32815 Africa Region Human Development Working Paper Series Decentralized Delivery of Primary Health Services in Nigeria Survey Evidence from the States of Lagos and Kogi Monica Das Gupta, Varun Gauri Stuti Khemani September 24, 2003 Development Research Group The World Bank 32815 Copyright © June 2004 Human Development Sector Africa Region The World Bank The findings, interpretations, and conclusions expressed in this re- port are entirely those of the author and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. Typography by The Word Express Cover design by the Word Express Cover photo: World Bank photolibrary ii Contents Foreword ......................................................................................................................................v Abstract ......................................................................................................................................vii Executive Summary ..................................................................................................................... 1 1. Introduction ..................................................................................................................... 7 2. Survey Results ................................................................................................................ 13 3. Emerging Issues.............................................................................................................. 63 4. Main Conclusions and Policy Lessons ............................................................................ 73 Bibliography .............................................................................................................................. 77 Boxes Box 1: Survey Instruments at the Health Facility Level ..................................................... 10 Box 2: Survey Instrument at the Local Government Level ................................................ 11 Tables Table 2.1.1: Ownership of Facilities.......................................................................................... 13 Table 2.1.2: Number of Facilities by State and Facility Type..................................................... 14 Table 2.1.3: Condition of Facilities by Type of Facility ............................................................. 15 Table 2.1.4: Percent Distribution of Main Source of Water by Facility Type ............................. 17 Table 2.1.5: Percent of Each Type of Facility with Various Amenities ....................................... 18 Table 2.1.6: How Vaccines are Stored....................................................................................... 19 Table 2.1.7: How Equipment Is Usually Sterilized .................................................................... 21 Table 2.1.8: Average Distance from LGA and Other Health Facilities by Facility Type .................................................................................................. 22 Table 2.2.1: Principal Decisionmakers for Facility Functioning................................................. 26 Table 2.2.2: Activities of Primary Health Care Management Committee .................................. 29 Table 2.2.3: Community Participation in Kogi and Lagos......................................................... 30 Table 2.2.4: Community Participation in Kogi across Facility Types......................................... 31 Table 2.3.1: Main Supplier of Facility Resources ...................................................................... 32 Table 2.3.2: Per Capita LGA Revenues ..................................................................................... 34 Table 2.3.3: Local Government Health Expenditure ................................................................. 37 Table 2.4.1: Number of Health Workers by Facility Type ......................................................... 39 iii Table 2.4.2: Designation of Health Staff in the Sample by State............................................. 40 Table 2.4.3: Personal Characteristics of Staff ......................................................................... 40 Table 2.4.4: Percent of Staff that Supplement Salary .............................................................. 41 Table 2.4.5: Household Condition of Staff ............................................................................ 41 Table 2.4.6: Monthly Salary by Designation and Facility Type .............................................. 42 Table 2.4.7: Percent of Staff Receiving In-kind Benefits by Designation ..................................... Table 2.4.8: Determinants of Monthly Wages ........................................................................ 43 Table 2.4.9: Average Number of Years Working in Current Health Facility ........................... 43 Table 2.4.10: Total Days Spent in Training in Previous Year by Designation and Facility Type ............................................................................................. 45 Table II.4.11: Total days spent in training in last year by designation and facility type.................................................................................................. 46 Table 2.4.12: Professional Attitudes......................................................................................... 46 Table 2.4.13: Availability of Drugs, Equipment, and Surveillance Records .............................. 47 Table 2.5.1: Percent of Facilities Providing Specific Services by Facility Type ............................................................................................... 49 Table 2.5.2: Average Number of Outputs by Facility Type between March and May 2002 ..................................................................................... 51 Table 2.5.3: Facility-Level Average Output per Staff in Categories 1­7 Facilities between March and May 2002 .................................................. 52 Table 2.5.4: Tracer and Immediately Notifiable Diseases, Percentage of Facilities by Facility Type ............................................................................ 53 Table 2.5.5: Percent of Staff Performing Various Duties in One Week by Category of Staff ........................................................................................ 54 Table 2.5.6: Number of Days Worked in One Week by Category of Staff .............................. 56 Table 2.5.7: Patients Seen Outside Facility per Week by Category of Staff ............................. 56 Table 2.5.8: Percent of Staff with Various Attributes by Category of Staff ............................. 57 Table 2.5.9: Number and Types of Sanitary Inspections Conducted in the Local Government Authorities between March and May 2002 ....................... 59 Table 2.5.10: Immunization during Special Drives and on Routine Basis ................................. 61 Table 3.1.1: Impact of Community Participation on Facility Productivity.............................. 65 Table 3.1.2: Impact of Community Participation on General Facility Characteristics .................................................................................... 66 Table 3.1.3: Impact of Community Participation on Availability of Essential Drugs ........................................................................................... 67 Table 3.2.1: Non-payment of Staff Salaries in Selected Local Government Authorities in Kogi .......................................................................................... 69 Table 3.2.2: Impact of Non-payment of Staff Salaries on Facility Performance ...................... 71 Figures Figure 2.3.1: Composition of Kogi State Revenues.................................................................. 35 Figure 2.3.2: Composition of Lagos State Revenues ................................................................ 36 Figure 2.3.3: Composition of Health Expenditures for Kogi, 2000 ......................................... 37 Figure 2.3.4: Composition of Health Expenditures for Lagos, 2000 ....................................... 37 Figure 3.2.1: Non-payment of Staff Salaries in Kogi and Lagos .............................................. 69 iv Foreword N igeria is one of the few countries in health service delivery depends critically on ex- the developing world that has system- isting institutions and incentives in the public sec- atically decentralized the delivery of tor. In recent years, public revenues in Nigeria basic services in health and education have increased substantially due to the boom in to locally elected governments and community- world oil prices, and some of this windfall is be- based organizations. This study uses an exten- ing channeled into increased spending on primary sive survey of primary health facilities and local health care. Yet, there remains a concern whether governments in the states of Lagos and Kogi to the institutions of public accountability in the analyze how local institutions function in prac- country will effectively allow these large spend- tice in delivering basic health services, and to ing programs to translate into improved services draw lessons for improving public accountabil- and outcomes. A major channel through which ity. The newly developed survey methodology increased public resources are expected to impact employed in the study enables rich analysis of basic health and education services in Nigeria is outcomes in primary health service delivery at that of spending by local governments that are the front lines, in terms of the range of services largely responsible for these services. It is there- provided, facility infrastructure, availability of fore important to delve deeper into the role of essential supplies and equipment, staffing pat- local governments and community organizations terns and provider incentives. Facility-level data in basic health service delivery. on service delivery outcomes has been linked to The relative roles of the three tiers of govern- data collected from local governments on the ment--federal, state, and the local government governance environment and financing arrange- authorities (LGAs)--in public service delivery has ments to shed light on the political economy of emerged as one of the most important topics of decentralized service delivery. open and vigorous debate in the new democratic It is increasingly being recognized that simply climate in Nigeria. There have been increasing allocating greater public resources to basic health calls for intergovernmental fiscal relations to be services is not enough to ensure that quality ser- reassessed in light of a widespread belief that al- vices are made available to the vast majority of though the states and local government authori- poor citizens in the developing world. The im- ties are assigned primary responsibility for the pact of public spending on actual outcomes in delivery of basic public services, they are not v equipped with adequate revenue resources to ful- primary health care services. Recently there have fill their expenditure obligations because the bulk been several governmental initiatives to strengthen of government revenues is retained by the fed- these institutions of community participation to eral government. This has culminated in an improve health services. amendment to the revenue-sharing formula in the This detailed survey evidence from frontline Federation which increases resources available to service delivery agencies and local governments the decentralized levels of governments. provides valuable knowledge on how the de- In addition to the constitutional provisions for centralized system of primary health service de- decentralization to locally elected governments, livery actually works in Nigeria. The findings health policy in Nigeria has been guided by the will enrich the policy dialogue within the coun- Bamako initiative of encouraging and sustaining try and beyond about innovative institutional community participation in primary health care solutions to improve public accountability. servicedelivery.Communityparticipationhasbeen institutionalized through the creation of Village Ok Pannenborg Development Committees and District Develop- Senior Advisor for Health, Nutrition, and ment Committees that are grass-roots organiza- Population tions expected to work closely with local Human Development Department governments in monitoring and supporting Africa Region vi Abstract T his report presents findings from a sur- services provided in primary health care fa- vey of 252 primary health facilities and cilities. However, the service delivery environ- 30 local governments carried out in the ments between the two states are strikingly states of Kogi and Lagos in Nigeria in different. In largely urban Lagos, public de- the latter part of 2002. Nigeria is one of the few livery by local governments is influenced by countries in the developing world to systemati- the availability of private facilities and prox- cally decentralize the delivery of basic health and imity to referral centers in the state. In largely education services to locally elected govern- rural Kogi, primary health services are pre- ments. Its health policy has also been guided by dominantly provided in public facilities, but the Bamako Initiative to encourage and sustain with extensive community participation in the community participation in primary health care maintenance of service delivery. The survey services. The survey data provide systematic evi- identified an issue which is highly relevant for dence on how these institutions of decentraliza- decentralization policies--the non-payment of tion are functioning at the level local--govern- health staff salaries in Kogi--which is sugges- ments and community based organizations--to tive of problems with local accountability deliver primary health service. when local governments are heavily depen- The evidence shows that locally elected gov- dent on fiscal transfers from higher tiers of ernments indeed do assume responsibility for government. vii viii Executive Summary Motivation and Objectives In addition to its analytical objectives, the conduct of this study was specifically designed T his study analyzes decentralized deliv- to promote evidence-based policy dialogue in ery of primary health services in two Nigeria by engaging the active participation of states in Nigeria, Lagos and Kogi, to un- the overarching government agency in the coun- derstand how existing institutional ar- try responsible for monitoring and supervising rangements work in practice and how they im- outcomes in primary health care service deliv- pact service delivery outcomes. Nigeria is one of ery--the National Primary Health Care Devel- the few countries in the developing world to have opment Agency (NPHCDA). The terms of systematically decentralized the delivery of basic reference for this study were developed in part- health and education services to locally elected gov- nership with NPHCDA, with the agency closely ernments. In addition, it's health policy has been involved at every stage right from study design guided by the Bamako initiative to encourage and to its implementation and subsequent analysis. sustain community participation in primary health care services. The study therefore focuses on the Methodology role of local governments and community based organizations in the delivery of primary health care The methodology adopted to address the objec- services. The outcomes of interest are measured tives of study is based on extensive and rigorous as performance of public health facilities, in terms survey work, at the level of frontline public ser- of actual service delivery outputs at the level of vice delivery agencies--the primary health care frontline delivery agencies--services provided, fa- facilities--and the local governments. Three ba- cility infrastructure, availability of essential sup- sic survey instruments of primary data collection plies and equipment, staffing patterns and pro- were used--one, administered to public officials vider incentives. Although these are not the ulti- at the local government level to collect informa- mate outcomes we care about, such as improve- tion on the governance environment and public ment in household health indicators, focusing on financing patterns; second, administered to the them may nevertheless enhance our understand- facility manager for general facility characteris- ing of what public budgets "buy" in terms of tics and services provided, including direct data intermediate service delivery outcomes. collection from facility records; and third, 1 2 Decentralized Delivery of Primary Health Services in Nigeria administered to individual staff at the facility level LGA as the principal decision-maker for most for data on working environment and incentives. service provision decisions at the facility level, as The survey was undertaken during June-August compared to the other two tiers of government. 2002, with data collected in 30 local govern- The state and federal governments were indicated ments, 252 health facilities, and from over 700 very infrequently as principal decision-makers for health workers, in Lagos and Kogi states. any area, and even then for only one or two areas Facility-level data on service delivery outcomes of decision-making in any individual facility. This was linked to data collected from local govern- evidence for the health sector is a striking con- ments on the governance environment and financ- trast to available evidence for service delivery in ing arrangements. This micro-level survey ap- other sectors--such as primary education, water proach has allowed a deeper investigation of ac- and sanitation--that are characterized by consid- tual outcomes in public service delivery at the erable overlap and confusion with regard to the frontline, and the impact on these outcomes of sharing of responsibilities between the three tiers, broader institutions of governance and financ- often at the expense of undermining LGA ing arrangements, than more aggregative tools responsibility and accountability. of public expenditure analysis. Community participation in primary health care service delivery has been institutionalized in Nigeria through the creation of Village De- Governance Environment and velopment Committees and District Develop- Financing Arrangements ment Committees. There are striking differ- ences in the sharing of responsibilities between A strict interpretation of the Constitution of the LGA and community development com- Nigeria with regard to the sharing of responsi- mittees in the two states of Lagos and Kogi bilities between the three tiers of government studied here. In Lagos, more than 80% of fa- implies that it is the state governments that have cility-level respondents indicated the LGA as principal responsibility for basic services such principal decision-maker in most areas of ser- as primary health and primary education, with vice delivery at the facility level, while in Kogi, the extent of participation of Local Government only about 50% indicated the LGA as princi- Authorities (LGAs) in the execution of these re- pal decision-maker. The remaining facilities in sponsibilities determined at the discretion of in- Kogi listed either the community development dividual state governments. The constitutional committees or the facility head or both as the existence of state-level discretion may lead to principal decision-makers. Community orga- disparities across local governments or across nizations are particularly active in Kogi in the states in the extent to which responsibility for areas of building maintenance, and acquiring primary health services is effectively decentral- drugs, medical supplies, and equipment for the ized. In the face of such constitutional ambigu- facilities. There is comparatively little commu- ity, the survey of LGAs and health facilities at- nity engagement in setting charges for drugs, tempted to assess the actual extent of decentrali- as was envisioned by the Bamako Initiative zation of services to local governments. and almost negligible in disciplining staff, The overwhelming majority of LGA respon- which is overwhelmingly indicated as the dents indicated the LGA as the principal decision- responsibility of local governments. maker for most of the areas of facility-level pro- Amongst government agencies the LGA is the vision of primary health services. There was no main source of financing of primary health ser- systematic variation across local governments in vice delivery at the facility level. Staff salaries, the extent of decentralization of responsibility. The facility building construction and maintenance, facility-level respondents similarly indicated the supply of drugs, equipment and other medical Executive Summary 3 commodities, are all predominantly provided by much higher proportion of public facilities are local governments in Lagos state. However, in of higher level, whereas in Kogi 80% of facili- Kogi, community-based organizations and fa- ties are health posts. Moreover, Lagos facilities cility staff are frequently indicated by facility are proximate to a much higher density of refer- respondents as the main source of drugs (for ral centers and private facilities than those in 28% of facility respondents), medical supplies Kogi, and are also much better provided with (31%), and building maintenance (57%). It is public amenities such as water and electricity. surprising to note that as many as 15% of fa- The data indicate that Kogi facilities succeed in cilities in Kogi indicate staff personal funds as functioning under very difficult circumstances the main source of facility resources, which if in terms of lack of basic amenities, and main- accurate probably implies that staff compensate tain public facilities better than those in Lagos, themselves from facility revenues. In Lagos, for despite their better endowments. A substantial the majority of facilities (over 85%) resources proportion of facilities in both states were in were either provided by the LGA or indicated poor repair. as not provided at all in the last twelve months. Giventherelativeshortageofalternativesources Staff salaries are almost exclusively provided by of care, Kogi health posts necessarily meet a much local governments in both states. wider range of the health care needs of the popu- Local governments in Kogi are overwhelm- lation they serve. For example, Kogi health posts ingly dependent on statutory allocations from provide a full range of services including antena- the Federation Account for their revenues, and tal and postnatal care, deliveries, and in-patient receive almost nothing from the state govern- malaria treatment, while those in Lagos concen- ment. Revenue sources of local governments in trate mostly on outpatient consultations (for chil- Lagos are more diversified--bulk of their rev- dren and adults) and immunizations. enues comes from two sources, the Federation The services provided in different types of Account and the VAT, but a significant amount facilities show a pattern consistent with the rela- is also internally generated from local tax bases. tive advantages of lower-level facilities in terms This is as one would expect given that Lagos of proximity to their patients. For example, the state is the urban center of Nigeria, while Kogi average number of home visits per staff declines, is a largely rural state. The consequences for the higher the type of facility. Only 30% of PHCs basic health service delivery between the two compared with 64% of health posts/dispensa- states is therefore clear--services in Kogi are ries conduct in-patient deliveries, and similar more vulnerable to external shocks that affect figures prevail for in-patient malaria treatment. oil prices, which is why, perhaps, communities One possible reason for this might be that staff in Kogi take a more active role in maintaining do not stay overnight in these facilities, unlike basic health services. Bulk of LGA health ex- health posts where staff reside on the premises penditures are allocated to staff salaries--in Kogi or very nearby. in 2000, LGAs on average spent 78% of health There is an impressive range of sanitary in- expenditures on salaries, while in Lagos, LGAs spections conducted in Nigeria. 