world bank document - documents & reports€¦ · world bank or any of its affiliated ... hope...

VV~~~~~~~~~W R Kn Nv G nP rAP r C SE R I CES

THE WORLD BANK

Better Hedith Outcomesfrom Limited Resources

Focusing on Priority Servicesin Mdldwi

-O C j

-E ill NJ -

j = X x \ __ K vY <- _

Ef \, :r~\- tl! I " -_ j0'

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Other Titles in This Series

Dynamic Risk Management and the Poor-Developing a Social Protection Strategy for Africa

Engaging with Adults-The Case for Increased Support to Adult Basic Education in Sub-Saharan Africa

Inclure les adultes - Pour un appui a l'education de base des adultes en Afrique subsaharienne

Enhancing Human Development in the HIPC/PRSP Context-Progress in the Africa Region during 2000

Early Childhood Development in Africa-Can We Do More for Less?A Look at the Impact and Implications of Preschools in Cape Verde and Guinea

AIDS, Poverty Reduction and Debt Relief-A Toolkit for Mainstreaming HIV/AIDS Programsinto Development Instruments

Systemic Shocks and Social Protection-Role and Effectiveness of Public Works Programs

Social Protection of Africa's Orphans and Vulnerable Children-Issues and Good PracticeProgram Options

Can Africa Reach the International Targets for Human Development? An Assessment ofProgress towards the Targets of the 1998 Second Tokyo International Conference onAfrican Development (TICAD II)

Education and Training in Madagascar-Towards a Policy Agenda for Economic Growth and PovertyReduction A SummaryEducation et Formation a Madagascar - Vers une politique nouvelle pour la croissance 6conomiqueet la r6duction de la pauvret6 Un r6sum6

Issues in Child Labor in Africa

Community Support for Basic Education in Sub-Saharan Africa

Le systeme 6ducatif mauritanien - El6ments d'analyse pour instruire des politiques nouvelles

Rapid Guidelines for Integrating Health, Nutrition, and Population Issues into the PovertyReduction Strategies of Low-Income CountriesInt6grer les questions de sante, de nutrition et de population aux strat6gies de reductionde la pauvret6 dans les pays a faibles revenus: quelques directives rapides

Deux 6tudes pour la scolarisation primaire universelle dans les pays du Sahel en 2015

Improving Health for the Poor in Mozambique - The Fight Continues

Skills and Literacy Training for Better Livelihoods - A Review of Approaches and Experiences

The Impact of Adult Mortality on Primary School Enrollment in Northwestern Tanzania

Africa Region Human DevelopmentWorking Paper Series

Better Health Outcomesfrom Limited Resources:Focusing on Priority Services in Malawi

Oscar F. Picazo

Africa RegionThe World Bank

ii AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES

© April 2002Human Development SectorAfrica RegionThe World Bank

The views expressed herein are those of the author anddo not necessarily reflect the opinions or policies of theWorld Bank or any of its affiliated organizations.

Cover photo by Oscar Picazo.Cover design by Tomoko Hirata.

BETTER HEALTH OUTCOMES FROM LIMITED RESOURCES

Contents

1 Introduction 1

2 Key findings and recommendations 3

The resource envelope and health sector performance 3The government budget 5Extra-budgetary sources offunds 7Private health expenditures 8Donorfinancing and expenditures 9Sectoral efficiency improvement 10

3 MOHP budget and expenditures 13

Allocation trends 13MOHP's Medium-Term Expenditure Framework 17Continuing challenges in the MTEF process 19Recommendations for improved resource allocation and use 21

4 Extra-budgetary sources of funds 23

Drug revolvingfunds 23Cost-sharing programs 23

5 Private health expenditures 26

Mission/CHAM facilities 26Private for-profit health providers 27Health insurance 27Household health expenditures 29

iV AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES

6 Donor financing and expenditures 31

Allocation vs. expenditures 32Project pipeline 34Health servicefocus of donor projects 34Uses of donor resources 34Geographic distribution of donor projects 38

7 Analysis of the Fourth National Health Plan 40

Infrastructure implications 40Personnel implications 41Recurrent cost implications 42Total cost of the NHP and optionsfor phasing 43

8 Efficiency improvement in the health sector 45

Hospital autonomy 45Improving pharmaceutical financing, distribution and use 47Health service decentralization 49Contracting of health services 52The role of nongovernmental organizations 53

Appendix - Unit cost of providing health services to the top five diseaseconditions in Malawi 55

References 59

List of Tables

1 Key socioeconomic and health indicators in Malawi: various years 22 MOHP recurrent and donor expenditures in the health sector: FY94/95-FY98/99 43 Malawi health spending and performance compared to other countries in Sub-Saharan Africa:

various years 54 MOHP recurrent expenditures and per capita expenditures, in current and real terms:

FY95/96-FY98/99 145 MOHP recurrent expenditures by level of institution: FY95/96-FY98/99 156 MOHP recurrent expenditures by level of cost center: FY98/99 167 MOHP recurrent expenditures by economic classification: FY95/96-FY98/99 178 MOHP staff breakdown: FY97 189 Recommended key actionsfor improved use of MOHP resources 2210 Recommended restructuring of selected hospital services in Malawi 2511 Estimated revenues from cost-sharing program based on alternative scenarios 2512 Health service providers in Malawi: mid-1990s 2613 Benefits and contribution rates of different types of health insurance plans offered by

the Medical Aid Society of Malawi 2814 Operating indicators of the Medical Aid Society of Malawi: 1993/94-1996/97 28

BETTER HEALTH OUTCOMES FROM LIMITED RESOURCES V

15 Annual household health expenditures in Malawi: 1990-91 2916 Reported paying patterns of poor and nonpoor households for health services

by urban and rural location: 1995 3017 Summary of donors' health sector allocations and expenditures in Malawi: FY94/95-FY97/98 3118 Budget allocation and expenditures of donor projects: as of end-FY97/98 3219 Donors' budget allocation and expenditures in the health sector in Malazvi: FY94/95-FY97/98 3520 Pipeline analysis of major donor projects in the health sector in Malawi: FY98/99-FY01/02 3621 Donor projects by health servicefocus in Malazvi: as of end-FY97/98 3622 Donor expenditures by major classification: FY96/97 and FY97/98 3723 Implementing agencies of donor projects in Malawi: as of end-FY97/98 3824 District location of donor projects in Malazvi: as of end-FY97/98 3825 Current vs. proposed vs. standard number of health facilities in Malawi 4026 NHP's "priority" infrastructure program and cost 4127 Current vs. proposed vs. standard number of health personnel in Malazvi 4128 Current vs. proposed number of personnel for health centers and district/rural hospitals in Malazvi 4229 Estimated recurrent cost requirements of the NHP's infrastructure program 4330 MOHP contracting out and rationalization of governmentfunctions 52

Appendix tables

1-A Average cost of care per patient, by type of care, by type of disease, and by type offacility: 1998 561-B Average cost of care per inpatient and per OPD visit by disease: 1998 57

BETTER HEALTH OUTCOMES FROM LIMITED RESOURCES Vii

Foreword

n order to identify appropriate policy and management. Reform in these areas are key toadministrative measures to address develop- improving the quality of health services in Malawi.ment issues in the health sector, it is helpful to The Government of Malawi has committed itself

have access to country-specific knowledge that to addressing the deficiencies in the health systemsheds light on the key weaknesses in the health under its proposed debt relief program. In order tosystem. In a large number of Sub-Saharan countries, do so, an infusion of additional resources, from bothhowever, such a knowledge base remains sparse, domestic and international sources, may well bereflecting the absence or sometimes systematic neg- needed. However, an equally important issue islect of analytical work in the past. The present ensuring that currently available resources are usedreport deals with health financing issues in Malawi efficiently and equitably The evidence suggests thatand analyzes trends in health expenditures in the in Malawi, as in many Sub-Saharan countries, sub-1990s, along with the prospects for improving stantial scope for progress exists in this regard. Theresource mobilization, allocation and use in the Government of Malawi indeed has begun formulat-health sector of that country. ing a reform program to improve the performance

Malawi's major public health issues include of the health system, including a "fast-track"HIV/AIDS, poor reproductive health and severe approach to produce more trained health workersconstraints in the availability of health personnel, (especially nurses), restructure the pharmaceuticaldrugs and other supplies, in conjunction with a distribution system, increase government financingvery limited capacity of the Government to define of key health sector recurrent inputs, gradualand implement good health policies. The most decentralization of health services, and a more coor-pressing challenges include defining a cost-effective dinated approach to donor assistance in the healthand sustainable package of health services that the sector. However, much more remains to be done,Government can commit itself to finance, reaching a particularly in tackling the institutional andconsensus on an appropriate division of responsi- human-resource capacity constraints.bility between the public and private sectors for The publication of this health expenditurefinancing and delivering health services, setting review for Malawi is intended to contribute to oursustainable levels of health worker remuneration, collective knowledge about the country's healthand creating institutional arrangements for efficient sector and the nature of the policy challenges, and

Viii AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES

to share that knowledge where possible. It is my Ok Pannenborghope that as new knowledge emerges in the course Senior Health Advisor and Sector Leader forof implementing the country's poverty reduction Health, Nutrition and Populationstrategy, this knowledge will be instrumental in Human Developmentovercoming the constraints in the health sector that Africa Regioncurrently impede poverty reduction in Malawi.

Better Health Outcomesfrom Limited Resources:Focusing on Priority Services in Malawi

Africa Region Human DevelopmentWorking Paper Series

Introduction

M alawi is one of the poorest countries in the AIDS epidemic. The economic slowdown thatIA world, with a per capita income of around began in 1994/95 and the continuing decline in the

US$180 in 1996. Its demographic profile international price of Malawi's key export (tobacco)exhibits a population that is high growth (2.8 per- on which a large proportion of Malawians depend,cent a year), young (47.5 percent are below 15 years have added to the complexity of making appropri-of age), mostly rural (77.7 percent), and has a very ate recornmendations to improve health sector per-high dependency ratio (97 dependents for every 100 formance.adults of working age). Malawi's public spending This health expenditure review (HER) provideson health has historically been high (at least until occasion to take stock of Malawi's performance in1993/94) relative to other Sub-Saharan countries the health sector. The paper reviews the status of theand developing nations with comparable GNP per country's health expenditures, identifies issues oncapita, but the country's living conditions are the level and quality of these expenditures, and pro-among the poorest in the world. Although physical vides recommendations to improve resource mobi-access to a health facility has improved over the lization, resource allocation, and organizational effi-years, access to functional health services continues ciency. Section 2 of the paper summarizes the keyto be limited as indicated by low provider-to-popu- findings and recommendations. The next four sec-lation ratios and often severe unavailability of tions deal with the various sources of health financ-drugs, contraceptives, and other supplies. As much ing: section 3 on government health expenditures,as 54 percent fell below the given household- section 4 on extra-budgetary funds, section 5 on pri-income poverty line in the mid-1990s. There is no vate sector financing, and section 6 on donor financ-recent update on poverty, but given the poor eco- ing. Section 7 analyzes the implications of thenomic performance throughout the 1990s, one can Malawi National Health Plan while section 8 exam-infer that the situation has not improved dramati- ines reform proposals related to sector efficiencycally. Life expectancy at birth has fallen from 45 and improvement.years in 1982 to 42 years in 1998, due largely to the

2 AFRICA REGION HUMAN DEVELOPMENT WORKING PAPER SERIES

Table IKey socioeconomic and health indicators indicators in Malawi: various years

Indicators Latest Year Available

Total population mid-1999 10.8 million

Avg. annual growth of pop. 1990-1998 2.8 percent

Percent urban population 1998 22.3 percent

Growth of urbanization 1990-98 9.5 percent

Total fertility rate 1998 6.4 children

Crude birth rate 2001 47 per 1,000 population

Crude death rate 2001 23 per 1,000 population

Life expectancy at birth 1998 42 years

Infant mortality rate 1998 134 per 1,000

Under-5 mortality rate 1998 229 per 1,000

Matemal mortality rate 1999 620 per 100,000

Adult HIV-1 seroprevalence per 100 adults End-1999 16.1

Children 0-1 immunized against DPT 1998 96 percent

Children 0-1 immunized against measles 1998 90 percent

Oral rehydration therapy use among under-5 children 1992-93 50 percent

Percent of infants with low birth weights 1991-94 10 percent

Population per physician 1990-96 33,344

Population per hospital bed 1990-96 625

Percent of births attended by trained health personnel 1990-95 57 percent

GNP per capita (Atlas dollars) 1999 US$180

Public health expenditures as percent of GDP 1990-97 3.3 percent

Source: World Bank African Developnment Indicators 2001.

2

Key findings andrecommendations

T his review highlights the need to further pri- devoted to health which compared favorably withT oritize the activities under the Malawi 4.2 percent for similar countries in Sub-SaharanNational Health Plan so that the plan will be Africa and 4.8 percent for developing countries with

a basis for government policy and budgetary com- similar GNP per capita. The economic contractionmitments and also an instrument to marshal and from 1994-95, however, ushered in the budget crunchorchestrate donor support to the sector. Once the that continues to this day, dramatically altering thegovernment has determined that its core function financing picture. Since FY94/95, government healthunder the Plan is the provision of an essential pack- spending has barely kept pace with inflation andage of health services, it needs to translate this population growth, with real per capita expenditurespolicy aim into budgetary allocations at both the of the Ministry of Health and Population (MOHP)central and district levels. It also needs to ensure actually declining from MK (Malawi Kwacha) 45.12that inputs, especially drugs, medical supplies, and to only MK40.91 in FY98-99 (Table 2). Investmentstrained staff, are made available to support the made in the first half of the decade (expansion ofdelivery of the package. Health services not physical infrastructure) are now wanting in recurrentincluded in the package should be subject to fees, costs. The government was slow to tap extra-budget-either on a modest cost-sharing or on a full-cost ary sources of funds (user fees) to arrest the declinerecovery basis. The fiscal crisis has underscored the in real per-capita public spending; it was also slow toimportance of fee revenues to cushion the impact of reconfigure its budget to focus on those services thatdeclining budget allocations to health facilities. provided the greatest health impact (promotive andFinally, the government needs to focus more on the preventive services), choosing instead to allocate thefinancing and delivery of district health services increasingly meager resources to a broad set of serv-and providing for their legal and administrative ices, and allowing donor-funded new capital invest-framework. ments that required additional running costs that

could not be provided adequately in the recurrentThe resource envelope and health sector budget.performance Donor resources have risen significantly in nom-

inal terms since FY95/96, which somehow cush-For the first half of the 1990s, Malawi's health sector ioned the adverse impact of a stagnant real perenjoyed relatively robust financing as the Govern- capita MOHP spending. However, since thesement of Malawi (GOM) deliberately increased public resources were not designed to directly support thefunding of social services. In 1993, as much as 7.4 budget except for limited amounts that went topercent of central government expenditures was departmental support, donor-funded projects were

3


Table 2MOHP recurrent and donor expenditures in the health sector: FY94195-FY98199

Items FY94195 FY95196 FY96197 FY97198 FY98199

MOHP recurrent expenditures (MK millions) 244.929 334.763 590.747 668.569 753.265

Donors' expenditures (US$ millions) 28.2 35.7 50.0 47.1 n.a.

Number of MK =US$1.0 8.7 15.3 15.3 17.5 43.0

Donors' expenditures (MK millions) 245.34 546.21 765.00 824.25 n.a.

Total expenditures (MK millions) 490.269 880.973 1,355.747 1,492.819 n.a.

Pop. of Malawi (millions) 9.96 10.25 10.57 10.91 11.24

MOHP recurrent expenditures per capita (MK) 24.59 32.66 55.89 61.28 67.02

Donors' expenditures per capita (MK) n.a. 53.28 72.37 75.55 n.a.

MOHP and donors' expenditures per capita (MK) n.a. 85.94 128.26 136.83 n.a.

Price index (1995=1.000) 0.545 1.000 1.354 1.517 1.638

MOHP recurrent expenditures per capita,at constant 1995 prices (MK) 45.12 32.66 41.27 40.40 40.91

Donors' expenditures per capita,at constant 1995 prices (MK) n.a. 53.28 53.45 49.80 n.a.

MOHP and donors' expenditures per capita,at constant 1995 prices (MK) n.a. 85.94 94.73 90.20 n.a.

Note: There is no consistency between the populaton figures used by the GOM and those reported in intemabonal bodies (such as Populabon Reference Bureau, the Wortd Bank). Thelatest Malawi Nabonal Census has not yet been finalized. This table uses the GOM population figures.Source: This review.

kept running well while the government health and to explore alternative financing and service-service delivery system remained underfunded. delivery modalities so that health services can beAlso, inflation has reduced the sizeable nominal improved.increase in donor spending, with real per capita Though the level of per capita health spending in

donor spending tending to decline. Malawi has declined since the mid-1990s, it stillThe difficult economic prospects for Malawi compares favorably with its neighboring countries.

require greater fiscal prudence and better resource Based on available data, government and donors'allocation in the health sector. Donors continue to health expenditures per capita in Malawi wasshow willingness to support the sector (with a large US$7.82 in FY97/98 while it was US$4.72 in Kenyapercentage of allocated amounts remaining undis- and US$8.90 in Tanzania in the mid-1990s (Table 3).bursed), but a new way of providing and managing As the HER will show, the allocation of Malawidonor assistance is imperative. Surveys also show health expenditures by levels of care is not undulyhousehold willingness to contribute to the financ- unbalanced and also compares favorably with theseing of health care. Finally, nongoverrment organi- countries. But sector performance data indicate thatzations (for-profit and nonprofit providers alike) Malawi's health spending seems not to be yieldingcan be tapped by the government in pursuit of the expected health outcomes commensurate withnational health goals. The challenge for the govern- the country's level and allocation of health expendi-ment is to orchestrate these domestic resources, to tures, or commensurate with the reported coveragecapitalize on the continuing goodwill of donors, rates of its health programs. Among four compari-

BETTER HEALTH OUTCOMES FROM LIMITED RESOURCES 5

Table 3Malawi health spending and performance compared to other countries in Sub-Saharan Africa:various years

Indicators Malawi Kenya Tanzania Zambia

GNP per capita in US$ (1997) 220 330 210 380

Population in millions (mid-1997) 11.24 28.0 33.3 9.4

Gov't expenditures on health per capita (US$) 3.50 3.56 4.23 8.12

Donor expenditures on health per capita (US$) 4.32 1.16 4.67 5.94

Gov't + donors' expenditures on health per capita (US$) 7.82 4.72 8.90 14.06

Percentage of HH expenditures per capita spenton health care (1993) 8.0 5.0 n.d. 3.0

Percentage of children immunized against DPT (1993-95) 98 40 79 76

Percentage of children immunized against measles(1993-95) 99 35 75 78

Percentage of births attended by trained health worker(1990-93) 41 n.d. n.d. 43

Population per physician (1990-94) 45,737 21,970 n.d. 10,917

Life expectancy at birth in years (1996) 43 58 50 44

Infant mortality rate per 1,000 live births (1992-93) 133 61 84 107

Childhood mortality rate per 1,000 live births (1992-93) 225 94 167 190

Maternal mortality rate per 100,000 live births (1993) 620 170 410 n.d.

