Sri Lanka - Population and Health Sector Report

Document of The W'o:rld Bank FOR OFFICIAL '6��-iLY CONFIDENTIAL Report No. 5218-CE SRI LANKA POPULATION AND HEALTH SECTOR REPORT September 20, 1984 Population, Health and Nutrition Department This document has a restricted distribution and m:1y be used by recipients only in the performance of their official duti1es. Its rontetats may not otherwiise be disclosed without World Bank authorization. FOR OFFICIAL USE ONLY CONFIDENTIAL Abstract Sri Lanka's rate of population growth remains high, despite a reduction in fertility levels. This reduction was due to increased age at marriage and contraceptive practice, caused mainly by educational and health improvements. Mortality levels are much better than for mos t low-income countries, but significant geographical differentials and morbidity still exist. The current report focuses on three particular population issues i) the need to strengthen the management and coordination of the family planning program; ii) the importance of expanding the mix of contraceptives available, in the context of a strategy to raise the contraceptive prevalence rate and to increase the proportionate use of modem contraceptive methods; and iii) the need to generate additional demand for contraception. In the health sector, the report i) emphasizes the need to improve the management�efficiency of the· existing services; ii) suggests the need for possible changes in the design of the �ew primary health care program; and iii) argues for a greater degree of cost recovery for publicly provide health services. · This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  SRI LANKA: POPULATION AND HEALTH SECTOR REPORT Table of Contents SUMMARY AND CONCLUSIONS i I. INTROOUCTION•••••••••••••••••••••••••••••••••••••••••••••• l II. POPULATION AND HEALTH SITUATION•••••••••••••••••••••••••••! A. P•lpulation •••••••••••••••• 1 0 • • • • • • • • • • • • • • • · ·�· · • • • • • • • Dynamics of Population Growth••••••••••••••••••••••••• l Proximate Determinants of Fertility •••••••••••••••••••4 Determinants of Fertility Decline••••••••••••••••••••• s Prospects for Future Fertility Decline••••••••••••••••6 Developmental Consequences ••••••••••••••••••••••••••••7 B. Health • •••••••••••••••••• o • o • • • • • • • • • • • • • • • • • • • • • • • • • • 10 General MortalitY••••••••••••••••••••••••••••••••••••lO Infant and Child Mortality •••••••••••••••••••••••••••11 1-forbidity•...........................................•12· Vulnerable Groups ....................... �···•••••••••••13 III. POLICIES •••••••••••••••••••• o••••••••••••••••••••••••••••l4 A. Population••••••••••••••••••••••••••••••••••••••••••••14 Migration ••••••••••••••••••••••••••••••••••••••••••••14 Redistribution •••••••••••••••••••••••••••••••••••••••14 Population Growth••••••••••••••••••••••••••••••••••••15 B. &alth••••••••••••••••••••••••••••••••••••••••••••••••16 IV. PROGRAM PERFORMANCE••••••••••••••••••••••••••••••••••••••17 A. Population ••••••••••••••••••••••••••••••••••••••••••••17 Program Structure and Management•••••••••••••••••••••17 Program Indicators•••••••••••••••••••••••••••••••••••19 Key Program Opportunities and Constraints••••••••••••23 B. &alth••••••••••••••••••••••••••••••••••••••••••••••••26 Organization and Management ••••••••••••••••••••••••• 26 Facilities and Services ••••••••••••••••••••••••••••••28 Manpower and Training••••••••••••••••••••••••••••••••30 Support Services •••••••••••••••••••••••••••••••••••••35 Non-Government Health Services•••••••••••••••••••••••36 Water and Sanitation •••••••••••••••••••••••••••••••••36 V. FINANCING AND EXPENDITURES ••••••••••••••••••••••••••••••36 Past Trends ••••••••••••••••••••••••••••••••••••••••••36 Present Situation••••••••••••••••••••••••••••••••••••39 Investment Program•••••••••••••••••••••••••••••••••••43 VI. ISSUES AND RECOMMENDATIONS •••••••••••••••••••••••• ••••••45 A. Population ..•.••••...•••...•••••••••••...•••••..•..••45 B. Health••••••••••••••••••••••••••••••••••••••••••••••• 47 STATISTICAL ANNEX LIST OF TABLES: Table 1: Summary of Population Projections under Three Different Assumptions of Fertility Decline, 1980-2025, Selected Years ••••••••••••••••••••••••••••9 Table 2: General Morbidity by Broad Disease Category of Cause of Death, Rank Order and Year-All Age Groups, 1971-1979••••••••••••••••••••••••••••••••••••ll Table 3: Contraceptive Prevalence Rates by Method 1979-1982 (Per_centage of Currently Married Women Using Contraception)•••••••••••••••••••••••••••••••••20 Table 4: A Comparison of Existing Health Care Delivery System and New PRC Model•••••••••••••••••••••••••••••27 Table 5: Distribution of Hospital Beds, 1981••••••••••••••••••29 Table 6: Utilization of Hospital Beds by Type of Facility, 1982•••••••••••••••••••••••••••••••••••••••30 Table 7: Distribution of Health Staff by SHS Division, 1981•••32 Table 8: Distribution of Health Staff by Type of Facility, 1981 •••••••••••••••••••••••••••••••••••••••33 Table 9: Workload of Health Staff by Type of Facility, 1981 •••34 Table 10: Government. Recurrent Health Estimates and Actual Expenditures, in Current and Constant Prices, and Per Capita, 1971/72-1983•••••••••••••••••••••••••••••38 Table 11: Expenditures on Health Care and Health-Related Activities, 1982•••••••••••••••••••••••••••••••••••••39 Table 12: Recurrent Expenditures on Health Care by Source of Finance , 198 2 ••••• t• ••••••••••••••••••••••••••••••41 Table 13: Unit Costs at Selected Types of Health Facilities, 1982•••••••••••••••••••••••••••••••••••••42 This report is based on the findings of a Population and Health Sector Mission carried out by the World Bank in June 1983. The rembers of the mission were Dr. John Hamilton (mission leader and public health specialist), Mr. Michael Mills (economist) and Dr. Jay Satia (consultant management and population specialist). Ms. Karen Lashman-Hall and Messrs. K.C. Zachariah and James Greene also contributed to the report. A preceding USAID mission had concentrated on demography and epidemiology, and its findings were made available to the World Bank mission, which concentrated on the main theme of cost-effectiveness and efficiency of existing health and family planning services and possibilities for cost saving. The mission is grateful for excellent cooperation from the Government, WHO, UNICEF, USAID, UNFPA and other agencies.  ABBREVIATIONS AMP Assistant Medical Practitioner CBR Crude Birth Rate CDR Crude Death Rate CPS Contraceptive Prevalence Survey OMO District Medical Officer ESCAP Economic and Social Commission for Asia and Pacific FHB Family Health Bureau FHW Family Heal th Worker. FPASL Family Planning Association of Sri Lanka IMR Infant Mortality Rate I PPF International Planned Parenthood Federation IUD Intra-Uterine Device MCH Maternal and Child Heal th MH Ministry of Health MOH Medical Officer of Health MPI Ministry of Plan Implementation NGO Non-Government Organization NIHS National Institute of Health Science NORAD Norwegian Agency for Development PHC Primary Health Care PHI Public Health Inspector PHM Public Health Midwife PHN Public Health Nurse SIDA Swedish International Development Agency SHS Superintendant of Health Services TFR Total Fertility Rate TMFR Total Marital Fertility Rate UNICEF United Nations Ch:lldren 's Fund UNFPA United Nations Fund for Population Activities USAID United States Agency for International Developmen t WDB Water Supply and Drainage Board WFS World Fertility Survey WHO World Health Organization  -i- Summary and Conclusions A. General Population and Health Situation 1. Sri Lanka has long been considered a mdel of social development among low-incoma countries. By key social performance indicator it rival s many countries at several times its income levels. Life expectancy at birth is 67 years for males and 71 years for females, the crude death rate is 6 per 1,000 population and infant mrtality is estimated at 32 per 1000 live births. Average caloric intake is 96% of estimated daily requirements. the adult literacy rate is 91% among males and 82% among females. 2. Between 1960 and 1982 the total fertility rate fell by a third from 5.4 to 3.4. The decline in the crude birth rate from 38 in 1941-50 to around 27 has challenged traditional economic assumptions that broad-scale e conomic development is a a prerequisite to such demographic change. Moreover, Sri Lanka has succeeded in developing a widespread and generally e fficacious health care system at relatively low cost to consumers. 3. However, the pace of improvement in Sri Lanka's population and health situation has begun to slow. The rate of natural population increase has remained virtually static for some years and demographic conditions now favor an undesirably higher rate of future population growth. Emerging inef ficiencies and inequities threaten the heal.th system and are likely to intensify as the result of persistent financial. c onstraints. B. Population Growth and Fertility 4. Sri Lanka's population increased from 12.7 million to 14.8 million over the 1971-81 intercensal period and is currently estimated a t 15.2 million. The crude birth rate has remained virtually level for a decade while the crude death rate has continued to fall and is now close to the projected low point of 5.5 per 1,000 population. The resultant high rate of natural increase (2.2% in 1980) has been mitigated by substantial emigration to produce an average annual growth rate of 1.7%, 1971-1981. However, emigration may well fall in· the future, if job opportunities abroad decline. Moreover, the proportion of women of child-bearing age is expected to continue to increase over the next 26 years, rising from 23.8% in 1971 to an estimated 27.2% in 2010, before beginning to decline. On the other hand, desired family size is relatively low, with a national mean o f 3.7 and a median of 2.9 chiidren per ever-married woman. 5. Several key factors which dominated the fertility decline over the past few decades are in a state of transition. There appears only limited potential to achieve further significant delays in marriage or  -ii- to raise levels of female literacy substantially. Female labor force participation is relatively high at about 17%. Moreover, the crude death and infant mortality rates are close to an expected plateau. Additionally, the total marital fertility rate has remained high, falling by only 5% - from 8.0 to 706 children - between 1963 and 1978. 6. Negative effects of population growth have appeared in the social sector. In education, student population per school has risen by 20%. In housing, population growth has almost eclipsed efforts to reduce shortages. In health, population growth has tmdermined efforts to expand service coverage. 7. Perhaps the most strongly felt consequences of population growth are in employment. Despite the expected creation of mre than 500,000 new jobs from major development projects, the employment situation is expected to remain grim for the rest of the century. According to USAID estimates, a total of 6.2 million people are 1:!xpected to enter the labor force between 1980 and 2005. 8. Alternative population projections show that the population would increase to more than 36 million by 2025 under constant fertilty. Under the mst optimistic scenario, with fertility declining to replacement level by 1990, the population wo�d rise to nearly 26 million by 2025. With m0derate fertility decline, population would grow to 27.3 million by that year. Under all three scenarios ·the working age population would increase to 13.5 million by the year _2000 and betw�en 17 and 22 million by 2025, an enormous employment challenge for the Government. C. Health Conditions 9. As a whole, Sri Lanka mortality rates and life expectancy are far better than elsewhere in South Asia and rival those of developed countries. Among SOIE sectors of the population, mortality patterns have some features typical of developed countries. Circulatory system diseases were the leading cause of death in 1979; malignancies were the seventh leading cause of deatho On the other hand, infections and respiratory diseases were still the second and third leading causes of death, respectively, in that year. 10. While lower than in mst developing countries, maternal mrtality was 7 per 10,000 registered live births, around five times the rate of many industrialized countries. While the precise level of infant mrtality is uncertain, it is exceptionally low for a country at as low an incoIE level as Sri Lanka. However, wide variations exist amng districts and the pattern of infant and child mortality continues to reflect the influence of poverty, an adverse physical environment and still relatively high fertility levels.  -iii- 11. Complications of pregnancy, respiratory diseases, infectious and parasitic diseases, and injuries and poisonings were the leading causes of hospitalization in 1981. However, the level of infectious diarrhoeal diseases has failed to decline significantly over the last 25 years despite substantial water, drainage and sanitation investments. Malaria morbidity and ID)rtality registered significant declines 1975-81, but there is concern about signs that possible resistance to the Malathion used in anti-malaria campaigns is developing. Malnutrition continues to be a major and possibly increasing problem, particularly for pregnant women and pre-school children. D. Policies and Programs 1. Population 12. The Government has long recognized the importance of population in economic and social development. It has tried to restrict immigration, repatriate.non-citizens and redistribute population from the densely­ occupied wet zone to the sparsely-inhabited dry zones. Since 1962, Sri Lanka al.so has had a policy commitment to deal with population problems. Since 1977, the Government has pledged itself to take all meaningful step s to curb popQlation growth, mainly through expanded service delivery activities emphasizing voluntary sterilization accompanied by acceptor incentive.s. The Government also has adopted tax policy measures explicitly to reduce desired family size and has continued to promote enhancement of the status of women, including education. 13. Family planning services are delivered through a combination of three channel's. The Ministry of Health (MH) is the main provider through the Family Health Bureau (FHB) but quasi and non-Governmental agencies, principally private voluntary agencies, also have important roles. Private practitioners have m:>re limited roles in family planning. 14. Government programs, delivered mainly by FHB workers, emphasize sterilizations and IUDs. Non-governmental. organizations provide non-clinical nethods, principally pills and condoms, and injectibles which are becoming increasingly popular. Private physicians trained iµ Western medicine provide sterilizations and other forms of ID)dern contraception on a fee-for-service basis but the large network of ayurvedic practitioners has yet to be involved actively in family planning activities. 15. Despite uneven performance, reflected in annual fluctuations in the numbers of new acceptors, the Contraceptive Prevalence Rate (CPR) has increased steadily from 32% in 1975 to 55% in 1982. Sterilizations account for around half the CPR increase between 1975 and 1982 but the main users are women who al.ready have ID)re children than they desire. IUD use fell during the period, consistent with worldwide trends. The use of pills increased slightly but remains at fairly low levels (2.6% in 1982); condom use also increased slightly. The use of withdrawal, rhythm and other traditional nethods remains unusually high in Sri Lanka, accounting for almost 45% of all contraception in 1982.  -iv- 16. The Government's family planning program faces a number of problems which will affect its future success. It is clear that the present program is unable to attract enough young low parity women to the use of modem contraception. There are still significant gaps in knowledge of where to obtain family planning supplies. The distribution network remains uneven. Some rural areas are under-served, the quantity of supplies is inadequate and the available mix is inappropriate for changing consumer preferences. Drop-out rates for pill users are high, possibly because of side effects from high-dosage pills. 17. Management and organizational problems also pose a major constraint to increasing the program's reach and effectiveness. Public health staff are too few in numbers particularly at the periphery, and lack motivation. Wide variation in acceptance rates among administrative divisions indicates the need to strengthen field supervision and n:onitoring and to adapt programs to Sri Lanka's ethnic and religious diversity. Finally, program efforts need to be better coordinated and managed at central levels through strong central leadership, since the family planning effort involves a number of government ministries, non-governmental organizations, private practitioners and external aid agencies. 2. Heal th 18. Sri Lanka committed itself in 1980 to the obje·ctive of "Health for All by the Year 2000" and began to plan a new system of primary heal th care (PHC). The Medium Term Investment Program (MTIP), 1981-85, stated that the new PHC model would include improved ru�al medical services delivered by midwives, multipurpose health workers and health aides. It would emphasise preventive measures, particularly through the family health program which includes family planning, and would require substantial increases in training and salary expenditures. However, the MTIP omitted quantitative objectives and specific targets for PHC implementation reflecting the larger absence of a comprehensive national plan for the health sector. 19. The MH is responsible for all government health and family planning services other than those managed by the Ministry of Women's Affairs and Teaching Hospitals. Intersectoral coordination and community participation take place through a national network headed by a National Health Council chaired by the Prime Minister with eight other Ministers as members. The council is supported by a National Health Development Committee, chaired by the Secretary of Health. Its membership consists of the secretaries of ministries on the Council, senior health officials, and representatives of selected international organizations. 20. Health and family planning services are organized and managed through 19 health divisions serving populations ranging in size from 220,000 to around 1.5 million. Divisions are further subdivided into 109 health areas. Urban and municipal councils are responsible for preventive  -v- health care in their areas. In the estates and plantations, the corporations provide some basic care through their own dispensaries and small hospitals. However, the standards vary widely because of shortages of medical manpower. Under the new PHC system, preventive and curative services are being integrated into a single administrative system, with around 38 formerly single-purpose instj_tutions such as maternity homes, augmented by 238 planned facilities, expected to take on multi-purpose functions as subdivisional. health centers. A network of around 4,000 Gra100daya Heal th Centers is proposed as the basis for community health care. 21. Vertical. programs for malaria control, immunization and food supplementat:ton al.so operate. Additionally, a program to manage diarrhoeal disease in children is being set up with UNICEF assistance. 22. A main problem of the health system is wide variation in the use of health .facilities of different types and among those of the same type. The general pattern, partly because of ,good transport, is for clients to bypass lower level primary and second·ary facilities in favor of mre advanced tertiary level ones. This situation reflects a general shortage of medical J>E!rsonnel, combined with expansion of tertiary institutions and relative shortages of financing for lower-level facilities. 23. A f1econd problem is both the absolute shortage of health personnel and their geographically-imbalanced deployment. The vacancy rate for Government doctors is 15%. Around 40% of key administrative posts in the divisions and Colombo are vacant. Districts with teaching hospitals have the largest concentrations of health manpower, while some districts are comparatlvely underserved with regard to doctors, nurses or both. The annual output of doctors is more than sufficient to meet the Ministry's needs, but around half of all physicians emigrate within ten years of graduation. 24. Other problems affecting government health services delivery include poor and underfunded maintenance and replacement of equipment and transport, weiaknesses in the analysis and use of data from the Health Information System and the need to improve the logistics and personnel position in regard to pharmaceuticals. Inappropriate and excessive use o f pharmaceuticals, particularly antibiotics, is common as well as costly and harmful because of the development of drug-resistant bacteria. 