70% of LGAs spent 65% on average on staff salaries. were reported to have undertaken food vendor certification in the past year, and all conducted most of the prescribed forms of sanitary inspec- Facility Characteristics and tion: of public water sources, of markets, house- Services Provided to house inspections for public health nuisances, and inspection of food sellers. Public health care facilities in Lagos and Kogi Immunization is provided through the function in quite different contexts. In Lagos, a regular health services on a routine basis, as 4 Decentralized Delivery of Primary Health Services in Nigeria well as through the National Immunization likely to have pharmaceutical products, such Drive. It appears that, despite the high-profile as chloroquine, paracetamol, antiobiotics, ORS pressure of the National Immunization Drive, sachets, and multivitamins. A likely explana- that the routine immunization effort may be tion for this is that whereas in Lagos alterna- more effective. For example, 37% of the LGAs tive suppliers are available, such as pharma- sampled did not carry out polio (NID) immuni- cies, in Kogi the public clinics effectively func- zation during the preceding three months. tion as pharmacies in which health staff sell Facilities for storing vaccines are poor, especially privately acquired products. It is not clear in Kogi. whether this health staff are responding to Although the majority of public health facili- shortages in public supply, or whether facility ties were observed to be clean and functioning owned products are being expropriated. In and providing a range of health services, there is Lagos, the public-private ownership correla- some suggestion of poor quality of services for tions for these products are also negative but some of the conditions that are reported as the much smaller. Vaccines were far more likely to main causes of mortality and morbidity among be available in Lagos facilities. That might sug- women and children, namely malaria, diarrhea, gest better public provision in Lagos but might and vaccine preventable diseases. For example, also be an artifact of differing delivery although malarial drugs were available in more schedules in the two states. than 60% of the facilities surveyed, there was no equipment nor expertise for testing for malaria in more than 90% of the facilities, thereby im- Staffing Patterns plying that drugs are administered on the basis of symptoms alone. This may be the best strat- The average health facility in the sample had egy available given the constraints, but is not 7.85 health workers, but the average for health optimal for effectively controlling and treating posts was 2.3 workers. Health facility types were the disease. unevenly located across the two states in the Strengthening of policies on preventive health sample: 93% of health posts in the sample were care is urgent in light of evidence that public located in Kogi state while 75% of the remain- health surveillance may be particularly poor in ing higher level facilities were located in Lagos. rural states--in Kogi, only 38% of facilities were As a result, while 61% of all facilities in the able to show records of tracer and immediately sample were in Kogi, 66% of the staff were from notifiable diseases to the survey interviewer, Lagos. Kogi had a mean of 4.0 staff per facility; compared to 94% of facilities in Lagos that pro- in Lagos there was a mean of 13.7 primary duced these records. health care staff per facility. Staff in Lagos had more clinical training. For example, while nurses make up about 10% of total staff in Kogi, nurses Availability of Essential Drugs constituted 20% of all staff in Lagos. Similarly, and Equipment 7% of Kogi staff were midwives, compared to 26% in Lagos. Many health facilities reported shortages of ba- The average age of staff was 41 years, but sic health equipment. For instance, 95% did doctors were younger than the rest of the cadre, not have microscopes, 59% did not have ster- with an average age of 30 years. A large major- ile gloves, 98% did not have a malaria smear, ity of health staff were women, with exceptions and 95% did not have a urine test strip. Lagos again being doctors (50%) and environmental facilities were six times more likely to have a health officers (21%). The large majority of staff generator, but Kogi facilities were much more in almost all categories had some amount of Executive Summary 5 post-secondary education. Only about 28% of Despite the fact that local conditions were staff were indigenous to the communities in not significantly affecting salaries, there was a which they are working, with percent indigene substantial amount of churning among health ranging from 0% for doctors to 41% for nurses. staff. The average length of tenure in the cur- Staff had on average 14 years of experience in rent facility was short, about 2.7 years. Medi- primary health care, but doctors had relatively cal officers had been working in the current fa- less experience, with an average of 2.6 years of cility for three months (given their age, most work in the field. Almost all staff (96%) were were probably just out of medical school), and employed by the LGA, though half of the 10 nurses and midwives had an average tenure in medical officers in the sample were employed the current facility of less than two years. Se- by the federal government and half by the LGA. nior and junior health education workers had Medical officers rarely worked in public facili- longer tenures (most were in Kogi, where staff ties ­ only one in four type 3 facilities had a typically had a longer average time in the physician on staff, and the sole type 4 facility in current facility than Lagos). the sample did not have any. The data on work done by different catego- About 45% of staff were supplementing their ries of staff indicate that most tasks are done by income in some way. The most common sources all the grades of staff. Nurse-midwives are the of supplemental income were agricultural work work-horses, a much higher percent of them than and commerce. About 17% of staff reported other staff do deliveries, immunizations, antena- supplementing their work with some form of tal care, and family planning. High percentages health-related activities, including clinical work, also do out-patient care and health education. home health care, or the sale of medicines. In-kind Substantial proportions of the staff, including benefits typically did not constitute a large element those in the non-professional grades, report see- of an average staff member's reimbursement pack- ing patients privately outside the facility -- and age: 63% of health staff did not receive any in- it can be assumed that such data is subject to kind benefits at all. For those who did receive in- under-reporting. Doctors do the most "moon- kind benefits, the most common benefit was free lighting": 90% of doctors compared with 50- health care, which 21% of staff received. 60% of other categories of staff report seeing The monthly salaries of health staff were, on patients outside the facility. average, 26306 Naira (about US$220), in 2001. The highest paid staff were midwives, CHOS, and nurses. Doctors, surprisingly, were the low- Impact of Governance and Financing est paid. The reason for the low pay of doctors Environment on Service Delivery is likely related to the fact that doctors were on average more than ten years younger than their A striking feature of public delivery of primary colleagues in other designations. An estimation health services in Nigeria was revealed through of staff salaries using a standard Mincerian earn- the survey--public resources, in fact, do not ings function found that, controlling for gen- appear to be reaching their intended destina- der, experience, education, state, local compe- tions. There is evidence of large scale leakage in tition, and unobserved LGA characteristics, civil public resources in Kogi, away from original service pay scale explained the largest fraction budget allocations. Although staff salaries ac- of health worker salaries. In other words, tra- count for 78% of health expenditures and 20% ditional civil service pay scales, rather than lo- of total LGA revenues, on average, the survey cally determined rewards for performance, re- of facility staff in Kogi revealed that 42% of mained the dominant element in the incentive them had not been paid their salaries for more environment for primary health care staff. than 6 months in the past year. Using the survey 6 Decentralized Delivery of Primary Health Services in Nigeria data, we estimated and compared actual staff suggest is that of strengthening local government costs per facility in each LGA with what the LGA accountability. We propose one major channel reported as budget allocations towards staff sala- for this purpose--providing citizens with greater ries per facility within its jurisdiction, and found information about the resources and responsi- that even when budget allocations were suffi- bilities of their local representatives, so they are cient to cover estimated actual costs, the staff empowered to hold them accountable for the survey showed non-payment of salaries for sev- delivery of basic services. eral months in the year before the survey. There We undertook some analysis of the impact of is, in fact, no significant correlation between community participation in Kogi on various local government revenues and resources bud- performance indicators at the facility level. The geted towards staff salaries with the non-pay- most striking result is that community partici- ment of salaries. Hence, the non-payment of pation is significantly associated with greater salaries cannot be explained by lack of resources productivity per staff in providing inpatient de- available to local governments. liveries, immunizations, and outpatient consul- The analysis also showed that the greater is tation. While an appealing interpretation of this the extent of non-payment of salaries, the higher association may be that greater community par- is the likelihood that facility staff in fact behave ticipation makes facility staff more responsive as private providers--with more services pro- to the health needs of the community they serve, vided outside the facility through home visits, there are alternative interpretations, and the and with essential drugs being privately pro- analysis undertaken here is too limited to draw vided, either funded by staff own resources or strong conclusions about the causal impact of expropriated from facility stocks. community participation on service delivery. This evidence suggests that there is a general There is also a significant negative correla- problem of accountability at the local govern- tion of community participation in facilities with ment level in the use of public resources that are record-keeping at the facility level for public transferred from higher tiers of government and health surveillance. A causal interpretation of about which, therefore, local citizens may not this would suggest that with more decentral- be well informed since they are not the tax-pay- ized management and monitoring of facilities ers. In and of itself, this analysis does not sug- by the immediate communities they service, gest that the counterfactual would be true--that some facility activities with beneficial spillovers is, more centralized delivery in the hands of the outside the community are likely to be under- state or federal government would be better. The provided. This underscores the need to analysis undertaken here cannot address this strengthen the role of local governments in im- question because we cannot compare outcomes proving public health management and closely across more or less decentralized systems. But coordinating activities in service delivery with the overall policy lesson that the analysis does community based organizations. CHAPTER 1 Introduction I t is increasingly acknowledged that only in countries with good governance, as there are weak links in the chain from measured by lower corruption and quality of public spending to actual outcomes the bureaucracy. The role of institutions and in making basic services available to incentives in the public sector in translating poor people. Simply increasing budget alloca- budgeted resources into actual outcomes is tions to essential services such as health and therefore critical. education is not enough to ensure that quality These issues are particularly important for the services are indeed delivered. Even when re- public delivery of basic health services in Nige- sources are appropriately allocated they may ria. The paramount issue in the health sector in not reach their intended destinations because Nigeria in the 1980s was the tightening finan- of organizational and incentive problems in cial constraints imposed upon public spending public agencies. Even when resources reach the in health following a fiscal crisis and decline in health clinic or the primary school the actual the country's oil revenues. In the last five years, service providers may have weak incentives or however, public revenues in Nigeria have in- capacities to deliver effectively. creased fivefold thanks to a boom in world oil Cross-country evidence suggests that total prices, and the new democratic government has public spending on health has had a surpris- been eager to use the windfall to deliver so-called ingly low impact on average health outcomes, "democracy dividends" to the people. In par- relative to other socio-economic characteristics ticular, spending on primary health care has in- such as income per-capita and female educa- creased substantially. Funding for the National tion (Filmer et al. 2000; Musgrove, 1996). New Programme of Immunization (NPI) for instance, empirical evidence that the impact of public has gone from N9 million in 1998, to close to spending on basic health outcomes depends N7.5 billion in 2001. Allocations for programs upon the overall governance environment pro- to control diseases such as malaria, guineaworm vides an important explanation for the observed and so on has increased to about 165 million in weak relationship between public spending and the 2001 budget. (The World Bank, 2001). Rev- outcomes. Rajkumar and Swaroop (2002) find enues of Local Government Authorities (LGAs) that greater public spending on health signifi- that are primarily responsible for public spend- cantly lowers child and infant mortality rates ing on primary health care, has increased from 7 8 Decentralized Delivery of Primary Health Services in Nigeria an average of 5 percent of GDP between 1990 the unit of analysis in enterprise or investment and 1998, to over 10 percent of GDP in recent climate surveys.1 Both qualitative and quanti- years after the oil price increase in 1999 tative data was collected at the facility level (IMF, 2001). through interviews of facility staff and directly However, there is still little known about from facility records. Facility-level data was whether the institutions and incentives in the linked to data collected from local governments public sector in Nigeria will actually allow large on the governance environment and financing spending programs to effectively deliver basic arrangements. services to the people. There are few concrete answers to questions such as--what has public spending achieved in terms of actual outcomes 1.1 Participation, Ownership, and in service delivery? Do public resources actu- Capacity-Building ally reach their intended destinations? How ac- countable are public service providers to their In addition to its analytical objectives, the con- expected beneficiaries? This study provides some duct of this study was specifically designed to answers to these questions through the analysis promote evidence-based policy dialogue in Ni- of extensive data on expenditures on primary geria by engaging the active participation of the health care and service delivery processes and overarching government agency in the country outcomes, that was collected through a survey responsible for monitoring and supervising out- of local governments and public primary health comes in primary health care service delivery-- care facilities in the states of Lagos and Kogi the National Primary Health Care Development over June-August 2002. Agency (NPHCDA). The terms of reference for The micro-level survey approach of this study this study were developed in partnership with has allowed a deeper investigation of actual NPHCDA, with the agency closely involved at outcomes in service delivery at the frontline, and every stage right from study design to its imple- the impact on these outcomes of broader insti- mentation and subsequent analysis. This part- tutions of governance and financing arrange- nership has facilitated ownership of the results ments, than more aggregative tools of public of the analysis, and is therefore more likely to expenditure analysis. A new survey tool, the allow for greater policy impact. In addition, both Quantitative Service Delivery Survey (QSDS), NPHCDA and the World Bank research team was employed, in which the facility or frontline worked closely with the local consultants that service provider is the main unit of analysis in implemented the study to build local capacity much the same way as a household is the unit in developing potentially useful survey instru- of analysis in household surveys, and a firm is ments and a methodology that would be useful for analyzing the impact of public expenditures on service delivery outcomes. It was agreed to undertake the study at the 1 Provider or facility surveys have been undertaken state-level, given the enormous diversity across from time to time in the context of large house- the different states in Nigeria and the decen- hold surveys such as the Living Standard Mea- tralized nature of service delivery in primary surement Study (LSMS) surveys, the Demographic health care. It was further agreed to focus on and Health Surveys (DHS), and RAND's Family only two states so that the survey instruments Life Surveys. However, these surveys do not focus and analytical methodology could be appro- on the incentive environment of and public ex- penditure flows to public providers. Lindelow and priately developed on a manageable scale, given Wagstaff (2002) provide a review of various health logistical difficulties with conducting field- facility surveys. work. NPHCDA expressed interest in Introduction 9 subsequently expanding the work to other within a state and across states in the extent of states based on success of the instruments and autonomy and the nature of responsibilities of interest expressed by state governments. Lagos local governments in the public delivery of pri- and Kogi states were selected for the study on mary health care services. Hence, an additional the basis of the interest expressed by their Com- interest of this study is to understand whether missioners of Health, and the possibility of there is variation in the extent of local govern- contrasting service delivery between largely ment autonomy and correlate it with variation urban and largely rural settings in two in service delivery outcomes at the level of the different geo-political regions. health facility. The National Health Policy also emphasizes the role of community participation in the deliv- I.2 Objectives of the study ery of primary health care services. It indicates that local governments shall mobilize com- It was agreed that the objective would be to study munis to participate in the provision maintenance the following three issues: 1) the flow of re- of health services, eliciting the support of various sources allocated in public budgets to the front- formal and informal community leaders. The line service delivery agency, that is, the primary study is also designed to examine the role of com- health care facilities, 2) analysis of provider be- munity participation in shaping outcomes in pri- havior and provider incentives in shaping out- mary health care service delivery. comes at the health facility level, and 3) the role Given these objectives it was decided to fo- of local governments and community participa- cus the study exclusively on the performance of tion in determining outcomes in public primary public health facilities, focusing explicitly on the health care service delivery. supply side of the determinants of actual The National Health Policy adopted by the present democratic government in 1999 lays out the roles and functions of each tier of govern- ment in primary health care. While the federal 2To quote: "With the general guidance, support and government is assigned the responsibility of over- technical supervision of State Health Ministries, all policy formulation, coordination, and adher- under the aegis of Ministries of Local Government, Local Government Councils shall design and ence to internationally recognized standards, the implement strategies to discharge the responsibili- state government with the active participation ties assigned to them under the Constitution, and of local governments is responsible for actually to meet the health needs of the local community." delivering primary health care services. How- (page 26, National Health Policy) ever, neither the National Health Policy nor the The Constitution in its turn is also not clear in Constitution of 1999 makes clear prescriptions its prescriptions. To quote from the Fourth Sched- ule which provides a list of functions to be per- about the delineation of responsibilities and formed by Local Government Councils: authorities between the states and local govern- "The functions of a local government council ments. Instead the official language seems to shall include participation of such council in the suggest that state governments have the ultimate Government of a State as respects the following responsibility for delivering primary health care, matters: (a) the provision and maintenance of while the role of local governments can vary primary, adult and vocational education; (b) the within a state and across states depending on development of agriculture and natural re- sources, other than the exploitation of minerals; particular state policies and local socio-economic (c) the provision and maintenance of health ser- conditions.2 vices; and (d) such other functions as may be This lack of clarity in constitutional prescrip- conferred on a local government council by the tions may, therefore, have led to large variations House of Assembly of the State." 10 Decentralized Delivery of Primary Health Services in Nigeria outcomes, as in improved health indicators, that workers, as well as direct collection of data we care about. Consequently, the study is not on inputs and outputs from facility records designed to address issues of public-private part- 2. Survey of local governments (under whose nership in delivering basic services, nor issues jurisdiction the health facilities reside)--in- of household demand and responsiveness to cluding interviewers of local government trea- public initiatives. surers for information on budgeted resources and investment activity, and interviews of pri- mary health care coordinators for roles, re- I.3 The Survey Approach sponsibilities, and outcomes at the local gov- ernment level The approach adopted to addressing these issues revolvesaroundextensiveandrigoroussurveywork, Box 1 and 2 summarize the type of informa- at the level of the primary health care facilities and tion collected through each type of instrument. thelocalgovernments.Twobasicsurveyinstruments The focus of the study is thus public service de- ofprimarydatacollectionwereagreedupon,based livery outcomes as measured at the level of front- oncollectinginformationfromgovernmentofficials line delivery agencies--the public primary health and public service delivery facilities: care facilities. We also originally planned to in- clude interviews of patients present at the health 1. Survey of primary health care facilities--in- facilities, to get the user's perspective on public cluding interviews of facility managers and service delivery, but found that difficult to Box 1 Survey instruments at the health facility level The facility level survey instruments were designed to collect data along the following lines: 1. Basic characteristics of the health facility: who built it; when was it built; what other facili- ties exist in the neighborhood; access to the facility; hours of service etc. 2. Type of services provided: focusing on ante-natal care; deliveries; outpatient services, with special emphasis on malaria and routine immunization 3. Availability of essential equipment to provide the above services 4. Availability of essential drugs to provide the above services 5. Utilization of the above services, referral practices 6. Tracking and use of epidemiological and public health data 7. Characteristics of health facility staff: professional qualifications; training; salary structure, and whether payments are received in a timely fashion; informal payments received; fringe benefits received; do they have their own private practice; time allocation across different services; residence; place of origin 8. Sources of financing--who finances the building infrastructure and its maintenance; who finances the purchase of basic equipment; who finances the purchase of drugs; what is the user fee policy; revenues from user fees; retention rate of these revenues; financing available from the community 9. Management structure and institutions of accountability: activities of and interaction with the local government and with the community development committees Introduction 11 follow-through given local capacity constraints vice delivery. The survey was undertaken dur- in implementing a survey of this kind. ing June-August 2002, with data collected in The survey instruments were developed 30 local governments in Lagos and Kogi states, through an iterative process of discussions be- 252 health facilities, and from over 700 health tween the World Bank team, NPHCDA, and workers. local consultants at the University of Ibadan, A multi-stage sampling process was employed over the months of March-May 2002. During where first 15 local governments were randomly May 2002, four questionnaires were finalized selected from each state; second, 100 facilities through repeated field-testing--1) Health Fa- from Lagos and 152 facilities from Kogi were cility Questionnaire: to be administered to the selected using a combination of random and health facility manager, and to collect recorded purposive sampling from the list of all public data on inputs and outputs at the facility level; primary health care facilities in the 30 selected 2) Staff Questionnaire: to be administered to LGAs that was provided by the state govern- individual health workers; 3) Local Govern- ments; third, the field data collectors were in- ment Treasurer Questionnaire: to collect local structed to interview all staff present at the government budgetary information; and 4) Pri- health facility at the time of the visit, if the total mary Health Care Coordinator Questionnaire: number of staff in a facility were less than or to collect information on local government ac- equal to 10. In cases where the total number of tivities and policies in primary health care ser- staff were greater than 10, the field staff were Box 2 Survey instrument at the local