Sources of data: World Bank African Development Indicators (1998199); World Development Report (1997); Malawi Human Resources and Poverty (1996); Health Policy in Eastem Africa:A Structured Approach (1997, draft); Malawi Health Expenditure Review (1999); Zambia Health Sector Expenditure Review (1995).

son countries in Sub-Saharan Africa, Malawi has autonomy, appropriateness and mix of skills, coun-the lowest life expectancy at birth, the highest infant terproductive coping mechanisms of staff), and inand childhood mortality rates, and the highest the overall institutional environment (the ability ofmaternal mortality rate. And although the country's MOHP to plan, marshal, and deploy resources toimmunization coverage is high, child mortality areas and services in greatest need). The quality ofremains high. health services being provided also need to be

There is clearly a conundrum that policymakers examined further, and remedial measures taken.in the health sector need to face. Poor sector per- This HER provides a broad macro/sectoral pictureformance cannot be blamed entirely on the low of the level and allocation of resources, but muchlevel of health spending (some comparison coun- more remains to be done to analyze and address thetries have lower spending per capita), or on the allo- micro/facility-specific problems plaguing the deliv-cation of these expenditures (other countries have ery of health services.more skewed resource allocation). Explanations forthis conundrum should be sought in technical effi- The government budgetciency (how well the system and individual facili-ties are managed), in the incentive structure for staff The GOM is the biggest provider of health services,and program managers (salary levels, degree of though funding has been eroded over the past few


years by a chronic budget crisis and a limited capac- below shows, the NHP's financial requirements areity to plan health services. In FY95, the MOHP far greater than the available resources forecast forbegan budgetary reform under the Medium-Term the medium term. There the focus of the Plan mustExpenditure Framework (MTEF) aimed at keeping be sharper.the scope and level of health services in line with Implementation of the NHP first requires translat-available budgetary resources. However, the MTEF ing the EHS into policy This entails (a) definition ofexperience up to 2000 indicates that the process and agreement on the package that the GOM is com-remains a mechanistic "budget cut-and-defend" mitted to fund; (b) the political and administrativeexercise that is not central to the MOHP's concep- approval of the EHS package; and (c) official policytion of its long-term strategic role in the sector. on the provision and funding of nonessential or non-Incremental budgeting continues to be practiced, package health services. The absence of such a policybudget protection is based on maintaining the cur- on nonessential services is tantamount to the gov-rent level of investments (especially hospitals), and errnent's acquiescing to continued unplanned infra-the scope of MOHP operations remains far wider structure construction and system expansion.than it can effectively fund and adequately super- The policy on the EHS package then needs to bevise. As a result, no dramatic budget reallocation translated into its programmatic requirements,towards more cost-effective preventive and promo- which entails (a) costing out the package itself;tive health services has occurred. In fact, there are (b) identifying and costing out the support servicesindications from the donor survey conducted for and other incremental inputs needed to implementthis study that donors may be crowding out the the package (staff training, information and educa-government from funding what should be its core tion campaigns, supervision and monitoring, infor-function (preventive and promotive health), and mation systems support); (c) adjusting the servicesthat as a result, a greater proportion of government included in the package on the basis of their costsresources are flowing to tertiary-care institutions. and available fiscal resources over the next three to

Admittedly, the level of government health allo- five years; and developing an implementation plancation is too low, but instead of focusing on cost- to carry out the approved strategy.effective services with large public-health impact, Finally, the EHS program costing exercise needsthe government has opted to finance all existing to be translated into the government budget, whichhealth programs and infrastructure including cost- entails actual government allocation for the pack-ineffective interventions. Some donors have con- age (both at central and district levels); governmenttributed to the crisis by offering infrastructure commitment not to divert the budget for nonpack-expansion with little regard for their recurrent-cost age services; and use of the refined NHP and itsimplications. The MOHP should use the budget implementation plan as the basis for marshallingcrisis and the MTEF process as an opportunity to additional external donor support.rethink its role, scope, and focus. Towards this end,the following actions are recommended: Tighten the coordination and institutional

locus of planning and budgetingFurther refine the National Health Plan asthe basis for priority setting, programming, The scarcity of resources demands more circum-and budgeting spect policymaking, planning, programming, budg-

eting, and releasing functions and tighter coordina-GOM launched the Malawi National Health Plan tion of these functions. Discussions are needed(NHP) in May 1999, embodying the government's within the govenmment on the budgetary implica-vision 2007 "to provide every Malawian with an tions-especially recurrent cost implications-ofaffordable package of essential health services each health policy, program, service, or function.(EHS) based on intensified community efforts and The MOHP must routinely undertake an exercise ofbacked by upgraded health centers and appropriate making alternative choices given alternative fund-district hospitals." As the funding gap analysis ing scenarios. Key actions in this area involve:


• Synchronizing the mandate and membership of worsen the recurrent-cost situation of the budget. Athe MTEF Committee and the Cash Budget Com- key tactic in solving the recurrent-cost problem is tomittee so that there is less variability in the plan- rationalize capital investment decisions.ning/budgeting and cash-releasing functions. Key actions in this area involve:Releases should be governed by the agreed-uponbudget priorities. * Exerting moral suasion at the highest levels in

donor headquarters so that uncalled-for medical* Decentralizing some of the MOHP functions to investments in Malawi are deferred unless the

districts and regional offices (such as monitoring donor is willing to fully or partly cover the run-and supervision) to free senior officers for more ning costs of the infrastructure in the mediumstrategic policy concerns, rather than having term, or until such time that the budget crisis hasthem occupied with day-to-day management. abated. Imposing policy conditions in donor

projects can also rationalize investments.* Integrating the recurrent and development

budgets, or at least making both of them as trans- * Enforcing a thorough and comprehensive finan-parent and comprehensive as possible. Until the cial analysis of the recurrent-cost implications ofmerger of the Ministry of Finance (MOF) and the any new government- or donor-funded project,National Economic Council (NEC), these two and developing stringent health-planning andactivities were disconnected, with recurrent public-finance standards to appraise these proj-budgeting mainly the purview of the MOF and ects. In this regard, the National Economic Coun-development budgeting mainly the purview of cil, working with the MOHP, should specifythe NEC. This bifurcation made it difficult to for- project-approval criteria that can then be used tomulate a rational forward-looking budget that set national investment priorities. In thetook account of the recurrent-cost implications of medium-term, MOHP also needs to explore theall capital investments. The merger of the two political feasibility of enacting a Public Healthministries into the Ministry of Finance and Eco- Investment Act to set the parameters for health-nomic Planning should at least provide the insti- sector investments, along the lines of "certificate-tutional locus for a more consistent budgeting of-need" or similar approaches.exercise.

Reducing or waiving counterpart funds, or* Strengthening the capacity of the MOHP Plan- designing counterpart-fund requirements so that

ning Unit to undertake health needs, costing, and they occur more in the out-years of projectscost-effectiveness analyses; to analyze budgetary, rather than up-front, or until such time that theservice-performance, demographic, and socioe- existing budgetary crisis has abated.conomic data and propose adjustments onexpenditure flows; and to analyze the recurrent- Extra-budgetary sources of fundscost implications of major health investments(donated or not) and recommend the best course The GOM's inability to finance the health needs ofof action on these proposed investments. Malawians through tax revenues should encourage

it to explore other financing modalities that areRationalize capital investments compatible with the population's ability to pay

Given a limited planning and regulatory capacity, a Expand program on drug revolving fundsMalawi NHP that is essentially a "wish-list" forinfrastructure projects undermines any rational Pilot-testing of drug revolving funds (DRFs) underapproach to sector investments because it leaves the International Development Association's (IDA)open the possibility of investors and donors offer- Population, Health and Nutrition (PHN) Projecting projects to the government, even if such projects has shown that they can be a viable source of sus-


tainable financing and can facilitate community (d) synchronize fee schedules among the differentaccess to basic pharmaceutical supplies. There are levels of care; and (e) specify waived or exemptedpotentially 600 villages that can participate in these health services, persons, or areas to protect theschemes. Key actions in this area involve (a) con- poor.ducting more DRF training programs on commu- The second key action is to, develop guidelinesnity organization and DRF management, and on the accounting, safekeeping, and utilization of(b) reviving the stalled reforms to restructure the fee revenues, including the phases of the fee pro-Central Medical Stores (CMS). The sustainability grams, the schedule of fees to be charged, staff tasksof DRFs hinges on the availability of drugs, which and responsibilities, and the planning and use ofin turn hinges on the institutional and financial generated revenues. The guidelines should alsosustainability of CMS, which procures drugs for define the waiver and exemption system at centralDRFs. and district hospitals and the procedures that staff

need to follow on means-testing, waiving, andImplement formal cost-sharing programs in exemption.government hospitals The third key activity is to restructure the Central

Medical Stores so that drugs are made available onDespite years of intent, government hospitals have a sustainable basis. Fee programs depend cruciallynot formally established institutionally sustainable on the availability of drugs, which is a major indi-fee programs. Existing efforts are haphazard and cator of quality of health services in Sub-Saharanuncoordinated; they have not been properly evalu- Africa.ated. Current disparities exist between MOHPfacilities, which do not formally impose fees, and Improve health insurance reimbursementdistrict-designated mission and local-authorityhealth facilities, which do. Partly because no fees Although health insurance coverage is smallare charged at the hospital level, and partly because nationwide, it represents a significant pool of thosegovernment primary-care facilities are ill-funded, with the ability to pay and thus provides a poten-all tertiary government facilities are clogged with tially major payment system for hospitals. GOMpatients who bypass the referral system. hospitals need to review their fee schedules and

To jump-start the fee initiative, the first key reimbursement rates to patients under medical aidaction is to conduct a thorough and comprehensive schemes or health insurance coverage to align themreview of existing formal and informal fee schemes. with actual costs and remove unnecessary govern-A Japanese PHRD grant is currently supporting the ment subsidy for these patients with the abilityreview of the cost-sharing programs of the central to pay.hospitals. In May 1999, a DflD-funded consultancy The government needs to study the administra-completed a draft health financing strategy for tive, financial, and economic feasibility of convert-Malawi (LATH: 1999). Based on the findings of ing the current government-funded Referral ofthese analyses, MOHP needs to develop a national Cases Abroad into a medical aid scheme for civilgovernment fee policy covering district hospitals, servants or a health insurance contract with a pri-district-designated Christian Hospital Association vate insurer.of Malawi (CHAM) facilities receiving governmentsubventions, and local-authority facilities. The Private health expendituresnational fee policy should permit (a) central anddistrict hospitals to impose fees and have private Private health expenditures are a significant,wards based on fees; (b) permit 100 percent though largely unstudied, source of health financ-retention of fee revenues at the hospital level; ing in Malawi. There is a well-established nonprofit(c) permit hospital use of fee revenues subject sector (Christian Hospital Association of Malawi orto specified guidelines from MOHP and spending CHAM) and a burgeoning for-profit sector, mainlyauthorities designated by the government; private clinics.


Learn from nonprofit health providers to pay for health services but do not) and under-coverage (households with scant capacity to pay

A third of CHAM resources are from user charges but who do) in health facilities across the country.and sale of drugs generated from modest fee sched- Informal fee programs without official sanction orules that do not deny poor patients access to care. with scanty supervision tend to worsen these prob-GOM should learn from these institutionalized sys- lems. Fee programs need to be formally established,tems and practices. GOM also needs to reconsider especially in urban facilities where nonpoor house-reviving the CHAM subvention program (which holds are mostly located. But the waiver andgrew from MK4.6 million in FY90/91 to MK22.2 mil- exemption systems of these programs should belion in FY94/95) but was eliminated with the onset strengthened, staff should be trained in them, and aof the budget crisis in FY95/96. It should be noted vigorous information campaign should be launchedthat CHAM facilities provide services to more than so that patients know which persons or conditionsa third of the Malawi population, mostly in rural are eligible for free care.areas where there are no government providers.

Donor financing and expendituresEncourage but regulate private for-profithealth providers Donors account for about half of health expendi-

tures in Malawi. As of the end of FY97, as much asSerious issues with respect to equity, pricing, and US$246.5 million were allocated to the sector, ofquality of care have emerged in the wake of the which only US$115.7 million had been expended.health-sector privatization and liberalization. GOM The slow disbursement rate (46.9 percent as of end-needs periodically to evaluate the impact of open- FY97) is a cause of concern, especially in the currenting the medical sector to private practice. budget crisis when donor resources ought to

The government should assist the Medical Aid So- cushion the impact of fiscal stringency. Here areciety of Malawi (MASM) address the issue of provi- several recommended actions to improve the use ofder-driven over-consumption of health services. This donor funds:problem is manifested through over-prescription,over-examination, and prolonged length of stay for Improve project pipeline management andpatients with health insurance coverage. It can be ad- disbursementdressed through better monitoring of providers andmore stringent professional guidelines and discipline. The administrative bottlenecks at MOHP headquar-

Finally, AIDS care should be addressed in the ters should be eased. Within government, delegationcontext of health insurance as well as public authority must be established for officials who arefinance. Though AIDS is a national epidemic, the away, and whose signatures are necessary to moveissue of financing AIDS care is being flagged more the paperwork. Government capacity has to beactively by private insurance schemes. These pro- strengthened to manage and follow up requiredgrams fear that they will collapse under the impact actions in order for donor funds to flow or for pro-of large-scale clamor for expensive treatment will curement to proceed. Both government and donorsraise premiums to an unaffordable level. The need to review time-consuming and onerous pro-increasing international clamor to provide financ- curement procedures, improve understanding ofing and coverage for expensive antiretroviral drugs these procedures, and train more staff in all aspectshighlights the importance of this issue. of procurement and financial management. A pro-

curement unit within MOHP may be called for.Promote household expenditures, but MOHP needs to draw up a project monitoring chartimprove protection systems for the poor for each of the donor-funded projects, monitor dis-

bursements more closely, and pay close attention toA UNICEF-commissioned survey in 1995 reveals slow-moving projects. Tighter coordination issubstantial freeloading (households with capacity required among the MOHP, the Ministry of Finance,


local governments, and nongovernmental organiza- donors should work together in (a) setting standardstions (NGOs). Periodic review and flagging of issues for consultant fees, technical assistance contracts,can speed compliance with legal, policy, program, or and contract provisions to achieve greater cost-effec-administrative requirements and conditions for dis- tiveness, quality control, and equity; (b) synchroniz-bursement. A regular portfolio review between the ing project design, supervision, and evaluationGovernment and the Bank appears to be yielding activities to save costs; and (c) drawing up a code ofpositive results; such a practice needs to be conduct for NGOs and contractors, including theirexpanded to other donors. financial responsibilities.

Tighten central coordination of donor- Explore the feasibility of a multidonorfunded projects budget or expenditure support program

The plethora of projects is beginning to 'balkanize' The sheer number of donor projects and activitiesMalawi's health sector as each donor stakes its (at least 34) overburdens the constrained capacity atclaim on a district or vertical project. The tendency central MOHP and may account for slow disburse-is for each donor-funded district to have its own ments. The design, appraisal, management, moni-health priorities, training thrusts, information toring, and evaluation requirements of individualrequirements, and financial systems, with little projects are immense and tend to overburden theregard for national-system requirements, eventual Malawian bureaucracy. In addition, individualdata aggregation, or service standards. This trend donor projects may in fact worsen the recurrent-costmust be stopped, but this can only be done with problem as GOM can ill-afford to provide the 10-15strong, central MOHP coordination. percent of government counterpart funding. The

The geographic distribution of donor-funded situation is worse for capital projects as GOM needsprojects must be rationalized. The inventory of to set aside their running costs, which are increas-these projects shows that donors tend to locate in ingly becoming unavailable in the current fiscalbetter-off areas, with remote districts having little squeeze. Finally, donor-funded projects in principleaccess to donor-funded services. GOM has to estab- do not provide salary supplements, yet the lowlish explicit criteria and guidance on where donor level of salaries is probably the single most impor-projects should locate to ensure greater equity. tant factor that accounts for the poor quality of

Similarly, donor-funded training needs to be health services. A way out of this conundrum ofrationalized. Donors are spending a staggering "cash-rich" donor projects and "cash-starved"amount (US$4.5 million annually) on long- and GOM may be a multidonor budget or expenditureshort-term training (workshops, conferences), with support program that can be made contingent uponlittle documentation of their impact on health serv- GOM's commitment to prioritize its budget to pro-ice quality. Future training activities should be vide an essential package of health services, and tobased on a sector-wide human resource develop- launch policy and administrative initiatives to sup-ment plan and should include (a) a program for port such service-delivery focus. The MOHP's pre-synchronizing and harmonizing existing training liminary discussions with donors on a sector-wideprograms and possibly combining similar ones; and approach (SWAp) is a step in this direction.(b) a mechanism for evaluating their impact.

Sectoral efficiency improvementImprove management of project managersand contractors In addition to budgetary reform, three other areas

can be pursued to improve efficiency in Malawi'sThere is an increasing trend towards privatizing or health sector: restructuring the Central Medicalcontracting the supervision and management of Stores (CMS); granting of hospital autonomy todonor projects. Given the perceived costliness of selected tertiary facilities; and further decentraliz-nongovernment management of projects, GOM and ing the health service. All of these improvements


involve profound systems change and institutional tional, management, and financial requirements fordevelopment and therefore have inherent risks and expanding their autonomy; (c) set quantitative tar-costs, especially during the transition. The govem- gets for the feasible reduction of government sub-ment needs to carefully weigh the options within sidy and for increased altemative revenue sourceseach reform area to minimize transition costs. such as fees, insurance reimbursement, official and

private philanthropic grants, and nonmedical rev-Pursue the stalled pharmaceutical enues; and (d) set service-delivery targets that can besector reforms used in the planned health service "contracts"

between the Govemment and the hospital boards.Drug availability is the lynchpin that joins major Second, on the basis of the results of the studyflanks of any health sector reform. If drugs are and stakeholder consultations, provide legal auton-unavailable, community drug revolving funds omy to the three central hospitals and seven "busy"cannot operate and fee-based cost-sharing pro- district hospitals (Rumphi, Mzimba, Kasungu,grams cannot succeed. Without a good revenue Dedza, Mangochi, Machinga, and Mulanje), andbase from fees to cushion the impact of reduced establish an enabling policy environment to makebudgetary support, tertiary hospitals cannot them self-financing. An Act of Parliament isbecome autonomous. Given the centrality of phar- required to grant full autonomy to these facilities.maceuticals in the whole health sector reform effort, The Act should also redefine the role of the MOHPthe government needs to pursue the stalled restruc- with respect to autonomous hospitals. In lieu ofturing of the CMS to make it more autonomous, very specific instructions, regulations, control, andsustainable, and efficient. Without these CMS supervision of most aspects of hospital manage-reforms, the financing, procurement, and distribu- ment, the MOHP should establish goals, targets andtion of drugs will continue to be imperiled. policies for hospital services to be provided, and

The supply-side reforms in pharmaceuticals then fund hospitals according to these requirementsneed to be supported by corresponding improve- through a health-service contract with each of them.ment in consumption pattems. This can be achieved This arrangement distinguishes between MOHP asthrough a variety of mechanisms including improv- purchaser and regulator of services and theing physician-prescription behavior through better autonomous hospitals as providers of services.training and monitoring, imposing partial or full- The Working Group on Hospital Autonomy'scost fees on prescription drugs at govemrment facil- third proposal is to establish a Council of Hospitalities, and establishing therapeutic committees and Boards comprising representatives from the individ-drug registers at hospitals to keep track of drug con- ual hospital boards. The Council should be perceivedsumption, reduce theft, and encourage the develop- as a trade organization acting in the interests of thement of demand-driven (rather than the current autonomous hospitals in their dealings with theinefficient supply-driven) system. MOHP in such areas as staff salaries, pensions, and

benefits; hospital fee schedules; staff training; bulkGrant hospital autonomy to selected procurement and equipment sharing; and coordina-hospitals tion of management systems and procedures.