25. Government health expenditures 1978-82 averaged 5.7% of the recurrent budget against 7.6% in 1971/2-1977. However, overall budget growth kept t.he per capita real value of those expenditures relatively constant despite inflation and population growth. The cutbacks were largely at the expense of community health services while administrative and training expenditures increased for hospitals, mainly in Colombo. On the capital account, government health expenditures rose in real and per capita terms over the past decade. In constant prices, total expenditures  -vi- rose at a compound growth rate of 18% yearly during that period and amounted to about US$9.8 per capita in 1982. The ratio of capital to recurrent expenditures rose from 2 . 4% to as much as 19.7% 1971/2-1982. Most capital expenditures in recent years have been on tertiary facilities, influenced by the availability of foreign aid. However, these outlays have intensified the problem of financing the operating costs of increased capital stock with limited recurrent funds. 26. Important inefficiences have emerged in health care expenditure patterns. Relatively low occupancy rates of primary dispensaries and maternity homes have contributed to their relatively high unit costs against the next higher level of care. Additionally, both the cost of outpatient and inpatient care rises sharply as one mounts to higher levels of care, leading to issues of cost ef fectiveness of providing services at higher levels which could be performed at lower levels. Expenditures on pharmaceuticals, dressings and surgical/supplies reflect those same patterns. Outlays for pharmaceuticals in Colombo hospitals accounted for 18-23% of their total expenditures and one-fourth of the total expenditures of the State Medical Stores in 1982; evidence suggests that possible savings on pharmaceutical usage could be made at all tertiary facilities. 27. Shortages of recurrent funds have constrained expenditures on maintenance and vehicle operations, while differentials in the cost of training various categories of health personnel suggest inefficiences in this area as well. 28. Both the family planning program and the health investment program for PRC are heavily dependent on foreign aid. Around 70% of capital expenditures for family planning in 1980 were externally-financed. The seven major MTIP health projects including PRC/family planning and the new Kotte hospital will require capital costs estiamted at US$175 million. Excluding Kotte, the PRC project will account for 68% of identified MTIP capital costs. 29. Donors are unlikely to finance the incremental recurrent costs of t he PRC project, estimated at US$6.1 million yearly, a relatively low amount since most required staff are already employed. However, the public health sector faces other recurrent cost increases. Kotte hospital will require around US$4 million yearly; contraceptives to replace UNFPA donations will cost another US$1.6 million yearly. Additionally, there are pressures to raise worker salaries and a need to increase funding for travelling budgets and allowances, building and equipment maintenance and vehicle operations. 30. Recurrent funds available to the health ministries will depend on the growth of GDP, the proportion of GDP in the recurrent budget and the health share of the budget. Assuming a middle course between the optimistic government projection of 5.6% GDP growth and a worst-case scenario of 3%, current projections of the proportion of the recurrent budget ·met through GDP, and a constant health share of the recurrent  -vii- budget, the additional real resources available to the health ministries in 1990 would bE� US$10.3 million in 1983 prices. This is around 10% less than the amount nE�eded only for PHC, Kotte hospital and contraceptives, excluding all other incremental requirements. Under the worst-case growth scenario, the additional recurrent funds available would be less than the estimated recurrent costs of the PHC project alone. 31. It is clear that the Government needs to make urgent efforts to (a) improve the efficiency of health resource allocation, ensuring that patients are seen and treated at appropriate levels of the delivery system and by restricting capital investments to high-priority projects; (b) improve management and thereby tighten up operational efficiency, and (c) extend cost recovery for certain categories of patients. E. Issues and Recommendations 1. Population 32. Sri Lanka still faces the prospect of considerable population growth. The rate of natural increase is high and has remained nearly static for ten years. Emigration is m:>re likely to decline than increase. Prospects for future fertility reduction apart from increased contraception_ are dim. 33. Fan1ily planning performance is still well below its potential, as evidenced by considerable unmet demand, while high contraceptive prevalence is dominated by sterilizations, mainly of those who already have m:>re children than they desire. The use of modern spacing methods and participation by younger and low-parity women are low. Around half of all contraception. is by relatively inefficacious traditional methods. 34. A basic framework for more effective management of population growth is in place but the program suffers from several major constraints. Field workers need to carry out family planning activities more intensively and systemat:l.cally. Physical and manpower resources are inadequate to meet demand for sterilizations, while those for IUD insertions are underutilized. Pills account for an unusually low proportion of the contraceptive mix. Despite their relatively low efficacy, traditional methods remai.n popular with younger, low-parity women. Opera,tional coordination remains weak among the principal institutional actors on the population scene. Improvements need to be made in three key areas: program management and coordination, program strategy and service mix, and demand generation and incentives. 35. Program Management and Coordination. It would be desirable for one official organization, preferably the FHB, to have full authority for comprehensive national programming, from planning through service delivery and performance evaluation. Measures flowing from that decision should result in a family planning work program which would delineate specific  -viii- role for each agency in the population sector. The program should take into account explicit demographic goals, which in turn would be translated into actions to monitor and evaluate performance. These would include specific targets by method and geographical area, to increase both the overall CPR and the use of modern spacing methods. Family planning operations should be further decentralized to ensure that the local program mix reflects the characteristics and needs of particular communities. 36. Program Strategy and Service Mix. Alternative approaches are needed to persuade younger, low-parity women to space and limit births. One prospective strategy would be to motivate women to shift fr om traditional to tmdern methods and then to sterilization when they have reached appropriate family size. Access to modern reversible methods would need to �pand, along with better follow-up, particularly to manage side-effects. The Government's role in such a strategy would be to concentrate on those clinical methods which depend largely on the health system - sterilization, IUDs and injectibles. High priority should be given to increasing the role of the private sector, particularly through commercial marketing channels and ayurvedic physicians, in delivering non-clinical contraceptives such as pills and condoms. 3i. Demand Generation and Incentives. Raising the CPR is difficult because most of the strongest demand for family planning has been met . already. Although there is still substantial unmet need,generating mre demand will be critical. Receptivity to incentives for sterilization suggests that incentives per adoption of other methods may be warranted. Additional p opulation-oriented policy measures such as building family size criteria into the selection process for such social benefits as education and housing, effectively instituted elsewhere, also may be considered. 2. Health 38. The health sector confronts three major issues in moving toward a system responsive to Sri Lanka's health problems: (a) management and efficiency of the public health services, and ways to improve them despite existing resource constraints; (b) the design of the new PRC program, and its anticipated effectiveness in addressing underlying health problems and priorities. and (c) the question of cost recovery in view of the need to raise additional operating funds for the sector, to reduce excessive use of publicly subsidized services, and to improve the equity of health service use. 39. Management Efficiency of Existing Health Services. Poor inter-ministerial coordination and weak planning capacity have resulted in the absence of a comprehensive national. health plan and strategy. The absence of sufficiently coordinated planning, along with weak financial controls, has led to over-funding of tertiary levels of health care at the expense of. the periphery. There is an urgent need for planning activities to be broadened to encompass the entire range of health services.  -ix- 40. Managerial weaknesses affect headquarters, field and hospital operations. The distribution of resources and responsibilities between the two line ministries concerned with health may be expected to cause problems in allocating funds and staff, coordination of planning and program implementation and in provision of technical support to lower levels of the h ealth system. Functional problems also hamper the management of health services. For example, superintendents of health services have wide responsibilities for public health, patient care institutions and technical and administrative services, yet lack commensurate authority to take action in such areas as personnel and financi� matters. 41. Thei existing' pattern of health service utilization is badly skewed by the bypassing of peripheral facilities and the underutilization of secondary ones in favor of the tertiary levels of care. A main reason is the relat:f.ve deterioration of peripheral services because of overall resource constraints accompanied by considerable expansion of tertiary-leve1 institutions which have high operating costs. Planning for and control c,ver the use of recurrent financial resources and restrictions on further capital investments on buildings, are urgently required. 42. Theire is also the need to set up a health manpower planning and training unit to review both the present distribution of health manpower and training programs especially for doctors, nurses and field workers. 43. New Primary Health Care Program. To ensure that the PHC program r eflects a real shift toward preventive care, the Ministry of Health needs to review the proposed role of PHC workers, to prioritize their tasks and to·plan a continuous in-service training program. Moreover, the management skills of mid-level supervisors also need to be upgraded and measures need to be taken to encourage doctors to study community medicine and accept Ministry of Health jobs in under-served areas. 44. Flexibility is needed to adapt the PHC program design and operations to differing population densities, local conditions and existing infrastructure. Priority should be given to implementing the program first in those districts which have relatively poor health coverage, and particularly an above-averge infant mortality rate. Construction of facilities and staff quarters could be scaled down and simplified to improve the program's cost-effectiveness. 45. Cost Recovery. The Government should further explore possibilities for greater cost recovery, particularly to improve resource allocation in the health sector. There are strong grounds for continuing to subsidize or provide at no charge such services as routine maternal and child health care, health education activities a�d treatment for such diseases as malaria. In contrast, there are equally strong grounds for charging those who can pay for other forms of out-patient, curative care to curtail excessive use of services. Since inpatient services are m:>st costly to the Government, some degree of cost recovery may be necessary.  -x- 46. There al.so are two particular areas in which heal.th charges should be raised or introduced on the grounds of equity. First p Government charges for pay-beds in hospitals are below the actual cost of services and should be raised at least to that level. Second p the new Kotte hospital. will be more costly to operate than the Colombo General Hospital., which is the country's top referral. facility; every effort should be made to minimize its net patient subsidy through both expenditure control and cost recovery. To keep subsidies at the two institutions equal p Kotte would need to raise about US$1.6 million yearly p implying at 80% occupancy a daily charge of US$8 per inpatient •.  -1- I. INTRODUCTION 1.01 Sri Lanka, an island republic with an estimated 1982 population of around 15.2 million persons, has long been c<?nsidered a model of social development in the Third World. By key social performance indicators, it rivals many countries at several times its income levels with a life expectancy at birth of 67 years for males and 71 years for females, a crude death rate of 6.0 per 1,000 population, an infant mortality rate estimated at 32 per 1,000 live births, an aveirage caloric intake of 96% estimated daily requ:lrements, and an adult literacy rate of 91% among males and 82% among females. 1.02 The total fertility rate fell by a third from 5.4 to 3.4 between 1960 and 1982 alone. The decline in the crude ·birth rate from 38 in 1941-50 to around 27 in 1981 has challenged traditional economic assumptions that broad-scale economic development is a prerequisite to such demographic change. Additionally, Sri Lanka has succeeded in deploying a fairly widespread and generally efficacious health care system at relatively low cost to consumers. 1.03 However, the pace of improvement in Sri Lanka's population and health situation has begtm to slow down. The rate of natural population increase ru1s remained virtually static for some years, and demographic conditions now favour an undesirably higher rate of future population growth. Emerging inefficencies and inequities threate_n the health system and are likely to intensify as the reisult of persistent financial constrainu. Since Sri Lanka is nc,w embarking on a major primary health care initiative, it is timely to review what has been achieved, to analyze the key prc>blem areas in population and health, and to suggest managerially and financ:lally feasible courses of action in these sectors to improve living conditions for the island's population. II. POPULATION AND HEALTH SITUATION A. POPULATION Dynamics o:f Population Growth 2.01 Sri Lanka's population increased from 12.7 million to 14.8 million ovE!r the 1971-1981 intercensal period, representing an average annual growth rate of 1.7%. This contrasts with the much higher r:ates of population growth occurring over the last several decades. Further, provisional. estimates by the Registrar General in Sri Lanka indicate that annual growth slowed to 1.5% in 1980. The UDSt striking feature of population dynamics in Sri Lanka over the past decade has been the virtual standstill in the Crude Birth Rate (CBR) at around 28 per 1000 population (27.6/1000 in 1980). This contrasts sharply with the continuing decline in the Crude Death Rate (CDR) from 7.5 in 1970 to 6.1 per 1000 population in 1980. This level is now close to the projected minimum of 5.5 per 1000 population, after which the increasing proportion of older persons can be expected to result in a rise in the CDR.  -2- 2.02 The resultant high rate of natural increase, 2.2% in 1980, has been significantly tempered, however, by substantial emigration, presently averaging 6.5 per 1,000 per annum. Historically, the important role of migration in Sri Lanka is well documented. Large-scale immigration since the 1800s, attracted by employment opportunities on the plantations, contributed substantially to population growth and accounted for 60% of intercensal growth in the 1891-1901 period alone. However, high domestic unemployment, coupled with the lure of jobs in the Middle East and Europe, and repatriation of Indian plantation workers have reversed the trend, with emigration dominating since 1953.. According to Bank estimates, the net outflow of 530,000 persons over i;he 1971-1981 period reduced population growth due to natural increase by 24%. 2.03 The Total Fertility Rate (TFR) decreased almost 25% from 5.0 to 3.9 children between 1963 and 1978, and is estimated to have fallen to about 3.4 by 1981. This general decline in fertility has been important in constraining what otherwise might: have been substantial increases in the CBR, given the rising proportion of women "at risk" over this period. The proportion of women 15-49 years of age rose from 23.8% in 1971 to 25.5% in 1981, and is expected to further increase to 26.8% in 2000 and to 27.2% by 2010, declining only from the year 2010 onwards. However, minor fluctuations in the CBR in recent years have raised concern about the continuation of fertility decline and that the process or demographic transition may be interrupted unless greater attention is given to current key determinants of fertility·. First, in contrast to the 1953 to 1973 period in which the trend in age-·specif�c fertility rates was uniformly·, consistently and significantly downward, over the last decade the decline has been erratic. In fact, data indicate small increases at prime childbearing ages (20-29 years) 1.n the latter part of the 1970s. Second, the average annual decline in thEi TFR slowed from 2.2% over the 1963-1971 period to only 1.3% over the 1973-78 period. And third, there are significant interzonal variations: in fertility, with a difference of almost one child in the average number of children ever born between the lowest (Zone I, Colombo, at 3.37) and highest (Zone IV) at 4.33) rates. I / !J The 1975 World Fertility Survey (WFS) sampling design divided the country into six zones on the basis of regional socioeconomic characteristics. The capital, Colombo, with the adjacent urban areas was taken as Zone I. Zone II was made up of the administrative districts of Kalutara, Galle,. Hatara and Colombo (excluding Metropolitan Colombo). Zone III consisted of the administrative districts of Anuradhapura, Polonnaruwa, Monaragala, Hambantota, Chilaw/Puttalam, and a part o,f Kurunegala (North) and Amparai (Sinhala Amparai). The administrative districts of Trincomalee, Batticaloa and Muslim Amparai became Zone IV. The three districts of Jaffna, Mannar and Vavuniya, the traditional homeland of the Sri Lanka Tamils, became Zone V. Kandy, Matale, Nuwara-Eliya, Badulla, Kegalle, Ratnapura and South Kurunegala, where m:>re than 85% of of the Indian Tamils are resident in estates, became Zone VI.  -3- 2.04 Nevertheless, desired family size is relatively low, with a national mean of 3.7 and median of 2.9 children per ever-married woman, according to a 1979 Bank study of the determinants of fertility in Sri Lanka and Kerala. These figures are below not only those for sub-Saharan Africa and Latin America, but also those of several countries in Asia. Significant. interzonal variations do exist, however, with a range of 3.32 to 3.95 desired children; while at the same time a high percentage of women in Zone 1, Colombo, actually desire not more than two to three children. In analyzing the principal determinants of desired family size, parity emerges as the most important factor, although age at marriage, current family planning practice, level of .female educational attainment and ethnic group all exert a strong influence. 2.05 Despite the noted general decrease in fertility and low desired family size, excess fertility (as measured by the difference between desired faudly size and net, or surviving, children) continues to exist in Sri Lanka. As the 1979 Bank study revealed, 16% of surveyed women had mre (excess), 4-2% equal (par), and 42% less (deficit) children than desired. The average excess of 2.7 children per respondent having excess fertility in Sri Lanka was in fact higher than the other two areas surveyed-­ Karnataka and Kerala, India. 2.06 E�amining characteristics of ever-married women with excess children, important for targeting family planning program efforts, �ank bivariate. analysis of ·the WFS data revealed the largest proportion of women with excess fertility were in the intervals between the 4th and.5th (10.7%), 5th and 6th (10.4%) and 7th+ (9.0%) births. Relatedly, excess fertility tended to be influenced strongly by: (1) age, increasing successively among ever-married women from just 1.7% in the 20-24 age group, and 15.9% in the 30-34 year group, to 29.3% in women 45-49; (2) age at first ma.rriage (excess fertility dropped from 22% of ever-married women who married at age 15 or younger to just 7% among those who married at age 23 or older); and (3) level of educational achievement of women (with 20% excess fertility among women with no schooling as contrasted to just 9% among those with 10 or m::>re years of schooling). 