government level The local government survey instruments were designed to collect data along the following lines: 1. Basic characteristics: when was the local government created, population, proportion ur- ban and rural, presence of an urban center, presence of NGOs and international donors 2. Number of primary health care facilities by type (types 1 and 2) and ownership (public-- local government, state, and federal government; private-for-profit; private-not-for-profit) 3. Supervisory responsibilities over the general functioning of the primary health care centers 4. Health staff: number of staff by type of professional training and civil service cadre; salary; 5. Monitoring the performance of health staff: how is staff performance monitored and by whom; are staff rewarded for good performance or sanctioned for poor performance, and how; instances when local government has received complaints; what disciplinary action was taken 6. Budget and financing: data on actual LGA revenues and expenditure from available budget documents; 7. Management structures: functioning of the Primary Health Care Management Committee (PHCMC), the Primary Health Care Technical Committee (PHCTC), and the community based organizations--the Village Development Committee (VDC) and the District Devel- opment Committee (DDC) 8. Health services outputs at the local government level: records of immunization, and envi- ronmental health activities 12 Decentralized Delivery of Primary Health Services in Nigeria instructed to randomly select 10 staff, but mak- tioning of health facilities in Kogi is a striking ing sure that one staff in each of the major ten result that will be discussed in this report. categories of primary health care workers was included in the sample. Health facilities were selected through a com- I.4 Organization of the Report bination of random and purposive sampling. First, all facilities were randomly selected from The rest of the report is organized as follows. the available list for 30 LGAs. This process re- Section II presents the evidence obtained from sulted in no facility being selected from a few the survey--II.1 provides a general overview LGAs. Between 1-3 facilities were then randomly of the characteristics of public facilities pro- selected from these LGAs, and an equal number viding primary health care services in Nige- of facilities were randomly dropped from over- ria; II.2 describes the governance environment represented LGAs, defined as those where the within which service delivery takes place; II.3 proportion of selected facility per LGA is higher discusses financing arrangements for essential than the average proportion of selected facilities services provided at the facility level; II.4 ana- for all sampled LGAs. A list of replacement fa- lyzes staff issues and availability of essential cilities was also randomly selected in the event of inputs at the facility level; II.5 describes the closure or non-functioning of any facility in the outputs and outcomes delivered at the health original sample. An inordinate amount of facili- facilities. Section III explores some emerging ties were replaced in Kogi (27 in total), some due issues in primary health services delivery in to inaccessibility given remote locations and hos- Nigeria--the role of community participation, tile terrain, and some due to non-availability of staff incentives, and decentralized manage- any health staff. The local community volunteered ment by local governments, in the performance in these cases that the reason there was no staff of health facilities. Finally, section IV presents available was because of non-payment of sala- the main conclusions of this study and the ries by the LGA. This characteristic of the func- policy lessons going forward. CHAPTER 2 Survey Results II.1 General description of the facilities very different picture--there, the majority (61%) of facilities were privately owned. 34% were T he public health care service system in owned by the LGAs, 3% by the state, and Nigeria is delivered through a tiered around 1% by the federal government. package of facilities. At the lowest rung A total of 252 facilities were sampled, 150 in of the tier are the Type I facilites, known Kogi and 100 in Lagos. 80% of the facilities as health posts/ clinics. These are village-level fa- sampled in Kogi were health posts/ dispensaries, cilities, typically staffed by a junior paramedic and 15%werePrimaryHealthCenters,4%wereCom- an assistant, with the most basic amenities. Type munity Health Centers, and one was of unclassi- II facilities, also known as Primary Health Cen- fied type. In Lagos, only 9% of the sampled facili- tres,arelargerfacilitieswithamorediversecomple- ties were health posts/ dispensaries, while 47% ment of staff and amenities (see tables below). were Primary Health Centers and 39% were Com- There are supposed to be at least one such facility munity Health Centers. The preponderance of in each health district. Type III facilities, called higher-level facilities in Lagos was underscored by Community Health Centres, are intended to be the presence of a tertiary hospital. 4 of the equippedandstaffedasmini-hospitals,andtoserve facilities in Lagos were of unclassified type. as referral centers for the facilities below them. The local government records indicate that the local government owns the great majority Table II.1.1 (71%) of public health care facilities in the re- Ownership of Facilities gions sampled. The private sector is the other major player, owning 25% of all facilities. The All Kogi Lagos state owned less than 3% of all facilities, while Ownership the federal government owned less than 1%. LGA 827 695 132 The two states are quite different in their pro- State 33 20 13 files. In Kogi, the LGAs own 90% of the facili- ties, while another 7% are state-owned and less Federal 8 2 6 than 1% are owned by the federal government. Private/Non- Only 7% are privately owned. Lagos presents a Government 291 56 235 13 14 Decentralized Delivery of Primary Health Services in Nigeria As discussed below, the Lagos facilities were Substantial proportions of all types of facili- also geographically more proximate to referral ties were in poor repair. Around half of facili- centers, as well as to a range of private facili- ties, of each type, had a leaking roof. Again, ties, than those in Kogi. Thus our study find- health posts in Kogi were in better shape than ings need to be understood in the context of very those in Lagos: 44% of the former but 56% of substantial differences in the nature of health the latter were reported by interviewers' own facilities available to people between the two assessment to have leaking roofs. There were states, as well as corresponding differences in no differences between the two states for the the context in which the facility staff function. other types of facilities: about half of all these For people in Kogi, health posts and dispensa- facilities had leaking roofs. ries necessarily have to meet a wider range of Around half of health posts and PHCs had health care needs for the population, regardless broken doors/windows, while nearly 70% of of the resources available to them. The profile CHCs had this--probably because the majority that emerges below in this report indicates that of them were in Lagos, where facility mainte- Kogi facilities succeed in functioning under very nance appears to be poorer. Around 40% of difficult circumstances in terms of lack of basic PHCs and CHCs had cracked floors, this was amenities. By contrast, people in Lagos have a 50% for health posts. Broken doors/windows variety of private facilities available to them, and were far more prevalent in all types of facility in public facilities appear to be less well maintained, Lagos, as compared with Kogi. Cracked floors despite having much better amenities such as presented a more complex picture: their preva- water and electricity. lence was slightly higher among health posts in Kogi (52%) than in Lagos (33%). PHCs were Condition of Facilities (Table II.1.3) fairly similar, but CHCs in Lagos were in much poorer repair than in Kogi: 46 % had broken Around 30% of health posts were classified by doors/windows, compared to only 17% in Kogi, the interviewers as "dirty/very dirty", while only and 74% had broken doors/windows, compared 10% of PHCs and CHCs were classified as such. to only 33% in Kogi. The overall impression is Health posts in Kogi were cleaner than those in that facilities are better maintained in Kogi than Lagos: 70% of the former but only 56% of the in Lagos. latter were reported by interviewers' own assess- Having a working toilet for patients was pre- ment to be "clean" or "very clean". The differ- dictably low (23%) for health posts, rising to ences between the two states were much smaller half of PHCs to three-quarters of CHCs. On for the other types of facilities. this dimension, Lagos has a better record than Kogi: 44% of health posts in Lagos as com- pared with 21% in Kogi had working toilets. Table II.1.2 PHCs were fairly similar, but amongst CHCs, Number of Facilities by State and Type of Facility 77% had working toilets in Lagos as compared with 33% in Kogi. This may be partly because Kogi Lagos of far higher availability of piped water in Lagos All Facilities 152 100 (see below). Type 1 Facilities 122 9 Most facilities have been working in the past 3 months (Table II.1.5), health posts perhaps Type 2 Facilities 23 47 a little less than others. Of those which had Type 3 Facilities 6 39 not been working in past 3 months, only a few Type 4 Facilities 0 1 health posts provided reasons. The reasons Type Unspecified 1 4 were all related to issues with health Survey Results 15 Table II.1.3 Condition of facilities, by type of facility (% ) Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) n=5 n=252 Dirty / very dirty 29 10 10 ­ 40 20 Clean / very clean 70 87 88 100 40 78 Cleanliness unspecified 1 3 2 ­ 20 2 Total 100 100 100 100 100 100 Leaking roof 45 49 51 100 40 47 Broken doors/windows 46 50 69 100 40 51 Cracked floor 50 37 42 ­ 40 45 Working toilet for patients 23 53 71 100 60 41 Health post/ Kogi dispensary PHC CHC Tertiary Unspecified All Dirty / very dirty 29 4 17 ­ 0 24 Clean / very clean 70 96 83 ­ 100 75 Cleanliness unspecified 1 0 0 ­ 0 1 Total 100 100 100 100 100 100 Leaking roof 44 48 50 ­ 0 45 Broken doors/windows 43 22 33 ­ 0 39 Cracked floor 52 43 17 ­ 0 49 Working toilet for patients 21 52 33 ­ 100 27 Health post/ Lagos dispensary PHC CHC Tertiary Unspecified All Dirty / very dirty 33 13 10 0 0 50 Clean / very clean 56 83 87 100 0 25 Cleanliness unspecified 11 4 3 0 ­ 25 Total 100 100 100 100 100 100 Leaking roof 56 49 51 100 50 51 Broken doors/windows 78 64 74 100 50 69 Cracked floor 33 34 46 0 50 39 Working toilet for patients 44 53 77 100 50 62 16 Decentralized Delivery of Primary Health Services in Nigeria personnel--non-payment of salary by the lo- especially among CHCS, 50% of those in Kogi cal government, strike organized by staff, de- had working laboratories, while only 5% of parture of staff in-charge, either voluntarily or those in Lagos had them. due to a dispute. Access to transport vehicle for emergencies is reportedly available to around a quarter of Facility Amenities (Tables II.1.4 and health posts, a third of PHCs and half of II.1.5) CHCs. While there are little inter-state differ- ences between health posts and CHCs in ac- Three-quarters of PHCs and CHCs have "pro- cess to vehicles for emergencies, a far higher tected" sources of water, compared with only a percentage of PHCs in Kogi had such trans- quarter of health posts. Two-thirds of health port available than in Lagos (61% and 23% posts rely on rivers/streams/open sources for respectively). their water supply. There is a very large gap be- Even more sharply than the data on the con- tween the states in availability of "protected" dition of the facilities, the data on amenities sug- water sources, with Kogi far less privileged than gests that Kogi has more active maintenance of Lagos not only in terms of total availability of health facility infrastructure under difficult cir- protected water, but also the sources thereof. cumstances, while Lagos is far better served in Between 82-88% of facilities of all types had terms of public infrastructure such as water and protected water sources in Lagos--in fact, the electricity--presumably because it is the com- higher figure of 88% pertains to the Lagos health mercial center of the country. posts, as compared with only 18% for Kogi Communication with the outside world is lim- health posts. For PHCs the figures were 65% ited to direct contact in both states, as almost and 83% respectively for Kogi and Lagos, and no facility has working radios or telephones. for CHCs they were 50% and 82%. Moreover, Most (91-97%) of PHCs and CHCs in both most of the protected water in Lagos came from states had been working in the past 3 months. piped water and boreholes, which are sources Amongst health posts, the percentage was lower, preferable to the covered wells which account especially in Lagos, where only 67% of facili- for a substantial proportion of Kogi's protected ties had been working in the past 3 months as water supplies. compared with 87% in Kogi. The gap is even wider between categories There is an average of two beds per health of facilities for working electricity connec- post and PHC, and 8 beds per CHC. The states tions: while most PHCs and CHCs (70% and differ little on this score. 89% respectively) do have this, only 15% of Two-thirds of CHCs have functioning fridges/ health posts do. Once again, Kogi is far less freezers, compared with one-third of PHCs and well-served. Only 11% of Kogi health posts few health posts. Consistent with the differences have working electricity connections, while 67% in the availability of functioning electricity con- of those in Lagos have them. For PHCs the fig- nections, a far higher proportion of Lagos fa- ures were 43% and 83% respectively for cilities had functioning refrigerators than Kogi. Kogi and Lagos, and for CHCs they were 33% A third of Lagos health posts had these, as com- and 97%. pared with only 1% of those in Kogi. Among There is a real shortage of working laborato- PHCs the figures were 40% in Lagos and 22% ries: upto the PHC level, almost no facilities have in Kogi, while among CHCs the figures were working laboratories, and only 11% of CHCs 74% in Lagos and only 17% in Kogi. do. So all treatment of malaria, for example, For storing vaccines (Table II.1.6), 40-50% must be on purely symptomatic grounds. Inter- of each type of facilities said they used cold estingly, Kogi seems to have the edge here: boxes / vaccine carriers. For health posts and Survey Results 17 Table II.1.4 % distribution of main source of water, by type of facility Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) n=5 n=252 Piped water 5 26 29 ­ 20 15 Borehole 8 34 33 ­ 20 20 Protected well 11 17 16 100 20 14 Unprotected well 8 9 9 ­ 20 8 Rain collection 2 1 ­ ­ ­ 2 River, stream, open source 63 9 4 ­ ­ 36 Other 4 3 7 ­ ­ 4 Unspecified 1 1 2 ­ 20 2 Total 100 100 100 100 100 100 Health post/ Kogi dispensary PHC CHC Tertiary Unspecified All Piped Water 2 17 0 0 4 Borehole 7 22 17 100 11 Protected Well 9 26 33 0 13 Unprotected Well 8 9 0 0 8 Rain Collection 2 4 0 0 3 River, Stream, Open S 67 22 33 0 59 Other 4 0 17 0 4 Unspecified Total 100 100 100 100 100 100 Health post/ Lagos dispensary PHC CHC Tertiary Unspecified All Piped Water 44 30 33 0 25 32 Borehole 11 40 36 0 0 34 Protected Well 33 13 13 100 25 16 Unprotected Well 0 9 10 0 25 9 ver, Stream, Open S 0 2 0 0 0 1 Other 0 4 5 0 0 4 Unspecified 11 2 3 0 25 4 Total 100 100 100 100 100 100 18 Decentralized Delivery of Primary Health Services in Nigeria PHCs, this was the main method of storage. Oddly, the proportions of CHCs and PHCs re- Nearly half of CHCs (44%) also used electric porting using fridges/freezers for storing vaccines fridges and freezers for storing vaccines. Non- is around 20% lower than the proportions re- electric refrigerators are virtually non-existent. porting having functional fridges/freezers (Table Table II.1.5 % of each type of facility, with various amenities Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) n=5 n=252 Working electricity connection (%) 15 70 89 100 80 45 Working laboratory (%) 2 3 11 100 ­ 4 Access to vehicle in emergency (%) 27 36 49 100 20 34 Working telephone/radio (%) 2 3 4 ­ ­ 2 % of facilities working in past 3 mths 85 92 98 100 80 90 Average no of beds 2 2 8 140 2 ­ Functional fridge/freezer 3 34 67 100 20 24 Health post/ Kogi dispensary PHC CHC Tertiary Unspecified All Working electricity connection (%) 11 43 33 ­ 100 18 Working laboratory (%) 2 4 50 ­ 0 4 Access to vehicle in emergency (%) 27 61 50 ­ 100 34 Working telephone/radio (%) 2 0 0 ­ 0 1 % of facilities working in past 3 mths 87 91 100 ­ 100 88 Average no of beds 2 3 9 ­ ­ 2 Functional fridge/freezer 1 22 17 ­ 0 5 Health post/ Lagos dispensary PHC CHC Tertiary Unspecified All Working electricity connection (%) 67 83 97 100 75 87 Working laboratory (%) 0 2 5 100 0 4 Access to vehicle in emergency (%) 33 23 49 100 0 34 Working telephone/radio (%) 0 4 5 0 0 4 % of facilities working in past 3 mths 67 94 97 100 75 92 Average no of beds 0 2 7 140 2 5 Functional fridge/freezer 33 40 74 100 25 53 Survey Results 19 II.1.5). Also oddly, over a third of health posts For sterilizing equipment (Table II.1.6), all and PHCs said the question of storing vaccines types of facilities rely heavily (67­87%) on was "not applicable"--hopefully by this they boiling. 11% of health posts use chemicals meant that they didn't store them for any length for sterilizing equipment. Although fairly of time because they used cold boxes and dis- similar percentages of facilities of different pensed the vaccines as soon as they received types reported "boiling" as the primary them. The differences between the various types method of sterilization, the results for CHCs of facilities was statistically significant at the were statistically significantly higher: at the 1% level. 10% level compared with health posts and Very few of the Kogi facilities, of any type, dispensaries, and at the 5% level compared used fridges/freezers for storing vaccines-- with PHCs. 17% of PHCs (and even 2% of they reported either using cold boxes/vaccine CHCs) said this question was "not appli- carriers, or that this question was "not appli- cable", which is not reassuring. Interestingly, cable". By contrast in Lagos, fridges/ freezers this response was concentrated in Lagos, were used for storing vaccines by 22% of where as much as 23% of PHCs reported health posts, 17% of PHCs and 49% sterilizing equipment to be "not applicable" of CHCs. to their situation. The tertiary facility uses Table II.1.6 How vaccines are stored (%) Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) n=5 n=252 Electric fridge/freezer 2 13 44 100 20 13 Non-electric fr/fr ­ 1 ­ ­ ­ ­ Cold box/vaccine carrier 48 46 38 ­ ­ 44 Non-refrigerated storage 3 1 2 ­ - 2 Not applicable 38 36 13 ­ 60 33 Unspecified 9 3 2 ­ 20 6 Total 100 100 100 100 100 100 Health post/ Kogi dispensary PHC CHC Tertiary Unspecified All Electric fridge/freezer 0 4 17 ­ 0 1 Non-electric fr/fr 0 4 0 ­ 0 1 Cold box/vaccine carrier 50 57 67 ­ 0 51 Non-refrigerated storage 3 0 0 ­ 0 3 Not applicable 38 35 0 ­ 100 36 Unspecified 9 0 17 ­ 0 8 Total 100 100 100 ­ 100 100 (continued on next page) 20 Decentralized Delivery of Primary Health Services in Nigeria Table II.1.6 How vaccines are stored (%) (continued) Health post/ Lagos dispensary PHC CHC Tertiary Unspecified All Electric fridge/freezer 22 17 49 100 25 31 Non-electric fr/fr ­ ­ ­ ­ ­ ­ Cold box/vaccine carrier 22 40 33 0 0 34 Non-refrigerated storage 0 2 3 0 0 2 Not applicable 44 36 15 0 50 29 Unspecified 11 4 0 0 25 4 Total 100 100 100 100 100 100 Table II.1.7 How usually sterilize equipment (%) Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) n=5 n=252 Autoclave 1 1 2 ­ ­ 1 Steam 3 6 7 100 ­ 5 Boiling 74 67 87 ­ 80 74 Chemicals 11 3 2 ­ ­ 7 Not applicable 7 17 2 ­ ­ 9 Other ­ 4 ­ ­ ­ 1 Unspecified 5 1 ­ ­ 20 ­ Total 100 100 100 100 100 100 Health post/ Kogi dispensary PHC CHC Tertiary Unspecified All Autoclave 1 0 0 0 ­ 1 Steam 3 4 33 0 ­ 5 Boiling 74 87 67 100 ­ 76 Chemicals 11 4 0 0 ­ 10 Not applicable 7 4 0 0 ­ 6 Other ­ ­ ­ ­ ­ ­ Unspecified 4 0 0 0 ­ 3 Total 100 100 100 100 ­ 100 (continued on next page) Survey Results 21 Table II.1.7 How usually sterilize equipment (%) Health post/ Lagos dispensary PHC CHC Tertiary Unspecified All Autoclave 0 2 3 0 0 2 Steam 0 6 3 100 0 5 Boiling 78 57 90 0 75 72 Chemicals 0 2 3 0 0 2 Not applicable 11 23 3 0 0 13 Other 0 6 0 0 0 3 Unspecified 11 2 0 0 25 3 Total 100 100 100 100 100 100 the more advanced technology of steam should keep only one of these measures (which- sterilization. ever measure is likely to be the more accurate), to reduce confusion. Availability of other facilities nearby (Table II.1.8) II. 2. Governance Environment Health posts are considerably further on aver- age from LGA HQ, and from the nearest refer- This section studies the governance environment ral center, than PHCs and CHCs. Most facili- in Nigeria within which primary health care ties have other health facilities available within (PHC) services are provided, focusing on two a 2-hour walking radius. But compared with striking characteristics: (i) the decentralization PHCs and CHCs, health posts have half (or less) of responsibility for PHC service delivery to lo- as many of these available on average, -- largely cal governments, and (ii) the institutionalization because PHCs and CHCs have a plethora of of community participation in PHC service de- small private clinics available (8-10 on average). livery through community-based health devel- In addition, the CHCs have an average of 3 pri- opment committees. vate secondary or tertiary facilities available within a 2-hour walking radius. Decentralization to Local Governments Kogi health posts are especially remote, with an average walking time of 9 hours to reach Nigeria has been organized as a federal country the LGA HQ -- as compared to just over half since 1954 with the responsibility for providing an hour in Lagos. Similarly, they have an aver- most public goods being concurrently shared be- age walking time of nearly 4 hours to the near- tween the federal and state governments. In 1976, est referral center, compared with only 1.6 local government authorities (LGAs) were estab- hours in Lagos. Lagos also has a far higher den- lished and recognized as the third tier of govern- sity of private facilities available near public ment, responsible for participating in the delivery facilities of all types, especially near PHCs of most local public services along with state gov- and CHCs. ernments, and entitled to statutory revenue allo- (Note that the ratio of distance to walking cations from both the federal and state govern- time is quite different for different questions. We ments for the discharge of their responsibilities. 