Hospital autonomy should be implemented inThe Working Group on Hospital Autonomy has the context of hospital financing reform based on aidentified three specific proposals, which should be realistic assessment of the budget and availablesupported. First, undertake a feasibility study to extra-budgetary resources; the essential and cost-underpin the hospital reform strategy. The study effective clinical services that these hospitals shouldshould (a) analyze the degree of autonomy that can- provide; and the cost-effective ways of providingdidate hospitals currently have with respect to man- ancillary services including contracting. Finally,agement, staffing and personnel, budgeting and these reforms should be underpinned by hospitalfinancial resource base, procurement, and quality renovation or refurbishment, staff retraining, drugimprovement; (b) discuss the legal, policy, organiza- availability, and other visible signs of quality im-


provement. (Any fee-based program supportive of * the revitalization of the subvention program and

hospital autonomy is bound to fail without requi- possible conversion of such traditional and infor-site improvement in quality.) mal relationship into a more formalized health-

service contract between the GOM as funder andCarry out planned activities to support mission hospitals as providers, with specifichealth service decentralization responsibilities and deliverables defined in terms

of target coverage, cost, and quality; andWith the passage of the Local Government Act, theholding of local elections, and the establishment of * the wider participation of CHAM and NGOs inlocal councils, the legal and political structure for service delivery in their respective health dis-devolution have been set in motion. he greater chal- tricts, and in health policy at the national level.lenge now is to institutionalize the fiscal frameworkfor decentralization so that central grants and trans- Monitor and evaluate the contracting offers are channeled appropriately to peripheral areas health servicesand key health priorities are adequately funded ontime. The flow of bilateral and multilateral assis- The Government is in the process of contracting outtance to the districts is particularly knotty since ancillary health services such as cleaning, transport,there are many donors, each with specific require- building and ground maintenance, laundry,ments. So far, no adequate mechanism has been security, catering, audit, and mortuary. These aredefined for this purpose. novel experiments in Malawi and need to be closely

The extent and pace of decentralization would monitored and evaluated in terms of efficiency anddepend on the capacity of local authorities to cost-effectiveness relative to the status-quo (that is,manage decentralized health services, as well as self-provision). So far, contracting of clinicalMOHP's ability to support and supervise these services is not yet under discussion, but given thefunctions. Key issues in this area are the availability burgeoning private medical practice, the Govern-of skilled staff at the local level, and the status of ment should consider it. Lessons learned fromcivil servants to be "absorbed" by local authorities. nearby countries (South Africa) should inform theThe roles and functions of private providers under design of contracts, price negotiation, and othera decentralized setup should also be defined. The considerations.key issues in this regard are:

3

MOHP budget andexpenditures

Allocation trends FY98, the MOHP's budget ceiling was pegged atMK753.3 million (Table 4). Reckoned in real prices,I n the first half of the 1990s, the GOM made a the Ministry's tight level of health spending is even

deliberate effort to increase public spending in more stark: while the absolute and per capita levelshealth, raising the health budget from MK83 of recurrent expenditures more than doubled nom-

million in FY91 to MK415 million in FY95. As a inally, at 1995 constant prices, per capita MOHPresult, the proportion of the MOHP budget to the spending has remained constant at around MK40total government budget increased from 9 percent during the past three years.in FY91 to 11 percent in FY95 (Mwambaghi 1996).Because of relatively modest inflation during the MOHP expenditures by level of institutionperiod, at 1990 prices the budget rose from MK72million in FY91 to MK108 million in FY95. Per Under the MTEF, dramatic changes were made incapita, the GOM health budget at 1990 prices MOHP budget accounting and presentation,increased from MK7.94 in FY91 to MK10.25 in FY95. making it difficult to analyze MOHP recurrent

These successes began to unravel towards the expenditures by level of institution. The data pre-middle of the decade in the wake of political uncer- sented in Table 5, therefore, should be interpretedtainties, the economic downtum in 1994-95, a large with care; they are meant to be impressionistic.currency depreciation, and the inflation that For FY98, MOHP requested a budget of MK753.3ensued. By 1996, forecasts made by the Ministry of million, of which more than a third (35.9 percent)Finance incorporating economic growth and was for central HQ, a minuscule 1.7 percent was forexpected tax and other revenues showed that liberal the three regional offices, around 17.1 percent assincreases in the health budget were no longer possi- allocated for the four central hospitals, and close toble (Marshall 1996). At that time, the health budget a half (45.3 percent) was devoted to the 24 districtwas expected to be MK760 million in FY97, MK865 health offices (inclusive of district hospitals, ruralmillion in FY98, and MK970 million in FY99. At hospitals, and preventive and promotive care serv-these levels, the health budget was expected to ices delivered in their catchment areas). Contrary toretain its 15 percent share of voted expenditures the stated policy thrust of greater decentralization,and 12 percent to 13 percent of total GOM budget. the central MOHP HQ dramatically increased its

These forecasts turned out to be optimistic; in share of allocation. On the other hand, the share offact, actual health expenditures (at current prices) resources devoted to district health offices shrankwere only MK334.8 million for FY95, MK590.7 mil- from more than a third (67.7 percent) of actuallion for FY96, and MK668.6 million for FY97. For expenditures in FY95 to only 45.3 percent of the

13


Table 4MOHP recurrent expenditures and per capita expenditures in current and real terms:FY95196-FY98199 (in million Malawi Kwacha)

Percent changeItems FY95196 FY96197 FY97198 FY98199 FY95198

MOHP recurrent expendituresin current prices (MK m) 334.763 590.747 668.569 753.265 125.0

Price index (1995=1.000) 1.000 1.354 1.517 1.638 63.8

MOHP recurrent expendituresin real prices, 1995=100 (MK m) 334.763 436.298 440.718 459.869 37.4

Population (m) 10.25 10.57 10.91 11.24 9.7

Per capita MOHP recurrentexpenditures in current prices (MK) 32.66 55.89 61.28 67.02 105.2

Per capita MOHP recurrent expendituresin real prices, 1995=100 (MK) 32.66 41.27 40.40 40.91 25.3

Notes: (a) Malawi's FY is from Apri I to March 30 the following year. (b) Populabon is based on mid-calendar year estimates. (c) FY95/96 to FY97/98 figures are actual expenditures;FY98199 figures are budget allocations submitted to Parliament as of June 30, 1998.Source of basic data: MOHP Planning Unft.

allocated amount in FY98. The share of resources health offices tend to underspend, indicating that

going to regional health offices was also down (4.4 the government's bias towards urban areas and cur-

percent in FY95 to 1.7 percent in FY98). Some of ative care has increased in a tight fiscal situation.

these drops might have been due to accounting

changes (for example, certain budget items were MOHP expenditures by level of cost center

"recentralized" for accounting purposes and cost-

center responsibility, or portions of the Regional For FY98, MOHP rewrote its budget to reflect iden-

and District Medical Offices budgets were trans- tified priority programs. The results of this exercise

ferred to central administration), but the magnitude are shown in Table 6. District health offices provide

of the numbers involved raises serious concerns. preventive/promotive services and primary/

The four central hospitals, which used to account secondary curative services, which together

for around 25 percent of MOHP's budget got a accounted for roughly 45.7 percent of the FY98

smaller share in FY98. Budgetary resources devoted budget allocation. The central hospitals provide ter-

to the two largest hospitals (Queen Elizabeth and tiary curative as well as rehabilitative health serv-

Lilongwe Central), however, remain large: one terti- ices, which together accounted for roughly 17 per-

ary hospital (with an average annual budget of cent of the FY98 budget. The remaining 35 percent

MK45 million) consumes as much as three district of the budget went for technical and administrative

hospitals (each with an average annual budget of services provided by central HQ and regional

MK15 million). The government needs to find alter- offices.

native ways of funding these large urban-based These budget figures, however, mask the real

institutions so that critical health services in the type of health services rendered by health facilities.

periphery are not crowded out of funding. The gov- For instance, hospital managers say that around

einment also needs to ensure that institutions stick three-fourths of patients at central hospitals are pri-

to their budgets. Recent experiences on allocations mary-care patients who should have been treated at

versus expenditures show that MOHP HQ and cen- lower level facilities. The ill-functioning referral

tral hospitals tend to overspend while district system (largely made inoperative by poorly funded


Table 5MOHP recurrent expenditures by level of institution: FY95196-FY98199(in million Malawi Kwacha)

Percent changeItems FY95196 FY96197 FY97198 FY98199 FY95198

Central HQ 21.653 83.921 73.125 270.409

Regional health offices (all three) 14.758 15.455 21.882 12.676 (16.4)

Average per regional health office 4.919 5.152 7.294 4.225

Central hospitals (all four) 71.540 142.428 166.619 129.114 80.5

of which Queen Elizabeth CH 29.000 n.a. n.a. 46.134 59.1

of Which Zomba CH 11.166 n.a. n.a. 28.031 151.0

of which Zomba Mental Hosp. 6.349 n.a. n.a. 10.842 70.8

of which Lilongwe CH 25.025 n.a. n.a. 44.107 76.3

District health offices (all 24) 226.813 348.942 406.945 341.066 50.4

Average per district health office 9.451 14.540 16.956 14.211

Total 334.763 590.747 668.569 753.265 125.0

of which percent Central HQ 6.5 14.2 10.9 35.9 -

of which percent regional health offices 4.4 2.6 3.3 1.7 -

of which percent central hospitals 21.4 24.1 24.9 17.1 -

of which percent distrct health offices 67.7 59.1 60.9 45.3 -

Notes: (a) FY96 and FY97 are revised actual tigures; (b) FY98 based on budget submitted to Parliament. The details of FY98 figures are not directly comparable to the details of the earlierfiscal years due to changes in definition of accounts and cost centers, e.g., some items previously under 'Central Hospitals" vere later placed under 'Central MOHP. (c) For this and othertables, details may not exactly add up to totals due to rounding.Source of basic data: MOHP Planning Unit.

rural health centers and dispensaries) perpetuates care needs are also unknown, it is difficult to assessthis massive and distorted health-seeking behavior, whether the government is providing enoughwith large financial implications. The preference of resources for them. Recently, the Working Group onpolicymakers, at least as revealed through annual Essential Health Package proposed the followingbudget allocations, has been to expand tertiary care services to receive core funding from the govern-to accommodate primary-care patients, rather than ment: Expanded Program on Immunization; com-to improve lower-level health facilities. Govern- munity-based primary health care (village healthment's long hesitance to institute a modicum of fees committee, income generating activities, traditionalat the highest levels of care, even as it encourages birth attendants); family planning; safe mother-fees for drugs under community revolving funds, hood; nutrition; control of malaria; prevention andhas also engendered a perverse incentive structure control of HIV/AIDS and other sexually transmit-in the health system. ted diseases; control of tuberculosis; control of diar-

The proportion of FY98 budget allocation to pre- rheal diseases; control of acute respiratory infec-ventive/promotive care was nominally 14.2 per- tions; and management of other common illnessescent. However, since the cost of providing such and conditions (such as eye infections, skin dis-services is unknown and the magnitude of primary- eases, and minor injuries). It behooves the MOHP to


Table 6MOHP recurrent expenditures by level of cost center: FY98/99 (in million Malawi Kwacha)

PercentItems Coverage Amount Share

a. Preventive and Human resources, child health, nutrition, reproductive health, environmental health, 107.025 14.2promotive services communicable disease control, noncommunicable disease control, health education

b. Primary curative Human resources, patient care, special services (ambulance, mortuary), 45.844 6.1services construction/rehabilitation

c. Secondary curative Human resources, patient care, special servicess (ambulance, mortuary), hospital 191.549 25.4services management, construction/rehabilitation

d. Tertiary curative Human resources, patient care, special services (ambulance, mortuary, CSSD), 121.713 16.2services hospital management, construction/rehabilitation

e. Rehabilitative health Human resources, physio/occupational therapy, psychiatric services, prosthetic 6.259 0.8services services, home-based care

f. Health technical Human resources, lab services, radiology services, pharmaceutical services, 223.425 29.7support services physical-assets maintenance

g. Administrative and Human resources, general administration, planning and evaluation, financial 47.204 6.3technical services management, clinical services, nursing services, support to preventive services

h. Support to other Subventions and subscriptions to local and international organizations 9.840 1.3institutions

i. Population services - 0.406 Negl.

Total All 753.265 100.0

Source of basic data: MOHP Planning Und.

fully fund these programs to avert expensive hospi- (51.8 percent) are low-skilled workers doing ancil-talization and thereby begin to gradually reduce lary duties (Table 8). By its sheer number, the cost ofallocations to the hospital sector in favor of lower- maintaining such workforce must be significant.level facilities. For higher-skilled workers, the vacancy rate is high

due to low salaries and the difficulty of filling staffMOHP expenditures by economic positions in rural areas where housing amenities areclassification often unavailable. Proposals have been made to

build staff houses at regional/district offices andPersonal emoluments accounted for more than a student accommodations at central hospital/train-third (36.9 percent) of the FY98 budget allocation ing institutions. A few donors are also mulling over(see Table 7). Personal emoluments include salaries, the possibility of providing salary supplements towages for non-established staff and temporary health staff in rural areas.employees, and a significant amount for housing The low salary of civil servants should beallowance. addressed. In the absence of a suitable resolution,

There are no basic staffing norms (range of skills some workers may be adopting counter-productiveand services to be provided at each level of care) in behavior to augment their meager salaries (atten-Malawi. As a result, the appropriate mix of health dance at as many workshops as possible, travelworkers required is unknown. Analysis of the under the guise of supervision). Though this issueMOHP plantilla, however, shows that the majority has been discussed at senior levels, a thorough-


Table 7MOHP recurrent expenditures by economic classification: FY95/96-FY98199(in million Malawi Kwacha)

FY95196 FY96197 FY97198 FY98199 Percent changeItemsa amount percent amount percent FY95196-FY98199

Personal emoluments 132.402 39.6 277.808 36.9 109.8

of which administration, technicalsupport and RHO 12.404 3.7 25.348 3.4

of which four central hospitals 39.298 11.7 66.927 8.9

of which district health offices 78.312 23.4 185.533 24.6

of which training institutionsb 2.388 0.7 0 0.0

Goods and services 191.200 57.1 452.078 60.0 136.4

administration, technical supportand RHO 14.042 4.2 243.903 32.3

of which, four central hospitals 39.504 11.8 58.866 7.8

of which, district health offices 131.615 39.3 149.309 19.8

of which, training institutionsb 6.039 1.8 0 0.0

Capital transfers 4.651 1.4 9.840 1.3 111.6

Capital formation 6.510 1.9 13.539 1.8 108.0

Total 334.763 100.0 753.265 100.0 125.0

Notes: (a) Expenditures for the three regional offices are included under admin. & tech. support. (b) Training insbtubons have been made autonomous and were taken off the MOHPbudget in 1997/1998.Source of basic data: MOH Planning unit.

going "right-sizing" of the bureaucracy is yet to be rationalize the MOHP's budget and public expendi-implemented. Meanwhile, personal emoluments as ture management.a share of total MOHP expenditures continued todecline (from 39.6 percent in FY95 to 36.9 percent in Development of sector objectivesFY98), reflecting the number of staff leaving thecivil service as well as those dying of AIDS and In FY96, the Ministry conducted a logical frameworkother diseases. exercise, developed sector objectives, and irtitiated

the move away from incremental budgeting andMOHP's medium-term expenditure framework towards a more progranmmatic approach. On the

basis of the existing program portfolio, however, itIn FY95, in response to the extremely tight budget appears that MOHP used the logframe exercise to jus-situation, MOHP was one of the four ministries to ttfy cramping all the existing programs within thedevelop and execute a Medium-Term Expenditure framework, rather than to prioritize. Thus, the log-Framework (MTEF) to keep health services in line frame became an "all-inclusive" device rather than anwith available budgetary resources. The aim of the exercise to streamline operations and do away withMTEF exercise to help ministries prioritize the serv- the lowest priorities. According to MOHP planningices they will provide. Despite the initial difficulties, staff, the elimination of "less priority" programs wascertain positive aspects have been incorporated to never discussed in the early years of the MTEF as the


Table 8MOHP staff breakdown: FY97/98

Number of Number of filled Percent to totalGrade estimated posts and funded posts filled and funded posts

Si

S2/P2 9 9 0.1

S3/P3 4 4 Negl.

S4/P4 31 31 0.3

S5/P5 42 43 0.5

S6/P6 5 5 Negl.

S7/P7 68 68 0.7

S8/P8 89 89 0.9

CEO/CTO/PO/AO 335 347 3.6

D1-8 87 89 0.9

SEO/STO 361 381 4.0

EO/TO 941 824 8.7

SCO/STA 520 520 5.5

CO/TAIDP3/2 2,177 2,172 22.8

SC1-IV 4,924 4,933 51.8

Total 9,493 9,515 100.0

Source of data: Budget subrrssion.

Ministry hoped that a more robust budget the next closer to the "Program" rather than to the "Activi-

year could be used to protect existing programs. In ties", the budget accounting system may be more

FY98, the budget crisis continued but the elimination performance-oriented rather than input-oriented,

of programs, services, and functions remained taboo but this conjecture needs to be verified by results.

within MOHP. Given the extremely linited budget MOHP also developed new program budget

provided by the Miristry of Finance, the relationship classifications in FY96 that made it relatively easier

between the Treasury and MOHP continued to take to see program priorities and were better than clas-

the form of a budget "cut and defend" approach. sifications in use until FY95. The old (transitory)

classifications were based on objectives that

New budget accounting system reflected accounting inputs; the new classifications

and classifications were based on programs, that is, they unified

accounting functions that were similar (preventive

Under the old/transitory accounting system, objec- care, curative care). The new classifications were

tives were defined in accounting terms and tended to considered more transparent.

lump together institutions doing different things.

The new accounting system attempts to lump similar Introduction of cash budgeting

institutions/functions based on a program; the defi-

nition of cost centers also clarified financial responsi- In theory, the introduction of cash-based budgeting

bility. It is possible that by bringing the "Cost Center" and lump-sum releases under MTEF should have


encouraged ministries to prioritize their spending. activities as rehabilitative care (physiotherapy andHowever, according to MOHP planning staff, the occupational therapy, psychiatric services, andcash budgeting system may unduly create adverse prosthetic services), cancer registry, oral health,incentives within MOHP to spend quickly; it also mental health, the maintenance of high-tech equip-engenders aggressive lobbying by program man- ment, treatment abroad, free mortuary, and trans-agers and hospital directors, which can subvert allo- port of the dead.cations based on strategic considerations. Some Analysis of the past budgets under MTEF showsMOHP informants believe the cash-based budget- MOHP prefers to run most if not all of its traditionaling system preserves the status quo, and in fact programs and to distribute the meager budgetdoes not lead to real reallocation based on cost- across all these programs, rather than cut away lesseffectiveness criteria. These adverse incentives are necessary programs and focus on those servicesexacerbated by the fact that the members of the that are justified epidemiologically, cost-effectively,MOHP's MTEF Committee (which prepares the and on the basis of equity. This broad rather thanbudget) do not also site on the Cash Budget Com- focused approach hurts all programs, poorly pro-mittee (which allocates funds made available by the tects cost-effective primary-care services, and leadsTreasury). Real power lies within the Cash Budget to the unfortunate drying up of key interventionsCommittee, whose priorities do not necessarily that are aimed more at the poor and rural areas,reflect those by the MTEF Committee. such as extremely limited budgets for key interven-

tions in malaria, tuberculosis, and other promo-Stopping virement tive/preventive services. There are also indications

that health services under local authorities, whichMOF has stopped all ministries from the practice of ought to provide first-level contact, are far morevirement, that is, using a line item to fund expendi- underfunded than MOHP hospitals.tures in another line item that has been exhausted.Virement prohibition is intended to instill greater Macro constraintsfiscal discipline, though under very extreme finan-cial situations, virement is often used to meet emer- The continuing breadth of MOHP operations, evengencies and contingencies. The prohibition of vire- under MTEF, can be explained by political, institu-ment, while admirable in itself, can be rendered tional, administrative, and technical problems thatineffective by the Treasury's and/or the Cash need to be addressed in order to realize the fruits ofBudget Comnmittee's delay of releases to specific better management of public expenditure. First,programs or budget items. A program could be MOHP has suffered because MOF does not offi-severely underfunded during the year if the Cash cially announce its budget ceiling early. The lateBudget Committee decides to use the available official announcement of ceiling has been going onfunds for something else, even one that is of less since FY94 and was true at least until 1998/99. Itpriority. appears that the MOF finds it politically difficult to

announce the ceiling early as it will come underContinuing challenges in the MTEF process very strong pressure from politicians and technical

ministries lobbying for higher ceilings.Given the tight budget situation, GOM cannot fully Second, the recurrent and capital developmentfund all health services that it has traditionally pro- budget processes have historically been bifurcated:vided. It appears that the government has not taken MOF is principally involved in the recurrent budgetthis to heart and is spreading its resources too thinly while the NEC takes care of the developmentby service and program area. Even under three to budget. In general, donor-funded expendituresfour years of MTEF, MOHP's scope of services and remain reckoned within the development budget,programs remains broad, ranging from preven- even though much of these expenditures aretive/promotive programs, to basic and specialist recurrent. Similarly, the Recurrent Budget thatcurative services, to such programs and funded MOHP prepares for MOF approval includes capital


formation, which ought to be under the Develop- To assess how the recurrent costs and counter-ment Budget. The MTEF exercise aims to consolidate part funding requirements exacerbate the MOHPthese two processes, but full consolidation remains budget crisis, this study calculated (a) the recurrent-to be completed. Since MTEF began, heavy focus has cost implications of health investments, especiallybeen placed on recurrent expenditures. Consolida- hospitals, using an R ratio - the ratio of recurrenttion of the recurrent and development budgets is expenditure requirements to total investment cost -extremely important as it will show how GOM and of 0.247, based on a Malawi study (Heller 1997); anddonor resources are being allocated and which (b) the required counterpart funding of donor-programs are relatively well-funded and which are funded projects in the PHN sector, which is typi-

not. Historical trends indicate GOM's heavy bias cally 10-15 percent. Based on these rough calcula-toward curative care while donors in general focus tions, the implied recurrent cost requirements foron preventive and promotive care. A sector ex- MOHP capital expenditures rises threefold frompenditure plan that lays out the anticipated MK24.1 million in FY95 to MK74.9 million in FY97.resources from both government and donors over On the other hand, based on a 12.5 percent counter-the next three to five years is clearly needed to part-fund requirement, MOHP needed at leastinform the annual MTEF exercise. The recent unifi- MK40 million a year in the past three fiscal yearscation of the Treasury and the NEC into the Ministry (1995-97).of Finance and Economic Planning should alsoaugur well for a more comprehensive and consistent Technical constraintsbudgeting exercise.

