 0.00 Pseudo R2 0.21 0.18 0.19 Logpseudo-likelihood -394.62 -2524.79 -3076.60 *p<0.05 **p