22 Decentralized Delivery of Primary Health Services in Nigeria In the late 1980s there was a national initiative to Local Government, Local Government overhaul the primary health care system through Councils shall design and implement strat- the adoption of a new national health policy, in egies to discharge the responsibilities as- the context of which the federal and state gov- signed to them under the Constitution, and ernments issued directives in giving LGAs full ju- to meet the health needs of the local com- risdiction over the delivery of PHC services munity." (page 26, National Health Policy) (Adeniyi and Oladepo, 2003). The current national health policy document, Yet, the current Constitution (1999) of Nige- revised in 1996, indicates that local governments ria is ambiguous with regard to the authority are expected to be the main implementers of PHC and autonomy of local governments in provid- policies and programs, with the federal govern- ing basic services, such as primary health, for ment responsible for formulating overall policy which they have been assigned responsibility and for monitoring and evaluation, and state gov- through sectoral directives. The Fourth Sched- ernments for providing logistical support to the ule of the Constitutions lists the functions of LGAs such as personnel training, financial assis- LGAs as follows: tance, planning and operations. To quote: "The functions of a local government coun- "With the general guidance, support and cil shall include participation of such coun- technical supervision of State Health Min- cil in the Government of a State as respects istries, under the aegis of Ministries of the following matters: (a) the provision and Table II.1.8 Average distance from LGA and other health facilities, by type of facility Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) n=5 n=252 Distance to LGA HQ (km) 24 9 6 5 5 ­ Walking time to LGA HQ (hours) 8 2 2 1 1 ­ Distance to referral center (km) 15 10 12 ­ 4 ­ Walking time to referral center (hours) 4 3 3 24 1 ­ Driving time to referral center (hours) 1 <1 <1 5 <1 ­ % with other health facilities within 2-hr walk /10km radius 87 97 96 100 80 91 Average # of (within 2-hr): Total 7 14 18 9 13 ­ Public PHCs 4 3 3 ­ 2 ­ Public sec/tertiary 1 1 1 ­ 1 ­ Small Private clinics 2 8 10 5 9 ­ Private sec/tertiary 1 1 3 4 1 ­ (continued on next page) Survey Results 23 Table II.1.8 Average distance from LGA and other health facilities, by type of facility (continued) Health post/ Unspec- Kogi N dispensary N PHC N CHC N Tertiary N ified N All Distance to LGA HQ (km) 121 25.2 23 8.9 6 1.8 ­ ­ 1 10.0 151 22 Walking time to LGA HQ (hours) 114 8.9 23 2.0 6 0.4 ­ ­ 1 1.3 144 7 Distance to referral center (km) 121 15.8 23 9.4 6 3.7 ­ ­ 1 10.0 151 14 Walking time to referral center (hours) 113 3.9 23 2.3 6 1.1 ­ ­ 1 1.3 143 3 Driving time to referral center (hours) 120 0.6 23 0.4 6 0.1 ­ ­ 1 0.2 150 1 % with other health facilities within 2-hr walk /10km radius 107 87.7 23 100.0 6 100.0 ­ ­ 1 100.0 137 90 Average # of (within 2-hr): Total 122 7.0 23 7.3 6 9.3 ­ ­ 1 9.0 152 7 Public PHCs 122 3.7 23 3.2 6 5.3 ­ ­ 1 4.0 152 4 Public sec/ tertiary 122 0.6 23 1.0 6 1.2 ­ ­ 1 1.0 152 1 Small Private clinics 122 2.2 23 2.7 6 2.5 ­ ­ 1 4.0 152 2 Private sec/ tertiary 122 0.5 23 0.5 6 0.3 ­ ­ 1 0.0 152 0 Distance to LGA HQ (km) 8 2.8 46 8.4 39 6.2 1 5.0 3 3.3 97 7 Walking time to LGA HQ (hours) 8 0.6 45 2.7 39 2.2 1 1.0 3 0.8 96 2 Distance to referral center (km) 8 7.3 46 11.0 39 13.8 1 0.0 3 2.0 97 11 (continued on next page) 24 Decentralized Delivery of Primary Health Services in Nigeria Table II.1.8 Average distance from LGA and other health facilities, by type of facility (continued) Health post/ Unspec- Kogi N dispensary N PHC N CHC N Tertiary N ified N All Walking time to referral center (hours) 8 1.6 45 3.6 39 3.7 1 24.0 3 1.4 96 4 Driving time to referral center (hours) 8 0.5 46 0.4 39 0.4 1 5.0 3 0.2 97 0 % with other health facilities within 2-hr walk /10km radius 7 77.8 45 95.7 37 94.9 1 100.0 3 75.0 93 93 Average # of (within 2-hr): Total 8 9.8 46 17.2 39 19.5 1 9.0 3 14.7 97 17 Public PHCs 8 3.0 46 3.0 39 2.4 1 0.0 3 1.7 97 3 Public sec/ tertiary 8 1.3 46 1.7 39 1.3 1 0.0 3 0.7 97 1 Small Private clinics 8 3.9 46 10.7 39 11.9 1 5.0 3 10.7 97 11 Private sec/ tertiary 8 1.6 46 1.8 39 3.9 1 4.0 3 1.7 97 3 maintenance of primary, adult and voca- or across states in the extent to which responsi- tional education; (b) the development of bility for PHC services is effectively decentral- agriculture and natural resources, other ized. In the face of such constitutional ambigu- than the exploitation of minerals; (c) the ity, the survey of LGAs and health facilities at- provision and maintenance of health ser- tempted to assess the extent of decentralization vices; and (d) such other functions as may of PHC services to local governments. be conferred on a local government coun- The survey asked respondents at both the cil by the House of Assembly of the State." LGA and facility level which agency, choosing one amongst the federal government, the state This implies that according to the Constitu- government, the LGA, community-based orga- tion, it is the state governments that have prin- nizations, and facility head or staff, was the prin- cipal responsibility for basic services such as cipal decision-maker for each of the following primary health and primary education, with the areas of PHC service provision in health extent of participation of LGAs in the execu- facilities: tion of these responsibilities determined at the discretion of individual state governments. The · Undertaking new construction, such as facil- constitutional existence of state-level discretion ity expansion may lead to disparities across local governments · Acquiring new equipment Survey Results 25 · Making drugs and medical supplies available ing in any individual facility. Table II.2.1 lists · Setting charges for drugs and treatment the frequency of responses for each agency by · Use of facility revenues from treatment and each type of service delivery decision area. There consultation is, therefore, no evidence from the survey of state · Disciplining staff governments being actively engaged in the pro- · Transferring staff between facilities vision of PHC services, as appears to be indi- cated in the Constitution. Amongst government The overwhelming majority of LGA respon- agencies, the LGA is overwhelmingly indicated dents indicated the LGA as the principal deci- as primarily responsible for PHC, with no sig- sion-maker for most of the areas of facility- nificant variation in responses across the LGAs level provision of PHC services. Of the 29 or between the two states surveyed. LGAs that responded to these questions, 21 In addition to the LGA, it was the commu- listed the LGA as the principal decision-maker nity development committees and the facility for all of the areas listed above. Of the remain- head and staff that were indicated as principal ing LGAs, 7 listed the LGA as the principal decision-makers in some specific facility deci- decision-maker for most service delivery activi- sions. For making drugs, supplies, and equip- ties, except one or two areas that were non- ment available, and/or setting charges of drugs, systematically assigned to other agencies--for and/or determining use of facility revenues, the example, the state government was cited by 2 community development committees and/or fa- LGAs as the principal decision-maker for un- cility head or staff was indicated for about 35% dertaking new construction, by 1 LGA for set- of the facilities surveyed.3 For decisions to un- ting charges of drugs and treatment, and by 1 dertake new construction or expansion, for decisions of transferring staff between fa- community development committees were indi- cilities, with all other decisions being princi- cated as principal decision-makers for 26% of pally determined by the LGA. Only one LGA, all facilities surveyed. However, decisions related Ibaji LGA in Kogi state, listed an agency other to staff discipline were overwhelmingly cited as than the LGA, namely, community based or- the responsibility of the LGA. ganizations, as the principal authority for ma- There are striking differences in the sharing jority of the decisions of day-to-day running of responsibilities between the LGA and com- of facilities. This LGA had been pointed out munity development committees in the two states during field-work for the survey as particularly of Lagos and Kogi studied here. Of the 97 remarkable for the extent of community participation in PHC service delivery. The facility-level respondents similarly indi- cated the LGA as the principal decision-maker 3 for most service provision decisions at the About 53 facilities reported community organi- zations as principal decision-makers and about 62 facility level, as compared to the other two tiers reported facility head/staff as principal decision- of government--the state and the federal gov- makers in one or more of the following areas-- ernment. Out of 249 facility-level respondents making drugs and medical supplies available, ac- that answered most of the questions related to quiring new equipment, setting charges for drugs, facility decision-making, 61% indicated the LGA and deciding what to do with facility user rev- as the principal decision-maker for all or most enues. About 28 of these 115 facilities reported both communities and facility head/staff as prin- activities listed earlier. The state and federal gov- cipal decision-makers for different activities in this ernments were indicated very infrequently as list. Hence, 87 facilities reported either communi- principal decision-makers for any area, and even ties or facility head/staff as principal decision- then for only one or two areas of decision-mak- makers in one or more of these areas. 26 Decentralized Delivery of Primary Health Services in Nigeria Table II 2.1 Principal Decision-Makers for Facility Functioning Undertaking New Construction/Repairs Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 1 0.4 Local Government 177 70.24 Community 65 25.79 Facility Head/Staff 2 0.79 Missing Response 6 2.38 Acquiring New Equipment Principal Decision-Maker Frequency Percentage Federal Government 0 0 State Government 0 0 Local Government 203 80.56 Community 28 11.11 Facility Head/Staff 17 6.75 Missing Response 3 1.19 Making Drugs Available Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 0 0 Local Government 182 72.22 Community 27 10.71 Facility Head/Staff 37 14.68 Missing Response 5 1.99 Making Medical Supplies Available Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 0 0 Local Government 199 78.97 Community 20 7.94 Facility Head/Staff 28 11.11 Missing Response 4 1.19 (continued on next page) Survey Results 27 Table II 2.1 Principal Decision-Makers for Facility Functioning (continued) Setting Drug Charges Principal Decision-Maker Frequency Percentage Federal Government 2 0.79 State Government 2 0.79 Local Government 173 68.65 Community 14 5.56 Facility Head/Staff 48 19.05 Missing Response 13 5.15 Setting Treatment Charges Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 3 1.19 Local Government 173 68.65 Community 13 5.16 Facility Head/Staff 47 18.65 Missing Response 15 6.96 Use of Facility Revenues from Treatment and Consultations Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 1 0.4 Local Government 176 69.84 Community 32 12.70 Facility Head/Staff 23 9.13 Missing Response 19 7.54 Staff Discipline Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 5 1.98 Local Government 210 83.33 Community 9 3.57 Facility Head/Staff 21 8.33 Missing Response 6 2.38 (continued on next page) 28 Decentralized Delivery of Primary Health Services in Nigeria Table II 2.1 Principal Decision-Makers for Facility Functioning (continued) Staff Transfers Principal Decision-Maker Frequency Percentage Federal Government 1 0.4 State Government 8 3.17 Local Government 233 92.46 Community 3 1.19 Facility Head/Staff 2 0.79 Missing Response 5 1.98 Source: Survey Data facilities in Lagos that responded to most of these Committee (PHCMC) and its technical arm, the responsibility questions, 74% indicated the LGA Primary Health Care Technical Committee as principal decision-maker, while of the 152 (PHCTC). In order to get a picture of the extent respondents in Kogi, only 52% indicated the of monitoring of health facilities by the PHCMC, LGA as principal decision-maker. The remain- the survey asked several questions related to the ing facilities in Kogi listed either the community activities of this committee which is shown in development committees or the facility head or Table II.2.2. PHCMCs appear to be quite active both as the principal decision-makers. Of the 53 in Kogi, with over 80 percent of the sample re- facilities in the sample that listed community porting that the committee visits the facility development committees as principal decision- regularly, monitors patient registers, drug stocks, makers for one or more of the following areas-- and equipment, and discusses medical protocol making drugs, supplies and/or medical equip- and administrative issues. In Lagos, PHCMCs ment available, setting charges of drugs, deter- appear less active, with more than 40 percent of mining use of facility revenues--48 belonged to the sample either not responding to the ques- Kogi, and only 5 to Lagos. Of the 65 facilities tions or reporting that the committee visits rarely where communities were reported as principal or never. In both states it is surprising to note decision-makers for undertaking new construc- the low frequency of responses for checking of tion, 61 belonged to Kogi and only 4 to Lagos. user receipts in the facility by the PHCMC--as Hence, while the LGA has predominant respon- we will discuss in the section on financing, most sibility for PHC service delivery in both states, facilities responded that revenues from user as compared to the state and federal govern- charges are supposed to be handed-over to the ments, in Kogi PHC service delivery appears to local government and not retained for general be characterized by active participation of facility purposes. communities and facility staff. In its implementation guidelines for primary Community Participation health care services the National Health Policy requires all local governments to establish com- The national health policy in Nigeria empha- mittees that will manage, monitor, and evaluate sizes active community engagement in the pro- health care programs and provide technical ad- vision of PHC services in the spirit of the Bamako vice to the local government council. These Initiative of 1987, when Health Ministers from are the Primary Health Care Management various African nations adopted resolutions for Survey Results 29 promoting sustainable primary health care ties that committee members engage in during a through community participation in financing, facility visit--discussion of medical protocol and maintenance, and monitoring of services. Com- administrative and staff issues leads the list for munity participation was institutionalized in both states. Nigeria through the creation of development Table II.2.3 also provides a picture of the ex- committees at the level of the district--district tent of active community participation in dif- development committee (DDC)--and the vil- ferent areas of service provision. Development lage--village development committees (VDC), committees in Kogi are particularly active in with explicit guidelines for their respective du- supporting service provision in the areas of car- ties and responsibilities. rying out repairs on facility structures (with 54% District or village development committees are of the facilities reporting action in this area in indicated as existing and engaging in various as- the past year), providing drugs to the facility pects of service provision at the facility level for (27%), and resolving administrative and person- 87% of the facilities surveyed. Of these, 84% nel management issues (25%). Development are indicated as active in the sense of meeting committees in Lagos are significantly less active, regularly to discuss facility operations either with less than 15% of the facilities reporting once a month or a few times a year, and 80% any action undertaken by the committees in the indicate that the committee members visit the past year. However, even in Kogi, there is very facility either once a month or a few times a little community engagement in setting charges year. Table II.2.3 shows the typical list of activi- of drugs which would be critical for maintain- Table II.2.2 Activities of Primary Health Care Management Committee KOGI LAGOS Frequency of Visits to Facilities: Once a month or a few times a year 85% 60% Once or twice a year 7% 6% Very rarely or never 8% 34% Activities of development committee during a facility visit: Checking patient register 90% 67% Checking stock cards 82% 44% Checking user charge receipts 69% 13% Discuss medical protocol 89% 71% Discuss administrative issues 89% 83% Hold an official staff meeting 76% 64% Checking equipment 83% 66% Values in the columns indicate the percentage of facility respondents that responded "yes" to the questions for com- munity participation; the respondents are 144 facilities in Kogi and 79 facilities in Lagos for the first question on fre- quency of visits; thereafter total number of respondents is between 135-140 for Kogi and 60-65 for Lagos Source: Survey Data 30 Decentralized Delivery of Primary Health Services in Nigeria Table II.2.3 Community Participation in Kogi and Lagos Kogi Lagos Existence of Village/District Development Committee 96% 74% Activities of development committee during a facility visit: Checking patient register 31% 29% Checking stock cards 24% 18% Checking user charge receipts 19% 4% Discuss medical protocol 47% 35% Discuss administrative issues 43% 39% Hold an official staff meeting 58% 25% Checking equipment 34% 18% Actions undertaken by committee in past year: Committee made disciplinary recommendations on staff 14% 3% Committee provided drugs to the facility 27% 4% Committee fixed the price of drugs in the facility 13% 7% Committee fixed user charges and fees 6% 3% Committee requested more vaccines 37% 9% Committee carried out structural repairs 54% 13% Committee provided fuel or other resources 13% 10% Committee repaired equipment 20% 12% Committee made new investments 17% 3% Committee resolved administrative issues 24% 12% Committee resolved staff personnel issues 25% 3% Values in the columns indicate the percentage of facility respondents that responded "yes" to the questions for com- munity participation; the respondents are the entire sample of 152 facilities in Kogi and 94 facilities in Lagos for the first question on the existence of a development committee, but thereafter total number of respondents is an average of 140 for Kogi and between 50 and 70 for Lagos ing a drug revolving fund, which is a particular while Lagos LGAs report a higher average of aspect of community participation emphasized 74%. Yet, the picture at the facility level sug- in Nigerian health policy. gests that if the sample is representative of an Respondents at the LGA level do not appear average facility in each of the states, then the to be well informed about the extent of com- chances of a Kogi facility having a village devel- munity participation within their jurisdiction, opment committee is 83%, while the chances of as evidenced by their responses on the existence a Lagos facility having a village development of development committees being in stark con- committee is only 56%. trast to what is reported at the facility level. Kogi Community participation in Kogi is concen- LGAs report on average that only 60% of vil- trated in the running of health posts/dispensa- lages have village development committees, ries (Type 1 facilities). Table II.2.4 shows the Survey Results 31 Table II.2.4 Community Participation in Kogi across Facility Types Type 1 Type 2 Type 3 No. of facilities reporting communities as principal decision-makers in making essential supplies available, and/or setting charges for drugs, and/or determining use of facility revenues 45 3 0 No. of facilities reporting communities as principal decision-makers for undertaking new construction 56 4 1 Total number of facilities in the sample 122 23 6 distribution of facilities in Kogi that reported medical commodities, are all predominantly pro- community development committees as princi- vided by local governments in Lagos state. How- pal decision-makers in undertaking construc- ever, in Kogi, community-based organizations tion and in one or more of the following ar- and facility staff are frequently indicated by fa- eas--making drugs and medical supplies avail- cilities as the main source of drugs (for 28% of able, acquiring equipment, setting charges for facility respondents) and medical supplies drugs, and deciding what to do with facility rev- (31%). With regard to building maintenance, enues--across the three types of health facili- 57% of respondents indicated that community- ties in the sample. Amongst the Type 1 facilities based organizations were the main suppliers in sampled in the state, 37% indicated communi- the last twelve months, as compared to only 24% ties as principal decision-makers in the latter of respondents that indicated the LGA as the category, and 46% indicated communities as main supplier. It is surprising to note that as principal decision-makers for undertaking new many as 15% of facilities in Kogi indicate staff construction. In contrast, only 15­17% of type personal funds as the main source of facility re- 2 facilities sampled in the state indicated com- sources, which if accurate probably implies that munities as principal decision-makers for mak- staff compensate themselves from facility rev- ing essential supplies available, or for under- enues. In Lagos, for the majority of facilities taking new construction. This may indicate that (over 85%) resources were either provided by given few choices for primary health care ser- the LGA or indicated as not provided at all in vices in rural areas of the state, communities the last twelve months. Staff salaries are almost decide to invest efforts in improving the quality exclusively provided by local governments in of services available in the public facilities in both states. their neighborhood. Hence, financing of day-to-day facility func- tioning is largely provided by local governments. However, the National Health Policy provides II.3. Financing Arrangements general guidelines to all three tiers of govern- ment to prioritize resource allocation in favor The survey evidence presented in Table II.3.1. of preventive health services and primary health shows that amongst government agencies the care, which is the cornerstone of the national LGA is the main source of financing of PHC program. In this spirit of prioritization, the fed- service delivery at the facility level. Staff sala- eral and state governments are expected to pro- ries, facility building construction and mainte- vide logistical and financial assistance to the nance, supply of drugs, equipment and other LGAs, primarily for programs of national 32 Decentralized Delivery of Primary Health Services in Nigeria Table II.3.1 Main Supplier of Facility Resources Kogi Lagos Who has been the main supplier of drugs to the facility in the last year? Federal Government 2.7% 1.1% State Government 4.7% 6.4% Local Government 53.7% 69.2% Community Development Committee 12.1% 2.1% Facility Funds 4.7% 1.1% Staff Personal Funds 15.4% 2.1% NGO/Donor/Individuals 5.4% 1.1% Not supplied in the last year 1.3% 17% Who has been the main supplier of other medical commodities to the facility? Federal Government 2% 0% State Government 4.7% 7.3% Local Government 50.3% 69.8% Community Development Committee 16.1% 0% Facility Funds 2% 0% Staff Personal Funds 14.8% 0% NGO/Donor/Individuals 6% 6.6% Not supplied in the last year 4% 16.7% Values in the columns indicate the percentage of facility respondents that responded "yes" for the agency listed to the left; the total respondents are an average of 140 respondents for Kogi and 94 respondents for Lagos Who has been the main supplier of new equipment to the facility in the last year? Federal Government 0.7% 0% State Government 1.5% 4.1% Local Government 55.8% 61.9% Community Development Committee 14.5% 0% Facility Funds 0.7% 0% Staff Personal Funds 8.7% 0% NGO/Donor/Individuals 3.6% 1% Not supplied in the last year 14.5% 33% (continued on next page) Survey Results 33 Table II.3.1 Main Supplier of Facility Resources (continued) Kogi Lagos Who has been the main supplier of equipment maintenance in the facility? Federal Government 0.7% 0% State Government 1.4% 2.1% Local Government 38.6% 61.5% Community Development Committee 15% 0% Facility Funds 2.1% 1% Staff Personal Funds 22.9% 2.1% NGO/Donor/Individuals 5% 2 Not supplied in the last year 14.3% 31.3 Values in the columns indicate the percentage of facility respondents that responded "yes" for the agency listed to the left; the total respondents are an average of 140 respondents for Kogi and 94 respondents for Lagos Who has been the main supplier of facility building maintenance in the last year? Federal Government 0.7% 0% State Government 0.7% 4.1% Local Government 23.8% 60.8% Community Development Committee 57.3% 1% Facility Funds 1.4% 0% Staff Personal Funds 3.5% 1% NGO/Donor/Individuals 1.4% 1% Not supplied in the last year 11.2% 32% Who pays staff salary? Federal Government 2.5% 0.9% State Government 0.4% 1.9% Local Government 94.58% 96.2% Community Development Committee 0% 1.1% NGO/Donor/Individuals/Other 2.5% 0% Values in the columns indicate the percentage of facility respondents that responded "yes" for the agency listed to the left; the total respondents are an average of 140 respondents for Kogi and 94 respondents for Lagos; Values in the columns for staff salary indicates the percentage of staff respondents that responded "yes" for the agency listed to the left; the total respondents are 240 staff for Kogi and 472 staff for Lagos 34 Decentralized Delivery of Primary Health Services in Nigeria importance such as the National Program of Im- LGA Finances munization, or controlling the spread of HIV/ AIDS. The federal budget in recent years has Local government expenditure responsibilities included programs of construction of PHC fa- are financed largely through statutory alloca- cilities in local governments. However, there are tions from the Federation Account, with LGAs no established rules or policies for the provision regularly receiving about 20 percent of total fed- of financial assistance from the higher tiers of eral resources in the divisible pool. Since oil rev- government, and it is not clear how well any enues are part of the Federation Account, LGAs assistance that is forthcoming is coordinated receive substantial revenues on account of this with LGA budgets and plans for PHC services. statutory allocation. LGAs are also entitled to a Although this survey has not provided any evi- share of federally collected VAT revenues (out- dence with regard to coordination between the side of the Federation Account). In addition, three tiers, more qualitative studies on the ex- LGAs are supposed to receive statutory alloca- tent of coordination between different agencies tions from state government revenues, but the have indicated that there is often lack of clarity, rules related to this are less strict and not al- wastage of resources, and lack of ownership of ways enforced. Total LGA revenues in the coun- local governments in efforts that require coor- try amounted on average to over 5 percent of dination between all three tiers of government GDP between 1990 and 1999, and over 10 per- (IMF, 2001). cent of GDP after the oil price increase in 1999. There is also some evidence from qualitative LGAs also have recourse to significant own tax work in Lagos state that local governments are bases, although studies have shown that these not able to assume full responsibility for non- have not been explored to full potential, and facility services such as water and sanitation, that internally generated revenues are a small which are an integral part of primary health care proportion of total LGA revenues (IMF, 2001). services, leading to situations where the state The survey collected data on LGA revenues government has to actively intervene and solve and health expenditures for 1999 and 2000 from critical problems (IMF, 2001). This survey does available budget documents. Table II.3.2 show not provide evidence on the functioning of local summary statistics on per capita revenues in the governments in this regard. two states. Average per capita revenues in both Table II.3.2 Per Capita LGA Revenues Mean Std. Dev. Minimum Maximum Kogi 1999 Per capita revenues 1018.6 599.6 443.4 2391.8 2000 Per capita revenues 2191.2 1218.2 1190.6 5634.8 Lagos 1999 Per capita revenues 1266.4 1623.1 465.1 6753.7 2000 Per capita revenues 2352.3 3428.1 582.8 14412.1 Source: Survey Data. 2000 data is for 15 LGAs in each state; 1999 data is for 13 LGAs in Kogi (missing values for Kogi and Lokoja LGAs) and 14 LGAs in Lagos (missing values for Ojoo). Survey Results 35 states more than doubled in 2000, owing to the is also internally generated from local tax bases. country-wide increase in oil revenues which led This is as one would expect given that Lagos to greater allocations to LGAs from the Federa- state is the urban center of Nigeria, while Kogi tion Account. The facility survey has therefore is a largely rural state. The consequences for been undertaken at a time when LGA revenues basic health service delivery between the two have been substantial and rising. Although the states is also clear--services in Kogi are much levels of per capita revenues are not significantly more vulnerable to external shocks that affect different across the two states, there is greater oil prices, which is why, perhaps, communities variation across LGAs in Lagos state, with the in Kogi take a more active role in maintaining richest LGA (Ibeju-Lekki) having more than basic health services. 