The health sector's MTEF exercise as practiced inInstitutional and political constraints the late 1990s has been more "strategic". MOHP

consulted with stakeholders (program managersThe government and MOHP policymakers appear and district health officials) but these consultationsunable to face the budget inadequacy squarely, were not done in the context of finalizing, budget-according to planning staff. Capital investments, ing for, and implementing the draft health sectorespecially in the hospital sector, are largely physi- strategic framework. Rather, budget consultationscian-driven or politician-driven or sometimes and cuts were made on the basis of existing pro-donor-driven, with the lobbying sponsor having grams and protecting the sunk costs of existinglittle regard for the recurrent-cost implications of investments. MOHP Planning staff admit that thethese investments or for the alternatives for which so-called MTEF, in spirit and in practice, is stillthe funds could be used, especially for ill-funded largely an incremental budgeting approach and ispreventive/promotive health programs. A few far from the desired standard of budgeting on thedonors are also prone to offer capital investments basis of sector objectives. The inertia of incrementalwith little regard for their recurrent cost implica- budgeting means that the MTEF process is simply ations, thereby forcing the government to budget cutting-and-defending-the-budget exercise, ratherfunds to maintain this new infrastructure rather than a means to give the GOM a way to strategi-than use those funds for the more cost-effective cally plan its core functions, namely, what it shouldinterventions. MOHP needs to be more aware of the be funding and providing cost-effectively tocost implications of these choices, to be prepared to Malawians.assess the projects that are offered, and to resist Limited technical data and skills hamper thepolitical pressures to accept them. Donors should MOHP from moving towards a more strategicalso be educated on the adverse recurrent-cost approach. There are no unit costs available for pro-impact of their capital investments. Due to the very gram costing exercises. Neither are effectiveness orfragile budget situation now and in the near future, output indicators available for programs, making itboth donors and GOM ought to be extremely wary difficult to make rational choices about health-serv-of the cost implications of even the most modest ice alternatives under alternative funding scenarios.capital investments in the sector. In the absence of this information, prioritization


becomes a matter of lobbying by program man- siderable, with significant impact on the MTEFagers, rather than of policymakers deliberately budget.making strategic choices based on epidemiologicalneeds, cost-effectiveness, and/or equity criteria. hI Recommendations for improved resourceaddition, there are no practical means of allocating allocation and usejoint costs across programs, for example, thesalaries of staff members involved in multiple pro- The changes in budget titles and subprograms ingrams have not been apportioned appropriately to the past few years make it difficult to compare thethose programs. Also, staff listed in one facility or annual consistency in government priorities. How-district may in fact be working in another facility or ever, based on principles of cost-effectiveness anddistrict or in the central office. the appropriate role of the government in the health

Many of these problems can be resolved sector, Table 9 provides a summary of recommen-if MOHP takes a purposeful approach to devel- dations for improved targeting and more effectiveoping its national health strategic plan that, among use of government resources. The following need toother things, (a) defines the Ministry's role and core be highlighted:

functions in health service provision and financing;(b) justifies these roles, functions and services in * Referral of cases abroad - These referrals costterms of well-established economic and cost- MK2.8 million in FY95/96. Under such a system,effectiveness criteria; and (c) based on these GOM self-insures by directly funding the referralcriteria, explicitly sets priorities based on the health services of senior civil servants. The eco-resource envelope available from the government nomics of this system (in terms of medical andbudget, extra-budgetary resources, and donors' administrative costs, program benefits, andcontributions. equity effects) compared to the alternative

system of health insurance contracts should be

Staff constraints analyzed and, if feasible, contracts should beimplemented to replace direct funding.

The low pay of MOHP civil servants (generallyUS$80-100 per month) spawns coping practices that * Mortuary services and public transport of thehave a deleterious effect on service delivery, such as dead - Unlike other countries in Sub-Saharanpilferage of drugs and other supplies, misuse of Africa, GOM provides free mortuary and trans-vehicles, reduced time in service, absenteeism, and port services, a policy that has probably preventedinformal fee charging. There are significant, albeit the development of a private-sector funeral indus-unquantified, inefficiencies engendered by these try. Mortuary and transport of dead bodies can bepractices. More stringent controls are needed to operated on a full or partial cost-recovery basis, ascurb them, but the long-term solution can only be they are in neighboring countries.

with pay reform that is tied to civil service andsystem-wide "right-sizing". Civil service retrench- * External travel and external traveling allowancement can generate savings that can be used to raise - MOHP does not have information on howthe salaries of the remaining work force, similar to many of these trips are necessary and how manywhat was done in neighboring Zambia. At present, can be postponed or eliminated. A recent 2001the MTEF budget protects personal emoluments, government circular, however, noted this prob-and GOM appears not yet ready for retrenchment. lem and has specified actions to assess the levelHowever, deaths from the AIDS epidemic and staff of importance of these trips.

resignations due to the low salaries are depletingthe MOHP civil service. This may not be the best * Fuels and lubricants and maintenance of motorapproach to MOHP right-sizing since the most vehicles - More stringent monitoring of the trans-skilled - and therefore marketable - staff often leave port fleet is needed to reduce possible theft offirst. The retraining costs of new staff are also con- fuels and lubricants and vehicle mismanagement.


Table 9Recommended key actions for improved use of MOHP resources

Action Illustrative programs, services or budget items

Budget items or programs Preventive and promotive services, primary curative services, secondary curative servicesto be protected * Disease outbreak investigation and management

* Purchase of drugs and vaccines, subject to establishment of full or partial cost-recovery program for drugsdispensed in hospitals.

* Maintenance of boreholes* Maintenance of plant and equipment* Water and sanitation* Subventions to local institutions providing preventive services promotive services

Budget items or programs * Personal emoluments (for 'excess' or redundant staff)to be reduced or down-scaled * Extemal travel and allowances

e Workshops, seminars, other training* Fuels and lubricants, maintenance of motor vehicles (wastage)* Subscriptions to local and foreign institutions* Tertiary curative and rehabilitative services, in general* Administrative services deemed less necessary

Budget items or programs * All hospital construction, unless recurrent costs for their operation are assuredto be deferred * Hosting of international meetings

Budget items or programs * Imposition of full cost-recovery for patients with healh insurance, patients utilizing high-techto be restructured medical services

* Imposition of full or partial cost-recovery for specialist physician consultation, prescription drugs, radiologyand lab services, hospital food, mortuary services and transport of the dead.

* Ambulance services, use of operating theater, physio/occupational therapy, and prosthetic services* Contracting out of hospital preventive maintenance and other ancillary/support services* Conversion of direct-funded foreign referrals into a health insurance program, subject to copayments from

members

Budget items or programs * Referral of cases abroadto be eliminated * Cancer registry

Source of data: This study.

Better fleet management is also called for, espe- * Pharmaceutical supply - Potentially large effi-cially for vehicles separately funded by donors. ciencies can be gained from better physician pre-

scription, monitoring, and fees for prescription* Construction and rehabilitation - All new clinic drugs. Informants cite the widespread free-drug

or hospital construction should be deferred, policy in hospitals as the major cause of leakageunless adequate recurrent costs are found to of drugs to the private sector. Partial or full costoperate them over the medium-term. recovery is recommended.

4

Extra-budgetarysources of funds

T he country's narrow tax base and less-than- Supervisory assessments of these funds indicateT optimistic forecast of economic growth their good performance in distributing SP, paraceta-leaves little scope for a dramatic increase in mol, aspirin, eye and skin ointments, and micronu-

health sector allocation in the medium-term. The trients. The communities keep good records, and forcentral government is clearly unable to afford the the items sold, the prices charged are far less thancontinued financing of free health care at all levels people would otherwise pay at private shops or theespecially with a rapidly growing population (Mar- health center (WB 1996). To expand this program,shall 1996). The Technical Working Group on MOHP plans to provide DRF seed money, give anHealth Financing, and more recently the DfID-com- initial stock of five basic drugs (anti-malarials, eyemissioned technical work on health financing strat- ointment, skin ointment, cotrimoxazole, and ironegy, have identified extra-budgetary financing as fortification) plus cotton wool and bandages; and toone way of alleviating the adverse impact of the organize and train village health committees on DRFfiscal crunch. The key initiatives that can be pur- management. Funds from drug sales will be used tosued include community drug revolving funds, fee- restock, improve physical condition of the clinics,based hospital revenue programs, and more aggres- and provide incentives to clinic staff.sive claiming of health insurance reimbursements The issues facing the DRF program are: (a) thefrom medical aid schemes. financing, intemational sourcing, and replenish-

ment of drug stocks once the PHN project ends;Drug revolving funds (b) the choice of drugs that will be under the revolv-

ing funds, and how specific needs of communitiesAbout 600 villages can be tapped to establish drug can be addressed; and (c) the stalled reforms withrevolving funds (DRF). The World Bank provides respect to the Central Medical Stores(CMS). Theassistance under the PHIN Project for the establish- sustainability of DRFs hinges on the availability ofment of DRFs operating under a 100 percent cost- drugs, which in turn depends on the institutionalrecovery target. Since June 1996, one community in and financial sustainability of CMS.each of the 10 priority districts under the PHN Pro-ject has begun DRFs and has distributed the initial Cost-sharing programsstock of sulfadoxine/pyrimethamine (SP), paraceta-mol and aspirin. The village health committees Statusmanaging these DRFs store and dispense the drugs,keep the funds generated, and order new supplies Government district hospitals operate generalfrom the health centers. wards and, in general, impose no fees. Discussions

23


on instituting user charges within government hos- MOHP recently permitted district hospitals topitals were at an advanced stage in the early 1990s introduce cost recovery in private wings. There are(Rep. of Malawi 1992) but they were never part of also discussions on the need to recover costs fromthe policy component of the IDA-funded PHN Pro- services such as mortuary, ambulance, radiologicalject, the government never introduced. Although and laboratory exams, prescription drugs, and usecost-sharing programs at district hospitals and of amenity wards. Full cost-recovery should be pur-health centers were initiated any formal fee pro- sued for members of medical-aid schemes and pri-grams (though fees for drugs are being charged in a vate insurance plans. Failure to do so results in anfew sites). Tertiary hospitals do operate payment ill-funded government facility subsidizing privatewards; these cost-sharing systems are currently insurance plans. Partial cost-sharing for other hos-being evaluated under a Japanese PHRD grant. pital and health-center services should be consid-

In contrast to government hospitals, CHAM ered. District hospitals should review fee schedulesfacilities have sliding-scale fees so that indigent and reimbursement rates to patients under medicalpatients are not necessarily denied treatment (Nga- aid scheme or health insurance to align them withlande 1994). Each hospital has a board that decides actual costs so that these patients are not undulyon the modest fee schedules. Though not all health subsidized by the government.units have cashiers, medical assistants and nurses Geographic or demographic targeting can pro-are trained to collect the money. Fees generated are tect the poor from fees. Poverty studies have wellused to purchase drugs and medical supplies. established the characteristics of the poor in

Local-authority health facilities that receive budg- Malawi. Poor households tend to live in rural areas,etary support from the Ministry of Local Govern- especially in dense central and southern regionsment also charge fees. The lack of common policy and along the Mozambican border; some live inbetween MOHP and local-authority facilities, and bomas (townships); others in pockets of cities espe-between government hospitals and subvented cially Lilongwe. Poor households are almost alwaysCHAM hospitals, has led to an anomaly where cer- headed by someone with little or no education; thetain districts have fee programs while others do not. poor also tend to have a high dependency ratio.

Assuming such a cost-sharing program gets imple-Proposals mented, what is the forecast for fee revenues? The

Liverpool Associates in Tropical Health (LATHMOHP proposals call for a gradual approach of 1999) developed a simple model for assessing theestablishing fee programs, first at the central hospi- potential revenue from user fees. The model's hypo-tals, then at the district hospitals, and finally at thetical district comprised one district hospital, 12health centers and clinics. The government has indi- health centers, four dispensaries, and 10 healthcated that urban middle- and higher-income groups posts. Table 11 shows five scenarios and their corre-should bear the cost of care at government hospitals. sponding revenue uptakes (net of assumed running

A cursory examination of the services provided by costs) for a typical district and nationwide.central hospitals indicates considerable scope for fee Note that the most optimistic scenario (5), if donerevenues (Table 10). MOHP and hospital management now, yields MK 7.9 million per district (or MK 189.6need to focus on (a) strengthening the private patients' million nationwide), which is equivalent to aboutpaying scheme, which requires private patients to pay 15 percent of the recurrent expenditure budget forthe actual cost of care received; (b) reviewing and FY2000/01. Note also that the above scenariosupdating reimbursement schedules for patients with exclude central hospitals (Queen Elizabeth, Zombaprivate health insurance coverage; and (c) costing the Central, Zomba Mental, Lilongwe Central, andhospital services that are amenable to fee imposition most recently Mzuzu Central), some of which haveand developing fee schedules for them (such as high- larger revenue potential than lower-level facilities.tech medical services, prescription drugs, radiology The above figures, therefore, should be viewed asand laboratory services, mortuary services, and phys- lower-bound estimates for potential user-fee rev-iotherapy and occupational therapy). enues of the entire government health system.


Table 10Recommended restructuring of selected hospital services in Malawi

Services Remarks Proposed restructuring

Patients with medical aid scheme WVith ability to pay Subject to full cost-recoveryor health insurance coverage

Prvate patients paying scheme With ability to pay Subject to full or partial cost-recoveryin amenity wards

High-tech medical services Expensive and probably socially Subject to full cost-recovery(CT-scan, dialysis, etc.) cost-ineffective

Specialist physician consultation Fee basis in most countries Subject to full or partial cost-recovery

Pharmaceutical supply Fee basis for prescription drugs in Subject to full cost-recovery; reduce wastage throughmost countries treatment protocols, drug registers, much-improved

monitoring, and physician education; reorganize intoHospital Drug Capitalization Program

Radiology and lab services Fee basis in most countries Subject to full or partial cost-recovery

Food provision Included in the per-diem cost in Reduce theft and wastage; subject to cost recovery;most countries provide alternative cooking facilities

Mortuary services Private business in most countries Subject to full or partial cost-recovery(funeral parlor)

Transport of the dead Private business in most countries Subject to full or partial cost-recovery(funeral parlor)

Ambulance services Fee basis in most countries Reduce abuse of transport; subject to full or partial cost-recovery

Use of operating theater Fee basis in most countries Subject to full or partial cost-recovery

Preventive maintenance, electromedical On contractual arrangement in Contract outengineering and physical asset many countriesmanagement

Physio/occupational therapy Fee basis in most countries Subject to full or partial cost-recovery

Prosthetic services Private business in most countries Privatize or subject to full or partial cost-recovery

Source: This study.

Table 11Estimated revenues from cost-sharing program based on alternative scenarios

Revenue yield Revenue yieldRevenue yield nationwide nationwide

per district (@ 24 districts (in US$ millionScenarios (in MK m) in MK m) @US$1=MK43)

Scenario 1: strengthening OPDl and private amenity vings MK 1.4 m MK 33.6 m US$ 0.78 m

Scenario 2: nominal fee and extended exemptions MK 3.0 m MK 72.0 m US$ 1.67 m

Scenario 3: nominal fee and limited exemptions MK 4.3 m MK 103.2 m US$ 2.40 m

Scenario 4: higher fee and extended exemptions MK 5.3 m MK 127.2 m US$ 2.96 m

Scenario 5: still higher fees and extended exemptions MK 7.9 m MK 189.6 m US$ 4.41 m

Source: LATH, 1999.

5

Private healthexpenditures

Mission/CHAM facilities enues, are managed by independent committees;fee schedules are made at the local level.N | ongovernmental organizations (NGOs) In districts where there are no government hos-N play a critical role in Malawi's health sector. pitals, CHAM facilities may be designated to fulfill

Mission facilities, most of which are organ- their functions. These district-designated CHAMized under the Christian Hospital Association of hospitals receive financial support (subvention)Malawi (CHAM), are the largest group, accounting from the Ministry of Finance covering most of theirfor 19 percent of all facilities, 38 percent of beds, and salary costs. CHAM subventions dramatically10 percent of outpatient visits (Table 12). Most of increased from MK4.6 million in FY90/91 tothese church-sponsored health institutions predate MK22.2 in FY94/95, but the budget support pro-the government health system and provide training gram dried up the following year due to the fiscalfor nurses and other health personnel. Though crisis. In the mid-1990s, CHAM facilities claimedorganized as a network, CHAM facilities operate that 42.0 percent of their operating funds come fromautonomously. Hospital finances, including fee rev- subvention grants, 25.5 percent from donations, and

Table 12Health service providers in Malawi: mid-1990s

Number of outpatientType of provider Number of facilities Number of beds visits per year (millions)

MOHP 339 8,156 15.8

of which Queen Elizabeth CH 1 1,057 About 0.4

of which Lilongwe CH 1 700 About 0.4

Local govemments 65 493 0.5

Missions/CHAM 164 5,648 2.1

For-profit providers and nonprofit NGOs 282 596 2.6

Total 850 14,893 21.1

Source: GOM and MOHR Basic Health Statstics 1994.