2.07 Overall, the data reflect a high, continuing unmet need for contraception. Estimating the actual number of women in need of family planning is always controversial because the level can vary substantially by the criterion of "need" adopted. Recent estimates of unmet need, based on the findings of a Contraceptive Prevalence Survey (CPS) carried out by Westinghouse in 1982, indicate tqat the unmet need may be as high as 24% of currently married women of reproductive age (as measured by the proportion of women who responded that they dia not want another child but were not contracepting). Taking into account women wanting to postpone their next birth for at least one year, 1982 CPS data indicate that at least 27% of currently married couples remain at risk of unwanted or untimed pregnancies. Based on the 1981 census, this represents an unmet need of over one-half million women (526,160). Further, if users of traditional methods are excluded, the estimated number of currently-married women in need would rise to 51% or 1.1 million persons.  -4- Proximate Determinants of Fertility 2.08 Increased Age at Marriage. Age at marriage has an important impact on family size in Sri Lanka because, as survey information shows, delaying first births is not common. Examining the determinants of the decline in the CBR since the early 1960s, the 1979 Bank study reveals that rising age at marriage accounted for 55% of the decline in the TFR from about 5 to 4 over the 1963-1978 period. Between 1963 and 1978 the average female age at first marriage increased from 22.1 to 23.4 years, and continued to rise to 24.6 years in 1981. (For males the average has been fairly constant since the 1920s, with a mean in 1981 of 27.9 years). At the same time the tradition of child marriages has virtually disappeared, with only 0.1% of girls under age 15 ever-married in 1981. 2.09 Fertility Behavior/Contraceptive Usage. In contrast to increased age at marriage, family planning appears to have played a relatively minor role in the early stage of fertility decline in Sri Lanka. The initiation of the decline clearly preceded the introduction of government-sponsored family planning efforts in 1965. Nevertheless, the service coverage provided by the public program, complemented increasingly by activities of Non-Governmental Organizations (NGOs) and sales through the commercial sector, contributed· to the noted acceleration of the decline. The demographic effect of the family planning program itself was highlighted by the 1975 WFS data, which revealed that those zones which had a higher percentage of currently married �.,omen using family planning between 1960-64 and 1970-74 had.higher reductions in marital fertility. The direct correlation between the family planning program performance and fertility decline is further supported by Bank analysis indicating that in the 1970-73 period, when the program's three-year mving average acceptance of pills, IUDs and sterilization increased by 85%, TFR declined from 4.3 to 4.0. Over the 1975-78 period, when there was a 25% decrease in acceptors, the TFR remained stationary. 2.10 Abortion. Abortion has also reduced births. While the illegality of induced abortion impedes collection of precise statistics, the UNFPA 1980 Basic Needs Report noted that in 1977 in one hospital alone--the University Hospital in Kandy--550 women were admitted for incomplete abortions. Further, between 1975 and 1979, 154 abortion-related deaths were registered in the country. Actual levels are assumed to be much higher. A more recent study, carried out by a member of the Sri Lanka Commission for the International Year of the Child, indicated that the national number of illegal abort:Lons in 1981 was about 30,000. 2.11 Lactation. Breastfeeding has traditionally been an important contributor to fertility control in Sri Lanka, and it continues to be virtually universally practiced in the early months of infancy. However, decline in the duration of breas1�feeding, particularly aioong younger, urban, and working mothers, has raised concern over the increased risk of pregnancy in this group. As highlighted in a 1981 WHO-supported survey, the proportion of urban and rural women breastfeeding begins to diverge significantly in the fourth to f:Lfth m:mth postpartum; and by the 12th month of infancy, only 50% of urban women are still breastfeeding their babies, whereas 89% of rural women are still doing so. A joint Sri Lanka  -5- Women's Bureau and UNICEF study in 1982 further documented a significant fall in brEiastfeeding among working mothers in ,.both the public and private (excluding plantation) sectors as soon as maternity leave ended, generally at two months postpartum. Determinants of Fertility Decline 2.12 While increase in age at marriage and contraceptive practice were the principal neans by which fertility was reduced, their underlying socioeconomic determinants were mainly improvements in education and heal th. 2.13 Educational Achievements. The increase in female educational levels has been a major contributing factor to fertility decline. As the 1971 censue1 revealed, while those with no schooling had an average of 6 CEB, women with higher education had on average just 3.2. The 1975 WFS revealed dmilar differentials; con-trolling for age at marriage, women with more than 10 years of schooling had an average parity of 4.9, while the sample average was 6. The strength of the inverse, independent effect of education on family size decisions and, thus, fertility was further corroboratEid by the Bank survey which found a ·variation in the average number of CEB from 3.0 for those with 10 or more years of schooling to 4.1 for those with no education, standardizing for marital duration. However, fertility differentials between women with higher educational attainment and those with little education have narrowed considerably in recent years, as a result:· of relatively higher rates of fertility decline among the less educated. For example, the 1979 Bank study indicated a cumulative decline of 13% between 1967 and 1977 among those with a 6th grade or less, education compared to 7% among those with mre than 6 years of schooling. 2.14 llealth Improvements. Reductions in infant and child m:>rtality have also been a major driving force behind the fertility decline. National death rates have been relatively low since the early 1900s compared to many other low-income countries. However, the demographic transition began early in Sri Lanka with a definite downward trend in death rates noticeable since the 1920s. The introduction of an anti-malaria program in 1946 is viewed by many as the single m:>st important catalyst of a new era of much lower mortality commencing in the quinquennium 1946 to 1950. The greatest single annual fall in deaths occurred between 1946 and 1947. The 28% decline in the number of deaths and the 29% decline in the CDR in this one year resulted in a nine-year increase in life expectancy at birth (from 4 3 years in 1946 to 52 years in 1947). This gain took most Western countries almost half a century to achieve. The largest contributor to these reductions was improved child survival, with the Infant Mortality Rate (IMR) falling 72% from an average of 131 per 1000 live births in the 1941-45 period to an estimated 37.1 per 1000 live births in 1978. 2.15 While direct statistical evidence of the m:>rtality-fertility linkage in Sri Lanka is hard to come by, indirect evidence is provided by data from recent fe·rtility surveys. For example, female contraceptive usage can be analyzed according to child death experience and birth intervals. An ESCAP study shows that women who had experienced child loss  -6- tend to use contraception less frequently than women who had no child loss (20% compared to 35% among two-parity women). Similarly, women who had a child loss are less likely than women whose children survived to say they did not want another child; hencE� child mortality was associated with a greater desire for another child. Birth interval analysis measures the short-term effect of infant mortality on fertility. An infant death cuts short the period of breastfeeding, reduces the period of postpartum amenorrhea, shortens the interval between births and thereby increases fertility. World Bank analysis of the WFS data indicates that there was a 3.5 month reduction in the birth interval if the child died before reaching 12 months. 2.16 Female Labor Force Participation. Female employment has increased as the level of female education has risen. According to the 1981 census, the number of economically active females increased at an average annual rate of 2.3% (as <�ompared to 1.4% for males) since 1971, with about 17% of the nation's fomales currently in the labor force (as contrasted to 50% of males). Further, a recent Bank study indicated a correlation between female work participation and delayed marriage. 2.17 Ethnic and Religious Diversity. _ There are distinct fertility differentials in terms of ethnic and religious groups. The Sinhalese, historically the largest ethnic group, comprised 74% of the national population in 1981, followed by Sri Lankan Tamils (13%); �oors (7%); Indian Tamils (6%); and Malays, Burghers, Eurasians, Veddahs, Europeans and others (less than 1%) The majority of the population are Buddhist, though most Tamils are Hindus. The differentials in completed fertility by ethnic group are significant, with a Bank study indicating that the completed mean number of CEB ranged from 4.77 among Sinhalese Christians to 7.18 among the Moors. Even adjusting for other factors, a range of 1.5 children still existed in the mean numbers of CEB of distinct ethnic groups. 2.18 Place of Residence. Residential factors also affect fertility. As revealed in the 1975 WFS, the lowest average parity was among estate women (5.2 children) as compared to an average of 5.5 and 6.2 children among urban and rural women, respectively. However, controlling for age at marriage, the urban and rural differentials close substantially, reflecting the larger proportion of younger marriages in rural areas. The relatively lower parity on the estates, as compared to the rest of the country, may be related to higher levels of fetal wastage, under-reporting of births, and/or abortion. In a 1981 study, however, no evidence was found that the low rate was due to either greater contraceptive usage or longer duration of breastfeeding. Similarly, women living in the Dry Zone districts had higher average CBRs in the 1970s than those living in the Wet Zone. Prospects for Future Fertility Decline. 2.19 Despite the substantial. fertility decline achieved to date, there are several reasons why populatic,n growth will remain rapid unless immediate actions are taken. First and foremost is the sheer mmentum for growth in such a youthful population, with a median age of 21.4 years in 1981. As mentioned above, the proportion of women in reproductive ages will increase from 24.8% in 1980 to 27.2% in 2010.  -7- 2.20 Second, examining only aggregate declines in birth and fertility rates masks the still small reductions achieved to date in marital fertility--·a critical yardstick for achieving fertility reductions. In contrast to a 24% reduction in the TFR between 1963 and 1978, the Total Marital Fertility Rate (TMFR) remained high over the period, falling only 5% from 8.0 to 7.6 children. This contrasts with an average TMFR of about 5.7 and in both Kerala and Karnataka, respectively, over the 1975 to 1979 period. 2.21 Third, several of the key factors which dominated the fertility decline ove.r the past few decades are in a state of transition. Limited potential a.ppears to exist to achieve further significant delays in marriage, given the already substantial increases in age at first marriage and improvements over the last decade in the pool of eligible men (with the ratio of males to females in the prime marriage ages increasing from 73 to 97 per 100 females). The economic constraint to early marriage, based on high tmemployment and dif ficulty of financing a dowry, appears to be easing as well. The substantial inflow of family income from migrants abroad has raised concern that the age of marriage may fall. If this happens, general rates of marriage and fertility would increase. 2.22 The direction of the relationship between under-five mortality and fertility may also be changing. Historically, m ortality decline has preceded fertility decline. However, the CDR is now close to plateauing. There is also strong evidence to sugge�t that further.reductions .in infant mortality, still hig}:l in several di.stricts, may now be dependent on the attainment of additional reductions in fertility, as high fertility is itself a ma. j or associated cause of both infant and maternal morbidity and mortality. 2.23 1he decline in the duration of breastfeeding may also raise fertility. At the same time, the existing capacity of the public health system, though a m::>del in the region, remains seriously strained to meet existing demands without the added challenge of the need to expand family planning services. As employment opportunitities abroad wane, and government attention increasingly focuses on the costly "brain drain", the currently s:ignificant net emigration is also expected to taper of f. 2.24 In summary, Sri Lanka is now in a dif ficult stage in its demographic transition. A demographic structure favorable to a decline in the population growth rate will emerge only in the next century. Until then, reduc:tion in population growth has to come largely through increased contracep�i.ve practice and reduced marital fertility. Development.al Consequences. 2.25 Sri Lanka was one of the first Asian countries to formally recognize t.he adverse consequences of rapid population growth. However, it has not been able to avoid some of them. According to USAID estimates, the current population density of 220 persons per square km (with 57% of the population concentrated in 1981 in the Wet Zone which comprises just 23% of  -8- the land area) has reduced the average size of landholdings and may have contributed to a worsening of the nutritional status of the rural population. 2.26 Social sector investments--long the pillar of Sria Lankan development-have been consistently and significantly eroded by rapid population growth, as USAID estimates highlight. In the health sector, efforts to expand service coverage have been undermined by population growth which has increased the number of consultations demanded and the ratio of population to hospital beds. At the saioo time, the large-scale emigration of health professionals has further raised the ratio of population to physicians. Similarly efforts to increase the level of primary school attendance amidst growth in this age group have resulted in a 20% increase in student population per school, seriously straining the existing system's capacity to provide quality education. In the housing sector, despite considerable investments to reduce housing shortages among the poor, government programs were virtually eclipsed by population growth. The total net gain of accommodation for only 145,000 persons contrasted sharply to the 1.4 million persons initially expected to be housed. 2.27 The negative effect of rapid growth has probably been most strongly felt in the area of employment. Using the 1981 census definition, between 1971 and 1981 unemployment in Sri Lanka rose from 16% to 18% of the economically active population; and in some districts, it accounts for _mre than one quarter of the labor force. Despite soIOO alleviation expected from the two major development projects--the Accelerated Mahaweli Development Scheme (an expected 450,000 jobs) and the Industrial Promotion or Free Trade Zone (with an expected 60,000 new jobs over the next decade)--the overall employment situation is expected to remain grim over the balance of the century. According to USAID estimates a total of 6.2 million people (or 249,000 annually) are expected to enter the labor force between 1980 and 2005. A third major development project--the Urban Renewal and Housing Program--has also been motivated by population pressure, which contributed to two thirds of Greater Colombo's population living in substandard housing. 2.28 Alternative population projections shown in Table 1 illustrate the effect of further reductions in fertility on the rate of population growth. Under the most optimistic scenario, with fertility reaching replacement level by the year 1990, the population would be over 20 million by the end of this century, and nearly 26 million by the year 2025. However, the population would be over 36 million by the year 2025 if fertility remains constant. Under all three scenarios the working age population increases to 13.5 million by the year 2000, and between 17 to 22 million by the year 2025, an enormous challenge for the Government to expand employment opportunities. Women in reproductive ages are expected to increase from 3.7 million to 5.6 million by 2000; but they will increase to 6.7 million by 2025 according to the scenario of mderate fertility decline, against 8.5 million under the assumption of constant fertility. The age group 0-14 years is most affected in the short run by fertility reduction. Under the assumption of rapid fertility decline, the size of this group would change very little between 1980 and 2025. However,  -9- Table 1: SUMMARY OF POPULATION PROJECTIONS UNDER THREE DIF FERENT ASSUMPTIONS OF FERTILITY DECLINE, 1980-·2025, SELECTED YEARS (OOOs) 1980 1985 1990 1995 2000 2025 Population Rapid decline 14738 16270 17683 18860 20091 25558 Moderate decline 14738 16270 17933 19560 21158 27313 Constant fertility 14738 16270 18134 20253 22512 36325 Dependenci Ratio (Per 100) Rapid decline 69.7 62.9 59.1 53.0 48.5 47.5 Moderate decline 69.7 62.9 61.4 58.6 55.6 48.2 Constant fertility 69.7 62.9 63.2 64.3 66.3 64.5 Population 0-14 Rapid decline 5442 5560 5695 5515 5299 5428 Moderate decline 5442 5560 5946 6216 6266 6082 Constant fertility 5442 5560 6146 6909 7720 11446 Population 15-64 Rapid decl ine 8683 9988 11113 12330 13533 17328 Moderate decline 8683 9988 11113 12330 13533 18429 Constant fertility 8683 9988 11113 12330 13533 22078 Women· 15-49 Rapid decline 3652 4186 4670 5162 5648 5400 Moderate decline 3652 4186 4670 5162 5648 6740 Constant fertility 3652 4186 4670 5162 5648 8532 Source: World Bank Projections. Note: The three variants have differing assumptions only about the pace of fertility decline; all three share the same m:>rtality and migration assumptionei. In the first variant, replacement level fertility (NRR=l) is reached in the year 1990; in the second, which is the World Development Report projection for Sri Lanka, NRR=l in 2005; and in the most pessimistic scenario, fertility remains constant at the current level. Life expectancy is expected. to improve to 74 years by 2005, and it is assumed that net migration �dll decline gradually to zero by the end of the century.  -10- B. HEALTH General Mortality 2.29 Mortal! ty rates and li:fe expectancy in Sri Lanka as a whole are far better than in other countriE�s of South Asia, and indeed rival those of developed countries. From a levitl of 19.8 in 1946, the CDR dropped to only 6.1 in 1979. The latter may be compared with corresponding figures of 13 in India, 16 in Pakistan and 18 in Bangladesh. Similarly life expectancy in Sri Lanka was then estimated at 66 years for males and 70 years for females--not far short of the average life expectancy of 75 years in industrialized market economies. 2.30 The pattern of m::>rtality in Sri Lanka has some features typical of developed countries. For example, the leading cause of death in 1979 was circulatory system diseases (see Table 2), which accounted.for 14% of all deaths (and 25% of all deaths in the 40-69 age group). Malignancies were the seventh leading cause of death. On the other hand, infectious diseases were the third leading cause of death, responsible for 9.0% of the total. Respiratory system diseases accounted for another 8.3% of the total, and infant perinatal diseases for 6. 9%. Maternal deaths occurred at a rate of seven per 10,000 registered live births, about five times the rate of many developed countries but lower than most developing countries. Around 4% of deaths among women in the childbearing age group were directly due to complications of pregnancy. Of these deaths, 40% were due to postpartum hemorrhage and sepsis, both considered to be preventable •. Given the age structure of the populat:Lon, with 50% below 22 years of age and 35% under 15 years of age, analysis C)f age specific m::>rtality rates and patterns, especially for infants and children, is important in order to identify priority needs.  -11- Table 2: GENERAL MORTALITY!/ BY BROAD DISEASE CATEGORY OF CAUSE OF DEATH, RANK ORDER2/ AND YEAR - ALL AGE GROUPS 1971 - 1979 Mean 1971 - 1979 1979 Cause of Death Rank Rate Rank Rate Diseases of the Circulatory System l 91.9 1 90.1 Infectious Diseases 2 81.1 3 59.5 Diseases of the Respiratory System 3 66.6 4 54.7 External Causes of Inj.ury 4 65.2 2 70.5 Diseases .of the Nervous System & Sense Organs 5 53.2 5 46.l Infant Perinatal Disorders 6 51.5 6 45.7 Malignancies 7 31.3 7 29.6 Hematologic Disorders 8 21.1 9 11.1 Avitaminoses & Other Nutritional Deficiencies 9 18.1 11 7.3 Diseases of the Castrointestinal Tract 10 18.1 8 14.2 Endocrine Disorders 11 10.3 10 7.6 Diseases of the Musculoskeletal System 12 8.8 13 5.8 Diseases of the Genito-urinary Tract 13 7.3 12 6.4 Diseases of the Skin 14 6. 