10 times the per capita revenues of the The survey attempted to collect budgetary poorer LGAs. data on health expenditures of local govern- Figures II.3.1 and II.3.2 show the composi- ments, which was a difficult exercise because tion of LGA revenues on average for each of the budget documents and categories across local two states. Local governments in Kogi are over- governments, both within and across states, are whelmingly dependent on statutory allocations not uniform. During the field testing of the sur- from the Federation Account for their revenues, vey instruments it was observed that numbers and receive almost nothing from the state gov- on total health expenditures were either not easy ernment. Revenue sources of local governments to find or simply not available in LGA budget in Lagos are more diversified--bulk of their rev- documents. However, three categories of expen- enues comes from two sources, the Federation ditures that appeared to show-up more consis- Account and the VAT, but a significant amount tently across LGAs were expenditures on health Figure II.3.1 Composition of Kogi State Revenues KOGI: Composition of 1999 Revenues KOGI: Composition of 2000 Revenues VAT Revenues VAT Revenues 17% Federal Statutory 9% Federal Statutory Allocation Allocation Internal Revenues 80% Internal Revenues 90% 2% 1% State Statutory State Statutory Allocation Allocation 1% 0% Source: Survey Data. Missing data for two LGAs (Kogi and Lokoja) in 1999 36 Decentralized Delivery of Primary Health Services in Nigeria personnel, overheads, and capital projects. These because of potential bias introduced by several also appeared to be exhaustive categories for the missing observations. Lower public expenditures budgeting of health expenditures. Hence, data on health in Lagos LGAs may be because of was collected on these three categories of health greater availability of private health care in the expenditure, which we add-up here to estimate substantially more urban state. total health expenditures by local governments. Figures II.3.3 and II.3.4 shows the average There are several missing values for this esti- composition of health expenditures in terms of mate of total health expenditures, arising when- capital, overheads, and personnel expenditure ever any one of the three categories--person- in 2000 for each of the states. Bulk of LGA nel, overheads, and capital--are missing. In to- health expenditures are allocated to staff sala- tal, we have missing values for total health ex- ries--in Kogi in 2000, LGAs on average spent penditures for 7 LGAs in Kogi and 1 LGA in 78% of health expenditures on salaries, while Lagos for the 1999 budget, and for 4 LGAs in in Lagos, LGAs spent 65% on average on staff Kogi and 3 in Lagos for the 2000 budget. salaries. Table II.3.3 shows summary statistics for per capita total health expenditures in the two states, Facility-level finances and the proportion of total local government revenues spent on health. For the sample for Although the survey asked several questions re- which data is available, Kogi LGAs spend more lated to fees charged at facilities for their ser- per capita and as a proportion of total revenues vices, the responses were often inconsistent on health than do Lagos LGAs. However, this across questions and the data are therefore hard comparison is to be interpreted with caution to interpret. About 43% of the facilities surveyed Figure II.3.2 Composition of Lagos State Revenues Lagos: Composition of 1999 Revenues Lagos: Composition of 2000 Revenues Internal Revenues 11% VAT Revenues Federal Statutory Federal Statutory 24% Allocation Allocation 44% 63% State Statutory Allocation 7% Internal Revenues 8% State Statutory VAT Revenues Allocation 38% 5% Source: Survey Data. Missing data for one LGA (Ojoo) in 1999 Survey Results 37 Figure II.3.3 Figure II.3.4 Kogi ­ Composition of Health Lagos ­ Composition of Health Expenditures, 2000 Expenditures, 2000 KOGI: Composition of Health Expenditures Lagos: Composition of Health Expenditures Capital Expenditure Capital Expenditure 7% Personnel Expenditure 16% 78% Overheads Expenditure Overheads Expenditure 15% 19% Personnel Expenditure 65% Table II.3.3 Local Government Health Expenditure Mean Std. Dev. Minimum Maximum Kogi Tot. Health Exp. Per Capita, 1999 240.7 235.5 92.4 800.2 Tot. Health Exp. Per Capita, 2000 379.5 261.6 191.8 1121 Proportion spent on health, 1999 26% 16% 13% 62% Proportion spent on health, 2000 22% 15% 6% 61% Lagos Tot. Health Exp. Per Capita, 1999 154.2 152.1 48.5 624.8 Tot. Health Exp. Per Capita, 2000 251.2 304 60.2 1162.7 Proportion spent on health, 1999 14% 7% 8% 37% Proportion spent on health, 2000 12% 9% 5% 41% Source: Survey Data. 2000 data is for 11 LGAs in Kogi and 12 in Lagos; 1999 data is for 8 LGAs in Kogi and 14 in Lagos. claim to be non-fee charging facilities. Yet, some prices for treatment. Hardly any facilities (less of these facilities that claimed not to charge fees than 5%) were observed to have permanent dis- for their services respond with non-zero values plays of user charges. 60% of respondents indi- for average charges for services, and "yes" to cate that the facilities do not charge standard the question of whether they charge standard prices for treatment. Only 24% of respondents 38 Decentralized Delivery of Primary Health Services in Nigeria claim that the facility has an exemption policy. the responses were often inconsistent across ques- Almost all surveyed facilities in Kogi reported tions. This provides some lessons for better design- positive fees for two services about which the ing these questions in future surveys, and under- question was asked: outpatient treatment (96%) scores the value of supplementing the survey with and inpatient deliveries (69%). Average fees in exit interviews of patients, particularly on their ex- Kogi for in-patient deliveries is Naira 330, rang- perience with user charges. However, conflicting ing from a minimum of Naira 20 to a maxi- responses at the facility level, and lack of visible mum of Naira 1500. Average fees for outpatient posting of user fees does suggest that fee policies treatment is Naira 116, ranging from a mini- are non-transparent and therefore leave room for mum of Naira 10 to a maximum of Naira 300. staff discretion and corruption. Far fewer facilities in Lagos responded to the question about fees for service (a little over 20%). Average fees for inpatient deliveries in II.4. PHC Staff, Incentives, and Lagos is Naira 572, ranging from a minimum Equipment of Naira 50 to a maximum of Naira 5000, while for outpatient treatment the average fee is Naira Introduction 107, ranging from a minimum of Naira 50 to a maximum of Naira 160. A complete account of public service delivery in Data on fee revenues collected by facilities is primary health care includes at least three caus- available for few facilities, with only 61% of ally related elements: financing and overall gov- Kogi facilities responding with positive rev- ernance, the resulting incentive environment in enues, and as few as 14% of Lagos facilities which staff deliver pharmaceuticals, diagnoses, responding with positive revenues. The Lagos treatment, and the other services that patients facilities reporting revenues are all Type 2 or seek, and the net impact of those services on Type 3 facilities, that is larger, more compre- health care quality, efficiency, and health out- hensive health care centers, but the Kogi facili- comes. The sections above characterized financ- ties reporting revenues are largely Type 1 fa- ing and governance, and the succeeding section cilities, that is, simple health posts or small clin- will assesses outcomes. This section analyzes the ics. For the responding facilities in Kogi, the crucial middle part of the causal chain ­ the work average facility monthly revenue is Naira 1694, of staff and the availability of health equipment. ranging from a minimum of Naira 70 to a maxi- The first part below describes staff in Nigerian mum of Naira 13,333. For the few facilities primary health care, the second analyzes the in- responding in Lagos, the average facility centive environment, and the third examines the monthly revenue is Naira 11,493, ranging from availability of crucial non-personnel inputs. a minimum of Naira 217 to a maximum of Naira 72, 360. Fees collected from users of fa- Staff Characteristics cilities are not reported as a systematic nor sig- nificant source of financing facility resources. Primary health care facilities in Nigeria continue Only about 45% of facilities report that rev- to be staffed by a variety of health care workers enues generated from sale of drugs and treat- organized in a civil service hierarchy. At the top ment provided at the facility may be used for are medical officers, or physicians. Below them general facility purposes. are community health officers (CHOs), nurses, Overall, the survey questions related to user fees midwives, senior and junior community health andfacility-levelfinancialmanagementdidnotyield education workers (SCHEWs and JCHEWs), useful answers--the response rate was low, raising and environmental health officers. A number of the risk of bias in interpreting sample averages, and health care staff in other categories also work Survey Results 39 Table II.4.1 Number of health workers by facility type All Facilities Type 1 Type 2 Type 3 Type 4 N=252 N=131 N=70 N=45 N=1 Designation Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. All 7.85 9.25 2.32 1.67 8.09 5.82 22.76 9.51 31.00 ­ Medical Officers 0.06 0.24 0.00 0.00 0.06 0.23 0.27 0.45 0.00 ­ CHOs 0.82 1.75 0.05 0.23 1.19 1.00 2.38 3.34 2.00 ­ Nurses 0.17 0.91 0.02 0.17 0.20 0.60 0.56 1.94 0.00 ­ Midwives 1.21 2.59 0.04 0.19 0.76 0.89 5.22 3.89 7.00 ­ SCHEWs 1.06 1.13 0.87 0.56 1.09 1.44 1.60 1.62 0.00 ­ JCHEWs 0.38 0.70 0.27 0.48 0.41 0.69 0.67 1.11 1.00 ­ Environmental Health Officer 0.71 1.56 0.01 0.09 0.83 1.57 2.42 2.21 5.00 ­ Other 3.43 4.28 1.07 1.37 3.56 3.13 9.64 4.59 16.00 ­ in the health facilities, including lab technicians, It is important to note that health facility types pharmacy technicians, medical records officers, are also unevenly located across the two states dental assistants, health attendants, and secu- in the sample: 93% of health posts in the sample rity guards. Those staff are grouped as "other" are located in Kogi state while 75% of the re- in the tables below. maining higher level facilities are located in Table II.4.1 shows the number of health care Lagos. As a result, while 61% of all facilities in workers by designation and facility type. In the the sample were in Kogi, 66% of the staff were sample, the average number of staff in a facility from Lagos. Kogi had a mean of 4.0 staff per was 7.85, with about half of these in the cat- facility; in Lagos there was a mean of 13.7 pri- egory of environmental health officers and oth- mary health care staff per facility. As Table II.4.2 ers. The typical facility in the sample had about shows, staff in Lagos have more clinical train- one midwife, one SCHEW, about one nurse, and ing. For example, while nurses make up about no doctors. These averages, however, conceal a 10% of total staff in Kogi, nurses constitute 20% wide variation in staffing levels across types of of all staff in Lagos. Similarly, 7% of Kogi staff facilities. The median number of staff in a facil- are midwives, compared to 26% in Lagos. ity, for instance, was three health care workers. Table II.4.3 below describes the personal Type 1 facilities, which are like health posts, characteristics of health workers in the sample. composed slightly more than half the sample, The average age of staff is 41 years, but doc- and on average had 2.3 workers, usually one tors are younger than the rest of the cadre, with SCHEW and one in the category of "other." an average age of 30 years. A large majority of Medical officers rarely worked in public facili- health staff are women, with exceptions again ties ­ only one in four type 3 facilities had a being doctors (50%) and environmental health physician on staff, and the sole type 4 facility in officers (21%). The large majority of staff in the sample did not have any. almost all categories have some amount of post- 40 Decentralized Delivery of Primary Health Services in Nigeria Table II.4.2: Designation of health staff in the sample, by state Kogi Lagos Designation N percent N percent Medical Officers 2 0.8 8 1.7 CHOs 24 9.9 95 20.0 Nurses 8 3.3 20 4.2 Midwives 16 6.6 125 26.3 SCHEWs 125 51.7 69 14.5 JCHEWs 36 14.9 20 4.2 Environmental Health Officer 2 0.8 75 15.8 Other 29 12.9 63 13.3 Total 242 100 475 100 Table II.4.3 Personal characteristics of staff Age Female Education Indigene Years in PHC LGA Designation N Mean S.D. (%) (%) (%) Mean S.D. (%) All Employees 717 40.5 7.8 70.4 86.5 27.9 14.2 7.1 95.6 Medical Officers 10 29.5 4.2 50.0 100 0.0 2.6 7.4 50.0 CHOs 119 43.9 5.7 78.2 98.3 21.9 16.8 5.3 95.7 Nurses 28 42.1 7.4 78.6 89.3 40.7 12.4 7.0 100 Midwives 141 43.8 6.5 99.3 99.3 16.4 15.6 6.9 96.5 SCHEWs 193 38.8 7.5 66.0 86.6 36.1 14.1 7.0 96.9 JCHEWs 56 35.9 9.4 83.9 67.9 39.3 11.6 7.2 90.9 Environmental Health Officers 77 38.6 7.4 20.8 96.1 14.3 12.8 7.5 97.4 Others 91 39.7 8.5 58.7 52.2 37.8 13.6 7.3 96.7 Education: Completed at least OND/HND degree. LGA: Percentage of staff employed by the LGA secondary education. Only about 28% of staff medical officers in the sample were employed are indigenous to the communities in which by the federal government and half by the LGA. they are working, with percent indigene rang- Tables II.4.4 and II.4.5 below report staff liv- ing from 0% for doctors to 41% for nurses. ing conditions. About 45% of staff supplement Staff have on average 14 years of experience in their income in some way. The most common primary health care, but doctors have relatively sources of supplemental income were less experience, with an average of 2.6 years of agricultural work and commerce. About 17% of work in the field. Almost all staff (96%) are staff reported supplementing their work with employed by the LGA, though half of the 10 some form of health-related activities, including Survey Results 41 Table II.4.4 Percent of staff that supplement salary Agricultural Comm/Petty Clinical Home Sale of Other Work Trade Work Health Serv. Medicines Activities Designation % % % % % % All 14.0 13.3 2.4 8.3 5.8 1.7 Medical Officers ­ ­ 30.0 10.0 ­ ­ CHOs 8.4 12.6 2.5 6.7 5.0 1.7 Nurses 14.3 17.9 ­ 10.7 3.6 ­ Midwives 5.0 17.7 ­ 4.3 3.6 2.1 SCHEWs 23.7 12.4 3.1 11.9 7.7 1.0 JCHEWs 25.0 19.6 1.8 21.4 14.3 5.4 Environmental Health Officer 5.2 10.4 2.6 1.3 1.3 2.6 Other 16.3 8.7 2.2 5.4 6.5 ­ Table II.4.5 Household Condition of Staff Own a Bicycle Own a Car Has Flush Toilet Number of Rooms Designation % % % Mean S.D. All 14.2 43.1 61.1 3.5 1.9 Medical Officers ­ 50.0 100.0 3.5 1.6 CHOs 14.3 62.2 82.4 3.9 1.9 Nurses 10.7 64.3 75.0 3.8 2.2 Midwives 12.9 68.6 87.9 4.2 1.7 SCHEWs 16.6 24.7 30.4 3.3 1.9 JCHEWs 18.2 23.2 33.9 3.0 1.7 Environmental Health Officer 13.3 39.5 80.5 2.8 1.2 Other 10.9 26.4 50.0 3.1 1.9 clinical work, home health care, or the sale of highest paid staff were midwives, CHOS, and medicines. (Staff might have been reluctant to nurses. Doctors, surprisingly, were the lowest paid. reveal the extent of their health-related moon- The reason for the low pay of doctors is likely re- lighting). Some 43% of health staff lated to the fact that doctors were on average more reported ownership of a car in their household, thantenyearsyoungerthantheircolleaguesinother and 61% had flush toilets in their homes. designations.Staffintype1facilitiesearnedlessthan their counterparts in higher level facilities, but no Salaries and Incentives systematic differences in staff salary levels were ap- parent among type 2, type 3, and type 4 facilities. The monthly salaries of health staff were, on aver- In-kind benefits typically did not constitute a age, 26306 Naira (about US$220), in 2001. The large element of an average staff member's 42 Decentralized Delivery of Primary Health Services in Nigeria Table II.4.6 Monthly salary by designation and facility type (naira) All Type 1 Type 2 Type 3 Type 4 Facilities Facilities Facilities Facilities Facilities Designation N Mean S.D. N Mean S.D. N Mean S.D. N Mean S.D. N Mean S.D. All 715 26306 12583 158 18914 9901 226 27055 11449 296 29733 12975 10 26666 12288 Medical Officers 10 18090 8124 ­ ­ ­ 1 21000 ­ 8 18113 9071 ­ ­ ­ CHOs 119 33456 11464 5 28494 10986 58 33297 9314 49 34962 13730 2 25724 2438 Nurses 28 31435 12383 1 22600 ­ 12 30432 13554 15 32826 11952 ­ ­ ­ Midwives 141 35521 10485 2 19775 12410 36 34342 11861 94 36400 9738 3 36507 10007 SCHEWs 191 20334 7258 100 20112 7298 48 20896 7357 40 20617 7297 ­ ­ ­ JCHEWs 56 13694 10651 28 14072 14871 14 13009 3699 13 13594 2578 ­ ­ ­ Environmental Health Officers 77 28085 9506 ­ ­ ­ 31 26739 8557 39 29630 10304 3 28865 11403 Other 91 21021 13351 21 16981 10367 26 21025 10600 38 22949 14549 2 9550 71 reimbursement package: 63% of health staff did decisions can undermine teamwork and overall not receive any in-kind benefits at all. For those morale. Health and education systems in devel- who did receive in-kind benefits, the most com- oping countries usually employ civil service staff, mon benefit was free health care, which 21% of but reform in agendas in several countries are staff received. Free medicines were available to attempting to incorporate more flexibility, in- 15% of staff, and free housing to 18%. Free cluding market-like mechanisms, in the incen- schooling or food items were rare. An analysis tive systems for public sector workers. The de- by state (not shown) found no significant differ- centralization agenda in Nigeria is consistent ence between Kogi and Lagos in receipt of with the reform program. An analysis of the earn- in-kind benefits. ings of primary health care staff sheds light on Given that salaries were the strongest extrin- how decentralization has affected the incentives sic motivation for health staff performance (apart that frontline staff face, and whether, therefore, from under-the-table payments, which were it is likely to have any impact on health care de- probably under-reported), how were salaries de- livery and the health status of the population. termined? The question is crucial for character- A simple Mincerian earnings function was izing the incentive environment that staff face. used to explore the sources of variation in staff Personnel systems can reward staff based on salaries. The results are presented in Table II.4.9 years of experience and civil service grade, or below. In the first estimation, in column (1), ex- they can be more flexible, with localities or fa- perience, experience squared, and education vari- cility managers using their own discretion to re- ables were all significant determinants of monthly ward or punish staff. The advantage of the lat- wages and display the expected signs. Gender, ter system is that local mangers use all informa- being indigenous to the community, and state of tion about staff performance in determining re- work were not significant. Staff in type 2 and wards, not just centrally determined criteria for type 3 facilities had significantly higher wages promotion. The disadvantage is that arbitrary than staff in type 1 facilities. (The omitted Survey Results 43 Table II.4.8 Percent of Staff Receiving In Kind Benefits by designation Health Care Medicines Schooling Housing Food Items Designation % % % % % All 20.7 14.6 0.6 18.1 5.2 Medical Officers 40.0 30.0 ­ 20.0 11.1 CHOs 15.1 10.9 ­ 20.2 5.1 Nurses 21.4 21.4 ­ 14.3 7.1 Midwives 18.4 12.8 ­ 14.9 5.0 SCHEWs 19.8 13.1 1.6 26.4 5.2 JCHEWs 20.0 7.3 ­ 17.9 7.3 Environmental Health Officers 23.4 18.2 ­ 9.2 2.6 Other 29.4 22.8 1.1 10.9 5.6 Table II.4.9 Determinants of Monthly Wages, Robust OLS (1) (2) (3) (4) (5) LGA dummies LGA dummies LGA dummies Grade level 3,008.62 3,037.37 3,077.28 (19.89)** (18.74)** (17.81)** Experience 1,422.65 1,659.81 146.518 (7.05)** (8.33)** ­1.06 Experience squared ­26.356 ­34.439 ­1.846 (3.77)** (5.10)** ­0.42 Secondary school 5,335.81 4,932.59 1,500.39 (2.05)* ­1.91 ­1.15 OND 10,140.29 10,035.91 ­493.993 (4.82)** (4.50)** ­0.47 University 12,449.66 12,102.01 ­5,526.66 (3.70)** (3.58)** (2.07)* Postgrad 18,026.20 18,434.12 3,362.24 (3.60)** (4.74)** (2.35)* Male ­623.515 ­887.099 ­590.876 ­0.77 ­1.12 ­1.24 Indigene ­2,232.12 ­2,108.78 ­485.832 (2.57)* (2.35)* ­0.84 Kogi 2,432.94 2,004.94 ­269.947 (2.02)* ­0.53 ­0.16 T2: Primary 7,148.74 6,565.03 ­42.38 (6.16)** (5.40)** ­0.05 (continued on next page) 44 Decentralized Delivery of Primary Health Services in Nigeria Table II.4.9 Determinants of Monthly Wages, Robust OLS (continued) (1) (2) (3) (4) (5) LGA dummies LGA dummies LGA dummies T3: Comprehensive 9,538.94 8,324.39 1,870.61 (7.09)** (5.80)** ­1.61 T4: Tertiary 6,542.09 7,470.24 3,905.71 (2.33)* (2.28)* (2.27)* Constant ­3,232.56 ­5,645.56 ­2,263.96 ­2,095.70 ­3,651.84 ­1.15 ­1.54 ­1.41 ­1.27 ­1.88 Observations 687 687 700 700 676 R-squared 0.31 0.37 0.6 0.65 0.7 Robust t statistics in parentheses * significant at 5%; ** significant at 1% category is T1: health posts). Adding categori- pay scales, rather than locally determined re- cal variables for LGA, in the estimation in col- wards for performance, remain the dominant umn (2), did not affect the magnitude or signifi- element in the incentive environment for primary cance level of the explanatory variables and in- health care staff. creased the variance explained by about 19%, Ideally, the incentives associated with uniform to an R2 of 0.37. The estimation in column (3) civil service pay scales establish a career path for used current civil service grade level to estimate staff. Opportunities for career advancement and monthly salaries. The estimate of the coefficient learning combined with job stability motivate on civil service grade was significant at 1%, and staff to remain in their positions for long peri- the variable by itself explained 60% of the vari- ods. The availability of promotions, the fact that ance in monthly wages. Adding LGA dummies, colleagues remain in place for long periods, and in column (4), hardly affected the significance the absence of short-term, competitive rewards level of its coefficient. This suggests that despite minimizes adversarial relations among staff and the decentralization that Nigeria has undergone, promotes the conditions for teamwork. For ex- LGAs were not exercising discretion to establish ample, high-powered incentives based on num- local pay rates for health staff, nor were they ber of patients a staff member sees might, while using it to prioritize health care. (Of the 29 LGA encouraging productivity, create incentives for dummies used in the column (4) estimation, only staff to steal patients from each other or hesitate three were significant). The estimation in col- to refer patients to one another. Low-powered umn (5) shows that when civil service grade is incentives, though they do not reward produc- added to the estimation of column (2), the expe- tivity to the same extent, avoid such problems. rience and education variables lose significance. Is the civil service payment system creating con- When a variable for local competition is included ditions for teamwork in Nigeria? Table II.4.10 in the estimations, the number of other health below shows the average length of time that health facilities within a 2 hour walk, the coefficient on staff have been working in their current facility. In the variable is not significant, and the results re- general, average length of tenure in the current fa- main unchanged (not shown). Overall, these es- cility is short, about 2.7 years. Medical officers have timations establish that traditional civil service been working in the current facility for three Survey Results 45 months (given their age, most are probably just training. Table II.4.11 below shows that staff out of medical school), and nurses and midwives spent an average of 7 days in training during have an average tenure in the current facility of the past 12 months. Midwives received more less than two years. Senior and junior health training on average than workers in the other education workers have longer tenures (most are cadres, and doctors received the least. There in Kogi, where staff typically have a longer aver- were no significant differences in time spent age time in the current facility than Lagos). The training across facility types, and no difference average length of time is similar for health posts, between the states (not shown). The most com- primary centers, and comprehensive centers, but mon form of training was for immunization and the one tertiary facility in the sample exhibited an vaccines (not shown). exceptionally high rate of turnover. These data in- Finally, teamwork and collaboration are ex- dicate that there is a lot of staff churning in the hibited in staff behavior and professionalism. health facilities in the two sampled states, and that The questionnaire asked staff a number of ques- the civil service system does not appear to be achiev- tions regarding views of the health system and ing one of its principal objectives ­ facilitating team- health management. Table II.4.12 below reports work and stability in health service delivery. individual and facility averages for three of Another important element of career-based those questions: the number of times in the last professional incentives are opportunities for month another health care worker watched the learning and professional growth, including staff member diagnose and treat a patient for training purposes, the number of patient cases discussed with another health staff member dur- Table II.4.10 ing the last month, and whether, if a staff mem- Average Number of Years Working ber witnessed egregious behavior on the part in Current Health Facility of another staff, he or she would bring it up in Years a staff meeting. Less than half of health profes- sionals would bring up egregious behavior on Individual staff the part of another staff member in staff meet- Medical Officers 0.25 ing. Of all health professionals, medical offic- CHOs 1.9 ers discuss patient cases most frequently (per- Nurses 2.4 haps with each other), but they are least likely to bring up issues in staff meetings (possibly Midwives 1.7 because they do not participate in or value the SCHEWs 3.3 meetings). Staff in Kogi facilities engage in fewer JCHEWs 2.9 professional interactions than staff in Lagos Environmental Health Officers 1.9 facilities, but this is related to the fact that there Others 5.0 are fewer patients and therefore fewer oppor- tunities to do so. Kogi facilities also are less All staff 2.7 inclined to raise issues in staff meetings, per- Facility averages haps because staff sizes are smaller and formal Kogi 3.4 meetings less frequent. Lagos 1.6 Availability of drugs, equipment, and Health posts 2.9 surveillance records Primary centers 2.7 Comprehensive centers 2.4 Many health facilities reported shortages of basic Tertiary facilities 1.0 health equipment. For instance, 95% did not have 46 Decentralized Delivery of Primary Health Services in Nigeria Table II.4.11 Total days spent in training in last year by designation and facility type All Type 1 Type 2 Type 3 Type 4 Facilities Facilities Facilities Facilities Facilities Designation Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. All 7.0 10.6 6.3 10.5 6.6 9.9 7.4 10.9 13.5 17.7 Medical Officers 2.1 6.6 ­ ­ 0.0 ­ 2.6 7.4 ­ ­ CHOs 9.4 12.3 6.6 4. 