26


32.5 percent from fees, sale of drugs, and other panded. As of the end of 1996, the Registry ofinternally generated revenues. Health Professionals listed 23 medical practitioners

All CHAM facilities charge fees. District-desig- and seven dentists, all practicing in cities and largenated CHAM hospitals have private wings that towns; 60 medical assistants; 30 clinical officers; 67generate revenues. Patients in general wards also registered premises for private practice; and 19pay but at a much-reduced rate. Fees are paid in licensed specialists, medical practitioners, and den-cash though in-kind payments are also accepted. tists in part-time practice. New regulations alsoCHAM waives payment only in exceptional cases. made it possible for government health workersThe larger hospitals have written criteria and regu- (doctors and medical officers) to have second jobslations on who can be waived or exempted, but as private practitioners. As a result, moonlightingthese are not very transparent in smaller facilities. A government doctors have shown up in cities andstudy on outpatient paying patterns commissioned moonlighting paramedics in periurban areas. Drugby UNICEF in 1995 indicates that 95 percent of all shops are also a convenient source not only of phar-CHAM patients pay treatment fees, 19 percent pay maceuticals but sometimes of medical advice. Theassociated health-service fees, and 22 percent pay College of Medicine graduates 20 doctors a year (96for transport costs. In contrast, only a negligible have graduated since it opened in 1991), and thepercentage of MOHP patients pay treatment fees, privatization of the medical sector could boost itsonly 12 percent pay associated health service fees, output. So far, there has been no evaluation of howand only 9 percent pay for transport costs. CHAM this liberalization of the health sector has affectedfees are modest: a typical patient in 1995 paid MK1 health-service access, equity, or quality of care.for an access ticket, MK6 for medicines, MK2 for Traditional healers are well-established, chargingconsultation, and MK 2 for lab exams. as much as MK1O-25 for eye problems, MK40-100

CHAM facilities receive donations (in cash or in for pneumonia cases, and MK60 for common ill-kind, such as, medical equipment and supplies). nesses. An estimated 13,000 traditional healersExpenditures on expatriate health workers and reli- practice in rural and urban areas. Some of themgious orders are unknown. report seeing as many as 200 patients a month; at

MK60 per patient, this patient volume translates toPrivate for-profit health providers MK12,000 in fees collected a month, more than the

salary of government doctors.Private estates and corporations, especially those infar-flung areas, usually provide on-site health serv- Health insuranceices to employees, their dependents, and sometimesnon-employees and non-dependents. The Nchima The slow structural transformation of the economyPlantation, for instance, operates a clinic staffed and a narrow base of formal employment constrainwith five nurses and one medical assistant and sup- the development of social/compulsory or private/plied with drugs from the UK, courtesy of the voluntary health insurance in Malawi. Neverthe-Nchima Trust. The clinic charges token fees, which less, unlike its neighbors Zambia (which made pri-are fixed for a course of treatment or service (MK25 vate health insurance illegal in the 1970s) and Tan-for deliveries; MK2 for pediatric treatment; MK1O zania (which followed a socialist path for much offor adult confinement; and MK5 for child confine- its history), Malawi has a small established privatement). GOM should encourage private estates and health insurance industry.corporations that do not currently provide health The self-insuring agencies operate their ownservices to their employees and their dependents to health insurance schemes - Federal Reserve Bank,set up such services. In spite of serious limitations the Post Office, National Bank, Inde Bank, Admarc,in the public sector, there is very little discussion of Levy Brothers, etc. The government encouragesthe role of these in-plant health services. parastatals to establish health insurance and med-

Since GOM allowed private medical practice in ical aid schemes for their employees. All parastatals1991, the for-profit sector has dramatically ex- currently have some form of health insurance for


senior staff, while a number operate clinics for a profitable enterprise and expects market expan-junior staff. Scant information exists on the organi- sion. Recent and prospective activities include:zation of these schemes, their number, coverage, (a) Acquisition a 15 percent share in the Mwaiwathubenefit packages, or utilization patterns. Hospital, probably the most modern hospital in

The Medical Aid Society of Malawi (MASM) is a Malawi, and (b) the plan to establish its own phar-subsidiary of the National Insurance Co. and was macy, actions that will make MASM evolve towardsestablished in 1983 as a nonprofit organization under a managed-care setup. Further, the self-insured com-the Trustees Incorporation Act. MASM provides panies have asked MASM to manage their schemes,health insurance coverage to employees (and de- in effect making MASM an "administrative servicespendents) of large companies and better-off self-em- only" contractor to these schemes.ployed professionals. Contribution rates of employees Moral hazard is a major issue facing MASM andand employers vary by company, but the benefit other private insurance schemes in Malawi. Con-packages are standardized as shown in Table 13. tracted private providers tend to over-use laborato-

MASM's operating indicators are shown in Table ries and x-rays, over-treat through prolonged con-14. Though the membership base remains small, it is finement, and over-prescribe drugs. This problem

Table 13Benefits and contribution rates of different types of health insurance plans offered by theMedical Aid Society of Malawi

Type of health Illustrativeinsurance plan Benefits contribution rates

Basic * 100 percent of the cost of inpatient confinement in government and CHAM hospitals. MK26/person/month

General * 100 percent of the cost of inpatient confinement in govemment and CHAM hospitals; MK106/person/month80 percent if in private hospital.

* 100 percent of cost of outpatient consultations + 80 percent of cost of drugs.

Extended * 100 percent of the cost of inpatient confinement in govemment, CHAM, and private hospitals. MK350/person/month* 100 percent of cost of outpatient consultations + 90 percent of cost of drugs.* 50 percent of the cost of inpatient confinement abroad (South Africa and Zimbabv),

with a cap of MK100,000 per member.

Source of data: MASM.

Table 14Operating indicators of the Medical Aid Society of Malawi: 1993/94-1996/97

Items 1993/94 1994195 1995196 1996/97

Estimated membership (including family members) 16,000 17,000 18,000 20,500

Contributions (MK Mn) 20.50 30.11 40.55 42.47

Domestic claims (MK Mn) 15.44 25.54 30.40 34.46

Foreign claims (MK Mn) 1.38 3.00 0.58 1.47

Administrative costs (MK Mn) 3.87 4.61 5.15 5.60

Income (MK Mn) - - 6.70 4.60

Number of staff 15 17 19 20

Source of data: MASM.


encouraged MASM to take greater control of the pro- Household health expendituresvision of care through direct equity investments in ahospital and a pharmacy. Table 15 shows indicators of annualized household

The second major problem has to do with AIDS. health expenditures based on a survey of one-Insurance contracts exclude AIDS care, but to avoid month expenditures conducted in 1990-91. Thelosing clients, MASM currently does not strictly expenditure pattern reflects the geographic patternenforce this clause. This approach can continue as of income and wealth (as well as the presence oflong as AIDS patients do not demand the expensive providers), with urban households spending 6.5cocktail of AIDS drugs available in Western coun- times as much as rural households on health serv-tries. As soon as they make such demands, the pre- ices. The Blantyre figures are remarkably high, withmium rates will have to be adjusted to such an a typical household spending MK107 (or US$40) onimpossibly high rate that private health insurance health care a year, in contrast to a typical ruralmay become unviable, throwing such patients back household which spends only MK8-12 (US$3-5) ainto the ill-funded government system. This is a year. Rural household underspending can be allevi-critical issue that GOM must tackle. ated with more intensive public spending, but the

Fiscal difficulties have triggered proposals to trend in MOHP expenditures does not point in thisexplore social health insurance for the formally direction.employed, but Malawi's National Health Plan pro- In March 1995, UNICEF conducted anothervides no specific actions to pursue it. survey (Table 16) showing the persistence of rural

Table 15Annual household health expenditures in Malawi: 1990-91 (in Malawi Kwacha)

Doctor Dental Fees forand and paramedical Pharma- Other Health

Hospital hospital optical and traditional ceutical medical insuranceLocation fees OPD fees fees healers expenses expenses contributions Total

Blantyre City 28.32 48.60 7.56 5.88 12.24 0.84 3.60 107.04

Lilongwe City 0.24 23.64 0.00 4.32 12.00 0.24 14.16 54.60

Zomba Municipality 1.92 22.68 24.60 5.40 16.44 0.60 0.00 71.64

Mzuzu City 5.76 24.00 0.60 5.28 15.36 0.84 2.52 54.36

Unweighted averagefor urban 9.06 29.73 8.19 5.22 14.01 0.63 20.28 71.91

South, rural 0.24 3.00 0.12 1.80 2.76 0.00 0.00 7.92

Central, rural 0.48 4.80 0.12 2.88 4.44 0.12 0.36 13.20

North, rural 0.12 1.08 0.12 3.12 7.68 0.00 0.00 12.12

Unweighted averagefor rural 0.28 2.96 0.12 2.60 4.96 0.04 0.12 11.08

Unweighted averagefor all 4.67 16.35 4.16 3.91 9.49 0.34 10.20 41.50

Note: Annualized from reported monthly figures.

Memo item: The exchange rate to the US$ was MK2.7 in 1990 and MK2.8 in 1991.

Source of basic data: National Statistics Office, 1994.


Table 16Reported paying patterns of poor and nonpoor households for health services,by urban and rural location: 1995

Income status Reported poor Reported nonpoor

Geographic location National Urban Rural National Urban Rural

Percent of householdsreported not paying 47 percent 58 percent 56 percent 61 percent 47 percent 63 percent

Percent of householdsreported paying 43 percent 42 percent 44 percent 39 percent 53 percent 37 percent

Average monthly MK 12 MK 31 MK 9 MK 23 MK 60 MK 15(Annual) amount paid (MK 144) (MK 372) (MK 108) (MK 276) (MK 720) (MK 180)by paying households

Memo item: The exchange rate to the US$ was MK15.3 in 1995.Source of basic data: UNICEF, March 1995.

households' underspending on health services means are paying for care. These findings strongly(MK180 a year) relative to urban households indicate the need to sharpen the focus of fee pro-(MK720 a year). Analysis of survey results also indi- grams (imposing fees on free-care urban facilitiescates the degree of leakage and undercoverage of where those with capacity to pay are mostlyfee programs in Malawi. At the national level, as located) and to establish clearer waiver and exemp-much as 61 percent of nonpoor households get tion systems in both urban and rural facilities. Thehealth services without paying (leakage) while as survey did not disaggregate respondents by themuch as 43 percent of poor households end up type of facility used (MOHP, CHAM, local authori-paying (undercoverage). Thus, many households ties, private for profit), making it difficult to recom-with capacity to pay for health services are free- mend institution-specific reforms.loading while just as many households with scant

Donor financing andexpenditures

M alawi hosts five bilateral and eight multi- ment", which garnered 21 percent of total donorN 4 lateral donors in the health sector.' From allocation during the period. Education and humanFY94-FY97, these had an aggregate coun- resources took 15 percent, and agriculture, 14 per-

try allocation of around US$1.4 billion (or around cent. The respective percentage shares of the rest ofUS$350 million a year), of which 19 percent was the sectors (natural resources, transport, energy,allocated to health. In terms of sector allocation, social development, and so on) were less than 6 per-health was eclipsed only by "economic-manage- cent each.

Table 17Donors' health sector allocations and expenditures in Malawi: FY94195-FY97198

Percent changeItems FY94/95 FY95196 FY96197 FY97198 FY94-FY97

Donors' country budget allocation (US$ m) 319.3 376.0 331.9 375.6 17.6

Donors' actual expenditures (US$ m) 259.5 307.0 262.2 349.2 34.6

Donors' health sector allocation (US$ m) 50.4 57.1 72.3 86.0 70.6

Donors' health sector expenditures (US$ m) 28.2 35.7 50.0 47.1 67.0

Percentage of health sector allocationto total allocation 15.8 15.2 21.8 22.9 -

Percentage of heafth sector expenditureto total expenditure 10.9 11.7 19.1 13.5 -

Population of Malawi (m) 10.25 10.57 10.91 11.24 9.7

Donors' health sector expendituresper capita (US$) 2.58 3.38 4.58 4.19 62.4

Note: Donors have varying fiscal years: (a) Jan. 1-Dec. 31 - GTZ, EU, UNAIDS, UNDP, UNFPA, UNICEF, WFP, and WHO; (b)Apr. 1-Mar. 31 -CIDA, Dftl, JICA, IDA; (c) Oct. 1-Sept. 30 -USAID. No attempt vwas made in this study to transform the original responses to conform with the Govemment of Malawi's fiscal year.Sourc of basic data: Survey of donors.

31


Allocation vs. expenditures per capita health expenditures in Malawi have notdeclined in the face of drastic reduction of govern-

Donors' allocation to health dramatically increased ment health spending.from 15.8 percent in FY94 to 22.9 percent in FY97 The multilateral agencies (IDA, UNAIDS,(Table 17), reflecting their collective commitment to UNFPA, UNICEF, the World Food Program, WHO,the sector. By FY97, the share of health expenditures and the European Union) account for around 48.4to total donor expenditures had reached 13.5 per- percent of donor allocations in the health sector; thecent. The level and pace of donor health expendi- bilateral donors (CIDA, DfID, GTZ, JICA, USAID)tures far exceeded population growth so that per contribute the remaining 51.6 percent. In all, therecapita donor health expenditures rose markedly are 34 discrete projects currently implemented byover the four-year period to reach US$4.19 in FY97, donors (Table 18). The projects vary in size, theafter peaking at US$4.58 in FY96. It is largely due biggest ones being DfID's Malawi Reproductivethis increased donor-funding of the sector that total Health Project; EU's Health Reform and Decentral-

Table 18Budget allocation and expenditures of donor projects: as of end-FY97/98(in thousand U.S. dollars)

Donor/project name Life of project* Budget allocation Expenditures to date

CIDA

Family health project 4/96-3/98 1,032 1,032

Social sector grant 4196-3197 1,845 1,845

Southern Africa AIDS training 5/97-12197 143 143

DflD

Contraceptive supply and reproductive heafh 7/93-12/97 1,416 1,369

Anesthetic training and support 10/93-3/98 700 692

Technical assistance to health sector reform 8/94-7/98 600 564

Malawi reproductive heafh project 9/94-3/01 16,518 6,912

Support to national AIDS coordination program 7/95-12/97 404 384

Interim contraceptive supply project 8/96-12197 1,180 972

Support to national tuberculosis control program 8/97-7/99 1,771 1,055

European Union

Rural health program (building and equipment) 12/87-n.d. 8,882 7,078

STD prevention project 1/93-n.d. 1,104 995

Health sector project identification 8/94-n.d. 170 162

Technical assistance to MOHP 10/96-n.d. 481 108

Health reform and decentralization 10/96-n.d. 17,211 76

HIV/AIDS prevention 7/97-n.d. 718 0

Table 18 continues on next page


Table 18 (continued)

Donoriproject name Life of project* Budget allocation Expenditures to date

GTZ

Strengthening of Machinga district health services 1991-2000 12,000 8,521

Strengthening of Zomba district heahh services 1997-2000 2,000 0

JICA

Community health science project 1994-1999 3,500 3,500

USAID

PHICS 6189-12/98 23,493 15,550

STAFH project 9/92-9/98 45,000 17,685

CHAPS project 9/95-9/00 15,000 53

COPE 1 project 7/95-9/97 539 538

IDA

PHN project 1991-1999 55,500 23,000

UNAIDS

Assistance to national AIDS control program 1/96-1/98 289 154

UNDP

Health component 1/93-12/98 980 980

UNFPA

Census: basic data collection 1992-2001 1,743 912

Demography training 1992-2001 849 512

Population policy 1992-2001 2,658 1,605

Family planning 1992-2001 9,355 5,974

IEC 1992-2001 5,213 3,985

UNICEF

Country program in heaHh n.a. 3,732 3,732

World Food Program

Vulnerable group feeding 1996-1998 10,161 5,305

WHO

Technical cooperation 1997-1998 300 300

Total - 246,484 115,693

* Months of initiation and terminabon of projects given when known.Source of basic data: Survey of donors.


ization Project; GTZ's strengthening of district priate delegation authorities within governmenthealth services in Machinga and Zomba; USAID's need to be established for officials who are away.PHICS, STAFH, and CHAPS Projects; IDA's PHNProject; and WFP's Vulnerable Group Feeding * Slow compliance with legal, policy, program, orProject. administrative requirements and conditions for

disbursement.Project pipeline

* Weak capacity to manage and follow up requiredAn analysis of the lives of these projects indicates actions for donor funds to flow or for procure-that most of them were planned to end within FY98 ment to proceed; loose coordination between theor in the next two fiscal years. However, due to the MOHP, Ministry of Finance, and local govern-slower-than-planned disbursement rates, it is likely ments; miscommunication between donors andthat they will be extended beyond their current government or NGO counterparts.closing dates. To date, the ongoing projects havecollectively expended only US$115.7 million, or 46.9 * Time-consuming and possibly onerous procure-percent of the aggregate budget allocation of ment procedures, lack of understanding of theseUS$246.5 million. In the period FY94-FY97, donors procedures, or lack of qualified staff knowledge-were able to spend only an average of 60.5 percent able in these procedures.of their annual health sector allocation (Table 19).

For existing projects with closing dates beyond * The sheer number of projects and activities - 34FY97, we calculated expected disbursements for FY98 at present - that overburdens the alreadyto FY01, and the results are shown in Table 20. The strained capacity at central MOHP HQ.analysis shows that even under the most optimisticmanagement scenarios, there will be significant Health service focus of donor projectsamounts of already-committed donor resources to bedisbursed: US$72.5 million in FY98, US$32.3 million Donor projects focus principally on primary, pre-in FY99, US$15.5 million in FY00, and US$10.5 million ventive, and promotive care. As Table 21 shows, 64.7in FY01. Given the projects now in the pipeline, as percent of the total number of projects, 84.2 percentwell as major projects being considered, designed, or of their budgets, and 87.3 percent of their expendi-negotiated (such as those of DflD, IDA, JICA, tures to date are oriented to primary health services.USAID), the government and its donor-partners The most frequently cited objectives of these projectsshould begin serious discussions about how donor are prevention of HIV/AIDS and other sexuallyhealth expenditures can be expedited, especially in transmitted infections, family planning, familylight of the very tight government fiscal situation now health, and central- and district-level support to pri-and in the foreseeable future. Donor resources should mary care interventions. In recent years, malaria andcome in handy during a budgetary crunch, but unless tuberculosis have received increasing focus.the government's capacity to absorb and donors' ownability to disburse are addressed, donor resources Uses of donor resourcesmay not be able to quickly cushion the impact of gov-ernment underspending in health. Donor response to the survey on the use of their

The following issues hindering faster disburse- health expenditures was poor; as a result, as muchments need to be addressed: as a third of the recurrent and 7 percent of the capi-

tal expenditures cannot be classified. Of their total* GOM's constrained ability to meet counterpart health expenditures in FY96 and FY97, donors

funding requirements. claimed that around 87 percent were for recurrentand around 13 percent were for capital expenditures

* Administrative bottlenecks at central MOHP (Table 22). Thus, even though donor-funded projectsHQ, especially the absence of signatories. Appro- are reported in the Development Account of the


Table 19Donors' budget allocation and expenditures in the health sector in Malawi: FY94195-FY97198(in thousand U.S. dollars)

Annual averageDonors FY94195 FY95/96 FY96197 FY97198 FY94-FY97

Budget allocation

CIDA 1,923 1,923 3,778 3,778 2,851

DfID 4,768 4,768 3,263 4,000 4,200

EU 9,572 9,572 27,520 28,560 18,806

GTZ 4,000 4,500 1,667 2,520 3,172

JICA 2,510 3,260 3,610 3,971 3,338

USAID 7,085 11,547 9,573 12,826 10,258

IDA 11,000 11,000 11,000 17,700 12,675

UNAIDS 0 0 69 220 72

UNDP 113 77 191 240 155

UNFPA 1,478 2,546 3,288 3,785 2,774

UNICEF 4,050 4.050 4,050 4,050 4,050

WFP 2,963 2,963 3,387 3,387 3,175

WHO 920 920 920 920 920

Total 50,383 57,126 72,316 85,956 66,445

Estimated expenditures

CIDA 1,923 1,923 3,778 3,778 2,851

DfID 1,716 1,716 2,542 2,900 2,219

EU 5,955 6,341 7,400 8,385 7,020

GTZ 2,000 2,500 2,000 2,000 2,125

JICA 2,510 3,260 3,610 3,971 3,338

USAID 2,878 5,103 6,731 9,277 5,997

IDA 2,300 4,770 12,633 5,827 6,383

UNAIDS 0 0 35 118 38

UNDP 113 77 191 240 155

UNFPA 1,478 2,546 3,288 3,784 2,774

UNICEF 3,732 3,732 3,732 3,732 3,732

WFP 2,593 2,770 3,177 2,128 2,667

WHO 920 920 920 920 920

Total 28,118 35,658 50,037 47,060 40,218

Ratio of annual expendituresto budget allocation (percent) 55.8 62.4 69.2 54.7 60.5

Source of basic data: Survey of donors.


Table 20Pipeline analysis of major donor projects in the health sector in Malawi: FY98/99-FYOI/02(in million U.S. dollars)

Donoriproject Closing date* Pipeline to date FY98199 FY99100 FYOOO1 FY011 02

DflD's Malawi reproductive heafth project 3/01 9.6 2.4 2.4 2.4 2.4

DfID's other projects 1999 1.0 .5 .5 - -

EU's health reform and decentralization project 2001 17.1 4.3 4.3 4.3 4.3

EU's other projects 2001 3.0 .8 .8 .8 .8

GTZ's strengthening of Machinga andZomba district health services 2001 5.4 1.4 1.4 1.4 1.4

USAID's PHICS project 12/98 7.9 7.9 - - -

USAID's STAFH project 9/98 27.3 27.3 - -

USAID's CHAPS project 9/00 14.9 5.0 5.0 5.0

IDA's PHN project 1999 32.5 16.3 16.3 -

UNAIDS's assistance to NACP 11/98 .1 .1 - - -

UNFPA's four population-related projects 2001 6.8 1.7 1.7 1.7 1.7

WFP's vulnerable group feeding project 1998 4.9 4.9 - - -

Total - 130.8 72.5 32.3 15.5 10.5

'Month of termination given when known.Source of basic data: Survey of donors.