1 14 3.6 Congenital .Anomalies 15 4.0 15 3.2 Maternal Perinatal Disorders 16 3.0 16 2.2 Psychiatric Disorders 17 2.1 17 2.1 Senility & Other lll defined causes (1) 233.3 (1) 199.4 Total 772. 7 765.4 Notes: 1/ Per 100,000 population 2/ Rank order determined by actual frequencies. Source: Department of Census & Statistics Infant and Child Mortality 2.31 There is SODE uncertainty about the precise level of the IMR in Sri Lanka. The 1984 World Development Report records 32 for 1982, and it is clear that the level is exceptionally low for a country as poor as Sri Lanka. In comparison, India has an estimated IMR. of 94, Pakistan 121 and Bangladesh 133. But the average national data mask the extreDE conditions of poverty still pervasive in many parts of Sri Lanka--for example, in the estates, the urban slums and shanties, and the new development areas, infant roortality rates are in the range 55-79 per 1000 live births. Even at the district level. one can see a wide variation in infant mortalitv  -12- rates--in 1979 seven districts (Hambantota, Mannar, Vavuniya, Amparai, Puttalam, Anuradhapura and Monaragala) had a rate in the 20 's � while four districts (Nuwara Eliya, Kandy, Badulla and Ratnapura) had rates of 55 or more. 2.32 The pattern of infant and childhood mortality largely reflects the influence of poverty, an adverse physical environment, and still relatively high fertility levels. In infancy and early childhood (1-4 year olds), infectious diseases were responsible for 12.6% and 24.3% respectively of all deaths in 1979. Nutritional deficiencies were registered as the cause of death in 7.4% of children 1-4 years of age; and in 1978 immaturity was listed as the leading cause of infant mortality, responsible for 13 deaths in every 1,000 live births, a cause rost probably related to maternal malnutrition. In 1981 neonatal tetanus was reported to, occur in one out of every 2,000 live births. Also of significance, in 1979 the ratio of neonatal deaths to post-neonatal deaths in Sri Lanka was 1.77 (again with a substantial variation between districts). This is in contrast to a typical ratio of about 3 for a developed country, where post-neonatal deaths are significantly-reduced ·through improved environmental factors, control of infectious diseases, and good nutrition. While the reported deaths in 1979 due to measles among children less than 5 years old represented only 0.5% c,f all deaths in this age group, the known synergism between measles and malnutrition suggests that measles may be a more serious problem than present�y reported. Morbidity 2.33 In the absence of comprehensive community-based surveys, hospital-based estimates are the major s�urce for data on UDrbidity. Major causes of hospitalization in 1981 were complications of pregnancy (23.3%), respiratory diseases (14.7%), infectious and parasitic diseases (13.2%), and injuries and poisonings (11.2%). With an expanding immunization program, there has been a fall iTl the incidence of diphtheria, pertussis and tetanus. However, the continuing presence of neonatal tetanus, when most babies are delivered in institutions, points to a breakdown in infant care. 2.34 The ptesence of infectj�ous diarrheal diseases has shown no significant decline over the last 25 years, despite substantial investments by the National Water Supply and Drainage Board. It is estimated that two million patients are treated annually as outpatients for gastrointestional diseases; 193,000 are admitted to hospitals (representing 8% of all admissions, and 68% of admissions for infectious diseases); and 1,800 die there each year from these diseases. Of other fecally transmitted diseases, infectious hepatitis continues to be a significant problem, while up to 85% of the rural population suffers from intestinal parasites. 2.35 Malaria outbreaks have occurred regularly, though in recent years the anti-malaria campaign has had some marked successes. For example, the reported incidence of malaria was reduced from 28/1000 population in 1975 to 3.2/1000 in 1981. In additiort, mortality due to malaria declined significantly and the proportion of falciparum fell from 16% in 1975 to 3% in 1981. There is concern, however, about possible signs that the vector  -13- is develop:f.ng resistance to malathion, and also about the potential for an epidemic ir.L the new Mahaweli development areas • .- Indeed, the number of reported cases of malaria rose from 38,566 in 1982 to 123,996 in 1983. 2.36 Malnutrition continues to be a major and possibly increasing problem. Comprehensive studies of child growth were undertaken in 1975/76 and again in 1979/82. Although the methodologies of the studies differed in many ways, a comparison of the results is useful to assess trends over time. Among children aged 6 to 59 months� the prevalence of stunting (from chronic malnutrition) declined fron1 35% to 23%. However, the prevalence of wasting (frOJil acute malnutrition) worsened, increasing from 7% to 9%. Concurrent wasting and stunting declined from 3% to 2%. The prevalence of stunting gemerally increases with age and is particularly high among estate children. In the second study, however, both the estate sector and all children ir.t their sixth year were excluded, and so the apparent improvement in chronic malnutrition implied by the 1979/82 data may not be significant. Malnutrition among pregnant women is al so common, and as a result _20% of the babies born at term are significantly underweight. Anemia is due mainly to iron deficieney, but it is aggravated by hookworm, protein deficiency, and to a lesser extent folic acid deficiency. Hemoglobin concentrations of less than 50% of normal are found in 5% of pregnant women in the estates where it is a significant direct cause of death. �i.ter caused by iodine deficiency occurs in up to half the population in the southwest coastal belt and the central regions. Vitamin A deficiency, on the other hand, now accounts for less than 2% of all cases of blindness, probably re'flecting the known high intake of fruits and vegetables. 2.37 Although diseases of poverty, such as diarrheas, respiratory diseases, a.nd malnutrition, continue to predominate, some new health challenges have arisen with the changing age structure and development taking place. These include cardio-vascular diseases, accidents, cancer and geriatric problems. Vulnerable Groups 2.38 Despite relatively high life expectancy and low infant lll)rtality there is still considerable preventable mortality and especially morbidity. Certain areas of the certain country and population groups are especially badly off. Health conditions are particularly bad in the estates and also in the urban slums and shanties. In view of its relative lack of social infrastructure, the Mahaweli area also deserves special concern. In terms of population groups, infants and children under 4 years of age face the highest risk of mortality (apart from adults aged over 45 years). In addition, ��men of fertile age are uniquely at risk of death from complications associated with pregnancy, labor and the puerperium. Pregnancies occurring in women under 17 years of age or over 35 years of age carry a higher risk of death for both the mother and her infant. Moreover, around 50% of ever married women of reproductive age do not want any more children while only one quarter are practicing a modern form of contraception, leading to unwanted pregancies. Therefore, particular attention needs to be given to improving the effectiveness of both family planning and maternal and child heal th services.  -14- Ill. POLICIES Ao POPULATION 3.01 The Government has long recognized the role of population in economic and social development. Three major interrelated issues have dominated the Government's attention to population over the past five decades--!) migration; 2) internal population distribution; and 3) m:>re recently, population growth. Migration 3.02 As early as the 1940s, the Government recognized the need to control the large number of permanent migrants, most of whom, following the path started in the 1830s, came to work on the plantation estates. Defining the status of the large number of Indian immigrants was viewed as an important first step. Intergovernmental ministerial discussions in 1964 led to agreement to repatriate about 525,000 persons and confer Sri Lankan citizenship to the remaining 300,000 Indian immigrants over the ensuing 15 years. The pace of implementation has been extremely slow, however, and among those who have stayed lack of formal political status has eontinued to have a detrimental effect on their access to key government programs, including heal,th and family planning. 3.03 In recent years, however, attention has dramatically shifted to the problem of massive out-migration of labor, first highly skilled but successively semi and unskilled. This "brain drain" over the 1971-1976 period alone accounted for approximately 15% of the professional and technical personnel in the country in 1971, and it clearly helped alleviate demographic and employment pressures. But it has been at a high cost when viewed from the perspective of the considerable national resources invested in their training, and the rapidly increasing need for skilled labor to meet national development objectives. Redistribution 3.04 Redistribution of population from the densely occupied wet zones to the much larger but sparsely inhabited dry zones has been a major thrust of national development programs since the 1950s. With the increasing control of malaria and communicable diseases, the Government has systematically moved to redistribute population into previously uninhabitable areas, investing substantial resources in both irrigation and infrastructural development to support massive internal resettlement schemes. The Mahaweli alone is expected to resettle some 145,000 families or about 5-7 percent of the national population. In Mahaweli, as elsewhere in Sri Lanka's dry zones, however, such internal migration, which tends to be dominated by younger age groups, is resulting in rapid population growth• Fertility increases in these resettlement areas are exace the still highly limited access to key social services, including rbate d by family planning. Improved demographic impact analysis and incorporation of population planning into such key development programs are therefore essential.  -15- Population Growth 3.05 While government concern over the problem of rapid population growth has long been enunciated and the first formal policy statement was made in 1965, the issue of population growth has taken on a particularly urgent tone over the last decade. Recognition of, and commitment to redress, the population problem were dominant features of the plat form of the present: Government which came into power in 1977. As stated in the M anifesto of the United National Party and later incorporated into the Presidential Throne Speech and Statement of Objectives, the Government formally pledged itself to take all meaning ful steps to curb unplanned growth of population. To this end, enhanced family planning services would be provided. by the state and financial incentives given to individuals to practice family planning. It was also decided that primary emphasis would be placed on service-oriented programs to enable mtivated couples and individuals to receive family planning services and to undergo vasectomy or sterilization voluntarily. 3.06 In 1982 these policies were again confirmed when the President stated the following at the Third Asian and Pacific Population Conference: "Our policy is to· provide enhanced family planning services to those who voluntarily seek such services. There is absolutely no compulsion or coercion, and the Government is committed to the provision of a variety of in formation, education and communication services as well as contraceptive and clinical services to meet the needs of those who voluntarily seek family planning." 3.07 These policy statements have been backed up by a number of practical measures. For example, the establishment in Septe�ber 1977 of a Population Division within the Mini.stry of Plan Implementation (MPI) formally res ponsible to the President for the formulation and implementation of population policy has been widely heralded as an demonstration of the broad political support accorded to population limitation. The Government has also undertaken several specific measures to improve service delivery: these include the consolidation in 1978 of family planning activities within the Family Health Bureau (FHB) under what was then a separate Project Ministry of Colombo H ospitals and Family Health, and the extension of authority to train non-physicians to dispense contraceptives and insert Intra-Uterine Devices (IUDs). 3.08 Of particular note have been the policy measures instituted by the Government explicitly to lower desired family size. For example, in 1979 the Government introduced substantial financial incentives to both providers and recipients of sterilizations. Also important was the "beyond family planning" incentive adopted in 1979: a revised, flat rate allowance of Rs.12,000 was given to each resident taxpayer (with average wages about Rs.500-600 per mnth), regardless of family size, as part of a broader planned tax policy reform geared to discouraging large families. However, only about 14% of the population currently pay tax. Government supported programs to raise health and educational levels and policies to enhance the status of women also have unquestionably contributed significantly to changes in desired family size and, relatedly, acceptance of family planning.  -16- 3.09 Nevertheless, there is need for a mre strongly oriented and streamlined population policy to ensure that all development endeavors support population objectives. Much room for improvement still exists for integrating explicit population considerations into the public investment program. Furthermore, more use needs to be made of demographic and contraceptive prevalence targets,. In 1981 the MPI assumed a demographic goal of a CBR of 26 per 1000 by 1984 in order to calculate the contraceptive requirements of th4:! Government; and according to the UNFPA Basic Needs Report of 1980, the FHB had set an annual target of reaching 1% of the population in each Superintendent of Health Services (SHS) division. ·However, detailed targets are not actively used for purposes of either planning or monitoring performance, even though they are an essential component of most well-managed programs. This reluctance to set and use targets has largely been due to the Government's concern about delicate ethnic issues. Nevertheless, the virtual lack of strong opposition by local interest and other political groups to population issues in general and specifically family planning would permit it to do so, as in other countries in Asia. B. HEALTH 3.10 The Government pledged itself in 1977 to "restore the high standard of health care and disease prevention that existed earlier, and make fur-ther improvements in our health services particularly in the rural areas through both ayurvedic and western systems." In the following year there was a joint Government-wHO health programming exercise; and in 1980 the Government formally· committed itself to the objective of "Health For All By The. Year 2000," and began to plan a new system of primary health care (PHC) • Under the new PHC model, the m::,st peripheral facility will be the gramodaya (village) health center, staffed by a Family Health Worker (FHW) and serving a population of 3,000. The existing cadre of Public Health Midwives (PHMs) are being retained in this role. FHWs will be expected to provide an expanded range of as yet unprioritized services, including: antenatal care, emerg1ency natal care, postnatal care, child care, family planning, immunizations, prevention and care of diarrheal diseases, nutrition, school health, family health and adolescent care, control of communicable diseases, management of minor ailments, health education, urine testing, blood filming for malaria, and screening of patients. FHWs will only be able to dispense chloroquine and contraceptives. An average of about 6 gramodaya health centers will be supervised by each subdivisional health center. The latter will also provide out-patient care, and be responsible for environmental sanitation and control of communicable diseases. A divisional health center will be established in each assistant government agent division, and it will undertake integrated preventive and curative services, including inpatient care dental care, minor surgery .including sterilizations, and medico-legal work. A new organizational stru,cture is also planned to increase intersectoral coordination and community participation. This proposed new PHC model is a significant departure from the previous structure in that preventive and curative services will no longer be provided separately.  -17- 3.11 The Medium-Term Investment Programme for the period 1981-1985 refers spedfically to the PRC program. It states that "the heal th strategy.... will, therefore, place greater emphasis on. aspects of primary health care! including the improvement of medical services in rural areas by providing a well organized field based domiciliary service using midwives as multipurpose health workers and the services of health aides. Increasing emphasis on preventive health, particularly under the family health programme, will require substantial increases in recurrent expenditures, particularly on staff salaries and training." However, the Medium Tert11 Investment Programme does not include quantitative objectives for the de11elopment of the health sector, nor targets for the implementation of the PHC model. Detailed planning has so far only been done for the needs of the lower level facilities--i.e. divisional health centers and below. The detailed requirements of the higher level hospital facilities and services have yet to be considered. There has also been little attempt to allocate recurrent budget resources in line with stated priorities.. This lack of a comprehensive national plan for the health sector in j�ts entirely could lead to haphazard development and inappropriate resource al.locations. IV. PROGRAM PERFORMANCE A. POPULATION Program Structure and Management 4.01 Although private family planning efforts sponsored by women's associations began in the 1930s and the first family planning program was formally launched in Sri Lanka in the 1950s with Swedish assistance, services were not generally available to the population at large until the late 1960s. Overall, family planning services are provided through three principal channels in Sri Lanka--the government; quasi- and nongovernmental agencies, primarily private voluntary agencies; and private practitioners. 4.02 The Public Sector. The official incorporation of family planning into national. maternal and child health (MCH) programs in 1965 extended a much needed: legitimacy to family planning practice. Since then, the Government has assumed the role of principal provider. Family planning is available j_n each of the 15 administrative SHS divisions of the public health system, via both fixed facilities and through field operations, under the c,verall guidance of the FHB and the local direction of the each district's medical officer-MCH. 4.03 l'he FHB has been assigned primary responsibility for m:>nitoring and evaluating the family planning program. Data are obtained through three main mechanisms: 1) required standardized reporting of new acceptors by all goveirnmental and NGO providers; 2) periodic, 3-4 year interval, stratified sample program impact surveys; and 3) follow-up studies to ascertain continuation rates. A system exists for regular feedback o1: information to planners and service providers, but in practice it does not work well.  -18- 4.04 Given the strong curat:lve orientation of the health system, the Government's family planning program has largely been geared towards the provision of principally two methods, sterilizations and IUDs, which are both facility based. By 1980, 1:here were 86 sterilization centers and 635 IUD clinics dispersed across pro11incial, base and district hospitals. In addition, several mobile polyclinics operate in areas where there are few health facilities. Yet availability of services varies considerably between and within districts. 