9.3 13.4 9.9 12.2 12.0 12.7 Nurses 6.0 8.3 17.0 ­ 3.5 4.1 7.2 10.1 ­ ­ Midwives 11.1 14.0 4.5 2.1 10.3 12.6 10.5 13.8 29.0 26.2 SCHEWs 5.7 8.9 6.1 10.3 5.1 6.8 5.5 7.3 ­ ­ JCHEWs 6.2 11.1 7.2 14.4 6.7 7.5 3.8 5.6 ­ ­ Environmental Health Officer 4.0 6.5 ­ ­ 2.7 4.2 5.8 7.9 0.0 0.0 Other 4.0 6.0 6.0 6.9 4.5 5.7 2.6 5.3 12.0 4.2 Table II.4.12 Professional Attitudes Been observed Discussed a case Bring up in staff (Times) (Times) Meeting (%) Individual staff Medical Officers 4.2 4.6 10.0 CHOs 5.8 2.6 50.0 Nurses 6.3 3.7 23.1 Midwives 5.6 3.1 51.9 SCHEWs 3.5 1.6 35.9 JCHEWs 3.0 1.5 40.0 Environmental Health Officers 2.1 1.3 18.5 Others 1.4 1.2 29.6 All staff 4.1 2.2 40.0 Facility averages Kogi 1.9 1.4 29.2 Lagos 4.0 1.9 38.7 Health posts 1.6 1.0 27.6 Primary centers 3.1 1.7 41.2 Comprehensive centers 5.0 2.8 34.9 Tertiary facilities 3.5 2.2 20.0 Survey Results 47 Table II.4.13 Availability of Drugs, Equipment, and Surveillance Records Public- Public- Private Private Facility Privately Kogi Lagos Correlation Correlation Equipment and Drugs Owned Owned Either (Either) (Either) Kogi Lagos Generator 12.8 1.9 12.3 4.6 24.0 0.39 ­ Blood pressure gauge 59.4 52.8 85.3 83.6 88.0 ­0.64 ­0.13 Child weighing scale 67.9 7.4 67.1 61.2 76.0 ­0.10 ­ Microscope 4.2 2.6 5.2 6.6 3.0 ­0.03 ­ Antiseptic 27.2 34.3 46.8 52.6 38.0 ­0.33 ­ Sterile gloves 22.3 31.2 40.9 49.3 28.0 ­0.33 ­0.19 Malaria smear 1.8 0.0 1.6 2.0 1.0 ­ ­ Urine test strip 5.4 0.6 5.2 5.9 4.0 ­0.02 ­ Chloroquine 48.9 37.9 67.9 90.8 33.0 ­0.82 ­0.26 Paracetamol 49.6 39.3 67.1 88.8 34.0 ­.079 ­0.24 Antibiotics 41.2 33.5 58.7 77.0 31.0 ­0.57 ­0.20 ORS sachets 22.5 17.2 31.0 40.1 17.0 ­0.21 ­0.15 Multivitamins 43.1 32.0 59.9 76.3 35.0 ­0.59 ­0.23 BCG vaccine 40.6 0.7 36.9 17.8 66.0 0.25 ­ Measles vaccine 42.4 0.7 38.5 17.8 70.0 0.25 ­ Condoms 29.3 9.7 31.4 32.2 30.0 ­0.07 ­0.09 Kogi Lagos Surveillance records 37.8 94.2 microscopes, 59% did not have sterile gloves, last three months a stock-out of one week or 98% did not have a malaria smear, and 95% did longer for each of the medicines and vaccines listed not have a urine test strip. As noted earlier, the below, as well as for condoms. absence of laboratory equipment and expertise Columns three to seven in the table below com- means that health staff must rely on syndromic pare the availability of key inputs in Kogi and treatment, which can be ineffective for treating Lagos. Lagos facilities were six times more likely malaria and preventing congenital syphilis. Some to have a generator, but Kogi facilities were much 69% of facilities did not have condoms available. more likely to have pharmaceutical products, such The most common items types of privately owned as chloroquine, paracetamol, antibiotics, ORS equipment available were blood pressure gauges, sachets, and multivitamins. A likely explanation antiseptic, sterile gloves, chloroquine, antibiotics, for this is that whereas in Lagos alternative sup- and multivitamins. Additional data (not shown) pliers are available, such as pharmacies, in Kogi demonstrated that when equipment was available, the public clinics effectively function as pharma- it was usually in good working order, and that cies in which health staff sell privately acquired about 40-50% of facilities experienced during the products. Further evidence for this is found in 48 Decentralized Delivery of Primary Health Services in Nigeria columns six and seven, which report correlations ­ is not being achieved. On the day of the sur- between the availability of facility owned and vey, most facilities were missing essential equip- privately owned equipment and drugs at the fa- ment, medications, vaccines, and supplies, and cility level. In Kogi, privately owned, curative stock outs of a week or longer were relatively pharmaceutical products for which there is sub- common. stantial private demand, such as chloroquine and paracetamol, are available whenever they are not provided by the facilities. (Correlations approach II.5 Outputs and Outcomes negative one). It is not clear whether this health staff are responding to shortages in public sup- Services provided and average output of ply, or whether facility owned products are being different types of facility (Table II.5.1 and expropriated. In Lagos, the public-private own- Table II. 5.2): ership correlations for these products are also negative but much smaller. A similar but less pro- As discussed in Section II.1, facilities in Kogi nounced pattern occurs for antiseptic and sterile and Lagos operate under quite different condi- gloves, which also exhibit private good charac- tions. Few facilities in Kogi had access to con- teristics and a negative correlation between pub- venient supplies of water and electricity, while lic and private ownership in Kogi, and less so in those in Lagos were served by much better pub- Lagos. Vaccines were far more likely to be avail- lic infrastructure. The great majority of facili- able in Lagos facilities. That might suggest better ties surveyed in Kogi were health posts/dispen- public provision in Lagos but might also be an saries, while Lagos had substantial proportions artifact of differing delivery schedules in the two of higher-level facilities. Moreover, the facilities states. Finally, the last row shows that whereas in Lagos were geographically proximate to re- 94% of Lagos facilities produced last month's ferral centers, as well as to a range of private report on tracer and notifiable diseases, only 38% facilities, while these were much fewer in Kogi. of Kogi facilities did so. That suggests that the Thus there are very substantial differences in the critical activity of public health surveillance is context in which facility staff function in the much stronger in Lagos than in Kogi. two states. In Kogi, health posts and dispensa- ries have to meet a wider range of health care Summary needs for the population, regardless of the re- sources available to them. By contrast, people The dominant element in the incentive envi- in Lagos have a variety of private facilities avail- ronment for primary health care staff contin- able to them, and need to depend less on public ues to be promotions based on standard civil facilities. service grades. LGAs, despite having assumed The data suggest that health posts/dispensa- responsibility for hiring and paying staff, are ries play a very important role in making health not using their powers to raise or lower the services available to people--and that they seek average pay of their staff based on their expe- to provide whatever services there is a demand rience or educational attainment, in response for, despite their lack of amenities. In both the to competition from other health providers, or states, they provide a very wide range of ser- for any other reason. At the same time, the vices, not much less than larger and better- average number of years that staff have worked equipped facilities. They are fairly similar to in their current facility is low, which means that PHCs in terms of the percentage of facilities pro- there is a lot churning of staff in the system, viding different types of services. The average and that one of the objectives of civil service number of health education sessions provided incentive schemes ­ collaboration and stability is similar across health posts, PHCs and CHCs. Survey Results 49 Table II.5.1 % of facilities providing specific services, by type of facility Health post/ dispensary PHC CHC Tertiary Unspecified All (n=131) (n=70) (n=45) (n=1) (n=5) (n=252) Under-5 consultations 99 97 98 100 80 98 Adult consultations 99 93 98 100 80 97 Antenatal consultations 79 70 98 100 80 80 Postnatal consultations 76 50 87 100 80 71 Family planning services 40 61 93 100 80 56 STI/STD services 45 49 51 100 60 48 Dentistry 10 3 2 ­ 20 7 BCG Immunization 81 80 96 100 80 83 Measles immunization 88 86 98 100 80 89 TT for pregnant women 86 87 98 100 80 88 Inpatient deliveries 64 30 91 100 40 59 Inpatient malaria treatment 72 36 60 100 40 59 Malaria lab tests 1 3 4 100 ­ 2 Anemia blood lab test 1 9 9 100 ­ 5 Health post/ dispensary PHC CHC Tertiary Unspecified All Kogi (n=131) (n=70) (n=45) (n=1) (n=5) (n=252) Under-5 consultations 100.0 100.0 100.0 100.0 100.0 Adult consultations 100.0 100.0 100.0 100.0 100.0 Antenatal consultations 85.3 91.3 100.0 100.0 86.8 Postnatal consultations 81.2 78.3 100.0 100.0 81.6 Family planning services 41.8 73.9 100.0 100.0 49.3 STI/STD services 48.4 47.8 83.3 100.0 50.0 Dentistry 10.7 4.4 0.0 100.0 9.9 BCG Immunization 84.4 87.0 100.0 100.0 85.5 Measles immunization 88.5 87.0 100.0 100.0 88.8 TT for pregnant women 86.1 95.7 100.0 100.0 88.2 Inpatient deliveries 68.9 73.9 100.0 100.0 71.1 Inpatient malaria treatment 75.4 69.6 83.3 100.0 75.0 Malaria lab tests 0.8 0.0 33.3 0.0 2.0 Anemia blood lab test 0.8 17.4 50.0 0.0 5.3 (continued on next page) 50 Decentralized Delivery of Primary Health Services in Nigeria Table II.5.1 % of facilities providing specific services, by type of facility (continued) Health post/ dispensary PHC CHC Tertiary Unspecified All Lagos (n=131) (n=70) (n=45) (n=1) (n=5) (n=252) Under-5 consultations 88.9 95.7 97.4 100.0 75.0 95.0 Adult consultations 88.9 89.4 97.4 100.0 75.0 92.0 Antenatal consultations 0.0 59.6 97.4 100.0 75.0 70.0 Postnatal consultations 0.0 36.2 84.6 100.0 75.0 54.0 Family planning services 11.1 55.3 92.3 100.0 75.0 67.0 STI/STD services 0.0 48.9 46.2 100.0 50.0 44.0 Dentistry 0.0 2.1 2.6 0.0 0.0 2.0 BCG Immunization 33.3 76.6 94.9 100.0 75.0 80.0 Measles immunization 77.8 85.1 97.4 100.0 75.0 89.0 TT for pregnant women 77.8 83.0 97.4 100.0 75.0 88.0 Inpatient deliveries 0.0 8.5 89.7 100.0 25.0 41.0 Inpatient malaria treatment 22.2 19.2 56.4 100.0 25.0 35.0 Malaria lab tests 0.0 4.3 0.0 100.0 0.0 3.0 Anemia blood lab test 0.0 4.3 2.6 100.0 0.0 4.0 The data seem to suggest that health posts/ fewer alternative facilities area available in Kogi, dispensaries are viewed by their clients as seri- but it could also indicate that Kogi facilities seek ous sources of care, and that their proximity to provide good services within the constraints outweighs the possible advantages of PHCs. they face. For example, Kogi health posts pro- Compared with PHCs, a higher percentage of vide a full range of services including antenatal health posts provide in-patient care for malaria and postnatal care, deliveries, and in-patient and deliveries, and postnatal consultations. They malaria treatment, while those in Lagos concen- also do more "home visits"--that is, seeing pa- trate mostly on outpatient consultations (for chil- tients in their homes. This is consistent with the dren and adults) and immunizations. It is espe- fact that health post staff are physically located cially surprising that Lagos PHCs provide little closer to their patients that the staff of PHCs. by way of antenatal consultations, family plan- However, health posts provide fewer antenatal ning, and in-patient deliveries. A substantially consultations per staff member than PHCs, higher proportion of Kogi health posts and PHCs fewer out-patient consultations, and fewer fam- do home visits than those in Lagos. The inter- ily planning consultations. The last point sug- state differences between PHCs are less extreme, gests that family planning is not a top priority but along the same lines. of the government, insofar as no stiff family plan- In Kogi, 10% of health posts provide dentistry ning targets seem to have been issued to local services, which is more than PHCs and CHCs. health personnel. This too suggests that they try to provide what- In Kogi, health posts/dispensaries appear to try ever there is a demand for, and one wonders what especially hard to meet a wide range of local the quality of their dentistry is. In Lagos no health health care needs. This may be partly because posts provide dentistry services, presumably Survey Results 51 Table II.5.2 Average number of outputs (between March-May), by type of facility Health post/ dispensary PHC CHC Tertiary (n= ) (n= ) (n= ) (n=1) Unspecified Antenatal consultations 11 105 219 214 134 Family planning visits 5 52 65 143 20 In-patient deliveries 4 4 26 45 32 BCG immunizations 38 160 257 56 150 Outpatient consultations 56 283 371 326 443 Health educ. group sessions 17 71 214 57 117 Homes visited 32 46 53 63 0 Health post/ dispensary PHC CHC Tertiary Unspecified All Kogi (n=122) (n=23) (n=6) (n=0) (n=1) (n=152) Antenatal consultations 12 188 74 11 46 Family planning visits 6 16 28 15 9 In-patient deliveries 4 8 16 6 BCG immunizations 40 83 121 40 52 Outpatient consultations 46 109 252 65 Health educ. group sessions 14 10 15 13 Homes visited 33 64 25 38 Health post/ dispensary PHC CHC Tertiary Unspecified All Lagos (n=9) (n=47) (n=39) (n=1) (n=4) (n=100) Antenatal consultations 0 38 242 214 176 157 Family planning visits 0 68 71 143 22 65 In-patient deliveries 0 0 28 45 32 18 BCG immunizations 15 197 279 56 186 218 Outpatient consultations 203 368 390 326 443 365 Health educ. group sessions 74 115 243 57 117 170 Homes visited 21 32 58 63 0 43 52 Decentralized Delivery of Primary Health Services in Nigeria Table II.5.3 Facility-Level Average Output per Staff in Categories 1­7 (March­May 2002) All Type 1 Type 2 Type 3 Type 4 Type Diagnostic/Procedure Facilities Facilities Facilities Facilities Facilities Unspecified Ante-Natal Consultations 15.7 10.5 22.2 18.8 14.3 21.5 Family Planning Visits 6.4 4.5 10.7 4.7 9.5 6.0 In-Patient Deliveries 2.8 3.7 2.1 1.9 3.0 4.6 BCG Immunizations 33.0 33.3 39.0 26.0 3.7 28.3 Out-Patient Consultations 51.5 46.4 69.4 38.8 21.7 62.9 Health Education (Group Sessions) 15.1 15.0 15.8 14.8 3.8 16.6 Home Visits 20.1 28.8 11.5 7.1 4.2 0.0 All Type 1 Type 2 Type 3 Type 4 Type Kogi Facilities Facilities Facilities Facilities Facilities Unspecified Diagnostic/Procedure Mean Mean Mean Mean Mean Mean Ante-Natal Consultations 15.7 11.0 33.7 22.6 11.0 Family Planning Visits 5.0 4.8 3.8 8.6 15.0 In-Patient Deliveries 3.7 3.9 4.0 1.7 BCG Immunizations 36.9 35.2 37.6 54.6 40.0 Out-Patient Consultations 42.7 40.3 38.9 95.8 Health Education (Group Sessions) 10.7 12.8 4.1 3.9 Home Visits 26.0 29.7 13.3 7.5 All Type 1 Type 2 Type 3 Type 4 Type Lagos Facilities Facilities Facilities Facilities Facilities Unspecified Diagnostic/Procedure Mean Mean Mean Mean Mean Mean Ante-Natal Consultations 15.6 0.0 12.7 18.2 14.3 25.1 Family Planning Visits 7.8 0.0 13.7 4.2 9.5 3.0 In-Patient Deliveries 1.4 0.0 0.3 2.0 3.0 4.6 BCG Immunizations 28.4 5.3 39.6 21.4 3.7 24.3 Out-Patient Consultations 64.2 135.5 84.3 29.6 21.7 62.9 Health Education (Group Sessions) 22.4 63.4 24.3 16.5 3.8 16.6 Home Visits (Houses) 8.6 12.5 10.0 7.0 4.2 0.0 Survey Results 53 Table II.5.4 Tracer and Immediately Notifiable Diseases, percentage of facilities by facility type All Type 1 Type 2 Type 3 Type 4 Type Facilities Facilities Facilities Facilities Facilities Unspecified (N=252) (N=131) (N=70) (N=45) (N=1) (N=5) Keep Monthly Records 78.6 71.8 84.3 91.1 100.0 60.0 Forward Monthly Records to LGA 76.6 68.7 82.9 91.1 100.0 60.0 Showed to Interviewer 48.8 28.2 61.4 86.7 100.0 60.0 All Type 1 Type 2 Type 3 Type 4 Type Facilities Facilities Facilities Facilities Facilities Unspecified Kogi (N=152) (N=122) (N=23) (N=6) (N=0) (N=1) Keep Monthly Records 73 70 87 83 0 Forward Monthly Records to LGA 70 67 87 83 0 Showed to Interviewer 28 25 35 67 0 All Type 1 Type 2 Type 3 Type 4 Type Facilities Facilities Facilities Facilities Facilities Unspecified Lagos (N=100) (N=9) (N=47) (N=39) (N=1) (N=4) Keep Monthly Records 87 89 83 92 100 75 Forward Monthly Records to LGA 86 89 81 92 100 75 Showed to Interviewer 81 78 74 90 100 75 because people can obtain these services elsewhere health posts where staff reside on the premises in Lagos. On the whole, dentistry services are little or very nearby. provided by the facilities surveyed. Almost all CHCs provide immunization, and The services provided in different types of the majority (80­87%) of PHCs and health posts facilities show a pattern consistent with the rela- do this as well. However, output per staff is lower tive advantages of lower-level facilities in terms in tertiary facilities for immunizations, as well of proximity to their patients. For example, the as for outpatient consultations, health educa- average number of home visits per staff declines, tion sessions, and home visits. PHCs provide the higher the type of facility. Only 30% of PHCs more outpatient consultations and family plan- compared with 64% of health posts/dispensa- ning per staff person, than other types of facili- ries conduct in-patient deliveries, and similar ties. The average number of antenatal consulta- figures prevail for in-patient malaria treatment. tions per staff person rises with type of facility. One possible reason for this might be that staff The proportion of facilities keeping monthly do not stay overnight in these facilities, unlike records and forwarding them to the LGA rises 54 Decentralized Delivery of Primary Health Services in Nigeria with type of facility, from around 70% of health conducted laboratory tests for malaria. Another posts to 90% of CHCs (Table II.5.4). A very low major concern, especially in the context of the proportion of health posts actually showed the AIDS epidemic as well as overall reproductive records to the interviewer, so it is difficult to as- health, is that only half of facilities (of any given sess the validity of their statement that they keep type) provide STI/STD care. and forward the records regularly. It is notable that high proportions of Lagos facilities of all Work done by the different categories of types (74­81%) showed the records to the inter- staff (Table II.5.5 to II.5.8): viewer. The performance of Kogi facilities is far poorer on this score, with only 25­35% of fa- Most tasks are done by all the grades of staff. cilities showing the records. Apparently disease Nurse-midwives are the work-horses, a much surveillance works more effectively in Lagos than higher percent of them than other staff do de- in Kogi--this could be another indication of good liveries, immunizations, antenatal care, and fam- public infrastructure as opposed to interest in pro- ily planning. High percentages also do out- viding good public clinical services. patient care and health education. As mentioned in chapter II.1 a major concern Everyone, and especially nurses, do non-health is that laboratory testing is virtually non-exis- duties. A higher proportion of Community Health tent in the facilities surveyed. This is of especial Officers report doing administrative work than concern for the effectively responding to the high others, except in Kogi where a similar percent of burden of morbidity from endemic malaria in nurse-midwives do administrative tasks. Lower Nigeria. While 70-83% of facilities of different grade staff do most things (including adminis- types treated malaria cases, only 1-4% of them tration), but with an emphasis on outpatient care Table II.5.5 Percent of staff performing various duties during the past week, by category of staff Nurse/ EnvHlth Lower Unknown MO CHO Nurse midwife SCHEW JCHEW Officer grades n=2 Outpatient care 100 95 86 84 91 80 ­ 53 100 Deliveries 10 12 18 35 10 13 ­ 7 50 In-patient care 30 28 39 45 33 25 ­ 15 ­ Immunizations 40 75 43 78 43 54 1 25 ­ Ante-natal care 30 50 46 63 43 39 ­ 15 50 Family planning 20 38 25 57 17 21 1 14 100 Health education 70 90 93 92 74 61 86 51 50 San. inspections/ home visits ­ 48 50 42 54 52 94 15 50 Laboratory exams 10 1 4 1 1 0 ­ 11 ­ Administration 20 61 43 45 38 21 36 23 100 Assigned non- health duties 10 9 25 9 9 7 13 11 ­ (continued on next page) Survey Results 55 Table II.5.5 Percent of staff performing various duties during the past week, by category of staff (continued) Nurse/ EnvHlth Lower Unknown Kogi MO CHO Nurse midwife SCHEW JCHEW Officer grades n=2 Outpatient care 100.0 91.7 75.0 87.5 86.4 75.0 58.6 100.0 Deliveries 50.0 16.7 100.0 50.0 14.4 19.4 100.0 13.8 50.0 In-patient care 100.0 54.2 37.5 31.3 35.2 36.1 27.6 Immunizations 50.0 50.0 25.0 68.8 26.4 36.1 31.0 Ante-natal care 50.0 58.3 50.0 68.8 52.0 47.2 24.1 50.0 Family planning 50.0 25.0 25.0 68.8 15.2 25.0 17.2 100.0 Health education 50.0 66.7 75.0 81.3 68.0 52.8 100.0 41.4 50.0 San. inspections/ home visits 41.7 37.5 37.5 55.2 50.0 100.0 24.1 50.0 Laboratory exams 50.0 12.5 0.8 17.2 Administration 62.5 25.0 62.5 48.0 22.2 37.9 100.0 Assigned non- health duties 16.7 6.3 11.2 5.6 3.5 Nurse/ EnvHlth Lower Unknown Lagos MO CHO Nurse midwife SCHEW JCHEW Officer grades n=2 Outpatient care 100.0 95.8 90.0 84.0 98.6 90.0 94.7 50.8 Deliveries 10.5 25.0 32.8 1.5 3.2 In-patient care 12.5 21.1 40.0 46.4 29.0 5.0 9.5 Immunizations 37.5 81.1 50.0 79.2 72.5 85.0 1.3 22.2 Ante-natal care 25.0 48.4 45.0 62.4 27.5 25.0 97.3 11.1 Family planning 12.5 41.1 25.0 56.0 20.3 15.0 1.3 12.7 Health education 75.0 95.8 100.0 93.6 84.1 75.0 85.3 55.6 San. inspections/ home visits 49.5 55.0 42.4 50.7 55.0 93.3 11.1 Laboratory exams 1.1 0.8 1.5 7.9 Administration 25.0 60.0 50.0 43.2 18.8 20.0 37.3 15.9 Assigned non- health duties 12.5 7.4 35.0 9.6 5.8 10.0 13.3 14.3 and health education. JCHEWs and SCHEWs do some administration). Some interstate dif- show a similar pattern of tasks. ferences are apparent: it is in Kogi that The only exception is the Environmental Environmental Health Officers apply themselves Health Officers, who tend to specialize in sani- to their assigned tasks of inspections and health tary inspections and health education (and also education sessions, although they also report 56 Decentralized Delivery of Primary Health Services in Nigeria Table II.5.6 Number of Days Worked Last Week, by category of staff Designation N Mean S.D. All 716 5.03 0.91 Medical Officers 10 4.00 1.56 CHOs 119 4.97 1.04 Nurses 28 4.79 1.17 Midwives 141 4.95 0.72 SCHEWs 194 5.18 1.02 JCHEWs 55 5.27 0.73 Environmental Health Officer 77 4.94 0.30 Other 90 5.08 0.86 Table II.5.7 Patients Seen Outside Facility per Week, by category of staff Designation ALL KOGI LAGOS All 1.73 1.69 1.75 Medical Officers 7.83 2.00 9.00 CHOs 1.78 1.89 1.75 Nurses 1.67 2.40 1.44 Midwives 2.48 2.80 2.44 SCHEWs 1.66 1.41 2.16 JCHEWs 1.76 2.25 0.69 Environmental Health Officer 0.10 1.00 0.05 Other 0.82 1.13 0.67 doing deliveries. In Lagos, Environmental Of- average 4 days a week. Doctors do the most ficers seem to dabble in many kinds of task, and "moonlighting": 90% of doctors compared with not all of them report working on their 50-60% of other categories of staff report see- scheduled tasks. ing patients outside the facility. It is noteworthy Sanitary inspections are in fact conducted not that even lower grades of staff (grades 8 and only by the designated Environmental Health Of- below) see patients outside the facility. Only En- ficers, but also by around half the staff from the vironmental Health Officers report doing little other categories. Doctors are the only category of this. Doctors report seeing an average of 8 of higher staff who don't do this, and this could patients a week outside the facility, while other be because of the relative paucity of doctors with categories of staff report seeing around 2 or less consequent high patient demand for their services. per week. Of course, it is extremely probably Most categories of staff reportedly worked that these figures are heavily under-reported. All around 5 days a week during the recall categories of staff report having very little of the period, except doctors who reported working on equipment they need. Survey Results 57 Table II.5. 