Table 21Donor projects by health service focus in Malawi: as of end-FY97/98 (in thousand U.S. dollars)

Number of Percentage Expenditures Percentage ofHealth service focus projects Total budget of budget to date expenditures

Primary, preventive and promotive care 22 207,605 84.2 101,021 87.3

Basic curative care/district health services 2 14,000 5.7 8,521 7.4

Tertiary care 1 700 0.3 692 0.6

Health reform, decentralization, census,technical cooperation not classifiedelsewhere, and other activities 9 24,179 9.8 5,459 4.7

Total 34 246,484 100.0 115,693 100.0


government budget, most of these are really recur- ect monitoring units, vertical disease programsrent rather than capital investment expenditures. (AIDS, tuberculosis, primary health care), rural

Of the US$43.6 million recurrent expenditures in health units, and the Queen Elizabeth Hospital. TheFY96, as much as 16.6 percent was provided by six level of this direct departmental and administrativedonors as departmental support to the MOHP, proj- support was maintained in FY97, and two more ver-


Table 22Donor expenditures by major classification: FY96/97 and FY97/98 (in thousand US dollars)

FY96197 FY96197 FY97198 FY97198Expenditure Classification amount percent share amount percent share Key Donors

Recurrent expenditures 43,574 87.1 40,523 86.1

Departmental support to MOHP 8,487 16.6 7,788 16.5 CIDA, DfID, EU, IDA, UNDP, UNFPAand administration

Subcontractors 1,106 2.2 1,951 4.1 DfID, USAID, UNFPA

Technical assistance 8,382 16.8 11,314 24.0 CIDA, DfID, JICA, USAID, IDA, UNAIDS,UNICEF

Research 205 0.4 245 0.5 JICA, USAID

Training 3,826 7.6 4,505 9.6 CIDA, DfID, JICA, USAID, IDA, UNFPA

Transport O&M 670 1.3 824 1.8 DfID, IDA, UNICEF

Drugs and contraceptives 4,867 9.7 1,378 2.9 CIDA, DfID, USAID, IDA, UNICEF

IEC 206 0.4 627 1.3 DfID, EU, IDA, UNICEF

Other operating expenses 603 1.2 3,098 6.6 DfID, EU, USAID, IDA, UNDP, UNFPA

Unclassified recuning expenses 15,222 30.4 8,793 18.7

Capital expenses 6,463 12.9 6,537 13.9

Building construction and rehabilitation 926 1.9 1,680 3.6 DfID, JICA, IDA, UNICEF

Vehicles 108 0.2 768 1.6 EU, IDA, UNICEF

Office equipment 460 0.9 1,079 2.3 IDA, UNFPA

Medical equipment 1,626 3.2 717 1.5 DfID, JICA, IDA

Unclass. cap. exp. 3,343 6.9 22 4.9

Total 50,037 100.0 47,060 100.0


tical programs received support (malaria and schis- under African/local NGOs (Table 23). Together,tosomiasis). Thus, there appears to be an increasing NGOs - defined broadly - oversee more than halftrend among donors to strengthen vertical disease (55.9 percent) of the ongoing projects.interventions. Training is also a major user of donor funds.

Subcontracts and technical assistance weighed Close to a tenth of all donor expenditures in FY97heavily, with as much as US$9.5 million devoted to went on training, which includes long-term trainingthese two expenditure items in FY96, and increasing mostly out of the country (US$856,000), short-termto US$13.3 million in FY97. Concern is being raised training mostly out of the country (US$600,000), in-of the costliness of these inputs, but in view of the country and out-of-the-country workshops (US$2.2weak implementation capacity of the government, million), and other unclassified training activities.technical and management assistance may be a The US$4.5 million annual cost of these trainingneeded supplement it. As many as 16 projects or activities is staggering for a country the size ofdiscrete activities are being managed by foreign Malawi, a bureaucracy the size of MOHP, and anagencies, NGOs, and contractors; three others are NGO community which is at best fledgling. (There


Table 23 using the same level of resources - would be forImplementing agencies of donor projects in sponsoring donors to provide these as direct salary

Malawi: as of end-FY97198 supplements. The US$4.5 million would translate to

Number of US$473 salary increase for one year for each of theType of agency Implementing agency projects 9,500 civil servants (about US$40 per month, or 50

National College of Medicine 1 percent of the US$80 monthly salary of a typicalgovemment District health offices 3 civil servant). Salary supplements would probablyagencies Family health unit 2 have a more visible impact on health service deliv-

MOHP 21 ery compared to training, which takes serviceMinistry of Local Government 1Ministry of Works l providers away from their respective posts, some-Nat' AIDS Control Program 3 times for prolonged and recurrent periods. (In thisNat'l Statistics Office 2 regard, the greater use of on-site training programsNat'l Tuberculosis Control Program 1 is called for.)Population Unit 3

African/local Africare 1 Geographic distribution of donor projectsNGOs Banja la Mtsogolo 1

CHISU 1 Absence of data precludes proper analysis of the

Foreign agencies, Canadian Public Health Assn. 2 geographic distribution of donor-funded projects.NGOs, and Centers for Disease Control 1 The donor survey, however, yielded the followingcontractors Project HOPE 2 illustrative information: Of the 34 donor-funded

International Eye Foundation 2 projects, 23 have nationwide application. Twelve

International Labour Org. 1 districts host two or more donor-funded projects

Marie Stopes Intemational 1 while six districts have at least one project each. SixPopulation Services Int'l 2 districts, however, do not host a single donor-Save the Children (UK) 1 funded project (their access to donor-funded activi-Save the Children (US) 3 ties is limited to those donor projects that are nation-

Total no. of projects/discrete wide in application). Donors are prone to locate inactivities 34 relatively nicer districts; it is also these districts that

have multiple donor projects: Mangochi, five proj-Note: Total does not tally with details due to double-counting, especially of MOHP hinvolvement. ects; Salima, four; Blantyre, two; Lilongwe, two;Source of basic data: Survey of donors.

Table 24

is also a separate line item for training in the MOHP District location of donor projects in Malawi:as of end-FY97/98

budget, in addition to the donor-funded training

expenditures; in addition, around 15 percent of Type of Number ofdonor allocations are for a separate Human district Names of districts districtsResources item, apart from these health-dedicated Districts with Mangochi (5 projects), Salima (4), 12

training programs). two or more Mzimba (4) Chikwawa (3), Kasungu (3),The continued sponsorship of these training pro- donor health Mulanje (3), Blantyre (2), Dowa (2),

grams in the absence of a human-resource develop- projects Lilongwe (2), Mwanza (2), Thyolo (2),

ment plan in the MOHP and in the NGO commu- and Zomba (2)nity, as well as the absence of any mechanism to Districts with Chiradzulu, Dedza, Karonga, Machinga, 6

measure the impact of these programs on health one donor Mchinji, and Rumphihealth project

service delivery, is a serious concern. It is common health_project

knowledge that these training programs provide Districts with no Nsanje, Ntcheu, Nkhotakota, Ntchisi, 6strong financial and other incentives to participants. donor health Nkhata Bay, and Chitipa

If the objective of these training programs is implic- projects

itly to reward civil servants, a better mechanism - Total 24


Zomba, two (See Table 24). Donors have much less ties in districts, and distribute donor healthtendency to locate in the northern and/or farther resources more equitably.districts (such as Nkhotakota, Ntchisi, Nkhata Bay,and Chitipa in the north; Nsanje in the far south).

These results on locational patterns of donor 1. In addition, the African Development Bank (ADB), theGerman Kreditanstalt fur Wiedenraufbau (KfW), and the Chi-

health projects need to be examined further. The nese goverment provide significant assistance, mostly ingovernment must provide specific guidance and capital construction. The European Development Fundgeographic criteria to donors in order to provide (EDF), under the Second Family Health Project, has provided

funds for functional literacy and [EC. Because they were notaccess to health services for populations in the most based in Malawi, these donors were not included in the

vulnerable areas, avoid duplication of donor activi- survey of Malawi-based donors; thus, their activities, alloca-tions, and expenditures were not quantified.

7

Analysis of the FourthMalawi National HealthPlan, 1999-2004

T he Fourth Malawi National Health Plan was NHP as requiring upgrading, rehabilitation, or newT launched in May 1999, but its implementation construction; and standard facility/populationhas been delayed and ongoing services are ratios as established in Better Health for Africa

being provided on the basis of the last plan (1984-94). (BHA). Using a population figure of 10 million forThe NHP does not clearly lay out how it relates to the Malawi, it is clear that the Plan severely understatesrecurrent and development budgeting process, and the need for health centers while giving relativeindeed it continues to lack an implementation preference to district/rural hospitals. In fact, thearrangement. Nevertheless, it is instructive to analyze current number (82) of district/rural hospitalsits implications on infrastructure, personnel, and already exceeds the BHA standard of 66, if oneoverall recurrent costs because, although it has not counts both GOM and CHAM facilities, which dobeen formally implemented, the NHP with improve- receive Government subvention and provide healthment could be used by the GOM to marshal govern- services on behalf of the GOM. In contrast, the cur-ment as well as donor resources under a SWAp. rent number of health centers is more than 50 per-

cent below the BHA standard. (Note, too, that theInfrastructure implications "current number" column excludes private and

NGO clinics). Even if GOM-only facilities are con-Table 25 lays out the current inventory of health sidered (excluding CHAM facilities), the NHP still

facilities; the number of facilities identified by the has strong preference for district/rural hospitals (38

Table 25Currant vs. proposed vs. standard number of health facilities in Malawi

BHA facility Number neededper population in Malawi

Facility type standard ratio (Pop.= 10 M) Current number NHP proposed number

Central hospitals No standard No calculation made 3 3

District/rural hospitals 1 per 150,000 66 82 (38 GOM + 5744 CHAM)

Health centers 1 per 10,000 1,000 417 (201 GOM + 195216 CHAM)

Sources: (a) Tables 16 to 19 of the NHP; (b) Better Heaflth in Africa.

40


existing vs. 66 needed, or 43 percent below stan- and proposed number of personnel only considerdard) relative to health centers (201 existing vs. GOM staff and exclude mission, NGO, and private1,000 needed, or 80 percent below standard). for-profit health personnel. The proposed personnelClearly, the infrastructure focus of the NHP must be additions would mean that the deficit in nursesreexamined. (and probably medical assistants) would have been

Of the total set of candidate facilities for upgrad- narrowed (probably even exceeded) relative to theing, rehabilitation, or new construction, the NHP required number if one counts similar personnel inidentifies the following "priorities" (52 facilities + the private sector. However, it would appear thatsupport infrastructure) with their estimated costs even with the NHP additions, doctors would still be(Table 26). Assuming that construction would be in short supply.spread out over a five-year period, this means that If one were to consider only the needs of healthan average of 10 facilities will be upgraded, rehabil- centers and rural/district hospitals, the estimateditated, or constructed per year. health personnel levels using the BHA standards

(three nurses and one support staff for one healthPersonnel implications center serving 10,000 population, and three doctors,

35 nurses, and 22 support staff for one district/ruralTable 27 lays out the current and the proposed hospital serving 150,000 population) would be aslevels of staff under the NHP. Note that the existing shown in Table 28. The problem with this and the

Table 26NHP's "priority" infrastrucutre program and cost

Infrastructure program Cost (US$ m)

Type A: 164 boreholes, 12 heath centers hooked to ESCOM, 206 solar paneling units installed, 185 radiocommunication units installed 5.45

Type B: 52 facilities (i.e., construct 2 new hospitals, replace 5 district hospitals, rehabilitate 9 district hospitals, rehabilitate3 central hospitals, upgrade 9 dispensaries to full health centers, upgrade 17 health centers to community hospitals,rehabilitate 5 rural hospitals, construct 2 new health centers) and 5 new 20-unit apartment flats 178.75

Type C: Upgrade basic technology and equipment in all district and central hospitals 13.8

Total 198.0

Source: NHP.

Table 27Current vs. proposed vs. standard number of health personnel in Malawi

Current Population NHP proposed Total number Population Number neededPersonnel type number per personnel addition of personnel per personnel in Malawi

(a) (b) (a+b) (*)

Doctors 315 31,746 251 566 17,668 1,111

Nurses 2,053 4,871 2,842 4,895 2,043 5,000

Medical assistants 442 22,624 383 3,825 2,614 -

Other staff 523 19,120 747 1,270 7,874

H.S.A.s 3,431 2,915 3,109 6,540 1,529 -

(*) Number of personnel needed is based on BHA standard population per personnel ratio of 9,000 for doctors and 2,000 for nurses. Assumed populaton of Malawi=10 million.Source: NHP and this study.


Table 28Current vs. proposed number of personnel for health centers and districtirural hospitalsin Malawi

Health centers District/rural hospitals

Personnel type Current number Needed number Current number Needed number

Doctors 0 0 158 200

Nurses 574 3,000 719 2,333

Medical assistants 236 - 159

Other staff 175 1,000 227 1,467

Source: NHP and this study.

other approach specified above, however, is that An even more serious shortcoming of the NHPthey do not consider the non-facility health pro- calculation is that it probably double-counts thegram approaches (such as outreach, community cost requirements of the health system by addingbased distribution, and mobile health teams) that the NHP infrastructure cost + GS cost estimatesmay prove to be more cost-effective and have wider derived by using the BHA standards + the PE costsreach than facility-based approaches. derived from the Managed Change Agency Report

of 1998 (as explained above). At first blush, thisRecurrent cost implications sounds reasonable enough, but the BHA standard

unit cost per capita already includes capital cost +The NHP shies away from calculating the would-be salary costs + nonsalary recurrent costs.1 There isrecurrent cost implications of the health infrastruc- some "netting out" in the NHP's application of theture proposed for construction and rehabilitation as BHA costing standard, but it is not clear how it waswell as the implied personnel complement. Instead, done. In any case, the combined PE and GS for thethe NHP focuses on (a) the calculation of PE based first year of the NHP will reach US$57.69 million, aon the Management Change Agency Report of 1998, staggering increase of almost double the US$26.9which assumes that two-thirds of the existing million expenditure estimate for district/rural hos-vacancies as of the report's date would be filled in pitals and health centers alone in FY98.the first year of the plan, and in the second year, the An alternative approach to calculating the recur-remainder of the vacancies would be filled; and rent cost implication would be to base such calcula-(b) the calculation of GS based on district health tion itself on the proposed infrastructure programfacility standards set in the BHA, and using Malawi using suitable r-ratios and add this figure to thepopulation figures. Thus, the so-called "recurrent recurrent budget of the existing health system.costs" were calculated outside (or independent of) Unfortunately, the NHP's infrastructure program isthe implied infrastructure network and personnel merely a listing of the facilities and supporting civilcomplement as proposed under the NHP. These are works and equipment, with no program phasing.critical shortcomings of the Plan (which means that For purposes of the r-analysis, it is assumed that thethe "development budget" implied in the infra- infrastructure program is spread out evenly acrossstructure plan is divorced from the "recurrent five years and, using very conservative r-ratios, arebudget," which was simply added using assump- as indicated in Table 29. The resulting estimatestions from different studies). Thus, the NHP financ- show that annual recurrent costs for these newing calculations of the development and recurrent investments would rise from US$6.41 million inbudgets remain bifurcated. Year 1 to US$32.05 mnillion in Year 5.


Table 29Estimated recurrent cost requirements of the NHP's infrastructure program(in million U.S. dollars)

Year Type A Type B Type C TotalR=0.05 R=0.17 R=0.10

Year 1 0.05 6.08 0.28 6.41

Year 2 0.10 12.16 0.56 12.82

Year 3 0.15 18.24 0.84 19.23

Year 4 0.20 24.32 1.12 25.64

Year 5 0.25 30.40 1.40 32.05

Total: 5 Years 0.75 91.20 4.20 96.15

Source: Estimated in this study.

Total cost of the NHP and options for phasing * Priority 2: Equipping of existing district/ruralhospitals, staffing, supply.

The NHP calculates the total cost of the plan tobe US$504.45 million,2 excluding the recurrent cost * Priority 3: Rehabilitation of existing district/implications of new investments, which it does not rural hospitals.calculate. One could make the assumption that theconstruction program would be spread out evenly * Priority 4: Selective rehabilitation of central hos-across the five years (that is, 10 facilities a year). In pitals.this case, the total cost translates to US$101 milliona year. If one were to add in the recurrent cost * Priority 5: New hospital construction.requirements of the new investments from the r-analysis above, the annual cost of the NHP would The NHP has been launched, but it is advisablebe US$107.41 million for Year 1, US$113.82 million to have an implementation plan to address thefor Year 2, US$ 120.23 million for Year 3, US$126.64 above concerns. In our view, the implementationmillion for Year 4, and US$ 133.05 million for Year 5. plan must have the following features so that it canClearly, from all indications, carrying out the NHP be made a basis for moving towards an eventualis a daunting task and the GOM ought to consider sector program/SWAp.options for phasing investments to achieve consis-tency with available resources and with the govern- * An infrastructure program that (a) is prioritizedment's implementation capacity. The NHP "wish- according to available resources; (b) reflects thelist" needs to be translated into a prioritized, country's ability to implement such a construc-phased, and properly costed set of activities before tion, upgrading and rehabilitation program; (c) isit can be presented to the donor community as a cognizant of the recurrent cost implications; andbasis to move forward with the SWAp. Based on an (d) takes account of the staffing and humananalysis of Malawi's urgent needs, cost-effective- resource requirements of rehabilitated/con-ness considerations, and the government's imple- structed facilities.mentation capacity, a simple prioritization rulecould be: * A supporting service delivery strategy that is

consistent with cost-effectiveness, burden of dis-* Priority 1: Rural health centers, equipping, staff- ease, and equity considerations (see Chapter 8).

ing, supply.


* A more detailed and transparent costing that laysout all the assumptions used. 1. "The overall cost was annualized by adding the yearly

costs of recurrent items, such as salaries and essential drugs,to the amortized costs of capital investments, such as build-

* A financing plan that has alternative scenarios of ings and equipment. Amortization is required to translate ini-resource availability from GOM+donors. The tial outlays of capital into an annual amount, thus yielding

fnnigpatwith the information about the yearly cost of paying off the outlaysfinancing plan must be consistent wlth the over time (assuming a loan was used to finance the capital

Medium-Term Expenditure Framework (see outlay). Capital costs were annualized on the basis of the eco-

Chapter 8). In this regard, since the NHP upholds nomic life of the assets at a 4 percent discount rate." (Betteruser fees as an alternative financing mechaniism, Health in Africa 1994, p.129).

2. Our own calculations show US$486.45 million; it is difficultthe implementation plan should also state a time- to check the accuracy of the figures because the NHP does notframe for the development of a supportive policy lay out the financing plan.

framework on cost sharing.

* An explicit underlying logical framework withsuitable monitoring indicators.