4.05 Major responsibility for family planning at the community level continues to be delegated to the FHWs (previously PHMs) under the supervision of the public health nurses (PHNs). While the widespread adoption of institutional deliveries has already begun to displace the PHM in her traditional role, she has continued to play an important educational role for· family planning: a 1977 survey indicated that 31% of all new family planning acceptors traced their motivation to the PHM. District level heal th workers receive commissions from subsidized sales of contraceptives in more than 1,200 "sales points" established in NGO outlets organized through municipal! ties., corporations and boards. Nevertheless, a 1978 WHO report indicated that dE!tection and registration of pregnancy, and also pre-natal and post-natal care then continued to be given priority over family planning. In fact, the report found that family planning education and follow-up tended to be concentrated only in the postpartum period. A main reason for this lack of attEmtion by FHWs to the provision of family planning services to the broader eligible population has been their weak supervision by PHNs. 4.06 The revised role of thE! FHWs under the new PHC model is still not clear. While the extension of. their functions to that of a mu1 tipurpose health worker can be expected to raise their credibility and status at the community level, it may also further dissipate their energies over such a broad range of activities that family planning could easily be given low priority unless ef fective supervision combined with systematic nnnitoring of established targets. 4.07 In a move to decentralilze operations, local district population committees have been created in 1!ach of the 24 districts in order to generate demand for family planning and improve local family planning programs. Yet the failure to allocate discretionary funds has se verely hampered their ability to institute innovative approaches. As many government agents also appear to lack strong belief in family planning activities, there is a need for programs to raise their awareness and commitment to these activities. 4.08 Non-Governmental Organizations. The growth of the family planning work by NGOs and privatE! providers has been impressive. Of particular note is the Family Pl.mning Association of Sri Lanka (FPASL), established in 1953 as an affiliate of the International Planned Parenthood Federation (IPPF). This organization remained the principal provider of family planning services until 1965, when the government program was launched. Its ef forts in mass media and interpersonal communication are complemented with direct service provision through a network of facilities including a family planning and a sub-fertility clinic in Colombo, se ven  -19- other. clin:l.cs, and a minilaparotomy and DX>bile ·vasectomy service. It also manages the highly successful contraceptive so�{al marketing program, launched in 1973. The latter provided a reported 95,727 couples (7% of them new acceptors) in 1979 with subsidized sales of condoms, oral contracept:1.ves and mre recently spermicides via 4, 611 private outlets including restaurants, grocery and drug stores, and texture and wholesale outlets. JCt also distributes orals to interested private doctors and pharmacists for prescribed use by their clients. Overall, the FPASL has reported an average cost of just $2.90 per couple protected. Since launching the social marketing scheme, it has been involved in a project to train and E1upply indigenous practicitioners. Currently, ef forts are targeted specifically at select groups under-and unserved by the national program, mc,st notably rural youth i.ncluding school drop-outs and the industrial and plantation sectors. Other NGOs actively involved in family planning include Community Development Services, the Sri Lankan Association for Voluntary Sterilization, the Lanka Mahila Samit!, the YWCA and YMCA, the Lions ciub, the Rotary Club and the Jaycees. Despite the activities of all these NGOs, however, there are still major gaps in the provision of family planning services. 4.09 Private Practitioners. The involvement of private physicians in family planning has been limited almost exclusively to provision of sterilizat:1.on and other contraceptive products on demand on a fee-for-service basis. The- 16,000 (including 11,000 formally trained and certified) ayurvedic practitioners_ have not been actively involved in family planning to date, though they represent the largest single network of medical services nationwide. In fact it is estimated that the vast majority of' population live within 15 minutes' walk of a traditional practitione!r and within 6 miles of an ayurvedic dispensary or hospital. Program Ind.icators 4.10 �,cceptor Rates, Methods, and Characteristics. The numerous service poi.nts belie the ·actual availability and use of services. Family planning pr-ogram performance in Sri Lanka, as measured by total acceptors, has been hi. ghly uneven. After a relatively slow and irregular start in the late 1960s, the number of acceptors of modern methods rose rapidly. Between 1968 and 1973 inclusive, acceptors in the Government's program increased at an annual compound rate of 14. 3%. The advent of orals and condoms marketed privately through the IPPF/community-based distribution (CBD) progr-am boosted acceptance rates; from 1974 through 1977, acceptors rose at a c.ompound annual rate of 37.6%. However, the growth of clinical methods - the mainstay of the Government's program - declined during those same years. New sterilizations fell by 55% and IUD insertions dropped by 23% between. 1974 and 1977. During the next three years the Government's sterilization program picked up substantially and virtually tripled to over 101,000 acc.eptors between 1979 and 1980 alone. IUD insertions registered mixed but generally declining performance during that same period. FWB statistics document continuing fluctuations in overall program performance, with the number of acceptors per 1000 married women in the reproductive age groups fluc.tuating from 49.1 in 1979, to 75.8 in 1980, and to 55.6 in 1981 .  -20- 4.11 Despite the annual fluctuations in the numbers of new family planning acceptors� there has been a steady increase in the Contraceptive Prevalence Rate (CPR) over the last ten years reaching 55% y b 1982 (see Table 3). Yet the national averages mask significant inter zonal variations in CPRs a1ll0ng zones of the country. As the 1975 WFS_ indicated the , proportion of currently-married ��men practicing family planning varied from just 19% in Zone IV (including the eastern costal. belt consisting of Trincomalee, Batticaloa and part of Amparai) to a high of 51% in Zone I {Metropolitan Colo�b,o). The wide interzonal differences have continued according to the 1979 Bank and 1982 CPS data. Table 3: mmw:::EP PMV� RA1ES BY f'filliO 1975-1982 � OF aJRROO'L OO!RI.EO W0£N US00 �Pl'I{N) 1-EtbJd I ' WFS Q.975) % l lrkl.d.d Bame (1979) %' FllB (1982) % I Westingrous (1982) % ) i .J ',run IPlll I I 1.5 4.7 I I 2.5 I I 2.5 t I 2.6 I I I I 3.2 I 3.1 I 2.5 I l I ( l I l ' ' 'I /OJ:rklom I 2.3 I 1.0 I 2.8 I 3.2 [ l I I ( Ste.rilization I 9.9 I 14.. 3, I 23.9 1 20.7 , (9.2) { {Female) I I (20.0) I (17�0) I I l I (Male) (0.7) ( r (3.9) t (3.7) I 1 fnjections a'kf otrer mnele I f I 1 ) I 'I (Sub Total �em) I scientific n:ethxls t r ---- 0.4 .I t f I ' f _!.4 _ I I _l_.!!__ I I I l (18.8) l (.21..0) I (33.7) f (30.. 4) l r , l i i I ) j I l I ' I With:frawal. I 1.5 n.a. I 3.6 t J I 4.7 t l ( ( � 't' ---- jRhythm I l 8.0 r J n.a. n.a.. t 10.3 r 13.0 t tOOer 'rradit:ionS ) } l 3.7 l i _Q•3._ I _E_.8 - r 'Sub 'Ibtal Trailtiona1) � t ! I 'r t l I (13.2) Q.3 ..3) (14.l) l (24.5) I t '- ' I ) I I ! t I ' ' !All &toods I 32.0 I 34.3 I 47 ..8 I 54.9 , I I I i ltbn-users 68.0 I 65.7 I 52.2 � 45.1 f Source: Y:irl.d Fertility Sur,.iey (1975); lbtld Bark �y (1979); Family �al.th "Bureai Survey (1982); am Westingoouse &a:,.,ey (1982).  -21- 4.12 There is concern about the present mix of contraceptives used, and especially the significant _and probably increasing use of traditional methods. lBased on the results of the 1982 Westinghouse Survey, the proportion of currently married women using tr aditional family planning methods wai3 24.5%, compared to only 13.2% in 1975. This heavy reliance on traditional family planning methods is problematic. The issue is whether both the t:Lming and number of desired births can be ensured given the continued high use of these methods, especially by young, lower- parity women. Th:Ls is particularly important given the significant proportion of women reporting that they did not want any more children (ranging from one quarter to one third of women aged 20-29 with two or mre children) or that they desired long birth intervals. 4.13 'I'he mix of modern contraceptives used has changed significantly over the years, shifting from reversible methods principally IUDs or orals, toward terminal methods, which havei dominated family planning efforts since 1974. The decline in IUD use since 1977 is consistent with a general worldwide trend. The utilization of oral contraceptives, the mainstay of numerous programs elsewhere, has remained low, despite significant steps taken by the Government to expand their availability by delegating authority .for prescribing and dispensing to midwives and nurses, and by establishing the approximately 1,200 "sales points" noted earlier. The 1982 CPS ri�vealed that only 2.6% oJ: currently married women were presently using this method. In fact, even the increase in the reported use of condoms since 1975 has been twice as rapid as that of the pill. 4 .14 In the period 1975-1979, the annual number of sterilizations performed was in the range 19,000-25,000. However, in 1980 it rose significant:ly to over 100,000; and the dif ferenti� results of the 1975 and 1982 surveys show that sterilizations contributed almost half of the increase in the CPR. Further, as 1982 CPS data reveal, female sterilizat:Lon remains the most wideily used single method by currently married women over the age of 25, cmd its use even in the 25-29 year age group (23.3%) exceeds that of the pill, IUD and condom combined. One of the main rHasons for the popularity of sterilization has been the financial incentives provided by the Government to both acceptors and providers. The incentives paid to acceptors is the same for both males and females, but the actual level has varied considerably since first introduced in 1980. From a basE? of Rs .100 plus 3 days employment leave instituted in January 1980, incentives have fluctuated from Rs .500 (October 1980 - February 1981), to Rs.200 (March 1981 - December 1981), to Rs. 300 (January 1982 - May 1983), and to Rs. Rs.500 again, or the equivalent of about US$20, on June 1, 1983. The potential income effect of the payments for providers is also substantial: doctors outside C:olombo with limited scope for private practice can increase their salaries by 50% or nnre in this way. 4.15 The number of sterilizations ( both male and female) performed has been extre»iely sensitive to the level of incentive payments, as best illustrated by the large increase in sterilizations in 1980. The apparent success of sterilization in attracting previous non-acceptors is shown by data from the 1982 CPS, as almost half of the women/husbands had never used another method. However, results of an 1981 FHB study corroborate earlier  -22- Bank study findings of the typically high parity of steril ization acceptors, with 86% of the femal E!S and 44% of the males having four or more children. The dominance of excess fertiiity as a principal incent ive for adopting sterilization underscorE!S the relatively limited fertility impact (around 1.7 births averted per ac:ceptor, based on 1975 data) inherent in the current acceptor profile. 4.16 Injectables are also increasingly popular. In the FPASL program, for example, it has become a preferred method, attracting 25% of new acceptors in 1979. Despite the limited availability of injectables in the public program, high effective dEimand for them has been sustained even when supplies have had to be obtained at relatively high cost (25-50 rupees per 3-month dose) in the private sector. The UNFPA has estimated a potential increase in demand from 280,000 to 500,000 doses within just a few years if supplies are increased. However 1, there is a need to study both whether increased availability of injectables would attract non-contraceptors or merely substitute for other methe>ds, and also the relative cost-effectiveness of this method. 4.17 In summary, the over emphasis of the family planning program on sterilization at the expense of reversible methods considerably reduces the ability of the program to attract: priority target groups including low-parity women and some ethnic minorities traditionally opposed to terminal methods. Moreover, widE!r availability of reversible methods is __particularly important given the significant demand ·for child spacing. This is shown by data from the 1982 cps· which found that 71% of respondents desired to space their next birth one or mre ye.ars, yet only 46% were contracepting and of these just e>ne quarter were using a modern method. �so, concerning service delivery, the rapid increase in 1980 in postpartum sterilizations requidng a 3-5 days hospital stay created major work pressures on the government hospitals; and these problems have recently been compounded by a reduction due to emigration in the numbers of doctors available to perform the operations. 4.18 The Fertility Impact of Contraceptive Use. In the absence of detailed statistics on acceptor patterns, the precise fertility impact of contraceptive use in Sri Lanka is difficult to measure. An estimated 88,617 births were averted in 1980, based on FHB figures which indicated a total of 350,261 women in the reproductive age groups (15-49) who either accepted sterilization, or were. using IUDs or orals in 1979-1980. Comparing this figure to the actual number of births in 1980 of 407,243, Bank analysis has concluded that family planning efforts accounted for an 18% reduction in the CBR (from a potential 34.9 to 28.7 per 1000 population). 4.19 The Sri Lanka family planning program has not achieved its full potential. Based on contraceptive usage patterns in 1980-81, Bank estimates indicate that only 28% of couples were effectively protected. While the coverage of the target population has clearly increased over the 1975-1980 period, according to Bank estimates about one half of currently-married women in the rEiproductive age groups remain to be reached. The typical profile of the potential acceptor (expressing a desire for no mre children but not contracepting) is a married women 35-44  -23- years of age, illiterate, with three or nnre surviving children, and of the Hindu religion. Key Program Opportunities and Constraints 4.20 Promising Signs. Despite the setbacks in the program over the past decadE!, there is no question that the program has the potential for considerably more impact than in the past. Many essential precondition s already exlst: (1) there is a high level of present contraceptive knowledge, with 99% of women knowing at least one method, and notably no significant difference by level of education; (2) there is also substantial manifest and latent demand for contraception with 67% of women overall indicating the desire for no mre children, including one out of every two women by age 30 and four out of every five by age 40; and (3) a substantial proportion (71%) of those wanting another child also want to postpone the next birth for at least one year, according to the 1982 CPS. 4.21 The present Government remains extremely supportive of population efforts and has taken several actions to both increase demand and expand service coverage. Of particular note has been the assignment of overall population policy to the highly visible and powerful MPI, under the direction of the President, and the introduction of strong incentives to reduce des:lred family size. The hi.gh priority assigned to education; particul!1rly female education, and efforts to enhance the overall status of women can also be expected to generate increased demand for family planning. In addition, the private sector has reached a significant and rapidly inc::reasing number of new acceptors, and this channel of delivery could sign:Lficantly raise the CPR. 4.22 Continuing Problems. Despite the potential for improvement, the challenges facing the family planni.ng program should not be underemphaaized. Over the last dec:ade numerous problems have plagued the :Qrogram, adversely af fecting both demand and supply of services. The future performance of the program, particularly in the public sector, will depend sub1;tantially on its ability to redress certain key problems. In the first ]Place, the present program is unable to attract young, low parity wom1m to use modern contracE!ptives. Clearly the reasons why this prime targ1!t group prefers to adopt traditional if any methods should be carefully :lnvestigated, and targetE!d efforts undertaken to increase demand for mdern methods. 4.23 Second, there are still significant gaps in knowledge of where to obtain fam:lly planning supplies. While knowledge of contraceptive methods is virtually universal in Sri Lanka, a substantial percentage of non-users in 1982 st:lll reported lack of information on where to obtain condoms (22.8%) and pills (14.5%), despite their wide availability through both {)ublic and non-governmental sourceE1. Even with regard to injections and IUDs, 11.1% and 8.4% o f non-users, respectively, indicated lack of knowledge of sources. And despite an extensive population education ·program utilizing not only mass media but increasingly interpersonal communication, some important consumer information gaps appear to exist. The 1982 CPS confirmed the 1975 WFS findings that disapproval was not a major caus,! of non-use (only 2%). Nevertheless, it did reveal a  -24- considerable (8%) fear of side effects; and rost (70%) of these respondents were within a prime target group, 25 to 40 years of age. Further, a 1982 follow-up survey by the FHB of program drop outs revealed that among the 39% who had discontinued use of pills and IUDs, almost half the discontinuations were due to side effects. Additionally, 24% of drop-outs felt that family planning methods were harmful to their health and 19% indicated husband disapproval. These data suggest the need for strengthening follow-up of acceptors, as well as for IEC activities targeted to both consumers and providers, with special attention given to gaining family planning acceptance among husbands. 4.24 Third, as the '1982 CPS data indicate, there is a continuing problem of accessibility. While the Government reportedly then provided more than 90% of national family planning services, including 93% of IUDs and 97% of female sterilizations, non-governmental sources accounted for 85% of supplies of condoms and 36% of supplies of pills. Despite the large public and rapidly growing private supply network, accessibility is still limited at the local community level, particularly in rural areas. As the 1982 CPS highlighted, more rural than urban women had to use some means of transportation to obtain contraceptives, whatever the method. Additionally, with the exception of sterilization, for which all women on average spent more than 60 minutes travel time, rural women generally spent more than 30 minutes in reaching a source of supply, as contrasted to less · than 15 minutes for urban women, Further, three times more rural than urban women indicated inconvenience in obtaining contraceptive supplies, with the exception of pills and condoms • . As the 1,200 sales points are limited to standard working hours, access to them is restricted for those in the labor force. These data in the aggregate indicate a continuing supply constraint which must be addressed for usage to increase. 4.25 Fourth, the quantity of contraceptive supplies is inadequate, and the available mix inappropriate. Weak logistics capability, including inadequate purchasing and supplies management have impeded the effective distribution and utilization of contraceptives and equipment. Prolonged delays in clearing goods through customs, and insufficient attention to matching and disseminating inventories according to needs result in stock imbalances and frequent shortages of high demand items such as injectables. 4.26 With regard to modern contraceptives, as the 1986 phase-out of external commodity support by UNFPA draws near, the Government must plan to assume full responsibility for procurement and strengthen local purchasing and supplies management. The SIDA training of four local officers prior to their phase-out of commodity support in 1980 was most useful, but additional efforts are needed. Of particular concern has been the recently reported discard of millions of rupees worth of expired pharmaceuticals. Improved market knowledge and inventory control are critical to ensure that this costly, foreign exchange intensive, part of program operations is handled efficiently. 4.27 Prices of contraceptives vary widely by source. The higher prices of condoms and pills through the social marketing scheme than via the government outlets (Rs.3.50 vs. Rs.0.50 per cycle of pills and Rs.0.33  -25- vs. Rs.0.05 per condo.m, respectively) does not appear to have constrained demand, with an estimated 6.8 million condoms and 0.37 million cycles of pills distributed through_the approximately 6,009 outlets of the social marketing schema in 1981. However, soma usage appears to be highly price elastic, with the demand for sterilization rising and falling as cash payments to both providers and consumers, have been adjusted over time. The effect of current prices on demand for contraceptives, particularly among the still substantial group of·non-users, should therefore be carefully studied. 