8 Percent of staff with various attributes, by category of staff Nurse/ EnvHlth Lower Unknown MO CHO Nurse midwife SCHEW JCHEW Officer grades n=2 Adequate equipment 10 11 4 6 5 14 1 4 0 Seen patients outside 50 31.93 28.57 27.66 28.87 32.14 2.6 10.87 50 Nurse/ EnvHlth Lower Unknown Kogi MO CHO Nurse midwife SCHEW JCHEW Officer grades n=2 Adequate equipment 50 4 0 6 6 14 0 3 0 Seen patients outside 50 45.83 25 62.5 32.8 41.67 50 13.79 50 Nurse/ EnvHlth Lower Unknown Lagos MO CHO Nurse midwife SCHEW JCHEW Officer grades n=2 Adequate equipment 0 13 5 6 3 15 1 5 Seen patients outside 50 28.42 30 23.2 21.74 15 1.33 9.52 Sanitary Inspections and Immunizations: water-source inspections. This may be partly due Outputs reported at the LGA level by the to the greater availability of piped water in Lagos, PHCoordinator (Tables II.5.9 to II.5.11): but as Table II.1. 4 shows, only 33-44% of health facilities are served by piped water--implying There is an impressive range of sanitary inspec- that the majority of households are dependent tions conducted in Nigeria. 70% of LGAs were on non-piped water. This implies that Lagos reported to have undertaken food vendor certi- LGAs may be considerably less conscientious fication in the past year, and all conducted most than those in Kogi in assuring the public's health. of the prescribed forms of sanitary inspection: An obvious issue with sanitary inspections is of public water sources, of markets, house-to that of corruption. It may be expected that the house inspections for public health nuisances, individual contacts involved in house and food and inspection of food sellers. seller inspections lend themselves to corruption, The data seem to indicate a fairly high level unless there are effective measures in place to of participation in the sanitary inspection pro- protect citizens from inspector's demands. On gram, as indicated also by the reports of staff the other hand, there may be fewer opportuni- activity discussed above. All LGAs had con- ties for graft when inspecting public facilities ducted inspections of private homes and indi- such as water sources, or when inspecting mar- vidual traders (food sellers) during the three kets where attempts at extortion would be wit- months preceding the survey, with the exception nessed by many. of one in Kogi which had not done food seller It is difficult to interpret the data without inspections. All but one LGA in Kogi had con- knowing the denominators involved, in terms ducted inspections of public water sources. In of how many public water sources, markets, Lagos, nearly half of LGAs had not done public houses, and food sellers there are to be inspected 58 Decentralized Delivery of Primary Health Services in Nigeria in each LGA. To provide some crude standard- that the health care system in Nigeria is orga- ization for the denominators and tease out some nized more in the "regular functioning" mode indication of whether graft exists in the system, than in the "campaign mode" of programs such we take the ratio of private (houses and food as the National Immunization Drive. For ex- sellers) to public inspections (public water ample, 37% of the LGAs sampled did not carry sources and markets). These ratios are sugges- out polio (NID) immunization during the pre- tive of some graft in some LGAs in Lagos, which ceding three months, and an additional 7% put seem to conduct an unusually high number of in a token effort with less than 1000 immuniza- private inspections relative to their public ones. tions conducted. Given that at least three doses Of course, this could be simply that these LGAs of polio vaccine need to be administered per genuinely have few public facilities to inspect child, this amounts to a fairly paltry figure of relative to the number of households or food around 300 children covered in the entire LGA. sellers in the LGA, but this could also be further By contrast, only 3% of LGAs failed to partici- indication that Lagos LGAs are less assiduous pate in the routine program of administering than they could be in assuring the public health. BCG during the recall period, and another 17% However, there is no such indication of graft of LGAs put in what is clearly just a token ef- in Kogi--where on the contrary, the ratios of fort of vaccinating less than 250 children dur- private to public inspections are low. This sug- ing the period. The remainder participated quite gests that LGAs give considerably greater atten- actively in the immunization program. tion to public inspections, than to private in- There are clear interstate differences in the spections (particularly of homes). The reasons implementation of the immunization programs, for this should be investigated. with Kogi performing much less effectively than Data on immunization were also collected at Lagos in this regard. As much as 60% of the LGA level, for the three months preceding the LGAs in Kogi failed to participate in the polio survey. Two types of data were collected: one NID during the recall period, and 13% put in a on special immunization drives, represented by token effort with less than 1000 immunizations the intensive campaign for polio immunization conducted. By contrast, only 13% of LGAs in during the National Immunization Drive; and Lagos failed to participate in the NID, and all the other on routine immunization, represented the rest put in more than a token effort. In the by the administration of BCG vaccine. routine BCG immunization, only 7% of LGAs It appears that, despite the high-profile pres- in Kogi failed altogether to participate, but an- sure of the National Immunization Drive, that other 27% put in just a token effort. In Lagos, the routine immunization effort may be more all LGAS participated, and all but 7% put in effective in some ways--which in turn suggests more than a token effort. Survey Results 59 Cum. 60 80 83.3 90 100 Cum. ­ ­ ­ 93.4 100.0 page) % % Ratio % 60 20 3.3 6.7 10 next Ratio % 93.4 6.7 ­ ­ ­ on N 18 6 1 2 3 N 14 1 ­ ­ ­ Inspection (Private/Public) Inspection (Private/Public) (continued Ratio <10 10­49 50­99 Ratio <10 10­49 50­99 100­199 200­299 100­199 200­299 Cum. 3.3 36.7 46.7 60.0 73.3 86.7 96.7 96.7 Cum. 6.7 ­ ­ ­ ­ ­ 100.0 73.4 86.7 100.0 % % % 3.3 33.3 10.0 13.3 13.3 13.3 10.0 0.0 3.3 % 6.7 ­ ­ ­ ­ ­ 66.7 13.3 13.3 Sellers Sellers N 1 10 3 4 4 4 3 0 1 N 1 2 2 ­ ­ ­ ­ ­ Food Inspections Food Inspections 10 Value 0 <20 20­49 50­99 5000+ 100­499 500­999 Value 0 <20 50 100 500 1000 2500 5000 5000+ 1000­2499 2500­4999 Cum. 0.0 13.3 23.3 56.7 60.0 60.0 70.0 90.0 Cum. 0.0 ­ ­ ­ ­ 100.0 20.0 33.3 93.4 100.0 2002 % % visits visits % 0.0 13.3 10.0 33.3 3.3 0.0 10.0 20.0 10.0 % 0.0 20.0 13.3 60.0 6.7 ­ ­ ­ ­ nuisances nuisances N 0 4 3 10 1 0 3 6 3 N 0 3 2 9 1 ­ ­ ­ ­ March-May health health House-to-house for House-to-house for LGA, Value 0 <20 20­49 50­99 100-499 5000+ 500­999 Value 0 <20 20­49 50­99 5000+ 100­499 500­999 the 1000­2499 2500­4999 1000­2499 2500­4999 in Cum. 0.0 6.7 56.7 66.7 73.3 83.3 96.7 100.0 Cum. 0.0 3.3 40.0 43.3 46.7 50.0 50.0 50.0 % % % 0.0 6.7 50.0 10.0 6.7 10.0 13.3 3.3 % 0.0 3.3 36.7 3.3 3.3 3.3 0.0 0.0 conducted Market N 0 2 15 3 2 3 4 1 Market N 0 1 11 1 1 1 0 0 Inspections Inspections Inspections Value 0 <10 10­19 20­29 30­49 50­99 150+ 100­149 Value 0 <10 10­19 20­29 30­49 50­99 150+ 100­149 Cum. ­ ­ Sanitary 26.7 46.7 56.7 63.3 70 90 93.3 100 Cum. 6.7 46.7 66.7 73.4 80.0 100.0 % % of ater % 26.7 20 10 6.6 6.7 20 3.3 6.7 ater % 6.7 40.0 20.0 6.7 6.7 ­ ­ 20.0 W W types Inspections Inspections and N 8 6 3 2 2 6 1 2 N 1 6 3 1 1 3 ­ ­ II.5.9 Public Public Source Source Table Number Value 0 <10 <10 10­19 20­29 30­49 50­99 150+ 100­149 Kogi Value 0 10­19 20­29 30­49 50­99 150+ 100­149 60 Decentralized Delivery of Primary Health Services in Nigeria Cum. 26.7 60.0 66.7 80.0 100.0 % Ratio % 26.7 33.3 6.7 13.3 20.0 N 4 5 1 2 3 Inspection (Private/Public) Ratio <10 10­49 50­99 100­199 200­299 Cum. 0.0 0.0 6.7 20.0 46.7 73.3 93.3 93.3 100.0 % % 0.0 0.0 6.7 13.3 26.7 26.7 20.0 0.0 6.7 Sellers N 0 0 1 2 4 4 3 0 1 Food Inspections Value 0 <20 20­49 50­99 5000+ 100­499 500­999 1000­2499 2500­4999 (continued) Cum. 0.0 6.7 13.3 20.0 20.0 20.0 40.0 80.0 100.0 2002 % visits % 0.0 6.7 6.7 6.7 0.0 0.0 20.0 40.0 20.0 nuisances N 0 1 1 1 0 0 3 6 3 March-May health House-to-house for LGA, Value 0 <20 20­49 50­99 5000+ 100­499 500­999 the 1000­2499 2500­4999 in Cum. 0.0 6.7 33.3 46.7 53.3 66.7 93.4 100.0 % % 0.0 6.7 26.7 13.3 6.7 13.3 26.7 6.7 conducted Market N 0 1 4 2 1 2 4 1 Inspections Inspections Value 0 <10 10­19 20­29 30­49 50­99 150+ 100­149 Cum. Sanitary 46.7 46.7 46.7 53.3 60.0 80.0 86.7 100.0 % of ater % 46.7 0.0 0.0 6.7 6.7 20.0 6.7 13.3 W types Inspections and N 7 0 0 1 1 3 1 2 II.5.9 Public Source Table Number Lagos Value 0 <10 10­19 20­29 30­49 50­99 150+ 100­149 Survey Results 61 Table II.5.10 Immunization during special drives and on routine basis Polio Immunization during NID BCG Imm. (Children) under Routine Imm. (April-May) Value N % % Cum. Value N % % Cum. 0 11 36.7 36.7 0 1 3.3 3.3 <1000 2 6.7 43.3 <250 5 16.7 20.0 20000 3 10.0 53.3 500 5 16.7 36.7 50000 4 13.3 66.7 1000 3 10.0 46.7 100000 1 3.3 70.0 1500 4 13.3 60.0 200000 6 20.0 90.0 2000 4 13.3 73.3 300000 1 3.3 93.3 4000 5 16.7 90.0 300000+ 2 6.7 100.0 4000+ 3 10.0 100.0 30 30 Kogi Polio Immunization during NID BCG Imm. (Children) under Routine Imm. (April-May) Value N % % Cum. Value N % % Cum. 0 9 60.0 60.0 0 1 6.7 6.7 <1000 2 13.3 73.3 <250 4 26.7 33.3 20000 2 13.3 86.7 500 5 33.3 66.7 50000 2 13.3 100.0 1000 2 13.3 80.0 100000 0 0.0 100.0 1500 1 6.7 86.7 200000 0 0.0 100.0 2000 1 6.7 93.3 300000 0 0.0 100.0 4000 0 0.0 93.3 300000+ 0 0.0 100.0 4000+ 1 6.7 100.0 15 15 Lagos Polio Immunization during NID BCG Imm. (Children) under Routine Imm. (April-May) Value N % % Cum. Value N % % Cum. 0 2 13.3 13.3 0 0 0.0 0.0 <1000 0 0.0 13.3 <250 1 6.7 6.7 20000 1 6.7 20.0 500 0 0.0 6.7 50000 2 13.3 33.3 1000 1 6.7 13.3 100000 1 6.7 40.0 1500 3 20.0 33.3 200000 6 40.0 80.0 2000 3 20.0 53.3 300000 1 6.7 86.7 4000 5 33.3 86.7 300000+ 2 13.3 100.0 4000+ 2 13.3 100.0 62 Decentralized Delivery of Primary Health Services in Nigeria CHAPTER 3 Emerging Issues T he survey evidence confirms what one health facilities, and 2) pervasive non-payment may have expected about the contrast- of salaries of staff providing services in the health ing environments for service delivery facilities. Hence, while the incentive environment in the two states under study, Lagos for public delivery of primary health services in and Kogi. In the largely urban and densely popu- Lagos is influenced by the availability of private lated environment of Lagos there is a signifi- facilities and proximity to referral centers in the cant private market even for primary health ser- state, the incentive environment in Kogi is char- vices, whereas in mostly rural Kogi, with dis- acterized by local government monopoly over persed settlements, health services appear to be health service provision, and community par- largely provided in public facilities.4 In Lagos, ticipation through local institutions. In this sec- the majority of all health facilities is privately tion we provide some preliminary analysis of owned (61%) whereas in Kogi only 7% belong the impact of community participation on the to the private sector. Furthermore, of all the performance of health facilities in Kogi, and the health facilities in Kogi recorded by the local issue of non-payment of staff salaries and its governments, 48% are recorded as health posts potential impact on service delivery.5 or dispensaries (Type 1 facilities) providing the most basic of health services. In contrast, the majority of health facilities in Lagos (69%) are recorded as primary health centers (Type 2 fa- cilities). These differences in the market for pri- mary health services in the two states are re- 4This analysis only refers to services provided flected in differences in staffing patterns and ser- through health facilities, and is deduced from the vices provided, as discussed in previous sections. data obtained on the population of registered The survey finds two additional striking fea- health facilities, by ownership, from the local gov- tures of service delivery in Kogi that are distinct ernment authorities. 5Future studies that include private facilities in the from Lagos and could potentially have substan- surveyed sample can fruitfully address the issue tial impact on the quality of health services pro- of ownership and performance in health service vided--1) extensive participation by community delivery which appears to be particularly impor- development committees in the functioning of tant in the context of service delivery in Lagos. 63 64 Decentralized Delivery of Primary Health Services in Nigeria III.1. Impact of community participation of service delivery by communities in Kogi, par- on facility performance in Kogi ticipation is in fact lacking in the key areas that would be critical for improving performance of There is a large and growing body of evidence facilities. With this caveat in mind, we under- that certain types of service delivery are en- took some analysis of the impact of community hanced with the active participation of the com- participation in Kogi on various performance munities they serve. As end-users of the services, indicators at the facility level--productivity of communities have a stake in ensuring that ser- staff (as measured by numbers of patients seen vices are well-provided, and also are well-posi- for various conditions per staff in the facility), tioned to monitor the quality of services. With record-keeping for public health surveillance, the benefit of local information, they can assess cleanliness and general maintenance of the fa- the specific obstacles facing facilities in provid- cility, and availability of essential drugs and ing services And they can seek to ensure that equipment. facilities have the necessary infrastructure, sup- We construct an indicator variable for the plies and staff motivation to provide the services extent of community participation in the func- they are supposed to provide. Some of this can tioning of a facility which equals 1 if the facility be done through volunteer efforts, such as do- head responds that the community is the princi- nations for buying supplies, but most of the ben- pal decision-maker in one or more of the fol- efits of community participation can only be lowing areas--making drugs, medical supplies harnessed if there are specific mechanisms in available, acquiring and maintaining equipment, place to enable them to do so. For example, setting charges for drugs, and determining the whether or not they are allowed to raise local use of facility revenues--and equals 0 otherwise. resources will affect their ability to ensure a About 145 facilities in Kogi, out of a total sample smooth flow of supplies. Similarly, whether or of 152, systematically responded to these ques- not they have a say in the evaluation and re- tions about principal decision-making over vari- wards/sanctioning of facility staff will affect the ous areas of facility functioning. Of these 145 extent to which they are able to translate their facilities, communities were indicated as princi- observation of staff behavior into improved staff pal decision-makers in one or more areas by 48 responsiveness to local needs. facilities, that is by 33% of the respondents. The In Section II.2 we described how community remaining facilities in Kogi indicated the local participation in primary health care service de- government as the principal decision-maker, livery has been institutionalized in Nigeria with some decisions determined by the facility through the creation of Village Development head or staff. Committees and District Development Commit- Table III.1.1 presents multivariate regression tees. These community organizations are par- estimates of the impact of community partici- ticularly active in Kogi in the areas of building pation, as defined above, on productivity as maintenance and acquiring drugs, medical sup- measured by the number of patients seen in plies, and equipment in the facilities. There is the last three months before the survey per comparatively little community engagement in staff--for antenatal care (Column 1), in-pa- setting charges for drugs, as was envisioned by tient deliveries (Column 2), BCG immuniza- the Bamako Initiative and almost negligible in tions (Column 3), out-patient consultations disciplining staff, which is overwhelmingly in- (Column 4), and home visits (Column 5). The dicated as the responsibility of local govern- point estimates suggest that facilities with ments. It may therefore be that despite the insti- greater community participation provide tutionalization of community participation in greater services of each type, although the ef- Nigeria, and active engagement in some areas fect is statistically significant at conventional Emerging Issues 65 Table III.1.1 Impact of Community Participation on Facility Productivity (1) (2) (3) (4) (5) Antenatal In-Patient BCG Outpatient Home Care Deliveries Immunizations Consultations Visits Community Participation 6.24 3.20** 31.79* 18.63* 1.94 Indicator Variable (5.26) (1.52) (20.15) (12.15) (9.44) Facility Type Indicator 13.71 0.89 13.51 9.00 ­13.60* Variable (9.28) (1.84) (18.98) (12.29) (7.67) Distance from LGA ­0.004 0.001 0.20 0.32 0.03 Headquarters (0.13) (0.02) (0.27) (0.20) (0.16) Number of facilities in ­0.35 ­0.09* ­1.68* ­0.69 ­0.16 the neighborhood (0.24) (0.05) (0.91) (0.44) (0.58) LGA population, 1999 0.000 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) LGA Internally Generated 1.02 0.05* ­0.24 ­0.12 ­0.12 Revenues Per Capita, 2000 (0.82) (0.03) (0.37) (0.21) (0.21) Constant ­16.91 ­2.23 31.69 11.40 18.97 (22.27) (2.46) (38.03) (22.15) (25.47) No. of Observations 98 82 91 120 112 R-sq 0.25 0.20 0.08 0.12 0.05 OLS regressions with robust standard errors (in parentheses). * Significant at 10% level; ** Significant at 5% level; *** Significant at 1% level levels only for inpatient deliveries, at the 1% 2), availability of equipment and supplies such level, and for immunizations and out-patient as blood pressure gauge (Column 3), antiseptic consultation at the 10% level. The estimated (Column 4) and sterile gloves (Column 5). Al- effects are large--facilities with greater com- though the point estimate suggests that the fa- munity participation have 3 additional deliv- cility is less likely to be clean if there is greater eries per staff in the last three months, which community participation, it is not statistically is about the average number of deliveries in significant. Record keeping for public health the sample; they have 32 more vaccinations surveillance is significantly less likely in facili- per staff, which is about the average number ties with greater community participation. Point of vaccinations in the sample; they have 19 estimates suggest that essential equipment are more outpatient consultations, which is about less likely to be not-available, but none are sta- half the average number of outpatient consul- tistically significant. tations in the sample. Table III.1.3 presents the estimated impact Table III.1.2 presents the estimated impact of on availability of essential drugs--chloroquine community participation on general facility (Column 1), paracetamol (Column 2), and cleanliness (Column 1), public health surveil- antibiotics (Column 3). Again, the point esti- lance through the keeping of records (Column mates suggest that these drugs are less likely to 66 Decentralized Delivery of Primary Health Services in Nigeria Table III.1.2 Impact of Community Participation on General Facility Characteristics (2) (3) (4) (5) (1) 1=Facility 1=Blood 1=Skin 1=Sterile 1=Facility Keeps Health Pressure Gauge Antiseptic Gloves is Clean Records Not Available Not Available Not Available Community Participation ­0.04 ­0.21*** ­0.06 ­0.09 ­0.07 Indicator Variable (0.09) (0.07) (0.07) (0.09) (0.09) Facility Type Indicator 0.20*** 0.06 ­0.17** ­0.16 0.12 Variable (0.07) (0.11) (0.07) (0.12) (0.12) Distance from LGA ­0.001 ­0.003*** 0.000 ­0.001 0.000 Headquarters (0.001) (0.001) (0.001) (0.002) (0.002) Number of facilities in 0.000 0.000 ­0.005 ­0.004 0.004 the neighborhood (0.006) (0.005) (0.004) (0.006) (0.006) LGA population, 1999 ­0.000 ­0.000 ­0.000 0.000 ­0.000 (0.000) (0.000) (0.000) (0.000) (0.000) LGA Internally Generated 0.004** 0.004** ­0.001 0.000 ­0.002 Revenues Per Capita, 2000 (0.002) (0.002) (0.002) (0.003) (0.003) Constant 0.69*** 0.24* 0.33*** 0.58*** 0.63*** (0.16) (0.13) (0.12) (0.16) (0.16) No. of Observations 144 144 144 144 144 R-sq 0.08 0.16 0.04 0.02 0.03 OLS regressions with robust standard errors (in parentheses). * Significant at 10% level; ** Significant at 5% level; *** Significant at 1% level be not-available with greater community par- for income levels in the neighborhood of a ticipation, but the results are not statistically facility. The estimated impact of community significant. participation on the outcomes is robust to the In estimating the effect of community par- exclusion of these control variables. ticipation, we control for the type of facility In summary, the most striking result is that (whether health post/dispensary or a primary community participation is significantly associ- health center), the distance of the facility from ated with greater productivity per staff in provid- local government headquarters (as a proxy for ing inpatient deliveries, immunizations, and out- the degree of urbanization, connectivity, or patient consultation. While an appealing interpre- population density), the total number of fa- tation of this association may be that greater com- cilities within a 10km walking radius (to con- munity participation makes facility staff more trol for the availability of other health care responsive to the health needs of the community choices, and as a proxy for population den- they serve, there are alternative interpretations, sity), and population and internally generated and the analysis undertaken here is too limited to revenues of the local government within draw strong conclusions about the causal impact whose jurisdiction the facility resides. This last of community participation on service delivery. control variable is the best available proxy For example, we are unable to properly control Emerging Issues 67 Table III.1.3 Impact of Community Participation on Availability of Essential Drugs (1) (2) (3) 1=Chloroquine 1=Paracetamol 1=Antibiotics Not Available Not Available Not Available Community Participation Indicator Variable ­0.02 ­0.06 ­0.07 (0.05) (0.06) (0.08) Facility Type Indicator Variable 0.19** 0.13 0.19* (0.09) (0.09) (0.12) Distance from LGA Headquarters 0.002** 0.001 0.000 (0.001) (0.001) (0.002) Number of facilities in the neighborhood ­0.01*** ­0.01*** 0.001 (0.003) (0.003) (0.006) LGA population, 1999 ­0.000 ­0.000 0.000 (0.000) (0.000) (0.000) LGA Internally Generated Revenues Per Capita, 2000 ­0.001 ­0.001 ­0.003* (0.001) (0.001) (0.002) Constant 0.11 0.21** 0.24* (0.08) (0.09) (0.13) No. of Observations 144 144 144 R-sq 0.11 0.07 0.05 OLS regressions with robust standard errors (in parentheses). * Significant at 10% level; ** Significant at 5% level; *** Significant at 1% level for community-level income and education, nor worrisome, and a causal interpretation (despite for community-level demand for health services, the caveats indicated above) would suggest that that would affect both the extent of community with more decentralized management and participation and outcomes measured at the level monitoring of facilities by the immediate com- of the health facility. Richer or more educated munities they service, some facility activities communities, for instance, are both more likely with beneficial spillovers outside the commu- to participate in the management of public health nity are likely to be under-provided. This is a facilities and have greater demand for health ser- classic "public-goods" problem for which vices, leading to higher productivity of health staff greater control and supervision needs to be as measured in our regression analysis. Hence, exercised by a higher tier of authority, such as the analysis does not inform us about the impact the local governments. of specific policy interventions that promote com- munity participation. An alternate and tailor- made research design is needed for such an im- III.2. Non-payment of staff salaries pact evaluation. The other significant association, namely the Despite substantial budgetary allocations to staff negative correlation with record-keeping at the salaries in Kogi, the survey of health facility staff facility level for public health surveillance is revealed that non-payment of salaries is a 68 Decentralized Delivery of Primary Health Services in Nigeria serious concern in the state--42% of staff respon- the estimated annual budget allocation for sala- dents in Kogi report not receiving any salary for ries per facility. This statistic by itself may sug- 6 months or more in the past year at the time of gest that the problem