8

Efficiency improvementin the health sector

n addition to reform proposals involving The major problems of the government hospitalsgreater resource mobilization, discussions are in Malawi concern structure, financing, and man-ongoing within GOM to improve health-sector agement. Although a nominal pyramidal structure

efficiency These are not adequately dealt with in exists, the referral system operates inefficiently. Ter-the NHP, and it is useful to review some of the key tiary hospitals devote most of their resources toissues here. The initiatives involve hospital auton- basic or level I health care. It is estimated thatomy, improving pharmaceutical procurement and onlylO to 15 percent of patients at these facilitiesdistribution, decentralization of health services, receive level II or III. The Working Group on Hospi-contracting of health services, and enhancing the tal Autonomy indicated that QECH is mainly pro-role of NGOs in the health sector. viding district health services to the Blantyre area

(WGHA 1997). The poor state of primary-care serv-Hospital autonomy ices and the absence of "mid-level" regional hospi-

tals between tertiary and district hospitals are majorThere are 42 MOHP hospitals which, together with bottlenecks in rationalizing the system of referral.stand-alone government clinics/health centers and In recent years, the number and size of govern-

subvented district-designated CHAM hospitals, are ment hospitals have grown beyond what the gov-organized nominally in a pyramidal structure of ernment budget can support. Such growth may bereferral. At the apex are the three tertiary hospitals due to faulty budget allocation criteria. Historically,which also act as training institutions: Lilongwe budgetary allocations to hospitals have been basedCentral Hospital (with the attached Lilongwe on cost-per-bed, with little concern for quality out-School of Health Sciences which trains nurses, comes, actual need, or efficient use of resources.clinical officers, and hospital administrators); This provides hospitals with the perverse incentiveQueen Elizabeth Central Hospital (QECH) in Blan- either to increase beds or to inflate costs; it does nottyre, the biggest of the three (with the attached Col- provide incentive for hospitals to be resource-effi-lege of Medicine which trains doctors and nurses); cient. Indeed lower cost-per-bed hospitals may beand Zomba Central Hospital (with the attached penalized in the following year's budget by receiv-Zomba School of Nursing). The fourth central facil- ing smaller allocations while higher cost-per-bedity is the Zomba Mental Hospital. Below these are hospitals may be unduly rewarded with bigger allo-the 28 district hospitals (with the seven bigger and cations. Though lack of data precludes rigorousbusier ones at Rumphi, Mzimba, Kasungu, Dedza, analysis of this phenomenon, the use of cost-per-Mangochi, Machinga, and Mulanje), and 10 rural bed as a budget allocation method certainly needshospitals. to be reformed.

45


Poor referral and overall underfinancing of hos- financing autonomous bodies, but the granting ofpital care may also be due to the absence of a formal full autonomy will require an Act of Parliament.fee system that can rationalize hospital use and gen-erate revenues for supplies, basic repairs, and other 2. Reconfiguring the referral system, possibly byimprovements. Though discussions on cost-sharing having feeder hospitals for the central hospitalsstarted in 1990, it has not been adopted as policy. so that they can focus on real tertiary care ratherThus, fees have not been institutionalized even at than act as district hospitals.tertiary facilities, where there is the biggest scopefor revenue generation. 3. Redefining the role of the MOHP with respect to

Hospital mismanagement is manifested in inade- autonomized hospitals. With autonomy grantedquate control of funds, drugs, medical and other to key hospitals, MOHP's role is expected tosupplies, and assets (building, equipment, and change. Instead of instructing, regulating, con-vehicles); low staff morale and poor attitude to trolling, and supervising most aspects of hospitalpatients; and generally poor standards of care. management, the ministry will instead establishThese symptoms originate from a deeper set of goals, targets and policies for hospital servicescauses, including hospital managers' lack of author- that it wants provided, and then pay hospitals ifity to make key decisions, poor accountability, lim- they meet these requirements (WGHA 1997). Theited managerial capacity, ineffective structures and specific roles of MOHP under this scenario are:management systems, and organizational culture of (a) to set national policies and plans for provid-top-down rules and regulations (WGHA 1997). ing equitable hospital care; (b) to determine the

GOM is looking more closely at the problem of level, type, and quantity of services that the min-hospitals and realizes the need to improve hospital istry will pay the autonomous hospitals tomanagement, increase their accountability, improve deliver; (c) to determine hospital investmentsthe quality of hospital services, and increase the level and capital works based on national priorities toof internally-generated revenue. The MOHP has or- be developed; (d) to establish hospital standardsganized the Working Group on Hospital Autonomy and to monitor compliance and performance;(WGHA) to come up with preliminary policy reform and (e) to provide incentives for quality and effi-and hospital restructuring proposals to achieve these ciency improvements. A key theme of the pro-objectives. The second health sector vision 2007 in posed reforms is the conversion of existingthe draft Strategic Health Plan is "to improve the MOHP-hospital relationships into "contracts"quality of hospital care from central and large district that make a clear distinction between MOHP ashospitals by decentralizing their management to purchaser/regulator of services and theautonomous Hospital Boards." In general, the autonomous hospitals as providers of services.reform proposals identified by Working Group onHospital Autonomy are in the right direction and 4. Improving the governance of autonomous hospi-should be supported. These four proposals are: tals. Hospitals granted autonomy will be man-

aged by individual boards, each of which is envi-1. Providing of legal autonomy to the three central sioned to be a legal corporate body. The Act of

hospitals and seven "busy" district hospitals Parliament will have to define the roles, compo-(Rumphi, Mzimba, Kasungu, Dedza, Mangochi, sition and appointment, powers and authorities,Machinga, and Mulanje) and enabling legal and accountabilities, and specific duties and respon-policy environment for this purpose. The sibilities of these boards, as well as their relation-autonomous hospital is defined in the context of ship with the hospital management, with theMalawi as a legal corporate entity under the communities they serve, and with MOHP. Thedirection and control of a hospital board that is Working Group also proposes the establishmentseparate from the MOHP but accountable to it, as of a Council of Hospital Boards consisting of rep-well as accountable to the communities it serves. resentatives from the individual hospital boards.The GOM's goal is to make these hospitals self- The Council is envisioned as a "trade organiza-


tion" acting in the interests of the autonomous should also consider the feasibility of privatizing orhospitals vis-a-vis the MOHP in such aspects as contracting out hospital ancillary services (such asstaff salaries, pensions, and benefits; hospital fee preventive maintenance and repairs).schedules; staff training; bulk procurement and Finally, hospital statistical databases should beequipment sharing; and coordination of manage- strengthened. There is a serious lack of informationment systems and procedures. on the status and performance of these hospitals

(bed capacity, other assets, personnel, budget allo-Hospital financing must also be reformed. This is cations and spending, catchment area and origin of

the weakest aspect of the proposed hospital restruc- patients, inpatient census, case-mix, average lengthturing in Malawi. While the Working Group recog- of confinement, outpatient attendance, other servicenizes the need to reform hospital financing, it pro- delivery statistics, patients' ability and willingnessvides little guidance on how to do it. There is no to pay, and so on). While hospital autonomy isdiscussion on reducing government subsidies going almost universally prescribed in the existing docu-to these facilities or the need to broaden their rev- ments on health reform, no in-depth study can beenue base. The proposal implies that the same scope done unless these data are generated. From theseand level of services will be provided and that the data, rational decisions can be made.quality of services will be improved. This is more To provide a more rational basis for developing aoptimistic than realistic, given constraints in the strategy of hospital reform, a study should be com-GOM budget. A more thorough hospital restructur- missioned to analyze the feasibility of extendinging is necessary, based on (a) a realistic assessment greater autonomy to the tertiary hospitals (and pos-of the budget and available extra-budgetary sibly the seven largest district hospitals). The scoperesources, (b) the most essential and cost-effective of the study should include analysis of (a) theclinical services that these hospitals should be pro- degree of autonomy these three Malawian hospitalsviding, and (c) the most cost-effective way to pro- currently have with respect to management,vide ancillary services. staffing and personnel; budgeting and financial

MOHP and hospital directors should explore resource base; procurement; and quality improve-opportunities for broadening the revenue base of ment; (b) legal, policy, organizational, management,these hospitals. As GOM reduces its subsidy to the and financial requirements for increasing/improv-three hospitals, they should be made more finan- ing their autonomy; (c) quantitative targets forcially viable through other financing mechanisms reducing government subsidy and for increasingsuch as (a) cost recovery programs; (b) health insur- alternative revenue sources such as fees, insuranceance reimbursements (an immediate activity should reimbursement, grants, nonmedical revenues, andbe the review and updating of reimbursement so on; and (d) service delivery targets.schedules for health insurance firms and medicalaid schemes); (c) corporate contracts to provide Improving pharmaceutical financing,care, including private amenity wings; (d) nonmed- distribution, and useical revenue programs (parking fees, cafeteria,training fees, research, etc.); and (e) direct grants Financing and distributionfrom external sources.

MOHP and hospital directors should also The Central Medical Stores (CMS) is set up as aexplore opportunities for cost-containment and effi- Treasury Fund to procure, store, and supply phar-ciency enhancement at the hospitals. These include maceuticals and medical supplies and equipment toreduction in drug theft; much improved use of MOHP and CHAM health facilities. It has an annualdrugs through better physician prescription behav- budget set in conjunction with the Treasury, and itior; better supervision of food purchases and distri- uses these funds to tender for goods that are requi-bution (around 20 percent of hospital budget is allo- sitioned by approved hospitals and clinics. Theoret-cated to food); better staffing; and better ically, CMS is zero-budgeted and is expected tosupervision of transport fleet. Hospital directors cover its operating costs by appropriately charging


its users (Norman and Dawbarn 1999). However, ment, which may require revision of the currentbecause of overall budget difficulties, poor dialogue Essential Drug List; (b) cash-limited system whichbetween MOHP and MOF regarding the release of ensures that drug supplies can be paid for, that is,funds for drugs, and the absence of a sustainable "cash and carry" system; (c) government commit-financing mechanism (including fees), MOHP hos- ment not to divert drugs to non-priority locationspitals are substantially behind in reimbursing CMS for any reason; (d) political acceptance of some formfor approved purchases. In effect, CMS is treated by of fees (cost sharing) for drugs; and (e) a coherentMOHP health facilities as a "bursary." As a result, information strategy for pharmaceutical planning,CMS is frequently in arrears. Delays in the release of procurement, and distribution. In addition to thesefunds, if not the general intermittent unavailability steps, the government must address all forms of pil-of funds, negate all potential benefits of bulk pro- ferage and commit to a transparent external report-curement. Aside from the core financing problem, ing of the CMS reform process, especially to donorsCMS also suffers from mismanagement due to who are financing drugs or providing them in-kind.weak leadership, lack of qualified staff, poor logis- In FY2000, the government raised the spendingtics, and weak support systems. Poor MOHP staff on drugs from the estimated FY98 level of US$0.80morale has also led to reported theft of drugs, either per capita to US$1.00, to be further increased infrom CMS warehouses or within the health facilities FY01. (The WHO/World Bank estimated require-themselves. These factors lead to chronic shortages ment is US$1.00 for district health services in Sub-of drugs all over the country. Saharan Africa; US$1.60 inclusive of STD drugs;

Organizationally, CMS is part of the MOHP and and US$2.00 if tertiary-level needs are included.)therefore lacks autonomy to make critical decisions. Though this increase in drug financing isAbsence of such autonomy has been seen to con- admirable, it should be underpinned by radicallystrain CMS' ability to operate as a commercial improved management of CMS to ensure greaterenterprise. On the basis of this diagnosis, a series of availability of these inputs at the patient level.DfID-funded consultancies (first in 1995-96 byDeloitte and Touche, and later in 2000 by another Provider and consumer behaviorset of independent experts) have proposed therestructuring or "transformation" of CMS, with Compared to the drug supply problem, thisconcomitant reforms in pharmaceutical financing, demand problem is not as well recognized. How-and multi-donor commitment to its recapitaliza- ever, without corresponding reforms in the waytion. Towards this end, MOHP submitted to Parlia- providers and consumers behave, improvements inment a draft CMS reform bill in 1996-97, but later the supply situation may not result in overall healthwithdrew it pending the resolution of certain legal sector efficiency. In fact, readily available drugstechnicalities raised at the Cabinet level. This may just lead to over-consumption. The issues inreform program was not implemented, and lack of this area involve: (a) the manner of dispensingprogress in this policy area has certainly con- drugs as most private doctors prefer to fill theirtributed to drug shortages nationwide. A number of own prescriptions; (b) prescription drugs beingdonors who are keen on providing drug financing made available in the open market and dispensedhave deliberately stayed away due to the absence of illegally; and (c) tendency of government physi-an enabling environment and political commitment cians, especially at central hospitals, to aggressivelyto CMS reforms. prescribe medicines even when cheaper therapeutic

Pharmaceutical sector reforms are being revived altematives are available. ("Over-prescription" ofwith more modest expectations, with DfID again antibiotics has been observed and reported anecdo-providing critical support. Cabinet endorsement to tally.) Given these problems, investigations arerestructure and improve the governance and man- needed on drug consumption and prescriptionagement of the CMS is critical, and the consultants behavior. Specific interventions can include the for-have proposed the following principles for reform: mation of hospital therapeutic committees, the(a) need-based and resource-based drug procure- development or update of drug formularies, the


requirement for hospitals to have drug and revenue cies in the local government area. It is expected thatregisters, and necessary training for hospital and local assemblies will take over the District Develop-clinic staff on the use of these registers and on better ment Committees (DDCs), which were primarilyprescription. responsible for social services at the local level prior

to decentralization.Health service decentralization

Fiscal frameworkThe government sees decentralization as bringinghealth resources and services closer to the periph- The main revenues for the local assemblies are to beery. However, the problem of pursuing decentral- traditional, locally generated revenues, central gov-ization of government and government-funded emient transfers, ceded nontax revenues, sectorhealth services in Malawi is that there are three grants, and donor financing.structures, not exactly parallel but often duplicativeand uncoordinated. These are MOHP's health a. Traditional, locally generated revenues include prop-system, MOLG-supported local authority health erty rates, ground rent, fees and licenses, com-services focusing on preventive care, and CHAM mercial undertakings, and service chargessubvented health services. MOHP, the largest including user fees for health services. The issueprovider, itself consists of the national office, four of user fees in health is particularly relevant incentral tertiary hospitals, and district health author- discussions on locally generated revenues. Thereities. In the late 1990s, the Ministry of Local Gov- is no formal MOHP user fee policy for healthernment (MOLG) also funded health services services in Malawi. However, according to thethrough 22 local authorities (LA), 19 dispensaries, Decentralization Secretariat, with devolution, the26 maternity units, and nine health centers in 22 dis- local assemblies can impose user charges, eventricts, although most of these now seem to have without a national policy on this issue.been turned over to the MOHP. In addition, thecities of Blantyre, Lilongwe, Mzuzu, and Zomba b. Under the decentralization law, the national gov-provide preventive services. Finally, CHAM oper- ernment is mandated to make available to dis-ates its own network of independently-managed trict assemblies at least 3-5 percent of nationalhealth facilities, though these often receive govern- revenues, excluding grants, as central government

ment subvention. The relationship of these various transfers. In FYOO/01, the government will notplayers under a decentralized arrangement remains effect the transfer of the whole 5 percent, due toto be clarified. capacity constraints in local assemblies. How-

ever, MK 1 million for each assembly will beLegal and administrative framework included in the budget. By the third year of the

decentralization process, it is anticipated that theThe Local Government Act No. 42 was passed by national government will make the 5 percentParliament in December 1998. The demarcation of transfer. It is not clear how much of thesewards has been completed, although these bound- resources will be devoted to health services. Aaries do not always match those of the catchment DANIDA-funded study on inter-governmentalareas of health facilities. "Cross-jurisdictional" transfer systems will assist in determining anflows of patients will be a real issue in this regard. effective mechanism for the allocation of these

Local elections were held in 2000, paving the way revenue to local assemblies.for the creation of 39 local assemblies, consisting ofan elected mayor and elected ward representatives. c. The formula for the distribution of ceded (nontax)Ex-officio nonvoting members include the tradi- revenues will be formulated by Parliament.tional authorities, five persons to be appointed fromspecial interest groups (gender, minorities, and d. Due to capacity constraints at local assemblies,technical expertise), and the M.P.s from constituen- sector grants (budgets) from FYOO/01 to FY03/04


for the assemblies will still be incorporated into districts: Mangochi, Mzimba, Salima, Mulanje,the line ministries' budget. However, Treasury Chikwawa, Polambe, Blantyre, Lilongwe,would disburse funds directly to the assemblies Thyolo, Chiradzulu, and Machinga;without passing through the line ministries. Thekey issues that remain to be resolved include the . WHO/UNICEF, which provide resources formechanisms for channeling these sector grants, IMCI in four "early use" districts: Kasungu,management of sector grants, basic performance Mzimba, Mwanza, and Blantyre;indicators for these sector grants, and prepara-tion of sector standards. * WHO support to reproductive health in eight

districts:2 Chitipa, Karonga, Rumphi, Nkho-e. Donor financing can be channeled through the takhota, Nchisi, Salima, Dowa, and Ntcheu;

District Development Fund (DDF), which gives andunconditional block grants for development pur-poses. The DDF is used by the U.N. Country * The World Bank's Population and FamilyDevelopment Fund and UNDP and the govern- Planning Project operating in three districts:ment's development budget. Individual donor Chitipa, Ntchisi, and Chiradzulu.funds are accounted for separately through indi-vidual receipting accounts. Despite this conven- f. Equalization grants are envisioned to correct dis-ience, other donors have been less enthusiastic in parities in districts, with the formula to be deter-channeling resources through DDF (or similar mined by the central government."uniform" channel). This is clearly an issue thatneeds to be resolved as there is a plethora of In advance of actual devolution, the MOHP hasdonors providing or planning to provide established each district as a "cost center" under theresources to districts including: MTEF exercise. Thus, the recurrent budget is now

reflected for each district. However, the develop-• The African Development Bank, which plans ment budget remains outside the district "cost

to focus on five districts; center"; thus, the process of disentangling the dis-trict development budget from the MOHP develop-

* The Dutch: Lilongwe; ment budget will be challenging.

* The European Union, which provides or plans Extent and pace of decentralizationto provide resources to four districts: Blantyre,Mulanje, Thyolo, and Chiradzulu; The devolution of health assets and functions is cur-

rently in Phase 1. As planned, in FY01, the "basic* GTZ: Machinga; peripheral services" will be devolved, including:

* JICA rehabilitation of district hospitals: Dowa, * Hospital services (that is, rural hospitals) otherKasungu, and Dedza; than hospitals providing referral and medical

training - Under Phase 1, district hospitals will* UNDP, which has been supporting six "local continue to be under the MOHP; thus, all doc-

impact area" districts since 1994: Nkhata Bay, tors, dentists, most nurses, and most other tech-Mangochi, Nsanje, Dedza, Thyolo, and Salima; nical staff (lab technicians, and so on) would

remain under MOHP in Phase 1. Under Phase 2,* UNFPA training of family planning and repro- district hospitals and relevant staff will be

ductive health coordinators at the districts; devolved to the local assemblies.

* USAID, whose CHAPS Project 1 partners * Health centers, dispensaries, sub-dispensaries,NGOs with four districts and eventually to 10 and first-aid posts.


* Maternity and child welfare services. will be maintained between the DHO and thelocal assembly. In Phase 2, the district hospitals

* The control of communicable diseases including will be devolved.HIV/AIDS, leprosy, and tuberculosis; the controlof the spread of diseases in the local government * MOHP is also working to strengthen financialarea. management systems at the district level. Bank

accounts have been opened for most districts.* Ambulance services.

Other overarching issues* Primary health care services, vector control, envi-

ronmental sanitation, and health education. How does decentralization affect MOHP's plan tohave a National Health Service Commission (NHSC),

For its part, the MOHP is expected to dispense as articulated in the Human Resource Developmentfunctions related to health policy formulation and Plan? It appears that, since all district hospitals and allenforcement, standard setting, quality control and central hospitals will continue to be under thequality assurance, supervision and monitoring, MOHP, a considerable number of health staff willtraining and curriculum development, international continue to be central government employees underand regional representation in health fora, and for- the civil service. It is possible that these "retained"mulation of the national health plan. civil servants can be organized as the NHSC.

To carry out the devolved functions, the local Devolved health staff are expected to becomeassemblies will be responsible for hiring and firing employees of the local assemblies and cease to behealth staff, planning for health programs at their members of the civil service. It is also expected thatjurisdictions, and managing and inspecting health district assemblies would be required to pay a min-facilities within their purview. Toward this end, the imum salary equivalent to the existing civil servicefollowing administrative actions are being initiated: rate for the particular level of the devolved staff,

and that devolved staff will convert from the exist-* A District Local Development Planning System is ing civil service pension program (noncontributory)

being introduced to enable local people to plan, to a contributory pension scheme under the localprioritize, and implement their own plans at the government assemblies. How this can be broughtdistrict level. about, given the limited revenues of districts, is a

serious issue.* The District Development Plan is a constitutional The role of CHAM facilities in decentralization is

requirement; it consolidates all sector initiatives unclear, and this issue has not been properly takenin the district. According to the MLG, two dis- into account. CHAM's 153 health units (21 of whichtricts (Nchinji and Dedza) have completed their are hospitals) account for 30 to 35 percent of healthdistrict development plans while another district services delivered in Malawi. In many districts, the(Thyolo) is about to complete its plan. CHAM facility is the only hospital, though such

facilities have not been formally designated as "dis-* At varying rates, all districts are in the process of trict hospital." In the past, the government con-

formulating their respective socioeconomic pro- tributed as much as 30 percent of the CHAM budgetfiles, which are the basis for their district plans. through subvention grants (now managed by the

Ministry of Finance), though the size of the grants* No health planning officers have been estab- decreased in the 1990s. The lack of a formal agree-

lished at the district level. The understanding is ment between the government and CHAM hasthat the District Health Officer (DHO) will be the caused confusion, especially in years when the sub-de facto health planner. Since the district hospi- vention grant was not forthcoming. Nevertheless,tals will not be devolved in Phase 1, the DHO by tradition, the District Health Officer has beenwill still be under the Ministry, but coordination expected to supervise CHAM facilities.