4.28 At the same time, given the high continued preference for traditional methods, further study should be undertaken of their effectivenE!ss. If appropriate these alternative mathods should be incorporated into the formal service delivery system. 4.29 Fifth, drop-out rates are high for pill users. Data on continuation, though limited largely to sample surveys by the FHB, indicate a high twelve-month dropout rate of 64% among pill users, with the highest proportion of discontinuation during the first three m>nths of use, largely due to reported adverse side effects probably mainly as a result of the use of high-douage pills. These data suggest that both initial education and closer follow-up are required if oral contraceptives are to achieve wider acc�ptance� In contrast, the discontinuation rate for IUDs was only 26% even after 24 months. The high continuation rates for injectables at both 12 and 24 months (75% and 55%, respectively) also suggest the value of exploring more carefully the potential for expanding their· still limited use within the national family planning program. 4.30 Sixth, public health staff are too few in number and lack motivation,. Serious understaffing of health facilities, particularly at the peripheral level, also has limited the availability of family planning services. Despite extensive inservice training programs in family planning, :Lt continues to be assigned low priority, being commonly perceived as an additional, unremunerated task among diverse family health responsibilities. The lack of strong program commitment and poor follow-up of drop outs has been abetted by weak supervision attributable to transport shortages, an inadequate program reporting system, and a reported "go-slow" policy adopted in recent years by medical officers. 4.31 Seventh, program efforts are inadequately coordinated and managed. The multitude of internal ministries, non-governmental organizations, private practitionet·s and external assistance agencies involved in population has also impeded efficient and effective program management 1, in the absence of strong, centralized leadership. The main coordinating body, the Steering Committee, was first established with UNFPA assistance in 1973. It consists of all major actors from line ministry representatives and domestic voluntary agencies to multilateral and bilateral donors, and is chaired by the Secretary for Plan Implementation, but it has become too large to respond rapidly to problems. 4.32 The fragmentation of policy and program planning, implementation, and evaluation activities across both the MPI and the Ministry of Health  -26- also has impeded the development of a cohesive, strong program. The delegation of all these interrelated program responsibilities to the FHB would be one way to permit expansion and upgrading of service quality. 4.33 Weak management remains a key bottleneck. Program performance remains highly uneven both over tire and within a given period across the 15 SHS divisions. Of the 8,535 new acceptors reported in the government program in January 1982, two divisions (Monaragala and Amparai) had less than 100. The wide variations in acceptance rates across SHS areas underscore the need for strengtheming field monitoring and supervision, and for adapting programs locally to respond to the ethnic and religious diversity of Sri Lanka. B. HEALTH Organization and Management 4.34 Until recently the government health services were organized under the Ministry of Health (MH) and two project ministries - the Ministry of Colombo Hospitals and Family Health, which included the FHB and was responsible for primary care and government family planning interventions, and the Ministry of Indigenous Medicine. In mid-1983, however, a new cabinet-level minj_stry wa;s created--the Ministry of Women's Affairs and Teaching Hospitals. The MH is still responsible for all other health services, both pr�ventive and curative, and in the new arrangement the FHB is now directly within it. 4.35 To stimulate intersectoral coordination and community participation, a national health development network has been established. It consists of a National Health Council, chaired by the Prime Minister and with 8 other ministers as members. It is supported by the National Health Development Committee, chaired by the Secretary of Health and consisting of the secretaries of the ministries represented on the Council, health officials, and representatives of' certain international organizations. Six standing sub-committees have been formed to provide technical support. 4.36 For the management of government health and family planning services, the country has been divided into 19 health divisions which range in population size from 220,000 to 1,480,000 (although Colombo and Gampaha combined have a total population of 3,320,000). Each division is under the authority of a Superintendent of Health Services (SHS). The divisions are further subdivided into health areas, of which there are 109 under the MH. Before introduction of the new PHC model, a Medical Officer of Health (MOH) headed the area team which included Public Health Inspectors (PHis), Public Health Midwives (PHMs) and Public Health Nurses (PHNs). The area team was responsible for preventive health, maternal and child health, and domiciliary care. On the other hand, curative medicine below the tertiary level was the responsibility of the District Medical Officer (OMO), who was in charge of the district hospital.. Under the new PHC model, the preventive and curative health services are being integrated intp a single system. The proposed changes in physical infrastructure are shown in Table 4.  -27- Table 4: A Cllfi>ARISON CF EXISTIN; HFAI.l'H C.ARE IELIVERY SYSTEM AND t-Ei PlC MODEL Existing Facilitles New POC M:ldel Level N:>. Conversim N:>. Facility Level Tertiary Te� lbspital 13 Te� I Si;ecialty lbspital � a lbspital I l L I Provincial H:>Spital 9 9 Provincial Referral. 1bspita1 I Ba9e lbspital 17 Upgrade ir I I 24 Ill.strict + N:!w 7 lbspital I J J L Secondary Ill.strict lbspl.tal 11� Peripheral Ulit -I -I 1 Upgrade 21 Illv1s1ona1. 09 L a1 I Rural lbspi ta1. 102 + N:!w 67r 285 12alth , Center II ::J I Matemities Central Ill.sp:msary L 538 Sub- I Illvisional L Upgrade � Rest lbne I L73o 12alth POC _I + New 2181 - Center I I ldopt ot'ter I Conm.tnity Health N:> 1:sse facilities -:::I lOOL I '397 Granooaya I 0 I + New 7 01 l2alth I 38 Center I Source: Ministr y of l2alth 4.37 Urban and municipal councils are responsible for preventive heal th care! and enforcement of public heal th regulations in their areas; in Colombo Municipality the council also provides soue basic care, mostl y to underprivileged areas. In the estates and plantations, the corporations provide sone basic health care through their own dispensaries and small hospitals. Due to shortages of uedical manpower, however, the standards of care in the! estates vary considerabl y.  -28- 4.38 Some special programs have al.so been established. First, the malaria program started in 1946, and rapidly reduced morbidity and mortality from the disease. Most peripheral health workers have received training in malaria control, but full integration of the program into the general health services is not planned because of the risk of a major epidemic if control activities falter. Second, there is an expanded program of immunization, which operates through the governmental heal th service structure. Third, the G<>vernment is also establishing a program with UNICEF assistance to reduce the incidence of infant and childhood mortality from diarrheal. diseases. This program will .include health and nutrition education, improved personal hygiene, and distribution of oral rehydration salts; in the long tE!rm, however reduction in case incidence will depend on improved sanitation and the provision of safe water. Fourth, the main government programs for food supplementation are the thriposha program and the school biscuit program, the latter of which is being terminated. The aim is to feed malnourished children and mothers, and estimates of beneficiaries vary between 550,000 and 850,000. Facilities and Services 4.39 The Government's patient care services are provided through ·a three-tier structure of instituttons: primary, consisting of about 340 dispensaries and 102 maternities; secondary, consisting of 114 district hospitals, 109 peripheral. units and 102 rural hospitals; and tertiary, consisting of 4 teaching hospitals, 8 specialty hospitals, 9 provincial hospitals and 17 b-ase hospitals. Overall, there are approximately 2.4 beds per 1,000 population, ranging from about 1.5 in Amparai and Puttalam/Trincomalee, to about 306 in Kandy (See Table 5).  -29- Table 5: DISTRIBUTION OF HOSPITAL BEDS, 1981 Secondary and Tertiary Level Estimated Beds per 1000 SHS Divinion Beds Po12ulation Population ('OOOs) Amparai 587 389 1.51 Anuradhapura 1,974 588 3.36 Badulla 1,631 643 2.54 Batticaloa 769 331 2.32 Colombo 3,753 1,698 2.21 Galle 1,782 815 2.19 Gampaha 2,365 1,398 1.69 Jaf fna 2,502 831 3.01 Kalutara 1,945 827 2.35 Kandy 4,052 1,126 3.60 Kegalle 1,550 682 2.27 Kurunegala 2,830 1,213 2.33 Matale/Pollonaruwa 1,504 629 2.39 Matara/Hambantota 2,003 1,068 1.88 Monaragala 635 280 2.27 Nuwara Eliya 1,075 522 2.06 Puttalam/Trincomalle 1,092 750 1.46. Ratnapura 2,353 796 2.96 Vavuniya/Mannar 678 281 2.41 Sri Lanka 35,060 14,850 2.36 Source: Mlssion estimates. 4.40 The utilization of health facilities varies considerably both among instlltutions of di f ferent types, and also among institutions of the same type. The general pattern, partly because of good transport is for cliEmts to bypass lower level facilities in favour of mre advanced ones. Tertiary level facilities are overcrowded on average, especially by inpatients., Indeed, more than half of the provincial hospitals have average occupancy rates over 100% (see Table 6), while in contrast the inpatient occupancy rate of secondary care facilities is only about 75%.  -30- About one fifth of all secondary care facilities have an occupancy rate below 50%, although another fifth have a rate about 100%. Table 6: UTILIZATION OF HOSPITAL B EDS BY TYPE OF FACILITY, 1982 .L I I !Standard I I Average I Deviation I Percent of Hoseitals Total Number I Bed t of Bed I Below 50% !Above 100% Ti:ee of Hoseital of Beds Occueanci: Occueanci:I Occupanci: 1occupanci: (%) I I Provincial I I 7,443 109 17.6 I 0 I 56 Base 5,396 99 21.8 I 0 t 35 District, Large 8,054 70 22.8 I 16 I 12 District, Small 4,126 79 29.1 17 18 I I Peripheral 4,326 71 38.7 I 23 I 21 Rural 2,442 75 52.3 I 28 I 26 I I Total 31,787 86 I 21 I 23 Note: Data on only about 90% of institutions were available. Source: Mission estimates. 4.41 A comparison of data bE!tWeen 1972 and 1981 shows that the total number of beds increased by about 10%, while the total number of inpatient discharges increased by about 14:t. The average duration of patient stay declined significantly for all categories of institutions, ranging from 1.5% in peripheral units to 44% in district hospitals. Consequently, the average number of inpatients per day in secondary care institutions declined from an estimated 20,800 to 13,800; whereas it increased from an estimated 12,800 to 14,000 in tertiary level institutions. In terms of bed utilization, therefore, the occupancy of secondary care institutions declined from about 125% in 1972 to the 1981 level of 73% •.This was partly due to a general shortage of medical manpower along with an expansion of tertiary care institutions, shortages of financial resources especially at secondary (and primary) levels, and the bypassing of lower level facilities by patients. The bypassing phenomenon was facilitated by a good transport system. Maneower and Training 4.42 The major categories of: health care providers are doctors, nurses, assistant medical practitioners (AMPs), PHMs, PHNs and PHis. Al though they all register themsE!lves after qualifying, registration lists are not updated later to reflect migration or retirement. It is therefore difficult to estimate the total number of health personnel active in the country. The total MH establishment was about 21,400 in 1982, having  -31- increased by about 20% since 1978. The approved cadre for doctors has not changed sin.ce 1967, and about 15% of the positions are vacant. About 40% of the key SHS positions are presently vacant, ·as are many of the MOH positions. About 10% of the positions for nurses and AMPs are also vacant. In terms of geographical balance, those districts which have teaching hc1spitals have the largest concentrations of hitE?h manpower• Five distd.cts -- Nuwara Eliya, Matar a, Amparai, Monaragala and Kegalle have signiHcantly fewer doctors and AMPs per 1,0_00 population than the national average; and four districts -- Vavuniya, Nuwara Eliya, Monaragala and Amparaj_ - have significantly fewer nurses (see Table 7). In terms of overall health manpower, therefore, Amparai and Nuwara Eliya are ovst under served.  -32- Table 7: DISTRIBUTION OF HEALTH STAFF BY SHS DIVISION, 1981 Assistant Public Medical Medical Health Supporting Officers Practitioners Nurses Staff Staff Attendants Total Rate Rate Rate Rate Rate Rate Rate per per per per per per per 100000 100000 100000 100000 100000 100000 100000 SHS Division No. Pop. No. Pop. No. Pop. No. Pop. No. Pop. No. Pop. No. Pop. Amparai 20 5.1 27 6.9 35 9.0 78 20.1 24 6.2 48 12.3 232 59.7 Anuradhapura 73 8.6 69 8.2 212 25.1 199 23.6 93 11.0 201 23.8 847 100.3 Badulla 78 12.1 41 6.4 234 36.4 158 24.6 57 8�9 178 27.7 746 116.0 Batticaloa 41 12.4 28 8.5 84 25.4 53 16.0 46 13.9 70 21.2 · 322 97.3 Colombo 596 35.1 52 3.1 1728 101.7 498 29.3 416 24.5 706 41.6 3996 235.3 Galle 133 16.3 48 5.9 299 36.7 290 35.6 101 12.4 224 27.5 1095 134.4 Gampaha 155 11.2 70 5.0 480 34.5 480 34.5 157 11.3 407 29.3 1749 125.9 Jaffna 139 16.7 66 7.9 242 29.1 233 28.0 140 16.8 278 33.4 1089 132.1 Kalutura 86 10.4 32 3.9 335 40.5 369 44.6 77 9.3 244 29.5 1143 138.2 Kandy 248 12.4 80 4.0 718 35.8 404 20.1 189 9.4 398 19.8 2037 101.5 Kegalle 51 7.5 41 6.0 244 35.8 286 41.9 63 9.2 202 29.6 887 130.0 Kurunegala 144 11.9 107 8.8 431 35.5 483 39.8 138 11.4 415 34.2 1718 141.7 Matale 29 4.7 45 7.3 158 25.5 153 24.7 69 11.1 176 28.4 630 101.6 Matara 63 5.9 36 3.4 222 20.8 329 30.8 82 7.7 288 27.0 1020 95.5 Monaragala 11 3.9 24 8.6 36 12.9 80 28.6 13 4.6 62 22.2 226 80.8 Nuwara Eliya 32 6.1 29 5.6 99 19.0 66 12.6 36 6.9 86 16.5 348 66.6 Puttalam 35 7.1 40 8.1 134 27.2 142 28.8 54 10.9 103 20.9 508 103.0 Ratnapura 93 11.7 44 5.5 284 35.7 229 28.8 83 10.4 267 33.5 1000 125.6 Vavuniya 34 12.1 40 14.3 43 15.3 70 25.0 38 13.6 90 32.1 315 112.4 Special Campaigns 76 0.5 5 o.o 220 1.5 169 1.1 99 0.7 161 1.1 730 4.9 Other 96 0.6 1 o.o 199 1.3 16 0.1 119 0.8 279 1.9 710 4.8 Sri Lanka 2233 15.0 925 6.2 6437 43.3 4785 32.2 2094 14.1 4883 32.92 21357 143.8 Source: Ministry of Health  -33- 4.43 It is estimated that there are about 3,430 doctors (one per 4,431 population or 22.6 per 100,000 population) of whom 2,230 are serving in government health facilities. This ratio of doctors to population is somewhat higher than the average for low-income economies as a group, as shown in the 1984 World Development Report. However, most of the government doctors are concentrated in tertiary institutions, with about 32% in teaching and specialist hospitals, and another 51% in provincial and base hospitals (see Table 8). Only about 17% of government doctors work in secondary facilities. Most rural hospitals and dispensaries are manned by AMPs. In addition, there are about 6,800 nurses (one per 2,180 population), with about 28% in teac.hing and specialist hospitals, 50% in provincial and base hospitals, and the remaining 22% in district hospitals. On the preventive health side, there are about 4,800 staff including 227 PHNs, 1,850 PHMs, and 925 PHis. Table 8: DISTRIBUTION OF HEALTH STAFF BY TYPE OF FACILITY, 1981 Medical I Assistant Medical I Type of Facility O fficers I Practitioners I Nurses Midwives I . I (%) I (%) I (%) (%) Colombo and. Specialist Hospitals I I 32o3 I 1.2 I 27.6 12.1 Provincial Hospitals 32.8 1.9 29.5 12.4 Base Hospitals 17.8 I 4.3 I 19.2 14.0 District Ho,spitals 11.4 I 17.0 I 21.5 27.6 Peripheral Units 5.2 I 15.3 I 1.9 16.5 Rural Hospitals 0.5 I 11.8 I 0.3 7.6 Dispensaries and I I I I II II Maternity Homes 0 48.5 0 9.8 Total 100 I 100 I 100 100 Note: Medical officers exclude specialists. Source: Mission estimates based on information from the Ministry of Health. 4.44 In 1981 among the tertiary level facilities as a whole there was one doctor for about 16 inpatients, and one nurse for about five inpatients. In contrast each doctor at the district hospital level had an average of 28 inpatients. On average, each doctor or AMP treated about 110 outpatients and about 10 inpatients daily. In dispensaries the daily attendance averaged 125 outpatients. However, each midwife in a maternity hone attended only about 76 births .annually, while each midwife in a provincial hospital attended about 320 births. This difference is due to the relative under-utilization of the maternity homes and indeed of secondary-level facilities in general.  -34- Table 9: WORKLOAD OF HEALTH ST AFF BY TYPE OF FACIL ITY, 1981 1-----·-------Inpatients-------------1 Live I Daily I Medical Nurses !Births per I Average Number I Of ficers per In- Year per Type of Facility I I and AMPs patient Midwi fe Colombo General Hospital I I 2527 : I per IP 10.1 3.2 Colombo Group of Hospital s! 2085 I 10.0 3.7 Provincial Hospitals I 8608 I 16.9 5.1 316 Base Hospitals I 5354 I 17.6 4.9 260 District Hospitals I 8926 I 27.6 7.3 259 Peripheral Units I 3056 I 14.2 235 Rural Hospitals I 1832 I 15.9 195 Maternity Homes I I 76 I I Total I 32388 I Source: Mission estimates. 4.45 A staf fing norm study was carried out by t�e Government in 1980. However, the appropriateness of the output of new health workers from existing training programs is dif ficult to assess in the absence of anticipated net emigration rates and estimated needs of the private sector. The output of nurses, PHis and laboratory technicians is planned by the Government to provide for replacements and estimated population growth, although apparently existing vacancies have not been taken into account. The annual output of doctors (about 380) is quite high, as it is equivalent to 23% of all doctors presently in government service. Indeed, in order to compensate for the past exodus of doctors, with about hal f of· all graduates emigrating within 10 years, the Government has increased the annual intake for medical training from 275 to 400, with the first enlarged batch due to qualify in 1984. The existing approved cadre for government doctors is 2270, although the 'staf fing norm study suggested a need for 2750. Even with 50% loss into the private sector or through emigration, the MH would still be able to fill the existing vacancies in about two years and its staf fing norms in about five years. 4.46 In addition to the professional. health staf f, the Government has also trained a relatively large number of health volunteers, mainly young girls, nearly 3,000 of whom were trained in 1981 alone. These volunteers are expected to serve as communicators and facilitators for various programs such as the expanded program of immunization and diarrheal. disease control. However, the system is informal and their ef fectiveness has not been assessed.  -35- Support Services 4.47 _!:quipment. The State Medical Stores purchases all medical equipment. Although health facilities generally have appropriate items of equipment, only about 70% are in working order. Maintenance problems arise due to anti.quated equipment, inadequate financial resources, lack of technical support staff, and cumbersome ordering procedures. For example, about 60% o,f all x-ray and ECG equipment is more than ten years old; and budget allc,cations for repair and maintenance work have averaged O. 6% of the total value of the equipment, whereas a norm of about 4.0% would be appropriate:. The Electro-Medical Engineering Unit located in .Colombo and three small peripheral units are responsible for the maintenance work, and for some ti.me the Unit lacked a chief engineer. Despite the continuing resource cc,nstr·aints, there have recently been some improvements in procedures and services. 4.48 Transport. There are al.so major inefficiencies in the transport system. Altogether the MH has about 900 vehicles and 400 motorcycles, but 50% of the fleet is m:>re than 5 years old. A maintenance workshop in Colombo, is about to be closed down, where upon a maintenance and repair allowances will then be provided to each SHS and private facilities will be used insteaLd. While the proposed new system may be more efficient than the present sys:tem, it is important for financial allocations. to be increased for both the operation of vehicles and also their upkeep. 