of non-payment of salaries the survey. Figure III.2.1 shows the distribution arises due to inadequate budgetary allocations.6 of staff against the months in the past year for Yet, a comparison, LGA by LGA, of the num- which their salary has not been paid for each ber of months staff on average reported salaries state--the distribution of Kogi staff, in contrast not being paid, and the ratio of estimated ac- to that of Lagos, clearly shows that non-payment tual costs to budgeted allocations reveals that of salaries is a pervasive problem in Kogi state. there are several LGAs where salaries were not Rough calculations were performed to esti- paid even when estimated budget allocations are mate whether this problem of non-payment of sufficient to cover estimated actual costs. Con- staff salaries in Kogi could be due to inadequate versely, there are LGAs where the estimated ac- budgetary allocations towards salaries in the tual costs are more than twice the estimated LGA health budget. We use the sample average budgeted allocations, and yet staff report only a of staff monthly salary from the salary reported couple of months of non-payment, which could by staff in each LGA as an estimate of average be due to administrative delays alone. monthly salary per staff in an LGA, and the Table III.2.1 reports the average number of sample average of total number of staff in a fa- months in each Kogi LGA that staff reported cility (as reported by the facility head in the fa- not having salaries paid in the past year before cilities surveyed in each LGA) as an estimate of the survey, against the ratio of our estimate of average number of staff per facility in an LGA. the average salary cost per facility in the LGA The product of these two sample averages mul- to our estimate of the average budgeted alloca- tiplied by 12 thus gives an estimate of the aver- tion towards salary cost per facility in the LGA. age annual salary cost per facility in each LGA. If the problem underlying non-payment of staff The average across the 15 LGAs in Kogi of this salaries is inadequate LGA budget allocations estimated salary cost per facility is 1.4 million Naira, ranging from a minimum of 0.3 million Naira to a maximum of 7.5 million Naira. We then estimate each LGA's average annual budget allocation towards staff salaries per facil- 6 Although, even this interpretation begs the ques- ity. From the LGA respondents data, we divided tion of why LGAs do not allocate more resources actual budgetary allocation to staff salaries for towards committed expenditures such as staff sala- the year 2000 by the number of facilities that the ries, or alternately restructure personnel hires in LGA reported as owning within its jurisdiction, line with available resources. Budget allocations to get an estimate of the average LGA budget towards health staff salaries in 2000 constituted allocation to staff salaries for a typical health fa- 20% on average of total LGA revenues, and 78% of total health expenditures. cility in the LGA. Data on budgetary allocation It may also be reasonable to expect that the towards salaries of health personnel was missing estimate for average actual salary cost per facility for one LGA in Kogi--Mopa Muro. The aver- is an overestimate of actual costs since the aver- age across the 14 LGAs, for which data is avail- age monthly salary is reported by staff of higher able, of the estimated budgetary allocation for grades that were selected for the interview, and salaries per facility is 1.2 million Naira, ranging then applied to all the staff in the facility. In fact, similar calculations for Lagos state, with no sig- from a minimum of 0.2 million Naira to a maxi- nificant problem of non-payment of salaries, show mum of 8.1 million Naira. the estimate of actual salary costs to be 1.3 times, On average across Kogi LGAs, the estimated on average, the estimate of budget allocations per actual annual salary cost per facility is 1.6 times facility. Emerging Issues 69 Figure III.2.1 Non-payment of staff salaries in Kogi and Lagos 20% KOGI 18% 80% LAGOS 16% KOGI TOTAL 240 14% LAGOS TOTAL 495 12% respondents 10% 8% staff of 6% % 4% 2% 0% 0 1 2 3 4 5 6 7 8 9 10 11 12 Months Salary Not Paid Table III.2.1 Non-payment of staff salaries in selected LGAs in Kogi (2) (1) Ratio of Estimated Average Salary Months in the year before the Cost per facility (2002) to survey that salary has not Estimated Average Budget been paid (2001­02) Allocation per facility (2000) Adavi LGA 3 1.9 Bassa LGA 9 0.8 Dekina LGA 10 2.0 Ibaji LGA 2 2.8 Igalamela/Odolu LGA 3 1.9 Idah LGA 5 0.7 Ijumu LGA 6 1.6 Kabba Bunu LGA 6 2.2 Kogi LGA 6 3.1 Lokoja LGA 6 0.5 Mopa Muro LGA 3 N/A Ogori Magongo LGA 1 0.9 Olamaboro LGA 8 1.2 Omala LGA 4 3.3 Yagaba West 4 1.1 Correlation between columns (1) and (2): ­0.15 70 Decentralized Delivery of Primary Health Services in Nigeria for this purpose, then we should see a strong ment may have in terms of closing-down of positive correlation between these variables-- health facilities. Field-work for the survey in the higher is the estimate of actual salary costs fact revealed that several facilities in Kogi had as compared to budgeted allocation the greater been closed down for months due to non-pay- should be the number of months of non-pay- ment of staff salaries (Adeniyi, Oladepo, and ment. In fact, the correlation between these two Soyibo, 2003). series is negative. Regression analysis of aver- This problem of non-payment of salaries of age months of non-payment of salary in an LGA health staff by local governments is reminiscent on the reported budget allocation for salaries of of a similar problem of non-payment of teacher health workers (or on total LGA revenues) also salaries in primary schools in the 1990s, when reveals no significant correlation.7 primary education was decentralized to local Furthermore, there are striking examples of governments. Following nation-wide agitations LGAs such as Bassa, Idah, Lokoja, and by teacher unions a policy of deducting primary Olamaboro where salaries were not paid for school teacher salaries from the revenue share more than 5 months in the year before the sur- of local governments in the Federation Account vey, yet estimates of salary costs in a typical fa- was adopted, with the salaries being directly cility are below or almost equal to what the LGA passed-on to the teachers. However, this issue of reports as budgetary allocations towards staff "deductions-at-source" has substantially under- salaries in a typical facility. These estimates sug- mined accountability for the delivery of primary gest that the problem of non-payment of staff education by local governments (see discussion salaries in Kogi may not be lack of budgetary in IMF, 2001), and may not be a solution to pur- allocations for this purpose but rather leakage sue for staff salaries in the health sector. in resource flows at the LGA level. The evidence presented here, correlating the Table III.2.2 reports regressions of the impact non-payment of salaries with budgeted of average number of months of salary non-pay- allocations for salaries, suggests that the prob- ment in a facility. The greater the average num- lem is one of general accountability of local gov- ber of months for which staff salaries are not ernments in managing substantial resource paid in a facility, the greater are the number of transfers from taxpayers outside their jurisdic- home visits by facility staff, the lower the likeli- tion. And therefore larger solutions that tackle hood of the facility being clean, and the greater this fundamental problem of accountability the probability that essential drugs (chloroquine, should be explored rather than "top-down" ini- paracetamol, and antibiotics) are privately pro- tiatives such as specific purpose transfers, espe- vided by facility staff rather than being facility cially through deductions at source. One idea owned. These results suggest that non-payment would be to widely publicize information about of staff salaries may lead staff to provide pri- the resource envelopes of local governments and vate health services, in exchange for remunera- their constitutional responsibilities in order to tion from their patients. The available data and evidence does not allow us to distinguish whether the essential drugs are provided by staff out of their personal funds or if they are expropriated from facility stocks for private sale. It should be indicated here that this impact 7 The point estimates on budget allocations in these of non-payment of staff salaries is being esti- regressions in fact have a positive sign, that is, mated for facilities that are still functioning and suggesting that greater budget allocations are as- therefore responding to the survey questions, sociated with more months of non-payment of and does not capture whatever impact non-pay- salaries. Emerging Issues 71 make the local electorate more aware of the ca- transferring budgeted resources to schools. A fol- pacities and duties of their local representatives. low-up survey in Uganda showed that this in- A similar information-dissemination strategy, formation dissemination had a substantial through public radio and other media, was impact in preventing leakage of public funds adopted in Uganda after survey evidence away from purposes intended in public budgets revealed that district governments were not (Reinikka and Svensson, 2001). Table III.2.2 Impact of Non-Payment of Staff Salaries on Facility Performance (3) (4) (5) (1) (2) 1=Chloroquine 1=Paracetamol 1=Antibiotics Home 1=Facility is Privately is Privately is Privately Visits is Clean Owned Owned Owned No. of months in past 2.27* ­0.02** 0.02** 0.03*** 0.03*** year salary not paid (1.31) (0.01) (0.01) (0.01) (0.01) Facility Type Indicator ­11.24* 0.19*** ­0.25** ­0.17* ­0.11 Variable (6.58) (0.06)* (0.09) (0.10) (0.10) Distance from LGA 0.04 0.000 0.000 0.000 0.001 Headquarters (0.13) (0.001) (0.002) (0.002) (0.002) Number of facilities in ­0.14 0.000 ­0.001 0.003 ­0.001 the neighborhood (0.59) (0.006) (0.005) (0.005) (0.005) LGA population, 1999 0.000 0.000 0.000 ­0.000 ­0.000 (0.000) (0.000) (0.000) (0.000) (0.000) LGA Internally Generated ­0.10 0.003 ­0.001 ­0.002 ­0.002 Revenues Per Capita, 2000 (0.18) (0.002) (0.002) (0.002) (0.002) Constant 15.50 0.74*** 0.27* 0.24 0.24 (21.69) (0.15) (0.16) (0.15) (0.16) No. of Observations 109 141 141 141 141 R-sq 0.08 0.11 0.10 0.12 0.11 OLS regressions with robust standard errors (in parentheses). * Significant at 10% level; ** Significant at 5% level; *** Significant at 1% level 72 Decentralized Delivery of Primary Health Services in Nigeria CHAPTER 4 Main Conclusions and Policy Lessons A s discussed in the introduction of this re- the survey--public resources, in fact, do not port, the motivation for this study was appear to be reaching their intended destina- primarily to explore broad issues of gov- tions. There is evidence of large scale leakage in ernance and public expenditure manage- public resources in Kogi, away from original ment that underpin the translation of budgeted al- budget allocations. Although staff salaries ac- locations to actual services for citizens. Hence, the count for 78% of health expenditures and 20% main conclusions and policy lessons emerging from of total LGA revenues, on average, the survey this study are cross-cutting and inform our under- of facility staff in Kogi revealed that 42% of standingofgeneralinstitutionsofaccountabilitythat them had not been paid their salaries for more determine the extent to which public expenditures than 6 months in the past year. Using the survey are effective in actually delivering services. data, we estimated and compared actual staff Inadditiontothegenerallessons,thesurveymeth- costs per facility in each LGA with what the LGA odology employed here specifically for the primary reported as budget allocations towards staff sala- health sector has resulted in the generation of rich ries per facility within its jurisdiction, and found evidenceonpublichealthservicedeliveryatthefront- that even when budget allocations were suffi- line, in terms of facility infrastructure, staffing pat- cient to cover estimated actual costs, the staff terns,availabilityofessentialsuppliesandequipment, survey showed non-payment of salaries for sev- andservicesprovided.Belowwesummarizethemain eral months in the year before the survey. There conclusions of the study, and policy lessons, in turn is, in fact, no significant correlation between for each of these areas--the general area of gover- local government revenues and resources bud- nance and public expenditure management, and the geted towards staff salaries with the non-pay- sector-specific area of primary health services. ment of salaries. Hence, the non-payment of salaries cannot be explained by lack of resources available to local governments. Governance and Accountability--role The analysis also showed that the greater is of local governments and communities the extent of non-payment of salaries, the higher is the likelihood that facility staff in fact behave A striking feature of public delivery of primary as private providers--with more services pro- health services in Nigeria was revealed through vided outside the facility through home visits, 73 74 Decentralized Delivery of Primary Health Services in Nigeria and with essential drugs being privately pro- Nigeria would benefit from the institutional ar- vided, either funded by staff own resources or rangement of decentralization to local govern- expropriated from facility stocks. ments.9 Yet, the survey evidence provided here This evidence suggests that there is a general reveals enormous problems of accountability problem of accountability at the local govern- of local governments. In and of itself, this analy- ment level in the use of public resources that are sis does not suggest that the counterfactual transferred from higher tiers of government and would be true--that is, more centralized deliv- about which, therefore, local citizens may not ery in the hands of the state or federal govern- be well informed since they are not the tax-pay- ment would be better. The analysis undertaken ers. A similar problem of non-payment of pri- here cannot address this question because we mary school teacher salaries by local govern- cannot compare outcomes across more or less ments in the 1990s was solved by the federal decentralized systems. government by deducting the total salary costs But the overall policy lesson that the analysis from the local government share in Federation does suggest is that of strengthening local gov- Account revenues and transferring these directly ernment accountability, and we propose one to the teachers through state government au- major channel for this purpose--providing citi- thorities. However, this "solution" may have zens with greater information about the resources undermined overall local government responsi- and responsibilities of their local representatives, bility for primary education, without address- so they are empowered to hold them account- ing the fundamental problem of accountability able for the delivery of basic services. There is in the use of all public resources (see discussion very little systematic research evidence on in IMF, 2001), and may not be a solution to whether information dissemination truly has an pursue for staff salaries in the health sector. It impact, or what forms of dissemination are likely may therefore be better to address this emerg- to have greater impact; yet, theoretically, it seems ing problem in the health sector as an overall to be a reasonable way to proceed. The condi- problem of local government accountability, and tions under which local governments, or any larger solutions that tackle this fundamental elected government for that matter, will have the problem of accountability should be explored right incentives to improve the delivery of basic rather than "top-down" initiatives such as spe- services have been explored in a large political cific purpose transfers, especially through economy literature, and one of the "solutions" "deductions at source". to these political constraints suggested by the lit- Nigeria is one of the few countries in the de- erature is greater information dissemination veloping world that has constitutionally decen- about the roles and responsibilities of govern- tralized revenue allocation and expenditure re- ment, and the outcomes of public resource allo- sponsibilities to locally elected governments. In cation (see Keefer and Khemani, 2003, for a re- many developing countries, even when locally view of the literature and suggested solutions). elected governments exist there are no regular Designing a rigorous impact evaluation compo- nor systematic channels of resource transfer, nent to policy experiments with information dis- and no well-defined responsibilities for service semination would therefore be valuable to en- delivery.8 In comparison, local governments in hance our understanding of what works and Nigeria receive large and substantial funds to what doesn't, and how best to design institutional perform their functions, given their share in the interventions to improve public accountability. federation's oil revenues, and are generally iden- Another channel of strengthening LGA incen- tified as the responsible government agency for tives that may be explored is that of providing primary social services. This might lead to an direct incentives to local governments to improve expectation that public delivery of services in performance through additional resource Main Conclusions and Policy Lessons 75 transfers (additional to their constitutionally de- and have greater demand for health services, lead- termined share in the Federation Account) con- ing to higher productivity of health staff as mea- ditional on actual improvements in service de- sured in our regression analysis. Hence, the analy- livery. However, the impact of such conditional sis does not inform us about the impact of spe- grants will also depend upon overall account- cific policy interventions that promote commu- ability, as it may not be feasible to provide fi- nity participation. An alternate and tailor-made nancial incentives that are large enough for lo- research design is needed for such an impact evalu- cal governments to change the fundamental way ation, particularly controlling for community-level in which they work, unless there is direct pres- income and education. sure from an empowered citizenry. The litera- There is also a significant negative correlation ture on conditional or matching grants from of community participation in facilities with other parts of the world usually takes as given record-keeping at the facility level for public that local governments are accountable to local health surveillance. A causal interpretation of this citizens, and the incentive component of the would suggest that with more decentralized man- grants is largely intended to make local com- agement and monitoring of facilities by the im- munities internalize potential spillover effects of mediate communities they service, some facility local investments for the national good. activities with beneficial spillovers outside the We found some evidence that active commu- community are likely to be under-provided. This nity participation in health service delivery may underscores the importance of local government make staff more responsive to community health responsibility for public health management, and needs and increase overall productivity of fa- proper coordination and sharing of responsibili- cilities. Communities were particularly active in ties with community based organizations. participating in health service delivery in the state of Kogi, whose population largely lives in rural areas, and depends heavily on public institutions Issues in Primary Health of service delivery. The most striking result is Service Delivery that community participation in Kogi facilities is significantly associated with greater produc- Although the majority of public health facilities tivity per staff in providing inpatient deliveries, were observed to be clean and functioning and immunizations, and outpatient consultation. providing a range of health services, there is some While an appealing interpretation of this asso- suggestion of poor quality of services for some ciation may be that greater community participa- of the conditions that are reported as the main tion makes facility staff more responsive to the causes of mortality and morbidity among chil- health needs of the community they serve, there dren, namely diarrhea, and vaccine preventable are alternative interpretations, and the analysis diseases. Simple treatments for easy to diagnose undertaken here is too limited to draw strong con- conditions such as childhood diarrhea, that is clusions about the causal impact of community ORS sachets, were not available in 70% of the participation on service delivery. For example, we facilities surveyed. The analysis reported here are unable to properly control for community- therefore suggests greater attention and empha- level income and education, nor for community- sis on policies for preventive health services and level demand for health services, that would af- simple treatments, than just for drugs-based cura- fect both the extent of community participation tive care. Strengthening of policies on preventive and outcomes measured at the level of the health health care is also urgent in light of evidence that facility. Richer or more educated communities, public health surveillance may be particularly for instance, may be both more likely to partici- poor in rural states--in Kogi, only 28% of facili- pate in the management of public health facilities ties were able to show records of tracer and 76 Decentralized Delivery of Primary Health Services in Nigeria immediately notifiable diseases to the survey in- level, and civil service grades are primarily de- terviewer, compared to 80% of facilities in Lagos termined by seniority. The average number of that produced these records. years that staff have worked in their current fa- Lack of cold storage equipment meant that cility is low, which means that there is a lot vaccines were not available in a majority of fa- churning of staff in the system, and that one of cilities (80%) in Kogi, and in more than 30% of the objectives of civil service incentive schemes facilities in Lagos, despite facilities in that state ­ collaboration and stability ­ is not being having greater access to cold storage. The effi- achieved. Facility staff appear to have a lot of cacy of national immunization campaigns on discretion in charging fees from patients, as user selected days, as a solution to the infrastructure fee policies are not established nor transparent, constraints for storing vaccines, is suspect given thus exposing communities to the risk of over- the low numbers of immunizations provided by charging by staff for a supposedly subsidized LGAs in the last three months before the sur- public service. vey, which included one of these campaigns. A These findings suggest that national, state, or more detailed study on the delivery of immuni- local government policies with regard to user zation through the national campaigns, and its fees should be made more transparent, with per- interaction with routine programs, therefore haps facilities being required to visibly post the seems to be warranted. information in their buildings. Strengthening of The survey evidence showed that the incen- existing institutions of community participation, tives of frontline service providers, that is the particularly in the management of health staff health staff, are typically blunt, in that there is with a view to providing greater incentives for no discernable local discretion to reward good quality service provision, is a potential policy performance. Most of the variation in salaries instrument to improve service delivery outcomes across staff is explained by civil service grade for any level of resource allocation. Bibliography Adeniyi, J., O. Oladepo, and A. Soyibo. 2003. Report 2004: Making Services Work for Poor "Survey of Primary Health Care Service De- People, Development Research Group, World livery in Lagos and Kogi: A Field Report." Bank, Washington, D.C. Processed. Mimeo, African Regional Health Education Lindelow, M. and A. Wagstaff. 2002. "Health Center, University of Ibadan. Facility Surveys: An Introduction." Develop- Filmer, D., J. Hammer, and L. Pritchett. 2000. ment Research Group, World Bank, Wash- "Weak Links in the Chain: A Diagnosis of ington, D.C. Processed. Health Policy in Poor Countries." World Musgrove, P. 1996. "Public and Private Roles Bank Research Observer, 15:2, pp. 199. in Health: Theory and Financing Patterns." Filmer, D. and L. Pritchett. 1999. "Child Mor- World Bank Discussion Paper 339. tality and Public Spending on Health: How Rajkumar, A. and V. Swaroop. 2002. "Public much Does Money Matter?" Social Science Spending and Outcomes: Does Governance and Medicine, 49. Matter?" Policy Research Working Paper No. International Monetary Fund (IMF). 2001. 2840, World Bank, Washington, D.C. Nigeria: Options for Reforming Intergovern- Reinikka, R. and J. Svensson. 2001. "Informa- mental Fiscal Relations, Washington, D.C. tion and Voice in Public Spending." Mimeo, Federal Republic of Nigeria, Federal Ministry Development Research Group, The World of Health. 1996. National Health Policy. Bank. http://econ.worldbank.org/files/ Abuja, Nigeria 15157_PSproject.informationandvoice.reinikka. Keefer, P. and S. Khemani. 2003. "The Political may1.pdf Economy of Public Expenditures." Back- World Bank. 2001. Nigeria: Public Expenditure ground paper for the World Development Review 77 78 Decentralized Delivery of Primary Health Services in Nigeria