Contracting of health services services or ancillary health services. Appropriateprocedures had to be developed; the relevant direc-

The government views contracting of health serv- tive from the OPC Secretary has been forwarded toices, provision of autonomy to selected health facil- the Implementation Committee on Civil Serviceities, and "privatization" of specified functions as Reform for implementation.inherent parts of the overall decentralization pro- A national competitive procurement process wasgram. Under the World Bank's Second Fiscal used. The Terms of References (TORs) and RequestsRestructuring and Deregulation Program, the gov- for Proposals (RFPs) were prepared with MOHPernment committed to contract out several health support, especially on matters of standard require-functions, as laid out in Table 30. According to ments. The short-list of firms (all Malawi firms) hasMOHP, contracting out these services will focus on been developed and are as follows: (a) Cateringcentral hospitals during the first phase of imple- services in the three central hospitals - three firms;mentation. (b) Laundry services - two firms; (c) Cleaning serv-

Responsibility for contracting out is handled by ices - four firms; and (d) Ambulance services (onlythe OPC with technical support provided by the in Zomba Central Hospital) - one firm. Negotia-Public Sector Change Management Agency tions are ongoing, but the tendering committee(PSCMA). According to PSCMA consultants, the needs to speed up the process. Contracts have beencontracting out process has gone on for three years, drafted for each type of service.the prolonged process due to Malawi's lack of expe- Affected hospital staff will be handled by therience in this new area. There were no government OPC's Rationalization Unit, which has the mandaterules or procedures dealing specifically with health to retrench staff, retrain and/or redeploy staff, or

Table 30MOHP contracting out and rationalization of government functions

Current number of Expected date ofName of function civil servants affected Type of rationalization rationalization

Regional heafh offices* 70 Abolition of offices Dec. 31,1998

Cleaning services 2,645 Contracting out Apr. 1,1999

Transport services 415 Contracting out Apr. 1, 1999

Building and ground maintenance 1,132 Contracting out Dec. 31, 1998

Laundry services 159 Contracting out Apr. 1,1999

Security services 740 Contracting out Dec. 31, 1998

Catering services 46 Contracting out Apr. 1, 1999

Audit services 5 Contracting out Apr. 1, 1999

Drama, band graphics 25 Contracting out and redeployment Dec. 31,1998

Mortuary services 29 Contracting out July 1, 1999

All establishment changes 5,097 Creations, upgrading of posts especially Apr. 1, 1999field staff, e.g., nurses and doctors

Total 5,237

This has been accomplished: All three regional health offices have been abolished and the functions transferred to the relevant district hospitals.Source: World Bank, Second Fiscal Restructuring and Deregulation Program Technical Assistance Project. November 10, 1998.


have contractor(s) absorb affected staff. In the eval- subvention program, so some CHLAM facilities areuation of proposals, it is estimated that 25 percent of reportedly closing down. In general, the MOHP-affected staff can potentially be absorbed by the CHAM subvention program is drying up. GOM hascontractors. 3 PSCMA has requested the OPC to to make a fiscal commitment to support this program,make adequate budgetary provision for the initial especially in areas where government health services(upfront) payment to contractors as well as the full are unavailable and CHAM is left to provide them.cost of the contract. While MOHP services are free, CHAM services

The issue of "appropriate cost" of contracted are chargeable, leading to an anomaly that has ledservices is rather complicated, especially in a setting to persistent calls for the establishment of free-serv-where these services are being poorly provided by ice MOHP units in CHAM areas. In spite ofthe central hospitals due to budgetary constraints of CHAM's fees, it continues to draw patronage. Inthe government. As a result, current services and general, CHAM facilities are reputed to be bettercurrent costs cannot be compared directly with con- run than MOHP facilities, and the quality of serv-tracted services (with standard specified require- ices are perceived to be better. The modest fees sup-ments) and their attendant costs. port quality improvements.

The OPC's Government Contracting Unit (GCU) As has been suggested, it would be wasteful forwill manage the contracts, but the MOHP and indi- GOM to build hospitals in areas where CHAM isvidual hospital management are expected to moni- already established. There is a clear need to syn-tor the services provided. All of the contracts have a chronize fee policies of MOHP and CHAM, and totime-frame of one year, renewable for another year, work towards a uniform fee policy.based on acceptable performance. In the case ofcontracts with "asset specific" elements, the one- Overall private sector policy environmentyear timeframe may be a problem. The mission wasinformed that, in the case of the ambulance contract GOM's policy is characterized as flexible; it allowsfor Zomba Hospital, the single contractor promised participation by different NGOs (CHAM, estates,to provide as many as five ambulances, in addition industries, licensed practitioners, and for-profitto the existing two ambulances of the hospital; the providers), and includes traditional health practi-contractor will then manage the entire fleet of seven tioners (Ngalande, and others 1994). Religious mis-ambulances. In the case of the catering and laundry sions preceded the government in health servicecontracts, the central hospitals retain ownership of delivery and remain a significant part of Malawi'sthe relevant equipment. Repairs will be made by the health care system. Despite the flexible environ-hospital (government) or, if the repairs are made by ment, the private sector seems to have very littlethe contractor, the hospital (government) will reim- input in overall decisionmaking in the health sector.burse the contractor for expenses incurred. NGOs complain that the "government makes poli-

cies and decisions on health issues with little or noThe role of nongovernmental organizations involvement of other health providers," and thatin health health service goals are being made "the sole

responsibility and prerogative of small selectSubvention program groups" (Gondwe 1996).

NGO representatives recommend that (a) everyThe Christian Hospital Association of Malawi service provider should play a role in policy formu-(CHAM) operates 148 health units, mainly in rural lation and decisionmaking; (b) MOHP should beareas. Though primarily curative in orientation, 52 more consultative, for example, by inviting the pri-percent or 77 of the health units provide primary vate sector to technical committee sessions, discus-health care services (Gondwe: 1996). CHAM facilities sions on health policies, and coordination meetings;get an estimated 30 percent of their operating funds and (c) district hospitals and the private sector infrom GOM subvention. This is used mainly for each district should be organized as teamssalaries of local staff. The fiscal crisis is drying up the (Gondwe 1996).


In view of the scanty management expertise in offices and the delay in introducing user fees at gov-the public sector, CHAM has suggested sharing its emient facilities.management expertise with the government.

Specifically, it has recommended that the country bedivided into 39 health delivery areas (HDAs) to be 1. USAID support to the first set of CHAPS districts ended in

FYO1, and an evaluation is being planned. Meanwhile,headed by 15 CHAM hospitals, 24 government dis- CHAPS II is being envisioned; new NGO proposals will betrict hospitals, and the three district health offices in sought. It is likely that CHAPS II will cover six districts -

Blantyre, Zomba, and Lilongwe, where the three including, but not necessarily, the four original districts. Ifmore funds are available, as many as 10 districts will be

tertiary central hospitals are located. These HDA covered.

hospitals would then be responsible for the adnmin- 2. Support focuses on TBA (traditional birth attendant) shel-istration of health services in their areas (MOH: ters, provision of seeds, water, and sanitation.1992). The Health Policy Framework (MOHP 1995) 3. However, PSCMA informed the mission that the affected

staff have not been informed, because, if the contracts do notrecognizes the need to define the concept of HDAs materialize, the central hospitals would have to operate thein light of the review of the role of regional health services "as usual" with the affected employees.

Appendix

Unit cost of providinghealth services to thetop five diseaseconditions in Malawi

his appendix provides information on the and lab, and other supplies. The study is an attemptunit cost of providing health services to the to generate actual resource costs; hence they reflecttop 5 disease conditions for adults and chil- scarcities experienced by government facilities as

dren in Malawi, namely, malaria, diarrhea, pneu- well as wastage often reported in both governmentmonia/respiratory conditions, tuberculosis, and and private hospitals and clinics. No attempt wasHIV/AIDS. The study was conducted by KPMG/ made to adjust for severity of illness.Blantyre for the MOHP's PHN Project. It relied on The following observations can be can be1998 data from 15 sample health facilities: gleaned from the results: First, private for-profit

hospital costs exceed central government hospital* Private Hospitals (2) - Blantyre Adventist Hosp., costs by a substantial factor: as much as 120 times

Chitawira Private Hospital for inpatient malaria, 38 times for inpatient diar-rhea, 30 times for inpatient palliative HIV/AIDS

* CHAM Mission Hospitals (3) - Madisi MH, care, 14 times for inpatient pneumonia, and eightEkwendedni MH, St. John MH times for inpatient tuberculosis. To be sure, these

cost differences can very well reflect quality differ-* Government Central Hospitals (2) - Queen Eliz- ences, but the orders of magnitude are such that

abeth CH, Lilongwe CH they call attention to either or both of the following:

* Government District Hospitals (2) - Salima DH, . Government facilities are severely underfunded.Kasungu DH A detailed analysis of the cost structure of these

facilities would indicate what specific items are* Government Rural Hospitals and Health Centers scarcely being provided (drugs, lab tests, x-rays,

(6) - Mponela RH, Mzuzu HC, Mitundu HC, other supplies, food), thus driving down theirChileka HC, Mpemba HC, Kasiya HC operated cost structures.by the local authority

* Private for-profit facilities may very well beFrom these facilities, a purposive sample of 3,909 creaming off those with ability to pay. The KPMG

patients suffering from one of the top five diseases study suggests that some private facilities maywas selected (excluding those with co-morbidities). be unduly prolonging length of stay to maximizeOf these, 2,183 were inpatients and 1,726 were out- on patients' accommodation fees. Insurance cov-patients. The following costs were taken into erage worsens these moral hazard problems andaccount: staff time, accommodation and food, x-ray they need to be addressed.

55


Second, while outpatient care is in general less tals, and rural hospitals and health centers. Costcostly than inpatient care, the cost difference in pri- being a measure of quality, this finding raises thevate facilities between these two types of care is so issue of the quality of health services being. providedlarge that it raises questions about the type of inpa- at central hospitals, as well as the types/severity oftient services being given there. In other developing patients these apex hospitals treat. In the case of inpa-countries, the cost of inpatient care is usually four to tient malaria and inpatient tuberculosis, district hos-five times that of outpatient care. In this study, the pitals had costs exceeding those of central hospitals.private-facility cost of inpatient care exceeds outpa-tient care by as much as 26 times for diarrhea, nine 1. The cost levels between CHAM and governmenttimes for HIV/AIDS, and seven times for pneumo- facilities, by disease and by type of care, are notnia. (See Table 30.) Such inflated inpatient costs can too different. If CHAM facilities indeed providetempt private facilities to admit and confine patients better services, as is frequently claimed, and ifthat can otherwise be treated on an outpatient set- indeed these cost patterns reflect efficientting. This conjecture needs to be looked into. resource use, then government facility managers

Third, among government health facilities, there need to learn more from their CHAM coun-are few marked differences between the costs terparts on how to deliver services more cost-incurred by central tertiary hospitals, district hospi- effectively.

Table A-1Average cost of care per patient, by type of care, by type of disease, and by type of facility:1998 (in Malawi Kwacha)

Patient Central Rural hospitalssample Private CHAM government District and health

Disease type, by size hospitals hospitals hospitals hospitals centerstype of care provided (n) (MK) (MK) (MK) (MK) (MK)

Malaria

Inpatient care 595 3,012 61 25 56 26

Outpatient care 1,004 623 30 80 31 15

Diarrhea

Inpatient care 475 4,290 128 112 66 47

Outpatient care 203 167 18 n.d. 24 n.d.

Pneumonia

Inpatient care 629 2,723 167 199 138 126

Outpatient care 444 403 32 40 24 n.d.

Tuberculosis

Inpatient care 361 3,432 n.d. 407 654 n.d.

Outpatient care 58 n.d. n.d. n.d. 14 n.d.

HIVIAIDS

Inpatient care 123 5,931 246 197 111 74

Outpatient care 17 634 29 80 10 n.d.

Source of basic data: KPMG/Blantyre, Analysis of Hospital Facility Expenditures and Unit Costs of Heaflth Services in Malawi. Draft dated Feb. 18, 1999.


Table A-2Average cost of care per inpatient and per OPD visit by disease: 1998(in Malawi Kwacha)

Patient Central Rural hospitalsDisease type, by sample Private CHAM government District and healthtype of care size hospitals hospitals hospitals hospitals centers

Cost per inpatient

Malaria, adult 356 3,137 73 33 65 32

Malaria, child 239 2,887 48 16 47 19

Diarrhea, adult 295 4,076 191 132 85 33

Diarrhea, child 180 4,503 95 92 47 61

Pneumonia, adult 405 2,957 178 207 166 173

Pneumonia, child 224 2,489 156 190 110 78

Tuberculosis, adult 350 3,432 - 621 545 -

Tuberculosis, child 11 - - 193 762

HIV/AIDS, adult 117 7,481 246 237 111 74

HIVIAIDS, child 6 4,381 - 156 - -

Cost Per OPD Visit

Malaria, adult 616 601 33 80 35 15

Malaria, child 388 644 26 80 26 15

Diarrhea, adult 116 159 21 - 28 -

Diarrhea, child 87 175 15 - 20 -

Pneumonia, adult 301 240 45 40 28 -

Pneumonia, child 143 566 18 40 20 -

Tuberculosis, adult 55 - - - 18 -

Tuberculosis, child 3 - - - 10 -

1-1IV/AIDS, adult 10 - 29 - 10 -

HIVIAIDS, child 7 634 - 80 - -

Source of basic data: KPMGIBIantyre (1999).

2. Cost analysis is in its infancy in Malawi and programs envisioned for government hospitals;should be made a regular analytical activity for (c) the development of reimbursement schedulesprogram managers in both government and pri- for patients with health insurance coverage; andvate facilities. This type of analysis is critical for (d) the identification of problems associated with(a) program costing required by the annual resource use of specific inputs, such as food andMTEF budget exercise; (b) the development of drug supplies.fee schedules for cost recovery or cost sharing

References

Banda, Aleke K. March 24, 1995. "Budget Statement Mbewe, Patrick, M.P. May 16, 2000. Speech delivered by1995." Delivered at the National Assembly, Zomba, the Honourable Minister of Local Government at theMalawi. Consultative Group Meeting, Le Meridien Capital

Deloitte & Touche. June 1996. "Transformation Plan for Hotel, Lilongwe, Malawi.the Malawi Central Medical Stores." Muva, T.D.C. 1996. "Health Sector Decentralization in

Ferster, Geoffrey; P.H. Van Kessel; and Yaw Adu-Boahene. Malawi." Paper presented at the workshop on healthMarch 27, 1991. "Strategic Framework of the Cost- sector issues held June 30-July 2 at Nkopola Lodge,Sharing Scheme for the Ministry of Health-Malawi." Mangochi.Work Paper No. 1. Mwambaghi, F.R. July 1-2, 1996. "Public Expenditure

Franco-Miller, Lynne and others. March 1995. "Commu- Review in the Health Sector: Critical Commentary."nity Financing Experience and Options: A Study to Background paper for the workshop on health plan-Provide Baseline Information for Ministry of Health ning issues in Malawi. Mangochi, Club Makokokla.Decision-Making." Study funded by UNICEF and Ngalande, Elias E. and others. 1994."The Public/PrivateUSAID. Mix in Health Care System in Malawi." Health Policy

Gondwe, Regent L.M. June 30-July 2, 1996. "New Roles of and Planning 9(1): 63-71.the Christian Health Association of Malawi." Work- Norman and Dawbarn. June 1999. "Pre-Investmentshop on health planning issues. Mangochi, Club Health Study Funded by the African DevelopmentMakokola. Fund." Submitted to the Government of the Republic of

Huddert, Jenny February 19-March 10, 1990. "Health Per- Malawi: Health Planning, Human Resources, Gender andsonnel." Report of an assignment conducted with Environment, Volume IV, Part 5.MOH/Malawi in preparation for the proposed World Odala, T.R. July 1, 1996. "Management Realities in theBank Health Sector Credit. Funded by ODA. Health Sector in Malawi: A Discussion Paper." Club

LATH (Liverpool Associates in Tropical Health). 1999. Makokola."Development of a Health Care Financing Strategy." Price Waterhouse. October 1990. "A Study on Decentral-Visit report submitted to DflD. ization of Management Authority for Ministry of

Malawi Institute of Management (MIM). March 14, 1991. Health.""A Study to Identify Options for Reducing the Republic of Malawi Department of District and LocalVacancy Rate of the Malawi Civil Service." Government Administration. May 2000. Decentraliza-

-May 1994. "A Report of Workshops on Decentral- tion Process in Malawi. Office of the President and Cab-isation of Health Services, Planning, Delivery and inet. Lilongwe.Resources Management in the Ministry of Health." Republic of Malawi Ministry of Health (MOH). Novem-Project No. 93031. ber 1992a. Introduction of Cost Sharing in Government

Marshall, N.R. June 1996. "Government Financing in the Health Facilities: Implementation Issues.Health Sector-Prospects and Constraints." EU Tech-nical Assistance, Economic Affairs Unit, Ministry ofFinance.

59


_- 1992b. "Strengthening of Health Services in Soka, L.D. July 1-3, 1996. MOHP Workshop on HealthMalawi: Report of a Study on Co-ordination and Col- Planning Issues in Malawi. Ministry of Local Govern-laboration Between MOH and the Private Hospital ment and Rural Development.Association of Malawi." Lilongwe, Malawi. World Bank. March 19,1996a. Malawi-Human Resources

. March-April 1994. "MOH Workshops on Action and Poverty: Profile and Priorities for Action. HumanPlanning for Strengthening Decentralisation Process." Resources Division, Southern Africa Department.Club Makokola. Washington, D.C.

Republic of Malawi Ministry of Health and Population . November 10-December 6, 1996b. "Aide Mem-(MOHP). n.d. Decentralization in Malawi: The Case of oire: Population, Health and Nutrition Sector CreditMinistry of Health and Population. Implementation Review Mission."

September 1995. Health Policy Framework. Lilongwe.July 1-2,1996. "Planning the Health Sector." Back-

ground paper for the workshop on health planningissues in Malawi. Mangochi, Club Makokola.

THE WORLD BANK

1818 H Street, N.W.Washington, D.C. 20433, U.S.A.

Telephone: 202 477 1234Facsimile: 202 477 6391Internet: www.worldbank.orgE-mail: [email protected]

Malawi's public spending on health has historically been high relative toother Sub-Saharan countries and developing nations with similar GNP percapita, but the country's living conditions are among the poorest in the world.Although physical access to a health facility has improved over the years, ac-cess to functional health services continues to be limited. Major public healthissues include a high HIV/AIDS prevalence rate, poor reproductive healthand severe constraints in the availability of health personnel and drugs andother supplies.

Better Health Outcomes from Limited Resources: Focusing on PriorityServices in Malawi takes stock of Malawi's performance in the health sector.It reviews the status of the country's health expenditures, identifies issuessurrounding the level and quality of these expenditures, and provides rec-ommendations to improve resource mobilization, resource allocation and or-ganization efficiency.

The publication of this health expenditure review for Malawi is intendedto contribute to our collective knowledge about the country's health sectorand the nature of the policy challenges, and to share that knowledge wherepossible. It is our hope that as new knowledge emerges in the course of im-plementing the country's poverty reduction strategy, this knowledge will beinstrumental in overcoming health sector constraints that currently impedepoverty reduction in Malawi.

world bank document - documents & reports€¦ · world bank or any of its affiliated ... hope...

Documents