4.49 Health Information System:• Data on facilities, staff and services are sent from all health institutions to Colombo. There are considerable demands on the field staff to provide data - for example, a PHM is expe:cted to keep about 13 registers and 5 different cards with data on clients receiving services. However, there is little analysis of information. at the level of the health division, let alone the health area, and virtually no feedback to field staff. However, a new health information system is now being pilot-tested in one district. 4.50 Pharmaceuticals. Around 90% o f all drugs and pharmaceuticals are imported. Procurement of them for the Government is the responsibility of the State Pharmaceutical Corporation. Its total procurement in 1982 was about Rs. 180 million, just over half of the country's total consumption of pharmaceuti.cals. Procurement is normally through competitive international bidding, which secures a good price but a heterogeneous origin and possibly uneven quality. Only 3% of the pharmaceuticals purchased by the Government are through local order. The State Medical Stores, which stocks about 8,800 items, distributes all pharmaceuticals for government facilities. The main problems in the system are weaknesses in logistical support and shortage of trained manpower. Inappropriate and excessive use of pharmaceuti.cals,. especially antibiotics is common, costly and harmful because of development of_ drug resistant bacteria. The Government plans soon to fin.alize a list of essential drugs suitable for each level of heal th care:, and this could lead to significant savings and efficiency improvements if properly used.  -36- Non-Government Health Services. 4.51 About 1200 doctors were practicing privately in 1981. Government medical personnel are also allowed to do private practice during their of f-duty hours. Approximately 55 small private hospitals have a total of about 1650 beds, 70% of them concentrated in Colombo and Jaf fna. The number of patients seeking private treatment has almost certainly been increasing in recent years-presumably at least partly as a result of dissatisfaction with government health services. Voluntary agencies are also active in health--most notably the Sarvodaya Movement, which trains community health workers and also runs pre-schools. There are about 10,700 traditional medical practitioners registered with the Sri Lankan Ayurvedic Medical Council, but the actual number providing curative services may be much larger. Water and Sanitation. 4.52 While Sri Lanka's social indicators place it in the top 5% of developing countries, its supply of safe water places it in the lowest third. In 1981 only 9% of the population received piped water through home connections, and only 10% more used public taps. The rest drew water from surface sources and wells. Saniltation is similarly limited. The Government has announced plans to provide adequate water supplies to all urban and estate populations, and to 50% of rural populations, and also to provide adequate santitation fac:llities for the entire population, by 1990. However, these targets appear to be beyond the technical and financial resources available. Considerable improvements also need to be made in the coordination of the various agencies involved -- the National Water Supply and Drainage_ Board, the estate boards, the MH and the municipal authorities. V. FINANCING AND EXPENDITURES Past Trends 5.01 Sri Lanka has achieved social progress far in excess of that of other countries with comparably low per capita income. One of the reasons for this was the high proportion of government expenditure devoted to the social sectors, including health, in the 1960s and much of the 1970s. In the late 1970s, declining GDP growth led to a reorientation of government expenditure to the productive sectors, and the proportion of the government recurrent budget devoted to the health sector dropped: whereas from 1971/72 to 1977 it had averaged 7.6%, in the period 1978 to 1982 it averaged only 5.7%. In nominal terms, the value of actual recurrent health expenditures by the Government rose considerably from Rs 252 million in 1971/72 to Rs 1,126 million in 1982; but if price escalation and population growth are both taken into account, the per capita real value of the Government's expenditures was quite similar in 1982 to what it had been 10 years earlier (see Table 10).  -37- 5.02 The relative cutback at the end of the 1970s in government heal th expend!turE�s was 1argely at the expense of community care services. In 1977 these expenditures accounted for 28.0% of.the total, but by 1980 they had fallen to 20.6%. In contrast, administrative and training expenditures rose from 5. 9% to 7. 5% o f the total , and expend! tu res on patient care services rose from 66.2% to 72.0%. ·Much of this trend can be explained by a relative decrease in expenditures on the malaria program; and between 1980 and 1983 the proportion of the budget devoted to community care services rose again to 24.1%. At the same time, however, there was an increase in the proportion of the government recurrent budget spent on hospitals in Colombo: in 1977 these hospitals had accounted for 14.1% of the total, whereas in 1983 they accounted for 16.7% (including the al lowance for the new Kotte Hospital). 5.03 In contrast to the Government's recurrent expenditures, ·the Government s capital expend! tures on the health sector rose in re al and per 1 ' capita terms through the entire period. In constant prices, total actual expenditurE!s rose at a compound growth rate of about 18% p.a. from 1971/72 to 1982. The ratio of actual capital expenditures to actual recurrent expenditurE�s therefore rose from 2.4% to as much as 19.7%. As the annual recurrent budget requirement s of social sector capital expenditures tend to be relativHly high (typically in the order of 20-30%), such a trend in this ratio has important implications for future health recurrent budgets.  --38- 'Dible 10: aM!Jffl! IIIDlUU! IE\Llll FSmtlll: NII lmlAL Elll'l!HffllRl!S, --IN-QJQIEHl' NO Clll6rNl PRICES, NO ffll CAPirA, l9n/n-l983 Year Hln1stry of Health Ministry of Cblomo Hlnistry of Imtp- leal.th Btpenl. - % ltoalth /dual. 'lbtal. Heal.th Keal.th /dual. Btpenl. i!stimii Actual lblpltal.e ,nl Fmil:[ lul.tb FstiJralll Actual ..,.. Hed1dne iWld De- !111tmsd.<Jf� Fstillllta Jc 1t>tal. Health HID1str1ea Fatillllta Jctual 1t>tal. Q),emzen Reaarmc l!'xi>endl.ture Fstimlte Acl>JjlJ. Ix!!!!!, a,r !":!!f!ta (0Jrnllt Prices) cecm= Actual. 11<1!!!!. A!r!":!!f!t.a Prices) (Cora= Prices) (Ba.If) (Ba.If) (Ra.If) (Ba.II) (Ba.II) (Ba.If) (Ba.14) (Ba.14) (Ba.If) (Ra.14) (Ba.If) (Ba.14) (Rs.) lffl/n 307.1 305.2 10.s 10.0 317.6 315.2 8�4 (1971/n) (245.7) (244.2) (8.4) (8.0) (254.1) (252.2) (8.4) (19.61) (233.5) (18.16) 1973 258.S 257,1 8.6 11.3 2&7.4 21i8.4 1.1 20.50 211,7 16.20 1974 287.6 288.2 9.0 9.0 296.6 297.2 7.6 7.1 22.11 186.0 14.00 1975 314.6 323.6 11.0 10.6 325.6 334.2 6.7 6.9 24.76 193.0 14.30 1976 173.9 392.2 14.3 14.3 J!B.2 «l6.S 1.1 7.8 29.63 222.1 16.19 1'117 405.6 449.3 13,5 13.4 419.1 462.1 1.0 7.6 33.19 214,5 15.)J 1978 477,7 510,0 13.4 13..3 181.1 523.,3 4.9 S.3 36.89 225,0 IS.86 1979 630.2 632.2 19.2 15.3 649.4 647.5 5,9 5.9 44.74 21o0.6 16.63 lCJID 104.6 740.5 :zr;.o 21.8 730.6 762.3 6.4 6.0 51.n 2ll.7 16.27 -· 1981 82ll.8 83217 :zr;.5 21.6 847.3 853.3 5.4 5.6 56.8) 222.1 14.82 1982 751.2 866.4 22ti.4 234.7 'J!J.7 25.2 1007.3 1121i.3 4.8 S.9 n.90 2ti7.4 17.31 1983 1005.8 280.0 30.4 1317.2 5.6 ne flgun,a 1n bra:lo!t ..., a t9ntn, � an all.._ 12 IIDl1l:h 'blllta. tn 19n/n tho � fl.aail. iear l.mlllll lS IDltlw, • thon - a� ID a:dDcldo tte fl.ocal :,ur ldch tte calenlar ,ur. - 1ndox uaal 1a tie illplldt QI' defl.a1Dr. Source: Qwe!mm\t -. Mlntatry of Ftnanal ml l'!a1n1n; ..i IIR> O:uttry lblncalc ltomnnla 1982 ,al 1983.  -39- Furthermore, most of the capital expenditures in the last few years have been on tertiary level facilities - particularly the new Kotte Hospital.. To a large extent the pattern of capital expenditures has been influenced by the availability of foreign aid. 5.04 Attempts to monitor the growth of private health expenditures depend on intermittent surveys •. Al though not entirely comparable, the results of the 1978/79 Consumer Finance Survey may be compared with those of the 1980/81 Labor Force and Soci<>-economic Survey. In this period, i.e., at a tine when the Government was adjusting its pattern of recurrent expenditures in line with its new priorities, expenditures in the private sector on western medicine were increasing at about 11% annually. Expenditures on traditional health care were not only a very small proportion of total private heal.th care expenditures, but they were al.so decreasing in this period. Since 1980/81 these trends may be expected to have accelerated further, with the rapid increase in the availability of private western health services. Present Situation 5.05 In 1982 approximately Rs 4,600 million was spent on health care and health related activities, including public and private, and recurrent and capital expenditures (see Table 11). About 3.5% GDP (2.8% GDP recurrent and 0.7% GDP capital) was spent on health care and family planning and 1.6% GDP (0.5% GDP recurrent and 1.1% GDP capital) was spent on health-related activities (primarily consisting of programs in nutrition, water supply, and sanitation). In per capita terms, expenditures on health care and family planning amounted to about Rs. 208 (US$9.8). In contrast to the situation in th� 1960's and much of the 1970's, this expenditure is now what might be expected for a country with a per capita GDP of Sri Lanka. Table 11: EXPENDITURES ON HEALTH CARE AND HEALTH RELATED ACTIVITIES, 1982 I Health Care I I _ j Health Related Activities . I TotalII _ I I� II I-I II I-I I I Amount I Amount I% of I Amount I Amount I% of !Amount IAmount I%of I I (Rs m) I percap I GDP (Rs m) I percap I GDP (Rs m) I percap GDP Recurrent I I lI (Rs) (Rs) II (Rs) . . II I I Expenditures I 2:590 I 168 12.a I 438 I 28 I o.51 3028 I 196 13. 3 I capital I I I I I I I I I I Expenditures! 621 I 40 10.7 I 979 I 63 I 1.11 1600 I 104 11.11 II II I II II II I II I,s.11I Total: 3211 208 ,3.5 1417 91 1.s , 4628 3 00 Note: Totals may not add up due to rounding. Source: Mission estimates  -40- 5o06 The total expenditures in 1982 may be analyzed by source of financeo About half of recurrent health expenditures was financed from general tax revenues (see Table L2)o Cost recovery by the Government is extremely limitedo Day -beds (mainly in Colombo General Hospital) provided the equivalent of only 006% of the Government 9 s recurrent health expenditureso In contrast» however» a large amount of money was paid by individuals in the private health sectoro Although the results of the 1981/82 Consumer Finance Survey are not yet available» an extrapolation of the results of previous surveys suggests that such private expenditures amounted to around Rso75 per capita or about Rs 1»160 million (about 45% of the total recurrent health care i�penditures) in 19820 Health insurance accounted for a negligible proportion of this amounto  -41- Table 12: RECURRENT EXPENDITURES ON HEALTH CARE BY SOURCE OF FINANCE. 1982 (RS m) I I Source of Finance I I Gen. 1For-1Em- jPriv. !Local 1Priv. I I I J Gov.leignjploy-jinsurjVol. jPay- jTotal Rev. Aid ers ance Bodies ments , l l I , l Service Provide:r I Government I I Ministry of Health!/ I 757 66 830 Ministry of Colombo I 7 Hosp. & Fa.mily Health I 235 235 Ministry of Indigenous I Medicine I I26 26 Buildings Department 2/ I20 20 Ministry of Plan Implem._!/ 58 58 Municipal Councils 67 67 Urban/Town Councils 50 50 Other Local Authorities 4 4 Integrated Projects 10 ·1 10 University Grants Comm. 60 I 60 Non-Government I I I Estate Employers 50 I 50 Voluntary Bodies 20 I 201 Private Practitioners I I and Nursing Homes 101 690 I 1001 Pharmaceutical.Suppliers I 260 I 2601 Ceremonial Practitioners I 40 I 401 Ayurvedic Practitioners I 150 I 1501 Medical Education I 10 I 10 I Total: 66 50 I 20 1157 125901 I I 1287 101 Notes: !/ Excluding expenditure (Rs. 36 m) on nutrition programs (here classified under health-related activities) 2/ For maintenance of health facilities 3/ For family planning incentive payments Source: Mission estimates  -42- Table 13: UNIT COSTS AT SELECTED TYPES OF HEALTH FACILITIES, 1982 !Assumed Cost Breakdown Estimated Unit Costs I Type of Health I I I Facilitl !Inpatients _jOutpatients I Inpatient !Outpatient I I % I % Day I Visit I Colombo General I I (Rs.) 72.00 I (Rs.) I I J Hospital Specialised Hosp. 80 20 23.64 16.41 Provincial and Base Hospitals 60 40 22.02 9.98 District and Cottage Hospitals 45 55 16.02 7.80 Peripheral Units & Rural Hospitals 40 60 12.53 4.11 Central Dispensa- ries & Maternity 14 86 17.49 3.08 Homes Source: Government Estimates and mission estimates. 5.08 The total value of pharmaceuticals issued (excluding local purchases) in· 1982 was Rs. 144 million (or 13% of the total actual expenditures of the Ministry of Health and Ministry of Colombo Hospitals and Family Health), against an amount of Rs. 165 million originally approved. In addition to pharmaceuticals, expenditures on dressings (Rs. 36 million), surgical supplies (Rs. 8 million), and x-ray supplies (Rs. 8 million) increased the total amount to Rs. 194 million. Of these expenditures made by the State Medical Stores, 25% went to hospitals in Colombo. Expenditures on pharmaceuticals in these hospitals accounted for 18-23% of their total expenditures, with the exception of the Lady Ridgeway Hospital where it was 28%. All evidence suggests that possible savings on pharmaceutical usage could be made there and at the other tertiary facilities. 5.09 Building maintenance is also severely constrained due to lack of funds. In 1973, the total value of the Government's health facilities was about Rs. 730 million. In 1982 prices, this amount was equivalent to about Rs. 2.4 billion. A reasonable allowance for annual building maintenance is 1.5% of the total value, and this would have been equivalent to about Rs. 37 million in 1982. In that year, however, only Rs. 20 million was allocated for building maintenance, one half the amount required. 5.10 Funds provided for the use of vehicles were also inadequate. For example, the budget for fuels and lubricants is based on the assumption that each vehicle would travel only 320 miles per month. Al though the need for vehicle use varies depending on the local terrain and distribution of facilities this low allowance must impede staff performance. In addition, an average yearly mileage of less than 4,000 is an inefficient use of .. hi ,. , •  -43- an average yearly mileage of less t.han 4,000 is an inefficient use of vehicles •. 5.11 The respective costs of training different types of health workers also suggest inefficiencies in the use of existing financial resources. According to recent estimates from the National Institute of Health SciEmces (NIHS) at Kalutara, the costs per graduate are as follows: AMP (2 1/2 year course) Rs. 19,740; PHI (1 year course) Rs. 7,896; PHN (1 year course) Rs. 7,896; and FHW (18 month course) Rs. 13,767. For tIM:?dical students, on the other hand, the Government spent about Rs. 50 million on its four o!dical schools in 1982, or roughly Rs. 33,000 per student/year. Allowing for a five-year course, the cost of training a doctor would, therefore, be approximately Rs. 165,000 each, or over eight times the cost of training an AMP. However, about 50% of doctors leave the government health services, and indeed the country, within 10 years of graduating. 5 .12 'l'he mst striking featurei of the financing of the family planning program ha13 been its continued heavy reliance on external assistance. Foreign ·aid accounted for about 70% of capital expenditures in 1980. For its part, the MH spent about Rs.39 million on the FHB project in 1982, which was equivalent to about 25% of the budget for community health services, 1)r about 6% of its total recurrent budget. By individual source of funds, two external agencies have dominated--SIDA and UNFPA. Additionally, the FPASL, the major private sector provider, has received numerous grants from several international agencies including its parent IPPF, the International Fertility Research Program and the International Projects Assistance for Voluntary Sterilization. Smaller grants have also been provided other local voluntary agency programs, such as Community Development Services and the Association for Voluntary Sterilization. More recently NORAD and SIDA have provided family planning financing through bilaterally assisted integrated area development projects. In addition, USAID has recently completed a review of the population sector and is considering financial assistance through a new project. Investment Program 5.13 The Government's Medium 'l'erm Investment Programme has identified seven major projects in the health sector: (a) National Institute of Health Sciences, Kalutara; (b) staff quarters; (c) primary health care, including famUy planning (d) Sri Jayewardenapura (Kotte) Hospital; (e) machinery and equipment; (f) State Pharmaceutical Corporation tablet and capsule plant; and (g) sterile product plant. The total capital cost of these projects is estimated to be Rs. 4,374 million. Excluding the Kotte Hospital the PHC project accounts for 68% of the total remaining identified capital costs. In a recent decision of the Government, this project has been given highest priority in health sector development. 5.14 The affordability of the capital costs of the primary health care project will depend primarily on the availability of foreign aid. However, any additional recurrent costs arising from the project are likely to be the responsibility of the Government itself. These latter costs were estimated by the Government at Rs. 108 million, but this figure has recently been raised to Rs. 153 million in order to allow for the costs of drugs and food for inpatients. If allowance is also made for the salaries of the 800 FHWs who are yet to be employed, then the additional recurrent cost would increase to Rs. 163 million. For a health project with such a  -44- high capital cost, the anticipat,ad additional recurrent cost is actually quite small, largely because most of the staff needed are already employed. 5.15 However, the recurrent budgets of the health ministries will face other new demands on them. For example, it is estimated that the Kotte Hospital will cost about Rs.100 million to run. Support from UNFPA for supplies of contraceptives will 1be phased out in 1986, and the additional cost to the Government may be about Rs.40 million. There is also a need to increase funding for certain particular items of expenditure--for example, travelling budgets and allowances, and maintenance of buildings and vehicles--and there is also considerable pressure on the Government to raise the salaries of civil servants. 5.16 The recurrent funds available to the health ministries will depend on the growth of GDP, the proportion of GDP going into the Government's recurrent budget, and the proportion of the Government's recurrent budget going to the health ministries. In 1983, GDP was about Rs. 124,000 million; the Government's recurrent budget accounted for 18.9% of it; .and the health ministries together received 5.6% of the Government's - recurrent budget. In the Medium Term Investment Programme, GDP is projected to grow at 5.6% p.a. It is also projected that the share of GDP accounted for by the Government's. recurrent budget will fall from 18.9% in 1983 to 16.8% in 1987. Assuming that the latter proportion continues to 1990, and also assuming that the health ministries continue to receive 5.6% of the Government's recurrent bud.get, then the additional real resources (in 1983 prices) available to these ministries in 1990 would be Rs. 396 million. However, the annual growth rate of 5.6% projected in the Medium Term Investment'Programme is considered by the Bank to be optimistic, and would depend on the Government adopting a comprehensive set of structural adjustment measures. Using an alternative scenario and assuming a low range 3% annual growth in GDP, the additional resources available to the health ministries in 1990 (again assuming that together they receive 5.6% of the Government's recurrent budget and that the latter accounts for 16.8% of GDP) would be only Rs. 121 million. 5.17 The additional recurrent funds available to the health ministries under the latter scenario would be less than the estimated additional recurrent costs of the PHC project alone, without allowing for any other project related or non-project related real increases in health recurrent expenditures. Even if the scenario is somewhat pessimistic, it is nevertheless clear that the Government must make urgent efforts to: (i) improve the efficiency of resource allocation in the sector, for example by taking measures to ensure that patients are seen and treated at appropriate levels of the delivery system, and by restricting capital investments to only top priority projects; (ii) tighten up operational efficiency through better management, particularly at those facilities which account for the largest proportions of the recurrent budgets; and (iii) extend cost recovery for certain categories o:f patients. However, even with those improvements additional recurrent health expenditure would still be needed to make the most efficient use of past health sector investments. This could be accomplished by cutting back capital investments in health and shifting some of those savings to the recurrent budget.  -45- VI. ISSUES AND RECOMMENDATIONS A. POPULATION 6.01 Sri Lanka faces the prospect of considerable population growth. The rate of natural increase is high at 2.06% yearly and has remained nearly sta1:ic over the decade. Emigration, which has reduced net population growth to around 1.98% yearly, is mre likely to decline than increase. Prospects for future fertility reduction apart from increased contracept:Lon are dim: the average age of marriage and educational levels are already high, the duration of breast-feeding is declining, and an increasing proportion of women will continue to enter the reproductive years for at least the next two decades. Parental dependence on children for financial support in old age remains a significant constraint to lowered fertility. 6. 02 Family planning performance is still well below its potential. as indicated "by excess fertility (defined in terms of women having 100re children than desired), representing considerable unmet demand for family planning. The annual number of new acceptors has fluctuated widely over the last t,�n years. Contraceptive prevalence (55%) is relatively high, but is dominat,�d by sterilizations, mainly of couples who already have m:>re children than they desire. The use of modern spacing methods and participat.lon by younger and high-parity women are low. Around half of all contracept:ion is by traditional spacing, whose impact on fertility is probably l,�ss than m:>dern methods. 6.03 Nevertheless, a basic framework for m:>re effective management of population growth is in place. The Government has made strong statements in support of family planning, population as an issue has been largely depolitici.zed, a reasonable data base has been developed, and a relatively widespread delivery system involving both public and private initiatives is in operation. Moreover, knowledge of contraception is almost universal. 6.04 However, the program suffers from several major constraints. Public health midwives (now family health workers) do not carry out family planning a,ctivities systematically or with sufficient commitment. Given high existlng financial incentives, physical and manpower resources are inadequate to meet the demand for sterilizations, while those for IUD insertions are underutilized. Pills account for an unusually low proportion of the contraceptive mix in spite of the formal existence of some 1,200 distribution points. Despite their relatively low efficacy compared ti:> properly applied modern methods, traditional spacing methods remain popular with younger, low-parity women. Moreover, operational coordinati,on remains weak among the FHB, the MPI, and the FPASL, the country's .largest voluntary organization in the population field. A number of specific improvements need to be made to revitalize the population program.  -46- 6.05 Program Management and Coordination. From an administrative perspective, it would be desirable to give one official. family planning organization the full authority for comprehensive national programming, from planning through service delivery and performance evaluation. The FHB would be the roost logical lead agency to guide such activities through both the government health system and increasingly important private (NGO and commercial) channels. An optional approach would be to demarcate clearer responsibilities for the MPI, FHB and FPASL to improve their·coordination under an overall program framework which the Government would develop. The Population Steering Committee should be revitalized and streamlined into a viable and ongoing working group of the principal agencies involved in family planning, ensuring that each concentrates on areas of comparative advantage. The above measures should result in the development of a time-phased family planning work program which would delineate the role of each agency. The program should take into ac count explicit demographic goals, which in turn would be translated into measures to monitor and evaluate performance, including specific targets by method and geographical area, to increase both the overall CPR and also the use of modern spacing methods. In aggregate, according to Bank estimates, it would be necessary to raise the present CPR from SSl: to about 67% by 1990, and to about 76% by the year 2000, in order to obtain the rapid fertility dec line population projection shown in Table 1. Even to achieve the moderate fertility decline population projection, the CPR would have- to increase to about 61% by 1990 and to about 71% by the year 2000. 6.06 Family planning operations should be further decentralized to ensure that the local program mix reflects the characteristics and needs of particular communities. The emerging district population committees need discretionary financing and authority to test innovative ways of adapting overall program design to local conditions through the development of family planning strategies for specific communities. 6.07 Program Strategy and Service Mix. Alternative approaches to persuade younger, low parity women to space and limit births are needed, as -most acceptors of sterilization a.re high parity and have never used another family planning method. Studies need to be carried out to define more precisely what services they want and how delivery systems can best respond. One prospective strategy would be first to move women from traditional to roore efficacious modern methods, and then to sterilization when they have reached appropriate family size. Access to modern reversible methods needs to be expanded, ac companied by better follow-up particularly for management of side-effects. Pills and condoms also need to be distributed unre widely. Both peripheral health workers and private outlets should be authorized to issue initial and follow-up supplies of pills without prior reference to or endorsement by physicians. The availability of injectables should be increased through increased supplies and additional outlets. If the popularity of IUDs failed to increase, low-performing clinics should emphasize other methods. Facilities for outpatient, as opposed to postpartum, sterilization also need to be expanded.  -47- 6.08 The Government's role in the above strategy would be to concentratE! on those methods which depend largely on the health system --principally sterilizations, IUDs, and injectables--and management of -side effec:t:s. The development of community-based approaches would also be essential t:o increase the coverage of-other 100dern spacing methods, such as pills and condoms, among the large and rapidly expanding number of young fertile couples. This is an area in which, at least for the short term, the public health system appears lE!ast able to respond because of personnel shortages at the periphery, and the new expanded role of the FHW. Given the urgency of reaching un-served and under-served population groups, high priority should be assigned to exploring the capacity, and directly supporting the expansion, of the private sector in prescribing and delivering non-clinical contraceptives, most notably through the promising commercial marketing scheme and USE! of ayurvedics. Strengthening such potentially self-supporting activiti.es is essential since externally-supplied "free" contraceptives will be phased out by 1986 in favor of full government financial responsibility for them. 6.09 Demand Generation and Incentives. Raising the CPR above the present level is difficult because most of the strongest demand for family planning already has been met. Although there is still substantial unmet demand, generating additional demand for family planning will be critical. The danger of equating desired births for family planning with actual fertility behavior is underscored by data from a 1979 subsample survey of WFS respondents which show that two out of five women not wanting another child had had at least one more birth since the 1975 survey. Receptivity of the population to incentives fo:r sterilization, as might be expected given the low per capita income, suggests that incentives for adoption of other methods may be warranted. The likely cost-effectiveness of different incentive levels for sterilization also should be explored. Additionally, the Government which has already taken key "beyond family planning" steps, may want to consider other population-oriented policy measures, effectively instituted elsewhere, including b�llding family size criteria into the selection process for highly valued social benefits such as education and housing. B. HEALTH 6.10 There are three major issues whose resolution would facilitate progress towards a health system responsive to the country's health problems. These issues concern: (i) the management and efficiency of the existing public health services, and ways to improve them despite existing resource c.onstraints; (ii) the design of the new PRC program, and its anticipated effectiveness in addressing the underlying health problems and priorities; and (iii) the question of cost recovery in view of the need to raise additional operating funds for the health sector, to cut down on excessive use of publicly subsidized services, and to improve the equity of health service use. 6.11 Management and Efficiency of Existing Health Services. The existing health services operate inefficiently in a number of important respects. In the first place, there is a lack of a comprehensive national health plan and strategy, due to weak planning capacity and poor  -48- coordination between ministries. While the MH is competent in the technical. aspects of health, it lacks sufficient expertise in management and financial analysis. There is al.so poor coordination between the planning and accounting sections of the ministry. The Ministry of Finance and Planning, on the other hand, is strong in financial analysis but weak on the health technical. side. The absence of sufficiently coordinated planning, together with weak financial controls, has led to a situation where resources have been al.lowed to be drawn into the tertiary sector at the expense of the periphery. Partly in response to this situation, the Planning Unit of the MH has concentrated almost exclusively on developing the new PHC model; but the exclusion of planning for the tertiary sector has led to considerable frustration among hospital. staff. There is, therefore, an urgent need to broaden planning activities to encompass the entire range of health services. 6.12 Managerial weaknesses affect both headquarters, and field and hospital operations. For example, the distribution of resources and responsibilities between the MH and the Ministry of Women's Affairs and Teaching Hospitals may be expected to cause problems in the al.location of staff and budgets, in the coordination of planning and program implementation, and in the provision of technical support to lower levels of the health system. The operational management of the health services is also hampered by functional problems. For example, the superintendents of health services have responsibili.ty for public health, patient care institutions, and technical and administrative services, and yet they do not have commensurate authority to take action, particularly in regard t� personnel and financial matters. Health facilities are infrequently monitored and do not receive systematic feedback based on the reports which they are required to file. 6.13 , The existing pattern of utilization of health services is al.so inefficient: peripheral. facilities tend to be bypassed; secondary facilities are underutilized, and tertiary facilities are overcrowded. One of the main reasons for this is the relative deterioration of services in the periphery due to overall resource constraints and the expansion of the higher level infrastructure with heavy demands on the recurrent health budget. Despite the stagnation in real per capita recurrent health expenditures by the Government between 1970/71 and 1982, a total of more than Rs. 800 million was spent on the capital budget in this period, mainly on the expansion of the tertiary level health infrastructure (for example, on Peradeniya and Galle Hospitals, and roore recently on the new Kotte Hospital). The result has been a lack of funds for such items as vehicle operation (both fuel and travelling allowances), and equipment maintenance, which needs to be corrected. An important implication of this is that additional recurrent funds available to the public health services must be used only for priority items, both for existing services and for any new projects. There would seem to be little justification for going ahead with plans to expand bed capacity at the Colombo General Hospital. when the additional recurrent costs (in 1983 prices) might be in the order of Rs. 20 million, especially with the opening of the new Kotte Hospital. Moreover, the emphasis at leading facilities on high technology/curative medicine for cardio-vascular diseases, accidents and other modern health problems is costly in comparison to the benefits from an approach which  -49- stresses prevention. Along with the need to re-examine priorities and improve the planning and control of the use of recurrent financial resources, there is also a need to review the p�esent distribution of health manpower and training programs, especially for doctors, nurses and FHWs. The MH needs to establish a manpower planning and training unit for this purpose. 6.14 New Primary Health Care Program . There are also some fundamental issues relating to the new PHC program. The proposed integration of preventive and curative services is a significant departure from the existing structure and may result 1.n the continued dominance of curative activities., This danger is exacerbated by the large number of tasks planned for FHWs, which could result amongst other things in neglect of family planning. It is therefore i.mportant for. the MH to review the proposed rc>le of these workers, to prioritize key tasks clearly within the overall context of the PHC program, and to plan a continuous in-service training program. 6.15 Effective implementation of the PHC program, which is now being tested through a project with Asian Development Bank support, will depend critically on appropriate supervision and management of field staff. A substantial. number of FHWs were initially deployed without any correspondi.ng increase in supervision and other support. This is an· extra reason for strengthening the managerial. skills of mid-level health personnel. However, measures are required to encourage aoctors to study comm.unity medicine and accept MH positions, particularly in underserved areas. A review of the curriculum and training in community medicine is already underway, but new incentives may also need to ·be considered to counterbalance the attractions of clinical medicine and especially private practice. 6.16 E'lexibility is needed to adapt the design and implementation of the PHC program to differing population densities, local conditions, and existing in.frastructure. In particular, priority should be given to implementing the program first in those districts which have relatively poor health indicators, and especially an IMR above the national average. It may also be possible to improve the cost-effectiveness of the program by reducing the unit cost estimates for the proposed facilities and by reducing the number of additional inpatient beds planned. For example, the design of staff quarters could be based on the assumption that the investment cost should be no mre than three times the staff member's annual salary (which is a typical criterion for housing standards elsewhere). Even if staff salaries are assumed to be twice their present level, because salaries in Sri Lanka are low by international. standards, the total capital cost of housing in the proposed project would then be more than halved from the present estimate (falling from Rs. 950 million to Rs. 416 million). Second, if plans for service buildings were scaled down and no additional beds were to be added at division al health centers, an estimated Rs. 89 million could be saved. These two measures would also reduce the estimated annual additional recurrent costs from about Rs. 163 million to about Rs. 100 million. While these particular curtailments may or may not be appropriate, the exam:ples suggest that the cost-effectiveness of the existing proposals should be carefully examined.  -so- 6.17 Cost Recovery. The Government should further consider substantially increased cost recovery for public health services. At present these are entirely free, with the exception of a small number of private pay-beds mainly at Colombo General Hospital. From 1973 to 1977 a nominal charge (Rs. 0.25) was levied on outpatients, but this proved to be politically unacceptable and was discontinued. The strain of limited resources has prompted the Government to consider patient fees once again. Levying such charges can: (i) rai.se additional revenue to cover costs; (ii) improve the efficiency of resourc:e allocation; and (iii) improve the equity of health service use. 6.18 These simple criteria could be used to design an appropriate pricing policy for Sri Lanka. Fi.rst, there are two types of heal th services which provide "positive externalities" or social benefits: preventive services in general (which reduce the later need for mre costly curative care), and curative services for communicable diseases. There are strong grounds, therefore, for continuing to subsidize or provide at no charge such activities as routine maternal and child health care, health education activities, and treatment for diseases such as malaria and sexually transmitted diseases. In contrast, however, there are grounds for charging for other types of outpatient, curative care. Such charges can curtail "excessive" use of services or duplicative consultations. If there is a fear that charges may deter the poor from legitimate use of health services, then it may be possible to waive fees for indigents. It may also be appropriate to introduce an additional charge for patients who make direct use of tertiary level ·services, but a prerequisite for this should be improvements in primary and secondary level services to reduce the present incentive to bypass them� Regarding inpatient and outpatient referrals, the quantity of services consumed by patients should at least in theory be determined by the health staff, so charges should not lead to better resource allocation. In practice, however, some patients may try to influence the health staff to admit them into hospital, and so a relatively small charge could have a deterrent effect. Inpatient services are the most costly for the Government to provide, and so some degree of cost • recovery for these services may al.so be deemed necessary. 6.19 There are two particular examples where health charges should clearly be raised or introduced on grounds of equity. First, the Government's charges for its pay-beds are below the actual cost of providing the service. The Government should, therefore, raise its charges for pay-beds at least to the cost of providing the service (and perhaps higher if it wants to cross-subs:Ldize services). This might entail doubling the existing charges. '.l'he second example relates to the new hospital at Kotte, for which the Government should make every possible effort to minimize its net subsidy through both expenditure control and cost recovery. The hospital has not been designed as a tertiary level facility, but rather as a sophisticated secondary level facility. Even if 100% cost recovery is impossible, there would seem to be no justification for the Government to subsidize the hospital m:>re than the existing Colombo General Hospital which is the country's top referral facility. The latter has 2,500 beds and its budget was about Rs. 98 million in 1983. To provide an equal subsidy per bed for the Kotte Hospital with 1,000 beds, the Government should, therefore, provide recurrent cost support of no m:>re  -51- than about Rs. 40 million per annum, and the balance (estimated at about Rs. 60 million) should be derived from patient fees. If the hospital were to have an 80% occupancy rate, this would imply the need to charge about Rs. 200 per inpatient day for every bed. 6.20 In conclusion, it is recommended that the Government further explore the possibilities for greater cost recovery, especially to improve resource allocation in the health sector. The design of an appropriate fee system should be done in conjunction with other efforts to improve the existing h,ealth services, especially at the peripheral and secondary level. Thie Government should also substantially raise its charges for private pay-beds, and should institute maximum cost recovery from the new Kotte Hosp:ltal.